Journal of Industrial and Engineering Chemistry 17 (2011) 377–381

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry

journal homepage: www.elsevier.com/locate/jiec

Short communication

Acetalisation of glycerol with acetone over zirconia and promoted zirconia

catalysts under mild reaction conditions
Padigapati S. Reddy, Putla Sudarsanam, Baithy Mallesham, Gangadhara Raju, Benjaram M. Reddy *
Inorganic and Physical Chemistry Division, Indian Institute of Chemical Technology, Uppal Road, Hyderabad 500 607, India

A R T I C L E I N F O A B S T R A C T

Article history: Acetalisation of biomass derived glycerol was performed with acetone over zirconia and promoted
Received 6 October 2010 zirconia catalysts to synthesize solketal. The catalytic experiments were carried out at room
Accepted 9 November 2010 temperature. It was observed that the promoted zirconia catalysts exhibit promising catalytic activity.
Available online 14 May 2011
The activity increased in the order of ZrO2 < WOx/ZrO2 < MoOx/ZrO2 < SO42 /ZrO2. All catalysts
exhibited an excellent selectivity of 97% towards solketal. The effect of different parameters, such as
Keywords: molar ratio of acetone to glycerol, catalyst wt.% and time-on-stream was studied over zirconia and
Biomass
promoted zirconia catalysts. The synthesized catalysts were characterized by XRD, BET surface area,
Glycerol
Acetalisation
ammonia-TPD and FT-Raman techniques. Characterization results revealed that the addition of
Solketal promoters strongly influences the surface acidity of ZrO2 and enhances the tetragonal zirconia phase.
Promoted zirconia catalysts ß 2011 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights
reserved.

1. Introduction glycerol by various catalytic processes [6,7] involving hydrogeno-

lysis to obtain 1,2-propanediol and 1,3-propanediol; oxidation to
Presently, global research is focused on the development of obtain dihydroxyacetone, glyceraldehyde, glyceric acid, glycolic
sustainable and renewable resources, driven by declining of acid and hydroxypyruvic acid; dehydration to yield acrolein or 3-
petroleum reserves and growing concern about atmospheric hydroxypropionaldehyde; fermentation towards 1,3-propanediol;
greenhouse gas concentrations [1]. Biomass represents an polymerization to obtain polyglycerols and polyglycerol esters [8–
abundant carbon-neutral renewable resource for the production 12]. Glycerol is also a good platform for other chemicals, especially
of bioenergy and biomaterials, as well as source for a variety of oxygenated compounds such as acetals and ketals. Moreover, at
chemical intermediates. The enhanced use of biomass would present time the catalytic conversion of glycerol into acetals and
contribute to several societal needs [2]. Biofuels, which are ketals has been receiving great attention [13]. Acetals and ketals
produced from biomass are attractive alternatives to the fossil are produced from the acetalisation of glycerol with an aldehyde or
fuels. Among variety of liquid biofuels like bioethanol, biodiesel, a ketone, respectively, in the presence of an acid catalyst [14].
biomethanol and Fischer–Tropsch diesel fuel, one such bio-based Particularly, these compounds have enormous potential applica-
fuel that is being widely commercialized is biodiesel [3]. Presently, tions as fuel additives [15]. Moreover, glycerol acetals are used as
global production of biodiesel has been increased significantly. surfactants, flavours, disinfectants and so on [16]. The acetalisation
Glycerol is a value-added byproduct produced (approximately of glycerol with acetone gives branched oxygenated compounds
10 wt.%) from transesterification reaction performed during namely (2,2-dimethyl-[1,3]dioxane-4-yl)-methanol (solketal) and
biodiesel manufacturing process [4]. So, with the growth of the 2,2-dimethyl-[1,3]dioxane-5-ol. Solketal is an excellent compo-
biodiesel industry, there will be more glycerol than the current nent for the formulation of gasoline, diesel and biodiesel fuels [17].
market can easily absorb. Therefore, the effective utilization of This product can improve the cold properties, reduce viscosity and
glycerol enlarges the economy of the whole biodiesel production also meet the established requirements for flash point and
process and the conversion of glycerol to value added chemicals oxidation stability of biodiesel [13,18].
could be in agreement with the perception of green chemistry in Generally, acetalisation is an acid catalyzed reaction and
terms of the exploitation of renewable resources [5]. In recent conventionally carried out using mineral acids as catalysts. In
years many studies have been focused on the transformation of view of environmental and economical requirements, the devel-
opment of effective and inexpensive solid acid catalysts is
essential. There are some solid acid catalysts that are reported
* Corresponding author. Tel.: +91 40 27191714; fax: +91 40 27160921.
in the literature for the acetalisation of glycerol which include
E-mail addresses: [email protected], [email protected] (B.M. Reddy). amberlyst, heteropolyacids, zeolites and others [14,19,20]. How-

1226-086X/$ – see front matter ß 2011 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.jiec.2011.05.008
378 P.S. Reddy et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 377–381

ever, most of the reports used a temperature of 343 K or higher for 2.3. Activity measurements
the production of solketal from the acetalisation of glycerol with
acetone. Nevertheless, solid acid catalysts such as ion exchange The catalytic activity for the acetalisation of glycerol was
resigns (amberlyst) and heteropolyacids exhibit drawbacks like carried at room temperature and atmospheric pressure. In a typical
poor thermal stability, poor regeneration ability and low specific experiment, 1 g of glycerol and 0.5–3.0 mL of acetone were taken
surface area [21–23]. Supported heteropolyacids having high in a 25 mL round bottom flask with 0.05 g of catalyst. Catalysts
surface area and better thermal stability are hindered by limited were pre-activated at 423 K for 2 h before catalytic runs. Samples
accessibility and incompetence [24,25]. In comparison to the were taken periodically and analysed by GC equipped with BP-20
aforesaid catalysts, promoted metal oxides offer several advan- (Wax) capillary column and a flame ionization detector.
tages, since they are stable, regenerable and active. In this work,
the synthesis of solketal from acetalisation of glycerol with acetone 3. Results and discussion
was carried out, for the first time, at mild reaction conditions, over
zirconia and promoted zirconia solid acid catalysts. These catalysts 3.1. Catalyst characterization
were characterized by BET surface area, X-ray diffraction (XRD),
Raman spectroscopy and NH3-temperature programmed desorp- The XRD patterns of zirconia and promoted zirconia catalysts
tion (TPD) techniques. The influence of various reaction param- are presented in Fig. 1. It can be observed that all catalysts
eters such as molar ratio of acetone to glycerol, catalyst loading and exhibited the characteristic peaks corresponding to a mixture of
time-on-stream over these catalysts was also studied. monoclinic and tetragonal zirconia phases [28,29]. In general,
incorporation of the promoters stabilizes the tetragonal phase of
zirconia [30]. Further, from XRD results it can be pointed out that
2. Experimental methods zirconia catalyst is in a poorly crystalline form with more
monoclinic phase. Whereas in the case of promoted zirconia
2.1. Catalyst preparation catalysts (SZ, MZ and WZ), the tetragonal phase dominates over the
monoclinic phase. Also, observed from the XRD lines that the
ZrO2 was obtained by co-precipitation technique in accordance molybdate shows relatively strong influence than tungstate on the
with the procedure described elsewhere [26]. The ZrOCl2 aqueous monoclinic to tetragonal phase transformation of ZrO2. The values
solution was dropped into a flask with constant stirring and the pH of specific surface area and amount of NH3 desorbed are presented
value was maintained at about 8 with dilute NH4OH. The obtained in Table 1. As can be noted from the Table 1, all the promoted
precipitate was carefully washed with distilled water until free catalysts show a high surface area than the unpromoted zirconia
from chloride ions and dried at 393 K for 48 h. The molybdated and catalyst. Among all the promoted zirconia catalysts, sulfated
tungstated ZrO2 were synthesized by adopting a wet impregnation zirconia catalyst showed high surface area due to the inhibition of
technique as per the previously reported procedure [24]. In a crystal growth by sulfate ions [31]. From the ammonia-TPD results
typical procedure for the preparation of molybdate and tungstate (Table 1) it is clear that the total number of acidic sites in the case of
promoted catalysts (6 mol%), the desired quantity of ammonium sulfate promoted catalyst is higher than that of molybdate and
heptamolybdate and ammonium metatungstate was first dis- tungstate promoted catalysts. It appears from the NH3-TPD results
solved in excess water separately. To this solution the powdered that sulfate and molybdate promoters exhibit a strong influence on
zirconium hydroxide was added and excess water was evaporated the surface acidity of the zirconia.
on a water-bath with vigorous stirring, oven dried and calcined in
air at 923 K for 5 h. The sulfated ZrO2 was prepared as per the [(Fig._1)TD$IG]
earlier reported literature [27]. In a typical procedure, the finely
powdered zirconium hydroxide was immersed in 0.5 M sulfuric
acid solution. The excess water was evaporated on a water-bath
with vigorous stirring, oven dried and calcined in air at 923 K for
5 h. The investigated ZrO2, WOx/ZrO2, MoOx/ZrO2 and SO42 /ZrO2
catalysts are referred to as Z, WZ, MZ and SZ, respectively.

2.2. Catalyst characterization

Crystalline phase identification was performed employing a

Siemens D-5000 diffractometer using Cu Ka radiation source and
scintillation counter detector. The XRD patterns were collected in
the 2u range of 20–708 at room temperature. The XRD phases
present in the samples were identified with the help of JCPDS data
files. The specific surface area of catalysts was measured by BET
method (Gemini 2360 instrument) at liquid nitrogen temperature,
using nitrogen as the adsorbate by taking 0.162 nm2 as the area of
cross section of N2 molecule. Raman spectra were recorded at
ambient temperature on a DILOR XY spectrometer equipped with a
CCD detector. The spectra were recorded in the range of 100–
1300 cm 1 and at a spectral resolution of 2 cm 1 using the
514.5 nm exciting line from an argon ion laser (Spectra Physics,
USA). The TPD measurements were carried on an AutoChem 2910
instrument (Micromeritics, USA). A thermal conductivity detector
was used for continuous monitoring of desorbed ammonia. Prior to
TPD studies, the samples were pretreated at 473 K for 1 h in a flow Fig. 1. X-ray diffraction patterns of ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ZrO2 (MZ) and
of ultra pure helium gas (40 mL min 1). SO42 /ZrO2 (SZ) catalysts. (u) t-Zirconia phase; (*) m-zirconia phase.
 P.S. Reddy et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 377–381 379

Table 1 3.2. Glycerol acetalisation

BET surface area and total acidity of ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ZrO2 (MZ) and
SO42 /ZrO2 (SZ) catalysts.
Acetalisation of glycerol with acetone has been represented in
S. no. Catalyst BET SA (m2/g) NH3 desorbed (mmol/g) Scheme 1. This reaction mainly yields the five-membered ring,
1 Z 42 0.21 ketal (solketal) and also forms the six-membered ring ketal, whose
2 SZ 100 0.69 relative formation depends on the acetalisation position within the
3 MZ 94 0.47
4 WZ 60 0.39
glycerol molecule [17]. While, glycerol acetalisation with acetone
favors the formation of the five-membered ring transition state,
[(Fig._2)TD$IG] which leads to the production of solketal [14,20]. In the present
study, all catalysts which are employed in the glycerol acetalisa-
tion reaction showed a maximum selectivity of 97% towards
solketal. The acetalisation of glycerol over Z, WZ, MZ and SZ
catalysts is illustrated in Fig. 3. The catalytic experiments were
carried out at room temperature, 90 min of reaction time and 6:1
molar ratio of acetone to glycerol. The highest catalytic activity was
obtained over the SZ catalyst. The glycerol conversion over Z, WZ,
SZ MZ and SZ catalysts is 10, 80, 88 and 98%, respectively. The
Intensity (a.u.)

highest catalytic activity of SZ catalyst can be explained as due to

its high surface area and more number of acidic sites (Table 1). It
can be noted that all promoted zirconia catalysts show 80% of
MZ glycerol conversion, but the zirconia catalyst alone shows <10% of
glycerol conversion. This can be explained as due to more number
of acidic sites on the promoted zirconia catalysts (Table 1). Also
WZ
there is no glycerol conversion observed in the absence of catalyst
for the acetalisation of glycerol reaction. The catalytic activity
Z
increased in the following order: Z < WZ < MZ < SZ. Additionally,
the effect of different parameters such as molar ratio of glycerol to
acetone, time-on-stream and catalyst loading on glycerol acet-
alisation over the present catalysts was also studied.
200 400 600 800 1000 1200
-1
Raman shift (cm ) 3.2.1. Effect of molar ratio
The influence of molar ratio of acetone to glycerol on the
Fig. 2. Raman spectra of ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ZrO2 (MZ) and SO42 /ZrO2 glycerol conversion over Z, WZ, MZ and SZ catalysts is shown in
(SZ) catalysts. Fig. 4. The reaction was studied by varying the acetone to glycerol
mole ratios from 1:1 to 6:1, at room temperature, 90 min of
reaction time and a catalyst weight of 5 wt.%. It can be observed
Fig. 2 represents the Raman spectra of the Z, WZ, MZ and SZ that an increase in the glycerol conversion with an increase in the
catalysts. From these results it can be observed that all catalysts molar ratio of acetone to glycerol was observed for all catalysts. In
show Raman bands corresponding to a mixture of monoclinic (188, contrast, change in molar ratio of acetone to glycerol does not
221, 331, 380, 476 and 637 cm 1) and tetragonal (148, 290, 311, influence the selectivity to solketal. According to these results it
454 and 647 cm 1) phases. In the case of zirconia catalyst the can be stated that the promoted zirconia catalysts show better
Raman lines due to the tetragonal phase are less intense than the glycerol conversion even at low acetone to glycerol molar ratio.
lines due to the monoclinic phase. Whereas in the case of promoted
zirconia catalysts, lines due to tetragonal phase are more intense. 3.2.2. Effect of catalyst wt.%
Raman bands in the range of 1000–1150 cm 1 for the SZ catalyst The effect of catalyst wt.% on glycerol conversion over Z, WZ, MZ
are due to the presence of different surface sulfated groups [29,32]. and SZ catalysts at room temperature for the acetalisation of
Further, the MZ catalyst exhibits two additional bands at around glycerol is shown in Fig. 5. The catalytic experiments were studied
820 and 990 cm 1 corresponding to the formation of Zr (MoO4)2 by varying the catalyst wt.% from 1 to 5 wt.%, the molar ratio of
compound between zirconia and molybdenum [33]. From XRD and acetone to glycerol was 6:1 and the reaction time was 90 min.
Raman results, the absence of all characteristic bands due to From these results it can be observed that the promoted zirconia
crystalline MoO3 and WO3 suggests that the impregnated catalysts show good conversion of glycerol even at low catalyst
molybdenum and tungsten oxides strongly interacted with the wt.%. Also observed that the conversion of glycerol increases with
zirconia. the increase in catalyst wt.%. This is due to an increase in the total

[(Schem_1)TD$FIG]

OH O Acid catalyst
O O
O +
HO OH + O
-H2O OH
OH
Glycerol Acetone (2,2-dimethyl-[1,3] 2,2-dimethyl-
dioxan -4-yl)-methanol [1,3]dioxan-5-ol
(Solketal)

Scheme 1. Producton of solketal from an acid catalyzed acetalisation of glycerol with acetone.
[(Fig._3)TD$IG]
380 [(Fig._5)TD$IG]
P.S. Reddy et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 377–381

100
90
80

Conversion (%)
70
60
Z
50 WZ
40 MZ
SZ
30
20
10
0
1 2 3 4 5
Catalyst wt.% (w.r.t. glycerol)
Fig. 5. The influence of increasing catalyst wt.% on the conversion of glycerol in the
acetalisation with acetone using ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ZrO2 (MZ) and
SO42 /ZrO2 (SZ) catalysts. Reaction conditions: molar ratio of acetone to
glycerol = 6:1; reaction time = 90 min.
[(Fig._6)TD$IG]
100
Fig. 3. Conversion of glycerol in the acetalisation with acetone at room temperature
over ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ZrO2 (MZ) and SO42 /ZrO2 (SZ) catalysts.
90
Reaction conditions: molar ratio of acetone to glycerol = 6:1; reaction
time = 90 min; catalyst amount = 5 wt.% (w.r.t. glycerol).
80
70
Conversion (%)

number of acid sites available. The selectivity towards solketal was

60
not found to change much up on changing the catalyst wt.%. Z
50 WZ
3.2.3. Effect of reaction time MZ
40
Fig. 6 shows the time-on-stream studies over Z, WZ, MZ and SZ SZ
catalysts, at room temperature with a molar ratio of acetone to 30
glycerol of 6:1 and a catalyst weight of 5 wt.%. The experiments
were studied at different time periods up to 120 min. It is 20
interesting to note that the promoted zirconia catalysts show the
10
[(Fig._4)TD$IG]
glycerol conversion of >60% even after 10 min of reaction time. At
0
100 10 20 30 40 50 60 70 80 90 100 110 120

90
Time (min)

80 Fig. 6. Effect of reaction time on the conversion of glycerol in the acetalisation with
acetone using ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ZrO2 (MZ) and SO42 /ZrO2 (SZ)
Conversion (%)

70 catalysts. Reaction conditions: molar ratio of acetone to glycerol = 6:1; catalyst

amount = 5 wt.% (w.r.t. glycerol).
60
50 Z 90 min of reaction time the Z, WZ, MZ and SZ catalysts showed a
WZ glycerol conversion of 10, 80, 88 and 98%, respectively. After
40 MZ 90 min of reaction time the conversion of glycerol was constant for
SZ the all catalysts. Alternatively, the selectivity towards solketal did
30 not vary with reaction time.
20
4. Conclusions
10
The acetalisation of glycerol with acetone was carried out over
0
zirconia and promoted zirconia solid acid catalysts. It was observed
1 2 3 4 5 6
that the promoted zirconia catalysts exhibit better catalytic
Mole ratio
o (Acetone/G
Glycerol) activity. The sulfated zirconia catalyst exhibited the highest
glycerol conversion of 98%. The effect of various parameters
Fig. 4. The influence of increasing acetone to glycerol molar ratio on the conversion
of glycerol in the acetalisation with acetone using ZrO2 (Z), WOx/ZrO2 (WZ), MoOx/ such as molar ratio of acetone to glycerol, catalyst wt.% and time-
ZrO2 (MZ) and SO42 /ZrO2 (SZ) catalysts. Reaction conditions: reaction on-stream was studied over all the catalysts. All catalysts exhibited
time = 90 min; catalyst amount = 5 wt.% (w.r.t. glycerol). good selectivity towards solketal of 97%. Characterization results
 P.S. Reddy et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 377–381 381

revealed that the addition of promoters enhances the tetragonal [12] C.H. Zhou, J.N. Beltramini, Y. Fana, G.Q. Lu, Chem. Soc. Rev. 37 (2008) 527.
[13] J.A. Melero, G. Vicente, G. Morales, M. Paniagua, J. Bustamante, Fuel 89 (2010)
zirconia phase and the surface acidity. The ammonia-TPD results 2011.
indicate that the impregnated sulfate ions show a strong influence [14] C.X.A. da Silva, V.L.C. Goncalves, C.J.A. Mota, Green Chem. 11 (2009) 38.
on the acidity of ZrO2, which is followed by molybdate. [15] J. Delgado, EP Patent 1331260 (2002).
[16] P.H.R. Silva, V.L.C. Gonçalves, C.J.A. Mota, Bioresour. Technol. 101 (2010) 6225.
[17] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martı́n, Green Chem. 12 (2010)
Acknowledgment 899.
[18] E. Garcia, M. Laca, E. Pérez, A. Garrido, J. Peinado, Energy Fuel 22 (2008) 4274.
[19] J. Deutsch, A. Martin, H. Lieske, J. Catal. 245 (2007) 428.
P.S.R., P.S., B.M. and G.R. thank Council of Scientific and [20] P. Ferreiraa, I.M. Fonsecab, A.M. Ramosb, J. Vital, J.E. Castanheiroa, Appl. Catal. B:
Industrial Research, New Delhi, for the award of research fellow- Environ. 98 (2010) 94.
ships. [21] B.R. Jermy, A. Pandurangan, Appl. Catal. A: Gen. 288 (2005) 25.
[22] S. Mallick, K.M. Parida, Catal. Commun. 8 (2007) 889.
[23] S. Wang, J.A. Guin, Chem. Commun. 24 (2000) 2499.
References
[24] B.M. Reddy, M.K. Patil, B.T. Reddy, Catal. Lett. 125 (2008) 97.
[25] P.S. Reddy, P. Sudarsanam, G. Raju, B.M. Reddy, Catal. Commun. 11 (2010) 1224.
[1] S. Fernando, S. Adhikari, K. Kota, R. Bandi, Fuel 86 (2007) 2806. [26] B.M. Reddy, M.K. Patil, G.K. Reddy, B.T. Reddy, K.N. Rao, Appl. Catal. A: Gen. 332
[2] C. Perego, D. Bianchi, Chem. Eng. J. 161 (2010) 314. (2007) 183.
[3] A. Demirbas, Energy Source A 31 (2009) 1770. [27] B.M. Reddy, M.K. Patil, K.N. Rao, G.K. Reddy, J. Mol. Catal. A: Chem. 258 (2006) 302.
[4] J.A. Kenar, G. Knothe, J. Am. Oil Chem. Soc. 85 (2008) 365. [28] Y. Bai, D. He, S. Ge, H. Liu, J. Liu, W. Huang, Catal. Today 149 (2010) 111.
[5] A. Brandner, K. Lehnert, A. Bienholz, M. Lucas, P. Claus, Top. Catal. 52 (2009) 278. [29] L.D. Sasiambarrena, L.J. Mendez, M.A. Ocsachoque, A.S. Cánepa, R.D. Bravo, M.
[6] S.H. Song, D.R. Park, S.Y. Woo, W.S. Song, M.S. Kwon, I.K. Song, J. Ind. Eng. Chem. 16 Gloria González, Catal. Lett. 138 (2010) 180.
(2010) 662. [30] A. Khodakov, J. Yang, S. Su, E. Iglesia, A.T. Bell, J. Catal. 177 (1998) 343.
[7] H.J. Cho, H.M. Kwon, J. Tharun, D.W. Park, J. Ind. Eng. Chem. 16 (2010) 679. [31] B.M. Reddy, P.M. Sreekanth, V.R. Reddy, J. Mol. Catal. A: Chem. 225 (2005) 71.
[8] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem. 10 (2008) 13. [32] C. Morterra, G. Cerrato, S. Ardizzone, C.L. Bianchi, M. Signoretto, F. Pinna, Phys.
[9] Y. Gu, A. Azzouzi, Y. Pouilloux, F. Jerome, J. Barrault, Green Chem. 10 (2008) 164. Chem. Chem. Phys. 4 (2002) 3136.
[10] T. Miyazawa, Y. Kusunoki, K. Kunimori, K. Tomishige, J. Catal. 240 (2006) 213. [33] S. Xie, K. Chen, A.T. Bell, E. Iglesia, J. Phys. Chem. B 104 (2000) 10059.
[11] E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catal. Commun. 8 (2007) 1349.
 Chemical Engineering Journal 178 (2011) 291–296

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Valorization of glycerol into fuel additives over zeolites as catalysts

H. Serafim a , I.M. Fonseca b , A.M. Ramos b , J. Vital b , J.E. Castanheiro a,∗
a
Centro de Química de Évora, Departamento de Química, Universidade de Évora, 7000-671 Évora, Portugal
b
REQUIMTE, CQFB, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal

a r t i c l e i n f o a b s t r a c t

Article history: Acetalisation of glycerol with butanal has been investigated over a range of zeolites with different
Received 18 July 2011 pore structures and acidity (USY with different Si/Al ratio, BEA and ZSM-5). The products of glycerol
Received in revised form acetalisation were (Z + E)-(2-propyl-1,3-dioxolan-4-yl) methanol (five-member ring acetal) and (Z + E)-
27 September 2011
2-propyl-1,3-dioxan-5-ol (six-member ring acetal). The results showed an optimum in activity for Si/Al
Accepted 3 October 2011
ratio in USY zeolite that corresponds to 30. BEA zeolite presented the highest catalytic activity of all
zeolites while ZSM-5 zeolite presented the lowest activity. All catalysts exhibited high selectivity to
Keywords:
five-member ring acetal product (77–82%).
Glycerol
Butanal
The effects of reaction temperature, catalyst loading and molar ratio of glycerol to butanal on this
Acetalisation reaction, over BEA zeolite, have been studied. It was found that the glycerol conversion increased with
Zeolite the catalyst loading, with the temperature and with molar ratio of glycerol to butanal.
BEA zeolite was recycled four times with the same catalyst sample. It was observed a stabilization of
the catalytic activity (after the fourth use, the zeolite showed 93% of its initial activity).
© 2011 Elsevier B.V. All rights reserved.

1. Introduction Zeolites have found wide application in oil refining, petrochem-

istry and in the production of fine chemicals. Besides the general
Biodiesel is produced from renewable biological sources, such advantages associated with heterogeneous catalysts, zeolites can
as vegetable oils and animal fats, and consists of alkyl esters be synthesized with different crystal structure, framework Si/A1
derived from either the transesterification of triglycerides (TGs) ratio and proton exchange level, which allows design of catalytic
or the esterification of free fatty acids (FFAs) with short-chained properties such as pore size, hydrophobicity, strength of acid sites,
alcohols, in the presence of acid or basic catalysts [1]. In the and their distribution [8–11]. Zeolite catalysts are capable of shape-
biodiesel production, the glycerol is a by-product. For every 9 kg selective catalysis, as a result of their uniform micropore systems.
of biodiesel produced, about 1 kg of a crude glycerol is formed. The zeolite surface accessible to reactants comprises a large inter-
In order to develop new uses for glycerol different catalytic nal surface (the micropore system) and a smaller external surface
processes, including oxidation, reforming, hydrogenolysis, ether- (i.e. outside the micropores). Due to its special pore system and sur-
ification and esterification, acetalisation reactions, have been used face acidity, zeolites had attracted much attention in the synthesis
in the transformation of glycerol [2–5]. Another catalytic process of many fine chemicals [12,13].
to valorization of glycerol is the condensation of glycerol with Different acetalisation procedures were carried out over zeo-
butanal, which provides a branched oxygen-containing compound lites as heterogeneous catalysts [14–17]. The glycerol acetalisation
and could be used as an additive in the biodiesel formulation, products over zeolites were (Z + E)-(2-propyl-1,3-dioxolan-4-yl)
improving the cold properties and lowering the viscosity. methanol (five-member ring acetal) and (Z + E)-2-propyl-1,3-
Traditionally, the acetalisation of glycerol with butanal is carried dioxan-5-ol (six-member ring acetal) (Scheme 1). Acetal formation
out over mineral acids, as catalysts. However, the effluent disposal is a reversible reaction which proceeds by a two-step mechanism.
leads to environmental problems and economical inconveniences. The first step is the formation of an hemiacetal, followed by the
These problems can be overcome by the use of heterogeneous cat- removal of a water molecule. Acetal formation is strongly affected
alysts. However, only Amberlyst-15 resin [6] and MoO3 supported by electronic and steric factors, but it is generally accepted that the
on silica [7] were used as heterogeneous catalysts in acetalisation rate determining step of acetalisation is the formation of a cation
of glycerol with butanal. from the protonated hemiacetal. In order to compensate for the
low rate of hemiacetal formation, the reaction media must be suf-
ficiently acidic to promote effective protonation of any hemiacetal
∗ Corresponding author. Tel.: +351 266745311; fax: +351 266744971. formed, and sufficiently polar to allow stabilization of the cationic
E-mail address: [email protected] (J.E. Castanheiro). intermediate. According to the reaction mechanism, acetalisations

1385-8947/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cej.2011.10.004
292 H. Serafim et al. / Chemical Engineering Journal 178 (2011) 291–296

OH O Table 1
Physicochemical characterization of catalyst sample.

HO OH Sample Si/Ala SBET b (m2 /g) VT c (cm3 /g) Acid site conc.d
+ (mmolH+ /g)
H
USY1 5.2 679 0.39 2.68
USY2 12 695 0.49 1.28
USY3 30 777 0.52 0.53
USY4 80 799 0.55 0.20
H+ BEA 40 624 0.37 0.41
ZSM-5 250 495 0.27 0.12
Dowex – 7 1.48 × 10−2 4.9
a
ICP.
b
BET.
c
(p/p◦ ) = 0.95.
d
Calculated based on SiO2 /Al2 O3 .

The X-ray diffraction (XDR) patterns of zeolites were obtained

by using a Bruker powder diffractometer with built-in recorder,
using Cu K␣ radiation, nickel filter, 30 mA and 40 kV in the high
O O voltage source, and scanning angle between 5◦ and 55◦ of 2 at a
+ O O + H2O scanning rate of 1◦ /min.
The amount of sulfonic acid groups in Dowex resin was mea-
sured by a classic titration. The sample was equilibrated with 10 ml
of 0.1 mol/l NaOH solution for 24 h. Then, the amount of H+ was
OH calculated from the reduction in alkalinity determined by titration.
OH

Scheme 1. Acid-catalyzed acetalisation of glycerol with butanal over acid catalysts. 2.3. Catalytic experiments

The catalytic experiments were carried out in a stirred batch

may involve relatively bulky intermediates that can limit the activ- reactor, at 70 ◦ C. In a typical experiment, the reactor was loaded
ity and selectivity of zeolite molecular sieves [18]. with 10 ml of butanal, 4.0 g of glycerol and 0.3 g of zeolite.
The purpose of this paper is to study the synthesis of bio- The stability tests of BEA zeolite were carried out by four con-
additives to fuel by acetalisation of glycerol with butanal over secutive experiments, at the same reaction conditions. Between the
zeolites. Different zeolites (USY, BEA and ZSM-5) were used as cat- catalytic experiments, the catalyst was separated from the reac-
alysts. In order to study the effect of Si/Al ratio of USY on catalytic tion mixture by centrifugation, washed with water and dried at
activity, different USY zeolites were used. These results were also 120 ◦ C overnight. After this step, the catalyst sample was calcined
compared with the catalytic activity of BEA and ZSM-5. The cat- at 500 ◦ C, during 3 h.
alyst with the highest activity was used to optimize the reaction Samples were taken periodically and analysed by GC, using a
conditions. The influence of various reaction parameters, such as KONIC HRGC-3000C instrument equipped with a 30 m × 0.25 mm
temperature, molar ratio of butanal to glycerol and catalyst loading DB-1 column. Helium was used as carrier gas. The injector and
on the activity of BEA zeolite, is also studied. The catalytic stability detector temperatures were, respectively, 180 and 300 ◦ C. The split
of BEA zeolite was investigated. ratio was fixed at 100. The oven temperature program was as fol-
lows: start at 40 ◦ C (6 min), ramp at 5 ◦ C min−1 to 300 ◦ C.

2. Experimental
3. Results and discussion
2.1. Preparation of the catalysts
The nitrogen adsorption–desorption isotherms of zeolites are
type I (according to IUPAC nomenclature), characteristic of microp-
Beta zeolite (Si/Al = 40) was synthesized according to Wan-
orous materials. The specific surface area was determined using the
dlinger and Kerr [19]. Briefly, 3.0 g of sodium aluminate (0.0365 mol
BET method. The catalysts and their properties are listed in Table 1.
NaAlO2 ), 110 g of 40% (w/w) tetraethylammonium hydroxide
The surface areas of the zeolites reported in Table 1 are consistent
(0.299 mol TEAOH) and 291.57 g of silica sol (1.461 mol Ludox LS
with literature values and the supplier data sheets. It was found
(30% SiO2 )) were mixed and transferred to a Teflon autoclave. After
that when the Si/Al ratio of USY zeolite increases, the surface area
crystallization (145 h at 150 ◦ C), the zeolite was collected by filtra-
(SBET ) and total pore volume (Vp ) increase as well (Table 1). This
tion, dried at 115 ◦ C and calcined at 500 ◦ C to remove the organic
is an indication that there is a change in the textural properties of
template.
USY zeolite after dealumination.
USY (CBV600, CBV712, CBV720 and CBV780) and ZSM-5 (CBV
The acid site densities were calculated from the SiO2 /Al2 O3
28014) zeolite were supplied by Zeolyst Int., and they were NH4 +
ratios and are given in Table 1. It was observed that the acid site con-
exchanged followed by calcination at 773 K during 3 h.
centration of USY zeolite decrease when the Si/Al ratio increases.
Dowex resin (commercial resin, D50W2, with 2% cross-linking
This behaviour can be explained due to the decrease of aluminium
degree) was used as catalyst.
atoms in zeolite framework.
Fig. 1 shows the X-ray diffraction patterns obtained for the
2.2. Characterization of the catalysts zeolites. Change of structure after synthesis and purification pro-
cedures was not observed.
The textural characterization of the catalysts was based on Fig. 2A compares the influence of Si/Al ratio of different USY zeo-
the nitrogen adsorption isotherm, determined at 77 K with a lites on the initial activity of glycerol acetalisation with butanal. It
Micromeritics ASAP 2010 apparatus. is observed that the initial activity increases with the Si/Al ratio
 H. Serafim et al. / Chemical Engineering Journal 178 (2011) 291–296 293

80
(A)

60

Activity (mmol/h.gcat)
Intensity (a.u.)

40
E
D
C
B 20
A

5 10 15 20 25 30 35 40 45 50
2θ/º 0

Fig. 1. XRD patterns of (A) USY1; (B) USY2; (C) USY3; (D) USY4; (E) BEA; (F) ZSM-5.
BEA USY3 ZSM-5 Dowex

100
of zeolite until a maximum, which is obtained with USY3. With (B)
higher Si/Al ratio, a decrease in the catalytic activity is observed. At
80
a low Si/Al ratio, the increase of framework Si/Al ratio leads to the

Conversion (%)
increase of activity. This behaviour can be explained due to the cat-
alysts become more hydrophobic by increasing of framework Si/Al 60
ratio, achieve a better simultaneous diffusion and a more optimal
concentration of the two reactants (butanal and glycerol) within the 40
zeolite pores, while the number of acid sites decreases (Table 1).

20
50
(A)
0
40 0 2 4 6 8
Time (h)
Activity (mmol/h.gcat)

30 Fig. 3. Acetalisation of glycerol with butanal over different zeolites. (A) Initial activ-
ities taken as the maximum observed reaction rate. (B) Conversion versus time (h):
( ) BEA; () USY3; (×) ZSM-5; () Dowex resin. Reaction conditions: molar ratio
of glycerol to butanal = 1:2.5; temperature = 70 ◦ C; catalyst loading = 0.3 g.
20

However, with the higher framework Si/Al ratio leads a decrease

10
of the catalytic activity. These results can be explained due to the
few amount of active sites on zeolite surface, and also probably
too hydrophobic. Climent et al. [20] showed that the acetalisation
0
of phenylacetaldehyde with glycerol can be carried out success-
USY1 USY2 USY3 USY4
fully over USY zeolite with Si/Al = 35. After 4 h of reaction, it was
100 observed that the glycerol conversion (%) is 84%, 76%, 72% and
(B) 45% for the USY3, USY2, USY4 and USY1 catalyst, respectively
(Fig. 2B).
80 Acetalisation of glycerol with butanal was also carried out over
BEA and ZSM-5 zeolites and Dowex resin. Fig. 3A compares the ini-
Conversion (%)

tial activity of BEA, USY3, ZSM-5 and Dowex resin. It was observed
60
that the ZSM-5 zeolite showed the poor catalytic activity. This
behaviour can be explained in terms of the narrow pore struc-
40 ture, which probably does not permit the acetalisation to occur
inside the pores. ZSM-5 zeolite shows the lower surface area and
total pore volume than USY3 and BEA zeolite (Table 1). Proba-
20 bly, the reaction takes place at the external surface or at the pore
entrance. It can be also observed that the ZSM-5 acid site concen-
0 tration is lower than BEA and USY3. BEA zeolite showed higher
catalytic activity than USY3. This behaviour can be explained due
0 2 4 6
to the hydrophilic/hydrophobic balance of BEA surface. da Silva et
Time (h) al. [21] studied the acetalisation of glycerol with aqueous formalde-
hyde solution over zeolites. The use of beta zeolite (Si/Al = 16) leads
Fig. 2. Acetalisation of glycerol with butanal over USY zeolites. Effect of framework to high conversion (95% within 60 min of reaction). Dowex resin
Si/Al ratio. (A) Initial activities taken as the maximum observed reaction rate. (B)
showed higher catalytic activity than ZSM-5 zeolite. This result can
Conversion versus time (h): ( ) USY1; () USY2; () USY3; (×) USY4. Reaction
conditions: molar ratio of glycerol to butanal = 1:2.5; temperature = 70 ◦ C; catalyst be explained due to the high amount of active site of Dowex resin
loading = 0.3 g. (Table 1). After 4 h of reaction, it was observed that the glycerol
294 H. Serafim et al. / Chemical Engineering Journal 178 (2011) 291–296

Table 2 Table 3
Conversion and selectivity to the glycerol acetalisation products with butanal over Conversion and selectivity to the products of the acetalisation of glycerol with
zeolites. butanal over BEA catalyst. Effect of catalyst loading.

Sample Conversiona (%) Selectivity (%) Catalyst Conversiona (%) Selectivity (%)
loading (g)
Five-member Six-member
ring acetal ring acetal Five-member Six-member
ring acetal ring acetal
USY1 45 81 19
USY2 76 77 23 m = 0.10 53 82 18
USY3 84 78 22 m = 0.20 79 80 20
USY4 72 80 20 m = 0.30 87 79 21
BEA 87 79 21 a
Glycerol conversion after 4 h of reaction.
ZSM-5 28 82 18
Dowex 66 80 20
a 100
Glycerol conversion after 4 h of reaction.

80
conversion (%) is 87%, 84%, 66% and 28% for the BEA, USY3, Dowex
and ZSM-5 catalyst, respectively (Fig. 3B).

Conversion (%)
Table 2 shows the glycerol conversion and the selectivity to the 60
products obtained by the acetalisation of glycerol with butanal over
zeolites, after 4 h of the reaction. It can be seen that the sample
BEA zeolite exhibits the highest conversion. After 4 h of reaction, 40
the glycerol conversion was 87%, with a selectivity of 79% to five-
member ring acetal and 21% to six-member ring acetal. The high 20
selectivity to five-member ring acetal can be explained due to the
fact that five-member ring acetal (1,3-dioxolane) is favored kineti-
cally. Similar results were also observed by Umbarkar et al. [7] and 0
Silva et al. [6]. 0 1 2 3 4 5 6
In order to optimize the reaction parameters (catalyst loading, Time (h)
molar ratio of glycerol to butanal and temperature) on acetalisa-
Fig. 5. Acetalisation of glycerol with butanal over BEA catalyst. Effect of tempera-
tion of glycerol with butanal, different catalytic experiments were
ture. Conversion (%) versus time (h): ( ) T = 70 ◦ C; () T = 50 ◦ C; () T = 30 ◦ C. Reaction
carried out over BEA zeolite. conditions: molar ratio of glycerol to butanal = 1:2.5; catalyst loading = 0.3 g.

3.1. Influence of reaction parameters Table 4

Conversion and selectivity to the products of the acetalisation of glycerol with
butanal over BEA catalyst. Effect of temperature.
3.1.1. Catalyst loading
Fig. 4 shows the effect of catalyst loading (BEA) on the glyc- Temperature (◦ C) Conversiona (%) Selectivity (%)
erol conversion. The ratio of glycerol:butanal and the temperature Five-member Six-member
were kept constant. The increase of catalyst load from 0.1 g to ring acetal ring acetal
0.3 g showed substantial effect on the glycerol conversion. This
30 40 81 19
behaviour can be explained due to the increase in the total number 50 68 80 20
of available active catalytic sites for the reaction. 70 87 79 21
Table 3 shows the effect of the catalyst loading in the selectivity a
Glycerol conversion after 4 h of reaction.
of the BEA to the different products of the glycerol acetalisation. It
was observed that all catalyst loading gives good selectivity values
to five-member ring acetal. 3.2. Reaction temperature

In order to evaluate the influence of reaction temperature on

100 the acetalisation of glycerol with butanal over BEA, different exper-
iments were carried out. The reaction temperature was varied from
30 ◦ C to 70 ◦ C, while the ratio glycerol:butanal and the amount of
80
catalysts were kept constant. As expected, the conversion increased
Conversion (%)

with the temperature (Fig. 5). The effect of temperature on glyc-

60 erol conversion and on the selectivity to the reaction products is
shown in Table 4. It was observed that when the acetalisation of
40 glycerol with butanal is carried out at 70 ◦ C, after 4 h, the selectivity
to five-member ring acetal is 79 at 87% of glycerol conversion.
20
3.3. Molar ratio of glycerol to butanal

0
In order to study the influence of the mixture chemical compo-
0 1 2 3 4 5 6
sition on glycerol conversion, the glycerol/butanal ratio was varied
Time (h) using the proportions 1:1, 1:2.5 and 1:6 at reaction temperature
of 70 ◦ C for sample BEA. Fig. 6 shows the effect of molar ratio of
Fig. 4. Acetalisation of glycerol with butanal over BEA catalyst. Effect of catalyst
load. Conversion (%) versus time (h): () m = 0.10 g; () m = 0.20 g; ( ) m = 0.30 g. glycerol to butanal on glycerol conversion. It was observed that the
Reaction conditions: molar ratio of glycerol to butanal = 1:2.5; temperature = 70 ◦ C. glycerol conversion increases when the molar ratio of glycerol to
 H. Serafim et al. / Chemical Engineering Journal 178 (2011) 291–296 295

100 reusability of the catalyst. After the reaction on fresh catalyst, the
zeolite was separated from the reaction mixture by centrifugation,
washed with water and dried at 120 ◦ C overnight. After this period,
80
the zeolite was calcined at 500 ◦ C for 3 h in the presence of air. Fig. 7
Conversion (%)

shows the initial activity of BEA zeolite obtained in different uses.

60 It is observed a slight decrease of the catalytic activity in the second
run (Fig. 7), which could be explained due to some reagent/product
adsorb on zeolite surface. However, similar catalytic activity was
40
observed after the third use. BEA zeolite only lost 7% of its initial
activity, after the fourth use.
20
4. Conclusions
0
0 1 2 3 4 5 6 Zeolite catalysts (USY, BEA and ZSM-5) were found to be highly
Time (h) active for the acetalisation of glycerol with butanal. USY zeolites
with different Si/Al ratio were used as catalyst. It was observed
Fig. 6. Acetalisation of glycerol with butanal over BEA catalyst. Effect of molar ratio that an optimum in initial catalytic activity of USY zeolite with
of glycerol to butanal. Conversion (%) versus time (h): () 1:1; ( ) 1:2.5; () 1:6.
the Si/Al ratio equal to 30. BEA showed the highest catalytic
Reaction conditions: temperature = 70 ◦ C; catalyst loading = 0.3 g.
activity (75.6 mmol/h gcat ) while the ZSM-5 showed the lowest
activity (9.0 mmol/h gcat ) of all zeolites. High values of selectivity
Table 5
Conversion and selectivity to the products of the acetalisation of glycerol with to five-member ring acetal were obtained with all zeolites (about
butanal over BEA catalyst. Effect of molar ratio of glycerol to butanal. 77–82%).
Different reaction parameters were studied and the optimum
Ratio of Conversiona (%) Selectivity (%)
glycerol:butanal
reaction conditions were established over BEA zeolite.
BEA zeolite showed good catalytic stability. After the fourth use,
Five-member Six-member
it was observed that the catalytic activity decreased only 7% of its
ring acetal ring acetal
initial value.
1:1 71 81 19
1:2.5 87 79 21
1:6 88 80 20 References
a
Glycerol conversion after 4 h of reaction.
[1] F. Ma, M. Hanna, Biodiesel production: a review, Bioresour. Technol. 70 (1999)
1–15.
butanal increases from 1:1 to 1:2.5. However, when the molar ratio [2] Y. Zheng, X. Chen, Y. Shen, Commodity chemicals derived from glycerol, an
important biorefinery feedstock, Chem. Rev. 108 (2008) 5253–5277.
of glycerol to butanal increases (from 1:2.5 to 1:6) an increase of [3] N. Rahmat, A.Z. Abdullah, A.R. Mohamed, Recent progress on innovative and
conversion is not observed. This behaviour can be attributed with potential technologies for glycerol transformation into fuel additives: a critical
the occupancy of butanal over the active sites and the availability review, Renew. Sustain. Energy Rev. 14 (2010) 987–1000.
[4] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Glycerol acety-
of alcohol molecules for further acetalisation. Similar results were lation over dodecatungstophosphoric acid immobilized into a silica matrix as
also observed by Jermy and Pandurangan [22]. catalyst, Appl. Catal. B: Environ. 91 (2009) 416–422.
It is also observed that indifferent to the ratio, there is high [5] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Valorisation of
glycerol by condensation with acetone over silica-included heteropolyacids,
selectivity to the five-member ring acetal (Table 5).
Appl. Catal. B: Environ. 98 (2010) 94–99.
[6] P.H.R. Silva, V.L.C. Gonçalves, C.J.A. Mota, Glycerol acetals as anti-freezing addi-
3.4. Catalyst stability tives for biodiesel, Bioresour. Technol. 101 (2010) 6225–6229.
[7] S.B. Umbarkar, T.V. Kotbagi, A.V. Biradar, R. Pasrichab, J. Chanale, M.K. Don-
gare, A.-S. Mamede, C. Lancelot, E. Payen, Acetalization of glycerol using
BEA zeolite used in the acetalisation of glycerol with butanal mesoporous MoO3 /SiO2 solid acid catalyst, J. Mol. Catal. A: Chem. 310 (2009)
was recycled four times in order to check the activity, stability and 150–158.
[8] A. Corma, From microporous to mesoporous molecular sieve materials and
their use in catalysis, Chem. Rev. 97 (1997) 2373–2419.
80 [9] R.A. Sheldon, R.S. Downing, Heterogeneous catalytic transformations for envi-
ronmentally friendly production, Appl. Catal. A: Gen. 189 (1999) 163–183.
[10] U. Freese, F. Heinrich, F. Roessner, Acylation of aromatic compounds on H-Beta
zeolites, Catal. Today 49 (1999) 237–244.
[11] M.S. Holm, E. Taarning, K. Egeblad, C.H. Christensen, Catalysis with hierarchical
60 zeolites, Catal. Today 168 (2011) 3–16.
Activity (mmol/h.gcat)

[12] P.B. Venuto, Organic catalysis over zeolites: a perspective on reaction paths
within micropores, Micropor. Mater. 2 (1994) 297–411.
[13] W.F. Höilderich, H. van Bekkum, Zeolites in organic syntheses, Stud. Surf. Sci.
Catal. 58 (1991) 631–726.
40
[14] A. Corma, M.J. Climent, H. García, J. Primo, Formation and hydrolysis of acetals
catalysed by acid faujasites, Appl. Catal. 59 (1990) 333–336.
[15] R. Ballini, G. Bosica, B. Frullanti, R. Maggi, G. Sartori, F. Schroer, 1,3-Dioxolanes
from carbonyl compounds over zeolite HSZ-360 as a reusable, heterogeneous
20 catalyst, Tetrahedron Lett. 39 (1998) 1615–1618.
[16] I. Rodriguez, M.J. Climent, A. Corma, S. Iborra, V. Fornés, Use of delaminated
zeolites (ITQ-2) and mesoporous molecular sieves in the production of fine
chemicals: preparation of dimethylacetals and tetrahydropyranylation of alco-
hols and phenols, J. Catal. 192 (2000) 441–447.
0 [17] M.J. Climent, A. Corma, A. Velty, M. Susarte, Zeolites for the production
1st use 2nd use 3rd use 4th use of pine chemicals: synthesis of the fructone fragrancy, J. Catal. 196 (2000)
345–351.
Fig. 7. Acetalisation of glycerol with butanal over BEA catalyst. Reusability of BEA [18] M.J. Climent, A. Velty, A. Corma, Design of a solid catalyst for the synthesis of a
zeolite. Initial activities taken as the maximum observed reaction rate, calculated molecule with blossom orange scent, Green Chem. 4 (2002) 565–569.
from the maximum slope of the glycerol kinetic curve. [19] R.L. Wandlinger, G.T. Kerr, US Pat. Appl. 28,341 (1975).
296 H. Serafim et al. / Chemical Engineering Journal 178 (2011) 291–296

[20] M.J. Climent, A. Corma, A. Velty, Synthesis of hyacinth, vanilla, and blossom [22] B.R. Jermy, A. Pandurangan, Al-MCM-41 as an efficient heterogeneous
orange fragrances: the benefit of using zeolites and delaminated zeolites as catalyst in the acetalization of cyclohexanone with methanol, ethy-
catalysts, Appl. Catal. A: Gen. 263 (2004) 155–161. lene glycol and pentaerythritol, J. Mol. Catal. A: Chem. 256 (2006)
[21] C.X.A. da Silva, V.L.C. Gonçalves, C.J.A. Mota, Water-tolerant zeolite catalyst for 184–192.
the acetalisation of glycerol, Green Chem. 11 (2009) 38–41.
 Fuel 117 (2014) 470–477

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Thermodynamic and kinetic studies of a catalytic process to convert

glycerol into solketal as an oxygenated fuel additive
Malaya R. Nanda a, Zhongshun Yuan a, Wensheng Qin b, Hassan S. Ghaziaskar c, Marc-Andre Poirier d,
Chunbao Charles Xu a,⇑
a
Institute for Chemicals and Fuels from Alternative Resources (ICFAR), Department of Chemical and Biochemical Engineering, Western University, London, ON N6A 3K7, Canada
b
Biorefining Research Initiative and Biology Department, Lakehead University, ON P7B 5E1, Canada
c
Department of Chemistry, Isfahan University of Technology, IR 84156-83111, Iran
d
Imperial Oil Limited, Supply Support and Environmental Group, ON N7T 8C8, Canada

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Thermodynamics and kinetics of

glycerol ketalization were studied for
the first time.
 External mass transfer resistance was
completely eliminated above
400 rpm.
 Moisture content demonstrated an
adverse effect on the yield of solketal.
 Kinetics of the reaction was modeled
by Langmuir–Hinshelwood rate
expression.
 The activation energy for the reaction
was determined to be 55.6 kJ mol1.

a r t i c l e i n f o a b s t r a c t

Article history: Glycerol is a byproduct of biodiesel industry and can be converted into high value-added applications. The
Received 5 August 2013 heterogeneous ketalization of glycerol with acetone was conducted over a solid acid catalyst of Amberlyst-
Received in revised form 13 September 35 in a batch reactor. The thermodynamics and kinetics of the ketalization reaction for the synthesis of
2013
solketal were investigated. The reaction equilibrium constants were determined experimentally in the
Accepted 19 September 2013
Available online 2 October 2013
temperature range of 293–323 K, with which the following standard molar properties (at 298 K) were
obtained: DH0 = 30.1 ± 1.6 kJ mol1, DG0 = 2.1 ± 0.1 kJ mol1, DS0 = 0.1 ± 0.01 kJ mol1K1. Effects of
various experimental conditions (stirring speed, catalyst addition amount, pressure, temperature,
Keywords:
Adsorption
moisture content and the feed composition) on the reaction kinetics (glycerol conversion and solketal
Batch reactor yield vs. time) were also investigated in this work. A two-parameter kinetic law based on a Langmuir–
Glycerol Hinshelwood rate expression was used. The activation energy of the overall ketalization reaction was
Ion exchange resin determined to be 55.6 ± 3.1 kJ mol1. The obtained solketal could be synthesized from renewable
Kinetics resources like bioglycerol and biomass derived acetone, and seem to be a good candidate for different
applications such as fuel additive and in pharmaceutical industries. The work is an important step for
further development of a technology for the continuous synthesis and separation of solketal from glycerol
and acetone.
Ó 2013 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +1(519)661 2111x86414; fax: +1 (519)661 4016.

E-mail address: [email protected] (C.C. Xu).

0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2013.09.066
 M.R. Nanda et al. / Fuel 117 (2014) 470–477 471

1. Introduction of different heterogeneous catalysts for the formation of cyclic ke-

tals by the condensation of glycerol with aldehydes. A high yield of
Glycerol (propane-1,2,3-triol) is the simplest trihydric alcohol, solketal (more than 90% with high selectivity) was reported at a
commercially known as glycerin. It is well known for its versatile high molar ratio (approx. 6:1) of acetone to glycerol in a batch
applications in diverse fields such as the food, pharmaceutical, reactor [17]. However, so far kinetics and thermodynamics of the
polymer and fuel industries [1]. ketalization reaction of glycerol with acetone over solid acid cata-
Glycerol is produced at a large amount as a byproduct or waste lysts have not been reported.
stream from biodiesel production via transesterfication reactions. It is of no question that to understand thermodynamics and
The biodiesel production generates approximately 10% of glycerol kinetics of the ketalization reaction is important for further devel-
by volume [2]. Due to the increased concerns over the environment opment of the glycerol ketalization technology. Thus, the main
and energy security associated with petroleum-based transporta- objective of this work is to thoroughly investigate thermodynamics
tion fuels, the interest in producing bio-fuels (bio-ethanol and and kinetics of the ketalization reaction of glycerol and acetone in a
biodiesel) has been intensified worldwide in last decade. The pro- batch reactor over solid acid catalyst of Amberlyst-35. The results
duction of biodiesel has increased exponentially all over the world. of this research would help to suggest a reaction mechanism and a
Hence a large amount of glycerol is expected to be generated from rate equation with experimentally measured kinetic parameters.
the biodiesel industry. It is predicted that by 2020 the global pro- Furthermore, it will be advantageous for the development of a con-
duction of glycerol will be 41.9 billion liters [3]. The large scale tinuous process for the industrial production of solketal from
producers are able to refine this waste stream for the industrial glycerol.
applications whereas small scale producers are unable to justify
refining costs and instead pay a fee for glycerol removal. The cur- 2. Experimental
rent crude glycerol price is as low as 10 cents/lb [4]. The predicted
generation of huge amount of glycerol will further lower the glyc- 2.1. Materials
erol price once it enters the market. Therefore proper utilization of
glycerol is required in different value-added applications for the Glycerol and acetone (both >99 wt.% purity) were procured
sustainability of the biodiesel industry. from Sigma Aldrich and used as received, and commercial grade
The presence of three hydroxyl groups in glycerol makes it ethanol was supplied from Commercial Alcohols Inc. Solketal
unsuitable to be used as a direct fuel component due to its low [(S-) (+)  1, 2-Isopropylideneglycerol, 99 wt.%] was also obtained
heating value. Various processes have been investigated for con- from Sigma Aldrich as a calibration standard for GC analysis. The
version of glycerol into fuel components. Condensation of glycerol solid acid catalyst: Amberlyst-35 DRY was obtained from Rohm
with aldehydes and ketones to cyclic acetals and ketals, respec- and Hass Co. (USA) and used as received. The characteristics of
tively, is often considered one of the most promising glycerol the catalyst are given in Table 1.
applications for fuels/chemical intermediates [5–8]. The ketaliza-
tion reaction between glycerol and acetone is given in Fig. 1, where
2.2. Ketalization reaction of glycerol and acetone
solketal (2,2-dimethyl-1,3-dioxolane-4-methanol) is formed as the
condensation product in the presence of an acid catalyst. Solketal
The glycerol ketalization reactions were carried out in a 100 mL
can be used as a fuel additive to reduce the particulate emission
three-neck glass reactor in a water bath equipped with a magnetic
and to improve the cold flow properties of liquid transportation
stirrer and a condenser (Fig. 2) in order to condense and reflux all
fuels [9]. It helps reduce the gum formation, improves the oxida-
the vapors keeping the reaction volume almost constant. The reac-
tion stability, and enhances the octane number when added to gas-
tion temperature was precisely controlled for an accuracy of
oline [10]. Maksimov et al. [11] reported its use as a versatile
±0.03 K with an external thermostat containing an external ther-
solvent and a plasticizer in the polymer industry and a solubilizing
mocouple placed inside the reacting mixture. In a typical run, the
and suspending agent in pharmaceutical preparations.
composition of reaction mixture was 22.83 g of acetone, 18.11 g
Traditionally ketalization of glycerol is carried out via a homo-
of glycerol, 9.06 g of ethanol, 0.1811 g of catalyst and the total vol-
geneous catalytic process using mineral acids like H2SO4, HCl, HF,
ume of the mixture was 55 mL. At the beginning of the experiment,
and p-toluenesulphonic acid, etc. [12,13]. These processes have
known amount of glycerol, ethanol and acetone were charged into
serious shortcomings such as corrosion and catalyst separation
the reactor. Amberlyst-35 catalyst was placed in a basket at the top
from the product stream, hence raising environmental and eco-
of the condenser and added to the reactor only after the stabiliza-
nomic concerns for the effluent disposal. Most of these problems
tion of temperature of the system (time zero). The use of ethanol as
could be addressed by using heterogeneous catalysts. Studies of
solvent was mainly to improve the solubility of glycerol in acetone
ketalization of glycerol using solid acid catalysts like Amberlyst-
and the homogeneity of the solution was observed by the
15 [14], Amberlyst-36 [5], Montmorrilonite K-10 [15], Zeolite
formation of a single phase (naked eye observation), also checked
[11,14], silica supported heteropoly acids [16], and mesoporous sil-
by the GC-FID. The reaction of acetone with ethanol to form
icates containing arylsulfonate groups [17] were reported. Among
these catalysts, Amberlyst has demonstrated its potential for the
synthesis of solketal. Deutsch et al. [5] investigated the reactivity Table 1
Characteristics of the solid acid catalyst used.

Catalyst properties

OH
H3C
CH3
Acidity (eq/kg)a 5
O Particle size (lm) 490
+ HO OH + H2O Average pore diameterb (nm) 30
H3C CH3 O Max. operating tempc (°C) 150
O
Pore volume b (mL/g) 0.35
BET surface areab (m2/g) 50
a
Determined by ammonia TPD.
OH b
Measured by N2 isothermal adsorption at 77 K.
c
Obtained from the catalyst supplier.
Fig. 1. Ketalization reaction scheme of glycerol and acetone.
472 M.R. Nanda et al. / Fuel 117 (2014) 470–477

Conversion ðmol%Þ
Reduction in moles of glycerolin the reaction
¼  100% ð2Þ
Initial moles of glycerol

3. Thermodynamic results

Thermodynamic studies of the glycerol ketalization reaction

were carried out in the 100 mL batch reactor in a relatively lower
temperature range of 293–323 K, as the reaction is exothermic so
thermodynamically unfavourable at higher temperatures [20]. In
this series of experiments, a high initial molar ratio of acetone to
glycerol (6:1) was used, and the catalyst loading in the batch reac-
tor was fixed at 5 wt.% of the mass of glycerol. To ensure equilib-
rium of the reaction, all the experiments were kept for an
enough long time while monitoring the glycerol conversion and
solketal yield vs. time until the measured results did not change
(indicator of the reaction equilibrium). The equilibrium composi-
tions at various experimental conditions are listed in Table 2.
The equilibrium constant (Kc) for the liquid phase reaction was
calculated using the following equation and the results are pre-
sented in Table 2:
½S½W
KC ¼ ð3Þ
½A½G
where [S], [W], [A], and [G] are the molar concentration of solketal,
water, acetone, and glycerol, respectively. As shown from Table 2,
with increase in the reaction temperature, the equilibrium constant
Kc decreases gradually, indicating the exothermicity of the reaction.
To achieve a higher equilibrium conversion of glycerol, a lower tem-
perature is preferred. However, as expected the lower the reaction
temperature the longer time is required to reach the equilibrium
Fig. 2. Batch reactor.
state.
The thermodynamic properties such as entropy and enthalpy
hemiketal/ketal under this experimental condition is highly unfa- can be predicted by plotting the experimental values of ln Kc vs.
vourable [18,19] and the yields of other undesired products were 1/T (K1). In a narrow temperature range in the vicinity of room
very small (<2%) [11]. The total mass of the reaction substrate temperature, the plot of ln Kc vs. 1/T would follow a linear correla-
(M) was maintained to be 50 g, unless otherwise specified. A small tion as displayed in Fig. 3:
amount of samples were withdrawn at regular intervals of time for
analysis by GC-FID. The reaction was kept at the specific tempera- DS0 DH 0 1
ln K C ¼  ð4Þ
ture until the equilibrium state was achieved (when the reactants/ R R T
products concentrations did not vary with time) which was where DS0 is the standard entropy at 298 K (kJ mol1 K1), DH0 is
checked by the GC-FID in a regular interval of time. the standard enthalpy at 298 K (kJ mol1), R is the universal gas
constant (J mol1 K1) and T is the reaction temperature (K). The
2.3. Product analysis linear fitting of experimental data (in Fig. 3) according to above
equation gives:
The main components in the product mixture were first deter-
mined with a gas chromatograph, equipped with a mass selective ln K C ¼ ð3615:4=TÞ  11:31 ð5Þ
detector [Varian 1200 Quadrupole GC/MS (EI), Varian CP-3800 0
By solving Eqs. (4) and (5), we get DH = 30.1 ± 1.6 kJ mol and1

GC equipped with VF-5 MS column (5% phenyl/95% dimethyl-poly- DS0 = 0.1 ± 0.01 kJ mol1 K1 from the slope and the intercept,
siloxane, 30 m  0.25 mm  0.25 lm)], using helium as the carrier respectively. The standard state Gibbs free energy change (DG0)
gas at a flow rate of 5  107 m3/s. The oven temperature was can be related to the standard state enthalpy and entropy changes
maintained at 120 °C for 2 min and then increased to 200 °C at a for the system as:
ramp rate of 40 °C/min. Injector and detector block temperature
were maintained at 300 °C. The components were identified using DG0 ¼ DH0  T DS0 ð6Þ
the NIST 98 MS library with the 2002 update. The concentration of 0 1
with the above, DG is found to be 2.1 ± 0.1 kJ mol , suggesting
the glycerol and solketal in the products was analyzed with a GC-
the reaction can take place at standard state (room temperature).
FID (Shimadzu-2010) under the similar conditions as used for the
The DG0 value obtained is similar to the result reported in the liter-
GC–MS measurement. In a particular run of 70% yield of solketal,
ature for the synthesis of acetal from butanol and acetaldehyde
a glycerol conversion of 71% was observed and the ratio of yield
[21].
to conversion was consistent throughout the experiment.
The solketal yield and glycerol conversion were calculated using
the following equations: 4. Kinetic results

Moles of solketal formed Referring to the mechanism proposed for the synthesis of acetal
Yield ðmol%Þ ¼  100% ð1Þ
Initial moles of glycerol in the presence of a homogeneous catalyst [22] and the mechanism
 M.R. Nanda et al. / Fuel 117 (2014) 470–477 473

Table 2
Experimental data of equilibrium composition and equilibrium constants.a

Temp (K) IA IG FA FG FS FW KC XE
298 0.6817 0.1145 0.5747 0.0075 0.1070 0.1070 2.6562 0.9345
303 0.694 0.1132 0.5907 0.0099 0.1033 0.1033 1.8247 0.9125
313 0.6823 0.1153 0.5807 0.0137 0.1016 0.1016 1.2975 0.8812
323 0.6827 0.1137 0.5854 0.0164 0.0973 0.0973 0.9861 0.8558
a
IA = initial mole of acetone; IG = initial mole of glycerol; FA = final mole of acetone; FG = final mole of glycerol; FS = final mole of solketal; FW = final mole of water;
KC = equilibrium constant; XE = equilibrium conversion.

1.40 kinetics (glycerol conversion and solketal yield vs. time) and are
summarized in Table 3. The results are presented as follows.

1.00 4.1. Mass transfer resistance

y = 3.6154x - 11.308
R2 = 0.9987
To investigate the effects of mass transfer on the reaction kinet-
ln K c

0.60 ics, a wide range of agitation (stirring) speeds (from 400 rpm to
1150 rpm) were tested in the experiments. The solketal yields vs.
time under two different stirring speeds (400 and 1100 rpm) are
0.20 illustrated in Table 3 (entry 1 and 2) and Fig. 5. Clearly, at the same
conditions (323 K, acetone to glycerol molar ratio (A/G) of 2, and
catalyst loading (Wcat) of 1 wt.% of glycerol) both tests under 400
-0.20 and 1100 rpm led to the same equilibrium yield of solketal (60%)
3.05 3.15 3.25 3.35 3.45 as well as the initial formation rate of solketal (determined from
1/T (x 103 K-1) the slope of the trend-line of the solketal yield vs. time at the
beginning of the experiment). Thus, no effect of the agitation speed
Fig. 3. Plot of ln Kc vs. 1/T. on the reaction rate was observable at >400 rpm. Hence, all further
experiments were carried out at 700 rpm to eliminate the external
mass transfer resistance. The catalyst used in this study was a mac-
proposed by Maksimov et al. [11] for the synthesis of ketals from roscopic ion exchange resin (Table 1). In a macroscopic resin, the
plant-derived diols, we proposed a similar mechanism for ketaliza- reactants are able to diffuse into the pores without any resistance.
tion of glycerol over a heterogeneous catalyst as illustrated in Hence no internal mass transfer resistance was expected [23,24].
Fig. 4. The most important steps in the mechanism are:
4.2. Addition of catalyst
(1) Reaction between the adsorbed acetone and glycerol to give
the hemi-acetal (Step 1). The effects of catalyst addition amount on the reaction kinetics
(2) Reaction to form water (Step 2 – considered to be the rate were investigated under the conditions of 313 K and A/G = 2 while
limiting step). with different catalyst addition amount (i.e., Wcat = 1 wt.% and 2
(3) Reaction to form solketal (Step 3). wt.% in relation to glycerol). The results are given in Table 3 (entry
6 and 8) and Fig. 6, from which essentially the increase in the
In this study, we investigated various experimental conditions catalyst addition amount from 1 wt.% to 2 wt.% does not change
(stirring speed, catalyst addition amount, pressure, temperature, the final (equilibrium) yield of solketal (64%) as expected by ther-
moisture content and the reactor feed composition) on the reaction modynamics. Under the same experimental conditions, a two fold

H HO H
H+ H H
O +
O CH2 O
C O CH2
+
C O CH2
HO CH CH2OH
HO CH CH2OH
HO CH CH2OH

Fig. 4. Scheme of mechanism for ketalization reaction of glycerol and acetone.

474 M.R. Nanda et al. / Fuel 117 (2014) 470–477

Table 3
Summary of the experiments at different conditions.

Entry number Catalyst loading (wt.% of glycerol) Stirring speed (rpm) Temperature (K) Acetone:glycerol:ethanol (mole) Solketal yield (%)
1 1 400 323 2:1:1 60
2 1 1100 323 2:1:1 60
3 1 700 298 2:1:1 72
4 1 700 303 2:1:1 70
5 1 700 308 2:1:1 67
6 1 700 313 2:1:1 64
7 1 700 323 2:1:1 60
8 2 700 313 2:1:1 64
9 1 700 298 1.48 :1:1 68
10 1 700 298 2.46 :1:1 74
11 1 700 298 2.04:1:1 72
12a 1 700 298 2:1:1 68
13b 1 High 298 2:1:1 72
14c 1 High 298 2:1:1 72
a
With 3.15 wt.% moisture.
b
Conducted in a pressure reactor at 1 atm.
c
Conducted in a pressure reactor at 54 atm.

70 4.3. Effect of pressure

60 The effects of pressure (1–54 atm) on the reaction was tested,

Solketal yield (mol%)

and it was found that pressure showed a negligible effect on either

50
the equilibrium product yields or the reaction rate, which is ex-
pected for liquid phase reactions (Table 3: entry 13 and 14). In
40
400 rpm the present study, the reactor pressure for all reported results
1100 rpm was fixed at 1 atm, where the maximum number of molecules in
30
gas phase was found to be very small (2.1%) at the maximum oper-
20 ating temperature and the maximum acetone equivalent.

10 4.4. Effect of temperature

0
Effects of temperature on ketalization of glycerol were studied
0 10 20 30 40 50 60 70
at various temperatures ranging from 298 to 323 K under the con-
Time (min)
ditions of A/G = 2 and Wcat = 1 wt.% of glycerol, as shown in Table 3
Fig. 5. Effects of reactor stirring speed on the solketal yield (other reaction
(entry 3–7) and Fig. 7. A higher temperature led to a lower equilib-
conditions: 323 K, acetone to glycerol molar ratio (A/G) of 2, catalyst loading (Wcat) rium product yield, typical of exothermic reactions, as evidenced
of 1 wt.% of glycerol). previously by the thermodynamics results. It is also clear that
the initial rate of the ketalization reaction increases with increas-
ing the reaction temperature as expected.
increase in the mass of catalyst approximately doubled the initial
reaction rate for solketal formation, suggesting that the reaction 4.5. Initial molar concentration of reactants
rate can be promoted by increasing the catalyst amount or number
of the active sites in the reactor system, as similarly observed in According to both reaction thermodynamics and kinetics, the
the literature for the synthesis of acetal from butanol and acetalde- initial molar concentration of a reactant would influence the
hyde [20]. equilibrium conversion and the reaction rate. In this work, we

80
70
2 wt%
60
Solketal yield (mol%)
Solketal yield (mol%)

60
50

40
40
T= 298 K T= 303 K
30
T= 308 K T= 313 K
20 1 wt %
20
T= 323
10

0 0
0 30 60 90 120 150 0 75 150 225 300
Time (min) Time (min)

Fig. 6. Effects of the catalyst addition amount on the yield of solketal (other Fig. 7. Influence of temperature on the yield of solketal (other reaction conditions:
reaction conditions: 313 K and A/G = 2). A/G = 2 and Wcat = 1 wt.% of glycerol).
 M.R. Nanda et al. / Fuel 117 (2014) 470–477 475

conducted the reaction at 298 K with catalyst loading amount of 1 80

wt.% of glycerol while varying the initial acetone-to-glycerol molar
ratio from A/G = 1.48–2.46. The results are presented in Table 3

Solketal yield (mol%)

(entry 9–11) and Fig. 8. As clearly shown in the figure, the reaction 60
thermodynamics and kinetics are strongly affected by A/G molar
ratio: a higher A/G ratio led to a higher reaction rate and larger
equilibrium yield of solketal. These results are actually expected, 40
and similar observations were reported by Agirre et al. [6]. We Dry catalyst
have also examined the effect of initial ethanol concentration on wet catalyst
the ketalization reaction at various ethanol–glycerol molar ratios, 20
but the effect was found to be negligible due to the minimal reac-
tion between acetone and ethanol under the present reaction con-
ditions (reaction time, temp, amount of catalyst used, amount of
0
glycerol and acetone used). 0 100 200 300 400
Time (min)
4.6. Moisture content
Fig. 9. Effect of moisture content on the yield (other conditions: 298 K, Wcat = 1
The role of water in the reaction was investigated (Table 3: en- wt.% of glycerol).
try 3 and 12, Fig. 9) by using 3.15 wt.% water in the solution for the
ketalization and the yield of solketal (66%) was compared to that of
without moisture experiment (72%) under similar reaction condi- where K1, and K3 are the equilibrium constants in step a and c,
tions (Temperature: 298 K, acetone to glycerol molar ratio (A/G) respectively. K2 and K2 are the rate constants for forward and
of 2, catalyst loading (Wcat) of 1 wt.% of glycerol). The lower yield backward reactions, respectively in step b. As reported in the liter-
of solketal in the later case may be attributed to the adsorption ature, the surface reaction for the formation of adsorbed water
of water on the catalyst surface which inhibits the forward reaction (WF) could be the rate determining step [29–31]. So the rate of
to form solketal [25]. the reaction (r) can be given as [20,32]

4.7. Kinetic model r ¼ K 2 ½HF½F  K 2 ½IF½WF ð10Þ

The general reaction rate expression for the ketalization of According to Langmuir adsorption isotherm, the adsorption
glycerol with acetone could be expressed in the form of Lang- equilibrium for the species i is given as
muir–Hinshelwood model with surface reaction as the rate
K i ½i
determining step [26–28]. The key reaction steps of this model hi ¼ PN ð11Þ
are given as follows: 1þ j¼1 K j ½j

where hi the fraction of sites occupied by i and K is the equilibrium

(a) The surface reaction between the adsorbed species of glyc-
adsorption constant.Combining the above equations and assuming
erol (GF) and acetone (AF) to give adsorbed hemiacetal (HF)
that water is the more adsorbed component on the catalyst, the rate
K1 of the reaction can be modified as [20,30,31]
GF þ AF $ HF þ F ð7Þ
½G½A  ½S½W=K c ½G
where F is the vacant site on the catalyst. r¼k ð12Þ
2
(b) Surface reaction for formation of adsorbed water (WF) f1 þ K w ½Wg
K2 where the parameters k, Kw and Kc are the kinetic constant of the
HF þ F $ IF þ WF ð8Þ
overall ketalization reaction, the equilibrium constant for water
where IF represents for the reactive intermediate formed. adsorption on the catalyst surface, and the equilibrium constant
(c) Formation of adsorbed solketal (SF) for the overall ketalization reaction of glycerol and acetone (Table 2),
K3 respectively.
IF þ GF $ SF þ F ð9Þ The mass balance in the batch reactor for solketal in the liquid
phase at constant temperature can be given as [7]
80
dnc
¼ W cat r ð13Þ
dt
Solketal yield (mol%)

A/G=2.46
60 where nc is the number of moles of solketal, t is the time, Wcat is the
mass of catalyst and r is the reaction rate with respect to the cata-
lyst mass.
40 The above equation can be modified using the initial moles
A/G=2.04 (nl,0), stoichiometric coefficient of limiting reactant (vl) and conver-
sion (X) as [20]
20
A/G=1.48
nl;0 dX
¼ rf½G; ½A; ½S; ½Wg ð14Þ
W cat jml j dt
0
0 80 160 240 320 The theoretical rate of the reaction (Eq. (14)) was fitted to the
experimentally measured rates at different temperature and is
Time (min)
given in Fig. 10. The values of k and Kw at different temperatures
Fig. 8. Effects of initial acetone-to-glycerol (A/G) molar ratio on the yield of solketal were calculated using non-linear regression method and are given
(other reaction conditions: 298 K and Wcat = 1 wt.% of glycerol). in Table 4.
476 M.R. Nanda et al. / Fuel 117 (2014) 470–477

1.2 5. Conclusions

Thermodynamic and kinetic studies for the synthesis of solketal

0.9 25 ˚C in liquid phase were carried out in a well-controlled batch reactor
Moles of glycerol

in the presence of an acid catalyst (Amberlyst-35). The thermody-

35 ˚C namic equilibrium constant Kc at various temperatures ranging
from 293 to 323 K was determined. The reaction is exothermic
0.6 50 ˚C
and the standard enthalpy, entropy and Gibbs free energies at
298 K were found to be 30.1 ± 1.6 kJ mol1, 0.1 ± 0.01 kJ mol1
K1 and 2.1 ± 0.1 kJ mol1, respectively. The kinetic studies of
0.3 the same reaction demonstrated that the rate of the reaction in-
creased with increasing temperature, the catalyst addition amount
and acetone-to-glycerol (A/G) molar ratio. In this batch study of
0 the liquid phase reaction, pressure showed negligible influence
0 75 150 225 300 on the reaction thermodynamics and kinetics as expected, and no
Time (min) effect of the agitation speed on the reaction rate was observable
at >400 rpm. Langmuir–Hinshelwood model demonstrated to be
Fig. 10. Experimental and theoretical curves (Other conditions: 313 K, Wcat = 1 wt.% useful for describing the kinetic mechanism of the ketalization
of glycerol).
reaction of glycerol with acetone. Based on the Langmuir–Hinshel-
wood model, the values of the activation energy (Ea) of the overall
1.4 reaction was determined to be 55.6 ± 3.1 kJ mol1.
The future study of this project includes the development of a
continuous reactor for the synthesis of solketal from both pure
0.4 and crude glycerol using the thermodynamic and kinetic concepts
Kw where the recycle (recovery and reuse) of unreacted reagents could
be considered as an added advantage to the system.
ln K/Kd

-0.6
Acknowledgements
k

-1.6 The authors want to acknowledge the financial support

provided by Imperial Oil via University Research Award and the
Discovery Grant from NSERC for Dr. Xu. We are also thankful to
Professors Yasuo Ohtsuka and Guus Van Rossum from Akita
-2.6
University (Japan) and University of Twente (the Netherlands),
3.06 3.14 3.22 3.30 3.38
respectively, for their invaluable suggestions on some aspects of
1/T(x 10 3 K -1) this research.
Fig. 11. Plots of kinetic modeling parameters ln k or ln Kd vs. 1/T.
References

Table 4 [1] Johnson DT, Taconi KA. The glycerine glut: options for the value added
Kinetic modeling parameters k and Kwa. conversion of crude glycerol resulting from biodiesel production. Environ Prog
2007;26(4):338–48.
Temperature (K) k (L moles1 min1) Kw [2] Karinen RS, Krause AOI. New biocomponents from glycerol. Appl Catal A
298 0.112 2.650 2006;306(7):128–33.
303 0.158 1.498 [3] OECD-FAO Agricultural outlook 2011–2020; 2012. Available online at http://
www.oecd.org/document//9/0,3746,en_36774715_36775671_45438665_1_1_
308 0.239 1.090
1_1,00 [Retrieved March 9].
313 0.329 0.726
[4] Sims B. Clearing the way for by product quality. Biodiesel Magazine; 2011.
323 0.630 0.335
Available online at http://www.biodieselmagazine.com/articles/8137/
a clearing-the-way-for-byproduct-quality [Accessed online October 2011].
k = kinetic constant; Kw = equilibrium constant for water adsorption on the
[5] Deutsch J, Martin A, Lieske H. Investigation on heterogeneously catalysed
catalyst surface.
condensations of glycerol to cyclic acetals. J Catal 2007;245(2):428–35.
[6] Agirre I, Garcia I, Requies J, Barrio VL, Guemez MB, Canbra JF, et al. Glycerol
The temperature dependence of k and Kw can be given by the acetals, kinetics study of the reaction between glycerol and formaldehyde.
Arrhenius equations: Biomass Bioenergy 2011;35(8):3636–42.
[7] Silva VMTM, Rodrigues AE. Synthesis of diethylacetal: thermodynamic and
  kinetic studies. Chem Eng Sci 2001;56(4):1255–63.
Ea
k ¼ kr exp  ð15Þ [8] Barros AO, Faísca AT, Lachter ER, Nascimento RSV, Gil RASS. Acetalization of
RT hexanal with 2-ethyl hexanol catalyzed by solid acids. J Braz Chem Soc
2011;22(2):359–63.
  [9] Delfort B, Duran I, Jaecker A, Lacome T, Montagne X, Paille F. Etherification of
DH a glycerol with ethanol over solid acid catalysts. US patent; 2005: 6, 890, 364.
K w ¼ K a exp  ð16Þ
RT [10] Mota CJA, da Silva CXA, Rosenbach N, Costa JJ, da Silva F. Glycerin derivatives
as fuel additives: the addition of glycerol/acetone ketal (solketal) in gasolines.
where kr and Ka are the Arrhenius constants for Eqs. (15) and (16), Energy Fuels 2010;24(4):2733–6.
[11] Maksimov AL, Nekhaev AI, Ramazanov DN, Arinicheva YA, Dzyubenko AA,
respectively. Ea and DHa are the activation energy of the overall
Khadzhiev SN. Preparation of high-octane oxygenate fuel components from
reaction and enthalpy of the water adsorption reaction, respec- plant-derived polyols. Pet Chem 2011;51(1):61–9.
tively. The predicted values of k and Kw are presented as a function [12] Cablewski T, Faux AF, Strauss CR. Development and application of a continuous
of temperature in plots of ln k or ln Kw vs. 1/T in Fig. 11. From the microwave reactor for organic synthesis. J Org Chem 1994;59(12):3408–12.
[13] Krief A, Laurent P, Alexandre F. Diastereoselective synthesis of dimethyl
plots, the values of Ea and DHa were determined to be 55.6 ± 3.1 cyclopropane-1,1-dicarboxylates from a (c-alkoxyalkylidene) malonate and
and 64.7 ± 4.3 kJ mol1, respectively. sulfur and phosphorus ylides. Tetrahedron Lett 1998;39(11):1437–40.
 M.R. Nanda et al. / Fuel 117 (2014) 470–477 477

[14] Silva CXA, Goncalves VL C, Mota CJA. Water-tolerant zeolite catalyst for the [23] Yadav GD, Thathagar MB. Esterification of maleic acid with ethanol over
acetalisation of glycerol. Green Chem 2009;11:38–41. cation-exchange resin catalysts. React Funct Polym 2002;52(2):99–110.
[15] Roldan L, Mallada R, Fraile JM, Mayora JA, Menéndez M. Glycerol upgrading by [24] Bart HI. Kinetics of esterification of acetic acid with propyl alcohol by
ketalization in a zeolite membrane reactor. Asia-Pac J Chem Eng heterogeneous Catalysis. Int J Chem Kinet 1996;28(9):649–56.
2009;4(3):279–84. [25] Rat M, Zahedi-Noaki MH, Kaloaguine S. Sulfonic acid functionalized periodic
[16] Ferreira P, Fonseca IM, Ramos AM, Vital J, Castanheiro JE. Valorisation of mesoporous organosilicas as acetalization catalysts. Microporous Mesoporous
glycerol by condensation with acetone over silica-included heteropolyacids. Mater 2008;112(1–3):26–31.
Appl Catal B Environ 2010;98(1–2):94–9. [26] Izci A, Bodur F. Liquid-phase esterification of acetic acid with isobutanol
[17] Vicente G, Melero JA, Morales G, Paniagua M, Martín E. Acetalisation of bio- catalysed by ion-exchange resins. React Funct Polym 2007;67(12):1458–64.
glycerol with acetone to produce solketal over sulfonic mesostructured silicas. [27] Akbay E, Atiokka MR. Kinetics of esterification of acetic acid with n-amyl
Green Chem 2010;12:899–907. alcohol in the presence of Amberlyst-36. Appl Catal A 2011;396(1–2):14–9.
[18] Solmon G. Organic chemistry. 7th ed. New York: John Wiley & Sons; 2001. [28] Kaufhold M, Ei-Chawi MT. Process for preparing acetaldehyde diethyl acetal.
[19] da Silva CXA, Mota CJA. The influence of impurities on the acid-catalysed US Patent 1996; 5, 527,969.
reaction of glycerol with acetone. Biomass Bioenergy 2011;35:3547–51. [29] Zhou L, Nguyen TH, Adesina AA. The acetylation of glycerol over amberlyst-15:
[20] Héctor G, Ose IG. Solketal: green and catalytic synthesis and its classification kinetic and product distribution. Fuel Process Technol 2012;104(1):310–8.
as a solvent; 2,2-dimethyl-4-hidroxymethyl-1,3-dioxolane, an interesting [30] Morrison R, Boyd R. Organic Chemistry. 4th ed. Boston: Allyn and Bacon; 1983.
green solvent produced through heterogeneous catalysis, Chimica Oggi [31] Jemmy BR, Pandurangan A. AI-MCM-41 as an efficient heterogeneous catalyst
2008;26(3):10–2. in the acetalization of cyclohexanone with methanol, ethylene glycol and
[21] Graca NS, Pais LS, Silva VVMTM, Rodrigues AE. Oxygenated biofuels from pentaerythritol. J Mol Catal A Chem 2006;256(1–2):184–92.
butanol for diesel blends: synthesis of the acetal 1,1-dibutoxyethane catalysed [32] Xiao Y, Gao L, Xiao G, Lv J. Transesterification of reaction catalyzed by solid
by Amberlyst-15 ionexchange resin. Ind Eng Chem Res 2010;49(15):6763–71. base in a fixed bed reactor. Energy Fuels 2010;24(11):5829–33.
[22] Wade LGJ. Organic chemistry. 6th ed. New Jersey: Pearson Prentice Hall; 2006.
 Catalysis Today 254 (2015) 83–89

Contents lists available at ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

Amphiphilic property of niobium oxyhydroxide for waste glycerol

conversion to produce solketal
Talita E. Souza a , Izabela D. Padula a , Mariana M.G. Teodoro b , Poliane Chagas a ,
Jarbas M. Resende a , Patterson P. Souza b , Luiz C.A. Oliveira a,∗
a
Department of Chemistry, Federal University of Minas Gerais, Belo Horizonte CEP 31270-901, Brazil
b
Federal Center of Technological Education of Minas Gerais, Belo Horizonte CEP 30421-169, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Amphiphilic catalysts are synthesized using NbCl5 in the presence of the CTAB (cetyltrimethylammonium
Received 4 September 2014 bromide) to generate partial hydrophobicity to the catalysts. The partial hydrophobicity of the niobium
Received in revised form 3 December 2014 oxyhydroxides improved the acetalization reaction of a residual glycerol from biodiesel production by
Accepted 4 December 2014
decreasing the interaction between the water molecules and the acid sites of the catalyst. A waste glycerol
Available online 12 January 2015
conversion of 73% with a selectivity to solketal (2,2-dimethyl-[1,3]dioxolan-4-yl)methanol of 95% was
obtained. Many reuses of the catalysts showed glycerol conversions between 70 and 80%.
Keywords:
© 2015 Elsevier B.V. All rights reserved.
Niobium
Amphiphilic properties
Acetalization
Waste glycerol

1. Introduction In heterogeneous catalysis, different types of catalysts are used

in the acetalization reactions of glycerol, mesoporous silicates [13],
Many possibilities for use of the residual glycerin have been meso-structured silica functionalized with sulfonic acid [14], ZrO2
investigated, especially employing catalytic processes [1,2]. Cat- [15], Amberlyst resins [14] and oxides [8,15], among others.
alytic conversion reactions of glycerol, which is a highly reactive Modified catalysts based on niobium compounds, have received
molecule, favor obtaining chemical products with high added value, special attention because of their catalytic activity in several impor-
which can be obtained by reactions such as selective oxidation and tant chemical processes [16–20], as a promoter effect in oxidation
hydrogenolysis, dehydration, steam reforming, thermal reduction reactions [21,22], supporting effect in hydrogenation reactions
in synthesis gas, oligomerization/polymerization, selective trans- [23], metathesis [24], hydrotreating [25] and acid catalysis [26].
esterification, etherification and epichlorohydrin synthesis [3–6]. The acetalization reaction occurs through the condensation of a
Among these, we highlight the acetalization and etherification of glycerol molecule with one molecule of acetone, resulting in the lib-
glycerol to obtain fuel additives, which have a lower polarity and eration of water. The formed water can favor the reverse reaction,
viscosity, and therefore greater volatility, which can improve the reducing the formation of solketal [14,27]. Alternatives for these
combustion efficiency [2,7]. reactions have been the use of distillation processes [27] after the
In acetalization reactions, through the reaction of glycerol with reaction, or hydrophobic catalysts [28], which decrease the inter-
acetone, the products are cyclic acetals with 5 or 6 ring members, action between water molecules and the acid sites, without losing
with the solketal ((2,2-dimethyl-1.3-dioxolan-4-yl)methanol) as catalytic properties.
major product [8,9]. These ketals can act as additives in order to Niobium oxide shows a good water tolerance property as
increase the octane level of gasoline. reported by Okuhara [28], and exhibited the highest activity among
To obtain solketal, glycerol and acetone catalytic reactions have the solid acids. Nair et al. [8] shows the water tolerance of niobia
been studied using homogeneous and heterogeneous processes. In acid sites in acetalization reactions. Studies reported by Okuhara
homogeneous catalysis p-toluene sulphonic acid [10], hydrochloric [28] and Nair et al. [8], employ niobium oxide calcinated until
acid and sulfuric acid [11,12] have been employed. 773 K. Above 773 K this activity decreases the water tolerance,
thus it can be suggest that structural changes in the niobium
oxide occur [28]. In this work, we report the synthesis of a new
∗ Corresponding author. Tel.: +55 31 34096384; fax: +55 31 34095700. series of catalysts based on niobium oxyhydroxide obtained from
E-mail address: [email protected] (L.C.A. Oliveira). the precursor of niobium chloride and surface modification by

http://dx.doi.org/10.1016/j.cattod.2014.12.027
0920-5861/© 2015 Elsevier B.V. All rights reserved.
84 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89

partial hydrophobization with surfactants groups. Preliminary 100

S4
results from our research group have shown success in hydropho- 4.79%
bization processes of niobium-based catalysts with excellent S4-2D
95
5.87%
results in many types of reactions [29–32]. S4-SS
7.66% 7.14%
90
12.79%

Weight Loss (%)

2. Materials and methods
85
2.34%
2.1. Synthesis and characterization of the catalysts
80
The catalysts were synthesized, starting from the precursor of 13.94% 2.95%
niobium chloride (NbCl5 99%, purchased from Sigma–Aldrich). A 75
water:oil:surfactant system was prepared, consisting of a mix-
ture of Milli-Q water, 1-butanol (99%, Vetec), and 1-hexanol (99%, 70
Vetec) in proportions of 17, 22 and 61%, respectively. 41.2 mmol
CTAB (cetyltrimethylammonium bromide) surfactant was added 65
and after solubilization 46.2 mmol of (NbCl5 ) was added at 60 ◦ C 0 100 200 300 400 500 600
and constant stirring. A solution of NH4 OH (5 mol L−1 ) was then o
Temperature ( C)
added, dropwise, under constant stirring for 12 h (S4 catalyst) or
48 h (S4-2D catalyst). Finally, the resulting white solid is filtered, Fig. 1. Thermogravimetric analyses for catalysts based on niobium oxyhydroxide.
washed and dried at 60 ◦ C for 12 h.
A catalyst without surfactant was also synthesized by dissolv-
ing NbCl5 in 50% alcoholic solutions at 60 ◦ C, after the addition of oven temperature profile was 80 ◦ C for 5.0 min increasing to 100 ◦ C
NH4 OH until it reached pH 7, a solid white is precipitated. This at 3 ◦ C min−1 , maintaining the final temperature for 1 min.
is kept under stirring for 12 h, thus this is filtered, washed, and The calculation of the conversion, selectivity and products yield
oven-dried at 60 ◦ C for 12 h. This procedure is from catalyst S4-SS. were determined by the calibration curve with standard glycerol PA
The structural properties of materials were investigated (Modern Chemical, 99.5%) and solketal PA (provided by Petrobras).
by different characterization techniques. The thermal analysis
measurements were performed on a Shimadzu-TGA50H ther-
3. Results and discussions
mobalance at 700 ◦ C in air atmosphere and flow of 100 mL min−1 ,
with a heating rate of 10 ◦ C min−1 . Analyses of infrared spec-
3.1. Characterizations of the catalysts
troscopy (FTIR) were obtained using a Perkin Elmer RXI FTIR
spectrophotometer. The spectra of the samples were obtained in
Thermogravimetric analyses of the catalysts, Fig. 1, showed
the 4000–500 cm−1 region with resolution of 4 cm−1 and an aver-
three mass loss events, which are associated with the surface
age of 64 scan signal. Analysis of X-ray diffraction (XRD) was
hydroxyl groups above 110 ◦ C and after 200 ◦ C mass losses asso-
performed in the range between 2 = 20 and 70 in continuous
ciated with the surfactant group.
scan mode of 2.0◦ min−1 using k␣Cu radiation ( = 1.542 Å) in a
A mass loss of 14 and 3% was observed related to surfactant
Siemens D5000 diffractometer. Acidity of the materials was char-
group decomposition for S4 and S4-2D, respectively. These results
acterized by NH3 -TPD in a Quantachrome ChemBET-3000, TCD
indicate a lower degree of hydrophobization of the catalyst S4-2D
detector with 150 mA current and attenuation value of 32. N2
compared to S4. Above 450 ◦ C, the event observed in three sam-
adsorption/desorption isotherms were measured at 77 K using a
ples is associated with the loss of hydroxyl groups from the bulk,
Quantachrome Autosorb IQ2 equipment. Data were analyzed by
as suggested by some authors [31,33]. The mass loss for the cata-
BET method and pore size distribution using the BJH model. Scan-
lyst SS-S4 suggests that initially, the event with 5.87% loss is due
ning electron microscopy (SEM) obtained the morphology of the
to the humidity and the two following events are related to OH
catalysts in a Quanta 200 FEG-FEI scanning electron microscope,
loss (11.74%), from surface and bulk. These data suggests that the
operating at 30 kV with a resolution of 1.6 nm.
molecular formula of the S4-SS catalyst is NbO2 (OH), referred to as
niobium oxyhydroxide [32,34].
2.2. Catalytic reactions Analyses by infrared spectroscopy are shown in Fig. 2. The cat-
alyst S4 presented bands related to ␯C-H at 2922 and 2852 cm−1 ,
Typical acetalization reactions were performed at 70 ◦ C with which is a strong evidence of the presence of surfactant groups on
constant stirring in a batch reactor with an acetone/glycerol molar the surface of the catalyst.
ratio of 2/1, 0.2 g catalyst and 1 h of reaction. Residual glycerol from The S4-2D catalyst presented similar profile compared to S4-SS.
biodiesel production and a desalinated one were used in the acetal- The inset of Fig. 2, shows OH bond stretching observed at 3405 cm−1
ization reactions, both provided by Petrobras. NMR and GC-MS related to surface hydroxyls and at 3123 cm−1 due to the OH from
analyzed the conversion of glycerol and product identification. bulk, as reported by Schwertman and Cornell [32,34]. This fact gives
1 H, 13 C and DEPT, experiments, were performed at 20 ◦ C on an evidence of the formation of niobium oxyhydroxide. The presence
RMN spectrometer Avance DRX 200 at the High Resolution Mag- of the surfactant over the catalysts gives them a hydrophobic char-
netic Resonance Laboratory of UFMG. With a 5 nm multinuclear acter, which can contribute to the glycerol acetalization reactions.
probe and procedure to optimize the field homogeneity (shim- X-ray diffraction analyses not show diffraction patterns for cata-
ming), as well as probe tuning in the frequency of 1 H and 13 C, in lysts S4, S4-2D and S4-SS. The amorphous character of the catalysts
order to have optimal resolution and signal detection. are expected, because these were not calcinated during the synthe-
Analyses by GC–MS in an Agilent TechnologiesTM Model 7890A, sis to preserve the surfactant group, as surface and bulk hydroxyl
coupled to a mass detector model 5975C, were performed on HP- groups, which are characteristic of oxyhydroxides.
5MS (5% polymethylphenylsiloxane), 30 m × 250 ␮m × 0.25 ␮m The strength of the acid sites and total acidity were determined
column. Helium gas was used as carrier gas at a flow of by NH3 -TPD. The catalysts showed similar NH3 desorption profiles,
1.4 mL min−1 , split ratio of 130/1, injector temperature of 25 ◦ C and as shown in Fig. 3.
 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89 85

The catalyst without hydrophobization (S4-SS) showed a greater

area under the ammonia desorption curve. This result was
expected, since the hydrophobic group is attached to the catalyst
surface, replacing the Brønsted acid groups [35]. Using a zeolite as
a standard material, the number of acid sites could be calculated:
ν O-H
90.1, 63.3 and 58.5 ␮mol g−1 for S4-SS, S4-2D and S4, respectively.
Absorbance (arb. unit.)

ν O-H

Surface Bulk ν O-H The low acidity of the catalysts with highest degree of hydropho-
3600 3300 3000 2700 bization confirms the thermal analysis data.
The N2 adsorption and desorption isotherms and pore size dis-
tribution of the S4-SS, S4-2D and S4 catalysts are shown in Fig. 4.
The N2 adsorption isotherms are classified as Type IV according
to IUPAC, as reported by Kruk and Jaroniec [36], for materials
S4 with hydrophobic surfaces (surfactant-containing), however, type
S4-2D II isotherm is better to classified the S4 catalyst. The S4-2D and S3-
S4-SS SS catalysts have isotherm profiles very similar, this is due a similar
structure, since TG analyses show that hydrophobization was 3% for
4000 3500 3000 2500 2000 1500 1000 500
-1 S4-2D catalyst.
Wavenumber (cm )
The hysteresis may be classified as H3, for S4 catalysts, which
Fig. 2. Vibrational spectroscopy in the infrared region for the catalysts S4, S4-2D does not level off at relative pressures close to the saturation
and S4-SS. vapor pressure and is reported for materials comprised of aggre-
gated (loose assemblages) platelike particles forming slitlike pores.
The hysteresis in S4-2D and S4-SS catalyst is like H2 hysteresis,
S4 A = 21040 these achieve the level off at relative pressures close to the sat-
100 uration vapor pressure, the shape triangular, such as observed for
many porous inorganics oxides and assigned to porous connectivity
effect, as reported by Liu et al. [37].
The catalyst without CTAB (S4-SS) showed low BET-specific
Sinal TPD-NH3

S4-2D A= 22782
area value (135 m2 g−1 ). In the case of S4 and S4-2D, the presence
of surfactant leads to the formation of a material with relatively
100 high specific surface area (167 and 198 m2 g−1 for S4 and S4-2D,
respectively). The isotherm profile shows a continuous adsorption
of N2 throughout the range of relative pressures, and hysteresis are
observed during desorption. This result suggests the presence of
S4-SS
A= 32433 micro- and mesopores, as shown by the pore size distribution in
inset of Fig. 4.
100
The morphology of the materials was investigated using scan-
100 200 300 400 500 600 700
ning electron microscopy (Fig. 5). The S4 material synthesized using
NbCl5 as a precursor, exhibited a morphology distinct from that of
Temperatura ( °C)
the other materials with square openings that form a well-defined
porous structure, possibly generating a material with a high specific
Fig. 3. NH3 -TPD profile of the catalysts: S4, S4-2D and S4-SS.

200
1.5
3 -1

S4
Dv (log d)/cm g

1.0

160 S4-2D
0.5
S4-SS
Volume of N2/cm3 g-1

0.0
120 10 20 40 60 80100 200 400 600
Pore Diameter / Å

80

Área = 167 m²/g

40 2
198 m /g
2
135 m /g

0
0.0 0.2 0.4 0.6 0.8 1.0
P/Po

Fig. 4. N2 adsorption/desorption isotherms and pore size distribution of catalysts.

86 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89

Fig. 5. Micrographs obtained for the catalysts.

surface area. The S4-SS and S4-2D showed a change in morphology 100
9%
as agglomerated particles. Water Loss Crude Glycerin
16%
80
Desalinated Glycerin
3.2. Catalytic studies
Weight Loss (%)

Glycerol acetalization reactions were performed using different 60

types of glycerine from biodiesel production (crude and desali-
nated glycerol) and commercial glycerol. Usually the crude glycerol
contains between 2–3% fatty acid residues derived from the trans- 40
esterification reaction for biodiesel production, 6–8% of NaCl and
10–15% water. The glycerol amount in the crude is between 75 and
80%. 20

Thermogravimetric analyzes were performed to determine the <1%

6% Salt
salt and water amounts present in the waste glycerol utilized in
0
this work, in order to estimate the concentration of glycerol for the
0 100 200 300 400 500 600 700 800
calculation of conversion and yield, Fig. 6. o
Temperature ( C)
As expected, the amount of salt and water present in crude glyc-
erin was 6% and 16% of water, respectively. In the case of desalinated Fig. 6. Thermogravimetric analysis for crude and desalinated glycerin.
glycerin has 1% of salt and 9% of water. The amount of glycerol were
considered 78% and 90% for crude and desalinated glycerin, respec-
tively, which were used to calculate the conversion and yield of
the reactions. Fig. 7 shows data of the conversion of glycerol using The best results were obtained using desalinated glycerin that pre-
catalysts S4-SS, S4 and S4-2D with different types of glycerin. sented conversion values of 22, 43, 40 and 55% for S4, S4-2D, S4-SS
Fig. 7 presents an overview of glycerin (crude, desalinated and and HY-340, respectively. Acetalization reactions using commercial
commercial glycerol) conversion using the synthesized catalysts and crude glycerin, showed a similar conversion for the different
S4, S4-SS and S4-2D compared to commercial niobic acid, HY-340. catalyst. The low conversion presented using commercial and crude
 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89 87

70 100
S4-SS
S4
S4-2D
60 S4-2D
S4
S4-SS 80
HY340
HY-340
50

Glycerol conversion (%)

Glycerol Conversion (%)

60
40

30
40

20

20
10

0 0
1:2
1 1:4 1:6 2:2
Commercial glycerol Crude glycerin Desalinated glycerin
Molar ratio glycerol/acetone
Fig. 7. Conversion of acetalization reactions using the catalysts S4-SS, S4, S4-2D and
HY-340 at 70 ◦ C for 60 min, molar ratio glycerol/acetone 1/2. Fig. 8. Conversion of glycerin at molar ratios equal to 1/2, 1/4, 1/6 and 2/2 of glyc-
erol/acetone.

Table 1
TOF values obtained from acetalization reaction using desalinated glycerin with a
2/1 molar ratio of acetone (70 ◦ C for 1 h).
number of moles of solketal formed and the number of acid sites
Catalysts TOF (h−1 )
determined by NH3 -TPD (1 h reaction). The high TOF value for the
HY-340 380 S4-2D catalyst explains the high catalytic activity with desalinated
S4 618
glycerin.
S4-2D 1106
S4-SS 716 Reactions with different glycerol/acetone molar ratios (Fig. 8)
were performed. The conversion of glycerol showed a significant
increase when the glycerol/acetone molar ratio was 1/4 (73% con-
glycerol is probably due to the presence of NaCl (crude glycerol) and version) for 1 h of reaction using the catalyst S4-2D. Other studies
the high viscosity of commercial glycerol [38]. have shown the influence of the 1//1, 1/2 and 1/4 glycerol/acetone
The S4-2D catalyst showed better conversion even under high molar ratio, and the results were the best with 1/4, with greater
viscosity conditions of the commercial glycerol and in the presence than 90% conversion after 5 h of reaction using catalysts based on
of large amounts of salt (crude glycerin). The high specific area and carbon functionalized with acid groups [39]. In the literature, it
number of acid sites can explain these results. Table 1 shows the can be found that the conversion with an acetone/glycerol molar
TOF (turnover frequency), which is the number of catalytic cycles ratio (1/1) using Amberlyst-15 resin and K-10 Montmorillonite was
in the active center of the catalyst per time unit, obtained from the approximately 60% after 60 min of reaction [2]. For mesoporous

Fig. 9. NMR spectra of reaction 1/4 molar ratio glycerol/acetone with S4-2D catalyst.
88 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89

Scheme 1. Acetalization reaction of glycerol with acetone under acid catalysis to form as the major product (2,2-dimethyl-[1,3]-dioxolan-4-yl)methanol.

100 interface, reducing the interference of water molecules on the cat-

alyst surface.
The different textural properties such as surface area and acid-
80 ity obtained in partially hydrophobized catalysts and with only
hydrophilic surface showed good conversion results, about 77%
Glycerol Conversion (%)

depending on the reaction conditions studied, and 95% selectivity

60 for solketal formation.
In reactions with a 1/4 glycerol/acetone molar ratio, the glycerol
conversion was still greater than 80% after some reaction cycles.
40 The results demonstrated sustainable use of desalinated residual
glycerin, production of solketal, and its application as fuel additives.

20 Acknowledgments

The authors thank Petrobras for funding, as well as the

0 Microscopy Center – UFMG, the infrastructure of the Chemistry
Reaction Resuse 1 Reuse 2 Reuse 3
Department of UFMG and CEFET. The authors also thank FAPEMIG,
S4-2D Yields (%) S4-2D HY-340 Yelds (%) HY-340 CNPQ and CAPES for financial support.

Fig. 10. Conversion and yield of the reuse of the catalysts S4-2D and HY-340 with
References
glycerol/acetone molar ratio of 1/4, at 70 ◦ C, 1 h.

[1] L.C.A. Oliveira, M.F. Portilho, A.C. da Silva, H.A. Taroco, P.P. Souza, Appl. Catal. B
silicates with a 1/1 molar ratio, there were conversions ranging 117–118 (2012) 29–35.
from 28 to 52% for reactions at 80 ◦ C 6 h [13]. [2] C.X.A. da Silva, V.L. Gonçalves, C.J.A. Mota, Green Chem. 11 (2009) 38–41.
[3] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C.D. Pina, Angew. Chem. Int. Ed.
In acetalization reactions the products formed are 5-membered 46 (2007) 4434–4440.
rings (2,2-dimethyl-[1,3]-dioxolan-4-yl)methanol, called solketals, [4] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 5253–5277.
and a 6-membered ring (2,2-dimethyl-1,3-dioxan-5-ol). However, [5] J. Barrault, F. Jerome, Eur. J. Lipid Sci. Technol. 110 (2008) 825–830.
[6] A. Behr, J.P. Gomes, J. Eur, Lipid Sci. Technol. 112 (2010) 31–50.
the selectivity of formation of each of these products depends on [7] N. Viswanadham, S.K. Saxena, Fuel 103 (2013) 980–986.
the acid strength of the catalyst, the higher the acid strength, the [8] G.S. Nair, E. Adrijanto, A. Alsalme, I.V. Kozhevnikov, D.J. Cooke, D.R. Brown, N.R.
higher the selectivity for solketal [2,8]. The reaction for obtaining Shiju, Catal. Sci. Technol. 2 (2012) 1173–1179.
[9] M.J. Climent, A. Corma, A. Velty, Appl. Catal. A263 (2004) 155–161.
solketal is shown in Scheme 1. [10] Organic Syntheses III, Collective Volume, Wiley & Sons, New York, 1955.
The selectivity presented by S4-2D catalyst after 1 h of reaction [11] N. Suriyaprapadiloka, B. Kitiyanana, Energy Procedia 9 (2011) 63–69.
was 95%; as reported by Souza et al. [32], the other products were [12] A. Krief, L. Provins, A. Froidbise, Tetrahedron Lett. 39 (1998) 1437–1440.
[13] L. Li, T.I. Koranyi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012) 1611–1612.
not identified by NMR and GC–MS, like shown in Fig. 9, only solketal [14] G. Vicent, J. Melero, G. Morales, M. Paniagua, E. Martín, Green Chem. 12 (2010)
has formed. 899–907.
In order to study the stability of the catalysts the reuse of mate- [15] P.S. Reddy, P. Sudarsanam, B. Mallessham, G. Raju, B.M. Reddy, J. Ind. Eng. Chem.
17 (2011) 377–381.
rials using desalinated glycerin was studied. The results are shown
[16] M. Ziolek, Catal. Today 78 (2003) 47–64.
in Fig. 10. [17] K. Tanabe, Catal. Today 78 (2003) 65–77.
The reuse of the catalysts was tested in many subsequent reac- [18] M. Ziolek, I. Sobczak, I. Nowak, P. Decyk, A. Lewandowsk, J. Kujawa, Microporous
Mesoporous Mater. 35–36 (2000) 195–207.
tions using desalinated glycerin (Fig. 10). The conversion remained
[19] I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603–3624.
constantly at approximately 70%, indicating the high stability of [20] K. Tanabe, S. Okazaki, Appl. Catal. A 133 (1995) 191–218.
the hydrophobized catalyst S4-2D. Interestingly, in addition to the [21] N. Marin-Astorga, J.J. Martínez, D.N. Suarez, J. Cubillos, H. Rojas, C.A. Ortiz, Curr.
catalyst remaining active, only the formation of solketal occurs, Org. Chem. 16 (2012) 2797–2801.
[22] H.S. Oliveira, L.D. Almeida, V.A.A. de Freitas, F.C.C. Moura, P.P. de Souza, L.C.A.
showing that the material maintains its properties in the acetal- Oliveira, Catal. Today 240 (2015) 176–181.
ization process. The yield of the reaction is greater than 80% for [23] T. Iizuka, Y. Tanaka, K. Tanabe, J. Mol. Catal. 17 (1982) 381–389.
S4-2D catalyst in the first two cycles of reuse. On the other hand, [24] R. Buffon, et al., J. Mol. Catal. 76 (1992) 287–295.
[25] C. Geantet, J. Afonso, M. Breysse, N. Allali, M. Danot, Catal. Today 28 (1996)
HY-340 catalyst presented 40% of yield in the second reuse. 23–30.
[26] J. Souza, P.M.T.G. Souza, P.P. de Souza, D.L. Sangiorge, V.M.D. Pasa, L.C.A. Oliveira,
4. Conclusions Catal. Today 213 (2013) 65–72.
[27] J.S. Clarkson, A.J. Walker, M.A. Wood, Org. Process Res. Dev. 5 (2001) 630–635.
[28] T. Okuhara, Chem. Rev. 102 (2002) 3641–3666.
In this work niobium based catalysts with high catalytic activity [29] P. Chagas, H.S. Oliveira, R. Mambrini, M. Le Hyaric, M.V. de Almeida, L.C.A.
due to the specific surface area and acidic properties were syn- Oliveira, Appl. Catal. A 454 (2013) 88–92.
[30] L.C.A. Oliveira, H.S. Oliveira, G. Mayrink, H.S. Mansur, A.A.P. Mansur, R.L. Mor-
thesized and characterized. The higher hydrophobization decrease eira, Appl. Catal. B 147 (2014) 403–412.
the number of acid sites of the catalysts as shown by NH3- [31] N.T. Prado, F.G.E. Nogueira, A.E. Nogueira, C.A. Nunes, R. Diniz, L.C.A. Oliveira,
TPD. However, the hydrophobic groups act at the glycerol/acetone Energy Fuels 24 (2010) 4793–4796.
 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89 89

[32] T.E. Souza, M.F. Portilho, P.M.T.G. Souza, P.P. Souza, L.C.A. Oliveira, Chem- [36] M. Kruk, M. Jaroniec, S.H. Ryoo, Chem. Mater. 12 (2000) 1414–1421.
CatChem 6 (2014) 2961–2969. [37] H. Liu, N.A.J.J. Seaton, Colloid Interface Sci. 156 (1993) 285–293.
[33] H. Landmesser, et al., Solid State Ion. 271 (1997) 101–103. [38] C.X.A. da Silva, C.J.A. Mota, Biomass Bioenergy 35 (2011) 3547–3551.
[34] R.M. Cornell, U. Schwertmann, The Iron Oxides, VCH, New York, 1996. [39] R. Rodrigues, M. Gonçalves, D. Mandelli, P.P. Pescarmona, W.A. de Carvalho,
[35] A. Corma, V. Fornes, M.T. Navarro, J. Perez-Pariente, J. Catal. 148 (1994) Catal. Sci. Technol. 4 (2014) 2293–2301.
574–769.
 Applied Catalysis A: General 496 (2015) 32–39

Contents lists available at ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

Design of highly efficient Zn-, Cu-, Ni- and Co-promoted M-AlPO4

solid acids: The acetalization of glycerol with acetone
Siqian Zhang, Zhenbo Zhao ∗ , Yuhui Ao ∗
Changchun University of Technology, Jilin Province Key Laboratory of Carbon Fiber Development and Application, Changchun 130012, China

a r t i c l e i n f o a b s t r a c t

Article history: Acetalization of glycerol with acetone was examined using M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni, or Co) solid
Received 25 August 2014 acid catalysts with enlarged surface areas. The relationship between the structural properties and the
Received in revised form 28 January 2015 catalytic performances was investigated. The synthesized catalysts were systematically analyzed using
Accepted 4 February 2015
various techniques, namely, XRD, SEM/TEM, EDX, BET surface area, pore size distribution, FTIR, UV–vis
Available online 24 February 2015
spectrophotometry and NH3 -TPD. The XRD results suggested the formation of mesostructured amor-
phous materials. The SEM/TEM results showed that these materials had an ordered structure. All of the
Keywords:
materials exhibited uniform pore size and high specific surface area. The strength of the acid sites of the
Glycerol
Acetalization M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni, or Co) solid acid catalysts was studied by NH3 -TPD. UV–vis spectropho-
Mesostructured tometry was used as a novel method to measure the exact amount of acid sites of the M-AlPO4 /xAlPO4
Solid acids catalyst (x = Zn, Cu, Ni, or Co) solid acid catalysts. The resulting materials were found to exhibit excellent catalytic
activity for the acetalization of glycerol with acetone. The reaction with M-NiAlPO4 showed the highest
yield (75.44%) and the best selectivity (75.12%).
© 2015 Elsevier B.V. All rights reserved.

1. Introduction di- and tri-acetyl glycerol esters [4], halogenation to yield 3-

hydroxypropionaldehyde [5], dehydration to yield acrolein [6], and
Over recent years, increasing attention has been devoted to the acetalization with aldehydes and/or ketones to yield glycerol acetal
problem of the conversion of natural resources for the reduction and solketal [7,8].
of energy consumption and exhaust emissions [1]. Biodiesel is an Among these reactions, the acetalization of glycerol with
important type of renewable energy, which presents the advan- ketones in the presence of an acid catalyst is important for
tages of enhanced biodegradation, reduced toxicity and a lower the synthesis of five- and six-membered cyclic products, which
emission profile. are (2,2-dimethyl-1,3-dioxolan-4-yl)-methanol and 2,2-dimethyl-
Biodiesel is considered a methyl ester produced by the trans- 1,3-dioxan-5-ol, respectively. Primarily, the products of the
esterification of vegetable oils and animal fats in the presence of acetalization of glycerol with ketones are widely used as bases, fla-
an acidic or basic catalyst. For every 9 kg of biodiesel produced, vors, scents, and additives in several industries. For example, the
approximately 1 kg of crude glycerol is formed as the main by- ketal of glycerol is a component in the diesel, biodiesel and gaso-
product. Therefore, the crude glycerol becomes a widely available line formulation that effectively enhances the cold properties and
raw material at a low cost. However, the crude glycerol cannot be reduces the viscosity [9].
directly applied to the pharmaceutical and food industries due to Green catalysts have been utilized in the chemical system to
its contaminant level. address the challenging environmental and economic problems.
Glycerol is an abundant raw material for the synthesis of Gemma Vicente [10] and coworkers have found that sulfonic
chemical intermediates, solvents, and additives, among others. acid-functionalized mesostructured silicas had excellent catalytic
Accordingly, certain catalytic reactions have been studied, such as behavior in two-step batches in the acetalization of glycerol with
reduction to obtain 1,3-propanediol, 1,2-propanediol [2], ether- acetone to produce (2,2-dimethyl-1,3-dioxane-4-yl)-methanol.
ification to produce polyglycerols [3], esterification to obtain The acetalization of glycerol has been researched by Ferreira [11]
and his coworkers using heteropolyacids immobilized on silica as
the catalyst. The reaction selectivity to solketal reached approxi-
∗ Corresponding author. Tel.: +86 0431 85716671. mately 78%. Faria et al. [12] have discussed a thermodynamic and
E-mail addresses: [email protected] (Z. Zhao), kinetic reaction of the acetalization of glycerol with a variety of
[email protected] (Y. Ao). catalysts. The conversion of glycerol can reach approximately 90%

http://dx.doi.org/10.1016/j.apcata.2015.02.006
0926-860X/© 2015 Elsevier B.V. All rights reserved.
 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39 33

using Amberlyst-15 as the catalyst. Baithy Mallesham et al. [13] 2.2. Characterization
have developed a green heterogeneous process to synthesize solke-
tal from the acetalization of glycerol over SnO2 -based solid acids; Powder X-ray diffraction (XRD) measurements were performed
the best catalyst was MO3 /SnO2 . Khayoon and Hameed [14] have by a diffractometer (RigaKu D/max-2550) with a Cu Ka radia-
developed the acetalization of glycerol by a mesoporous 5% Ni – tion source from 0.02◦ to 5.0◦ (small angle) and 10.0◦ to 80.0◦
1% Zr/AC catalyst through a solventless process. Under the opti- (wide angle). The measurements were conducted using a cur-
mal reaction conditions, glycerol was converted to the final product rent setting of 20 mA for small angle XRD and 40 mA for wide
containing 76% and 24% of five- and six-membered ring acetals. angle XRD, with a step size of 0.02◦ . The morphology of the cat-
Moreover, researchers have discovered that the use of hetero- alytic materials was obtained by a scanning electron microscope
geneous catalysts, including heteropolyacids [15], ion exchange (SEM: JSM-5600, JEOL) and a transmission electron microscope
resins (Amberlyst-15, Amberlyst-36) [16,17] and zeolites (HSZ- (TEM: JEM-2000EX, JEOL). BET surface areas, pore volumes, and
36, ZSM-5) [18], is an excellent choice because of simple product average pore sizes of the samples were obtained from Nitro-
separation and catalyst recycling. However, zeolites possess dif- gen adsorption–desorption measurements at −196 ◦ C using a
fusion problems relative to the reactants within the microporous Micromeritics ASAP2020 system. The samples were pretreated
network caused by transfer resistance, whereas heteropolyacids under vacuum at 200 ◦ C for 2 h before the tests. The average
suffer from certain disadvantages, such as low thermal stability, pore size data were calculated using the Barrett–Joyner–Halenda
separation problems from reaction mixtures and solubility. The (BJH) method over an absorption and desorption period. Fourier
modified metal oxide is a potential heterogeneous catalyst because transform infrared (FT-IR) spectra were recorded on KBr pellets
of its many excellent characteristics, such as good thermal stability, by a Nicolet (6700) infrared spectrometer with the wavenum-
renewability and strong acidity. ber from 4000 to 500 cm−1 . Temperature programmed desorption
Among the modified metal oxides, AlPO4 has attracted con- of ammonia (NH3 -TPD) was performed on a FINESORB 3010. In
siderable attention [19,20]. However, the low effective surface a typical experiment for the TPD measurement, 0.05 g M-AlPO4
area and the structural collapse of the material are the primary was pretreated at 300 ◦ C for 2 h in helium and then cooled to
disadvantages [21,22]. Since the discovery of mesoporous silica 100 ◦ C. A flow of 2.5% NH3 /He was then passed over the pretreated
MCM-41 [23] with large specific surface areas, large pore volumes materials for 60 min. Following this ammonia adsorption proce-
and uniform pore sizes, different types of mesoporous materi- dure, the reactor was purged with helium for 30 min to remove
als, especially many non-siliceous mesoporous materials [24–26], residual/physisorbed ammonia. The sample was then heated to
have been widely utilized in catalytic fields [27–30]. These ordered 900 ◦ C at a rate of 10 ◦ C/min in 20 cm3 /min helium, and the
mesoporous materials have many free sites accessible to the reac- ammonia desorption signals were continuously recorded by the
tant molecules. It is expected that M-AlPO4 can be synthesized with thermal conductivity detector (TCD). The amount of surface acid
these prominent advantages. The catalytic properties of samples in the samples was quantitatively measured via UV spectrome-
are also controlled by the introduction of a heteroatom. Therefore, try with adsorbates of pyridine (Py) in cyclohexane solution. The
we introduce the Zn, Cu, Ni or Co atom into M-AlPO4 . UV–visible diffuse reflectance spectra were recorded on a Cary
To date, most of the synthesis methods are time consuming and 5000 UV–Vis–NIR. The detailed steps are as follows: (1) Py was
involve multiple steps with small amines and bulky organosilane dissolved in cyclohexane to obtain 0.632 ml/l(C0 ) solution. (2)
surfactants [31]. Choi and Ryoo [32] have synthesized mesoporous 0.05 g sample was transferred into the Py solution, which was
aluminophosphates with crystalline microporous frameworks by then placed in an ultrasonic oscillator for 2 h. (3) The Py solu-
the addition of organosilane surfactants into the conventional tion was measured using a UV–vis spectrophotometer at 251 nm.
synthesis composition. In this study, we attempted the synthe- The concentration of Py (C) was calculated based on a standard
sis of M-AlPO4 through a simple one-step method using a single curve.
template. The resultant materials with an ordered mesoporous
structure are applied to the acetalization of glycerol with acetone. Qi = (C0 − C) × V/(m × 1000)
Therefore, we synthesized the novel M-AlPO4 green solid acid.
These catalysts were characterized via certain methods, such
as X-ray diffraction (XRD), ammonia-temperature programmed where C0 (ml/l) and C (ml/l) are the concentration of Py before and
desorption (NH3 -TPD), Fourier Transform infrared spectroscopy after the absorption, respectively, V (ml) is the volume of the Py
(FTIR), nitrogen adsorption–desorption measurements, UV–vis solution into which the sample was transferred, and m (g) is the
spectrophotometry, transmission electron microscopy (TEM), and mass of the sample.
scanning electron microscopy (SEM). The results revealed that the
modified M-AlPO4 showed good catalytic ability.
2.3. Acetalization of glycerol with acetone

2. Experimental Glycerol acetalization was performed in a round bottom flask

with a reflux condenser. Into a reactor was placed the solid acid
2.1. Preparation of the catalysts catalyst (0.2 g), glycerol (98%, Aldrich) and acetone with a molar
ratio of glycerol to acetone of 1:8 (glycerol, 5 g; acetone, 25.2 g).
In a typical procedure for the synthesis of M-xAlPO4 (x = Zn, Cu, The mixture was magnetically stirred at 80 ◦ C for 1 h. A portion of
Ni, or Co), 4 g (CH3 )4 NOH·5H2 O was dissolved in 20 ml methanol. the reaction mixture was separated after the reaction time by cen-
When (CH3 )4 NOH was completely dissolved, 6 g Al [OCH·(CH3 )2 ]3 , trifugalization, which was then analyzed by a GC equipped with a
4 g (CH3 CHCH2 )4 NBr (TPABr), 3.2 g NH4 H2 PO4 and 0.005 mol BP-20 (wax) capillary column and an FID. The glycerol conversion
x(NO3 )2 (x = Zn, Cu, Ni or Co) were added into the above solution and ketal yield were determined based on the ketal/glycerol area
with vigorous stirring. The final solution was stirred for 1 h. The ratio. M-AlPO4 was reused five times to examine the catalyst’s recy-
transparent solution mixture was washed with an ample amount clability. After each catalytic run, M-AlPO4 was recovered from the
of ethanol and dried at 150 ◦ C for 24 h. Finally, the white powder reaction solution by filtration, washed with ethanol, and calcined
mixture was calcined at 500 ◦ C for 5 h in a muffle oven to remove at 300 ◦ C in air for 2 h to remove the strongly adsorbed organic
the template (TPABr). species.
34 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39

particles became more regular, smoother and smaller in the image

of M-NiAlPO4 . The morphology, as observed in the SEM images,
also showed a significant difference between the as-synthesized
M-AlPO4 and the M-AlPO4 calcined at 500 ◦ C. After being calcined
at 500 ◦ C, the size of the M-AlPO4 particles increased because of the
aggregation of the M-AlPO4 particles. In contrast, neither a struc-
tural nor a morphological change was observed over M-NiAlPO4
after being calcined at 500 ◦ C. These findings indicated that Ni-
doping inhibited M-AlPO4 sintering and particle aggregation, most
likely by stabilizing the M-AlPO4 network; thus, the initial textural
structure can be better preserved compared with that of M-AlPO4 .
This result is in good agreement with the characterization of the
BET surface area. The EDS of M-NiAlPO4 is shown in Fig. 2h in
which the exclusive peaks of the Ni, Al, P and O elements were
obviously observed, implying that all of the elements were suc-
cessfully introduced into the wall of the pores.

3.1.3. TEM analysis

The catalytic materials were characterized using the TEM
technique. The transmission electron microscopy (TEM) images
(Fig. 3) revealed a sponge-like structure. The average particles were
obtained after calcination at 500 ◦ C. The size of the particles in the
M-NiAlPO4 sample was similar to the other samples, but the linkage
between the different particles became more numerous.

3.1.4. BET measurement

Nitrogen adsorption–desorption isotherms and the average
Fig. 1. Small-angle X-ray diffraction of M-AlPO4 (red data plot), M-ZnAlPO4 (green pore size distribution curves of the M-AlPO4 /xAlPO4 (x = Zn, Cu,
data plot), M-CuAlPO4 (blue data plot), M-NiAlPO4 (cyan data plot), and M-CoAlPO4 Ni, or Co) samples are shown in ESI Figs. 2–6. The nitrogen
(magenta data plot). (For interpretation of the references to color in this figure adsorption–desorption isotherms of M-AlPO4 /xAlPO4 (x = Zn, Cu,
legend, the reader is referred to the web version of the article.)
Ni, or Co) were characterized as a mixed type (a combination of
types I, II, and IV), according to the IUPAC classification. The doped
3. Results and discussion heteroatom eventually did not change the shape of the isotherms
of the samples. The hysteresis of M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni, or
3.1. Characterization of the catalyst Co) was associated with the occurrence of the well-defined cap-
illary condensation. The tendency for the capillary condensation
3.1.1. XRD patterns steps of the isotherm to shift to higher relative pressures suggested
The low angle XRD peaks exhibited by M-AlPO4 /xAlPO4 (x = Zn, that the pore size increased with the addition of the various het-
Cu, Ni, or Co) are presented in Fig. 1, which reveals the presence of eroatoms [35,36]. A sharp average pore size distribution curve of
mesoporous structures [33]. The low angle reflection peaks shifted M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni, or Co) was presented. The sharp
to a lower 2 region (from 0.1588 to 0.1077) with the addition of distribution curve reflected the presence of a uniform pore struc-
the various heteroatoms, indicating a larger pore size and unit cell ture in the mesoporous materials. The textural parameters obtained
parameter. This fact proved the beneficial role of the heteroatom from the N2 adsorption isotherms are summarized in ESI Table 1.
toward M-AlPO4 , which can be further demonstrated by the BET Generally, the specific surface areas of the mixed oxides were con-
measurement. siderably larger than those of the individual oxides. There was a
The wide-angle XRD patterns of the different catalytic materi- gradual increase in the BET surface area (from 35.5 to 143.1 m2 /g)
als, including the M-AlPO4 /M-xAlPO4 (x = Zn, Cu, Ni, or Co) samples, after immobilization of the heteroatom, which was beneficial for
are shown in ESI Fig. 1. The lack of clear diffractions in the M- catalytic activity [37]. Similar trends were also observed in the
AlPO4 sample suggested the presence of amorphous materials [34]. alteration of the pore size (from 5.97 to 9.25 nm) and pore vol-
The absence of a crystalline state of M-AlPO4 may be due to the ume (from 0.10 to 0.42 cm3 /g). The specific surface area reached a
high calcination temperature for removal of the template. There maximum of 143.1 m2 /g after doping of the Ni atom into M-AlPO4 .
was also no obvious effect on the diffraction peaks after adding High thermal treatment is essential to achieve better physi-
the various heteroatoms. This result confirmed that the framework cochemical properties during the preparation of M-AlPO4 /xAlPO4
structure of M-xAlPO4 (x = Zn, Cu, Ni, or Co) was consistent with that (x = Zn, Cu, Ni, or Co). However, pure metal oxide typically under-
of the M-AlPO4 material. In addition, the lack of clear diffractions in goes agglomeration and sintering at high temperatures, eventually
the wide-angle range suggested that the heteroatom species were leading to the smaller specific surface area. In contrast, the het-
highly dispersed on the mesoporous surface, even after calcination eroatom is immobilized in the pure metal oxide, and the produced
at 500 ◦ C. mixed oxide exhibits significant resistance toward agglomeration
and sintering due to the cooperative nature of the cations of the
3.1.2. SEM/EDX analysis mixed oxide. This process can lead to a higher BET specific surface
To obtain the structure and morphology of the different cata- area. Most catalytic reactions occur on the surface of the catalysts,
lysts, the SEM characterizations were performed. Fig. 2a–e shows and the reaction rate is typically in proportion to the surface area
the SEM images of M-AlPO4 /M-xAlPO4 (x = Zn, Cu, Ni, or Co), and of the catalysts. Therefore, it can be expected that the enhanced
Fig. 2f–g indicates the SEM images of the as-synthesized M-AlPO4 surface area of M-NiAlPO4 is one of the primary reasons for achiev-
and M-NiAlPO4. Therefore, it was found that tight surfaces were ing a better catalytic activity in the acetalization of glycerol with
formed in M-AlPO4 /M-xAlPO4 (x = Zn, Cu, Ni, or Co). Moreover, the acetone.
 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39 35

Fig. 2. SEM images of (a) M-AlPO4 , (b) M-ZnAlPO4 , (c) M-CuAlPO4 , (d) M-NiAlPO4 and (e) M-CoAlPO4 and TEM images of (f) as-synthesized M-AlPO4 and (g) as-synthesized
M-NiAlPO4 ; EDX measurement of (h) M-NiAlPO4 .
36 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39

Fig. 3. TEM images of (a) M-AlPO4 , (b) M-ZnAlPO4 , (c) M-CuAlPO4 , (d) M-NiAlPO4 and (e) M-CoAlPO4 .

The specific surface area increased in the following order: the template. ESI Fig. 8 displays the FTIR spectra of the pure and
M-AlPO4 < M-ZnAlPO4 < M-CuAlPO4 < M-CoAlPO4 < M-NiAlPO4 . heteroatom-doped M-AlPO4 samples. M-AlPO4 /xAlPO4 (x = Zn, Cu,
This result proves that M-NiAlPO4 may be an efficient catalyst, Ni or Co) showed several infrared absorption bands at approx-
corresponding to the following reaction results. imately 1110, 1650, and 3500 cm−1 . The broad bands at 3500
and 1650 cm−1 were assigned to the OH stretching and bending
3.1.5. FTIR measurement of the water molecule associated with phosphorus that was per-
The FTIR spectra of the as-synthesized and calcined M-NiAlPO4 turbed by a hydrogen bridge bond from a surface hydroxyl band
are shown in ESI Fig. 7. Compared with the as-synthesized sam- [38,39]. The bands at 1110 and 468 cm−1 were attributed to the
ple, the absorption bands at 3010, 1500 and 1360 cm−1 , which triply degenerate P O stretching vibrations and the triply degen-
were assigned to the C H stretching vibration of TPABr, disappear erate O P O bending vibrations of tetrahedral PO43− , respectively
after the treatment at 500 ◦ C, indicating the complete removal of [40].
 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39 37

Fig. 5. Influence of the species of catalysts on acetalization. (Reaction conditions:

glycerol, 5 g; acetone, 25.2 g; reaction time = 1 h; catalyst amount = 0.2 g; reaction
temperature = 80 ◦ C).

which can increase the number of surface-exposed Ni atoms. This

phenomenon indicated that doping the Ni atom into the sam-
ples was a good choice, and the mixed oxide showed substantially
stronger acidic properties than the single oxide.

3.2. Catalytic activities of various catalysts in the acetalization of

glycerol

3.2.1. The catalytic performances of different catalysts

The prepared samples were tested for the acetalization of glyc-
erol with acetone at 80 ◦ C with the molar ratio of glycerol to acetone
of 1:8 (glycerol, 5 g; acetone, 25.2 g) for 1 h. The reaction profiles
using 0.2 g of catalyst are shown in Fig. 5. It was found that the addi-
tion of the heteroatom influenced the catalytic activity. The yields
of ketal were 57.33%, 66.24%, 69.37%, 75.44% and 67.20% when
M-AlPO4 , M-ZnAlPO4 , M-CuAlPO4 , M-NiAlPO4 and M-CoAlPO4
were employed as the catalysts, respectively. The yield increased
Fig. 4. NH3 -TPD profiles of M-AlPO4 (red data plot), M-ZnAlPO4 (green data plot),
M-CuAlPO4 (blue data plot), M-NiAlPO4 (magenta data plot), and M-CoAlPO4 (cyan
in the following order: M-AlPO4 < M-ZnAlPO4 < M-CuAlPO4 < M-
data plot). (For interpretation of the references to color in this figure legend, the CoAlPO4 < M-NiAlPO4 , corresponding to the specific surface area.
reader is referred to the web version of the article.) The highest selectivity of solketal (75.12%) was obtained by M-
NiAlPO4 .
Because the lifetime of the catalysts strongly affects the cost of
3.1.6. TPD measurement the overall process, the reusability and stability of the catalyst are
NH3 acts as an excellent probe molecule for the examination of key importance for practical applications. Thus, M-NiAlPO4 was
of the acidic properties of a heterogeneous solid acid. The strong recycled five times under the optimized reaction conditions. There
basicity and small molecular size of NH3 facilitates the simple was no significant decline in catalytic activity. Fig. 6 reveals that
detection of acidic sites present in the narrow pores of the samples M-NiAlPO4 showed a continual reduction in activity during mod-
[41]. The NH3 -TPD profiles of the M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni or erate cycling, and M-NiAlPO4 showed 80% loss of its initial activity
Co) materials are depicted in Fig. 4 to evaluate the acidity of the after five catalytic cycles (ketal conversion: 76.12% → 17.78%). The
catalysts [42]. The temperature of the desorption peak increased, considerable reduction in activity was attributed to the substantial
and the area of the desorption peak enhanced due to the addition loss of acid sites during the repeated use of the catalyst.
of the heteroatom into the M-AlPO4 sample. M-NiAlPO4 exhib- M-NiAlPO4 is a promising solid acid catalyst in terms of glycerol
ited a desorption peak at approximately 280 ◦ C (corresponding to conversion and selectivity of the products. A primary reason for the
weak acid sites) and a desorption peak distributed at approximately catalytic ability of M-NiAlPO4 is that the enlarged surface area can
650 ◦ C (corresponding to strong acid sites). The characteristic TPD increase the number of surface-exposed Ni atoms, thus affording an
patterns of the curve of the other samples showed two or three increased number of catalytically active acid sites. Another possible
obvious peaks at different temperatures due to the existence of reason is the improved accessibility of organic reactants (glycerol
weak, medium and strong acid sites. This result indicated that dif- and acetone) to the active acid sites [44–49].
ferent acidic sites were present in M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni
or Co) [43]. The calculated amounts of acid sites via the UV spec- 3.2.2. Optimized reaction conditions for the acetalization of
trometry measurement are shown in ESI Table 1, which provided glycerol
sharply increasing acid amounts of M-AlPO4 /xAlPO4 . M-NiAlPO4 The optimized reaction conditions for the acetalization of glyc-
(120 ␮mol/g), providing at least two times higher acid amounts erol, such as the catalyst amount, reaction temperature and molar
compared with M-AlPO4 (59 ␮mol/g). The highest acid amount of ratio of glycerol to acetone, were investigated by simple variable
M-NiAlPO4 may be due to the enlargement of the surface area, methods (Figs. 7 and 8). Initially, the reaction was performed with
38 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39

Fig. 6. Reusability of M-NiAlPO4 (reaction conditions: glycerol, 5 g; acetone, 25.2 g; Fig. 9. Influence of the catalyst amount on acetalization over the M-NiAlPO4 cata-
reaction time = 1 h; catalyst amount = 0.2 g; reaction temperature = 80 ◦ C). lyst. (Reaction conditions: glycerol, 5 g; acetone, 25.2 g; reaction time = 2 h; reaction
temperature = 80 ◦ C).

different temperatures ranging from 60 ◦ C to 90 ◦ C, which played

a dominant role for achieving a high catalytic efficiency. The yield
of ketal initially increased with the increasing reaction tempera-
ture. The maximum yield (75.96%) was obtained with the reaction
temperature up to 80 ◦ C. Moreover, using the catalyst under the
solvent-free condition for the experiment, we investigated the
appropriate molar ratio of glycerol to acetone to further optimize
the reaction conditions. With the increase of the molar ratio of glyc-
erol to acetone from 1:2 to 1:8, the yield of ketal increased from
39.36% to 76.19%. Moreover, the effect of the catalyst amount on
the catalytic performance was investigated, as shown in Fig. 9. This
reaction can reflect the production capacity of the catalyst. Con-
ducting the reaction with an amount higher than 0.2 g of catalyst
had no significant effect on the yield. However, when the amount
of the catalyst was reduced to 0.15 g and 0.10 g, the yields of ketal
were reduced to 55.44% and 45.79%, respectively. Therefore, 0.2 g
of catalyst provided the best result for the model reaction, result-
ing in a conversion of 75.39%. Finally, the effect of the reaction time
Fig. 7. Influence of reaction temperature on acetalization over the M-NiAlPO4 cata- was also considered in the acetalization of glycerol. The reaction
lyst. (Reaction conditions: glycerol, 5 g; acetone, 25.2 g; reaction time = 1 h; catalyst time was examined from 0.5 h to 2.0 h. The catalyst activities plot-
amount = 0.2 g). ted against the reaction time showed typical bell curves; obviously,

Fig. 8. Influence of the molar ratio of glycerol to acetone on acetalization over the M-
NiAlPO4 catalyst. (Reaction conditions: catalyst amount = 0.2 g; reaction time = 2 h; Fig. 10. Influence of the reaction time on acetalization over the M-NiAlPO4 catalyst.
reaction temperature = 80 ◦ C). (Reaction conditions: glycerol, 5 g; acetone, 25.2 g; catalyst amount = 0.2 g; reaction
temperature = 80 ◦ C).
 S. Zhang et al. / Applied Catalysis A: General 496 (2015) 32–39 39

the highest yield of ketal (74.14%) was obtained at the reaction time [14] M.S. Khayoon, B.H. Hameed, Appl. Catal. A: Gen. 464–465 (2013) 191–199.
of 1.0 h. The reaction result may be because the rates of the reverse [15] I.V. Kozhevnikov, Chem. Rev. 98 (1998) 171–198.
[16] B. Delfort, I. Durand, A. Jaecker, T. Lacome, X. Montagne, F. Paille, U.S. Patent,
reactions increased with a reaction time longer than 1.0 h (Fig. 10). 2003, 163, 949.
[17] J. Deutsch, A. Martin, H. Lieske, J. Catal. 245 (2007) 428–435.
4. Conclusions [18] C.X. Da Silva, V.L. Goncalves, C.J. Mota, Green Chem. 11 (2009) 38–41.
[19] P. Sreenivasulu, D. Nandan, M. Kumar, N. Viswanadham, J. Mater. Chem. A 1
(2013) 3268–3271.
In summary, the M-AlPO4 /xAlPO4 (x = Zn, Cu, Ni, or Co) cata- [20] Z. Miao, L. Xu, H. Song, H. Zhao, L. Chou, Catal. Sci. Technol. 3 (2013) 1942–
lysts were synthesized by a simple one-step method using a single 1954.
[21] M. Tiemann, M. Schulz, C. Jäger, M. Fröba, Chem. Mater. 13 (2001) 2885.
template. The obtained catalytic materials were characterized by [22] L. Wang, B. Tian, J. Fan, X. Liu, H. Yang, C. Yu, B. Tu, D. Zhao, Microporous
XRD, SEM/TEM, EDX, BET surface area, pore size distribution, FTIR, Mesoporous Mater. 67 (2004) 123.
UV–vis spectrophotometry and NH3 -TPD techniques. These cata- [23] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature 359 (1992)
710–712.
lysts had high thermal stabilities, regular mesoporous structures
[24] P. Yang, D. Zhao, D.I. Margolese, B.F. Chemlka, G.D. Stucky, Nature 396 (1998)
and high surface areas. Furthermore, acetalization of glycerol with 152–155.
ketones was successfully performed with M-AlPO4 /xAlPO4 (x = Zn, [25] F. Schüth, Chem. Mater. 13 (2001) 3184–3195.
[26] B. Tian, X. Liu, B. Tu, C. Yu, J. Fan, L. Wang, S. Xie, G.D. Stucky, D. Zhao, Nat. Mater.
Cu, Ni, or Co). The catalyst amount, reaction temperature, reaction
2 (2003) 159–163.
time and molar ratio of glycerol to acetone were further optimized. [27] A. Taguchi, F. Schüth, Microporous Mesoporous Mater. 77 (2005) 1–45.
The recycling experiment showed that the catalyst remains active [28] Q. Yuan, Q. Liu, W.G. Song, W. Feng, W. Pu, L. Sun, Y. Zhang, C. Yan, J. Am. Chem.
after being used five times. Therefore, these mesoporous solid cat- Soc. 129 (2007) 6698–6699.
[29] L. Xu, H. Song, L. Chou, ACS Catal. 2 (2012) 1331–1342.
alysts were efficient for the conversion of glycerol to ketal, which [30] L. Xu, H. Song, L. Chou, Catal. Sci. Technol. 1 (2011) 1032–1042.
will be useful for the future biomass utilization. [31] M. Choi, R. Srivastava, R. Ryoo, Chem. Commun. (2006) 4380.
[32] M. Choi, R. Srivastava, R. Ryoo, Chem. Commun. (2006) 4380–4382.
[33] N.S. Kabisatpathy, J. He, L.A. Lee, J. Rong, L. Yang, G. Sikha, B.N. Popov, T.S. Emrick,
Appendix A. Supplementary data T.P. Russell, Q. Wang, Nano Res. 2 (2009) 474–483.
[34] A.V. Biradar, S.B. Umbarkar, M.K. Dongare, Appl. Catal. A 285 (2005) 190–195.
Supplementary data associated with this article can be [35] M. Kruk, M. Jaroniec, A. Sayari, Langmuir 13 (1997) 6267–6273.
[36] M. Kruk, M. Jaroniec, Y. Sakamoto, O. Terasaki, R. Ryoo, C.H. Ko, J. Phys. Chem.
found, in the online version, at http://dx.doi.org/10.1016/j.apcata. B 104 (2000) 292–301.
2015.02.006. [37] L. Katta, P. Sudarsanam, G. Thrimurthulu, B.M. Reddy, Appl. Catal. B 101 (2010)
101–108.
[38] H.N. Kim, S.W. Keller, T.E. Mallouk, J. Schmitt, G. Decher, Chem. Mater. 9 (1997)
References 1414–1421.
[39] B. Tyagi, K.B. Sidhpuria, B. Shaik, R.V. Jasra, J. Porous Mater. 17 (2009) 699–709.
[1] G.D. Yadav, P.A. Chandan, D.P. Tekale, Ind. Eng. Chem. Res. 50 (2011) 8558–8570. [40] (a) K. Mtalsi, T. Jei, M. Montes, S. Tayane, J. Chem. Technol. Biotechnol. 76 (2001)
[2] R.S. Karinen, A.O.I. Krause, Appl. Catal. A: Gen. 306 (2006) 128–133. 128;
[3] J.M. Clacens, Y. Pouilloux, J. Barrault, Appl. Catal. A: Gen. 227 (2002) 181–190. (b) J.M. Campelo, M. Jaraba, D. Luna, R. Luque, J.M. Marinas, A.A. Romero, J.A.
[4] M.S. Khayoon, B.H. Hameed, Appl. Catal. A: Gen. 443/434 (2012) 152–161. Navio, M. Macias, Chem. Mater. 15 (2003) 3352.
[5] H. Noureddini, U.S. Patent, 2001, 501, 6174. [41] B.M. Reddy, G.K. Reddy, K.N. Rao, L. Katta, J. Mol. Catal. A: Chem. 306 (2009)
[6] A. Alhanash, E.F. Kozhevnikova, L.V. Kozhevnikov, Appl. Catal. A: Gen. 378 62–68.
(2010) 11–18. [42] X. Ma, J. Gong, S. Wang, F. He, X. Yang, G. Wang, G. Xu, J. Mol. Catal. A: Chem.
[7] A. Sinhamahapatra, N. Sutradhar, M. Ghosh, H.C. Bajaj, A.B. Panda, Appl. Catal. 218 (2004) 253–259.
A: Gen. 402 (2011) 87–93. [43] M.V. Landau, L. Titelman, L. Vradman, P. Wilson, Chem. Commun. (2003)
[8] H. Serafim, I.M. Fonseca, A.M. Ramos, Chem. Eng. J. 178 (2011) 291–296. 594–595.
[9] A. Jaecker-Voirol, I. Durand, G. Hillion, B. Delfort, X. Montagne, Oil Gas Sci. [44] C.-L. Chen, S. Cheng, H.-P. Lin, S.-T. Wong, C.-Y. Mou, Appl. Catal. A: Gen. (2001)
Technol. 63 (2008) 395–404. 21–30.
[10] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martín, Green Chem. 12 [45] C.-L. Chen, T. Li, S. Cheng, H.-P. Lin, C.J. Bhongale, C.-Y. Mou, Microporous Meso-
(2010) 899–907. porous Mater. 50 (2001) 201–208.
[11] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Appl. Catal. B: [46] X.-R. Chen, Y.-H. Ju, C.-Y. Mou, J. Phys. Chem. C 111 (2007) 18731–18737.
Environ. 98 (2010) 94–99. [47] S. Liu, H. Zhang, X. Meng, Y. Zhang, L. Ren, F. Nawaz, J. Liu, Z. Li, F.-S. Xiao,
[12] R.P.V. Faria, C.S.M. Pereira, V.M.T.M. Silva, J.M. Loureiro, A.E. Rodrigues, Ind. Eng. Microporous Mesoporous Mater. 136 (2010) 126–131.
Chem. Res. 52 (2013) 1538–1547. [48] Y. Du, Y. Sun, Y. Di, L. Zhao, S. Liu, F.-S. Xiao, J. Porous Mater. 13 (2006) 163–171.
[13] B. Mallesham, P. Sudarsanam, G. Raju, B.M. Reddy, Green Chem. 15 (2013) [49] L. Fuxiang, Y. Feng, L. Yongli, L. Ruifeng, X. Kechang, Microporous Mesoporous
478–489. Mater. 101 (2007) 250–255.
 Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54

Contents lists available at ScienceDirect

Journal of Molecular Catalysis A: Chemical

journal homepage: www.elsevier.com/locate/molcata

Room temperature synthesis of solketal from acetalization of glycerol

with acetone: Effect of crystallite size and the role of acidity of beta
zeolite
Pandian Manjunathan, Sanjeev P. Maradur, A.B. Halgeri, Ganapati V. Shanbhag ∗
Materials Science Division, Poornaprajna Institute of Scientific Research (PPISR), Bidalur Post, Devanahalli, Bangalore 562110, Karnataka State, India

a r t i c l e i n f o a b s t r a c t

Article history: Room temperature synthesis of solketal from acetalization of glycerol with acetone was carried out
Received 21 July 2014 over various types of Brønsted solid acid catalysts in the liquid phase. Among the catalysts screened,
Received in revised form H-Beta zeolite showed the best performance in less time period with 86% glycerol conversion and 98.5%
16 September 2014
selectivity to solketal. The chemical and structural properties of modified and unmodified beta catalysts
Accepted 21 September 2014
Available online 30 September 2014
were studied by X-ray diffraction, AAS, SEM, NH3 -TPD and FTIR-pyridine adsorption. The H-Beta catalyst
with lower crystallite size gave better conversion and solketal selectivity compared to H-Beta with higher
crystallite size. The effect of acidity of the catalyst on acetalization of glycerol was studied by modified
Keywords:
H-Beta zeolite beta catalysts of varying acidities. Glycerol conversion decreased with decrease in total acidity of beta
Crystallite size catalysts. Strong to weak acidity ratio of the catalysts was found to have a direct correlation with catalyst
Acidity performance.
Glycerol © 2014 Elsevier B.V. All rights reserved.
Solketal

1. Introduction commercial interest to design green and chemoselective catalysts

for these processes. Due to the presence of three hydroxyl groups,
Due to medium-term depletion of fossil fuel reserves and grow- glycerol exhibits versatile and high chemical reactivity, and can be
ing scarcity of fossil hydrocarbon, researchers around the world converted into value-added chemicals and fuel additives [1–4].
are motivated to search for an alternative source of energy. Experts Glycerol reacts with ketone, aldehyde and alcohol forming
considered biomass as the only sustainable source of future energy ketals, acetals and ethers respectively. These products can be used
and it has the potential to displace the petroleum feedstock. as fuel additives which improve the quality of diesel by reduc-
Among which vegetable oil is one such source used for the syn- ing the emission of carbon monoxide and unregulated aldehydes
thesis of biodiesel from transesterification reaction. Glycerol is [5]. Glycerol reacts with benzaldehyde to produce 5-hydroxyl-2-
obtained as a byproduct during transesterification of vegetable phenyl-1,3-dioxane which can be used as a starting material for
oil which accounts for one tenth of every gallon of biodiesel pro- the production of 1,3-propanediol [6]. The ketalization of glycerol
duced. Increasing availability of glycerol has made it an inexpensive with acetone yields solketal and acetal. Solketal can be used as a
and abundant raw material to synthesize value-added chemicals cold flow improver to enhance cold weather performance of diesel
and hence it is a challenge for researchers to utilize this waste fuel which also reduces its viscosity. Solketal blended with regular
glycerol to transform into commercially important products. The gasoline in 1, 3 or 5 vol% decreases the gum formation and also it
major direct utilities of this compound are in cosmetics, per- enhances the octane number up to 2.5 points [7,8].
sonal care products and soaps. However, these sectors may not be Heterogeneous catalysts are preferred for these reactions due
able to consume tons of glycerol produced from biodiesel plants. to several advantages over conventional homogeneous catalysts
Moreover, it is well known that glycerol can be converted into like high recyclability, thermal stability and process friendly
chemicals by catalytic processes namely, acetalization, oxidation, nature. Among which, zeolites [9], supported metal oxides [10],
hydrogenation, esterification, etherification, transesterification, hydrotalcite [11], sulphonic functionalized resins and mesoporous
dehydration, oligomerization and pyrolysis. Hence, there is a great silica [9,12] are used for glycerol transformation reactions. Zeo-
lite is an aluminosilicate which plays an important role in
petrochemical industries due to its acidic sites, high thermal sta-
∗ Corresponding author. Tel.: +91 27408552; fax: +91 23619034. bility and shape selectivity toward desired molecules with tunable
E-mail address: [email protected] (G.V. Shanbhag). pore size [13]. Among different forms of zeolite, nanocrystalline

http://dx.doi.org/10.1016/j.molcata.2014.09.028
1381-1169/© 2014 Elsevier B.V. All rights reserved.
48 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54

zeolite plays a vital role in catalytic reactions because of high sur- calcined at 540 ◦ C for 4 h under static air at a heating rate of ramp
face area, lower diffusion path length and more exposed active 5 ◦ C min−1 . The obtained sample was labeled as Cu/H-Beta [19].
sites. Nanocrystalline zeolite beta and Y showed higher activity The Na form of beta catalyst was prepared by stirring 10 g
than microcrystalline zeolite beta and Y toward palm oil crack- of H-Beta-1 with 100 ml of 0.5 M aqueous sodium nitrate solu-
ing for the production of biofuels [14]. N-alkylation of amines tion at 80 ◦ C for 4 h, followed by filtration, washing with distilled
with alcohols over nanosized zeolite beta was found to be more water and drying at 120 ◦ C for 2 h. The above procedure was
active than bulk beta due to shorter path length in nanosized form repeated to convert the protonic form into Na-form of beta zeo-
[15]. lite. Furthermore, copper ion exchange of Na-Beta was carried
Acetalization of glycerol with acetone was studied with various out by stirring 5 g of Na-Beta with 75 ml of 0.1 M aqueous cop-
acid catalysts such as amberlyst-15, montmorillonite K-10, zeolite per nitrate trihydrate solution at 80 ◦ C for 4 h and then it was
beta, HUSY, zeolite ZSM-5, p-toluenesulfonic acid [9], SO3 H-SBA-15 filtered, washed with distilled water and dried at 120 ◦ C for 2 h.
[12], sulphonated carbon-silica composite [16] and silica-induced The above procedure was repeated twice for maximum copper ion
heteropolyacids [17]. However, these catalysts are reported to be exchange. The obtained catalyst was calcined at 540 ◦ C for 4 h under
active at reaction temperatures higher than 65 ◦ C. Among them, static air at a heating rate of 5 ◦ C min−1 and was designated as
amberlyst-15 was found to be more active than all the other solid Cu-Beta.
acid catalysts because of the presence of sulphonic functionality
and high amount of acidity. Silva et al. reported acetalization of 2.2.3. Synthesis of H-Beta zeolite of higher crystallite size
glycerol with acetone using beta zeolite at 70 ◦ C with high cata- H-Beta zeolite of higher crystallite size (Beta-2) was synthesized
lyst concentration of 37.5 wt% with respect to glycerol which gave under template free condition with modification of the reported
glycerol conversion of 95% in 60 min [9]. method [20]. The synthesis was carried out in a 100 ml teflon-lined
The aim of this work is to explore a suitable solid acid catalyst stainless-steel autoclave with the following procedure. A 0.18 g of
for room temperature synthesis of solketal from acetalization of sodium aluminum oxide was dissolved in 37.8 ml of distilled water,
glycerol with acetone. Solid acid catalysts differing in properties followed by the addition of 1.56 g of sodium hydroxide. After stir-
(nature of acidity, porosity, etc.) are screened to find out an effi- ring for 30 min, 3.6 g of fumed silica was added to the above solution
cient catalyst for this reaction. Two decisive catalyst properties viz. followed by the addition of 12.2 ml distilled water into the mixture
crystallite size and acidity of zeolite beta are correlated with cat- with continued stirring. After stirring for 10 min, 0.18 g of H-Beta-1
alytic activity. Reaction parameters such as catalyst concentration zeolite was added into the mixture for seeding and stirred for 5 min.
and mole ratio are also optimized. The resulted mixture was then sealed and allowed to crystallize in
an oven at 120 ◦ C for 7 days. The obtained solid product was filtered,
2. Experimental washed with distilled water and dried at 80 ◦ C for 12 h. The obtained
Na-form of beta zeolite was converted to H-form by ion exchange
2.1. Materials treatment using 0.5 M NH4 NO3 (20 ml of 0.5 M NH4 NO3 solution/g
of zeolite) under reflux condition for 12 h. Then the sample was
Glycerol, acetone, methanol, isopropanol, sodium nitrate, filtered, washed with distilled water, dried at 80 ◦ C and calcined
sodium hydroxide and copper nitrate trihydrate were purchased at 450 ◦ C for 5 h at a heating rate of 1 ◦ C min−1 . The ion exchange
from Merck India Ltd. NH4 -Beta zeolite (SAR = 25) was obtained procedure was further repeated twice to remove the sodium ions
from Nankai, China (here onwards Beta-1). NH4 -ZSM-5 zeolite present in beta zeolite. The obtained sample was labeled as H-Beta-
(SAR = 23) and H-Y zeolite (SAR = 30) were purchased from Zeolyst 2.
International. H-Mordenite zeolite (SAR = 16) was kindly donated Other Brønsted solid acid catalysts like 10 wt% molybdenum
by Sud-Chemie India Pvt. Ltd. Montmorillonite-K10, tetraethy- oxide supported on silica (MoO3 /SiO2 ) and cesium salt of phos-
lorthosilicate and solketal (as a GC standard) were purchased from photungstic acid Cs2.5 H0.5 PW12 O40 (CsHPW) were prepared from
Sigma–Aldrich. Amberlyst-15, ammonium molybdate and fumed the reported literature [21,22].
silica were purchased from Alfa Aesar. Oxalic acid was purchased
from Nice Chemicals, Kochi, India. 2.3. Catalyst characterization

2.2. Catalyst preparation Powder X-ray diffraction patterns of unmodified and modified
beta catalysts were recorded with Bruker D2 phaser X-ray diffrac-
The ammonium form of zeolites were calcined in flowing air at tometer using Cu K␣ radiation ( = 1.542 Å) with high resolution
540 ◦ C for 4 h with the ramp rate of 5 ◦ C min−1 to convert ammo- Lynxeye detector. All the samples were scanned in the 2 range
nium form into protonic form. The obtained samples were labeled of 5–60◦ with step size of 0.02◦ /s. Crystallinity of unmodified and
as H-Beta-1, H-ZSM-5 and H-Mordenite. modified beta catalysts was evaluated from the integrated intensity
of the signal at 2 = 22.4◦ [23].
2.2.1. Preparation of dealuminated H-Beta-1 zeolite The Brønsted and Lewis acidic sites of unmodified and modi-
Mild dealumination was performed by stirring 4 g of H-Beta-1 fied beta catalysts were investigated by pyridine adsorption study
with 80 ml of two different concentrations (0.01 M and 0.05 M) of using FT-IR (alpha-T, Bruker). The self supported wafers of the cat-
oxalic acid solution at room temperature for 1 h and the samples alysts were prepared by a pellet press instrument. The wafer was
were filtered, washed with distilled water and dried at 120 ◦ C for then calcined at 500 ◦ C for 1 h, cooled and placed in a desiccator to
12 h. Then the dealuminated samples were calcined in flowing air maintain moisture free condition. Then the samples were saturated
at 540 ◦ C for 4 h at a heating rate of 1 ◦ C min−1 [18]. The obtained with pyridine and heated at 150 ◦ C for 1 h to remove physisorbed
samples were labeled as H-Beta-1-A and H-Beta-1-B respectively. pyridine. FTIR spectra were recorded in absorbance mode in the
wavelength range from 1400 to 1600 cm−1 . The spectrum obtained
2.2.2. Preparation of copper ion exchanged Beta-1 zeolite after pyridine treatment was subtracted with that of pyridine
The copper ion exchanged H-Beta-1 zeolite was prepared by untreated sample to get the peaks only due to pyridine–acid inter-
refluxing 4 g of H-Beta-1 with 60 ml of 0.1 M aqueous copper nitrate action.
trihydrate solution for 12 h. It was then filtered, washed with dis- The acidity of unmodified and modified beta catalysts were
tilled water and dried at 120 ◦ C for 12 h. The obtained zeolite was determined by temperature programmed desorption of ammonia
 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54 49

(NH3 -TPD) equipped with thermal conductivity detector using

helium as a carrier gas. In all the experiments, 150 mg of sample
was out-gassed at 550 ◦ C for 1 h in flowing helium (25 ml min−1 )
and then cooled down to 100 ◦ C. At 100 ◦ C, the sample was sat-
urated with 10% of ammonia/helium stream for 45 min. After the
saturation of ammonia, the sample was flushed with helium for 1 h
at 100 ◦ C to remove physisorbed ammonia, and then desorption of
ammonia was recorded with a temperature program from 100 to
600 ◦ C with a heating rate of 8 ◦ C min−1 .
To study the characteristic properties of the catalyst H-Beta-1
after the reaction, the catalyst was filtered, washed with methanol
to remove adsorbed reactants and products. The catalyst was dried
and calcined at 540 ◦ C before its characterization by XRD and NH3 -
TPD.
Elemental analysis of unmodified and modified beta catalysts
were determined by Perkin Elmer AAnalyst 200 atomic absorption
spectrophotometer (AAS). The solution for AAS analysis was pre-
pared by digesting 100 mg of zeolite powder with 1 ml of HCl (36%),
0.5 ml HF (48%), 1 ml of HNO3 (70%) and 4 ml of distilled water. After Fig. 1. XRD patterns of unmodified and modified beta catalysts.
digestion, 2 g of boric acid was added to neutralize HF and the solu-
tion was diluted with distilled water to desired concentration to
analyze the elements [24]. concentrations of elemental contents (silica/alumina and copper)
Scanning electron microscope (SEM) images of H-Beta-1 and H- are listed in Table 1. The SAR increased from 25 to 45 with increase
Beta-2 catalysts, were recorded on Zeiss microscope to investigate in the concentration of dealuminating agent (0–0.05 M). Copper
the crystallite size and morphology. content in Cu-Beta (3 ion exchanges with Na-Beta) was 1.8 times
higher compared with Cu-H/Beta (1 time ion exchange with H-
Beta-1).
2.4. Catalytic activity studies

The catalytic reactions were carried out in liquid phase batch 3.3. Pyridine FT-IR
reactor at room temperature. In a typical experiment, the required
amounts of glycerol (27 mmol) and acetone (54 mmol) were taken The nature of acidic sites in unmodified and modified Beta-1
in 25 ml glass reactor along with preactivated catalyst (5 wt% was investigated by pyridine-FTIR (Fig. 2). The relative concentra-
referred to glycerol mass). The reaction mixture was magneti- tions of Brønsted and Lewis acidic sites of catalysts were calculated
cally stirred at room temperature (28 ◦ C) for 1 h. Samples were from the intensity of PyH+ and PyL bands (1545 and 1450 cm−1
withdrawn at certain intervals of time, dissolved in methanol and respectively) and the B/L ratios obtained from the intensity mea-
centrifuged to separate out the catalyst from the liquid phase. The surements are listed in Table 2. Dealuminated H-Beta-1 catalysts
obtained liquid was analyzed by gas chromatography (Thermo sci- showed a relative increase in B/L ratio with increase in dealu-
entific Trace GC 700) equipped with a capillary column (0.25 mm mination. B/L ratio of dealuminated H-Beta-1-A and H-Beta-1-B
I.D. and 30 m length, Stabilwax, Restek) and flame ionization detec- are found to be 1.3 and 1.7 times higher compared with parental
tor. All the products were confirmed by gas chromatography using H-Beta-1 respectively. Mild dealumination with oxalic acid also
the product standard. The glycerol conversion and selectivity to removed extra-framework, Lewis acidic Al sites which led to the
solketal were determined from calibration lines obtained from increase in B/L ratio. Pyridine-FTIR study of copper ion exchanged
standards in wt%. The product yield was calculated by the GC anal- beta catalysts reveals that B/L ratio decreases with increase in
ysis using the formula, the degree of copper ion exchange. Cu/H-Beta showed higher B/L
ratio (0.47) compared to Cu-Beta (0.30) indicating the presence
glycerol converted (wt%) × selectivity (wt%)
Product yield (wt%) = . of higher amount of Lewis acidic sites in the latter than in the
100
former.

3. Results and discussion 3.4. NH3 -TPD

3.1. X-ray diffraction The NH3 -TPD profiles of unmodified and modified H-Beta are
depicted in Fig. 3. Weak and strong acidity of the catalysts are
XRD patterns of unmodified and modified beta catalysts showed quantified by deconvolution of peaks and are listed in Table 2.
a characteristic peak of beta zeolite at 2 = 22.4◦ and confirmed that
all the samples were highly crystalline with a marginal decrease in
Table 1
crystallinity after dealumination and copper ion exchange (Fig. 1). Characterization of unmodified and modified beta catalysts by AAS and XRD
Dealumination of H-Beta-1 resulted in decrease of crystallinity techniques.
from 100 to 85% with increase in the concentration of dealuminat-
Catalyst SiO2 /Al2 O3 a Cua (mmol/g) Crystallinityb
ing agent. Crystallinity of parent H-Beta-1 and Na-Beta catalysts
decreased by ∼10% after Cu ion exchange as determined based on H-Beta-2 14.2 – 100
H-Beta-1 25 – 100
the decrease in peak intensity of 2 = 22.4◦ (Table 1).
H-Beta-1-A 30 – 92
H-Beta-1-B 45 – 85
3.2. Atomic absorption spectroscopy Cu/H-Beta – 0.16 90
Cu-Beta – 0.30 91
Silica alumina ratios (SAR) of unmodified and modified beta cat- a
Calculated from AAS results.
b
alysts were determined by atomic absorption spectroscopy. The Calculated from XRD patterns.
50 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54

H-Beta-1 resulted in decrease of total acidity with decrease

in strong acidic sites as expected. The strong to weak acidity
ratio decreased from 1.5 to 1.2 with increase in dealumination.
Total acidity of H-Beta-1 decreased after Cu2+ ion exchange. The
strong acidic sites in Cu-Beta were 1.3 times higher compared
with Cu/H-Beta. TPD profile of Cu/H-Beta and Cu-Beta catalysts
showed a shift in ammonia desorption peak toward left (lower
temperature) compared to H-Beta-1. It could be due to the
weak Lewis acidic Cu2+ present in Cu/H-Beta and Cu-Beta cata-
lysts.

3.5. Microscopic analysis

SEM analysis of H-Beta-1 and H-Beta-2 revealed the crys-

tallite size and morphology as shown in Fig. 4. The crystallite
size of H-Beta-1 varied from 60 to 200 nm with an average
crystallite size of 135 nm with spherical morphology. H-Beta-2
crystallite size was in the range of 200–800 nm with an average
crystallite size of 450 nm with irregular morphology. The crys-
tallite size of H-Beta-1 was found to be 3.3 times smaller than
H-Beta-2.
Fig. 2. Pyridine-FTIR spectra of unmodified and modified beta catalysts.

3.6. Catalytic activity study

3.6.1. Catalyst screening

Acetalization of glycerol was carried out with acetone to
glycerol mole ratio of 2:1 at room temperature over different
types of Brønsted solid acid catalysts namely, ion exchange resin
(amberlyst-15), zeolites (medium and large pore zeolites like H-
ZSM-5, H-Mordenite, H-Y, H-Beta), clay (Montmorillonite K-10),
MoO3 /SiO2 and cesium salt of phosphotungstic acid (Fig. 5). Prior
to catalytic activity studies, blank run was carried out without a cat-
alyst, which resulted in very low glycerol conversion (0.01%). All the
screened catalysts were compared with their performance after 1 h
reaction time. Stronger Brønsted acidic catalyst, amberlyst-15 gave
73.1% glycerol conversion with 91% solketal selectivity whereas
acidic montmorillonite K-10 clay exhibited low glycerol conver-
sion of 27.3% with 89% selectivity for solketal. Stronger Brønsted
acidic catalyst, cesium salt of phosphotungstic acid showed glycerol
conversion of 67.3% with high solketal selectivity of 99% whereas
MoO3 /SiO2 showed 46.8% glycerol conversion with 90% selectivity
for solketal.
Fig. 3. Temperature programmed desorption profile of unmodified and modified Medium pore zeolite, H-ZSM-5 and dual pore zeolite, H-
beta catalysts.
mordenite resulted in low glycerol conversion (19.5 and 8%
respectively) which could be attributed to diffusion limita-
Acidity measurement revealed that the total acidity of H-Beta-2 tion due to pore size restriction. Interestingly, mordenite gave
was 1.2 times higher than H-Beta-1 zeolite. Higher acidity in H- lowest selectivity for solketal (80%) with considerable forma-
Beta-2 could be attributed to higher aluminum content compared tion of a six membered cyclic acetal (Scheme 1). Zeolite H-Y
to other beta catalysts. Weak acid sites in H-Beta-2 were found showed higher glycerol conversion (74.2%) and high solketal
to be 1.5 times higher compared with H-Beta-1. Dealumination of selectivity (98.2%) compared to medium pore zeolites which

Table 2
Acidity measurement of unmodified and modified beta catalysts measured by pyridine-FTIR and NH3 -TPD.

Catalyst Pyridine-FTIR NH3 -TPD

Brønsted Lewis acidity B/L (IB /IL ) Strong acidity (S) Weak acidity (W) Total acidity S/W
acidity (IB ) (IL ) (mmol
NH3 g−1 )

Peak max Amount Peak max Amount

temp (◦ C) (mmol NH3 g−1 ) temp (◦ C) (mmol NH3 g−1 )

H-Beta-1 0.39 0.28 1.39 364 0.90 212 0.61 1.51 1.48
H-Beta-2 – – – 360 0.87 197 0.93 1.80 0.94
H-Beta 1-A 0.41 0.23 1.78 361 0.73 198 0.58 1.31 1.26
H-Beta 1-B 0.35 0.15 2.33 371 0.60 202 0.50 1.10 1.20
Cu/H-Beta 0.27 0.57 0.47 314 0.52 191 0.56 1.08 0.93
Cu-Beta 0.26 0.87 0.30 318 0.68 192 0.64 1.32 1.06
 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54 51

Fig. 4. SEM images: (a) H-Beta-1 zeolite and (b) H-Beta-2 zeolite.

is attributed to its high acidity as well as 3-dimensional large 3.6.2. Effect of crystallite size of H-Beta zeolite
pore structure. Among all the catalysts, H-Beta-1 was found The effect of crystallite size on catalytic activity in glycerol con-
to be the most active catalyst which converted 86% of glyc- densation with acetone at room temperature was studied with
erol into products with 98.5% selectivity for solketal. The high H-Beta-1 and H-Beta-2 as shown in Fig. 6a. H-Beta-1 with smaller
activity of H-Beta-1 could be attributed to easy diffusivity crystallite size, gave 86% glycerol conversion of with 98.5% solke-
of molecules in the large pores and short path length due tal selectivity whereas H-Beta-2 zeolite with bigger crystallite size
to small crystallite size (Fig. 4). Also, the catalytic activity of showed only 38% glycerol conversion with 94.6% solketal selectiv-
H-Beta-1 was higher at the initial reaction time (10 min) com- ity. Even though the acidity of H-Beta-1 is 1.2 times lower than
pared with all the other catalysts screened. The results show H-Beta-2, the catalytic activity was much higher for H-Beta-1. This
a decrease in catalytic activity in the following order: H-Beta- indicates that crystallite size effect is predominant over acidity in
1 > H-Y > Amberlyst-15 > CsHPW > MoO3 /SiO2 > Montmorillonite H-Beta catalyst for this reaction. Interestingly for H-Beta-1 zeo-
K-10 > H-ZSM-5 > H-Mordenite. lite, the average crystallite size is 3.3 times smaller but turn over

Scheme 1. Plausible mechanistic pathway of acetalization glycerol with acetone.

52 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54

sites by oxalic acid treatment. Glycerol conversion decreased from

86 to 75.8% upon mild dealumination in comparison with parental
H-Beta-1 zeolite (86%). However, decrease in acidity did not affect
the selectivity as all the three catalysts gave identical solketal selec-
tivity of 98.5% as shown in Fig. 6b. It is evident that strong acidic
sites have a major role in the catalytic activity of glycerol transfor-
mation to solketal rather than weak acidic sites.
The behavior of Lewis acidic sites in beta catalyst was studied
using Cu2+ ion-exchanged Beta-1 catalyst. There was a decrease
in glycerol conversion after Cu2+ ion exchange of H-Beta-1 cat-
alyst. Cu/H-Beta catalyst containing (strong/weak (S/W) acidity
ratio = 0.93) gave 67.5% glycerol conversion with 98% solketal selec-
tivity in 60 min (Fig. 6c). To get more insight into this, Brønsted
acidic sites in H-Beta-1 zeolite were removed by converting pro-
tonic form into sodium form (Na-Beta) by sodium ion exchange.
Interestingly, Na-Beta zeolite and copper oxide showed no activ-
ity which may be attributed to their basic nature [25]. Cu-Beta
catalyst with S/W ratio = 1.06 gave higher glycerol conversion of
71.5% with 98.7% solketal selectivity in 60 min compared to Cu/H-
Beta. The higher activity of Cu-Beta could be due to increase in
strong acidic sites compared with Cu/H-Beta as shown by NH3 -
TPD.
The correlation between glycerol conversion with strong to
weak acidity ratio of unmodified and modified Beta-1 catalysts are
represented in Fig. 6d. The glycerol conversion decreased (from 86
to 67.5%) with a decrease in strong to weak acidity ratio (from 1.48
to 0.93). In other words, the decrease in strong acid sites of modi-
fied beta catalysts (Table 2) leads to decrease in glycerol conversion.
It indicates that the glycerol acetalization reaction was predomi-
nantly catalyzed by strong acidic sites under the reaction conditions
employed.

3.6.4. Influence of reaction conditions

The effect of catalyst concentration toward acetalization of
glycerol was studied by loading H-Beta-1 with different catalyst
amounts from 1 to 7.5 wt% (referred to glycerol) as shown in
Fig. 7a. As the catalyst amount increased, the glycerol conversion
was significantly increased from 78.7 to 87.1% due to increase
in the number of accessible active sites. The solketal selectivity
increased marginally from 97.4 to 98.5% as the catalyst amount
increased from 1 to 7.5 wt%. Solketal selectivity was found to be
almost the same for catalyst concentration from 2.5 to 7.5 wt%.
Fig. 5. Glycerol acetalization with acetone using various solid acid catalysts. Reac- Catalyst amount of 5 wt% with respect to glycerol was used for fur-
tion conditions: catalyst amount = 5 wt% referred to glycerol, acetone/glycerol mole ther studies as it was found to give high activity and selectivity for
ratio = 2, room temperature. H-Beta-1 (), H-Y (䊉), Amberlyst-15 (), CsHPW (), solketal.
Montmorillonite K-10 (), MoO3 /SiO2 (), H-ZSM-5 (hexagon), H-Mordenite (pen-
The effect of reactants mole ratio of acetone to glycerol was
tagon).
studied at room temperature using 5 wt% of H-Beta-1 catalyst
with respect to glycerol. The mole ratio of acetone to glycerol
was varied from 1:1 to 3:1 as shown in Fig. 7b. The conver-
frequency (activity) is 3.1 times higher compared with H-Beta-2 sion of glycerol increased with the increase in mole ratio from
zeolite. It is thus evident that glycerol conversion and selectivity 1:1 to 2:1 due to the increase in accessibility of acetone with
depends on crystallite size of the catalyst. Higher activity for zeo- glycerol. Glycerol conversion remained almost the same with fur-
lite beta with smaller crystallite size could be due to faster diffusion ther increase in mole ratio to 3:1. The selectivity to solketal
of molecules through smaller path length of zeolite channels. remained almost same at 98.5% for all the mole ratios. Based on
the above results, acetone to glycerol mole ratio of 2:1 was found
3.6.3. Influence of acidity of beta catalysts on their catalytic to be the optimum for the catalytic transformation of glycerol to
activity solketal.
Influence of nature and amount of acidity of modified zeolite
beta catalysts on catalytic activity was studied by dealuminated 3.6.5. Reaction mechanism
beta zeolite with varied acidity and Lewis acidic Cu2+ exchanged Based on the above results, a plausible mechanism is proposed
beta catalysts. The effect of acidity in H-Beta-1 zeolite was stud- for H-Beta zeolite catalyzed acetalization of glycerol with acetone
ied at room temperature with acetone to glycerol mole ratio of 2. as shown in Scheme 1. In the first step, the carbonyl group in
The catalytic activity of beta zeolite decreased with increase in SAR acetone was activated by acid sites of the catalyst. In the second
which is correlated with the acidity determined by NH3 -TPD as step, OH group of glycerol attacks the carbonyl carbon and forms
shown in Table 2. H-Beta-1-A and H-Beta-1-B catalysts showed a an intermediate called hemiacetal. This intermediate undergoes
decrease in glycerol conversion due to the decrease in strong acidic cyclization by the attack of lone pair of electrons present on the
 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54 53

Fig. 6. (a) Catalyst performance of H-Beta catalyst with two different crystallite sizes. (b) Catalyst performance of H-Beta catalyst of different SAR. (c) Glycerol acetalization
with Cu modified beta and CuO catalysts. (d) Catalyst performance v/s strong to weak acidity ratio of unmodified and modified Beta-1 catalysts. Common reaction conditions:
catalyst amount = 5 wt% referred to glycerol, acetone/glycerol mole ratio = 2, room temperature.

adjacent or terminal hydroxyl group on the tertiary carbon atom as shown in Fig. 8. The XRD pattern of spent H-Beta-
leading to the formation of solketal or 6-membered cyclic acetal 1 matched well with that of fresh catalyst indicating
with the elimination of water molecule in the last step. no change in the structure after the reaction. Further
investigation on the acidity of the spent catalyst was
3.7. Study of spent H-Beta-1 catalyst studied with NH3 -TPD. The acidity of spent H-Beta-1 catalyst
was found to be the same as that of fresh catalyst (∼1.5 mmol/g).
The characterization of the H-Beta-1 catalyst after the Further, the spent catalyst showed similar activity compared to
reaction was studied with X-ray diffraction and NH3 -TPD fresh H-Beta-1 catalyst.

Fig. 7. Influence of reaction conditions: (a) Effect of catalyst concentration. Reaction conditions: catalyst = H-Beta-1 zeolite, acetone/glycerol mole ratio = 2, room temperature.
(b) Effect of reactants mole ratio: reaction conditions: catalyst = H-Beta-1 zeolite, catalyst amount = 5 wt% referred to glycerol, room temperature.
54 P. Manjunathan et al. / Journal of Molecular Catalysis A: Chemical 396 (2015) 47–54

Fig. 8. Characteristic properties of spent H-Beta-1 catalyst. (a) XRD patterns of fresh and spent H-Beta-1. (b) TPD profiles of fresh and spent H-Beta-1.

4. Conclusions [3] L.N. Silva, V.L.C. Gonçalves, C.J.A. Mota, Catal. Commun. 11 (2010) 1036–1039.
[4] J.A. Sanchez, D.L. Hernandez, J.A. Moreno, F. Mondragon, J.J. Fernandez, Appl.
Catal. A 405 (2011) 55–60.
H-Beta zeolite showed higher glycerol conversion compared to [5] M.D. Gonzalez, P. Salagre, E. Taboada, J. Llorca, Y. Cesteros, Green Chem. 15
other porous and non-porous catalysts screened at room temper- (2013) 2230–2239.
ature due to its large pore size and easy diffusivity of molecules. [6] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 5253–5277.
[7] J.A. Melero, G. Vicente, G. Morales, M. Paniagua, J. Bustamante, Fuel 89 (2010)
H-Beta-1 zeolite with lower crystallite size (average: 135 nm) 2011–2018.
gave higher catalytic activity than H-Beta-2 (average crystallite [8] M.J. Climent, A. Corma, S. Iborra, Green Chem. 16 (2014) 516–547.
size 450 nm) due to lower diffusion path length. Dealumination [9] C.X.A. da Silva, V.L.C. Goncalves, C.J.A. Mota, Green Chem. 11 (2009) 38–41.
[10] B. Mallesham, P. Sudarsanam, G. Raju, B.M. Reddy, Green Chem. 15 (2013)
of beta zeolite resulted in decrease of strong acid sites which
478–489.
decreased the catalytic activity. Copper ion exchange of zeolite beta [11] S. Sandesh, G.V. Shanbhag, A.B. Halgeri, RSC Adv. 4 (2014) 974–977.
also decreased the catalytic activity by ∼18%. A good correlation [12] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martin, Green Chem. 12
(2010) 899–907.
between catalytic activity and strong to weak acidity ratio in beta
[13] A.B. Halgeri, J. Das, Appl. Catal. A 181 (1999) 347–354.
catalysts was obtained. Thus, H-Beta zeolite of small crystal size and [14] N. Taufiqurrahmi, A.R. Mohamed, S. Bhatia, J. Nanopart. Res. 13 (2011)
high acidity with strong acid sites is an efficient catalyst for acetal- 3177–3189.
ization of glycerol with acetone at room temperature to produce [15] M.M. Reddy, M.A. Kumar, P. Swamy, M. Naresh, K. Srujana, L. Satyanarayana, A.
Venugopala, N. Narender, Green Chem. 15 (2013) 3474–3483.
solketal in high yields. [16] D. Nandan, P. Sreenivasulu, L.N.S. Konathala, M. Kumar, N. Viswanadham,
Micropor. Mesopor. Mater. 179 (2013) 182–190.
Acknowledgements [17] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Appl. Catal. B 98
(2010) 94–99.
[18] J.W. Kim, D.J. Kim, J.U. Han, M. Kang, J.M. Kim, J.E. Yie, Catal. Today 87 (2003)
PM acknowledges Admar Mutt Education Foundation (AMEF), 195–203.
Bangalore for providing scholarship and thankful to Manipal [19] W. Kiatkittipong, S. Wongsakulphasatch, N. Tintan, N. Laosiripojana, P.
Praserthdam, S. Assabumrungrat, Fuel Process. Technol. 92 (2011) 1999–2004.
university for permitting this research as a part of the Ph.D. pro- [20] B. Yilmaz, U. Muller, M. Feyen, S. Maurer, H. Zhang, X. Meng, F.S. Xiao, X. Bao,
gramme. Authors are thankful to CENSE, Indian Institute of Science, W. Zhang, H. Imai, T. Yokoi, T. Tatsumi, H. Gies, T. De Baerdemaekerj, D. De Vos,
Bangalore for SEM analysis. Catal. Sci. Technol. 3 (2013) 2580–2586.
[21] A.P. Amrute, A. Bordoloi, N. Lucas, K. Palraj, S.B. Halligudi, Catal. Lett. 126 (2008)
286–292.
References [22] Y. Izumi, M. Ono, M. Kitagawa, M. Yoshida, K. Urabe, Micropor. Mater. 5 (1995)
255–262.
[1] J.C.S. Ruiz, R. Luque, A.S. Escribano, Chem. Soc. Rev. 40 (2011) 5266–5281. [23] M.D. Gonzalez, Y. Cesteros, P. Salagre, Appl. Catal. A 450 (2013) 178–188.
[2] A. Behr, J. Eilting, K. Irawadi, J. Leschinsk, F. Linder, Green Chem. 10 (2008) [24] V.N. Shetti, J. Kim, R. Srivastava, M. Choi, R. Ryoo, J. Catal. 254 (2008) 296–303.
13–30. [25] G.V. Shanbhag, T. Joseph, S.B. Halligudi, J. Catal. 250 (2007) 274–282.
 Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80

Contents lists available at ScienceDirect

Journal of Molecular Catalysis A: Chemical

journal homepage: www.elsevier.com/locate/molcata

Utilization of renewable resources: Condensation of glycerol with

acetone at room temperature catalyzed by organic–inorganic hybrid
catalyst
Swetha Sandesh, A.B. Halgeri, Ganapati V. Shanbhag ∗
Materials Science Division, Poornaprajna Institute of Scientific Research (PPISR), Bidalur Post, Devanahalli, Karnataka State, Bangalore 562110, India

a r t i c l e i n f o a b s t r a c t

Article history: Organic–inorganic hybrid catalyst prepared from organic ammonium salt and heteropoly acid is reported
Received 3 October 2014 as highly active and selective heterogeneous catalyst for the condensation reaction of glycerol with
Received in revised form 23 January 2015 acetone at room temperature. The product formed during the reaction, solketal is a highly potential
Accepted 15 February 2015
compound applied majorly in petroleum and pharmaceutical industries. The (C3 H7 )4 N+ /PWA catalyst
Available online 17 February 2015
performed better than other conventional solid acid catalysts like H-beta, amberlyst-15, montmoril-
lonite K-10 and cesium salt of phosphotungstic acid with 94% glycerol conversion and 98% selectivity for
Keywords:
solketal. The high activity of (C3 H7 )4 N+ /PWA catalyst can be explained by its acidity and pseudo liquid
Glycerol
Solketal
behavior. An independent study on the influence of water on catalyst deactivation was performed by
Condensation adding a small amount of water (glycerol: water (1:1)) during the reaction. The (C3 H7 )4 N+ /PWA cata-
Heteropoly salt lyst showed a remarkable resistance toward deactivation due to water with only a marginal decrease
Acid catalyst in conversion (∼3%) compared to other conventional solid acid catalysts like amberlyst-15, H-beta and
montmorillonite K-10. (C3 H7 )4 N+ /PWA catalyst was truly heterogeneous and showed good reusability
for 3 catalyst recycles.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction Conventionally condensation of glycerol with acetone was

carried out using homogeneous acid catalyst sulphuric acid or
Glycerol has recently gained importance as an inexpensive p-toulenesulphonic acid [15]. To overcome the environmental
and abundant raw material for the synthesis of variety of value- problems, many solid acid catalysts like arenesulfonic acid-
added chemicals. There are several routes for the transformation of functionalized mesostructured silica (Ar-SBA-15) [16], Nb2 O5 [5],
glycerol such as transesterification, etherification, acetylation, oxi- SO4 2− /ZrO2 [17], activated carbon [18], Si induced heteropoly
dation, condensation, hydrogenolysis and dehydration [1–7]. One acid [19], Hf-TUD-1 [20] and zeolites [21] have been reported
such economically viable reaction is the condensation of glycerol for condensation reaction of glycerol and acetone. However, all
with acetone to give 5 membered ring compound, solketal (2,2- these catalysts have been reported to work at reaction temper-
dimethyl-4-hydroxymethyl-1,3-dioxolane) (ESI: Scheme S1). ature ∼70 ◦ C. Recently, H-beta has been studied as a catalyst for
Solketal is a colorless and odorless liquid, completely soluble in acetalization of glycerol at room temperature, in which, effect of
water and stable under normal temperature and pressures. Due to crystallite size and acidity has been investigated in detail by our
its low toxicity, it has been widely used as a solvent for maintain- group [22]. In this study, quaternary ammonium salt of heteropoly
ing single phase mixture of gasoline and ethanol [8], for cleaning acid is reported for the first time as heterogeneous acid catalyst for
metallic, electrical and plastic substrates [9] and pharmaceutical glycerol transformation. Since heteropoly acids (HPA) are soluble
injections as water miscible solvent [10,11]. It is also used in cos- in polar compounds they can be heterogenised by ion exchange
metics industry as an additive in ointments [12] in the chemical with Cs, Rb and Sm [23,24]. Alkyl ammonium chloride and nitrate
industry as an additive in low temperature transfer fluids [13] and salts were used as homogeneous acid catalysts for aldol conden-
as a fuel additive [14]. sation and alkylation reactions [25,26] whereas alkylammonium
exchanged silicotungstate salt was found to act as heterogeneous
catalyst for epoxidation reactions [27]. Furthermore, sulfonated
∗ Corresponding author. Tel.: +91 80 27408552; fax: +91 80 23619034.
organic ammonium salt of phosphotungstic acid was also reported
E-mail address: [email protected] (G.V. Shanbhag).
as heterogeneous catalyst in polar medium [28].

http://dx.doi.org/10.1016/j.molcata.2015.02.015
1381-1169/© 2015 Elsevier B.V. All rights reserved.
74 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80

Fig. 1. a) FT-IR spectra of (C3 H7 )4 N+ /PWA, PWA and (C3 H7 )4 N+ Br and b) TGA plot of (C3 H7 )4 N+ /PWA catalyst.

There has been an increasing interest in recent years to Süd-Chemie India Pvt. Ltd. All the chemicals were of research grade
develop novel solid catalysts and modify them for selective syn- and used without any further purification.
thesis of value added chemicals. In the present communication,
we report tetrapropylammonium salt of phosphotungstic acid
2.2. Catalyst preparation
[(CH3 CH2 CH2 )4 N)3 PW12 O40 ] as an efficient heterogeneous catalyst
for the condensation reaction of glycerol with acetone under room
The organic–inorganic hybrid catalysts were prepared by adding
temperature (30 ◦ C). Here, heteropolytungstate is preferred as a
drop wise, the required amount of aq. 1 M solution of (C3 H7 )4 N+ Br
precursor for this catalyst over heteropolymolybdate because of
to aqueous solution of HPA hydrates (PWA, SWA, PMoA) (0.02 M)
their stronger acidity, higher thermal stability and lower oxidation
at room temperature with stirring. The obtained slurry contain-
[21].
ing precipitate was stirred for 2 h, filtered, washed with water
and dried in an oven at 120 ◦ C for 4 h and activated at 150 ◦ C
before use. The catalysts were designated as (C3 H7 )4 N+ /PWA,
2. Experimental section
(C3 H7 )4 N+ /SWA, (C3 H7 )4 N+ /PMoA for respective HPAs. Different
organic salts on PWA were prepared by following the same pro-
2.1. Materials
cedure as mentioned above by varying the organic salt and are
designated as C8 H20 N+ /PWA and C19 H42 N+ /PWA. Cs/PWA and
Glycerol, acetone and ammonium carbonate were purchased
NH4 /PWA catalysts were prepared by literature methods [29,30]
from Merck India Ltd. Phosphotungstic acid (PWA), silicotungstic
and other solid acid catalysts were activated at 150 ◦ C prior to the
acid (SWA), phosphomolybdic acid (PMoA), tetraethyl ammo-
reaction.
nium bromide ((C2 H5 )4 N+ Br− ), tetrapropyl ammonium bromide
and ((C3 H7 )4 N+ Br− ) and cetyltrimethyl ammonium bromide
((C16 H33 )(CH3 )3 N+ Br− ) were purchased from SD fine chemicals. 2.3. Catalyst characterization
Amberlyst-15 (AB-15) was obtained from Alfa Aesar, USA. The
montmorillonite K-10 clay (hereafter K-10 clay) was purchased FT-IR (Alpha-T, Bruker) spectra were obtained in the range of
from Sigma–Aldrich, USA. H-beta (SAR-25) was kindly donated by 4000–600 cm−1 to study the chemical property of the catalyst. The

Fig. 2. Catalytic activity of different solid acid catalysts. Conditions: Temperature = 30 ◦ C, Glycerol: acetone = 1:6, Catalyst weight = 3 wt% of total reactant weight.
 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80 75

Fig. 3. Catalytic activity of tetrapropylammonium exchanged HPAs. (Conditions as mentioned in Fig. 2).

KBr pellets of samples prepared were directly used for the FT-IR An independent experiment was conducted to understand the
analysis. PXRD and SEM data are represented in ESI (Fig. S1 and effect of water on the catalytic activity during the reaction. For this,
S2, respectively). Thermogravimetric analysis (TGA) was carried the reaction was conducted in the absence of water for 15 min.
out under inert atmosphere, with a heating rate of 10 ◦ C min−1 in Then, 1 mole of water was added to the reaction mixture (glyc-
the range 35–800 ◦ C. The C H N analysis was performed to evalu- erol: water = 1:1) and monitored the effect of water on the catalytic
ate the extent of exchange of tetrapropylammonium ions with H+ activity.
of HPAs. Acidity of all catalysts were measured by potentiometric Leaching test was carried out for condensation reaction to inves-
titration [31]. About 5 mg of the catalyst was suspended in 5 ml tigate the heterogeneity of the catalyst. The reaction was carried out
of n-butylamine solution (0.05 N) in acetonitrile and sonicated for using 0.5 wt% of (C3 H7 )4 N+ /PWA catalyst at reaction temperature
5 min to attain uniform dispersion. Then the above solution was of 30 ◦ C with glycerol to acetone mole ratio of 1:6. Reaction was
suspended in excess of acetic acid (90 mL) and potentiometrically stopped after 30 min and reaction mixture containing the catalyst
titrated against perchloric acid (0.1 N) in acetic acid. Prior to sample was filtered. The reaction was continued without the catalyst for
titration, a blank titration of acetic acid and n-butyl amine against next 5 h and sample was drawn on every one hour interval and
perchloric acid was carried out to check the acidity contribution analysed.
from the solutions used.
3. Results and discussion
2.4. Catalytic activity
Organic–inorganic hybrid catalysts were characterized by FTIR,
The condensation reaction of glycerol with acetone was carried TGA and C H N elemental analysis studies. FT-IR (Alpha-T, Bruker)
out in 100 ml two-necked glass reactor equipped with a magnetic spectra were obtained in the range of 4000–600 cm−1 to study
stirring bar, a Liebig condenser, and a thermometer. The required the chemical property of the catalyst. The KBr pellets of samples
amount of glycerol and acetone were taken in the reactor and prepared were directly used for the FT-IR analysis. IR spectra of
stirred at 1800 rpm for 15 min before the addition of pre-activated PWA (H3 PW12 O40 ), (C3 H7 )4 N+ /PWA and (C3 H7 )4 N+ Br are shown
catalyst. The reaction was performed at 30 ◦ C for 2 h and then the in Fig. 1(a). Four bands at 700–1100 cm−1 region corresponding to
mixture was taken out and centrifuged for 10 min to separate the Keggin unit (HPW) structural vibrations are observed for PWA and
catalyst from the liquid phase. The obtained product was analyzed
by gas chromatography (Shimadzu, GC-2014) with flame ionization
Table 2
detector (FID) equipped with capillary column (0.25 mm I.D and Potentiometric analysis of acidity and catalytic activity of all the catalysts.
30 m length, Stabilwax, Restek). All the products were confirmed
Catalysts Acidity (mmol/g) Glycerol conversion (mol%) TON
by gas chromatography with mass spectroscopy (Shimadzu, GCMS
QP 2010). (C3 H7 )4 N+ /PWA 0.60 94 120
AB-15 0.95 89 72
H-beta 1.5 88 23
K-10 clay 1.1 87 61
Table 1 Cs/PWA 1.88 38 24
Elemental analysis of organic–inorganic hybrid catalysts. (C3 H7 )4 N+ /SWA 0.65 90 106
(C3 H7 )4 N+ /PMoA 0.57 87 117
Catalysts N (wt%) C (wt%) H (wt%) (C2 H5 )4 N+ /PWA 0.53 83 120
+
(C3 H7 )4 N /PWA 1.4 12.5 2.4 (C16 H33 )(CH3 )3 N+ /PWAa 0.56 77 106
(C3 H7 )4 N+ /SWA 1.8 15.6 3.0 Conditions: Temperature = 30 ◦ C, Glycerol: acetone = 1: 6, Catalyst = 3 wt% of total
(C3 H7 )4 N+ /PMoA 2.1 18.1 3.4 reactant weight, Time = 120 min, TON = moles of glycerol converted per mole of acid
(C16 H33 )(CH3 )3 N+ /PWA 1.3 18.4 3.4 sites.
(C3 H7 )4 N+ /PWA 3rd recycle 1.9 14.2 3.1 a
Homogeneous phase.
76 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80

Fig. 4. The effect of addition of water during the reaction. Conditions: Temperature = 30 ◦ C, Glycerol: acetone: water = 1:6:1, Catalyst weight = 3 wt% of total reactant weight,
Water was added after 15 min of reaction time.

(C3 H7 )4 N+ /PWA suggesting that the framework of primary Keg- respectively [32]. In addition to Keggin unit, (C3 H7 )4 N+ /PWA
gin structure remained unaltered after modification of PWA with exhibits other 3 characteristic peaks corresponding to tetrapropy-
ammonium salt. The peaks corresponding to Keggin anion vibra- lammonium ion at 1467, 1400 cm−1 (C H stretching) and
tion are as follows. The stretching frequency of P O in the central 1180 cm−1 (C H bending). The peak at 1063 cm−1 is attributed to
PO4 tetrahedron is at 1081 cm−1 . The peak at 987 cm−1 is due to C N stretching vibration [33].
the terminal W O vibration in the WO6 octahedron and the peak Thermo gravimetric analysis (TGA) was carried out under inert
at 897 and 815 cm−1 were assigned to W Ob W and W Oc W atmosphere, with a heating rate of 10 ◦ C min−1 in the range
bridges, respectively. Weaker peaks appearing at 605 and 518 cm−1 35–800 ◦ C. Thermogravimmetric analysis of (C3 H7 )4 N+ /PWA cat-
due to bending vibrations of the type O P O and W O W bonds, alyst showed a marginal weight loss (∼ 1%) up to 120 ◦ C due to the

Table 3
Catalytic activities of NH4 + /PWA, organic salt of PWA and its precursors.

Catalyst Phase Glycerol conversion (mol%) Solketal selectivity (mol%) Solketal yield (mol%)

NH4 + /PWA Hetero 30 73 22

(C2 H5 )4 N+ /PWA Hetero 83 96 82
(C3 H7 )4 N+ /PWA Hetero 94 98 93
(C16 H33 )(CH3 )3 N+ /PWA Homo 77 94 75
(C3 H7 )4 N+ Br Homo 0.2 98 0.15
PWA Homo 90a 98 88

Conditions: Glycerol:acetone = 1:6, Temperature = 30 ◦ C, Catalyst = 3 wt% of total reactant weight, Time = 120 min.
a
Time = 10 min.
 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80 77

Fig. 5. Effect of conditions on performance (a) catalyst wt% (Conditions: Glycerol to acetone mole ratio = 1:6, temperature = 30 ◦ C) and (b) Glycerol to acetone mole ratio
(Conditions: Catalyst weight = 3 wt% of total reactant weight, Temperature = 30 ◦ C).

removal of water of hydration. Further, there was no weight loss protic reactants (such as glycerol in this study) or by reduction
up to 400 ◦ C which indicates that the catalyst is thermally stable of metal cations [35–38]. The structure is preserved even upon
up to that temperature. Above 400 ◦ C, there was a rapid decrease complete substitution of tetraalkyl ammonium ions and manifests
of weight by 15% up to 450 ◦ C which can be attributed to the loss itself to exhibit extremely high proton mobility and a pseudoliquid
of organic moiety (Fig. 1(b)). The weight loss of 2.2% from 550 to phase.
600 ◦ C is due to the decomposition of Keggin heteropoly anion into
WO3 and P2 O5 [20]. 3.1. Catalytic activity study
C H N elemental analysis showed the extent of exchange of
tetrapropyl ammonium ions with protons of HPA. It is found that Condensation of glycerol with acetone was carried out over
protons are exchanged completely with alkyl ammonium ions in organic–inorganic hybrid catalysts and compared with different
the heteropoly salts as presented in Table 1. These alkyl ammonium types of solid acid catalysts namely, H-beta zeolite, montmoril-
salts of HPA showed the acidity in the range of 0.5–0.7 mmol/g mea- lonite K-10 (K-10 clay), amberlyst-15 (AB-15) and cesium salt of
sured by potentiometric titration which could be attributed to the PWA (Cs/PWA) as shown in Fig. 2. A blank reaction in the absence
presence of protons in the catalyst. The reason for generation of of catalyst showed only trace conversion of glycerol.
protons in neutral salts of HPA is still not understood completely Although glycerol conversion varied with different hetero-
[34]. The solid tetraalkyl ammonium salts of HPA compounds geneous acid catalysts, selectivity toward 5-membered ring
possess discrete ionic structures, comprising fairily mobile struc- compound, solketal was invariably high for all the catalysts. Solid
tural units; heteropoly anion and the counter cation. Completely acid catalysts viz. H-beta, AB-15 and K-10 clay showed high glyc-
substituted heteropoly salts gain protons upon interaction with erol conversion of 89, 87 and 82%, respectively, for initial one
reaction medium either by dissociation of coordinated water/polar hour and then remained almost constant. These highly acidic

Fig. 6. (a) Leaching test and (b) Confirmation of heterogeneity from FT-IR.
78 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80

Fig. 7. (a) Reusability of (C3 H7 )4 N+ /PWA catalyst. Conditions: Temperature = 30 ◦ C, Glycerol: acetone = 1:6, Catalyst weight = 3 wt% of total reactant weight and (b) XRD
patterns of (C3 H7 )4 N+ /PWA fresh and 3rd recycled catalyst.

catalysts, although showed high initial activity, might be deac- exhibits bulk type I catalysis [28]. In bulk type I catalysis of HPA
tivated due to the blockage of active sites by water molecules and its salts, e.g., acid-catalyzed reactions of polar molecules at rel-
formed during the reaction. Cs+ exchanged PWA catalyst with atively lower temperatures, the reactant molecules are absorbed
high acidity 1.88 showed low glycerol conversion of around 40% in the inter polyanion space of ionic crystal, undergo reaction and
with 85% solketal selectivity. Compared to all the above acid cata- then products desorb from the solid bulk. Polar molecules like glyc-
lysts (C3 H7 )4 N+ /PWA catalyst with acidity of 0.6 mmol/g gave high erol absorb into the solid bulk and expand the distance between
glycerol conversion of 94% with almost 98% solketal selectivity in the polyanions. This type of catalysis is known as pseudo liquid
120 min of reaction. Both (C3 H7 )4 N+ /PWA and (C2 H5 )4 N+ /PWA cat- catalysis in presence of polar reactants [38–42].
alysts showed the highest turnover number of 120 followed by The condensation of glycerol with acetone was carried out
other organic–inorganic hybrid catalysts made of SWA and PMoA with different tetrapropylammonium exchanged heteropoly acids
(117 and 106, respectively) (Table 2). It is observed that solke- namely, (C3 H7 )4 N+ /PWA, (C3 H7 )4 N+ /SWA, (C3 H7 )4 N+ /PMoA and
tal selectivity increases as the reaction time increases for all the the results are represented in Fig. 3. Among these catalysts,
reactions. The other product formed is a 6-membered cyclic acetal, (C3 H7 )4 N+ /PWA showed highest glycerol conversion of 94% with
5-hydroxy-1, 3-dioxane as shown in ESI: Scheme S1. Initially, side 98% selectivity to solketal in 120 min reaction time, whereas
product acetal formation was high, probably because the catalytic (C3 H7 )4 N+ /SWA and (C3 H7 )4 N+ /PMoA showed marginally lower
sites are fresh and more active in the initial phase of the reaction. conversion (89.8 and 86.8%, respectively) with the selectivity same
In (C3 H7 )4 N+ /PWA catalyst, along with acidity (0.6 mmol/g) the as that of (C3 H7 )4 N+ /PWA. The similarity in performance of these
combination of rigid molecule, PWA and the flexible alkyl ammo- catalysts which differ in the type of heteropoly anion shows that
nium ion contributes to the high glycerol conversion. In this study, these catalysts contain similar active sites. Different organic ammo-
both reactants (glycerol and acetone) are polar and the reactant nium salts like (C2 H5 )4 N+ , (C3 H7 )4 N+ and (C16 H33 )(CH3 )3 N+ were
molecules can penetrate into the secondary structure of heteropoly exchanged with protons of PWA and applied as catalysts for solke-
anions to come in contact easily with active sites of (C3 H7 )4 N+ /PWA tal synthesis. While (C2 H5 )4 N+ /PWA and (C3 H7 )4 N+/ PWA catalysts
catalyst. Wang and co-workers observed that for esterification acted as purely heterogeneous catalysts with glycerol conversion
reaction, the synthesized alkyl quaternary ammonium salt of PWA of 83.4 and 94%, respectively, (C16 H33 )(CH3 )3 N+ /PWA with a long
alkyl chain of 16 carbon atoms acted as homogeneous catalyst with
a low conversion of 77% (Table 3). The ammonium salt of PWA
without alkyl groups (NH4 + /PWA) was applied as a catalyst for this
reaction and the activity was compared with (C3 H7 )4 N+ /PWA. The
NH4 + /PWA catalyst acted as heterogeneous catalyst but gave low
glycerol conversion of 30% with 73% selectivity for solketal. Inter-
estingly, PWA without ion exchange acted as homogeneous catalyst
with glycerol conversion of 90% within 10 min of reaction. Pure
(C3 H7 )4 N+ Br salt was also used as a homogeneous catalyst which
showed negligible activity for this reaction.
It is known that certain catalysts deactivate due to water formed
as a byproduct during the reaction. As it is difficult to confirm
this effect for the batch reactions where conversions get stable
at longer reaction period, independent experiments were con-
ducted by adding small amount of water (glycerol: water = 1:1)
after 15 min of reaction (Fig. 4). It is observed that K-10 and H-
beta were highly affected due to the addition of water and glycerol
conversion decreased remarkably by 30 and 35% compared with
neat reaction. AB-15 showed comparatively more resistance to the
effect of water (about 15% decrease in conversion). Among the 4
catalysts taken for this study, (C3 H7 )4 N+ /PWA catalyst showed a
Scheme 1. Plausible reaction mechanism for condensation reaction of glycerol with
remarkable resistance toward deactivation due to water with only
acetone.
 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80 79

a marginal decrease in conversion (∼3%). This study shows that hybrid catalyst also exhibited a remarkable resistance toward deac-
byproduct water can block the active sites for certain catalysts and tivation due to water compared to other conventional solid acid
highly hydrophilic catalysts like H-beta and K-10 are more affected catalysts like amberlyst-15, H-beta and montmorillonite K-10. Dif-
by deactivation of acidic sites due to water. ferent Keggin heteropoly salts with same alkyl ammonium cation
showed similar catalytic activity. (C3 H7 )4 N+ /PWA catalyst was
3.2. Influence of reaction conditions truly heterogeneous and showed good reusability for 3 catalyst
recycles.
The effect of catalyst weight percent and mole ratio of glycerol to
acetone was studied for (C3 H7 )4 N+ /PWA catalyst at room temper- Acknowledgements
ature and depicted in Fig. 5(a). The glycerol conversion increased
from 10 to 79.8% in 15 min of reaction as the catalyst concentration SS acknowledges CSIR, New Delhi for providing Senior Research
increased from 0.3 to 3 wt % (with respect to total weight of the Fellowship and also thankful to Manipal University for permitting
reactants). Further increase in the catalyst concentration from 3 to this research as a part of the PhD programme.
7 wt% did not improve the glycerol conversion.
The effect of mole ratio of glycerol to acetone was studied at
Appendix A. Supplementary data
room temperature using (C3 H7 )4 N+ /PWA catalyst. The mole ratio of
glycerol to acetone was varied from 1:1 to 1:8 as shown in Fig. 5(b).
Supplementary data associated with this article can be
The glycerol conversion increased with increase in mole ratio. The
found, in the online version, at http://dx.doi.org/10.1016/j.molcata.
conversion was 57% for 1:1 mole ratio after 15 min reaction and
2015.02.015.
it increased to 79.5% upon increasing mole ratio from 1:1 to 1:8
whereas selectivity toward products remained almost the same
(98%) with different acetone concentrations. References
Based on the above results, the optimized conditions using
(C3 H7 )4 N+ /PWA are 30 ◦ C (room temperature), 3 wt% catalyst with [1] J. Dubois, M. Aresta, A. Dibenedetto, C. Ferragina, F. Nocito, U.S. Pat., 245,513
A1 2011.
respect to total reactant weight and glycerol to acetone mole ratio [2] A.M. Ruppert, J.D. Meeldijk, B.W.M. Kuipers, B.H. Ern, B.M. Weckhuysen,
of 1:6. Based on activity of the catalyst, a plausible mechanism is Chem. Eur. J. 14 (2008) 2016–2024.
proposed. The acid sites in (C3 H7 )4 N+ /PWA facilitate the reaction by [3] S.B. Troncea, S. Wuttke, E. Kemnitz, S.M. Coman, V.I. Parvulescu, Appl. Catal. B:
Environ. 107 (2011) 260–267.
activating the carbonyl group of acetone forming a carbocation. The [4] Z. Chun-Hui, J.N. Beltramini, L. Chun-Xiang, X. Zhi-Ping, G.Q. Lu, A. Tanksale,
hydroxyl group of glycerol then attacks the carbocation to form an Catal. Sci. Technol. 1 (2011) 111–122.
intermediate which then undergoes cyclization with the removal [5] G.S. Nair, E. Adrijanto, A. Alsalme, I.V. Kozhevnikov, D.J. Cooke, D.R. Brown,
N.R. Shiju, Catal. Sci. Technol. 2 (2012) 1173–1179.
of water to form solketal (Scheme 1). [6] Y. Nakagawa, K. Tomishige, Catal. Sci. Technol. 1 (2011) 179–190.
Leaching test was carried out to confirm the heterogeneity of [7] A. Ulgen, W.F. Hoelderich, Appl. Catal. A: Gen. 400 (2011) 34–38.
the catalyst [Fig. 6(a)]. The catalyst was removed from the reac- [8] S.M. William, U.S. Pat., 4,390,344, 1983.
[9] R.G. Henry, U.S. Pat., 5,801,136, 1998.
tion mixture after 30 min (at 48% conversion) and the reaction was [10] B. Hassel, S. Leiner, WO/1995/026191A1, 1995.
continued without the catalyst for 5 h. The conversion remained [11] F. Mottu, S. Marie-Jose, D.A. Rufenacht, E. Doelker, J. Pharm. Sci. Technol. 55
the same at 48% even after five hours of reaction time, indicat- (2001) 16–18.
[12] A. Laurent, F. Mottu, R. Chapot, J.Q. Zhang, O. Jordan, D.A. Rufenacht, E.
ing no leaching of active sites into the reaction media. It is further
Doelker, M. Jean-Jacques, J. Pharm. Sci. Technol. 61 (2007) 64.
confirmed by measuring FT-IR spectra of the catalyst and product [13] A. Mendoza, U.S. Pat., 5,484,547, 1996.
mixture. The band between 1200 and 500 cm−1 corresponding to [14] G.A. Somorjai, U.S. Pat., 4,390,344, 1983.
[15] B. Bruchmann, H. Gruner, K. Haberle, M. Hirn, U.S. Pat., 599, 1705, 1999.
the Keggin structure of PWA was absent in the product mixture
[16] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martin, Green Chem. 12
[Fig. 6(b)]. (2010) 899–907.
Catalyst recyclability test was performed for (C3 H7 )4 N+ /PWA [17] P.S. Reddy, P. Sudarsanam, B. Mallesham, G. Raju, B.M. Reddy, J. Ind. Eng.
by employing washed (with acetone) and dried (at 120 ◦ C) cat- Chem. 17 (2011) 377–381.
[18] L. Li, T.I. Koranyi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012)
alyst under optimized reaction conditions. The catalyst showed 1611–1619.
good recyclability with marginal decrease in activity after 3 recycles [19] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Appl. Catal. B:
(Fig. 7(a)). XRD of fresh and 3rd spent catalyst showed no change in Environ. 98 (2010) 94–99.
[20] I.V. Kozhevnikov, J. Mol. Catal. A: Chem. 305 (2009) 104–111.
the phase purity of the catalyst (Fig. 7(b)). Moreover, C, H, N analysis [21] X.A.C. Silva, L.C.V. Goncalves, J.A.C. Mota, Green Chem. 11 (2009) 38–41.
of recycled catalyst showed marginally higher percentage of C, H [22] P. Manjunathan, S.P. Maradur, A.B. Halgeri, G.V. Shanbhag, J. Mal. Catal. A:
and N compared to fresh catalyst viz. C = 14.2 wt%, H = 3.1 wt% and Chem. (2015), http://dx.doi.org/10.1016/j.molcata.2014.09.028.
[23] V.V. Costa, K.A.S. Rocha, I.V. Kozhevnikov, E.F. Kozhevnikov, E.V. Gusevskaya,
N = 1.9 wt%. Higher percentage of C and H could be due to small Catal. Sci. Technol. 3 (2013) 244–250.
amount of adsorbed organics remained even after regeneration. [24] M.H. Haider, N.F. Dummer, D. Zhang, P. Miedziak, T.E. Davies, S.H. Taylor, D.J.
However, from % of nitrogen, it is confirmed that all the ammonium Willock, D.W. Knight, D. Chadwick, G.J. Hutchings, J. Catal. 286 (2012) 206.
[25] P. Dziedzic, A. Bartoszewicz, A. Cordova, Tetrahedron Lett. 50 (2009)
groups are fully preserved in the recycled catalyst (Table 1).
7242–7245.
[26] S. Shi-Kai, B. Xiang, Z. Li-He, Z. Yuxin, C.N. Jiao, Chem. Commun. 47 (2011)
5007–5009.
4. Conclusions [27] S. Uchida, K. Kamata, Y. Ogasawara, M. Fujita, N. Mizuno, Dalton Trans. 41
(2012) 9979–9983.
[28] Y. Leng, J. Wang, D. Zhu, Y. Wu, P. Zhao, J. Mol. Catal. A: Chem. 313 (2009) 1–6.
Organic–inorganic hybrid catalysts such as (C3 H7 )4 N+ /PWA [29] L. Pesaresi, D.R. Brown, A.F. Lee, J.M. Montero, H. Williams, K. Wilson, Appl.
were synthesized and applied as highly active and selective Catal. A: Gen. 360 (2009) 50.
heterogeneous catalysts for glycerol condensation reaction with [30] Y. Izumi, M. Ogawa, K. Urabe, Appl. Catal. A: Gen. 132 (1995) 127–140.
[31] G.H. Jeffery, J. Bassett, J. Mendham, R.C. Denney, Vogel’s textbook of
acetone to give solketal. Among all the tetrapropylammonium ions quantitative chemical analysis, 5th ed, pp. 590.
exchanged HPA catalysts, (C3 H7 )4 N+ /PWA gave high glycerol con- [32] Y. Leng, H. Ge, C. Zhou, J. Wang, J. Chem. Eng. 145 (2008) 335–339.
version of 94% with 98% solketal selectivity. The high activity of [33] R.M. Silverstein, G.C. Bassler, T.C. Morrill, Spectrometric Identification of
Organic Compounds, 4th ed., John Wiley and Sons, New York, 1981.
(C3 H7 )4 N+ /PWA catalyst can be explained by its acidity and pseudo [34] A.I. Chikin, A.V. Chernyak, Z. Jin, A.E. Ukshe, V.I. Volkov, Y.A. Dobrovolsky, J.
liquid behavior. The catalyst showed better performance compared Solid State Electrochem. 16 (2012) 2767–2778.
to other conventional solid acid catalysts. This organic–inorganic [35] I.V. Kozhevnikov, Appl. Catal. A: Gen. 256 (2003) 3–18.
80 S. Sandesh et al. / Journal of Molecular Catalysis A: Chemical 401 (2015) 73–80

[36] Y. Ono, J.M. Thomas, K.I. Zamaraev (Eds.), Perspectives in Catalysis, Blackwell, [39] M. Misono, Catal. Rev. Sci. Eng. 29 (1987) 269–321.
London, 1992, p. 431. [40] M. Misono, Chem. Commun. 1 (2001) 1141–1152.
[37] L. Matachowski, A. Zieba, M. Zembala, A. Drelinkiewicz, Catal. Lett. 133 (2009) [41] N. Mizuno, T. Watanabe, H. Mori, M. Misono, J. Catal. 123 (1990)
49. 157–163.
[38] J.E. Sambetha, G. Romanellia, J.C. Autino, H.J. Thomas, G.T. Baronetti, Appl. [42] K. Tanabe, M. Misono, H. Hattori, Y. Ono, J. Catal. 51 (1989).
Catal A: Gen. 378 (2010) 114.
 Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Catalytic conversion of glycerol for sustainable production

of solketal as a fuel additive: A review
Malaya R. Nanda a, Yongsheng Zhang a, Zhongshun Yuan a, Wensheng Qin b,
Hassan S. Ghaziaskar c, Chunbao (Charles) Xu a,n
a
Department of Chemical and Biochemical Engineering, Western University, London, ON, Canada N6GA5B9
b
Department of Biology, Lakehead University, ON, Canada P7B5E1
c
Department of Chemistry, Isfahan University of Technology, 84156-83111 Isfahan, Iran

art ic l e i nf o a b s t r a c t

Article history: The exponential growth of biodiesel industries all around the world has produced a large amount of
Received 18 May 2015 glycerol as a byproduct, which must be valorized for the sustainability of the biodiesel industry. Keta-
Received in revised form lization of glycerol with acetone to synthesize solketal-a potential fuel additive is one of the most pro-
11 November 2015
mising routes for valorization of glycerol. In this article, state-of-the-art of glycerol ketalization is
Accepted 3 December 2015
Available online 29 December 2015
reviewed, focusing on innovative and potential technologies towards sustainable production of solketal.
The glycerol ketalization processes developed in both batch and continuous reactors and performance of
Keywords: some typical catalysts are compared. The mechanisms for the acid-catalyzed conversion of glycerol into
Biodiesel solketal are presented. The main operation issues related to catalytic conversion of crude glycerol in a
Glycerol continuous-flow process and the direct use of crude glycerol are discussed.
Ketalization
& 2015 Elsevier Ltd. All rights reserved.
Solketal
Continuous-flow reactor
Crude glycerol

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023
2. Recent progress in the reaction processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023
2.1. Historical context. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023
2.2. Improvement in batch process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1024
2.3. Development of continuous processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
3. Catalysis – the important parameter in ketalization reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1026
3.1. Influence of catalyst acidity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1026
3.2. Development and performance of transition metal catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027
4. Reaction models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027
4.1. Reaction mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027
4.2. Reaction kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1028
5. Key operation issues of flow reactors and use of crude glycerol as feedstock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1028
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1029

Abbreviations: pTSA, p-toluenesulfonic acid; Ar-sBA-15, Arenesulfonic acid-functionalized mesostructured silica; PW, Tungsto-phosphoric acid; SiW, Tungsto-silisic acid;
PMo, Molybdo-phosphoric acid; SiMo, Molybdo-silisic acid; CMR, Continuous microwave reactor; FM, Fluidic modules; WHSV, Weight hourly space velocity; Pr-SBA-15,
Propylsulfonic acid-functionalized mesostructured silica; Ar-SBA-15, Arenesulfonic acid-functionalized mesostructured silica; HAr-SBA-15, Hydrophobised arenesulfonic
acid-functionalized mesostructured silica; MPV, Meerwein–Ponndrof–Verley reduction; G, Glycerol; A, Acetone; F, Vacant adsorb sites; S, Solketal; FTIR, Fourier Transfor-
mation Infrared Spectroscopy; CTAB, Cetyltrimethyl ammonium bromide; TEOS, Tetraethyl orthosilicate; Zr-TUD-1, Three dimensional mesoporous Zirconium containing
catalyst; Hf-TUD-1, Three dimensional mesoporous Hafnium containing catalyst; Sn-MCM-41, Mesoporous tin containing catalyst with mobil composition of matter number
41; Sn-MCM-1, Mesoporous tin containing catalyst with mobil composition of matter number 1; A/G, Acetone to glycerol molar ratio; Pr-SO3H-SiO2, Silica bonded-
propylsulfonic acid; T-SiO2, Silica bond-tosic acid; SAC-13, Nafion silica composite
n
Corresponding author. Fax: þ1 519 661 4016.
E-mail address: [email protected] (C. Xu).

http://dx.doi.org/10.1016/j.rser.2015.12.008
1364-0321/& 2015 Elsevier Ltd. All rights reserved.
 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031 1023

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1030
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1030

1. Introduction The ketalization reaction between glycerol and acetone is given

in Scheme 2, where solketal (2, 2-dimethyl-1, 3-dioxolane-4-
The depletion of non-renewable fossil fuels and their envir- methanol or 1,2-isopropylideneglycerol) is formed as the con-
onmental impacts are among the main factors that have drawn densation product over an acid catalyst. Solketal can be used as a
increasing attention towards biofuels, mainly bio-ethanol and fuel additive to reduce the particulate emission and to improve the
biodiesel. Biodiesel is mainly produced by the transesterification of cold flow properties of liquid transportation fuels [28]. It helps to
animal fats or vegetable oils (triglyceride) with a mono-alcohol
reduce the gum formation, improves the oxidation stability, and
(usually methanol) in presence of alkalies as shown below
enhances the octane number when added to gasoline [29]. Mak-
(Scheme 1) [1–3]. This biodiesel can be used directly or after
simov et al. reported its use as a versatile solvent and a plasticizer
blending with fossil-based diesel fuel.
in the polymer industry and a solubilizing and suspending agent in
In the transesterification process, glycerol is formed as the
principal byproduct. It is estimated that 10 wt% amount crude pharmaceutical preparations [30].
glycerol is generated for each amount of biodiesel produced [4]. This review paper mainly over-views the state-of-the-art of the
With the continued increase in the production of biodiesel, an sustainable production of solketal by catalytic reaction of glycerol
excessive amount of glycerol is expected to accumulate. It is pre- with acetone. Different types of processes and catalysts developed
dicted that by 2020 the global production of glycerol will be 41.9 and their performances are compared. Fundamentals of reaction
billion liters [5]. The crude glycerol produced form biodiesel mechanisms for the acid-catalyzed conversion of glycerol into
industry contains impurities such as water, inorganic salts (sodium solketal are presented. The main operation issues related to cata-
or potassium salts), methanol, fatty acids, and esters etc. [6–8], lytic conversion of crude glycerol in a continuous-flow process and
hence it is commonly treated as the waste stream of biodiesel the direct use of crude glycerol are discussed.
industry. It is economically viable for the large biodiesel producers The review article aims to (1) introduce various applications of
to refine this waste stream for the industrial applications, whereas solketal in different industries including polymer, pharmaceutical
for small biodiesel producers, they are unable to leverage the and cosmetics, food, and fuel industries, (2) highlight some major
treatment costs and instead they pay for glycerol removal. Due to challenges for industrial production of solketal, and (3) demon-
the excessive amount generated, the current crude glycerol price is strates promise of some new processes for utilization of crude
as low as 0.04–0.09 $/lb [9]. The predicted rapid growth of bio-
glycerol as feedstock for the production of solketal.
diesel production will further lower the glycerol price once it
enters into market [10]. Therefore, new and economical ways of
using glycerol must be developed to increase the value of crude
glycerol to enhance the sustainability of biodiesel industries. 2. Recent progress in the reaction processes
That being said, glycerol has diverse applications in different
fields especially in the pharmaceuticals, food, cosmetics, and 2.1. Historical context
polymer industries [11–13]. The versatility of glycerol is mainly
due to its physical and chemical properties. The presence of three It is well-known that ketals can be prepared by the reaction of
hydroxyl groups in glycerol makes it completely soluble in water an alcohol with a ketone in presence of an acid catalyst. Based on
and alcohols whereas insoluble in hydrocarbons. Furthermore, the the public sources of literature, Fischer first prepared the solketal
inter and intramolecular hydrogen bonds due to the presence of from acetone and glycerol in a batch rector catalyzed by hydrogen
hydroxyl groups lead to the high boiling point of glycerol (290 °C)
chloride [31]. 25 years later Fischer and Pfahler reported ketali-
at ambient pressure and high viscosity (1.412 Pa s) at room tem-
zation of glycerol using hydrogen chloride and anhydrous sodium
perature [14].
sulfate in a similar process [32]. Later, in 1948, Renoll and New-
On the other hand, catalytic and biological conversion of gly-
mann published their work on the synthesis of solketal in a three
cerol offer a tremendous potential to produce value-added che-
micals such as propanediols, acrolein, dihydroxyacetone, glyceric neck flask with reflux equipped with a sealed mechanical stirrer
acid, tartonic acid, epichlorohydrin, hydrogen, syngas, ethers, [33]. The authors chose petroleum ether as the reaction medium
esters, etc. [15–21]. Therefore glycerol can be considered as a and p-toluenesulfonic acid (pTSA) monohydrate as the catalyst to
platform chemical. A selection of these possibilities were reviewed achieve a high yield of solketal (87–90%). After the reaction, the
recently [12,22,23]. Production of cyclic acetals and ketals from products were separated by distillation under reduced pressure;
glycerol with aldehydes and ketones, respectively, is believed to be however the reaction time was very long (21–36 h). These early
one of the most promising glycerol applications as fuel/chemical studies on the synthesis of solketal remained without further
intermediates [24–27]. advances until the end of the 20th century when massive amount
of cheap glycerol was produced from biodiesel industry.

HO

OH
O O
H3C
+ HO OH + H2O
H3C CH3
H3C
O
Acetone Glycerol Solketal

Scheme 1. Glycerol as byproduct during biodiesel production. Scheme 2. Ketalization reaction between glycerol and acetone.
1024 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031

Nomenclature Kc Equilibrium constant for the reaction

r Rate of the reaction
[CpIrCl2]2 Pentamethylcyclopentadienyl iridium(III) chloride Kw Equilibrium adsorption constant for water
k Rate constant Ea Activation energy

2.2. Improvement in batch process reaction continuously. Removing water produced from solketal
synthesis is an effective way to break the thermodynamic barriers.
A Spanish patent was filed in 1981–1982 aiming to utilize a To remove the water from the reaction mixture, entrainers have
large volume of glycerol [34]. The inventors studied the reaction of been used in different processes [39]. Benzene is not a preferable
glycerol with acetone at the molar ratio of 1:1.1 in a batch reactor entrainer for this process as acetone is removed by distillation
over acid catalyst without a water entrainer. In the process, by- before benzene. Other entrainers for this process can be petroleum
product water was removed under reduced pressure (10 Torr) at ethers and chloroform [39]. However, the efficiency of these
equilibrium. However, the solketal yield never exceeded 80%, entrainers is not great either because their boiling points are still
which was the major disadvantage of this process. In addition, a higher than acetone. Acetone co-distillation creates the problem of
designed apparatus is required to work under reduced pressure to low efficiency in azeotropic water removal. This phenomenon was
conduct the experiment. A very similar process was reported in evident from its very long reaction time when using petroleum
literature where the authors heated glycerol with an excess of ether as entrainer [16]. The use of phosphorous pentoxide and
acetone over pTSA and reported a maximum of 56% solketal yield sodium sulfate as catalysts as well as desiccants for the removal of
[35]. The low solketal yield is ascribed to the presence of water in water generated from the system has also been reported [39], but
the reaction. high consumption of the catalysts in this case increased the
The major issue in the ketalization reaction of glycerol is the for- operation costs. More recently, molecular sieves were used for this
mation of by-product water, which creates a thermodynamic and purpose [40]. All these processes are not economical on an
kinetic barrier for high glycerol conversion to solketal. Different industrial scale.
processes were developed to overcome this issue. The above problems could be addressed more effectively by
Mushrush et al. studied the ketalization reaction using toluene using excess acetone, which not only acts as a reactant but acts as
as solvent [36]. In their experiment, 4.5 mol (232 g) of acetone was an entrainer for the system. The excess acetone could be distilled
added to 1.1 mol (100 g) of glycerol and 3.0 g of pTSA with 255 g of off and reused in the same or other processes. Roldan et al.
5 Å molecular sieves in a two neck round-bottomed flask (2 L), modified the batch reactor to a membrane batch reactor to remove
equipped with a mechanical stirrer and a refluxing condenser. The the water from the reaction system [41]. The authors conducted
reaction mixture was heated under gentle reflux for 33 h using a the experiment by refluxing a mixture of glycerol, anhydrous
heating mantle. The acidic reaction mixture was then neutralized acetone and heterogeneous acid catalyst, Montmorillonite K-10
with 3.0 g of sodium acetate and distilled to give solketal at a yield (total weight 1 g) in a three-neck flask (250 mL) equipped with a
of 88%. reflux condenser, a septum cap and a zeolite membrane fixed in
Garcia et al. studied the reaction with acetone-to-glycerol the central mouth (Fig. 1). The membrane allowed permeation of
molar ratio (A/G) of 3 over pTSA monohydrate [37]. The mixture small sized water vapor instead of pervaporation. A maximum
was heated to reflux for 16 h. During the process wet acetone was solketal yield of 82% was achieved by the authors using a very high
distilled off and dry acetone was simultaneously introduced to the A/G (20:1) for 2 h of reaction. As expected, a negligible effect of the
reactor to maintain the liquid concentration. The yield of solketal catalyst on the solketal yield was observed in this work.
was about 90% and no purification was required after solvent Recently, Vicente et al. attempted to remove water con-
removal. Considering the fact that pTSA monohydrate is soluble in tinuously from the reaction system by carrying out the reaction in
the reactants, the process can be classified as homogeneous cat- a two-step batch mode operation [42]. In the first step, the reac-
alysis, which causes a difficulty for catalyst recovery-typical tion mixture (glycerol, acetone and a heterogeneous catalyst) was
drawback of reaction systems employing homogeneous catalysts. stirred under reflux at 70 °C in a 100 mL flask and in the second
In fact, the use of homogeneous acid catalysts for chemical reac-
tion processes has many serious shortcomings in addition to cat-
alyst recovery, such as corrosion of the reactor, and the environ-
mental and economic concerns over the effluent disposal. Hence, it
is of significance to explore heterogeneous acid catalysts for the
glycerol ketalization process. Deutsch et al. reported the use of
Amberlyst-36 (an arenesulfonic acid polymer)-a heterogeneous
acid catalyst in a batch reactor with organic solvent (dichlor-
omethane) [24]. The authors conducted the experiment in the
presence of the solid catalyst in a 100 mL flask equipped with a
refluxing condenser. A Dean-Stark trap was used to remove the
formed water continuously. The maximum yield of solketal was
88% (w.r.t. glycerol) (reaction conditions: 0.1 mol glycerol, 0.15 mol
acetone, 17.5 mol dichloromethane, 0.5 g Amberlyst-36, 8 h reac-
tion time at room temperature).
It is well known that the ketalization reaction has a very low
equilibrium constant [37]. Therefore, to get high conversions of
glycerol it is necessary to shift the equilibrium towards the for-
mation of solketal. This could be achieved by either feeding excess Fig. 1. Membrane reactor for synthesis of solketal (adopted with copyright per-
amount of acetone or by removing the water generated during the mission [40]).
 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031 1025

step, the water produced along with the excess amount of acetone 2.3. Development of continuous processes
was removed by vaporization under vacuum at 70 °C and fresh
acetone was added to maintain the liquid level to start a new cycle. As discussed earlier, the majority of the studies on synthesis of
After three consecutive cycles (each cycle has two steps), a max- solketal were operated in batch reactors although using hetero-
imum solketal yield of 90% was achieved under the following geneous catalysts such as Zeolites, Amberlysts, montmorillonite,
reaction conditions; 70 °C, 5 wt% (w.r.t. glycerol) loading of Ar- silica induced heterolpoyacids, nafion, etc. [41–43,30] However, a
sBA-15 catalyst, and 30 min for each step. batch process has various limitations of which the main ones are a
To search for an effective heterogeneous catalyst for the keta- long time of reaction (usually exceeding 2 h) hence relatively
lization process, Ferreira et al. studied the reaction in a stirred lower efficiency, and the difficulty in process scale-up [46]. Pro-
batch reactor over a series of silica-induced heteropolyacid cata- duction of solketal in a continuous-flow reactor using hetero-
lysts, i.e., tungsto-phosphoric acid (PW), tungsto-silisic acid (SiW), geneous catalysts is thus much more advantageous because the
molybdo-phosphoric acid (PMo), and molybdo-silisic acid (SiMo) continuous-flow process enables better heat and mass transfer
[43]. The reported catalytic activities for the catalysts are in the efficiency, and easy scaling-up of the process from laboratory to
order of: SiMooPMo oSiWSo PWS, mainly owing to the increase industrial scale as well as more environmental and economical
in acidity [43]. The authors reported glycerol conversion of more benefits [47–50]. The continuous operation of the process also
than 97% with a very high selectivity of 99% towards solketal at the offers constant quality of the end product.
reaction conditions: 70 °C, A/G of 12:1, catalyst (PW) loading of The use of a continuous microwave reactor (CMR) for the
0.2 g, and 2–3 h. The high yield of solketal in this work was synthesis of solketal was reported [51]. In this CMR process, a
attributed to the strong acidity of the catalyst that promoted the solution of acetone, glycerol and pTSA as a homogeneous catalyst
reaction kinetics and to the high A/G (12:1). Good catalytic stabi- was mixed and pumped into the reaction coil (inside the micro-
lity was also observed, as the catalyst lost its activity by  15% after
wave cavity) to react at a desired temperature (process similar to
four consecutive batch runs using the same catalyst.
Fig. 3). The authors reported a maximum 84% yield of solketal at A/
Glycerol is poorly miscible with acetone in normal conditions
G of 13.5, in the presence of pTSA under the reaction conditions of
(25 °C and 1 atm) (only 5 wt% of glycerol is soluble in acetone),
132 °C, 1175 kPa, 1.2 min residence time and of 20 mL/min feeding
which is the major disadvantage for the synthesis of solketal.
rate. However, the system was restricted only to homogeneous
Royon et al. proposed to use the supercritical acetone with better
catalysts. Moreover, this technique would not be appropriate for
solubility for glycerol to synthesize solketal without using any
conducting the reaction at a low temperature or for reactants that
catalyst [44]. The authors carried out the experiment at 508 K and
are not compatible with microwave energy.
48 bar in a batch reactor, where acetone was at its supercritical
Clarkson et al. used a multi-tray reactive distillation column with
state. However, a maximum of 28% glycerol conversion with a
deep reaction stages containing catalyst (Amberlyst DPT-1) in sus-
selectivity of 80% towards solketal was observed after 4 h reaction
pension for the synthesis of solketal [52], as illustrated in Scheme 3.
at the A/G of 410. The low glycerol conversion and solketal yield
In their process, glycerol was preheated at 90 °C before feeding into
might be due to the lack of active acid sites in acetone at super-
critical condition. Hence, the result was not very encouraging. the reaction column. An extra amount of acetone was added in the
Since ketalization is an exothermic process [25], temperature is reaction stage to drive the reaction towards the production of
another important factor that affects the equilibrium conversion. solketal and the process has a long reaction time (more than 4 h).
To seek highly active catalysts at low temperature is another With this, the process is actually a semi-continuous process (con-
strategy to enhance the economy of the solketal production. tinuous operation with respect to acetone, but batch mode for
Menezes et al. reported the highest ever solketal yield obtained in glycerol). The process was found to be difficult to operate at a lower
a batch reactor (95%) at ambient conditions [45], over 10 mol% of temperature due to the high viscosity of glycerol. A continuous glass
stannous chloride (SnCl2) (w.r.t. glycerol) by reacting 6 M ratio of flow reactor (Fig. 2) made of several glass fluidic modules and
A/G for 0.5–2 h in presence of methyl cyanide (CH3CN) solvent. connected in series has been reported by Monbaliu et al. [53]. In
Table 1 presents a summary of the performance of various cata- their work, the total volume of the reactor is 72 mL and the first two
lysts for ketalization in batch reactors. From the Table, irrespective fluidic modules (FM01 and FM02) were used for feeding, preheating
of the catalysts, a usual long reaction time was observed (0.5–33 h) and premixing of the reactants. Glycerol (feed 1) was preheated (on
for the solketal yield in the range of (82–96%). However, all the FM01) and reacted with acetone in all other modules (FM03–FM09)
batch processes described above have common limitations in for the solketal product. Acetone (feed 2) and sulfuric acid (feed 3)
terms of the difficulty in scaling up for production of solketal on a were premixed and preheated in the fluidic module FM02. The
large scale. Thus, the advances in glycerol ketalization with main challenges of this reactor system include: a high residence
continuous-flow processes are discussed in the following section. time of the reactants, unsuitable for using heterogeneous catalysts,

Table 1
Performance of various catalysts for glycerol ketalization in batch reactors.

Experimental conditionsa Catalyst Glycerol conversion (%) Solketal selectivity (%) Solketal yield (%) Ref.

b
Temperature (˚C) A/G ratio Reaction time (h)

Refluxed 4 33 pTSA 98 88 35
Refluxed 3 16 pTSA – 90 36
38–40 1.5 8 Amberlyst 36 89 99 88 24
Refluxed 20 2 Montmorillonite K10 83 99 82 24
70 6 42.5 Ar-SBA 15 91 99 90 42
70 12 2–3 PW 99 97 96 43
235 11 4 – 28 80 22 44
Ambient 6 0.5 SnCl2 97 98 96 45

a
Reaction pressure is not available.
b
Acetone-to-glycerol mole ratio; Ar-SBA-15: Arenesulfonic acid functionalized mesostructured silica; PW: Tungstophosphoric acid; pTSA: p-Toluenesulfonic acid.
1026 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031

Glycerol
Reactor Profile
Acetone/water
(preheated to 90 ˚C) (to fractionator) Reaction mixture composition Temperature
(wt%) Conversion rate
G A W S
Feed stage 1 73.00 10.00 17.00 0.00
Reboiler Feed stage 2 74.69 12.00 17.00 2.31
Reaction stage 1 66.04 20.00 6.00 7.96
Reaction stage 2 53.86 26.00 3.00 17.14
Reaction stage 3 14.38 35.00 1.80 48.82
Reaction stage 4 3.67 45.00 1.20 50.13
Reaction stage 5 1.11 45.00 0.80 53.09
Reaction stage 6 0.33 45.00 0.50 54.17
Reaction stage 7 0.10 45.00 0.20 54.70
Reaction stage 8 0.03 45.00 0.20 54.77
Reaction stage 9 0.02 45.00 0.20 54.78
Acetone
(preheated to 70 ˚C) Reaction stage 10 0.01 45.00 0.20 54.79
Distillation stage1 0.02 4.50 0.10 95.38
Distillation stage 2 0.02 0.45 0.00 99.53
Distillation stage 3 0.02 0.05 0.00 99.94

0 40 80 120 160
Solketal
(product) Conversion rate(%stage)
Temperature(˚C)

Scheme 3. Schematic diagram of a multi-tray reaction column for glycerol ketalization adopted from Clarkson et al. [51].

Feed 1’
FM01

Feed 1

Feed 3 FM02 FM03 FM04 FM05 FM06 FM07 FM08 FM09

Feed 2 Product

He in

He out

Fig. 2. Schematic diagram of a continuous glass flow reactor developed by Monbaliu et al. [52] (used with copyright permission).

difficulty in conducting the reaction at low temperature, and

separation issues for the final product after neutralization, etc.
P
Maksimov et al. reported a continuous reactor for the pre-
Reactor
paration of high-octane oxygenate fuel components from plant-
derived polyols, however no description of the reactor was given
in the literature [30]. Recently, Nanda et al. have developed a P
Pump
continuous-flow reactor based on the concept of “Novel Process Feedstock

Windows” with respect to temperature, pressure and/or reactant

concentration to enhance the intrinsic kinetics of the reaction for Product
collector
an optimum yield [54–57]. The reactor is a continuous down-flow
tubular reactor (Inconel 316 tubing, 9.55 mm OD, 6.34 mm ID and Fig. 3. Process scheme of continuous-flow reactor reported in literature [51,54].
600 mm length) heated with a tube furnace. A similar flow process
obtained [58]. A summary of the performance of various catalysts
diagram is given in Fig. 3. The feed, a homogeneous solution of
for glycerol ketalization in flow reactors is given in Table 2. From
reactants (acetone and glycerol) with the solvent (ethanol) mixed
the table, it is clear that high solketal yields (84–94%) are feasible
at a selected molar ratio, was pumped into the reactor using a
in the range of 0.02–0.5 h reaction time in a continuous reactor.
HPLC pump at a specific flow rate. The reactor was maintained at a
Therefore, development of continuous-flow processes is promising
desired temperature and pressure. In each run, a pre-determined
for production of solketal on a large scale.
amount of catalyst was preloaded into the catalytic bed, where the
catalyst particles were supported on a porous Inconel metal disc
(pore size: 100 mm) and some quartz wool. The amount of catalyst
in each run was determined by the selected weight hourly space 3. Catalysis – the important parameter in ketalization reaction
velocity (WHSV). This flow reactor can operate in a wide range of
temperature and pressure using different heterogeneous catalysts. 3.1. Influence of catalyst acidity
Amberlyst-36 wet was used to optimize the process, and the
optimum process conditions are: 25 °C, 500 psi, A/G of 4, WHSV of As discussed later, the ketalization reaction proceeds via an
2 h  1, under which a very high yield of solketal (94 72 wt%) was acid catalyzed mechanism, hence catalysts with stronger acidity
 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031 1027

Table 2
Performance of various catalysts for glycerol ketalization in continuous reactors.

Experimental conditions Catalyst Glycerol conversion (%) Solketal selectivity (%) Solketal yield (%) Ref.

T (˚C) P (psi) A/G ratioa Reaction time (h)

132 170 13.5:1 0.02 pTSA – – 84 51

40 600 6:1 0.25 H-β zeolite 87 98 84 54
40 600 6:1 0.25 Amberlyst 36 wet 91 97 88 54
40 600 6:1 0.25 Amberlyst 35 90 97 86 54
40 600 6:1 0.25 ZrSO4 79 97 77 54
25 500 4:1 0.5 Amberlyst 36 wet 97 97 94 54

pTSA: p-Toluenesulfonic acid;

a
Acetone-to-glycerol mole ratio.

been reported by Ferreira et al. in ketalization of glycerol by

Table 3
Influence of catalyst acidity on solketal yield.
acetone [43].

Active phase Reaction Acidity BET Pore Yield (%) Ref. 3.2. Development and performance of transition metal catalysts
conditions (meq/g) (m2/g) size
(˚C, A/G, Tr) (nm)
Transition metal catalysts have demonstrated good catalytic
H-β Zeolite 40, 6:1, 0.25 5.7 480 2 84 54 performance in glycerol ketalization [59]. In fact, iridum catalyzed
Amberlyst-36 wet 40, 6:1, 0.25 5.6 33 24 88 54 ketalization reactions are promising and have been well studied
Amberlyst 35 40, 6:1, 0.25 5.4 35 16.8 86 54 among other transition metal catalysts [60–64]. The most active
ZrSO4 40, 6:1, 0.25 – – – 77 54
catalyst for the ketalization reaction was [CpIrCl2]2 (Cp¼ penta-
Polymax 40, 6:1, 0.25 – – – 60 54
Montmorillonite K10 40, 6:1, 0.25 4.6 264 5.5 68 54 methylcyclopentadienyl) [59], with a glycerol conversion of 87% and
Amberlyst 36 38–40, 1.5:1, 5.4 19 20 88 24 98% selectivity towards solketal in a batch reactor (other experi-
8 mental conditions were: 40 °C, [Ir]¼3.0  10  3 M, [glycerol]/[Ir]¼
Pr-SBA-15 70, 6:1, 0.5 0.94 721 8 79 42
500, and 1 h reaction time). Li's group specifically studied the per-
Ar-SBA-15 70, 6:1, 0.5 1.06 712 9 83 42
HAr-SBA-15 70, 6:1, 0.5 1.04 533 8 80 42 formance of mesoporous substituted silicates [65], in which the
Amberlyst 15 70, 6:1, 0.5 4.8 53 30 85 42 metal atoms were incorporated in the silicate framework. The
Pr-SO3H-SiO2 70, 6:1, 0.5 1.04 301 2–20 77 42 authors reported that the Zr-TUD-1 and Hf-TUD-1 were prepared by
T-SiO2 70, 6:1, 0.5 0.78 279 2–20 73 42 a one-pot sol–gel procedure, where triethanolamine was used as
SAC-13 70, 6:1, 0.5 0.12 4200 4 10 74 42
chelating and template agent and zirconium propoxide and haf-
Pr-SBA-15: Propylsulfonic acid-functionalized mesostructured silica; Ar-SBA-15: nium chloride as the metal precursors. Another catalyst Sn-MCM-41
Arenesulfonic acid-functionalized mesostructured silica; HAr-SBA-15: Hydro- was prepared by hydrothermal synthesis in a procedure similar to
phobised arenesulfonic acid-functionalized mesostructured silica; SAC-13: Nafion that of Li et al. [65], using cetyltrimethylammonium bromide (CTAB)
silica composite; T-SiO2: Silica bonded tosic acid; Pr-SO3H-SiO2:Silica bonded
as the template in a gel formed from a solution of tetraethyl
propylsulfonic acid.
orthosilicate (TEOS), SnCl4  5H2O and tetraammonium silicate [66].
The conversion of glycerol reached around 64%, 65% and 62% for Zr-
(relatively more number of acid sites per unit mass) might lead to
TUD-1, Hf-TUD-1 and Sn-MCM-1 catalysts, respectively, with almost
higher glycerol conversion. The influence of catalyst acidity on the
100% selectivity towards solketal in a batch reactor under the
solketal yield is shown in Table 3. It is clear that the catalyst acidity
experimental conditions of: 80 °C, 6 h reaction time, and A/G of 2:1.
is a crucial parameter influencing the catalytic performance.
Vicente et al. compared the performance of a series of catalysts
with different acidities (ranging from 0.12 to 4.8 meq/g) for keta- 4. Reaction models
lization of glycerol: propylsulfonic acid-functionalized mesos-
tructured silica (Pr-SBA-15), arenesulfonic acid-functionalized Establishing reaction paths for any process and for ketalization
mesostructured silica (Ar-SBA-15), hydrophobised arenesulfonic in particular is very crucial in the design of a catalyst. In addition,
acid-functionalized mesostructured silica (HAr-SBA-15), Amber- establishing the reaction rate equations helps in designing the
lyst-15, silica bonded-propylsulfonic acid (Pr-SO3H-SiO2), silica reactor as well. The reaction mechanism and the kinetic models
bond-tosic acid (T-SiO2), and Nafion silica composite (SAC-13) [42]. developed for the glycerol ketalization process are discussed below.
They obtained a solketal yield of 74% for SAC-13 catalyst (acid
strength 0.12 meq/g) and 85% for Amberlyst-15 (acidity 4.8 meq/ 4.1. Reaction mechanism
g). Thus, a catalyst with a stronger acidity would likely perform
better in the ketalization of glycerol with acetone. On the other As discussed previously, the relative acidity of the catalysts has
significant effects on the glycerol conversion and the product yield. It
hand, the results as shown in the table imply that surface area and
is thus of significance to discuss the reaction mechanism for the
the pore volume/size of a catalyst have negligible influence on the
glycerol ketalization reaction catalyzed by acid catalysts. The con-
catalytic activity for the ketalization of glycerol. A recent study by
densation reaction of glycerol with acetone leads to the formation of
Nanda et al. also revealed similar results in a continuous flow both five membered and six membered rings (ketals) [67]. However
reactor [54]. The authors observed that the activity of catalysts was the six membered ring ketal is less favorable because one of the
in the order of Amberlyst wetEH-beta zeoliteE Amberlyst methyl groups in the final product is in axial position of the chair
dry4 zirconium sulfate4montmorillonite4Polymax, which fol- conformation (Fig. 4) [30,68]. So the resulting product has a ratio of
lows the same order of the catalytic acidity (Table 3). Similar 99:1 for five membered ring (4-hydroxymethyl-2,2-dimethyl-1,3-
correlation between the catalyst acidity and the product yield has dioxolane, or solketal) to six membered ring (5-hydroxy-2,
1028 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031

OH
OH

O O
O O
Scheme 4. Mechanism proposed by Li et al. [64] for the reaction of acetone and
glycerol over Lewis acid catalyst (M is the metal atom).
CH3 H3C CH3
CH3 parameters (kinetic constant; k and water adsorption constant; Kw
Fig. 4. The cyclic acetals from the reaction between glycerol and acetone: and ketalization equilibrium constant; Kc) are to be estimated at
5-hydroxy-2,2-dimethyl-1,3-dioxane (a) solketal i.e 4-hydroxymethyl-2,2-dime- each temperature to find the rate of the reaction. The estimated
thyl-1,3-dioxane (b). values of these parameters are given in Table 4. Based on the var-
iation of kinetic constant with temperature, the activation energy
2-dimethyl-1,3-dioxane). For the ketalization reaction catalyzed by (Ea) of the reaction has been reported to be 55.673.1 kJ mol  1.[38]
Brønsted acids, the five membered ring solketal is dominantly
formed through a mechanism involving a short-lived carbenium ion
as an intermediate [65,69]. Li and co-workers proposed a similar 5. Key operation issues of flow reactors and use of crude gly-
mechanism for the ketalization reaction over Lewis acid catalysts cerol as feedstock
[65]. According to this mechanism, the Lewis acid metal sites play a
role similar to the MPV reduction (Meerwein–Ponndrof–Verley) or As discussed earlier, the ketalization reaction proceeds via an acid
Oppenauer oxidation reactions, by coordinating and activating the catalyzed mechanism, which means catalysts with stronger acidity
carbonyl group of the acetone. Then the carbon atom of the carbonyl might lead to higher glycerol conversion. However, catalysts with
group is attacked by the primary alcoholic group of glycerol accom- strong acidity would enhance fouling. Nevertheless, since the reaction
panied by the formation of a bond between the carbonyl oxygen is exothermic and carried out at a low temperature (usually below
atom and the secondary carbon atom of glycerol followed by dehy- 80 °C), the deactivation of catalyst due to fouling can be avoided.
dration to form the five membered ring solketal. The detail Nanda et al. investigated the catalytic deactivation process of different
mechanism is displayed in Scheme 4. heterogeneous acid catalysts such as H-beta zeolite, Amberlyst-35 dry
Nanda et al. have also used a reaction framework (Scheme 5) for and Amberlyst-36 wet in a continuous-flow reactor and observed a
the ketalization reaction proceeding via acidic catalytic mechanism slight reduction in the activities of these catalysts after 24 h on-stream
involving 3 steps [38]. The first step involves the surface reaction as compared to that of the fresh catalyst [54]. To better understand
between the adsorbed acetone and glycerol over the catalyst surface to these phenomena, they measured the textural properties and acidity
form the hemi-acetal. The next step is the removal of water leading to of the spent catalyst (Amberlyst-36 wet) after 24 h on-stream and
the formation of a carbocation on the carbonyl carbon atom, and the compared to the results of the fresh catalyst. The slight reduction in
last step is the removal of the proton to form solketal. the activity of the spent catalyst was attributed to the loss of active
acid sites during the reaction, not due to fouling. In order to regain the
4.2. Reaction kinetics initial activity of the catalyst, the spent catalyst was regenerated and
the regenerated catalyst demonstrated almost the same activity
The general reaction rate for the ketalization reaction has been (493% yield ) as that of the fresh catalyst [70]. However, after a long
expressed in form of Langmuir–Hinshelwood model with surface time (days or months) operation of a continuous-flow reactor using
reaction as the rate determining step [38]. The key reaction steps heterogeneous catalysts, reactor clogging might occur, caused by fine
of this model are given as follows: particles of disintegrated catalysts [54]. This problem can be effectively
alleviated by diluting the catalyst with glass beads and/or by
a) The surface reaction between the adsorbed species of glycerol decreasing the catalytic bed height.
(GF) and acetone (AF) to give adsorbed hemiacetal (HF) The price of glycerol depends on the technical grade. The
GF þ AF2HF þ F ð1Þ refined pure glycerol is currently expensive, costing around US$
500–600 per ton [70]. Crude glycerol is available for only US$ 40–
90 per ton [9]. Thus, use of crude glycerol for the production of
where F is the vacant site on the catalyst.
value-added products is crucial for achieving a sustainable and
b) Surface reaction for formation of adsorbed water (WF)
economical production of solketal. However, as mentioned earlier,
HF þ F2IF þ WF ð2Þ crude glycerol contains impurities including water, potassium or
sodium salts, esters, fatty acids and alcohols. Therefore, the direct
where IF represents for the reactive intermediate formed. use of crude glycerol as feedstock may cause problems such as
c) Formation of adsorbed solketal (SF) deactivation of catalyst (by poisoning the active sites by the
impurities) or plugging of reactor (due to deposition of high
IF þ GF2SF þF ð3Þ
boiling organic compounds or inorganic salts). To facilitate the use
The simplified rate expression for the reaction is given as [38]: of crude glycerol, da Silva and Mota investigated the effect of
½G½A  ½S½W=K c ½G impurities on the production of solketal in a batch reactor [71].
r¼k  2 ð4Þ They added impurities such as 10% water, 15% NaCl and 1%
1 þ K w ½W  methanol (assuming that these are the common impurities pre-
sent in crude glycerol) to pure glycerol and conducted the ketali-
where Kw is the equilibrium constant for water adsorption on the zation experiment in presence of heterogeneous catalysts such as
catalyst surface. According to the above kinetic model three Amberlyst-15 and H-beta zeolite. They observed significant
 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031 1029

Scheme 5. Mechanism used by Nanda et al. [37] for the reaction of acetone and glycerol over acid catalyst.

Table 4 catalyst (Amberlyst-36 wet) could be successfully regenerated for four

Kinetic model parameters at different temperatures. consecutive cycles (96 h) with acceptable reduction in the solketal
Temperature (K) KC k (L mol  1 min  1) KW
yield (from 92% to 81%) [70]. For the regeneration of the catalyst
(Amberlyst-36 wet) in the guard reactor, a 0.5 M sulfuric acid solution
298 2.66 0.11 2.65 was used to pass through the guard reactor, followed by washing the
303 1.82 0.16 1.50
regenerated catalysts with methanol solution and drying the bed with
308 1.51 0.24 1.09
313 1.30 0.33 0.73 nitrogen for 5 h.
323 0.98 0.63 0.34

Kc: Equilibrium constant for the reaction; k: Kinetic constant; KW: Equilibrium
constant for water adsorption on the catalyst surface 6. Conclusions

This review paper over-views the state-of-the art of the sus-

100
tainable production of solketal by catalytic reaction of glycerol
with acetone. Different types of processes and catalysts developed
and their performances are compared. Fundamentals of reaction
80 mechanisms and kinetics for the acid-catalyzed conversion of
glycerol into solketal are presented. The main operation issues
Solketal yield(%)

related to catalytic conversion of crude glycerol in a continuous-

60 flow process and the direct use of crude glycerol are discussed.
1 wt% NaCl +2 wt% H2O

Some key conclusions are summarized below:

1 wt% Nacl + ethanol

2 wt% H2O + ethanol

(1) Conversion of glycerol to solketal can proceed either using a

40
homogeneous or heterogeneous catalyst; nevertheless the use
of heterogeneous catalysts is preferred, as there are many
Methanol

+ ethanol

shortcomings for using homogeneous catalysts, e.g., difficulty

Ethanol

20 in catalyst recovery, corrosion to the reaction systems, and the

environmental and economical concerns over the effluent
disposal. Hence, it is of significance to explore heterogeneous
acid catalysts for the glycerol ketalization process.
0
(2) The ketalization reaction has a very low equilibrium constant.
A B C D E
In order to reach high conversions of glycerol it is necessary to
Fig. 5. Deactivation of catalyst by impurities in the glycerol feed. shift the equilibrium towards the formation of solketal, by
either feeding excess amount of acetone or by removing the
reduction in glycerol conversion (from 95% to 47% for Amberlyst- water generated during the reaction continuously.
(3) All the batch processes have common limitation in terms of
15 and from 90% to 50% using H-beta zeolite) while switching the
the difficulty in scaling up for production of solketal on a large
feed from pure glycerol to the impurities-added glycerol. A similar
scale. Compared with operation in a batch reactor, a
result has also been observed by Nanda et al. in a continuous-flow
continuous-flow process results a similar product yield with
reactor (Fig. 5) [58].
relatively shorter reaction time. Therefore, development of
Nanda et al. moved one step further and developed a modified
continuous-flow processes is promising for production of
continuous-flow reactor consisting of guard reactors allowing online solketal from glycerol on a large scale.
removal of impurities in the glycerol feedstock and online regenera- (4) The best yields of solketal were achieved by catalysts like
tion of deactivated catalysts (Fig. 6) [70]. Using crude glycerol and the Amberlyst-15, Amberlyst-35, Amberlyst-36, Ar-SBA-15, Zeo-
modified continuous-flow reactor, a significant yield of solketal lites, and SnCl2. The preferred reaction conditions are: cata-
(78%) was obtained after 1 h on-stream. Moreover, the authors have lysts with higher acidity, higher A/G, and lower temperature
carried out an on-line regeneration of the deactivated catalysts in the (o70 °C). Using Amberlyst-36 wet catalyst, a very high yield
guard reactor and ketalization experiment simultaneously using pur- of solketal (94 72 wt%) was obtained at 25 °C, 500 psi, A/G of
ified crude glycerol (E96% purity) as the feedstock and found that the 4, and WHSV of 2 h  1.
1030 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031

Fig. 6. Novel flow-reactor consisting of guard reactors allowing online removal of impurities in the glycerol feedstock and online regeneration of deactivated catalysts.

(5) The ketalization reaction proceeds via acidic catalytic [3] Lin L, Cunshan Z, Vittayapadung S, Xiangqian S, Mingdong D. Opportunities
mechanism, hence catalysts with strong acidity might lead to and challenges for biodiesel fuel. Appl Energy 2011;88:1020–31.
[4] Johnson DT, Taconi KA. The glycerin glut : options for the value-added con-
high glycerol conversion. version of crude glycerol resulting from biodiesel production. Environ Prog
(6) Heterogeneous catalysts for glycerol ketalization in a 2007;26:338–48.
continuous-flow reactor can be deactivated, attributed to the [5] Organization for Economic Co-operation and Development (OECD) and the
loss of active acidic sites during the reaction, not due to Food and Agriculture Organization (FAO) of the United Nations. Available on:
〈http://www.agri-outlook.org/48202074.pdf〉; 2011–2020 [accessed 9.04.15].
fouling. For a long time (days or months) operation, however, [6] Thompson JC, He BB. Characterization of crude glycerol from biodiesel pro-
the reactor clogging might occur, caused by fine particles of duction from multiple feedstocks. Appl Eng Agric 2006;22:261–5.
disintegrated catalysts. [7] Liu X, Jensen PR, Workman M. Bioconversion of crude glycerol feedstocks into
(7) Direct use of crude glycerol as feedstock may cause problems ethanol by Pachysolen tannophilus. Bioresour Technol 2012;104:579–86.
[8] Liang Y, Cui Y, Trushenski J, Blackburn JW. Converting crude glycerol derived
such as deactivation of catalyst (by poisoning the active sites from yellow grease to lipids through yeast fermentation. Bioresour Technol
by the impurities) or plugging of reactor (due to deposition of 2010;101:7581–6.
high boiling organic compounds or inorganic salts). [9] Biodiesel magazine – the latest news and data about biodiesel production.
Available on: 〈http://www.biodieselmagazine.com/articles/1123/combating-
(8) The review article introduces various applications of solketal
the-glycerin-glut/〉 [accessed 30.01.15].
in different industries including polymer, pharmaceutical and [10] Haas MJ, McAloon AJ, Yee WC, Foglia TA. A process model to estimate biodiesel
cosmetics, food, and fuel industries, and highlights some production costs. Bioresour Technol 2006;97:671–8.
major challenges for industrial production of solketal. In [11] Callam CS, Singer SJ, Lowary TL, Hadad CM. Computational analysis of the
potential energy surfaces of glycerol in the gas and aqueous phases: effects of
addition, this review article demonstrates the promise of new
level of theory, basis set, and solvation on strongly intramolecularly hydrogen-
processes for utilization of crude glycerol as feedstock for the bonded systems. J Am Chem Soc 2001;123:11743–54.
production of solketal. [12] Da Silva GP, de Lima CJB, Contiero J. Production and productivity of 1,3-pro-
panediol from glycerol by Klebsiella pneumoniae GLC29. Catal Today
2015;257:259–66.
[13] Zhou C-HC, Beltramini JN, Fan Y-X, GQM Lu. Chemoselective catalytic con-
Acknowledgments version of glycerol as a biorenewable source to valuable commodity chemicals.
Chem Soc Rev 2008;37:527–49.
[14] Pagliaro M, Ciriminna R, Kimura H, Rossi M, Della Pina C. From glycerol to
The authors want to acknowledge the financial support pro-
value-added products. Angew Chem In Ed 2007;46:4434–40.
vided by Imperial Oil via University Research Award as well as the [15] Zhu S, Gao X, Zhu Y, Zhu Y, Zheng H, Li Y. Promoting effect of boron oxide on
NSERC Discovery Grant (RGPIN-2014-05463) and CFI-LOF awarded Cu/SiO2 catalyst for glycerol hydrogenolysis to 1,2-propanediol. J Catal
2013;303:70–9.
to Dr. Xu. We are also thankful to Professors Yasuo Ohtsuka and [16] Chen L, Liang J, Lin H, Weng W, Wan H, Védrine JC. MCM41 and silica sup-
Guus Van Rossum from Tohoku University (Japan) and University ported MoVTe mixed oxide catalysts for direct oxidation of propane to acro-
of Twente (the Netherlands), respectively, for their invaluable lein. Appl Catal A 2005;293:49–55.
[17] Hirai T, Ikenaga N, Miyake T, Suzuki T. Production of hydrogen by steam
suggestions on some aspects of this bibliographic study. reforming of glycerin on ruthenium catalyst. Energy Fuels 2005;19:1761–2.
[18] Cassel S, Debaig C, Benvegnu T, Chaimbault P, Lafosse M, Plusquellec D, et al.
Original synthesis of linear, branched and cyclic oligoglycerol standards. Eur J
Org Chem 2001:875–96.
[19] Kenar JA. Glycerol as a platform chemical: sweet opportunities on the horizon.
References Lipid Technol 2007;19:249–53.
[20] Len C, Luque R. Continuous flow transformations of glycerol to valuable pro-
[1] Subramaniam R, Dufreche S, Zappi M, Bajpai R. Microbial lipids from renew- ducts: an overview. Sustain Chem Process 2014;2:1–10.
able resources: production and characterization. J Ind Microbiol Biotechnol [21] Rastegari H, Ghaziaskar HS. From glycerol as the by-product of biodiesel
2010;37:1271–87. production to value-added monoacetin by continuous and selective ester-
[2] Behr A, Eilting J, Irawadi K, Leschinski J, Lindner F. Improved utilisation of ification in acetic acid. J Ind Eng Chem 2015;21:856–61.
renewable resources: new important derivatives of glycerol. Green Chem [22] Pagliaro M, Rossi M. Future of glycerol. 2nd ed. Cambridge: Royal Society of
2008;10:13–30. Chemistry; 2010.
 M.R. Nanda et al. / Renewable and Sustainable Energy Reviews 56 (2016) 1022–1031 1031

[23] Zheng Y, Chen X, Shen Y. Commodity chemicals derived from glycerol, an [55] Hessel V, Vural Gürsel I, Wang Q, Noël T, Lang J. Potential analysis of smart
important biorefinery feedstock. Chem Rev 2008;108:5253–77. flow processing and microprocess technology for fastening process develop-
[24] Deutsch J, Martin a, Lieske H. Investigations on heterogeneously catalysed ment: use of chemistry and process design as intensification fields. Chem Eng
condensations of glycerol to cyclic acetals. J Catal 2007;245:428–35. Technol 2012;35:1184–204.
[25] Agirre I, García I, Requies J, Barrio VL, Güemez MB, Cambra JF, et al. Glycerol [56] Razzaq T, Kappe CO. Continuous flow organic synthesis under high-tem-
acetals, kinetic study of the reaction between glycerol and formaldehyde. perature/pressure conditions. Chem Asian J 2010;5:1274–89.
Biomass Bioenergy 2011;35:3636–42. [57] Illg T, Löb P, Hessel V. Flow chemistry using milli- and microstructured
[26] Silva MTM, Rodrigues AE. V. Synthesis of diethylacetal: thermodynamic and reactors-from conventional to novel process windows. Bioorg Med Chem
kinetic studies. Chem Eng Sci 2001;56:1255–63. 2010;18:3707–19.
[27] Barros AO, Faísca AT, Lachter ER, Nascimento RSV, San Gil RAS. Acetalization of [58] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu C (Charles). Cata-
hexanal with 2-ethyl hexanol catalyzed by solid acids. J Braz Chem Soc lytic conversion of glycerol to oxygenated fuel additive in a continuous flow
2011;22:359–63. reactor: process optimization. Fuel 2014;128:113–9.
[28] Pariente S, Tanchoux N, Fajula F. Etherification of glycerol with ethanol over [59] Crotti C, Farnetti E, Guidolin N. Alternative intermediates for glycerol valor-
solid acid catalysts. Green Chem 2008;11:1256–61. ization: iridium-catalyzed formation of acetals and ketals. Green Chem
[29] Mota CJ a, da Silva CX a, Rosenbach N, Costa J, da Silva F. Glycerin derivatives 2010;12:2225–31.
as fuel additives: the addition of glycerol/acetone ketal (solketal) in gasolines. [60] Sülü M, Venanzi LM. Acetalization and transacetalization reactions catalyzed
Energy Fuels 2010;24:2733–6. by ruthenium, rhodium, and iridium complexes with {2-{{Bis[3-(tri-
[30] Maksimov a L, Nekhaev a I, Ramazanov DN, Arinicheva Y a, Dzyubenko a a, fluoromethyl)phenyl]phosphino}methyl}-2-methylpropane- 1,3-diyl}bis[bis
Khadzhiev SN. Preparation of high-octane oxygenate fuel components from [3-(trifluoromethyl)phenyl]phosphine] (MeC[CH2P(m-CF3C6H4)2]3). Helv Chim
plant-derived polyols. Pet Chem 2011;51:61–9. Acta 2001;84:898–907.
[31] Fischer E. Ueber die verbindungen der zucker mit den alkoholen und ketonen. [61] Jiang Q, Rüegger H, Venanzi LM. Some new chain-like terdentate phosphines,
Eur J Inorg Chem 1895;28:1145–67. their ruthenium(II) coordination chemistry and the activity of the cations [Ru
[32] Fischer E, Pfahler E. Uber glycerin-aceton und seine verwendbarkeit zur rein- (MeCN)3(PhP{CH2CH2P(p-X-C6H4)2}2)]2 þ (X ¼ H, F, Me and OMe) as acet-
darstellung von α-glyceriden: uber eine phosphorsaure verbindung glykois. alization catalysts. Inorg Chim Acta 1999;290:64–79.
Eur J Inorg Chem 1920;53:1606–21. [62] Barbaro P, Bianchini C, Oberhauser W, Togni A. Synthesis and characterization
[33] Newman MS, Renoll M. Improved preparation of isopropyledene glycerol. J of chiral bis-ferrocenyl triphosphine Ni(II) and Rh(III) complexes and their use
Am Chem Soc 1945;67 1621–1621. as catalyst precursors for acetalization reactions. J Mol Catal A Chem
[34] Bruchmann Bernd, Haberle Karl, Helmut Gruner MH. Preparation of cyclic 1999;145:139–46.
acetals and ketals. Patent US5917059; 1999. [63] Gorla F, Venanzi LM. Cationic palladium(II), platinum(II), and rhodium
[35] Mushrush GW, Hardy D. Fuel system icing inhibitor and deicing composition. (I) complexes as acetalisation catalysts. Helv Chim Acta 1990;73:690–7.
Patent US5705087 A; 1998. [64] Cataldo M, Nieddu E, Gavagnin R, Pinna F, Strukul G. Hydroxy complexes of
[36] Mushrush GW, Stalick WM, Beal EJ, Basu SC, Slone JE, Cummings J. The palladium(II) and platinum(II) as catalysts for the acetalization of aldehydes
synthesis of acetals and ketals of the reduced sugar mannose as fuel system and ketones. J Mol Catal A Chem 1999;142:305–16.
icing inhibitors. Pet Sci Technol 1997;15:237–44. [65] Li L, Korányi TI, Sels BF, Pescarmona PP. Highly-efficient conversion of glycerol
[37] Garcı E, Laca M, Pe E, Garrido A. New class of acetal derived from glycerin as a to solketal over heterogeneous Lewis acid catalysts. Green Chem
biodiesel fuel component. Energy Fuel 2008;22:4274–80. 2012;14:1611–9.
[38] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu CC. Thermodynamic [66] Corma A, Navarro MT, Nemeth L, Renz M. Sn-MCM-41—a heterogeneous
and kinetic studies of a catalytic process to convert glycerol into solketal as an selective catalyst for the Baeyer–Villiger oxidation with hydrogen peroxide.
oxygenated fuel additive. Fuel 2014;117:470–7. Chem Commun 2001;21:2190–1.
[39] Ag B. Cyclic acetal or ketal preparation from poly:ol and aldehyde or ketone. [67] Vol’eva VB, Belostotskaya IS, Malkova a V, Komissarova NL, Kurkovskaya LN,
Patent DE19648960 A1; 1998. Usachev SV, et al. New approach to the synthesis of 1,3-dioxolanes. Russ J Org
[40] He DY, Li ZJ, Li ZJ, Liu YQ, Qiu DX, Cai MS. Studies on carbohydrates X. A new Chem 2012;48:638–41.
method for the preparation of isopropylidene saccharides. Synth Commun [68] Foster AB, Randall MH, Webber JM. 615. Aspects of stereochemistry. Part XX.
1992;22:2653–8. Reaction of 3-O-methyl-D-glucitol with acetone and benzaldehyde. J Chem
[41] Roldan L, Mallada R, Fraile JM, Mayoral JA, Menendez M. Glycerol upgrading by Soc (Resumed) 1965:3388–94.
ketalization in a zeolite membrane. Asia-Pac J Chem Eng 2009;4:279–84. [69] Garcfa H, Garcia OI, Fraile JM. Solketal: green and catalytic synthesis and its
[42] Vicente G, Melero J a, Morales G, Paniagua M, Martín E. Acetalisation of bio- classification as a solvent: 2,2-dimethyl-4-hidroxymethyl-1,3-dioxolane, an
glycerol with acetone to produce solketal over sulfonic mesostructured silicas. interesting green solvent produced through heterogeneous catalysis. Chim
Green Chem 2010;12:899–907. Oggi 2008;26:10–2.
[43] Ferreira P, Fonseca IM, Ramos a M, Vital J, Castanheiro JE. Valorisation of [70] Nanada MR, Catalytic Conversion of Glycerol to Value-Added Chemical Pro-
glycerol by condensation with acetone over silica-included heteropolyacids. ducts, Electronic Thesis and Dissertation Repository, Western University, 2015,
Appl Catal B 2010;98:94–9. London, Canada, Paper 3215.
[44] Royon D, Locatelli S, Gonzo EE. Ketalization of glycerol to solketal in super- [71] Da Silva CX a, Mota CJ a. The influence of impurities on the acid-catalyzed
critical acetone. J Supercrit Fluids 2011;58:88–92. reaction of glycerol with acetone. Biomass Bioenergy 2011;35:3547–51.
[45] Menezes FDL, Guimaraes MDO, da Silva MJ. Highly selective SnCl2 – catalyzed
solketal synthesis at room temperature. Ind Eng Chem Res 2013;52:16709–13.
[46] Green D, Perry R. Perry's chemical engineers' handbook. 8th ed. . Columbus,
OH: McGraw Hill Professional; 2007. Glossary
[47] Noël T, Buchwald SL. Cross-coupling in flow. Chem Soc Rev 2011;40:5010–29.
[48] Hartman RL, McMullen JP, Jensen KF. Deciding whether to go with the flow:
evaluating the merits of flow reactors for synthesis. Angew Chem Int Ed Biodiesel: vegitable oil/or animal fat based diesel having long chain alkyl esters;
2011;50:7502–19. Ketalization: reaction between a ketone and excess of alcohol to form a ketal;
[49] Chevalier B, Lavric ED, Cerato-Noyerie C, Horn CR, Woehl P. Microreactors for Valorization: route to enhance the value of the raw material;
industrial multi-phase applications : test reactions to develop innovative glass Solketal: a compound formed by the reaction between acetone and glycerol;
microstructure designs. Chim Oggi 2008;26:38–42. Transesterification: an organic process of exchanging the alkyl group between an
[50] Jähnisch K, Hessel V, Löwe H, Baerns M. Chemistry in microstructured reac- alcohol and an ester;
tors. Angew Chem Int Ed 2004;43:406–46. Crude glycerol: impure glycerol formed during the production of biodiesel;
[51] Cablewski T, Faux AF, Strauss CR. Development and application of a continuous Entrainer: reagent which can carry water into the vapor phase;
microwave reactor for organic synthesis. J Org Chem 1994;59:3408–12. Pervaporation: a process used to separated a mixture of liquids by partial vapor-
[52] Clarkson JS, Walker AJ, Wood MA. Continuous reactor technology for ketal for- ization through a membrane;
mation : an improved synthesis of solketal. Org Process Res Dev 2001;5:630–5. Homogeneous catalyst: catalyst that exists in the same phase as the reactants in the
[53] Monbaliu J-CM, Winter M, Chevalier B, Schmidt F, Jiang Y, Hoogendoorn R, reaction;
et al. Effective production of the biodiesel additive STBE by a continuous flow Heterogeneous catalyst: catalyst that exists in a different phase from the reactants
process. Bioresour Technol 2011;102:9304–7. in the reaction.
[54] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu C (Charles). A new
continuous-flow process for catalytic conversion of glycerol to oxygenated
fuel additive: catalyst screening. Appl Energy 2014;123:75–81.
 http://dx.doi.org/10.5935/0103-5053.20160066
J. Braz. Chem. Soc., Vol. 27, No. 10, 1832-1837, 2016.

Article
Printed in Brazil - ©2016 Sociedade Brasileira de Química
0103 - 5053 $6.00+0.00

Atmospheric Pressure Continuous Production of Solketal from the Acid-Catalyzed

Reaction of Glycerol with Acetone

Paula A. Oliveira,a Rodrigo O. M. A. Souzab and Claudio J. A. Mota*,a,c,d

a
Escola de Química, bGrupo de Biocatálise e Síntese Orgânica and cInstituto de Química,
Universidade Federal do Rio de Janeiro, 21949-901 Rio de Janeiro-RJ, Brazil
d
Instituto Nacional de Ciência e Tecnologia de Energia e Ambiente, Universidade Federal do Rio de
Janeiro, 21941-909 Rio de Janeiro-RJ, Brazil

A continuous-flow process at atmospheric pressure was designed for the conversion of glycerol
to solketal, an oxygenated fuel additive, through the acid-catalyzed reaction of glycerol with
acetone. Process optimization was performed by checking the influence of different variables
on the conversion and selectivity. The variables examined were: residence time (12, 24, 60 and
120 min), catalyst type (Amberlyst-15 and K-10 Montmorillonite), catalyst loading (7, 3, and
1 g), reaction temperature (50, 40 and 30 ºC), molar ratio of the reactants (1:2, 1:5, 1:10, 1:15 and
1:20) and solvent used to homogenize the system (dimethylsulfoxide or dimethylformamide).
The highest conversion (92%) was observed with 7.0 g of Amberlyst-15, reaction temperature of
50 °C, molar ratio of glycerol to acetone of 1:20 and dimethylformamide as solvent. In all cases,
solketal isomers (five and six-membered ring ketals) were the only product observed. The results
of solketal formation of this study, carried out at atmospheric pressure, were similar to other
studies with pressures of up to 120 bar. The utilization of higher catalyst loading and molar ratio
of reactants compensate the use of atmospheric pressure to achieve high conversion levels and
selectivity to the desired product.

Keywords: glycerol, solketal, flow chemistry, acid catalysis

Introduction sulfur and aromatic molecules, presents lower emission of

carbon monoxide and particulates and its cetane number is
Oil, natural gas and its derivatives are fuels that allow high, improving the engine efficiency.
the existence of fast and efficient means of transport, as In Brazil, biodiesel is presently blended with the
well as much of the industrial activities in the present days. petrodiesel in 7 vol%. A great challenge of the Brazilian
However, it is very likely that they will not last longer than biodiesel program is to find economical and environmentally
a few decades, because their reserves are finite, security of friend uses for the glycerol produced as byproduct. Personal
supply is problematic in many countries that import them care products, soaps, pharmaceuticals and food are the
and their use is the main source of CO2, a gas associated main sectors that make regular use of glycerol.3 However,
with climate changes and global warming. Thus, it is they cannot absorb alone, all the glycerol produced from
essential to search for substitutes for these fuels, especially the biodiesel industry. Thus, it is necessary to find new
from renewable organic matter (biomass). In this context, applications for this excess of glycerin produced by the
biodiesel appears as one of the most promising candidates. biodiesel industry.
Biodiesel is produced through the transesterification The use of glycerin as a raw material to produce
of vegetable oils or animal fat with methanol, under base other chemicals is gaining interest in recent years.4-7
catalysis conditions.1,2 In this process, glycerol or glycerin Hydrogenolysis over metal catalysts affords 1,2- and
is formed as byproduct in approximately 10 wt.%. This 1,3-propanediols, which can be used in the production of
biofuel can be used pure or blended with petrodiesel and polymers.8,9 Dehydration over acidic catalysts produces
features a number of environmental benefits. It contains no acrolein,10-13 an important intermediate in the synthesis of
acrylic acid and other chemicals. Synthesis gas, a mixture
*e-mail: [email protected] of CO and H2, can be produced from glycerol aqueous
Vol. 27, No. 10, 2016 Oliveira et al. 1833

phase reforming.14 Reaction with dimethyl carbonate can Solketal formation via glycerol ketalization with
lead to the formation of glycerol carbonate.15 Glycerol can acetone was studied in an Asia 110 continuous flow
also be converted in ethers, acetals/ketal and esters, all of system using a fixed bed of catalyst (Amberlyst-15 or
them potential fuel additives. For instance, the reaction K-10 Montmorillonite). Three packed bed columns, each
of glycerol with isobutene affords tert-butyl-glycerol with its respective volume (including the bed of catalyst),
ethers.16 Glycerol ethers can also be produced through were used. The column used with the highest catalyst
the acid-catalyzed reaction of glycerol with alcohols.17,18 loading (7.0 g) has a volume of 12.4 mL, the one used with
Glycerol acetals and ketals can be produced through 3.0 g of catalyst has a volume of 5.5 mL and the column used
the acid-catalyzed reaction with aldehydes and ketones, with 1.0 g of catalyst has 2.4 mL. The operational variables
respectively.19,20 These compounds can be used as fuel were flow rate, temperature, catalyst loading, type of
additives.21 Acetylation of glycerol affords the acetins or catalyst (Amberlyst-15 and K-10), molar ratio of reactants
glycerol acetates.22 and solvent used to homogenize the glycerol/acetone
We have recently shown23 that solketal, the ketal mixture. The solvent was used in the same molar proportion
produced in the reaction of glycerol with acetone, of glycerol in the mixture. The feed containing acetone,
improves the octane number and reduces gum formation glycerol and the solvent (dimethyl sulfoxide, DMSO or
in gasolines, either with and without ethanol. Therefore, it dimethylformamide, DMF) at room temperature was
can be a potential oxygenated gasoline additive, reducing pumped continuously at different flow rates into the reactor.
gum formation and improving the octane number. In Initially, only pure acetone was pumped through the column
addition, solketal is easily hydrolyzed in excess water,24 until the catalytic device reached the desired conditions.
indicating that it will not contaminate water sources, as The reaction was performed at proper temperature using
methyl‑tert‑butyl ether (MTBE) does, and this has been the heating plate device provided by the Asia system
the main reason for this additive to be phased out, because (USA) and atmospheric pressure. Eight samples, at the
it shows carcinogenic properties.25 end of each space time (0.2 h) processed, were taken and
Continuing our interest on the development of analyzed by capillary gas chromatography coupled with
continuous-flow protocols, 26 we wish to report an a mass spectrum (Agilent 5975, USA), operating with
atmospheric pressure continuous flow study for the electron ionization (EI) at 70 eV to obtain the conversion
production of solketal, by the acid-catalyzed ketalization and selectivity. A capillary column (30 m × 0.25 mm) with
of glycerol (Scheme 1). a 0.25 mm of methyl-phenyl-silicone stationary phase was
used to separate the products, with helium as a carrier gas.
Aliquots (0.2 mL) were injected using a heating program
from 70 to 200 oC at 10 oC min-1. Table 1 shows the different
variables used in this study.
Conversion was calculated using 1,4-dioxane as internal
standard with the use of a calibration standard curve.
Scheme 1. Continuous-flow production of solketal under atmospheric Conversion is defined as the amount of glycerol before
pressure by acid catalyzed ketalization of glycerol. and after the reaction, divided by the amount of glycerol
before reaction.
Experimental The productivity of solketal was calculated taken the
molar ratio of glycerol to acetone, without considering
General procedure the solvent, and calculating the molar concentration of

Table 1. Different parameters used for the continuous flow process of

Two catalysts were used in the experiments: Amberlyst‑15,
solketal production
(particle size < 300 μm, Sigma-Aldrich, USA) and K-10
Montmorillonite (powder, Fluka, USA). Both were used Parameter Used values
at the particle size provided by the suppliers. Prior to the
Flow rate / (mL min ) -1
0.1; 0.2; 0.5 and 1.0
catalytic tests, the Amberlyst-15 was activated at 120 oC for
Temperature / ºC 30; 40 and 50
2 h and K-10 Montmorillonite at 150 oC for 2 h on an oven
Catalyst loading / g 1.0; 3.0 and 7.0
and carefully transferred to the flow system. Amberlyst-15
is reported to have 4.7 mmol of acid sites per g, whereas Type of catalyst Amberlyst-15; K-10

K-10 has an acidity of 0.007 mol g-1, determined by Glycerol/acetone molar ratio 1:2; 1:5; 1:10; 1:15 and 1:20
adsorption/desorption of n-butylamine.22 Solvent DMSO and DMF
1834 Atmospheric Pressure Continuous Production of Solketal from the Acid-Catalyzed Reaction of Glycerol with Acetone J. Braz. Chem. Soc.

glycerol in the mixture by using the density and the molar in the reactor system, showed only 2% of conversion at
weight. Then, the conversion was taken to calculate the the base case conditions.
molar concentration of solketal formed, assuming the
stoichiometry of the reaction. The molar concentration
of solketal was transformed to mass concentration per
hour using the molar weight and the space time of each
experiment.

Results and Discussion

We began our studies evaluating the effect of the flow

rate on the conversion. We have used, as a first tentative
guess, the following reaction conditions: 50 oC, 7.0 g of
Amberlyst-15, glycerol/acetone molar ratio 1:2; DMSO as
solvent. The choice of this reaction condition was based on
the previous experience of our group in the related reaction
(Table 2). It can be seen that no significant change was
Figure 1. Effect of process time on glycerol conversion (T = 50 oC, flow
observed among the flow rates studied, with the glycerol rate of 1.0 mL min-1, glycerol/acetone (1:2), DMSO as solvent); (¿)
conversion being around 70%, with a selectivity of 100% Amberlyst-15; (p) K-10 Montmorillonite.
to solketal. This result shows that equilibrium is being
reached even at the highest flow rate used and the number Evaluation of catalyst loading was also taken into
of acid sites was the same in all experiments. account but, at this time, we have decided to move forward
only with Amberlyst-15 which present better results than
Table 2. Effect of the flow rate on the conversion and selectivity at 50 oC, K-10 Montmorillonite. Figure 2 shows the effect of catalyst
7.0 g of Amberlyst-15, glycerol/acetone molar ratio 1:2; DMSO as solvent
loading (Amberlyst-15) on the conversion. At steady state
conditions (after 30 min), the highest catalyst loading
Flow rate / Conversion / Selectivity /
entry (7.0 g) showed the highest conversion. These results may
(mL min-1) %a %a
1 1.0 70 > 99 be interpreted in terms of the number of acid sites. As the
mass of catalyst in the system increases, the number of
2 0.5 70 > 99
active sites on the catalyst does also go up, favoring the
3 0.2 70 > 99
reaction rate.
4 0.1 70 > 99
Conversion and selectivity measured by GC analysis.
a

After this initial assessment, the next step was

the evaluation of process stability of the packed bed
filled with Amberlyst-15 and K-10 Montmorillonite
as an alternative catalyst. Figure 1 shows that glycerol
conversion was significantly higher on Amberlyst-15 than
on K‑10 Montmorillonite. In addition, this later catalyst
showed a slight decrease in conversion with increasing
process time, indicative of deactivation. Experiments at
batch reaction conditions, using a constant number of acid
sites in the medium, also indicated the higher activity of
Amberlyst-15, but the difference with respect to K-10 was
Figure 2. Effect of catalyst (Amberlyst-15) loading on the glycerol
not so significant.20 Nevertheless, in the present study, conversion (T = 50 oC, flow rate of 1.0 mL min-1, glycerol/acetone (1:2),
the loading of catalyst was kept constant (7.0 g of each DMSO as solvent); (¿) 7g; (¢) 3g; (p) 1g.
catalyst), instead of the amount of acid sites. Thus, since
Amberlyst-15 has more acid sites per g than K-10, the Temperature was also evaluated as an important variable
conversion over this catalyst was significantly higher. It is for process optimization and Figure 3 shows the effect of
important to stress that blank reactions, with no catalyst different temperatures on the conversion of glycerol towards
Vol. 27, No. 10, 2016 Oliveira et al. 1835

soketal over Amberlyst-15. As the experiments were carried The basicity of the two solvents on aqueous sulfuric acid
out at atmospheric pressure, a limit of the temperature, are similar and slightly higher than acetone.27,28 Therefore,
slightly lower of the boiling point of acetone (56 oC), had equilibrium among protonated DMF, DMSO and acetone
to be established. At 50 oC the glycerol conversion was must occur, as well as with glycerol. Another possibility
around 72%, but decreased as the temperature goes down. would be the different solvation or interaction of DMF and
DMSO with the reactants, influencing the reactivity and the
energy barrier. DMF has two heteroatoms that may interact
through hydrogen bonding with two different molecules.
Scheme 2 shows possible pictures of the role of DMF
and DMSO in the nucleophilic attack of glycerol to the
protonated acetone. A DMF molecule can interact with the
protonated acetone and with the secondary hydroxyl group
of glycerol. This makes the primary hydoxyl more reactive
toward nucleophilic attack of the protonated carbonyl
group. On the other hand, with DMSO this situation
does not occur. One may infer that DMSO may help the
stabilization of the protonated acetone through hydrogen
bonding, whereas the glycerol molecule undergoes an
intramolecular hydrogen bonding between the secondary
and primary hydroxyl groups, which reduce the reactivity
Figure 3. Effect of temperature on the glycerol conversion (flow rate of
1.0 mL min-1, glycerol/acetone (1:2), 7.0 g of Amberlyst-15, DMSO as toward nucleophilic attack in the carbonyl group, explaining
solvent); (¿) 50 oC; (¢) 40 oC; (p) 30 oC. the lower conversion compared with DMF.

Solvent choice is a very important task during the

development of a continuous flow protocol, since all
reagents need to be soluble in order to be able for pumping
through the packed bed reactor. Figure 4 shows the effect
of solvent (DMSO or DMF) on the glycerol conversion.
Acetone and glycerol are not completely soluble, requiring
the use of a solvent to avoid phase separation. We have
previous shown21 that DMSO can be used to dissolve
glycerol with aliphatic aldehydes. DMSO and DMF are Scheme 2. Pictorial view of the role of DMF (a) and DMSO (b) in
polar aprotic solvent widely used in chemistry. It can be nucleophilic attack of glycerol to protonated acetone.
seen on Figure 4 that DMF led to slightly higher conversion
(77%) than DMSO did (72%). Ketalization is a reversible reaction and can thus be
shifted towards products by increasing the molar ratio
of reactants or withdrawing water from the medium. We
studied the effect of glycerol to acetone molar ratio on the
conversion of the continuous flow ketalization; the results
are presented on Figure 5. As expected, the increase in the
molar ratio of reactants led to an increase in conversion,
which jumped from 72% at 1:2 glycerol/acetone molar ratio
to 87% at 1:20 glycerol/acetone molar ratio. At higher molar
ratios, the two isomers, five and six-membered ring ketals,
could be observed with the former being in 96% selectivity.
Considering that DMF gave slightly better results
than DMSO, we decided to run an experiment at the best
parameter conditions, which are: 7.0 g of Amberlyst-15,
1.0 mL min-1 of flow rate, at 50 oC, 1:20 glycerol/acetone
Figure 4. Effect of the solvent on the glycerol conversion (T = 50 oC,
7.0 g of Amberlyst-15, glycerol/acetone (1:2), flow rate of 1.0 mL min-1); molar ratio and DMF, instead of DMSO, as solvent. Under
() DMSO; (Ú) DMF. these conditions, the glycerol conversion was 92%, the
1836 Atmospheric Pressure Continuous Production of Solketal from the Acid-Catalyzed Reaction of Glycerol with Acetone J. Braz. Chem. Soc.

of kg of solketal produced per liter and per hour, was also

calculated, taking into account the conversion, space time
and molar ratio of acetone to glycerol, as in some articles
the volume of solvent used was not clear. The use of higher
catalyst loading and molar ratio of reactants, overcome
the need for using high pressures. In one experiment,
Shirani et al.31 reported results at 1 bar, 60 °C and 1.5 g
of catalyst loading, achieving 64% conversion. This value
is even lower than our results at 50 oC and 1:2 glycerol/
acetone molar ratio. Therefore, the continuous liquid phase
production of solketal can be achieved with over 90%
conversion, by proper choice of catalyst loading and molar
ratio of reactants.
Figure 5. Effect of molar ratio of glycerol to acetone on the glycerol
conversion (T = 50 oC, 7.0 g of Amberlyst-15, flow rate of 1.0 mL min‑1; In terms of productivity, the space time has a significant
DMSO as solvent); () 1:2; (×) 1:5; (¿) 1:10; (¢) 1:15; (p) 1:20. effect. Even though we have used a large excess of acetone,
our productivity in kg per liter and hour was higher than
highest observed in this study at atmospheric pressure, and the one calculated for the Shirani’s work. However, if we
about 5 points higher than in DMSO. consider our lowest glycerol to acetone molar ratio (1:2) and
In recent years, some studies of continuous reactive the 77% conversion observed at this condition (taken form
distillation,29 as well as continuous flow production of Figure 4), we may calculate a productivity of 2.337 kg L-1 h-1.
solketal, from the reaction of glycerol and acetone, using Thus, if one wants to upscale the process, it might be better to
homogeneous30 and heterogeneous catalysts31-33 appeared in use a lower acetone to glycerol molar ratio, whistle working
the literature. Most of the heterogeneous catalysts used in at reduced space time and higher catalyst loading.
the flow production of solketal were sulfonic acid resins, but
the studies were carried out at high pressures to increase the Conclusions
solubility of glycerol in acetone. This allows the reaction to
proceed more efficiently on the catalyst surface, but usually Solketal was efficiently and continuously produced
involves higher costs for pressurizing the system. Table 3 through the reaction of glycerol with acetone using
shows a comparison of published works on the continuous heterogeneous catalyst. At the best reaction conditions,
solketal formation over sulfonic acid resins at high 1.0 mL min-1 of flow rate, 7.0 g of Amberlyst-15 as catalyst,
pressures and the present study, at atmospheric pressure. It 50 oC, 1 bar, 1:20 molar ratio of glycerol to acetone and
can be seen that at our best reaction conditions, the glycerol DMF as solvent, the conversion was 92%, with a selectivity
conversion was similar to the reported in the literature with of 96% to the five-membered ring isomer and 4% of the
similar catalysts, but using pressures in the range of 34 to six-membered ring ketal. No other product, but the ketal
120 bar. In addition, the productivity, expressed in terms isomers, was observed at the reaction conditions used in this

Table 3. Comparison of studies of continuous flow glycerol formation

Parameter Shirania Nandab Nandac Present work Present work

Space time / h
d
1 0.25 0.5 0.2 0.2
Temperature / ºC 20 40 25 50 50
Pressure / bar 120 ca. 41 ca. 34 ca. 1 ca. 1
Type of catalyst Purolite® PD206 Amberlyst-36 Amberlyst-36 Amberlyst-15 Amberlyst-15
Catalyst loading / g 0.77 – 2.0 7.0 7.0
Molar ratio 1:5 1:6 1:4 1:20 1:2
Solvent ethanol ethanol methanol DMF DMF
Conversion / % 95 88 94 92 77
Solketal productivitye / (kg L-1 h-1) 0.288 0.928 0.670 0.425 2.337
a
Reference 31; breference 32; creference 33; dspace time is defined as the time required to process a volume of alimentation, corresponding to the reactor
volume; ecalculated taken only the molar ratio of glycerol to acetone, conversion and space time.
Vol. 27, No. 10, 2016 Oliveira et al. 1837

work. This result, using atmospheric pressure, is similar to 13. Possato, L. G.; Diniz, R. N.; Garetto, T.; Pulcinelli, S. H.;
other studies of solketal formation under flow conditions, Santilli, C. V.; Martins, L.; J. Catal. 2013, 300, 102.
but at significantly higher pressures of up to 120 bar, which 14. Soares, R. R.; Simonetti, D. A.; Dumesic, J. A.; Angew. Chem.,
may impact the operational costs. The utilization of higher Int. Ed. 2006, 45, 3982.
catalyst loading and molar ratio of reactants compensate 15. Nogueira, D. O.; Souza, S. P.; Leão, R. A. C.; Miranda, L. S. M.;
the use of atmospheric pressure to achieve high conversion Souza, R. O. M. A.; RSC Adv. 2015, 5, 20945.
levels and selectivity to the desired product. In addition, 16. Klepacova, K.; Mravec, D.; Bajus, M.; Appl. Catal., A 2005,
the lower space time led to an increased productivity of 294, 141.
solketal, compared with other studies. The continuous 17. Gu, Y.; Azzouzi, A.; Pouilloux, Y.; Jerome, F.; Barrault, J.; Green
flow production of solketal may open the possibility of Chem. 2008, 10, 164;
producing this important derivative in high volumes and 18. Silva, C. R. B.; Gonçalves, V. L. C.; Lachter, E. R.; Mota,
low costs. C. J. A.; J. Braz. Chem. Soc. 2009, 20, 201.
19. Deusch, J.; Martin, A.; Lieske, H.; J. Catal. 2007, 245, 428.
Acknowledgments 20. Silva, C. X. A.; Gonçalves, V. L. C.; Mota, C. J. A.; Green Chem.
2009, 11, 38.
Authors acknowledge financial support from CNPq 21. Silva, P. H. R.; Gonçalves, V. L. C.; Mota, C. J. A.; Bioresour.
and FAPERJ. Technol. 2010, 101, 6225.
22. Gonçalves, V. L. C.; Pinto, B. P.; Silva, J. C.; Mota, C. J. A.;
References Catal. Today 2008, 133-135, 673.
23. Mota, C. J. A.; Silva, C. X. A.; Rosenbach Jr., N.; Costa, J.;
1. Meher, L. C.; Sagar, D. V.; Naik, S. N.; Renewable Sustainable Silva, F.; Energy Fuels 2010, 24, 2733.
Energy Rev. 2006, 10, 248. 24. Ozório, L. P.; Pianzolli, R.; Mota, M. B. S.; Mota, C. J. A.;
2. Luque, R.; Lovett, J. C.; Datta, B.; Clancy, J.; Campelo, J. M.; J. Braz. Chem. Soc. 2012, 24, 931.
Romero, A. A.; Energy Environ. Sci. 2010, 3, 1706. 25. Nadim, F.; Zack, P.; Hoag, G. E.; Liu, S. L.; Energy Policy 2001,
3. Mota, C. J. A.; Silva, C. X. A.; Gonçalves, V. L.; Quim. Nova 29, 1.
2009, 32, 639. 26. Souza, R. O. M. A.; Miranda, L. S. M.; Rev. Virtual Quim. 2014,
4. Zhou, C. H.; Beltramini, J. N.; Fan, Y. X.; Lu, G. Q.; Chem. 6, 34.
Soc. Rev. 2008, 27, 527. 27. Bagno, A.; Scorrano, G.; J. Am. Chem. Soc. 1988, 110, 4577.
5. Pagliaro, M.; Ciriminna, R.; Kimura, H.; Rossi, M.; Pina, C. D.; 28. Wada, G.; Bull. Chem. Soc. Jpn. 1969, 42, 890.
Angew. Chem., Int. Ed. 2007, 46, 4434. 29. Clarkson, J. S.; Walker, A. J.; Wood, M. A.; Org. Process Res.
6. Behr, A.; Eilting, J.; Irawadi, K.; Leschinski, J.; Lindner, F.; Dev. 2001, 5, 630.
Green Chem. 2008, 10, 13. 30. Monbaliu, J. C. M.; Winter, M.; Chevalier, B.; Schmidt, F.;
7. Jérôme, F.; Pouilloux, Y.; Barrault, J.; ChemSusChem 2008, 1, Jiang, Y.; Hoogendoorn, R.; Kousemaker, M. A.; Stevens, C. V.;
586. Bioresour. Technol. 2011, 102, 9304.
8. Dasari, M. A.; Kiatsimkul, P. P.; Sutterlin, W. R.; Suppes, G. J.; 31. Shirani, M.; Ghaziaskar, H. S.; Xu, C. C.; Fuel Process. Technol.
Appl. Catal., A 2005, 281, 225. 2014, 124, 206.
9. Kusunoki, Y.; Miyazawa, T.; Kunimori, K.; Tomishige, K.; 32. Nanda, M. R.; Yuan, Z.; Qin, W.; Xu, C. C.; Fuel 2014, 128,
Catal. Commun. 2005, 6, 645. 113;
10. Chai, S. H.; Wang, H. P.; Liang, Y.; Xu, B. Q.; J. Catal. 2007, 33. Nanda, M. R.; Yuan, Z.; Qin, W.; Xu, C. C.; Appl. Energy 2014,
250, 342. 123, 75.
11. Tsukuda, E.; Sato, S.; Takahashi, R.; Sodesawa, T.; Catal.
Commun. 2007, 8, 1349. Submitted: October 8, 2015
12. Pestana, C. F. M.; Guerra, A. C. O.; Ferreira, G. B.; Turci, C. C.; Published online: March 3, 2016
Mota, C. J. A.; J. Braz. Chem. Soc. 2013, 24, 100.
 Fuel 172 (2016) 310–319

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Heterogeneous catalytic conversion of glycerol to oxygenated fuel

additives
Vadim O. Samoilov a, Dzhamalutdin N. Ramazanov a, Andrey I. Nekhaev a, Anton L. Maximov a,⇑,
Leonid N. Bagdasarov b
a
Topchiev Institute of Petrochemical Synthesis, Russian Academy of Sciences, Leninsky pr. 29, Moscow 119991, Russia
b
Gubkin State University of Oil and Gas, Leninsky pr. 65, Moscow 119991, Russia

h i g h l i g h t s g r a p h i c a l a b s t r a c t

Reaction of glycerol with tert-butyl

alcohol and acetone in flow reactor.

Zeolite-BEA converts glycerol
completely into the mixture of
solketal and STBE.

Yield of STBE in 1-mono-GTBE
ketalisation depends inversely from
temperature.

The introduction of solketal into
hydrocarbon oil improves its
antiwear properties.

Solketal is the most effective antiwear
agent from all the ethers formed.

a r t i c l e i n f o a b s t r a c t

Article history: Glycerol alkylation with tert-butyl alcohol and ketalisation of glycerol 1-mono-tert-butyl ether with
Received 23 August 2015 acetone, as well as the combined ketalisation–alkylation process, has been studied in a fixed-bed flow
Received in revised form 26 November 2015 reactor. It has been shown that a continuous, one-step process for the quantitative conversion of glycerol
Accepted 5 January 2016
into a mixture of ethers can be accomplished under mild conditions (atmospheric pressure and temper-
Available online 13 January 2016
atures of 40–70 °C) over a zeolite BEA catalyst. Furthermore, the effects of the glycerol ether additives on
the antiwear properties of low-viscosity base oil have been characterised.
Keywords:
Ó 2016 Elsevier Ltd. All rights reserved.
Sustainable chemistry
Renewable fuel additive
Solketal
Glycerol
Solketal tert-butyl ether

1. Introduction in the performance characteristics of biodiesel and the comprehen-

sive processing of bioglycerol, which is a byproduct of the manu-
The progressive increase in the world production of biodiesel facture of fatty acid methyl esters. Compared with petroleum
confronts, among others, two important issues: the improvement fuel, biodiesel increases the emission of nitrogen oxides during
engine operation and experiences low-temperature pumpability
problems. Many ways to utilise bioglycerol have been proposed;
⇑ Corresponding author. Tel.: +7 495 9554309. they are primarily aimed at the manufacture of petrochemicals
E-mail address: [email protected] (A.L. Maximov).

http://dx.doi.org/10.1016/j.fuel.2016.01.024
0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
 V.O. Samoilov et al. / Fuel 172 (2016) 310–319 311

and the intermediates that are produced from propylene, propy- was reported to be realized in a flow reactor with the use of metha-
lene oxide, or 1,2- and 1,3-propanediols as feedstocks [1,2]. More- nol as a co-solvent [11]. Because these ketals bear a hydroxyl group
over, the problems with sustained bioglycerol utilisation and the in their molecules, it is reasonable to convert it into an alkoxy sub-
quality improvement of biodiesel fuel blends can be resolved stituent. A ketal molecule modified in this manner should exhibit
simultaneously because the bioglycerol processing method herein higher hydrophobicity, oxidation stability, and heat of combustion.
suggests that glycerol can be used to manufacture additives for Monbaliu et al. [28,29] reported the effective production of
motor fuels [3]. solketal tert-butyl ether (STBE), including its synthesis in continu-
Glycerol acetals can be easily synthesized and are used as addi- ous flow systems. STBE was obtained by glycerol ketalisation fol-
tives for alternative and petroleum fuels or their blends [3–11]. In lowed by solketal O-alkylation with isobutylene; sulphuric acid
particular, ethers are known to reduce the emission of hazardous was used as a catalyst. STBE is known as a promising motor fuel
substances during fuel combustion [12]. 2,2-Dimethyl-4-hydroxy additive. The introduction of STBE at a concentration of no more
methyl-1,3-dioxolane (known as solketal, the acetone ketal of glyc- than 5% into diesel leads to a reduction in nitrogen oxides, carbon
erol) is quite a promising additive for fuels and lubricants. It has monoxide, and soot emissions [2]. When added to gasoline, STBE
been recently shown [13] that solketal can be obtained with presumably exerts an octane-boosting effect (solketal has a blend-
100% selectivity and a 95% yield by performing the reaction in sub- ing octane number of 98) [7].
critical acetone on PuroliteÒ PD 206 cation-exchange resin. When The synthesis of solketal ethers by reacting solketal with benzyl
added to motor gasoline (including bioethanol–gasoline fuel), alcohol [30] or olefin oxides [31] has been previously described.
solketal increases the octane number of the fuel and reduces The published method for the two-step synthesis of STBE from
gum formation [7]. In diesel and biodiesel or their blends, acetals glycerol involves the acetone ketalisation of glycerol as the first
reduce the exhaust emissions of nitrogen oxides, carbon monoxide step, followed by alkylation of the product solketal with isobuty-
and soot, and they improve cold flow properties but noticeably lene [28,29].
decrease the oxidation stability [1,4]. In addition, glycerol acetals The aim of this work was to study the synthesis of STBE in one
including solketal exhibit antiwear properties when added to stage in a flow system and to outline the basic parameters of its
motor fuels or lubricating oils [14]. effective production process. The use of isobutylene as an alkylat-
Glycerol ethers of monohydric alcohols also improve the prop- ing agent causes problems associated with impeded mass transfer,
erties of diesel and biodiesel fuels. The production of glycerol tert- the amenability of isobutylene to side reactions in the presence of
butyl ethers (GTBEs) was thoroughly investigated in a number of strong acid catalysts, and the need to use a compressor to pump
studies [15–21] and summarised in a review [22]. Ozbay and the olefin. Therefore, we chose tert-butyl alcohol as the alkylating
coworkers [23] have recently reported the results of the reaction agent.
of glycerol with tert-butanol (TBA) on various industrial solid acid The efficiency of an industrial-scale process always decreases
catalysts in batch and flow reactors. Behr et al. [1] presented a flow with an increasing number of stages. Therefore, we focused on
chart of GTBE production by glycerol alkylation with isobutylene. developing a one-step process for glycerol conversion into substi-
The feasibility of GTBE production by reactive distillation was also tuted derivatives. Because STBE can be produced from either solke-
described [24]. tal (by alkylation with TBA) or glycerol 1-mono-tert-butyl ether
It should be noted that the abovementioned glycerol tert- (mono-1-GTBE) by ketalisation with acetone, the following pro-
butylation experiments were performed in order to evaluate the cesses were investigated in a flow system:
activity of the catalyst in the particular reaction, glycerol conver-
sion and the products distribution. The literature indicates that 1. glycerol alkylation with TBA;
GTBE production industrial process was developed by Behr and 2. acetone ketalisation of 1-mono-GTBE; and
Obendorf [25,26]. The authors used a cascade of stirred batch reac- 3. combined alkylation–ketalisation of glycerol with a mixture of
tors and p-toluenesulfonic acid as the catalyst. Moreover, there is alcohol and ketone.
information that an industrial plant (500.000 ton of GTBEs per
year) was opened in 2010: Behr [25] quoted that there is a 2. Materials and methods
patented industrial process developed by ARCO Chemical Technol-
ogy. Thus, the industrial synthesis of GTBEs seems to have been 2.1. Materials
already developed and realized on the industrial scale.
The possibility of the involvement of GTBE mixtures into motor The catalyst used in the study was commercially available zeo-
fuels remains debatable. The main problem is associated with the lite beta (Zeolyst International, Kansas City, Kansas, USA) CP811Tl
hydrophilicity of the ethers because the glycerol monoethers exhi- in the H+-form (SiO2/Al2O3 = 40; 0.05% Na; specific surface area,
bit good miscibility with water and the diethers are amphiphilic 725 m2/g; acidity, 0.81 mmolH+/g; pore volume, 0.58 cm3/g; pore
liquids. The synthesis of tri-tert-butyl ethers is fraught with great diameter, 0.69 nm). The powdered catalyst was tabletted (com-
difficulties because the formation of a structure of this type is pacting pressure of 3 t/cm2), ground, and sieved to select a fraction
impeded by significant steric hindrance. The ways to apply the of 0.63–1.60 mm.
GTBE mixtures have already been considered in [1], and such mix- The reactants were glycerol, acetone, and tert-butyl alcohol
tures do not have an unambiguously favourable effect on the prop- (Aldrich, USA). All chemicals were used without further
erties of formulated motor fuels. Mono- and diethers of glycerol do purification.
not possess the optimal volatility for their blending with gasoline
[2]; however, when mixed with biodiesel, they significantly 2.2. Equipment
improve its low-temperature and environmental properties [2,4,6].
We believe that it is better to use GTBE as a minor fuel additive Flow reactor experiments were performed on a unit with a
together with the major components, STBE and solketal. In this sce- ‘‘structured-regime reactor” that was specially designed for run-
nario, the aforementioned problems would not be so acute. ning such processes (Fig. 1). The design of the unit included a
In [27], we reported success in acetone ketalisation of glycerol heated plug-flow reactor with a fixed catalyst bed and a heated
in a flow reactor, yielding a mixture of solketal and 2,2-dimethyl- receiver–vaporiser connected to the reactor inlet by a separate
5-hydroxy-1,3-dioxane, the solketal isomer differing in the num- heated line. This design allows unreacted acetone and
ber of atoms in the cyclic moiety. Moreover, glycerol ketalisation tert-butanol to be continuously distilled off and recycled into the
312 V.O. Samoilov et al. / Fuel 172 (2016) 310–319

solved by addition of a homogenizer – ethanol [11] or ethylene gly-

col [9]. Introduction of TBA into the mixture raises the miscibility
of glycerol. Some control experiments were performed; it turned
out that glycerol-TBA mixture (1:3.4 mol.) was completely misci-
ble in acetone at any ratio. All the ternary mixtures were homoge-
neous at temperatures from ambient to 50 °C. Moreover, the
reactants were fed into the reactor as a homogeneous mixture
formed by a mixer which was a part of the flow system.
A schematic of the laboratory flow unit is pictured in Fig. 1. Ace-
tone (pump 1) and the glycerol mixture with TBA (pump 2) are fed
in a given ratio at a required space velocity (0.5–1.5 h–1) to mixer 3
and further to the inlet of catalytic Section 4, which is packed with
b-zeolite pellets. The water coming into the jacket of catalytic Sec-
tion 4 is maintained at a temperature of 30–70 °C.
The catalysis products mixture comes through condenser 5 to
arrive at receiver 6, as well as the unreacted acetone and TBA.
The receiver jacket is fed with water at 90 °C to distill the acetone
and TBA off. Through heated line 7, the unreacted acetone and TBA
enter catalytic Section 4 again to react with glycerol. Condenser 8
serves for venting.
The system has two major features. First of all, the system is not
closed – it is connected with the atmosphere through condenser 8.
Hence, the pressure in the reactor is close to atmospherical pres-
sure. Correspondingly, a gravity flow of the reactants mixture
through the catalyst bed takes place. The second feature is a con-
tinuous recycle of unreacted acetone and TBA allowing to create
a greater excess of acetone and TBA in the reactor and to use the
reactants more fully. Consequently, the effective excess of acetone
and TBA inside of the catalyst bed is greater than in the feed mix-
ture. Nevertheless, the influence of reactants recycle is significant
only at high temperatures (usually above acetone boiling point).
At moderate temperatures with no recycle the products distribu-
tion slightly differs: glycerol conversion is somewhat lower (max-
Fig. 1. Laboratory flow system. imum for 5%); m-GTBEs:di-GTBEs ratio is also somewhat lower.
Yields of solketal and STBE are not affected much. However, the
experiments were mostly performed with recirculation of acetone
Table 1
and TBA. In this regime, recycle acetone and TBA contained some
Feed stream compositions in the combined alkylation–ketalisation experiments.
moisture since they were separated from the reaction products
Glycerol TBA Acetone Alcohol mixture to mixture by a simple distillation. Nevertheless, no conversion fall
acetone flow volume ratio
Molar ratio was observed. This suggests that such a manner of the reactants
5.3 50/50 recycle may be feasible.
1 3.4 10.0 35/65 Apparently this system does not allow carrying out the reaction
18.0 25/75 at temperatures higher than reactants boiling points. The lowest
one is acetone boiling point which amounts to 56 °C. This limita-
tion forced us to find out a very mild regime of glycerol conversion.
reactor. The reactor feed system was comprised of two pumps and The material balance in the experiments was maintained at an
a standard Gilson mixer. The reaction temperature was monitored accuracy within 5 wt%, despite the observed evolution of a small
with a thermocouple mounted in the reactor. The reactor volume amount of isobutylene, which was not taken into account.
was 15 mL, and the catalyst charge was 10 cm3. In control experiments, a glass batch reactor was charged with
Control experiments were performed using a stirred glass reac- 50 mg of the catalyst and a magnetic stirrer. When the tempera-
tor with a heating jacket equipped with a magnetic stirrer and a ture of the jacket heating water reached the set level, a feed mix-
reflux condenser. The reactor internal volume was 25 mL. ture was loaded and the time countdown was started. For control
experiments, reaction mixtures from the flow unit underwent vac-
uum distillation to yield a fraction of monoethers (90 mol.% of m-
2.3. Experimental procedure GTBEs, 10 mol.% of di-GTBEs) and a fraction of diethers (90 mol.% of
di-GTBEs, 10 mol.% of m-GTBEs).
A homogeneous mixture of glycerol with TBA was used in all
flow reactor experiments. Glycerol to TBA molar ratio was 2.4. Analytical methods and calculation
1:3.44; the mixture was prepared by mixing one volume part with
2.75 volume parts of TBA. In the experiments on the combined To determine the product composition, cumulative catalysates
alkylation–ketalisation of glycerol, the pump system of the unit obtained during the 4-h operation of the unit were sampled and
simultaneously supplied acetone and the alcohol mixture in vari- studied by GLC on a Kristalyuks-4000 M chromatograph with a
ous volume ratios. Table 1 shows the correspondence of the ratios flame ionisation detector, using a Supelcowax 10
of pump flow rates to the molar ratios of the reactants. (30 m  0.25 mm) column. The temperature program included an
TBA is a good co-solvent for acetone–glycerol system. Glycerol isothermal hold at 60 °C (5 min), a temperature rise from 60 to
is known to be poorly miscible in acetone. The problem is usually 230 °C at a heating rate of 5 °C/min, and a hold at 230 °C for
 V.O. Samoilov et al. / Fuel 172 (2016) 310–319 313

10 min. For quantitative analysis, toluene was used as the internal 3. Results and discussion
standard. GC calibration was carried out with the use of pure com-
pounds recovered from the reaction products mixtures by means 3.1. Alkylation of glycerol with tert-butyl alcohol
of vacuum rectification (residual pressure – 1 mmHg). Three pure
calibration samples were obtained: solketal (98 mass% by GLC), We chose zeolite beta as the catalyst for the reaction of glycerol
the mixture of m-GTBEs (97 mass% by GLC) and the mixture of with tert-butanol because it had been shown to be the best zeolite
di-GTBEs (99 mass% by GLC). catalyst for this reaction [32]. Of the industrial catalysts,
The structure of the reaction products was determined by gas Amberlyst-15 has been studied to the greatest extent [19,33]. Gon-
chromatography–mass spectrometry on a Finnigan MAT 95 XL zález et al. [32] performed an extensive study of the properties of
instrument (Varian VF-5 ms capillary column, 30 m various types of zeolite catalysts, including zeolite beta. They
length  0.25 mm inner diameter, film thickness 0.25 lm) with explained the high activity of zeolite beta in the glycerol etherifica-
helium as the carrier gas. For the operating mode of the column, tion reaction by citing the high surface concentration of Brønsted
the injector temperature was 270 °C and the samples was isother- acid sites and appropriate acidity of the catalyst. Zeolite ZSM,
mally held for 5 min at an initial column oven temperature of which also has satisfactory acidity, shows lower activity because
30 °C, followed by heating at a rate of 10 °C/min to 300 °C. The of its narrower pores, a factor that is certainly the most important
mass spectrometer operating mode was: ionisation energy of as the GTBE molecules are rather bulky.
70 eV, source temperature of 230 °C, scanning over the range of In our earlier study [27], we also came to the conclusion that
20–800 Da at a scan rate of 1 s per decade and 1000 resolution. zeolite beta is optimal for use as an ketalisation catalyst. The cata-
The components were identified with the use of reference mass lyst chosen should be active to both etherification and ketalisation
spectra available from the NIST/EPA/NIH 11 database. to provide solketal tert-butyl ether formation. Moreover, a load of
Glycerol conversion (XGly), products selectivities (Si) and yields zeolite beta was on stream (glycerol tert-butylation and ketalisa-
(Yi) were calculated in the following manner: tion) for approximately 350 h. We did not observe any significant
C ethers activity fall: after 350 h on stream, the results for glycerol etherifi-
X Gly ¼ ; cation were about the same (±3–5% of glycerol conversion). The
C ethers þ C glycerol unreacted
only evident change was that zeolite colour altered from white
C ethers ¼ C STBE þ C Solketal þ C GTBEs ;
to light beige. Presumably some slight acetone self-condensation
where C STBE , C Solketal , C GTBEs , and C glycerol unreacted are molar concentra- took place. It was previously reported [34] that during glycerol
tions of STBE, solketal, GTBEs and free glycerol in a reaction prod- ketalisation zeolite BEA water tolerance was quiet high in compar-
ucts mixture correspondingly. Molar concentrations were ison with ion-exchangers (particularly sulfonated styrene-
calculated from GC analysis results. Glycerol and its compounds divynilbenzene catalysts). Enhanced zeolites water tolerance is
were considered to be absolutely non-volatile. presumably due to the surface hydrophobicity preventing irre-
C STBE versible adsorption of water [35]. Also, it was quoted that in a flow
SSTBE ¼ ; system glycerol tert-butylation proceeded more effectively than in
C ethers
a batch reactor possibly due to possible water ‘‘wash-away”
Y STBE ¼ X Gly  SSTBE :
inhibiting catalyst deactivation [23]. Finally, sulfonated ion-
Other selectivities and yields were determined similarly. exchangers were found to be easily reactivated [23] with no con-
siderable activity loss [23]. Since zeolite BEA was reported to be
2.5. Antiwear testing of glycerol derivatives more stable than ion-exchangers, the same reactivation efficiency
may be expected.
To evaluate their antiwear properties, the glycerol ethers were However, there are data [20] that indicate that zeolite beta dis-
mixed with heavy cycle oil manufactured at the Novoyaroslavsky plays a very low activity in the glycerol etherification reaction with
refinery (Slafneft’, Russia). The cycle oil was cleaned of dissolved monohydric alcohols.
hydrogen sulphide by alkalisation and dried by using 5 wt% calci-
nated molecular sieves (5 Å). The drying was performed for 24 h.
The dried cycle oil was mixed with the glycerol ethers in different 70
proportions, and the antiwear properties of the mixtures were then
determined according to ASTM D 2266-01 (Standard Test Method 60
for Wear Preventive Characteristics of Lubricating Fluid) at a load
of 196 N over one hour. 50
To blend the cycle oil, the following three additives were used:
40
1. Solketal with a purity of at least 97 wt% (by GLC data) isolated %
from the flow-unit catalysates via multistage vacuum 30
distillation.
2. A mixture of solketal and STBE with a purity of at least 97% (by 20
GLC data) and a components mass ratio of 70/30, as isolated
from the flow-unit catalysates via multistage vacuum 10
distillation.
3. A mixture of 1,2- and 1,3-di-GTBE with a purity of at least 97% 0
(by GLC data) isolated via multistage extraction from the cata- 45 50 55 60 65 70 75 80 85
lysate obtained by glycerol alkylation with tert-butanol in a stir- T, °C
red reactor. The extraction was performed in a n-hexane–water
system. - 0.5 h-1, -1.0 h-1, - 1.5 h-1.

Fig. 2. Temperature dependence of the total yield of the tert-butyl glycerols in the
The additives were mixed with the cycle oil immediately before reaction of glycerol with TBA at different feed space velocities. Glycerol:TBA molar
testing. ratio = 1:3.4.
314 V.O. Samoilov et al. / Fuel 172 (2016) 310–319

Fig. 2 shows the temperature dependence for the yield of tert- highly active catalyst for the tert-butylation of glycerol, exhibiting
butyl ethers of glycerol in the glycerol reaction with TBA. Among a higher selectivity for diethers than zeolite Y. In the catalysis by
the products, glycerol monoethers and diethers were detected, beta zeolite CP814E (T = 90 °C, isobutylene to glycerol molar ratio,
with the monoether to diether molar ratio varying from 4.0 to 4:1; batch reactor, 7.5 wt% catalyst), up to 21 mol.% of isobutylene,
6.0. No glycerol tri-tert-butyl ether was found in a detectable depending on the reaction time, is consumed for the formation of
amount in the product. diisobutylenes. Moreover, the maximal conversion of glycerol cor-
An increase in the space velocity at temperatures up to 70 °C responds to a dimerisation yield of 0.8 mol.% based on butylene,
leads to a decrease in the yield of the glycerol ethers due to the and the operation in the mode of maximal diether yield gives the
decrease in the space time. dimers with a yield of 5.4 mol.%.
The same behaviour was observed when the reaction was per- The catalysates obtained in this work did not contain apprecia-
formed under steady-state conditions [19]. At a temperature of ble amounts of isobutylene oligomers (nor was the intense evolu-
70 °C, the total yield of the glycerol ethers reached a maximum tion of isobutylene observed), suggesting that reaction selectivity
value at a space velocity of 0.5 h–1. The most dramatic decrease is higher when tert-butanol is used as an alkylating agent.
in the yield of GTBEs at 75 °C or higher took place at a low space Alternatively, tert-butanol is hydrated isobutylene, which therefore
velocity. This nonmonotonic pattern of dependence can be due to inherently affects the thermodynamic equilibrium in the reaction
that the reaction temperature approaches the boiling point of system because water reduces the equilibrium constant of the
TBA (82 °C) or its azeotropic mixture with water (80 °C); as a reaction and it is present in the feedstock, although in bound form.
result, the equilibrium concentration of TBA in the liquid phase It is the presence of water that is responsible for the inhibition of
and the heterogeneous-catalyst adsorption layer drastically the oligomerisation reaction in our system. We assumed that to
decreases. achieve optimal performance characteristics of the process, it
Because the formation of diethers is a two-step reaction, the was necessary to ‘‘adjust” the water content in the reactants. In
product composition is greatly affected by the contact time of [38], tert-butyl alcohol was shown to be produced in small
the reactants with the feedstock. For example, the proportion of amounts simultaneously with the dimers. The use of a mixed alky-
glycerol diethers increases with a decrease in the space velocity. lating agent (isobutylene–tert-butanol) will resolve this dilemma;
The highest glycerol conversion (70%) and the highest yields of however, the selectivity for oligomers will be lower than with iso-
GTBEs (mono-GTBE, 60 mol.%; di-GTBE, 9 mol.%) are attained at a butylene alone, but the conversion will be higher than with TBA.
temperature of 70 °C and a space velocity of 0.5 h–1.
Notably, the results obtained in this study are consistent with 3.2. Acetone ketalisation with 1-mono-GTBE
the data reported by Viswanadham and Saxena [36] for a similar
system. The yields of GTBEs produced in the zeolite BEA- The production of STBE through solketal alkylation with TBA
catalysed reaction at 0.5 MPa and 110 °C with a TBA molar excess was described by Monbaliu et al. [28,29], and the development of
of 4 are somewhat higher than those in the present study. Thus, an the two-stage processes involving ketalisation with glycerol as
increase in pressure is leverage for the process, preventing TBA the first step was the outcome of these thorough investigations.
from evaporating and inhibiting its intramolecular dehydration Note that the first step of the two-step STBE synthesis [29] is glyc-
reaction to form isobutene. Furthermore, the space velocities used erol ketalisation, and the second step is solketal etherification with
in [36] are below ours by a factor of 2 or greater, thereby allowing isobutylene. Sulphuric acid was used as a homogeneous catalyst.
for a more complete conversion of glycerol. At the same time, at 50 °C glycerol reacts with both with ace-
For the selectivity of the process, the choice of alkylating agent tone and TBA to form solketal and GTBEs correspondingly. This
is of particular interest. Noureddini and co-workers [37] reported suggests that it is possible to obtain STBE in one step by a reaction
that during the reaction of glycerol with isobutylene under similar of glycerol with a mixture of acetone and TBA.
conditions (isobutylene to glycerol molar ratio 3:1, batch reactor, In this system two different two-step reactions of STBE forma-
5 wt% Amberlyst-15, reaction times of 2–4 h) in the temperature tion can take place (Fig. 3). The first route (route A) is consisted
range of 80–93 °C, the mass fraction of isobutylene dimers in the of glycerol ketalisation (reaction A1) and subsequent solketal
catalysate was on the order of 10–18%. However, this ratio can tert-butylation with TBA (reaction A2). The second route (route
be substantially altered by varying the operating parameters of B) involves formation of 1-mono-GTBE (reaction B1) followed by
the process. ketalisation of the ether to form STBE (reaction B2). Presumably
Klepáčová et al. [38] performed an exhaustive study of the glyc- STBE can be formed via both pathways, but it is possible to mark
erol reaction with isobutylene and showed that zeolite beta is a out the prevailing way?

Fig. 3. Possible pathways of STBE formation.

 V.O. Samoilov et al. / Fuel 172 (2016) 310–319 315

90 During the ketalisation of the monoether STBE and water are

formed and the denominators of equilibrium constants become
80 greater. As the reaction runs, the numerator of the KB2 (Eq. (2))
becomes substantially less than the denominator: there is neither
70 free TBA nor solketal in the mixture, whereas water is formed in a
considerable amount as a result of ketalisation. The foregoing
60
poses a new question: why is it stated that solketal is formed from
STBE, whereas it can be also obtained by hydrolysis of 1-m-GTBE to
50
% mol.

glycerol followed by ketalisation of the latter? The ether ketalisa-

40 tion proceeds rapidly, whereas in the reaction of 1-m-GTBE hydrol-
ysis (with water molar excess) after 30 min glycerol yield was
30 about 10 mol.% (50 °C, water:ether = 4:1). Apparently there is no
moment during 1-m-GTBE – acetone interaction to observe excess
20 of water to the ethers. This suggests that 1-m-GTBE is first rapidly
ketalised and then slowly dealkylated to form solketal. The latter
10 statement is in concordance with the data displayed in Fig. 4: STBE
concentration slowly decreases in favour to solketal, which is more
0
0 30 60 90 120 150 180 thermodynamically stable. It should be noted that partial hydroly-
sis (or deprotection) of solketal benzyl ether yielding glycerol ben-
Reaction time, min
zyl ether was previously reported [30].
Fig. 4. 1-m-GTBE batch-reactor ketalisation with acetone: components molar Moreover, it turned out that di-GTBEs react in the same manner
concentrations vs reaction time. T = 50 °C, 1-m-GTBE:acetone = 1:10 mol. (r) – 1- to form STBE, monoethers and solketal (Fig. 6). We consider the
m-GTBE, (j) – di-GTBEs, (d) – STBE, (s) – solketal. elucidation to be the same as for 1-m-GTBE: the equilibrium of
GTBEs hydrolysis is shifted to the right due to low concentrations
of products (glycerol and TBA) and high concentrations of reactants
100 (starting di-GTBEs and water). It should also be noted that di-
GTBEs conversion rate is determined by the rate of de-O-tert-
butylation.
80 Unfortunately, from the data analysis it cannot be explicitly sta-
ted which pathway (A or B) is favoured. To be answered, this ques-
tion needs a more detailed kinetic study to determine the relative
60 rates of every elementary reaction step.
% mol.

The reaction temperature has a substantial effect on the selec-

tivity of the process for different products: lower temperatures
40
facilitate ketalisation [40], whereas a greater amount of GTBEs is
produced at elevated temperatures. Based on the results of glycerol
flow reactor alkylation with tert-butanol (45 °C and space velocity
20
of 0.5 h–1, total yield of the monoethers was approximately 15 wt%
based on glycerol), we assumed that the synthesis of STBE could be
0
run at a reduced temperature. Therefore, both reaction routes
0 30 60 90 120 (solketal alkylation by TBA and acetone ketalisation of 1-mono-
Reaction time, min GTBE) can contribute to the formation of STBE.
Experiments on the flow-reactor ketalisation of 1-m-GTBE
Fig. 5. 1-m-GTBE batch-reactor hydrolysis: components molar concentrations vs (Table 2) have shown that in the region of moderately low
reaction time. T = 50 °C, 1-m-GTBE:water = 1:4 mol. (r) – 1-m-GTBE, (j) – di- temperatures (30–45 °C), the conversion of 1-mono-GTBE results
GTBEs, (N) – glycerol.
100
The results of 1-m-GTBE batch-reactor ketalisation experiments
are displayed in Fig. 4. The ether was rapidly ketalised to form
STBE: after 30 min its conversion amounted to about 89%. More- 80
over, considerable amount of solketal was detected in reaction
mixtures. At first glance the latter fact looks strange regarding
60
the absence of glycerol in the reaction mixture. Nevertheless, the
% mol.

following simple elucidation can be applied here: solketal is

formed as a result of partial hydrolysis of STBE. Glycerol ketalisa-
40
tion is a reverse reaction [39]. Theoretically, interaction between
glycerol and TBA to yield GTBEs is also a reverse reaction; to con-
firm this statement a control experiment (1-m-GTBE hydrolysis) 20
was performed (Fig. 5). The equilibrium constants for reactions
A2 and B2 are as follows:
0
0 30 60 90 120 150
½Solketal½TBA
K A2 ¼ ; ð1Þ Reaction time, min
½STBE½H2 O
½1  m  GTBE½acetone Fig. 6. di-GTBEs batch-reactor ketalisation with acetone: components molar
K B2 ¼ : ð2Þ concentrations vs reaction time. T = 50 °C, di-GTBEs:acetone = 1:14 mol. (j) – di-
½STBE½H2 O
GTBEs, (d) – STBE, (s) – solketal, (N) – glycerol.
316 V.O. Samoilov et al. / Fuel 172 (2016) 310–319

Table 2 The conversion of glycerol decreases with an increase in tem-

1-m-GTBE ketalisation in the flow reactor: reaction mixtures composition.
perature, but a large excess of acetone makes compensation for
1-m-GTBE:acetone (vol.) T, °C vb, h1 Molar concentrationa, % the decrease possible. The dramatic change in the dependence of
1-m-GTBE STBE Solketal the product yields in the vicinity of acetone and tert-butanol boil-
30 0.5 42 54 4
ing points is apparently due to the phase equilibrium of reactants
30 1.0 41 53 6 vaporising from the reaction space. For example, at 40 °C,
30:70
30 1.5 41 54 5 VHSV = 0.5 h1 and reactants ratio 1:3.4:5.3 glycerol was fully con-
45 0.5 33 46 21 verted to yield approximately 80% of solketal and 20% of STBE.
30 0.5 41 46 13 Temperature elevation to 50 °C led to the conversion decrease
35:65 45 0.5 39 44 17 (84%) and formation of GTBEs (about 15 mol.%) was observed. At
a
Water and TBA are ignored. the same conditions and 75 °C glycerol conversion fell down to
b
Volume hourly space velocity. 35% within yields of solketal, STBE and GTBEs of 7.0, 4.7 and
22.8 mol.% correspondingly.
As seen from Fig. 7a–d, the temperature dependence of the pro-
in a mixture of solketal and STBE. Note that at higher temperatures duct yields is straightforward. The yield of solketal decreases with
a greater yield of solketal was observed due to a faster moving an increase in the temperature, a change that is consistent with the
towards the thermodynamical equilibrium. Therefore, we suppose earlier data on the decline in the rate of ketalisation with temper-
that the formation of solketal can be inhibited by introducing tert- ature [40] due to the decrease in acetone concentration in the reac-
butyl alcohol into the system. tion zone and to the shift of the thermodynamical equilibrium. The
yield of tert-butyl ethers increases with temperature, a fact that
3.3. Combined alkylation–ketalisation of glycerol by tert-butanol and also confirms the reliability of the data obtained in the previous
acetone studies [17,19,23,41]. At different reactant ratios, the temperature
dependence curves for the STBE yield exhibit a maximum each
3.3.1. Influence of temperature time that presumably indicates the existence of a regime in which
The temperature dependence of the glycerol conversion and the both of the competing reactions (ketalisation and etherification)
product yields during the combined etherification–ketalisation occur at a maximal overall rate. Thus, it may be assumed that at
process are presented in Fig. 7a–d. As follows from Fig. 7a–d, an a temperature corresponding to the extremum in the curve, STBE
increase in the temperature in the reactor can have an effect on is produced with the highest efficiency because the system has
the process, causing the following changes in the parameters: the maximally effective combination of unequal glycerol ketalisa-
tion and etherification rates.

Decrease in the degree of ketalisation.

Increase in the formation rate of the tert-butyl glycerols. 3.3.2. Influence of reactant ratio

Increase in the proportion of vapourised acetone (when The data presented in Fig. 7a–d indicate that the initial alcohols/
approaching 56 °C) and tert-butanol (when approaching acetone ratio has a substantial effect on the composition of the
82 °C), resulting in a reduction in glycerol conversion. resultant mixture. As the acetone excess decreases, the product

Increase in the mass transfer rate. distribution varies toward a higher STBE/ketals ratio. This change

100 100
a b
80 80

60 60
% mol.

% mol.

40 40

20 20

0 0

30 40
c d
30
20
% mol.

% mol.

20

10
10

0 0
40 45 50 55 60 65 70 40 45 50 55 60 65 70
T, °C T, °C

Fig. 7. Temperature dependence data for product yields during the combined etherification–ketalisation process. a and c – glycerol:TBA:acetone = 1:3.4:5.3 mol.; b and d –
glycerol:TBA:acetone = 1:3.4:10.0 mol.; (r) – solketal yield; (N) – GTBEs yield; STBE yield (d); blue – v = 0.5 h–1, red – v = 1.0 h–1, black – v = 1.5 h–1. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
 V.O. Samoilov et al. / Fuel 172 (2016) 310–319 317

Table 3
Various regimes of the one-step conversion of glycerol in the reactions with acetone and TBA.

Run no. Temperature, v, h–1 Glycerol: Glycerol STBE Total GTBE

T, °C TBA: conversion, % yield, yield, mol.%
acetone mol.%
molar ratio
1 40 37
0.5
2 50 1.0:3.4:18.0 36
0
3 50 1.0 100 28
4 45 30
0.5 1.0:3.4:10.0
5 55 36 5
6 45 0.5 29 0
1.0:3.4:5.3
7 60 1.0 84 28 6

seems reasonable because the ketal formation rate is proportional the absence of free glycerol on the reaction products mixtures
to the acetone concentration in the reaction volume. However, a observed by GLC.
side effect of the decrease in acetone excess is a reduction in the Reaching full glycerol conversion is of great importance with
glycerol conversion in the ketalisation reaction, leading to the pres- respect to the products separation. First, unreacted acetone (b.p.
ence of unconverted glycerol in the catalysates obtained at identi- 56 °C), TBA (b.p. 78 °C) and water (b.p. 100 °C) formed should be
cal temperatures from different feed mixtures. For example, the distilled off. If the residue contains free glycerol, the second step
glycerol conversion was 86% within the yield of GTBEs of about should be carried out involving vacuum distillation of the ethers
15 mol.% at 50 °C, VHSV = 0.5 h1 and a glycerol:TBA:acetone from glycerol followed by significant energy consumption; if not,
molar ratio of 1:3.4:5.3, whereas unreacted glycerol is absent from the separation comes to nothing more than atmospheric-pressure
the catalysates with a higher acetone excess: at the same condi- distillation of volatile compounds to be recirculated. Therefore,
tions and reactants molar ratio 1:3.4:10.0 glycerol was converted glycerol should be converted fully to provide the easiness and
quantitatively; GTBEs yield was only about 2.5 mol.%. Process ‘‘greenness” of the whole process.
selectivity shifted towards ketalisation: in the first regime yields
of solketal, STBE and GTBEs amounted 57.7, 11.0 and 15.3 mol.%
correspondingly in comparison with 84.2, 14.4 and 1.5 mol.% at a 3.4. Study of the antiwear properties of glycerol ethers
glycerol:TBA:acetone ratio of 1:3.4:10.0.
The glycerol ketals/STBE ratio in this case is determined by the Within the scope of this study, we examined the influence of
interplay of two opposite factors, the acetone excess and the tem- glycerol ethers as additives on the antiwear properties of a low-
perature. Elevation of the temperature leads to a decrease in selec- viscosity hydrocarbon oil fraction. An atmospheric distillation resi-
tivity for the glycerol ketals, even at a large excess of acetone. due of cycle oil was used as the base oil to be doped with the addi-
tives; the physicochemical characteristics of the cycle oil are given
in Table A.1.
3.3.3. Influence of space time
The choice of the deeply desulfurised fraction free of additives
The data in Fig. 7a and b shows that the yield of solketal
and impurities was determined by the following circumstances:
increases with an increase in space velocity, and this trend is
it was necessary to rule out, first, the presence of organic sulphur
observed over the entire temperature range of 40–70 °C. Most
compounds that exhibit wear-preventive properties and, second,
likely, the ketalisation reaction is faster than the mono-GTBE for-
the possibility of the chemical transformation of the oxygenates
mation reaction and, consequently, the selectivity for solketal
that would be introduced into the fuel. In addition, a low-
increases with an increase in the space velocity.
viscosity fraction allows the friction mode to be brought maximally
Notably, under conditions that ensure a significant yield of
close to the boundary friction. Table 4 presents the results of the
solketal and its tert-butyl ether, the yield of mono-GTBE is below
antiwear testing of the heavy cycle oil with the glycerol ether
that of di-GTBE, a difference that is due to the consumption of
additives.
the monoether in both the diether and the STBE formation reac-
The data in Table 4 show that solketal has the most significant
tions. From the curves for the dependence of the STBE yield on
effect on the antiwear properties; its tert-butyl ether and tert-butyl
the feed space velocity (Fig. 7c and d), it is clear that the yield of
the desired product at 40–50 °C is proportional to the space time
required for the two-step STBE formation reaction. Moreover, this
dependence can be reversed or even become nonmonotonic as the Table 4
temperature increases. Effect of glycerol ether additives on the antiwear properties of heavy cycle oil (ASTM
D 2266-01 test method).
Thus, the feed space velocity is the third leverage (together with
the temperature and reactant ratio) for the process performance Run no. Additive Average DWSDb,
parameters. The existence of three degrees of freedom makes the WSDa, mm %
Type Amount, ppm
system quite flexible; a high selectivity and a high yield of the 1 Additive-free cycle oil 0.94 –
desired product can be attained at different combinations of tem- 2 460 0.71 25
Solketal
perature regime, feedstock composition, and reactor productivity. 3 980 0.61 35
Various appropriate combinations of these parameters are given 4 22470 0.54 43
5 Mixture of di- 5250 0.76 19
in Table 3. Please note that raw glycerol was converted into the
6 GTBEs 1200 0.84 11
mixture of STBE and solketal with the predominance of the latter. 7 440 0.88 6
In some cases GTBEs (no more than 5–6 mol.% counting on the raw 8 490 0.87 7
STBE/solketal,
glycerol) and unconverted glycerol were found among the prod- 9 1242 0.82 13
70/30
ucts; no other glycerol derivatives were detected. The difference 10 5039 0.61 35

between glycerol conversion and the sum of STBE and GTBE yields a
Wear spot diameter.
b
equals to solketal yield. Glycerol conversion of 100% corresponds to Relative change to additive-free cycle oil WSD.
318 V.O. Samoilov et al. / Fuel 172 (2016) 310–319

glycerols also improve the wear parameter, but to a much lesser [10] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier MA, Xu CC. Thermodynamic
and kinetic studies of a catalytic process to convert glycerol into solketal as an
extent. Apparently, the effectiveness of solketal stems from its
oxygenated fuel additive. Fuel 2014;117:470–7. http://dx.doi.org/10.1016/
increased surface activity that is associated with the presence of j.fuel.2013.09.066.
the free hydroxyl group. The replacement of the solketal hydroxyl [11] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier MA, Xu C. Catalytic
group by TBA gives the bulky hydrocarbon radical that impedes the conversion of glycerol to oxygenated fuel additive in a continuous flow
reactor: process optimization. Fuel 2014;128:113–9. http://dx.doi.org/
adsorption of the molecule on metal surfaces by steric hindrance as 10.1016/j.fuel.2014.02.068.
well as altered polarity and surface activity of the molecule. In [12] Arteconi A, Mazzarini A, Di Nicola G. Emissions from ethers and organic
addition, the spatial configuration of the ketal molecule, which carbonate fuel additives: a review. Water Air Soil Pollut 2011;221:405–23.
http://dx.doi.org/10.1007/s11270-011-0804-y.
contains a cyclic moiety, may have a favourable effect on the [13] Shirani M, Ghaziaskar HS, Xu C. Optimization of glycerol ketalization to
adsorption. produce solketal as biodiesel additive in a continuous reactor with subcritical
However, it should be noted that solketal has low hydrolytic acetone using Purolite PD206 as catalyst. Fuel Process Technol
2014;124:206–11.
stability [39]. The antiwear properties of the cycle oil containing [14] McDougall PJ. Lubricating composition containing 1,3-dioxolane-4-methanol
approximately 0.1% solketal are 35.1% better than those of the compounds. 2012/0129744 A1; 2012.
additive-free oil. A tightly closed flask with cycle oil was left to [15] Chang JS, Chen DH. Optimization on the etherification of glycerol with tert-
butyl alcohol. J Taiwan Inst Chem Eng 2011;42:760–7.
stand for a week at room temperature. After this delay, the cycle [16] Di Serio M, Casale L, Tesser R, Santacesaria E. New process for the production of
oil was tested on a four-ball machine again, and it showed 24.5% glycerol tert-butyl ethers. Energy Fuels 2010;24:4668–72. http://dx.doi.org/
worse results compared with the initial material. 10.1021/ef901230r.
[17] Frusteri F, Arena F, Bonura G, Cannilla C, Spadaro L, Di Blasi O. Catalytic
etherification of glycerol by tert-butyl alcohol to produce oxygenated
4. Conclusions additives for diesel fuel. Appl Catal A Gen 2009;367:77–83. http://dx.doi.org/
10.1016/j.apcata.2009.07.037.
[18] Lee HJ, Seung D, Jung KS, Kim H, Filimonov IN. Etherification of glycerol by
For interaction of glycerol with acetone and TBA in the presence isobutylene: tuning the product composition. Appl Catal A Gen
of a zeolite catalyst, the dependency of the feedstock conversion, as 2010;390:235–44.
well as the product yields, for the reaction temperature, the feed [19] Pico MP, Romero A, Rodríguez S, Santos A. Etherification of glycerol by tert-
butyl alcohol: kinetic model. Ind Eng Chem Res 2012;51:9500–9. http://dx.doi.
space velocity, and the reactants ratio have been described. The org/10.1021/ie300481d.
possibility of STBE formation during the ketalisation of both [20] Roze M, Kampars V, Teivena K, Kampare R, Liepins E. Catalytic etherification of
mono-GTBEs and di-GTBEs was shown. The optimal process glycerol with alcohols. Mater Sci Appl Chem 2013;28:67. http://dx.doi.org/
10.7250/msac.2013.011.
parameters where STBE can be produced with maximally high [21] Xiao L, Mao J, Zhou J, Guo X, Zhang S. Enhanced performance of HY zeolites by
selectivity and molar yield of 30% under the most mild conditions acid wash for glycerol etherification with isobutene. Appl Catal A Gen
possible (T = 45 °C, v = 0.5 h–1, glycerol:TBA:acetone molar 2011;393:88–95.
[22] Izquierdo JF, Montiel M, Palés I, Outón PR, Galán M, Jutglar L, et al. Fuel
ratio = 1:3.4:10.0) were determined. The introduction of the glyc-
additives from glycerol etherification with light olefins: state of the art. Renew
erol derivatives in a concentration of 0.046–2.25 wt% into hydro- Sust Energy Rev 2012;16:6717–24.
carbon oil can improve the antiwear properties by 42%. [23] Ozbay N, Oktar N, Dogu G, Dogu T. Activity comparison of different solid acid
catalysts in etherification of glycerol with tert-butyl alcohol in flow and batch
reactors. Top Catal 2013;56:1790–803. http://dx.doi.org/10.1007/s11244-013-
Acknowledgment 0116-0.
[24] Kiatkittipong W, Intaracharoen P, Laosiripojana N, Chaisuk C, Praserthdam P,
The authors are grateful to R.S. Borisov for assistance in con- Assabumrungrat S. Glycerol ethers synthesis from glycerol etherification with
tert-butyl alcohol in reactive distillation. Comput Chem Eng
ducting the GC–MS analysis. 2011;35:2034–43.
[25] Behr A, Obendorf L. Development of a process for the acid-catalyzed
etherification of glycerine and isobutene forming glycerine tertiary butyl
Appendix A. Supplementary material ethers. Engennier Life Sci 2003;2:185–9.
[26] Behr A, Obendorf L. Process development for acid-catalyzed etherification of
Supplementary data associated with this article can be found, in glycerol with isobutene to yield glycerol tert-butyl ethers. Chem Ing Tech
2001;73:1463–7.
the online version, at http://dx.doi.org/10.1016/j.fuel.2016.01.024. [27] Maksimov AL, Nekhaev AI, Ramazanov DN, Arinicheva YA, Dzyubenko AA,
Khadzhiev SN. Preparation of high-octane oxygenate fuel components from
References plant-derived polyols. Pet Chem 2011;51:61–9.
[28] Monbaliu JC, Winter M, Chevalier B, Schmidt F, Jiang Y, Hoogendoorn R, et al.
Feasibility study for industrial production of fuel additives from glycerol. Chim
[1] Behr A, Eilting J, Irawadi K, Leschinski J, Lindner F. Improved utilisation of
Oggi 2010;28:8–11.
renewable resources: new important derivatives of glycerol. Green Chem
[29] Monbaliu JC, Winter M, Chevalier B, Schmidt F, Jiang Y, Hoogendoorn R, et al.
2008;10:13. http://dx.doi.org/10.1039/b710561d.
Effective production of the biodiesel additive STBE by a continuous flow
[2] Tanugula SK. Synthesis of glycerol-based fuel additives to reduce NOx
process. Bioresour Technol 2011;102:9304–7. http://dx.doi.org/10.1016/j.
emissions from diesel engines operated on diesel and biodiesel fuels by
biortech.2011.07.007.
SNCR. Technischen Universitaet Carolo-Wilhelmina zu Braunschweig; 2010.
[30] Suriyaprapadilok N, Kitiyanan B. Synthesis of solketal from glycerol and its
[3] Rahmat N, Abdullah AZ, Mohamed AR. Recent progress on innovative and
reaction with benzyl alcohol. Energy Proc 2011;9:63–9. http://dx.doi.org/
potential technologies for glycerol transformation into fuel additives: a critical
10.1016/j.egypro.2011.09.008.
review. Renew Sust Energy Rev 2010;14:987–1000.
[31] Selifonov S. Glyceryl ethers compouns and their use. 8318814 B2; 2012.
[4] De Torres M, Jiménez-Osés G, Mayoral JA, Pires E, De Los Santos M. Glycerol
[32] González MD, Cesteros Y, Salagre P. Establishing the role of Brønsted acidity
ketals: synthesis and profits in biodiesel blends. Fuel 2012;94:614–6.
and porosity for the catalytic etherification of glycerol with tert-butanol by
[5] García E, Laca M, Pérez E, Garrido A, Peinado J. New class of acetal derived from
modifying zeolites. Appl Catal A Gen 2013;450:178–88.
glycerin as a biodiesel fuel component. Energy Fuels 2008;22:4274–80.
[33] Vlad E, Bildea CS, Bozga G. Design and control of glycerol-tert-butyl alcohol
[6] Melero JA, Vicente G, Morales G, Paniagua M, Bustamante J. Oxygenated
etherification process. Sci World J 2012;2012:180617.
compounds derived from glycerol for biodiesel formulation: influence on EN
[34] Da Silva CXA, Mota CJA. The influence of impurities on the acid-catalyzed
14214 quality parameters. Fuel 2010;89:2011–8. http://dx.doi.org/10.1016/
reaction of glycerol with acetone. Biomass Bioenergy 2011;35:3547–51.
j.fuel.2010.03.042.
http://dx.doi.org/10.1016/j.biombioe.2011.05.004.
[7] Mota CJA, Da Silva CXA, Rosenbach N, Costa J, Da Silva F. Glycerin derivatives as
[35] Da Silva CXA, Gonçalves VLC, Mota CJA. Water-tolerant zeolite catalyst for the
fuel additives: the addition of glycerol/acetone ketal (solketal) in gasolines.
acetalisation of glycerol. Green Chem 2009;11:38. http://dx.doi.org/10.1039/
Energy Fuels 2010;24:2733–6. http://dx.doi.org/10.1021/ef9015735.
b813564a.
[8] Maximov AL, Nekhaev AI, Ramazanov DN. Ethers and acetals, promising
[36] Viswanadham N, Saxena SK. Etherification of glycerol for improved production
petrochemicals from renewable sources. Petrol Chem 2015;55:1–21. http://dx.
of oxygenates. Fuel 2013;103:980–6.
doi.org/10.1134/S0965544115010107.
[37] Noureddini H, Dailey WR, Hunt BA. Production of ethers of glycerol from crude
[9] Ramazanov DN, Dzhumbe A, Nekhaev AI, Samoilov VO, Maximov AL, Egorova
glycerol – the by-product of biodlesel production the by-product of biodiesel
EV. Reaction between glycerol and acetone in the presence of ethylene glycol.
production. Pap Biomater 1998;18.
Petrol Chem 2015;55:140–5. http://dx.doi.org/10.1134/S0965544115020152.
 V.O. Samoilov et al. / Fuel 172 (2016) 310–319 319

[38] Klepáčová K, Mravec D, Kaszonyi A, Bajus M. Etherification of [40] Lorette NB, Howard WL, Brown JH. Preparations of ketone acetals from linear
glycerol and ethylene glycol by isobutylene. Appl Catal A Gen 2007;328: ketones and alcohols 1959; 24: 1731–3. doi:http://dx.doi.org/10.1021/
1–13. jo01093a028.
[39] Ozorio LP, Pianzolli R, Mota MBS, Mota CJA. Reactivity of glycerol/acetone ketal [41] Melero JA, Vicente G, Morales G, Paniagua M, Moreno JM, Roldán R, et al. Acid-
(solketal) and glycerol/formaldehyde acetals toward acid-catalyzed catalyzed etherification of bio-glycerol and isobutylene over sulfonic
hydrolysis. J Braz Chem Soc 2012;23:931–7. http://dx.doi.org/10.1590/ mesostructured silicas. Appl Catal A Gen 2008;346:44–51. http://dx.doi.org/
S0103-50532012000500019. 10.1016/j.apcata.2008.04.041.
 Renewable and Sustainable Energy Reviews 62 (2016) 804–814

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

Glycerol acetals and ketals as possible diesel additives. A review

of their synthesis protocols
Ancuţa Roxana Trifoi a,b,n, Paul Şerban Agachi b, Timea Pap a
a
Research Institute for Auxiliary Organic Products , Carpati St, No. 8, RO-551022, Mediaş, Sibiu, Romania
b
Babes-Bolyai University, Faculty of Chemistry and Chemical Engineering, Arany Janos St., No. 11, RO-400028 Cluj-Napoca, Romania

art ic l e i nf o a b s t r a c t

Article history: With the rapid development of the biodiesel industry all over the world, a large surplus of glycerol has
Received 18 April 2015 been created, so economic uses of glycerol for value-added products are critical for the sustainability of
Received in revised form the biodiesel industry. One of the main interests in recent years for glycerin utilization is its conversion in
16 August 2015
acetals and ketals, with the potential to be used as fuel additives. Glycerol acetals and ketals can be
Accepted 3 May 2016
synthesized through the acid-catalyzed reaction of glycerol with aldehydes and ketones, respectively.
Available online 17 May 2016
This paper reviews different approaches and techniques used to obtain glycerol acetals and ketals,
Keywords: regarding the reactor design, catalyst design and the effect of different parameters in the reaction system.
Glycerol & 2016 Elsevier Ltd. All rights reserved.
Biodiesel
Acetal
Ketal

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804
2. Reaction mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
3. Chemical conversion of glycerol to acetals and ketals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
3.1. Catalyst design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 805
3.2. Influence of the reactant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 808
3.3. Influence of temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
3.4. Influence of the stirring speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 809
3.5. Reactor design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 812
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 813

1. Introduction The market for biofuel expanded recently as it is used on

increasingly large scales in the transport sector, especially for road
Nowadays one of the global interests focuses on renewable vehicles, in aviation, to generate electricity, for cooking and in
energy sources due to fossil fuel depletion and environmental maritime transport [3]. In 2013, the global biodiesel production
degradation. Biodiesel is considered as an attractive alternative to reached more than 110 billions of liters (see Fig. 1).
diesel fuels, because it is renewable, biodegradable, non-toxic and By 2022, the world production of bioethanol is estimated to
has almost the same properties as diesel fuel [1,2]. reach 168 billion liters and the production of biodiesel, 141 bil-
lion liters, with the contribution of Renewable Fuel Standard
(RFS2) in the United States, along with EU-RED. This will trigger a
n
Corresponding author at: Research Institute for Auxiliary Organic Products, rise in the global production of bioethanol and biodiesel by 70
Carpati St, No. 8, RO-551022, Mediaş, Sibiu, Romania.
E-mail addresses: [email protected],
per cent until 2022, as compared to the average from 2010 to
[email protected] (A.R. Trifoi). 2012. [4].

http://dx.doi.org/10.1016/j.rser.2016.05.013
1364-0321/& 2016 Elsevier Ltd. All rights reserved.
 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814 805

2. Reaction mechanism

To prepare glycerol acetal/ketal, the other component, aside

glycerol, is an aldehyde or a ketone. Acetal formation is a rever-
sible reaction via a two-step mechanism. Scheme 1 presents the
general mechanism without specifying any reaction conditions.
The first step is the formation of a hemiacetal, followed by the
removal of water molecules. Acetal formation is strongly affected
by electronic and steric factors, but it is generally accepted that the
rate determining step is the formation of a cation from the pro-
tonated hemiacetal. In order to compensate for the low rate of
hemiacetal formation, the reaction media must be sufficiently
acidic to promote effective protonation of any hemiacetal formed
and sufficiently polar to allow stabilization of the cationic inter-
mediate [14].
For the inferior aldehydes and ketones, the equilibrium is
Fig. 1. Increasing production of biodiesel since 2005 to 2013 [3].
shifted to the right.
Cyclic acetals or ketals are prepared by reacting a polyol with
the appropriate aldehydes or ketones [16]. When glycerol is used
as polyol, the 1,3 dioxane and 1,3 dioxolane structures are formed.
Cyclic ketals which result from the reaction of glycerol and
ketones have a structure of 1,3 dioxolane (5 membered ring) and
not a 1,3 dioxane (6 membered ring) due to the steric blocking of
radicals and hydrogen atoms from positions 4 and 6 of the
molecule, especially if taken into account that the C–O bond is
shorter (1.43 Å) than the C–C bond (1.54 Å). Glycerol reacting to
aldehydes results in 1,3 dioxane, with equatorial radical R [17]. The
reaction mechanism for glycerol acetals/ketals synthesis is pre-
sented in Scheme 2.
Acetalization/ketalization of glycerol is highly dependent on
the experimental conditions. The different conditions and perfor-
mances reported in the literature for glycerol acetals/ketals
synthesis are presented as follows.

Fig. 2. Crude glycerol price trend [8]. 3. Chemical conversion of glycerol to acetals and ketals

Although biodiesel represents a renewable, convenient and 3.1. Catalyst design

environmental friendly alternative to fossil fuel, the economic
feasibility of the biodiesel production must be improved. This can Traditionally, the reaction of glycerol with aldehydes/ketones is
be done by acting on three elements of the process: the raw carried out over mineral acids as catalysts like H2SO4, HCl, HF,
materials, the synthesis itself and the byproducts [5]. H3PO4, and p-toluenesulphonic acid, etc [7,9,17,19,20] and over
Biodiesel is produced by the transesterification of vegetable oils homogenous Lewis catalysts [18]. Recently, several studies were
or animal fat with methanol, in base catalysis conditions [6]. conducted regarding the use of heterogenous acid catalysts for
As for the byproducts, glycerol is obtained at high concentra- acetalysation of different carbonyl compounds with glycerol: ion
tion in a weight ratio of 1/10 (glycerol/biodiesel). Historically, exchange resins, zeolites, montmorillonite, metal oxide based
disposal of crude glycerin has been by incineration: the byproduct catalysts, activated carbons, mesoporous silicaheteropoly acids,
has not historically been used as a raw material for secondary mesoporous silicates with arylsulphonate group, rare-earth tri-
flates, nanoparticles supported on multi-walled carbon nanotubes
reaction [7]. As such, it could be especially advantageous to con-
or ion liquids. When studying the influence of the catalyst on the
vert it to added-value products, especially because the growing
reactant conversion, the research focused on the type of catalyst,
market of biodiesel has generated a glycerol oversupply and con-
acidity, porosity, catalyst loading, different techniques of catalyst
sequently the glycerol commercial price fell down from 25 cents/lb
preparation and its reusability. Many researchers studied only the
in 2004, to less than a half in 2011, approximately 8 cents/lb, as
influence of the catalyst loading, but there are also studies on
seen in Fig. 2 [8].
different types of catalysts.
One of the promising utilizations of low value glycerol is to be
Starting with homogenous catalysts (strong mineral acids and
transformed into fuel additives. Among ethers and esters, glycerol lewis catalysts) which cause corrosion and environmental pro-
acetals and ketals constitute excellent compounds for diesel and blems, and complicate the products’ separation and purification,
biodiesel blends [9]. They improve the octane number and the cold many studies were conducted with heterogenous catalysts that
flow properties, reduce particulate emission and gum formation solve most of the shortcomings of homogenous catalysts. The
[10–13]. efficiency and easy removal of the catalyst are not the only
The paper has collected and analyzed data about the progress important aspects, but also the accessibility and the price of the
and research findings on glycerol acetals/ketals synthesis in the catalyst and a good activity in mild conditions.
scientific literature. It focuses on aspects which include the reactor Ruiz et al. [18] studied the acetalization of glycerol with for-
design, catalyst design and the effect of different parameters in the maldehyde over three different types of acid catalysts: zeolites and
reaction system. organic resins - Bronsted acid, gold salts - Lewis catalysts and a
806 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814

Scheme 1. Acetal/ketal formation mechanism [15].

reusable gold complex. They obtained the highest yield of acetal USY3 for a glycerol conversion of 84%. They compared the activity
(up to 93%) over AuCl3 catalyst in mild conditions (T ¼80 °C, of USY 3 with BEA and ZSm-5 zeolite and with Dowex resin and it
reaction time ¼4 h). They also studied the reusability of the gold resulted that BEA zeolite showed a higher catalytic activity
catalyst and found that it may be reused several times with a (75.6 mmol/h gcat) than USY 3. They also studied the influence of
490% yield. With the other catalysts, the yield never exceed 80%. catalyst loading for BEA zeolite and they observed that increasing
Deutsch et al. [21] is another team which studied the acet- the catalyst load from 0.1 g to 0.3 g showed a substantial effect on
alization reaction over different types of catalysts. They tested acid glycerol conversion. When studying the BEA catalyst stability, they
polymers: Amberlyst-36 and Nafion–H NR 50, a clay mineral: showed that the zeolite lost only 7% of its initial activity after the
Montmorrilonite K-10 and a zeolite H-BEA. Due to the high activity fourth use.
and the small amounts used in the glycerol condensation, Oprescu et al. [24] synthesized a supported mixed oxide solid
Amberlyst -36 was selected as the catalyst for the reaction. The acid catalysts (SO42  /SnO2), which is water tolerant and stable at
yield in glycerol acetal, over Amberlyst 36 was up to 94% for high temperatures. The catalyst was prepared by using the method
glycerol-benzaldehyde condensation, 77% for glycerol–for- described by Chavan and partners [25,26]. The highest yield of
maldehyde condensation and 88% for glycerol–acetone 97.5% was obtained for glycerol ketals of cyclohexanone.
condensation. SnO2- based solid acids were also synthesized and tested by
Maksimov et al. [22] studied the reaction between glycerol and Mallesham et al. [27]. They improved the surface proprieties of
acetone over heterogenous catalysts (ion exchange resins and SnO2 by addition of WO3, MoO3 and SO42  . The best performances
different zeolites). The yield in solketal was significantly higher were achieved by SO42  /SnO2, with a glycerol conversion with
over zeolite catalysts than on polymer based catalyst. acetone and furfural of 98% and 99%.
De Torres et al. [23] selected the Montmorrilonite K10 catalysts Under optimum conditions Fan et al. [28], obtained a glycerol
for condensation of glycerol with ciclopentanone out of the other conversion to acetal of up to 95%, over TiO2–SiO2 mixed oxides
studied catalyst (Nafion NR-50 and Nafion SAC-13- acid resins), which naturally consist only of Lewis acidic sites and can produce
due to its good activity and price. With all the studied catalysts, a the Brønsted acidic sites after adsorbing water.
high solketal yield of 81.4–100% was obtained by the condensation Zirconia based mixed oxydes catalysts, impregnated with
of glycerol with ciclopentanone. molybdate ions (Mox/TiO2–ZrO2), exhibited superior glycerol
Acetalization of glycerol with butanal over a range of zeolites conversion with benzaldehyde. The conversion was up to 74%, in a
with different pore structures and acidity was studied by Serafim short reaction time- 30 min [29].
et al. [14]. Different zeolites - USY with different Si/Al ratio, BEA Ferreira et al. [30] demonstrated the efficiency of silica-
and ZSM-5, were used as catalysts. They found that by increasing included heteropolyacids for the acetalisation of glycerol with
the Si/Al ratio of USY zeolite, the surface area and the total pore acetone. They prepared the catalysts by the sol–gel method
volume increase as well. The maximum activity was obtained with according to Izumi et. al [31]. The catalysts were prepared from
 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814 807

Aldehyde/ketone

Scheme 2. Reaction mechanism of glycerol with aldehydes/ketones [18].

conditions (30 °C, glycerol:acetone molar ratio of 1:6, 3wt% cata-

lyst with respect to total reactant weight) exhibited a glycerol
conversion of 94% in 2 h of reaction. The catalyst also showed a
remarkable resistance toward deactivation due to water, with a
decrease in conversion of only 3%.
In their study, Da Silva et al. [34] tested various catalysts:
amberlyst 15, K10 montmorilonite, p-toluene sulfonic acid and
zeolite beta, for glycerol condensation with aqueous formaldeyde
and acetone.
Among these catalyst the best results were obtained with
zeolite Beta Si/Al ¼16:1, which also showed a resistance toward
deactivation due to water. The glycerol conversion was 95% in
60 min of reaction.
In their study, Khayoon et al. [35] compared the catalytic
activity of various loaded activated carbon (AC) catalysts with
different amounts of Ni and/or Zr during glycerol acetalization
with acetone. For the catalysts loaded with Ni (from 1 to 5%), the
Fig. 3. Effect of catalyst weight on glycerol conversion (G:Ac¼1:8; T¼ E °C, t ¼ 3 h) conversion of glycerol increased from 65% to 98%, but the Zr loa-
[35]. ded catalyst (from 1 to 5%) showed a lower conversion (54% to
67%) than those obtained with Ni. They prepared a catalyst for-
tungstophosphoric acid (PW), tungstosilisic (SiW), molybdopho- mula X%Ni-Y%Zr/AC (X and Y values from 1 to 5), and the catalyst
sphoric acid (PMo) and molybdosilisic acid (SiMo) immobilized in with the formula 5%Ni-1%Zr/AC showed the best activity, with
silica. They also studied the stability of the catalysts and after the 100% conversion for glycerol. The effect of catalyst amount was
forth usage, the catalysts lost only 10–13% of their initial activity. significant for 0.2 g of catalyst compared to 0.15 g (Fig. 3). Based on
Sulfonic acid functionalized meso structured silicas show a good their previous studies Khayoon and his team [36] developed other
activity in the acetalization reaction of glycerol and acetone and Ni-type of catalyst, based on multi-walled carbon nanotubes
the initial activity of the catalyst is preserved up to three utiliza- (MWCNT’s), due to the thermal stability, efficiency and affordable
tions, without regeneration treatment [32]. price as compared to ion exchange resins. A high conversion of
An efficient heteropolyacid catalyst, for the condensation 96% with respect to glycerol, resulted in a 3 h reaction time, at
reaction of glycerol with acetone at room temperature, was pre- 40 °C when using a catalyst containing 1,8% Ni on MWCNT’s.
pared and tested by Sandesh et al. [33]. The organic–inorganic Umbarkar et al. [37] studied the reaction of glycerol with
hybrid catalyst is the tetrapropylammonium salt of phoshotungstic benzaldehyde using a molybdenum oxide catalyst supported on
acid, [(CH3 CH2 CH2)4N)3]PW12O40] and for the optimized reaction SiO2, synthesized by the sol–gel technique. They varied the
808 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814

molybdenum oxide molar concentrations (1,5,10,15,20) and they highest conversions and yields were achieved over the rare-earth
discovered the MoO3/SiO2 catalysts with 20% Mo, the most acidic triflate catalysts [39], the catalysts loaded with Ni, that developed
and highly active catalyst in the series (72% conversion). by Khayoon and his team [35,36], and over natural material,
Adam et al. [38] studied the acetalization reaction of glycerol Montmorrilonite K [23], with a glycerol conversion up to 100%.
with benzaldehyde, over silica supported acid catalyst. They Other catalysts with good activity (with a 4 90% conversion
immobilized sulfonic acid onto silica obtained from rice husk ash, achieved for the glycerol) were the ionic liquids catalysts [42], the
which is a cheap silica source. The conversion of benzaldehyde gold catalysts [18], ion exchange resins [23,43] etc. The efficiency
was up to 78%, but after a long reaction time of 8 h. Also the and easy removal of the catalyst are not the only important
temperature was 120 °C and they used an argon atmosphere. aspects, when selecting a catalyst, but also the operating condi-
Rare-earth triflate catalysts were tested by Pierpont et al. [39] tions for industrial catalysts which are different from those used in
for the acetalization of glycerol with acetone because they are laboratory studies. Therefore, industrial catalysts must have spe-
recyclable and water compatible and also a 100% conversion of cial requirements which can be reduced to the following:
glycerol and acetone into solketal can occur at room temperature.
Indium (III) triflate (In(OTf)3) catalyst was used in mild condi- – the catalyst should not be very sensitive to poisoning and
tions by Smith et al [40] for the transacetalization reactions of impurities because the reagents used at industrial scale are not
glycerol with acyclic acetals to generate the corresponding cyclic such pure as those used in the laboratory.
acetals. – the production capacity should not be very small.
Other researchers who studied the acetalization of glycerol – they must have thermal and mechanical resistance.
with acetone at room temperature are Manjunathan et al. [41]. – they should be recyclable.
They screened the properties (acidity, porosity and cristallite size)
of different types of acid zeolites and the best results were 3.2. Influence of the reactant
obtained for the H-Beta zeolite with an average of crystallite size
of 135 nm. A glycerol conversion of 86% was obtained at room Most works were done with different types of aldehydes or
temperature, an acetone to glycerol molar ratio of 2:1 and 1 h of ketones, with the variation of molar ratios of reactants. There were
reaction time. investigated the effect of adding solvent and impurities to glycerol
Ionic liquids can be used as catalysts for the acetalization conversion.
reactions. Wang et al. [42] studied the reaction between glycerol Almost all the studies presented were conducted strictly
and benzaldehyde, on ionic liquids catalyst, and they obtained an experimentally, without a preliminary calculation of the process
acetal yield of 99.8% at room temperature. parameters.
Another class of catalysts with good activity at room tem- The reaction between glycerol and formaldehyde was studied
perature are silicotungstate types. Narkhede et al. [43] synthesized by Ruiz et al. [18], Deutsch et al. [21], Agirre et al. [50] and Cole-
and tested silicotungstate catalysts for glycerol and benzaldehyde man et al. [7].
reaction and they observed the selectivity towards dioxolane The general reaction between glycerol and formaldeyde and
derivates when using this type of catalysts. the resulting reaction products are presented in Scheme 3.
The selectivity for the five-membered cyclic ketals was also Coleman et al. [7] developed a solvent-free technology for
observed by Crotti et al. [44] when using an organoiridium glycerol formal preparation, from paraformaldehyde and crude
catalyst. glycerol. After the reaction, the mixture is cooled and neutralized
Niobium catalysts also present good activity for the con- and the products are separated through the vacuum distillation
densation reaction of glycerol with various aldehydes and ketones. method.
Souza et al. reported in their studies [45,46] the characteristics and Different mole ratios between glycerol and formaldehyde were
the catalytic activity of niobium oxyhydroxyde for the glycerol investigated by Agirre et al. [50]. They performed the experiments
condensation with acetone. The glycerol conversion was 74%, for a at 353 K, using Amberlyst 47 catalyst, varying the molar ratio of
glycerol:acetone molar ratio of 1:4, at 70 °C, after 1 h of reaction. glycerol:formaldehyde from 1:1; 1:2 and 1:3. By increasing the
Good performances were achieved by niobia catalysts for glycerol molar ratio of glycerol, the equilibrium conversion of the for-
acetalisation by Nair et al. [47], with high conversions of maldehyde increases as well. In another study, Agirre et al. [51]
about 80%. found that for the reaction of glycerol with acetaldehyde, the
Nair et al. [47] studied the condensation reaction of glycerol overall conversion reaches 100% in all cases when using glycerol in
with various aldehydes and ketones (paraformaldehyde, benzal- excess. Similar observation was reported by Nanda et al. [9] in the
dehyde, fufural and acetone) over aluminosilicate-supported iron case of glycerol and acetone condensation, that is, the reaction
oxide nanoparticles. A high glycerol conversion (99%) was thermodynamics and kinetics are strongly affected by acetone:
obtained at 100 °C, over Fe/Al-SBA-15 catalyst. glycerol molar ratio. For a molar ratio of acetone:glycerol of 1.48:1
AlPO4 based green solid catalysts, were synthesized and tested the solketal yield is 68% and for a molar ratio of acetone:glycerol of
by Zhang and his coworkers [48]. Different catalysts: M–AlPO4, 2.46:1 the solketal yield is 74%. They also studied the effect of
Zn–AlPO4, Cu–AlPO4, Ni–AlPO4 and Co–AlPO4 were tested for the ethanol as a solvent, mainly to improve the solubility in acetone,
acetalization of glycerol with acetone at 80 °C. The highest yield but the effect was found to be negligible. Ferreira et al. [30] stu-
was obtained with Ni–AlPO4 catalyst (75.44%). died the influence of acetone in excess on glycerol conversion. By
The acetalization of glycerol with aldehydes and ketones is increasing the molar ratio glycerol:acetone (from 1:3 to 1:12) an
generally carried out using acid catalysts, but Prakruthi et al. [49], increase in glycerol conversion was observed, but did not affect the
used Mg–Al–LDH as base catalyst for the acetalization of glycerol
and it exhibited maximum conversion of glycerol into 5-
membered cyclic acetal.
For glycerol acetalization with aldehydes or ketones various
types of catalysts has been tested: natural catalysts (silicates and
aluminosilicates), supported catalysts (impregnated catalysts:
salts, oxides, acids and metals), catalysts obtained by precipitation
(salts and oxides), organic catalysts (ion exchange resins) etc. The Scheme 3. Reaction between glycerol and formaldehyde.
 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814 809

selectivity to solketal. Khayoon et al. [35] used 5%Ni–1%Zr/AC Serafim et al. [14] varied the temperature from 30 °C to 70 °C
catalyst and asserted that by increasing the molar ratio of glycerol: and, as expected, the conversion of glycerol when reacting with
acetone from 1:4 to 1:8, not only did the glycerol conversion butanal increased with the temperature (from 40% to 87%) and the
increase from 52% to 100%, but it also promoted the formation of selectivity to five–membered cyclic acetal also increased. The same
6 membered cyclic ketal. observation was made by Khayoon et al. [35] for glycerol and
In case of glycerol and n-butyraldehyde, Guemez et al. [53] also acetone reaction: there was a steady increase in the conversion
reported that for a reaction time of 100 min, at 353 K, the con- with the increase temperature and the reaction favored towards
version of n-butyraldehyde increased from 88% to 98% when the the formation of the five membered cyclic acetal. Guemez et al.
initial glycerol:butyraldehyde increased from stoichiometric ratio [53] carried out the reaction between glycerol and butyraldehyde
(1:1) to the highest molar ratio studied (3:1). The conversion of in a temperature range of 323 K–353 K using stoichiometric feed
glycerol was 100% after 40 min of reaction time when glycerol: ratio and over Amberlyst 47 catalyst. With the increase of tem-
butyraldehyde molar ratio is 0.2, whereas in the same studied perature, increased the overall reaction rate, but in the absence of
reaction period, it varies between 93% and 98% (G:B ¼ 0.5). a catalyst, the temperature increase had no influence on the final
Experiments showed that when glycerol is the limiting reagent equilibrium conversion. The same was observed by Agirre et al.
and it is almost totally exhausted in the reaction medium, small [50] in their study for glycerol with acetaldehyde and
amounts of 2,4,6-tripropyl-1,3,5 trioxane were detected. formaldehyde.
Serafim et al. [14], also studied the influence of molar ratio of In contrast to the above, Shirani et al. [58] showed that the
glycerol/butanal on the conversion of glycerol. Over BEA zeolite solketal yield is reduced when the temperature and the acetone to
catalyst, at 353 K, the conversion of glycerol after 4 h of reaction glycerol molar ratio are increased. When the temperature increa-
increased from 71% to 87% for a molar ratio of glycerol/butanal of ses acetone is vaporized while the glycerol is in the liquid phase on
1:1 and 1:2.5, respectively. But by increasing the molar ratio to 1:6, the surface of the catalyst, therefore, there will be a lesser effi-
the increase of conversion is no longer observed (88%). ciency of the interaction of the gaseous acetone molecules with
Faria et al. [54] studied the influence of different solvents on the liquid glycerol molecules on the catalyst surface. This leads to a
the production of glycerol ethyl acetal, through the acetalization of reduction in the solketal production and a lower yield and con-
glycerol with acetaldehyde, in a simulated moving bed reactor, version. Nanda et al. [9] also observed that a higher temperature
using Amberlyst 15 as a catalyst. Between dimethylsulfoxide, N,N led to a lower product yield, typically for exothermic reactions.
dimethyl formamide and acetonitrile, dimethylsulfoxyde proved to Only the initial reaction rate is positively influenced by the
be the most suitable solvent for the reaction, due to its char- increase of the reaction temperature.
acteristics: it is unreactive, miscible with the reaction medium and Deutsch et al. [21] found that for catalytic condensation of
presents an adsorption capacity for the catalyst adsorbent. glycerol with paraformaldehyde over Amberlyst -36, a mild reac-
An efficient, environment friendly, simple and available tin- tion temperature is beneficial to increase the quantity of the 6-
based catalyst was used in the glycerol ketalization with different membered cyclic acetal in the mixture.
ketones at room temperature by Da Silva et al. [55]. The conver- Contradictory conclusions are drawn about the impact of the
sions of ketones were between 40% (for 4-methyl 2 pentanone) temperature towards the reaction rate or selectivity of the
and 98% (for cyclohexanone). products.
The same group of researchers [56] studied the influence of
different impurities (methanol, water and NaCl) found in crude 3.4. Influence of the stirring speed
glycerol, on the reaction of glycerol with acetone. The results
showed that 150 g water/kg glycerol drastically decreased the The literature reported many studies which investigated the
conversion of glycerol from 900 to 690 g/kg. The same dramatic effects of mass transfer on the reaction kinetics. It was shown that
effect has the NaCl and the concomitant addition of water and the stirring speeds have no influence on the reaction kinetics. For
NaCl, which decreases the glycerol conversion up to 500 g/kg. example, Nanda et al. [9] carried out the experiments under two
Another group of glycerol based acetals which has received stirring speeds 400 and 1100 rpm, and there was no influence on
attention in recent years is the furyl-1,3- dioxacyclanes, derived the reaction time.
from furans, furfural and hydroxymethylfurfural (biomass mate- The same conclusion was reached by Agirre et al. [51] when
rials). Wegenhart et al. [52] synthesized acetals derived from reacting glycerol with acetaldehyde, at 700 rpm and 1250 rpm and
glycerol and furfural over homogeneous Lewis acid catalysts and a glycerol:acetaldehyde molar ratio of 3:1.
heterogeneous solid acids. They studied the reaction in the pre- It is known that the mixing operation is influenced by various
sence and in the absence of nitrogen flow and obtained increased parameters: the reactants’ properties (physical properties, the
yields for nitrogen atmosphere. molar ratio between them), the product’s properties (density,
The reactions were conducted in the presence or in the absence viscosity, solubility, etc.) and the parameters related to the
of a solvent. Solvent free processes are preferred because of the operation itself (power of the agitation, continuous or dis-
environmental limitations. Most of the experiments were con- continuous mode, the equipment used, etc.) [59].
ducted using excess of reagents and the same conclusion was For example, from these two articles [9] and [51], it can be
drawn by many researchers, that the conversion increases when concluded that agitation speed has no influence on the reaction
the molar ratio of the reactants increases as well. An important kinetics if the reactants are in high excess. When working in
concern was emphasized by Da Silva et al. [56], that he impurities stoichiometric ratio, Agirre et al. [50] reported the same, but for
from glycerol have a high influence on glycerol conversion. high stirring speeds over 1200 rpm. It is not reported what hap-
pens for lower stirring speeds.
3.3. Influence of temperature Xong et al. [60] observed that the rate of conversion is inde-
pendent of the stirring rate only above 800 rpm. For a stirring
One of the most important parameter which determines the speed of 500 rpm, the conversion is almost 50% lower than the
equilibrium of the reaction is the temperature. The impact of the conversion at 800 rpm.
temperature on the reaction rate, conversion and selectivity has Some researchers report that stirring speed has no influence
been discussed in most reports. and some that it has. Different conclusions were drawn because
810 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814

Table 1
Reaction parameters reported by different studies for the glycerol acetalization with aldehydes and ketones.

No. Catalysts Reaction conditions Results Comments Reference

Homogenous catalyst
1 PTSA Glycerol:formaldehyde ¼ 1:1 Acetal yield ¼ 80% - [18]
T ¼80 °C
t¼ 9 h
2 PTSA Glycerol:acetone ¼ 1:2; 1:4;1:6 Glycerol conversion¼ 60– - [20]
t¼ 12 h 82,7%
3 PTSA Glycerol:benzaldehyde ¼ 1:2 Glycerol conversion¼ 67% Microwave assisted reaction; [65]
T ¼140 °C power¼ 600 W;
t¼ 15 min The conversion without catalyst
was 95%
4 H2SO4 Glycerol:formaldehyde ¼ 1:1-1,5:1 Acetal yield ¼ 69–89% - [7]
T ¼100 °C
t¼ 4 h
5 H3PO4 Glycerol with isobutyraldehyde High purity acetal (95%) - [19]
N2-atmosphere
Heterogenous catalyst
Micro- and mesoporous materials
6 Zeolite beta Glycerol:formaldehyde ¼ 1:1 Acetal yield ¼ 25% - [18]
T ¼100 °C
t¼ 2 h
7 Zeolite beta Glycerol:acetone ¼ 1:2 Glycerol conversion¼ 90% - [56]
T ¼70 °C
t¼ 1 h
8 Zeolite beta CP814E Glycerol:acetone ¼ 1:2; 1:6 Glycerol conversion¼ 60– Flow reactor [22]
T ¼35 °C 82%
t¼ 4 h
9 Zeolite beta CP811T1 Glycerol:acetone ¼ 1:2; 1:6 Glycerol conversion¼ 62– Flow reactor
T ¼35 °C 85%
t¼ 4 h
10 Zeolite HY Glycerol:acetone ¼ 1:2; 1:6 Glycerol conversion¼ 21– Flow reactor [22]
T ¼35 °C 37%
t¼ 4 h
11 H beta zeolite Glycerol:acetone ¼ 1:1; 1:2; 1:3 Glycerol conversion¼ 62– - [41]
room temperature t¼ 2 h 86%
12 Zeolite USY Glycerol:butanal ¼ 1:2,5 Glycerol conversion¼ 45– - [14]
T ¼70 °C 72%
t¼ 4 h
13 Zeolite BEA Glycerol:butanal ¼ 1:2,5 Glycerol conversion¼ 87% - [14]
T ¼70 °C
t¼ 4 h
14 Zeolite ZSM Glycerol:butanal ¼ 1:2,5 Glycerol conversion¼ 28% - [14]
T ¼70 °C
t¼ 4 h
15 Zeolite H BEA Glycerol:butanal ¼ 1:2,5 Yield ¼94% Chloroform-solvent [21]
T ¼70 °C
t¼ 4 h
16 Montmorrilonite K10 Glycerol:benzaldehyde ¼ 1,1:1 Yield ¼95% Chloroform-solvent [21]
t¼ 6 h
17 Montmorrilonite K10 Glycerol with acetone Zeolite membrane reactor [66]
18 Montmorrilonite K10 Glycerol:benzaldehyde ¼ 1:2 Glycerol conversion¼ 84% Microwave assisted reaction; [65]
T ¼140 °C power¼ 600 W;
t¼ 15 min The conversion without catalyst
was 95%
Polymers
19 Nafion SAC 13 Glycerol:benzaldehyde ¼ 1:2 Glycerol conversion¼ 81% Microwave assisted reaction; power [65]
T ¼140 °C ¼ 600 W;
t¼ 15 min The conversion without catalyst
was 95%
20 Dowex Glycerol:butanal ¼ 1:2,5 Glycerol conversion¼ 66% - [14]
T ¼70 °C
t¼ 4 h
21 Amberlyst 36 Glycerol:acetone ¼ 1:1,5 Glycerol conversion¼ 88% Dichlormethane - solvent [21]
T ¼40 °C
t¼ 8 h
Glycerol:benzaldehyde ¼ 1,1:1 Glycerol conversion¼ 94% Chloroform - solvent [21]
T ¼56-64 °C
t¼ 4 h
Glycerol:formaldehyde ¼ 1,1:1 Glycerol conversion¼ 58% Toluene- solvent [21]
T ¼84-111 °C
t¼ 2 h
Glycerol:formaldehyde ¼ 1,1:1 Glycerol conversion¼ 62% Benzene- solvent [21]
T ¼69-88 °C
t¼ 4 h
Glycerol:formaldehyde ¼ 1,1:1 Glycerol conversion¼ 77% Chloroform-solvent [21]
T ¼56-61 °C
 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814 811

Table 1 (continued )

No. Catalysts Reaction conditions Results Comments Reference

t¼ 6 h
Glycerol:formaldehyde ¼ 1,1:1 Glycerol conversion ¼74% Dichlormethane-solvent [21]
T ¼ 38-40 °C
t¼ 14 h
22 Amberlyst 36 Glycerol:formaldehyde ¼ 1:1 Yield ¼ 55% - [18]
T ¼ 100 °C
t¼ 4 h
23 Amberlyst 15 Glycerol:formaldehyde ¼ 1:1; Glycerol conversion Reactive distillation process [64]
1:1,5:1:2 almost 100%
T ¼ 75 °C
t¼ 2 h
24 Amberlyst 15 Glycerol: acetone¼ 1:2 Glycerol conversion - [56]
T ¼ 70 °C almost 95%
t¼ 1 h
25 Amberlyst 15 Glycerol: acetaldehyde Glycerol conversion Simulated moving bed reactor [54]
almost 95%
26 Amberlyst 15 Glycerol:benzaldehyde ¼1:2 Glycerol conversion ¼80% Microwave assisted reaction; [65]
T ¼ 140 °C power¼ 600 W;
t¼ 15 min The conversion without catalyst
was 95%
27 Amberlyst 35 Glycerol:acetone ¼ 1,48:1; 2:1; Yield ¼ 60–74% ethanol-solvent [9]
2,48:1
T ¼ 25-45 °C
28 Amberlyst 47 Glycerol:formaldehyde ¼ 1:1; 2:1; Formaldehyde con- - [50]
3:1 version¼60–75%
T ¼ 80-100 °C
t¼ 3 h
29 Amberlyst 47 Glycerol:acetaldehyde ¼1:1; 2:1; Acetaldehyde con- - [51]
3:1 version¼85–90%
T ¼ 10-50 °C
t¼ 4 h
30 Amberlyst 47 Glycerol:butiraldehyde ¼ 1:1; 3:1 Benzaldehyde con- - [53]
T ¼ 80 °C version¼66–95%
t¼ 100 min
Glycerol:butiraldehyde ¼ 0,2:1; Glycerol conversion ¼98– - [53]
0,5:1 100%
T ¼ 60 °C
t¼ 4 h
31 KU-2 Glycerol: acetone¼ 1:2; 1:6 Glycerol conversion ¼55– Flow reactor [22]
T ¼ 60 °C 85%
t¼ 4 h
32 Purolite PD 206 Glycerol: acetone¼ 5:1 Acetone conversion¼ 95% Continuous reactor [58]
T ¼ 20 °C Acetone- solvent
P¼ 120 bar

33 Ni–activated carbon Glycerol:acetone ¼ 1:8 Glycerol conversion ¼65– - [35]

T ¼ 45 °C 98%
t¼ 3 h
34 Zr–activated carbon Glycerol:acetone ¼ 1:8 Glycerol conversion ¼54– - [35]
T ¼ 45 °C 67%
t¼ 3 h
35 X%Ni-Y%Zr/activated carbon Glycerol:acetone ¼ 1:8 Glycerol conversion ¼68– Reduction of catalytic activity from [35]
T ¼ 45 °C 100% the 1st to the 4th run by 3%
t¼ 3 h
36 Ni-MWCNT’s (multi-walled carbon nanotubes) Glycerol:acetone ¼ 1:6 Glycerol conversion ¼63– Reduction of catalytic activity from [36]
T ¼ 40° 96% the 1st to the 4th run by 5%
t¼ 3 h
37 Aluminosilicate supported iron oxide T ¼ 100 °C Glycerol conversion ¼90% -
38 M–AlPO4 Glycerol:acetone ¼ 1:8 Glycerol conversion: Reduction of catalytic activity of [48]
M–ZnAlPO4 T ¼ 80 °C M-NiAlPO4 from the 1st to the 5th
M–CuAlPO4 t¼ 1 h run by 80%
M–NiAlPO4
M–CoAlPO4 57,33%–75,44%
Metal oxides
39 SO4/SnO2 [24]
40 SO4/SnO2 Glycerol with acetone and furfural Glycerol conversion ¼98– - [27]
99%
41 TiO2–SiO2 Glycerol:acetone ¼ 1:2; 1:3; 1:4; Glycerol conversion ¼72– Up to 95% to 5-membered ring ketal [28]
1:5; 1:6 98% selectivity
T ¼ 40–90 °C
t¼ 3 h
42 Mox/ TiO2–ZrO2 Glycerol:benzaldehyde ¼1:1 Glycerol conversion max- - [29]
T ¼ 60-100 °C imum 74%
t¼ 90 min
43 MoO3/SiO2 Glycerol:benzaldehyde ¼1,1:1 Conversion¼ 37–72% Using toluene as solvent [37]
T ¼ 100 °C
812 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814

Table 1 (continued )

No. Catalysts Reaction conditions Results Comments Reference

t¼ 8 h
44 Niobium oxyhydroxyde Glycerol:acetone ¼ 1:4 Conversion¼ 74% - [45,46]
T ¼40 °C
t¼ 1 h
45 Nb2O5 Glycerol:acetone ¼ 1:3 Conversion¼ 80% 4 times reused catalyst [47]
T ¼70 °C
t¼ 6 h
Other catalysts
46 Rare earth triflate Glycerol with acetone room Conversion 100% - [39]
temperature [40]
47 Ion liquids Glycerol:benzaldehyde room Yield ¼99,8% - [42]
temperature
48 Organic–inorganic hybrid catalyst (tetra- Glycerol:acetone ¼ 1:6 Glycerol conversion 94% Resistance of the catalyst toward [33]
propylammonium salt of phosphotungstic T ¼30 °C deactivation due to water
acid) t¼ 3 h

the mass transfer may have influence on the reaction kinetics, only To remove the water from the system without using a solvent,
in certain conditions. Roldan et al. [66] studied the ketalization of glycerol in a zeolite
membrane reactor, over K10 montmorillonite catalyst. The
3.5. Reactor design improvement in this case is the reduction in the excess of acetone,
and the water removal from the system without using a solvent.
Most of the laboratory studies for glycerol condensation with Innovative technologies with process intensification for acetals
aldehydes/ketones were undertaken in batch reactors. A batch production were developed by Perreia et al. [67] using Simulated
process has some limitation for scale up, such as poorly defined Moving Bed Reactor (SMBR) and Simulated Moving Bed Mem-
interfacial areas as they depend on the stirring intensity. Some brane Reactor (PermSMBR), for acetalization of ethanol and 2-
technologies were studied for the glycerol acetalization in a semi- butanol with acetaldehyde. Although they require additional
batch or continuous reactors. capital and operational costs, these reactors lead to high pro-
Monbaliu et al. [61] developed a continuous flow process for ductivity and high conversion at low temperatures (10 °C–50 °C).
glycerol based additive, but the disadvantage of the process is the This is an interesting approach that may be applied to glycerol
homogenous catalyst used. acetalization.
The synthesis of glycerol–acetone solketal in a semi-batch The most promising reactor design seems to be reactive dis-
reactor was developed by Clarkson et al. [62]. The acetone was tillation with an almost 100% conversion of the glycerol, because
fed continuously, while glycerol, because of its high viscosity at from the other reactors proposed it seems to have no drawbacks
low temperatures, was fed in batch. They improved the process by such as reactor clogging with the catalyst, or high capital costs.
recycling the unconverted reactant, creating a solvent-free process Because the reaction and distillation take place at a time, the
and using of heterogeneous catalyst. operational costs are also reduced.
Pathak et al. [57] studied the conversion of glycerol to acetal via Table 1 summarizes the conversion conditions of glycerol with
cracking reaction, in a fixed bed reactor, using heterogenous cat- various aldehydes and ketones. It was structured based on the
alyst and nitrogen as carrier gas. They had very good results, a catalyst design.
glycerol conversion up to 98%, in a 60 min reaction, at low cracking
temperature (350–500 °C) and at atmospheric pressure.
Recently, Nanda et al. [63] used a continuous flow reactor for 4. Conclusions
glycerol ketal synthesis. Using different types of catalysts, they
obtained reasonable conversions of glycerol, up to 73%, but in mild Glycerol based acetals and ketals have been identified to have
conditions. One of the drawbacks of this system, as they also particular qualities as fuel additives (improve the octane number
mention, is the clogging of the reactor with the catalysts. and cold flow properties, reduce particulate emission and gum
An optimization for continuous process ketalization of glycerol formation).
with acetone, in a stainless steel continuous reactor was developed This paper summarizes the approaches and strategies of the
by Shirani et al. [58], over Purolite PD 206 catalyst. The reactants: researchers for the glycerol–aldehydes and ketones systems and
glycerol, ethanol as co-solvent and acetone as solvent (under the progress in obtaining higher conversions and higher yields of
subcritical conditions) were pumped continuously into the reactor. the products of interest. These aspects include not only the effi-
Under optimum conditions (evaluated using central composite ciency of catalyst design, reactants, temperature, reactor design on
design –CCD), the conversion was 95% (t¼ 20 °C, p ¼120 bar, gly- conversion and selectivity of the glycerol based acetals/ketals, but
cerol:acetone¼ 5:1; flow rate¼ 0.1 ml/min and catalyst mass ¼ also different approaches for an environmental and inexpensive
0.77 g). process. None of the studies present the costs and environmental
The reactive distillation is another alternative to reactor design benefits for different processes, in amount or percentage . Most of
for glycerol acetalization, used by Hasabnis et al. [64]. They eval- the researches have been conducted over heterogenous catalysts
uated the feasibility of the reactive distillation by experiments and because can be easily separated from the reaction mixture (by
simulations and they found that reactive distillation is a good filtration or centrifugation), do not require neutralization proce-
alternative to glycerol acetalization, and for the proposed config- dure and they can be used several times.
uration of the column resulted an almost 100% conversion. One gap in the literature is that only a few of the researchers
Pawar et al. [65] obtained high yields of cyclic acetals and ketals made their studies using crude or partially purified glycerol, most
by microwave assisted acetalization under catalyst and solvent of the studies used commercial glycerol. So the synthesis protocols
free conditions. will not be fully adapted for the crude glycerol because it is well
 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814 813

known that crude glycerol contains impurities that affect the [22] Maksimov AL, Nekhaev AI, Ramazanov DN, Arinicheva YA, Dzyubenko AA,
catalyst activity. Khadzhiev SN. Preparation of high octane oxygenate fuel components from
plant derived polyols. Pet Chem 2011;51:61–9.
Another gap is that no one made a difference in amount or [23] De Torres M, Jiménez-osés G, Mayoral JA, Pires E, Santos M. Glycerol ketals:
percentage between the costs of the processes. synthesis and profits in biodiesel blends. Fuel 2012;94:614–6.
From the above chapters it can be observed that the trend in [24] Oprescu EE, Stepan E, Dragomir RE, Radu A, Rosca P. Synthesis and testing of
glycerol ketals as components for diesel fuel. Fuel Proces Technol
the glycerol acetalization process is to find the mildest reaction 2013;110:214–7.
conditions possible. Solventless process, low temperatures-even [25] Jaturong J, Boonyarach K, Pramoch R, kunchana B, Lalita A, Peesamai J.
room temperatures, cheap and reusable catalysts. Transesterification of crude palm kernel oil and crude coconut oil by different
solid catalysts. Chem Eng J 2006;116:61–6.
From the presented papers, it can be concluded that valoriza- [26] Subhash PC, Zubaidha PK, Shubhada WD, Keshavaraja A, Ramaswamy AV,
tion of glycerol to glycerol acetals and ketals is of current interest Ravindranathan T. Use of solid superacid (sulphated SnO2) as efficient catalyst
for the researchers who try to find the best operating conditions, in facile transesterification of ketoesters. Tetrahedron Lett 1996;37:233–6.
[27] Mallesham B, Sudarsanam P, Reddy BM. Eco-friendly synthesis of bio-additive
by testing different catalysts, reactants, temperature and fuels from renewable glycerol using nanocrystalline SnO2- based solid acids.
reactor type. Catal Sci Technol 2014;4:803–13.
[28] Fan CN, Xu CH, Liu CQ, Huang ZY, Liu JY, Ye ZX. Catalytic acetalization of
biomass glycerol with acetone over TiO2–SiO2 mixed oxides. React Kinet Mech
Cat Lett 2012;107:189–202.
Acknowledgments [29] Sudarsanam P, Mallesham B, Prasad AN, Reddy PS, Reddy BM. Synthesis of
bio–additive fuels from acetalization of glycerol with benzaldehyde over
molybdenum promoted green solid acid catalysts. Fuel Process Technol
This work was possible due to the financial support of Sectorial 2013;106:539–45.
Operational Program for Human Resources Development 2007– [30] Ferreira P, Fonseca IM, Ramos AM, Vital J, Castanheiro JE. Valorisation of
2013, co-financed by European Social Fund under the project glycerol by condensation with acetone over silica-included heteropolyacids.
Appl Catal B: Environ 2010;98:94–9.
number POSDRU/159/1.5/S/132400 with the title “Young success- [31] Izumi Y, Hisano K, Hida T. Acid catalysis of silica-included heteropolyacid in
ful researchers – professional development in an international and polar reaction media. Appl Catal A: Gen 1999;181:277–82.
interdisciplinary environment”. [32] Melero JA, Morales G, Paniagua M, Martin E. Acetalisation of bio-glycerol with
acetone to produce solketal over sulfonic mesostructured silicas. Green Chem
2010;12:899–907.
[33] Sandesh S, Halgeri AB, Ganapati V, Shanbhag GV. Utilization of renewable
References resources: condensation of glycerol with acetone at room temperature catalyzed
by organic–inorganic hybrid catalyst. J Mol Catal A: Chem 2015;401:73–80.
[34] Da Silva CXA, Goncalves VLC, Mota CJA. Water tolerant zeolite catalyst for the
[1] Fazal MA, Haseeb ASMA, Masjuki HH. Biodiesel feasibility study: an evaluation acetalisation of glycerol. Green Chem 2009;11:38–41.
of material compatibility; performance; emission and engine durability. [35] Khayoon MS, Hameed BH. Solventless acetalization of glycerol with acetone to
Renew Sustain Energy Rev 2011;15:1314–24. fuel oxygenates over Ni–Zi supported mesoporous carbon catalyst. Appl Catal
[2] Saxena P, Jawale S, Joshipura MH. A review on prediction of properties of A: Gen 2013;464-465:191–9.
biodiesel and blends of biodiesel. Proced Eng 2013;51:395–402. [36] Khayoon MS, Abbas A, Hameed BH, Triwahyono S, Jalil AA, Harris AT, et al.
[3] The global biofuels market: energy security, trade and development. In: Pro- Selective acetalization of glycerol with acetone over nickel nanoparticles
ceedings of the United Nations Conference on Trade and Development; 2014; supported on multi-walled carbon nanotubes. Catal Lett 2014;144:1009–15.
No 30 October. [37] Umbarkar SB, Kotbagi TV, Biradar AV, Pasricha R, Chanale J, Dongare MK, et al.
[4] SSI Review. Biofuels market; 2014, 6. p. 121. Acetalization of glycerol using mesoporous MoO3/SiO2 solid acid catalyst. J
[5] De Torres M, Jimenez-oses G, Mayoral JA, Pires E. Fatty acid derivates and their Mol Catal A: Chem 2009;310:150–8.
use as CFFP additives in biodiesel. Bioresour Technol 2011;102:2590–4. [38] Adam F, Batagarawa MS, Hello KM, Al-Juaid SS. One-step synthesis of solid
[6] Rincon LE, Jaramillo JJ, Cardona CA. Comparison of feedstocks and technolo- sulfonic acid catalyst and its application in the acetalization of glycerol: crystal
gies for biodiesel production: an environmental and techno-economic eva- structure of -5-hydroxy-2-phenyl-1,3-dioxane trimer. Chem Papers
luation. Renew Energy 2014;69:479–87. 2012;66:1048–58.
[7] Coleman T, Blankenship A. Process for the preparation of glycerol formal. US [39] Pierpont AW, Batista ER, Martin RL, Chen W, Kim JK, Hoyt CB, et al. Origins of
Patent 0094027; 2010. the regioselectivity in the lutetium triflate catalyzed ketalization of acetone
[8] Clomburg JM, Gonzalez R. Anaerobic fermentation of glycerol: a platform for with glycerol: a DFT study. ACS Catal 2015;5:1013–9.
renewable fuels and chemicals. Trends Biotechnol 2013;31:20–8. [40] Smith BM, Kubczyk TM, Graham AE. Metal triflate catalysed acetal ex-
[9] Nanda MR, Yuan Z, Qin W, ghaziaskar HR, Poirer MA, Xu CC. Thermodynamic change reactions of glycerol under solvent-free conditions. RSC Adv
and kinetic studies of a catalytic process to convert glycerol into solketal as an 2012;2:2702–6.
oxygenated fuel additive. Fuel 2014;117:470–7. [41] Manjunathan P, Maradur SP, Halgeri AB, Shanbhag V. Room temperature
[10] Delfort B, Duran I, Jaecker A, Lacome T, Montagne X, Paille F. Diesel fuel synthesis of solketal from acetalization of glycerol with acetone: effect of
compounds containing glycerol acetal. US Patent 890364; 2005. cristallite size and the role of acidity of beta zeolite. J Mol Catal A: Chem
[11] Garcıa E, Laca M, Pe´rez P, Garrido A, Peinado J. New class of acetal derived 2015;396:47–54.
from glycerin as a biodiesel fuel component. Energy Fuels 2008;22:4271–80. [42] Wang B, Shen Y, Sun J, Xu F, Sun R. Conversion of platform chemical glycerol to
[12] Mota CJA, da Silva CXA, RosenbachJr N, Costa J, da Silva F. Glycerin derivatives cyclic acetals promoted by acidic ion liquids. R Soc Chem Adv 2014;4:18917–23.
as fuel additives: the addition of glycerol/acetone ketal (solketal) in gasolines. [43] Narkhede N, Patel A. room temperature acetalization of glycerol to cyclic
Energy Fuels 2010;24(4):2733–6. acetals over anchored silicotungstates under solvent free conditions. R Soc
[13] Dhepe PL, Sahu R. A solid-acid-based process for the conversion of hemi- Chem Adv 2014;4:19294–301.
cellulose. Green Chem 2010;12:2153–6. [44] Crotti C, Farnetti E, Guidolin N. Alternative intermediates for glycerol valor-
[14] Serafim H, Fonseca IM, Ramos AM, Vital J, Castanheiro JE. Valorization of ization:iridium-catalyzed formation of acetals and ketals. Green Chem
glycerol into fuel additives over zeolites as catalysts. Chem Eng J 2010;12:2225–31.
2011;178:291–6. [45] Souza TE, Portilho MF, PMTG Souza, Souza PP, Oliveira LCA. Modified niobium
[15] Hendrickson JB, Cram DJ, Hammond GS. Chimie Organica Bucuresti: Editura oxyhydroxide catalyst: an acetalization reaction to produce bio-additives for
stiintifica si enciclopedica; 1976. sustainable use of waste glycerol. Chem Cat Chem 2014;6:2961–9.
[16] Bruchmann B, Hiiberle K, Gruner H, Hirn M. US preparation of cyclic acetals or [46] Souza TE, Padula ID, Teodoro MMG, Chagas P, Resende JM, Souza PP, Luiz CA,
ketals, 5917059; 1999. Oliveira LCA. Amphiphilic property of niobium oxyhydroxide for waste gly-
[17] Bărbulescu E, Bărbulescu N, Greff C. Reactii ale compusilor organici. Bucuresti: cerol conversion to produce solketal. Catal Today 2015;254:83–9.
Edituratehnica; 1972. [47] Nair GS, Adrijanto E, Alsalme A, Kozhevnikov IV, Cooke DJ, Brown DR, Shiju NR.
[18] Ruiz VE, Velty A, Santos LL, Perez LA, Sabater MJ, Iborra S, et al. Gold Catalysts Glycerol utilization over niobia catalysts. Catal Sci Technol 2012;2:1173–9.
and solid catalysts for biomass transformations: valorization of glycerol and [48] Zhang S, Zhao Z, Ao Y. Design of highly efficient Zn-, Cu-, Ni- and Co-promoted
glycerol – water mixtures through formation of cyclic acetals. J Catal M-AlPO4solid acids: the acetalization of glycerol with acetone. Appl Catal A:
2010;271:351–7. Gen 2015;496:32–9.
[19] Sato S, Araujo AS, Miyano M. Process for the production of glycerol acetals. US [49] Prakruthi HR, Chandrashekara BM, Prakash BSJ, Bhat YS. Microwave rehy-
Patent 0207927; 2008. drated Mg-Al-LDH as base catalyst for the acetalization of glycerol. Catal Sci
[20] Suriyaprapadilok N, Kitiyanan B. Synthesis of Solketal from glycerol and its Technol 2015;5:3667–74.
reaction with benzyl alcohol. Energy Proced 2011;9:63–9. [50] Agirre I, Garcia I, Requies J, Barrio VL, Guemez MB, Cambra JF, et al. Glycerol
[21] Deutsch J, Martin A, Lieske H. Investigation on heterogeneously catalysed acetals, kinetic study of the reaction between glycerol and formaldehyde.
condensation of glycerol to cyclic acetals. J Catal 2007;245:428–35. Biomass Bioenergy 2011;35:3636–42.
814 A.R. Trifoi et al. / Renewable and Sustainable Energy Reviews 62 (2016) 804–814

[51] Agirre I, Guemez MB, Uguarte A, Requies J, Barrio VL, Cambra JF, et al. Glycerol [60] Hong X, McGiveron O, Kolah AK, Orjuela A, Peereboom L, Lira CT, et al.
acetals as diesel additives: kinetic study of the reaction between glycerol and Reaction kinetics of glycerol acetal formation via transacetalization with 1,1-
acetaldehyde. Fuel Process Technol 2013;116:182–8. diethoxyethane. Chem Eng J 2013;222:374–81.
[52] Wegenhart BL, Liu S, Thom M, Stanley D, Abu-Omar MM. Solvent-free meth- [61] Monbaliu JCM, Winter M, Chevalier B, Schmidt F, Jiang Y, Hoogendoorn R,
ods for making acetals derived from glycerol and furfural and their use as a et al. Effective production of the biodiesel additive STBE by a continuous flow
biodiesel fuel component. Am Chem Soc Catal 2012;2:2524–30. process. Bioresour Technol 2011;102:9304–7.
[53] Guemez MB, Requies J, Agirre I, Arias PL, Bario L, Cambra JF. Acetalization [62] Clarkson JS, Walker AJ, Wood MA. Continuous reactor technology for ketal
reaction between glycerol and n-butyraldehyde using an acidic ion exchange formation: an improved synthesis of solketal. Org Process Res Dev
resin. Kinet Model Chem Engi J 2013;228:300–7. 2001;5:630–5.
[54] Faria RPV, Pereira CSM, Siva V MTM, Loureiro JM, Rodrigues AE. Glycerol [63] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier MA, Xu C. A new
valorization as biofuels: selection of a suitable solvent for an innovative pro- continuous-flow process for catalytic conversion of glycerol to oxygenated
cess for the synthesis of GEA. Chem Eng J 2013;233:159–67. fuel additive: catalyst screening. Appl Energy 2014;123:75–81.
[55] Da Silva MJ, De Oliveira GM, Julio AA. A highly regioselective and solvent-free [64] Hasabnis A, Mahajani S. Acetalization of glycerol with formaldehyde by
Sn(II)-catalyzed glycerol ketals synthesis at room temperature. Catal Lett 2015. reactive distillation. Ind Eng Chem Res 2014;53:12279–87.
[56] Da Silva CXA, Mota CJA. The influence of impurities on the acid-catalyzed [65] Pawar RR, Jadhav SV, Bajaj HC. Microwave-assisted rapid valorization of gly-
reaction of glycerol with acetone. Biomass Bioenergy 2011;35:3547–51. cerol towards acetals and ketals. Chem Eng J 2014;235:61–6.
[57] Pathak K, Reddy KM, Bakhshi NN, Dalai AK. Catalytic conversion of glycerol to [66] Roldan L, Mallada R, Fraile JM, Mayora JA, Menendez M. Glycerol upgrading by
value added liquid products. Appl Catal A: Gen 2010;372:224–38. ketalization in a zeolite membrane reactor. Asia Pac J Chem Eng 2009;4
[58] Shirani M, Ghaziaskar HS, Xu CC. Optimization of glycerol ketalization to (3):279–84.
produce solketal as biodiesel additive in a continuous reactor with super- [67] Perreia CSM, Rodrigues AE. Process intensification: new technologie (SMBR
critical acetone using Purolite PD 206 as catalyst. Fuel Process Technol and PermSMBR) for the synthesis of acetals. Catal Today 2013;218-219:
2014;124:206–11. 148–52.
[59] Bratu EA. Operatii unitare in ingineria chimica. . Bucuresti: Editura tehnica;
1984.
 Fuel 181 (2016) 46–54

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Highly selective acetalization of glycerol with acetone to solketal over

acidic carbon-based catalysts from biodiesel waste
Maraísa Gonçalves a,b,⇑, Raphael Rodrigues a, Thalita S. Galhardo a, Wagner A. Carvalho a
a
Centre of Natural and Human Sciences, Federal University of ABC, Santa Adélia Street, 166, Bangu, Santo André, SP 09210-170, Brazil
b
Science and Technology Institute, Federal University of São Paulo, Talim Street, 330, São José dos Campos, SP 12231-280, Brazil

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Green carbon-based solid with high Biodisel production byproduct (glycerol) is transformed in a environmental acid carbon which is used as
catalytic activity in the glycerol catalyst in glycerol acetalization.
acetalization reaction.
 High selectivity of solketal products
of the acetalization reaction.
 Carbon-based acid preparation: an
efficient approach for the valorization
of a biodiesel waste, glycerin.
 High stability of the acid group
surfaces present on the carbon-based
catalyst surface.

a r t i c l e i n f o a b s t r a c t

Article history: The glycerol acetalization reaction in the presence of acetone was evaluated using acidic carbon-based
Received 7 February 2016 catalysts from biodiesel waste as catalysts. Acidic carbon-based catalysts were prepared with different
Received in revised form 24 March 2016 glycerin:sulfuric acid mass ratios at 457 K for 24 h. The amount of acid groups as well as the acidity
Accepted 20 April 2016
(mainly attributed to sulfur groups) enhance with an increase of glycerol:sulfuric acid ratio (1:2 and
Available online 29 April 2016
1:3, named GC-1:2 and GC-1:3, respectively), reaching 0.9 mmol g 1 for sulfur and 3.8 mmol g 1 for total
acidity. Very promising results were obtained when these carbon-based catalysts were tested as catalysts
Keywords:
in glycerol acetalization, reaching about 80% of glycerol conversion and about 95% of solketal selectivity,
Waste recovery
Acidic carbon-based catalysts
this product has been applied as a fuel additive. Furthermore, the carbon-based catalysts surface groups
Fuel additive showed high stability under the reaction conditions tested in this study. These results indicated that
Glycerol conversion acidic carbon-based catalysts derived from glycerin are very attractive catalysts for glycerol acetalization.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Currently, global warming and environmental pollution have

⇑ Corresponding author at: Science and Technology Institute, Federal University led to increased concern among the population about what to do
of São Paulo, Talim Street, 330, São José dos Campos, SP 12231-280, Brazil. to minimize these impacts. Renewable energy has become a
E-mail address: [email protected] (M. Gonçalves). promising fuel to reduce the use of fossil energy. Among the

http://dx.doi.org/10.1016/j.fuel.2016.04.083
0016-2361/Ó 2016 Elsevier Ltd. All rights reserved.
 M. Gonçalves et al. / Fuel 181 (2016) 46–54 47

different renewable energies we highlight biodiesel. The promi- based catalysts from glycerin waste as a catalyst in glycerol acetal-
nent superiority of biodiesel over fossil diesel, regarding health ization, which is the innovation of this work.
and environmental concerns, has attracted its use as an alternative Thus, taking into account these premises, the aim of this work
fuel. Generally, biodiesel is produced from the transesterification was to evaluate the acidic carbon-based catalysts from biodiesel
of vegetable oils with alcohols in the presence of an alkaline- waste as catalyst in glycerol acetalization. The effect of the temper-
based catalyst. This process generates a large amount of glycerin ature, the glycerol:acetone molar ratio, the catalyst amount, as
waste (1:9 glycerin:biodiesel weight ratio). Currently in Brazil, well as the reusability of the acidic carbon-based catalysts, were
the use of 7% biodiesel mixed with fossil diesel is mandatory. Thus also evaluated.
in 2015 the production of glycerin waste was approximately
360,000 m3 [1]. Due to the increase of this waste, glycerol produc-
2. Experimental part
tion has also increased, and its price has reduced drastically. More-
over, glycerin usually contains between 2% and 3% fatty acid
2.1. Preparation of the carbon-based catalysts
residues, 6% and 8% NaCl, 10% and 15% water and 75% and 80%
glycerol. Taking this into account, it is of paramount importance
Biodiesel waste, glycerin, from Oxiteno-Brazil, was used as pre-
to develop technologies able to convert this waste into value added
cursor. The glycerin characterization showed this residue content:
products through different strategies and/or approaches.
water and alcohols = 20 ± 3% and metals, such as K or Na were not
Several reactions for glycerol conversion into value added
detected and the measured pH was 6.5.
products have been reported in the literature, e.g., hydrogenoly-
The acidic carbon-based catalysts (GC) were prepared by
sis [2], acetylation [3], reforming [4], etherification [5,6], among
hydrothermal carbonization of a mixture of glycerin and sulfuric
others [7]. Many of these processes are reported to be favored in
acid (96% v/v, Carlo Erba) at different mass ratios under 423 K for
the presence of Brønsted acid catalysts, especially, the acetaliza-
24 h, in a procedure adapted from previous works of our research
tion reaction for the production of cyclic acetals or ketals [8].
group [6]. The carbon-based catalysts were named GC-X, where
Typically, acetalization of glycerol with either aldehydes or
X is the glycerin:acid mass ratio (mglycerin:mH2SO4) used: 3:1, 2:1,
ketones is carried out using a homogeneous acid catalyst and
1:1, 1:2 and 1:3. After carbonization the carbon-based catalysts
mineral acids, such as HCl, H2SO4 and HF, is extensively
were repeatedly washed in a Soxhlet system with distilled water
employed [9,10]. Nevertheless, lately several studies have been
until neutral pH. Indeed, the evidence of washing for soluble acid
published describing many suitable heterogeneous catalysts,
removal was provided using a turbidity test with BaCl2. After
including Amberlyst resins [11], mixed oxides [12], niobia [13],
washing, all carbon-based catalysts were oven-dried at 393 K for
zirconia [9] and functionalized activated carbons [8,14], aiming
24 h.
at sustainability and an environmentally friendly process. As
described in the literature, the products obtained from the
acetalization of glycerol with acetone are mainly 2,2-dimethyl- 2.2. Carbon-based catalyst characterization
1,3-dioxolane-4-methanol, well known as solketal, and a six
member ring (6MR) 2,2-dimetyl-[1,3]dioxan-5-ol. Solketal has Textural properties of the carbon-based catalysts were deter-
been applied as a solvent, low temperature heat-transfer fluid, mined by nitrogen adsorption measurements at 77 K in an
flavoring agent, surfactant and fuel additive [14,15]. Autosorb-1MP device (Quantachrome Instruments). Prior to the
Another important route for the conversion of biodiesel wastes measurements, the samples were outgassed at 400 K and
would be their transformation into acidic carbon-based catalysts, a 1.3  10 3 Pa for 4 h, to remove moisture. The surface chemistry
new catalyst with high added value. There are several research of the solids was studied by Fourier transform infrared spec-
papers about the good catalytic activity and high surface group sta- troscopy (FTIR) analysis using a Varian 3100 FT-IR Spectrometer.
bility of different acidic carbon-based catalysts obtained from The analyses were performed mixing dried samples with potas-
waste materials and/or sugars. Janaun and Ellis obtained good sium bromide (KBr) in 1:100 weight ratio and ground into fine
results with carbonized sucrose as a catalyst in glycerol etherifica- powder. This mixture was dried at 373 K for 24 h and thin pellets
tion [16]. Sanchez and co-workers, using a sulfonated carbon- were made in manual equipment. The spectra were then acquired
based catalysts prepared from sucrose carbonization as catalyst at this temperature by accumulating 100 scans at 4 cm 1 resolu-
in glycerol etherification, obtained a glycerol conversion as high tion in the range of 400–4000 cm 1. The amount of sulfur on the
as 99% [17]. Khayoon and Hameed reported high catalytic activity carbon-based catalysts surface was determined by Energy Disper-
for sulfonated activated carbon in glycerol etherification, reaching sive X-ray Spectrometry (EDS). The analysis was performed on a
ca. 91% of glycerol conversion [14]. Liu and co-workers prepared JEOL JSM-6701F field emission scanning electron microscope oper-
one sulfonated activated carbon that showed high catalytic activity ating at 10.0 kV and 10.0 mA.
for the etherification of aliphatic acids [18]. Tao and co-workers For the evaluation of density of acidic groups on the solid sur-
reported the use of a carbon obtained from biomass as catalyst faces, acid-base titration was performed, as related by Boehm
with high activity in glycerol esterification [19]. Furthermore, our [23]. For the test, 0.3 g of each carbon-based catalyst sample were
research group has already demonstrated that sulfonated carbons added to 25 mL of previously prepared basic solutions (NaOH and
from agro-industry residues are promising catalysts for the glyc- NaHCO3 0.1 mol L 1, for determination of total acidity and
erol etherification and esterification reactions [20]. Brønsted acid). Sulfonic groups typically are stronger acids than
On the other hand, there are few research works in the litera- the corresponding carboxylic acids, thus are also neutralized with
ture describing the use of carbon-based catalysts from glycerin NaHCO3 as Brønsted acid (ACOOH and ASO3H) [24]. Afterwards,
waste as a catalyst. Prabhavathi Devi and co-workers prepared the solutions containing the carbon-based catalysts were stirred
an acidic carbon-based catalyst from glycerol pitch using 4:1 and for 24 h, at room temperature, and filtered prior to titration with
3:1 acid:glycerol ratios which showed good catalytic activity in HCl (Vetec) 0.1 mol L 1 in an automatic titrator (Metrohm 905
the conversion of several organic molecules and in biodiesel prepa- Titrando).
ration [21,22]. Gonçalves and co-workers reported the use of an Carbon, oxygen and sulfur content were also analyzed in an ele-
acidic carbon-based catalyst from glycerin waste as catalyst in mental analyzer (Thermo Finnigan FLASH EA 1112). The surface
glycerol etherification and obtained good results [6]. Moreover, stability of the carbon-based catalysts was verified by Thermo-
there are no reports in the literature on the use of acidic carbon- gravimetric analysis (TGA) in a Q500 TGA device, TA Instruments
48 M. Gonçalves et al. / Fuel 181 (2016) 46–54

out under N2 atmosphere (gas flow of 50 mL min 1) with a heating 3. Results and discussion
rate of 10 K min 1 in the range 300–1023 K. The ash content of the
carbon-based catalysts was determined by TGA analysis under oxi- 3.1. Characterization of carbon-based catalysts
dation atmosphere (synthetic air flow of 50 mL min 1) at
30 K min 1 from 298 K to 1023 K and maintained at this tempera- The BET surface areas of all carbon-based catalysts, determined
ture for 10 min. The turnover number (TON) was calculated based by N2 adsorption, were lower than 10 m2 g 1. These results are
on the quantity of Brønsted acid (ACOOH and ASO3H) surface related to the methodology used in preparation of these materials.
groups determined by acid-base titration (TON = mols of prod- After the hydrothermal carbonization, these carbon-based cata-
ucts/mols of ACOOH and ASO3H). lysts were not subjected to any activation process, required to
develop a high surface area.
Another important characteristic of the carbon-based catalyst
2.3. Catalytic activity of carbon-based catalysts from glycerin is that it is a metal-free material, as can be confirmed
by the very low ash content, <0.3 wt%, that was obtained by TGA
The activities of all carbon-based catalysts were evaluated in analysis in air atmosphere. This result suggests that there is no
the glycerol (99% pure reagent from Sigma) acetalization in pres- metal present in the composition of these carbon-based catalysts.
ence of acetone (99%, Sigma). In a typical run, approximately Similar results were observed by Ribeiro and co-workers when
0.93 g of glycerol (99%, Sigma), 2.3 g of acetone and 0.13 g of 1,4- analyzing a carbon from biodiesel waste [25].
dioxane (as GC internal standard) were placed in a 10 mL glass vial The surface group amounts and stability are essential for this
containing 3% of catalyst (wt% of glycerol). The mixture was con- type of carbon-based catalyst application in catalysis. In order to
tinuously stirred (600 rpm) for 4 h at room temperature, as confirm the carbon surface group stabilities, another set of exper-
described in previous work [8]. Extra experiments were carried iments was performed under nitrogen atmosphere (Fig. 1).
out using different parameters, such as temperature (313 and There are several studies in the literature about specific surface
338 K), catalyst amounts (0.75 wt%, 1.5 wt% and 5 wt%) and glyc- group decomposition temperatures, although these publications
erol:acetone molar ratios (1:1; 1:2 and 1:3). are often controversial regarding the accuracy of the decomposi-
The reaction products quantification was performed by Gas tion temperature. However, some general trends about the decom-
Chromatography (Agilent 7890A, FID, DB-Wax 30 m  0.25 mm  position temperature ranges have been established. There is a peak
0.25 lm) by the internal standard method (1,4-dioxane). The resulting from carboxylic and sulfonic group decomposition at low
repeatability of the analyses was ±5% for GC/FID. The mass balance temperatures and, at higher temperatures, peaks from lactone,
of the reactions analyzed presented values between 95% and 108%. phenol, ether and carbonyl group decomposition [26].
For all carbons a small weight loss was observed at about 373 K,
assigned to moisture. As observed, the largest amount of moisture
2.4. Surface carbons stability
occurs for carbon-based catalysts prepared in the presence of high
amounts of sulfuric acid (GC-1:1 GC-1:2 and GC-1:3). At the high-
The stability of the GC was investigated by removing the cata-
est temperature the stability of these carbon-based catalysts is
lyst from the reaction medium before completion of the reaction
very similar, with weight loss from 60% to 70%. On the other hand,
and maintaining the filtrate at the reaction temperature in order
the weight loss reaches about 35% in the GC-3:1. Therefore, it can
to test its activity. Afterwards, this solution was used for further
be concluded that these carbon-based catalysts are more stable
glycerol acetalization in order to determine the reactivity of the
than those prepared in the presence of low amounts of sulfuric
possible leached acid groups. The surface groups of GC-1:2 after
acid.
reaction was analyzed by Fourier transform infrared spectroscopy
Meanwhile, the decomposition temperature of the different
(FTIR) analysis using a Varian 3100 FT-IR Spectrometer as
surface groups is unclear in the thermograms obtained. Only a
described above.
slight and continuous variation in weight loss can be seen. To
The catalytic stability was also evaluated by performing consec-
improve the visibility of the decomposition temperature for differ-
utive batch runs under the same operating conditions, in order to
ent surface groups, the derivative thermogravimetric analysis
investigate the reusability of these catalysts. After each reaction
(TGA) was obtained for GC-1:3 (Fig. 1(b)).
run, the catalyst was removed from the reaction solution, repeat-
As can be seen from results shown in Fig. 1, the carbon-based
edly washed with alcohol to remove the adsorbed reactants from
catalyst exhibits one peak at 500 K, characteristic of ASO3H and
the surface and dried at 383 K for 12 h.

100 100 0.5

90 (a) 90
(b)
80 80 0.4
Deriv. weight (%/ K)

70
Weight loss (%)

70
Weight loss (%)

GC-3:1
60 GC-2:1 60 0.3

50 GC-1:1 50
GC-1:2
40 40 0.2
GC-1:3
30 30

20 20 0.1

10 10

0 0 0.0
300 350 400 450 500 550 600 650 700 750 800 300 350 400 450 500 550 600 650 700 750 800
Temperature (K) Temperature (K)

Fig. 1. Thermogravimetric analysis under nitrogen atmosphere of the carbons-based (a) and derivative thermogravimetric analysis of GC-1:3 sample (b).
 M. Gonçalves et al. / Fuel 181 (2016) 46–54 49

ACOOH decomposition groups [27,28]. Another large peak At this point it is noteworthy to mention that sulfur was not
between 550 and 730 K can be attributed to lactones and phenol detected only on the surface of the GC-3:1 sample. For the other
groups. In addition, there is one peak at 723–800 K attributed to carbons, the amount of sulfur determined by EDS analysis
carbonyl group decomposition [27]. increases with increasing percentage of acid used in the hydrother-
As demonstrated from results obtained by scanning electron mal carbonization process, ranging from the 0.24–0.90 mmol g 1.
microscopy coupled with Energy Dispersive X-ray (SEM-EDS) anal- It is assumed that the glycerin carbonization with sulfuric acid
ysis (Fig. 2), the carbon, oxygen and sulfur are the only atoms pre- incorporates sulfur on the surface of these carbons, thus yielding
sent on the surface for GC-1:2. Furthermore, the EDS analysis ACASO3H and ACAOASO3H groups. Likewise, the total sulfur
suggests that these elements are well dispersed on the surface of obtained by elemental analysis (CHNS) presented the same behav-
this carbon-based catalyst. ior. The amount of total sulfur found was 0.6%; 0.9%; 1.7%; 2.9% and
The amount and nature of acid sites becomes very important in 2.9% for GC-3:1; GC-2:1; GC1:1; GC1:2 and GC-1:3, respectively.
acetalization reactions. Among them, Brønsted acid sites, mainly These results are very similar to those related in the literature
the sulfonic groups, can play a crucial role as active sites for the for other acidic carbons-based catalysts [6,29]. Taking this into
aforementioned reaction. In this sense, Fig. 3(a) reports the total account, we can consider that the carbon-based catalysts prepared
acidity, the amount of carboxylic and sulfonic groups and the sul- in presence of higher amount of sulfuric acid are more hydrophilic,
fur content for all carbon-based catalysts. which may be related to the increased polarity of their surface due

Fig. 2. Micrographic (a), energy dispersive spectroscopy (b), carbon (c), oxygen (d) and sulfur (e) patterns for GC-1:2.

4.0 4.0

(a) (b)
3.5 3.5
GC-3:1
Surface groups (mmol g-1)

3.0
Transmitance (a.u.)

3.0
Total acidity (mmol g-1)

total acidity GC-2:1

2.5 2.5
-COOH + -SO3 H
sulfur
2.0 2.0
GC-1:1
1.5 1.5
GC-1:2
1.0 1.0 GC-1:3

0.5 0.5
-COOH -C=C- -SO3H

0.0 0.0 2000 1800 1600 1400 1200 1000 800 600

GC-3:1 GC-2:1 GC-1:1 GC-1:2 GC-1:3 Wavenumber (cm-1)

Fig. 3. (a) Total acidity, carboxylic and sulfonic groups and sulfur groups presents on the surface for all carbons and (b) infrared spectrum for all carbon-based catalysts.
50 M. Gonçalves et al. / Fuel 181 (2016) 46–54

to the incorporation of more oxygen and sulfur containing func- of ASO3H groups were observed [31]. Another important charac-
tional groups. This information is corroborated by the FTIR analy- teristic that should be mentioned regarding the carbons is the
sis, elemental analysis and titration (results showed in Fig. 3). presence of sulfur (ASO3H groups), whose amount increased when
As can be seen, the amount of total acid surface groups is higher increasing the acid proportion. Consequently, for the GC-3:1 these
for GC-1:3 compared to the GC-3:1 carbon-based catalysts. How- bands are not significant, corroborating the EDS results, in which it
ever, this difference is not as pronounced as that observed for was not possible to detect surface sulfur.
the amount of strong acid groups. The amount of total acidity for In summary, these results show that it is possible to success-
GC-1:3 is about 32% higher than that found for GC-3:1, whereas fully produce acidic carbon-based catalysts from biodiesel waste
for the stronger acid (carboxylic and sulfonic groups) this differ- by hydrothermal synthesis utilizing a low temperature, 423 K. This
ence was as high as 80%. It seems that the excess of sulfuric acid is an interesting finding since the acid groups on the surface of
used during the simultaneous carbonization and polymerization these carbon-based solid can act in many catalytic reactions, such
processes has an important effect on the final surface chemistry as glycerol acetalization. Thus, here we have an interesting way to
of the synthesized carbon-based catalysts, due to the insertion of add value to biodiesel waste.
sulfur groups on the carbon surface [30]. Consequently, it increases
the surface acidity of these catalysts, which may make them 3.2. Catalytic tests
promising catalysts for glycerol conversion reactions.
At this point it is noteworthy that the amount of acidic surface Both nature and the concentration of acid sites are crucial in the
groups for the carbon-based catalysts, mainly, GC-1:2 and GC-1:3, glycerol conversion process via acetalization with acetone. Typi-
is similar or even higher than that reported in the literature for cally, this reaction takes place over Brønsted acid sites and the pop-
other such catalysts. An acid carbon prepared by Khayoon and ulation of these sites seems to play a substantial role in the solketal
Hameed exhibits about 0.9 mmol g 1 of acid density [14]. Galhardo production [9,32]. Therefore, the carbon-based catalysts previously
and co-workers prepared a carbon from rice husk with acidity presented in this work are seen as promising materials for catalyz-
varying from 1.5 to 5.8 mmol g 1 [20]. ing the investigated process, based on their acidic properties.
In order to identify the kind of acid surface groups present, Fig. 4 presents the conversion of glycerol in function of time
infrared analysis (FTIR) was performed for all carbon-based cata- as well as the selectivity to solketal and 6MR in the reactions
lysts and the results are shown in Fig. 3(b). catalyzed by GC-2:1, GC-1:1, GC-1:2 and GC-1:3.
As can be seen from the FTIR spectrum, all carbon-based cata- Likewise, the behavior of GC-3:1 was evaluated and the result
lysts showed a band at 1580 cm 1 that can be attributed to AC@CA showed negligible activity (less that 5% glycerol conversion), which
stretching, related to bonds formed in the carbonization process. is most likely related to both low ACOOH concentration and the
Another band at 1690 cm 1 is characteristic of carboxylic group absence of ASO3H groups. On the other hand, the carbons contain-
(ACOOH) stretching modes [30]. Furthermore, characteristic bands ing sulfonic groups over the surface displayed high catalytic activ-
of asymmetric (1030 cm 1) and symmetric (1180 cm 1) stretching ity, reaching up to 82% of glycerol conversion and almost full

100 100 100 100

(a) (b)
90 90 90 90

80 80
Glycerol conversion (%)

80 80
Glycerol conversion (%)

70 70 70 70

Selectivity (%)
Selectivity (%)

60 60 60 60

50 50 50 50

40 40 40 40

30 30 30 30

20 20 20 20

10 10 10 10

0 0 0 0
0 25 50 75 100 125 150 175 200 225 250 0 25 50 75 100 125 150 175 200 225 250

Time (min) Time (min)

100 100 100 100

(c) (d)
90 90 90 90
Glycerol conversion (%)
Glycerol conversion (%)

80 80 80 80

70 70 70 70
Selectivity (%)
Selectivity (%)

60 60 60 60

50 50 50 50

40 40 40 40

30 30 30 30

20 20 20 20

10 10 10 10

0 0 0 0
0 25 50 75 100 125 150 175 200 225 250 0 25 50 75 100 125 150 175 200 225 250
Time (min) Time (min)

Fig. 4. Glycerol conversion and selectivity for acetalization products (closed symbol d: solketal and open symbol s: 6MR) in the presence of carbons: (a) GC-2:1; (b) GC-1:1;
(c) GC-1:2 and (d) GC-1:3. Reaction conditions: Glycerol:acetone molar ratio 1:4; 3% of catalyst and room temperature.
 M. Gonçalves et al. / Fuel 181 (2016) 46–54 51

selectivity to the product of interest (solketal). As expected, the selectivity results presented in the present work. This compound
concentration of the strongest acid sites is highly related to the car- always has a methyl group in the axial position, resulting in a steric
bon activity throughout the reaction. In this sense, GC-2:1 was hindrance with the axial hydrogen from the ring. In the case of the
somewhat active at the beginning of the reaction and the equilib- five membered ring, solketal, the hindrance is lower and over-
rium state is never reached after 250 min. GC-1:1 followed the comes repulsion of the eclipsed bonds. These isomer energies were
same trend, even though equilibrium is noted upon 180 min of calculated by Ozorio and co-workers and indicated that solketal is
reaction. Both catalysts display reasonable activity at the end of 1.7 kcal mol 1 more stable than the six membered ring (6MR) [36].
the evaluation time with glycerol conversion of 62% and 69%, Regarding the selectivity results, the mechanism and the ther-
respectively. modynamic stability of the products, it is expected that the 6MR
On the other hand, the catalysts containing a higher ASO3H isomerizes to solketal throughout the reaction and the population
population (GC-1:2 and GC-1:3) showed catalytic activity as high of available acid sites seems to play an important role in this
as 60% and 40% of glycerol conversion after 20 min of reaction, process.
respectively, and the equilibrium is achieved in 60 min with glyc- Moreover, the textural properties seem to be irrelevant to the
erol conversion of approximately 82%. In order to confirm the real process, since all acid carbon-based catalysts present a negligible
role of the sulfur groups in the investigated process, a single run surface area (<10 m2 g 1), as well as porosity. Furthermore, the
was performed using H2SO4 in a homogeneous catalytic system. turnover number (TON) was calculated based on the quantity of
The amount of catalyst was defined based on both number of carboxylic and sulfonic groups. The results found were 380 and
sulfur-containing groups identified over GC-1:2 and mass of cata- 300 for GC-1:2 and GC-1:3, respectively. Thus, GC-1:2 was used
lyst weighed in a typical catalytic run. In addition, for a homoge- as the catalyst for the investigation of the effect of glycerol:acetone
neous system, using sulfuric acids as catalyst, 72% glycerol molar ratio, reaction temperature, catalyst amount, as well as the
conversion was obtained after 30 min of reaction, evidence thus, catalyst surface stability. The results for the glycerol:acetone molar
the high catalytic activity of the carbon-based catalysts. ratio and catalyst amount used in the glycerol acetalization are
Therefore, it is noteworthy that the concentration of sulfonic presented in Fig. 6.
groups plays a major role in the acetalization of glycerol with ace- Fig. 6(a) presents the effect of excess acetone in the reaction
tone, as can be observed from the results shown in Fig. 5. Indeed, medium. As widely described in the literature [8,37], the decrease
there was no meaningful variation in the amount of carboxylic in the glycerol:acetone molar ratio provided a relevant increase in
groups in GC-2:1, GC-1:1, GC-1:2 and GC-1:3 (see Fig. 3(a)), the glycerol conversion reaching up to 76% for the 1:4 M ratio.
whereas, the ASO3H concentration gradually increased. Although Based on the proposed reaction mechanism, and disregarding
both ACOOH and ASO3H sites are highly reactive, the acidity of any competition between glycerol and acetone for the active sites,
the former is remarkably lower based on pKa, [33] which could jus- it is expected that the excess acetone shifts the equilibrium toward
tify the activity trend highlighted in Fig. 5. The small difference the products and facilitates their chemisorption over the active
between the amount of sulfur in carbon-based catalysts GC-1:2 centers of the catalysts, which is the first step of the catalytic cycle,
and GC-1:3 can be explained by the maximum glycerol conversion justifying the rise in glycerol conversion levels observed in this
be reached for CG-1:2. work.
The overall mechanism accepted for this reaction corresponds It can be noted in Fig. 6(b) that the increase in the glycerol con-
to the classic acid catalyzed mechanism of ketal formation. Ini- version is proportional to the amount of catalyst due to the obvious
tially, the acetone is protonated by the catalyst and a nucleophilic increase of the total available acid sites. However, no further cat-
attack from the primary glycerol hydroxyl group to the chemi- alytic activity gain was observed with a catalyst amount higher
sorbed acetone takes place yielding a hemiketal. The hemiketal is than 3%, which is expected since the equilibrium conversion is
dehydrated to a tertiary carbenium ion, which is stabilized by res- reached, at the same conditions, in the process containing 3% GC-
onance with the non-bonded electron pairs of the adjacent oxygen 1:2 (see Fig. 4(c)), therefore, any shift in the equilibrium with the
atom, and a fast nucleophilic attack from the secondary hydroxyl excess of acid sites in the reaction medium is obviously most
group occurs in order to form the five-membered ketal ring (solke- unlikely.
tal) [34,35]. The six membered ring formation is most likely asso- The effect of the temperature on the glycerol acetalization with
ciated to the nucleophilic attack of the primary hydroxyl group, but acetone catalyzed by CG-1:2 was evaluated. Nanda and co-workers
this occurs much less frequently, which is in agreement with the conducted kinetic and thermodynamic studies and found that glyc-
erol acetalization with acetone is an exothermic process [38].
Herein, it can be observed that the reaction carried out at room
100
temperature is slightly slower (60% glycerol conversion) in com-
parison to the process conducted at 313 K and 338 K (75% glycerol
90
conversion) at 15 min of reaction. On the other hand, there is no
80 meaningful difference in glycerol conversion upon 30 min of reac-
Glicerol conversion (%)

70 tion (about 75% glycerol conversion), indicating that the process

reaches equilibrium at the same level of conversion regardless
60
the temperature and that the exothermic nature of the acetaliza-
50 tion reaction plays no relevant role.
40

30
3.3. Stability of the catalyst
20

10 In order to investigate the stability of the acidic groups on the

0 carbon-based catalyst surfaces, leaching tests were carried out by
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 removing the carbon from the reaction medium after 15 min
Sulfur groups (mmolg )-1 allowing the liquid mixture to react in the absence of the heteroge-
neous catalyst for another 240 min. The results are presented in
Fig. 5. Influence of glycerol conversion in presence of sulfur groups. Fig. 7.
52 M. Gonçalves et al. / Fuel 181 (2016) 46–54

100 100
90
(a) (b)
90

Glycerol conversion (%)

80 80

Glycerol conversion (%)

70 70
60 60
50 50
40 40
30 30
20 20
10 10
0
0
1:1 1:2 1:3 1:4 0.75 1.50 3.00 5.00
Glycerol:Acetone (mol:mol) Catalyst amount (%)

Fig. 6. Glycerol conversion in the presence of GC-1:2 as catalyst at room temperature and 30 min reaction time: glycerol:acetone molar ratio (a) and influence of catalyst
amount (b).

100 100

90 (a) 90 (b)
80 80
Catalyst removal 70 Catalyst removal
70

60
Yield (%)

60
Yield (%)

50 50

40 40

30 30

20 20

10 10

0 0
0 25 50 75 100 125 150 175 200 225 250 0 25 50 75 100 125 150 175 200 225 250

Time (min) Time (min)

100 100
(c) 90
(d)
90
Catalyst removal
Catalyst removal
80 80

70 70
Yiled (%)

60 60
Yield (%)

50 50

40 40

30 30

20 20

10 10

0 0
0 25 50 75 100 125 150 175 200 225 250 0 25 50 75 100 125 150 175 200 225 250

Time (min) Time (min)

Fig. 7. Glycerol acetalization in the presence of GC-2:1 (a), GC-1:1 (b), GC-1:2 (c) and GC-1:3 (d) as a function of time and after catalyst separation from the reaction medium.
Reaction conditions: Glycerol:acetone molar ratio 1:4; 3% of catalyst and room temperature.

It can be observed that, for all carbon-based catalysts, the reac- Furthermore, for the confirmation that the sulfur groups were
tion is interrupted after the removal of the catalysts from the reac- not leached, infrared analysis was performed for GC-1:2 fresh
tion medium, which indicates that there is no appreciable leaching and after reaction (GC-1:2 used). The results are shown in Fig. 8.
of any active groups present over the surface of the solids. This sug- As can be observed from results, similar surface groups are
gests that the reaction is strictly dependent on the presence of the obtained for both catalyst, fresh and used, confirming that the
solid catalyst, indicating the heterogeneous nature of the process. leaching did not occur.
Even if sulfur leaching to the solution had occurred, the soluble Another important feature for heterogeneous catalysts is the
species would not have been active in homogeneous phase since reusability in consecutive reactions. In this direction, a recycling
the conversion of glycerol remains roughly constant in the absence test was performed by removing the catalyst after each catalytic
of the carbons. run and washing the material three times with ethanol in order
 M. Gonçalves et al. / Fuel 181 (2016) 46–54 53

only for glycerol acetalization reactions, but also for other catalytic
processes involving active acidic sites.

GC-1:2 used 4. Conclusion

Transmitance (a.u.)

The acidic carbon-based catalyst showed good activity, glycerol

conversion was fast and reached about 84% after one hour of reac-
tion time. The selectivity for solketal, the product of interest used
GC-1:2 fresh as a fuel bioadditive, was about 90%. The presence of acid groups,
mainly sulfonic groups, was likely the key factor for the improved
catalytic performance.
Additionally, synthesized carbon-based catalysts exhibit a high
stability and a good recycling capability after different reactions.
Consequently, this study shows an environmentally friendly alter-
2000 1800 1600 1400 1200 1000 800 600 native route for the conversion of glycerin, a by-product of biodie-
Wavenumber (cm-1) sel production, into value added materials.

Fig. 8. Infrared spectrum for GC-1:2 fresh and GC-1:2 used (after four hours first Acknowledgment
reaction).

This work was supported by FAPESP (Project No. 2011/22264-4,

2013/21160-6 and 2014/18529-0).
100

90
References
Glycerol conversion (%)

80
[1] ANP (Agência Nacional do Petróleo). A evolução dos biocombustíveis no Brasil;
70
2015. Available from: <http://www.anp.gov.br/?id=470> [cited 2015 Jan. 27].
60 [2] Rodrigues R, Isoda N, Gonçalves M, Figueiredo FCA, Mandelli D, Carvalho WA.
Effect of niobia and alumina as support for Pt catalysts in the hydrogenolysis of
50 glycerol. Chem Eng J 2012;198:457–67.
[3] Kim I, Kim J, Lee D. Sulfonic acid functionalized deoxycellulose catalysts for
40 glycerol acetylation to fuel additives. Appl Catal A 2014;482:31–7.
30 [4] Martínez T LM, Araque M, Centeno MA, Roger AC. Role of ruthenium on the
catalytic properties of CeZr and CeZrCo mixed oxides for glycerol steam
20 reforming reaction toward H2 production. Catal Today 2015;242:80–90.
[5] González MD, Salagre P, Taboada E, Llorca J, Molins E, Cesteros Y. Sulfonic acid-
10 functionalized aerogels as high resistant to deactivation catalysts for the
0 etherification of glycerol with isobutene. Appl Catal B 2013;136:287–93.
1st run 2nd run 3rd run 4th run 5th run [6] Gonçalves M, Mantovani M, Carvalho WA, Rodrigues R, Mandelli D, Silvestre
Albero J. Biodiesel wastes: an abundant and promising source for the
Catalytic runs preparation of acidic catalysts for utilization in etherification reaction. Chem
Eng J 2014;256:468–74.
Fig. 9. Glycerol acetalization promoted by GC-1:2 in successive catalytic runs. [7] Velasquez M, Santamaria A, Batiot-Dupeyrat C. Selective conversion of glycerol
Reaction conditions: Glycerol:acetone molar ratio 1:4; 3% of catalyst, and 120 min to hydroxyacetone in gas phase over La2CuO4 catalyst. Appl Catal B
of reaction time. 2014;160:606–13.
[8] Rodrigues R, Goncalves M, Mandelli D, Pescarmona PP, Carvalho WA. Solvent-
free conversion of glycerol to solketal catalysed by activated carbons
functionalised with acid groups. Catal Sci Technol 2014;4:2293–301.
to remove the remaining reagents or products physisorbed on the [9] Reddy PS, Sudarsanam P, Mallesham B, Raju G, Reddy BM. Acetalisation of
catalyst surface. The obtained material was then dried at 373 K for glycerol with acetone over zirconia and promoted zirconia catalysts under
mild reaction conditions. J Ind Eng Chem 2011;17:377–81.
24 h and finally reused in a new process. The results are shown in
[10] Sudarsanam P, Mallesham B, Prasad AN, Reddy PS, Reddy BM. Synthesis of bio–
Fig. 9. additive fuels from acetalization of glycerol with benzaldehyde over
Although a slight difference in the glycerol conversion is molybdenum promoted green solid acid catalysts. Fuel Process Technol
observed in the successive runs, it is possible to conclude that 2013;106:539–45.
[11] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu C. Catalytic
the activity of all the catalysts is roughly maintained in the five conversion of glycerol to oxygenated fuel additive in a continuous flow
evaluated experiments. The differences between the cycles are reactor: process optimization. Fuel 2014;128:113–9.
probably due to the site blocking, in the initial runs, by adsorbed [12] Zhang S, Zhao Z, Ao Y. Design of highly efficient Zn-, Cu-, Ni- and Co-promoted
M-AlPO4 solid acids: the acetalization of glycerol with acetone. Appl Catal A
substrates and/or products which were not removed during the 2015;496:32–9.
washing step and/or to the deactivation of the strong acid sites [13] Souza TE, Portilho MF, Souza PMTG, Souza PP, Oliveira LCA. Modified niobium
through reaction of ACOOH and ASO3H, present on the carbon- oxyhydroxide catalyst: an acetalization reaction to produce bio-additives for
sustainable use of waste glycerol. ChemCatChem 2014;6:2961–9.
based catalyst surface, with the AOH of the glycerol. Likewise, [14] Khayoon MS, Hameed BH. Acetylation of glycerol to biofuel additives over
Gonçalves and co-workers related an acid carbon obtained from sulfated activated carbon catalyst. Bioresour Technol 2011;102:9229–35.
biodiesel wastes with high stability in the glycerol etherification [15] Silva LN, Gonçalves VLC, Mota CJA. Catalytic acetylation of glycerol with acetic
anhydride. Catal Commun 2010;11:1036–9.
[6,39] and Rodrigues and co-workers ascribe the slow steady deac- [16] Janaun J, Ellis N. Glycerol etherification by tert-butanol catalyzed by
tivation of the functionalized activated carbon to both the deacti- sulfonated carbon catalyst. Journal of Applied Sciences 2010;10:2633–7.
vation process (as mention in this work) and the remaining [17] Sánchez JA, Hernández DL, Moreno JA, Mondragón F, Fernández JJ. Alternative
carbon based acid catalyst for selective esterification of glycerol to
physisorbed molecules [8]. It must be highlighted that the catalyst
acetylglycerols. Appl Catal A 2011;405:55–60.
activity, after five runs, remains as high as 74%, which is more than [18] Liu X-Y, Huang M, Ma H-L, Zhang Z-Q, Gao J-M, Zhu Y-L, et al. Preparation of a
90% of the initial activity. Similar results are described by Goswami carbon-based solid acid catalyst by sulfonating activated carbon in a chemical
and co-workers for an acid carbon catalyst [40]. reduction process. Molecules 2010;15:7188–96.
[19] Tao M-L, Guan H-Y, Wang X-H, Liu Y-C, Louh R-F. Fabrication of sulfonated
In summary, these advantages make acidic carbon-based carbon catalyst from biomass waste and its use for glycerol esterification. Fuel
catalysts from biodiesel glycerin highly attractive materials, not Process Technol 2015;138:355–60.
54 M. Gonçalves et al. / Fuel 181 (2016) 46–54

[20] Galhardo TS, Simone N, Gonçalves M, Figueiredo FCA, Mandelli D, Carvalho [31] Zhao WYB, Yi C, Lei Z, Xu J. Etherification of glycerol with isobutylene to
WA. Preparation of sulfonated carbons from rice husk and their application in produce oxygenate additive using sulfonated peanut shell catalyst. Ind Eng
catalytic conversion of glycerol. ACS Sustain Chem Eng 2013;1:1381–9. Chem Res 2010;49:12399–404.
[21] Prabhavathi Devi BLA, Vijai Kumar Reddy T, Vijaya Lakshmi K, Prasad RBN. A [32] Manjunathan P, Maradur SP, Halgeri AB, Shanbhag GV. Room temperature
green recyclable SO3H-carbon catalyst derived from glycerol for the synthesis of solketal from acetalization of glycerol with acetone: effect of
production of biodiesel from FFA-containing karanja (Pongamia glabra) oil in crystallite size and the role of acidity of beta zeolite. J Mol Catal A: Chem
a single step. Bioresour Technol 2014;153:370–3. 2015;396:47–54.
[22] Prabhavathi Devi BLA, Gangadhar KN, Siva Kumar KLN, Shiva Shanker K, [33] Okamura M, Takagaki A, Toda M, Kondo JN, Domen K, Tatsumi T, et al. Acid-
Prasad RBN, Sai Prasad PS. Synthesis of sulfonic acid functionalized carbon catalyzed reactions on flexible polycyclic aromatic carbon in amorphous
catalyst from glycerol pitch and its application for tetrahydropyranyl carbon. Chem Mater 2006;18:3039–45.
protection/deprotection of alcohols and phenols. J Mol Catal A: Chem [34] Silva CXA, Goncalves VLC, Mota CJA. Water-tolerant zeolite catalyst for the
2011;345:96–100. acetalisation of glycerol. Green Chem 2009;11:38–41.
[23] Boehm HP. Some aspects of the surface chemistry of carbon blacks and other [35] Mallesham B, Sudarsanam P, Raju G, Reddy BM. Design of highly efficient Mo
carbons. Carbon 1994;32:759–69. and W-promoted SnO2 solid acids for heterogeneous catalysis: acetalization of
[24] Kirk-Othmer encyclopedia of chemical technology. 5th ed. Wiley-Interscience; bio-glycerol. Green Chem 2013;15:478–89.
2007. [36] Ozorio LP, Pianzolli R, Mota MBS, Mota CJA. Reactivity of glycerol/acetone ketal
[25] Ribeiro RS, Silva AMT, Pinho MT, Figueiredo JL, Faria JL, Gomes HT. (solketal) and glycerol/formaldehyde acetals toward acid-catalyzed
Development of glycerol-based metal-free carbon materials for hydrolysis. J Braz Chem Soc 2012;23:931–7.
environmental catalytic applications. Catal Today 2015;240:61–6. [37] Fan C-N, Xu C-H, Liu C-Q, Huang Z-Y, Liu J-Y, Ye Z-X. Catalytic acetalization of
[26] Figueiredo JL, Pereira MFR, Freitas MMA, Órfão JJM. Modification of the surface biomass glycerol with acetone over TiO2–SiO2 mixed oxides. React Kinet Mech
chemistry of activated carbons. Carbon 1999;37:1379–89. Catal 2012;107:189–202.
[27] Ribeiro RFL, Soares VC, Costa LM, Nascentes CC. Production of activated carbon [38] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu CC. Thermodynamic
from biodiesel solid residues: an alternative for hazardous metal sorption from and kinetic studies of a catalytic process to convert glycerol into solketal as an
aqueous solution. J Environ Manage 2015;162:123–31. oxygenated fuel additive. Fuel 2014;117:470–7.
[28] Rocha RP, Silva AMT, Romero SMM, Pereira MFR, Figueiredo JL. The role of O- and [39] Goncalves M, Castro CS, Oliveira LCA, Carvalho WA. Green acid catalyst
S-containing surface groups on carbon nanotubes for the elimination of organic obtained from industrial wastes for glycerol etherification. Fuel Process
pollutants by catalytic wet air oxidation. Appl Catal B 2014;147:314–21. Technol 2015;138:695–703.
[29] Valle-Vigón P, Sevilla M, Fuertes AB. Sulfonated mesoporous silica-carbon [40] Goswami M, Meena S, Navatha S, Prasanna Rani KN, Pandey A, Sukumaran RK,
composites and their use as solid acid catalysts. Appl Surf Sci 2012; et al. Hydrolysis of biomass using a reusable solid carbon acid catalyst and
261:574–83. fermentation of the catalytic hydrolysate to ethanol. Bioresour Technol
[30] Kang S, Ye J, Zhang Y, Chang J. Preparation of biomass hydrochar derived 2015;188:99–102.
sulfonated catalysts and their catalytic effects for 5-hydroxymethylfurfural
production. RSC Adv 2013;3:7360–6.
 Fuel Processing Technology 167 (2017) 670–673

Contents lists available at ScienceDirect

Fuel Processing Technology

journal homepage: www.elsevier.com/locate/fuproc

Short communication

Highly efficient iron(III) molecular catalysts for solketal production T

a,b a,b a,b a,b
Roberto Esposito , Maria Elena Cucciolito , Angela D'Amora , Rossella Di Guida ,
Fabio Montagnaroa, Francesco Ruffoa,b,⁎
a
Dipartimento di Scienze Chimiche, Università di Napoli Federico II, Complesso Universitario di Monte S. Angelo, via Cintia 21, 80126 Napoli, Italy
b
Consorzio Interuniversitario di Reattività Chimica e Catalisi, via Celso Ulpiani 27, 70126 Bari, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: This communication describes extremely efficient homogeneous iron(III) catalysts for the synthesis of solketal
Iron from glycerol and acetone. The activities are the highest ever reported so far for this type of reaction, with TOFs
Solketal up to 105 h− 1 at negligible catalyst loading.
Glycerol
Acetone

1. Introduction [9,20] and Lewis acids, such as SnCl2 [11] and iridium complexes [7],
either used at considerable concentrations (1%) and in the presence of
It is well acknowledged that the choice of the auxiliary substances is additional solvents, or based on an extremely expensive metal. Very
strategic to reduce the impact of a chemical manufacture [1]. Within recently, a catalyst based on a Brønsted acid ionic liquid has been
this context, the selection of the solvent is of crucial importance be- proposed [21], which combines the benefits of homogenous and het-
cause the traditional VOCs are source of concern and main cause of erogeneous catalysis, thanks to its possible reuse. In this case, the ob-
pollution, due to their volatility, toxicity and flammability. stacle to a large scale application seems to be the high cost of the re-
Recently, the processing of renewable biomass has made available a action medium. Therefore, it seems that a fully convincing catalytic
wide range of alternative solvents, whose evaluation is the subject of system for the production of solketal has not yet been optimized.
important discussion and intensive research [2]. Among them, solketal Recently, our research group became interested towards the appli-
(Scheme 1) attracts increasing interest, especially due to the valoriza- cation of Lewis acids in the convenient conversion of vegetable oils. In
tion of glycerol, the by-product of biodiesel manufacture [3], and its this context, it was successfully developed a family of zinc-based cat-
applicability as fuel additive. alysts for the production of biodiesel [25–27], along with a process
Not surprisingly, new catalytic systems for the sustainable conver- based on tungsten(VI) for the oxidative cleavage of oleic acid [28].
sion of glycerol into solketal are constantly proposed [4–24]. Most of With the aim of further contributing to this flourishing field of re-
them are heterogeneous, and generally imply the use of solid acid search, our attention has been directed towards the production of
catalysts such as Amberlyst [4–6,12,13], heteropolyacids supported on solketal. This activity has resulted in the screening of simple iron(III)
silica [8], Zr- and Sn-mesoporous substituted silicates [10], carbon salts along with a class of new complexes containing pyridin-2-imine
functionalized with Brønsted acid groups [14,18], zeolites [15,24], ligands (Fig. 1).
metal aluminum phosphates M-AlPO4/xAlPO4 (x = Zn, Cu, Ni, or Co) The latter ones were selected aiming to discover any beneficial ef-
[16], mixed Al/Nb oxides [19], and montmorillonite [22]. In these fect arising from the presence of the modular ancillary ligands 1-R.
cases, complete conversion of glycerol is rarely achieved [20], and a Furthermore, unlike many simple salts of iron(III), they are not hy-
substantial amount of solid catalyst is generally required (5–40%). groscopic and can be easily handled in air, which greatly facilitates the
By far less investigated are homogeneous catalysts, despite they experimental manipulations.
commonly warrant mild reaction conditions, high efficiency, better Herein, we communicate preliminary results of our study, that in-
control of the reactive sites, and an easier understanding of the reaction troduce unique novelties in this field of research thanks to the use of
mechanism. In the face of these advantages, the difficult separation simple catalysts based on an economical and non-toxic metal, which
from the product often limits their application. The homogenous cata- promote the formation of solketal with unprecedented TOFs values (up
lysts so far described for the production of solketal are both Brønsted to 105 h− 1). Their extraordinary efficiency ensures expedient reaction

⁎
Corresponding author.
E-mail address: francesco.ruff[email protected] (F. Ruffo).

http://dx.doi.org/10.1016/j.fuproc.2017.08.018
Received 4 May 2017; Received in revised form 14 August 2017; Accepted 14 August 2017
Available online 17 August 2017
0378-3820/ © 2017 Elsevier B.V. All rights reserved.
R. Esposito et al. Fuel Processing Technology 167 (2017) 670–673

Scheme 1. Manufacture of biodiesel and solketal production.

Table 1
Screening of catalystsa.

Entry Catalyst Catalyst, % (mol/mol)b Conversion, %c

1 – – < 20
2 FeCl3 0.050 > 99
3 FeCl3 0.0050 > 99
4 FeCl3 0.0010 93 ± 1
5 FeCl3 0.00050 92 ± 1
6 [FeCl3(1-NO2)] 0.050 > 99
Fig. 1. Catalysts of type [FeCl3(1)]. UV (in acetone, λmax): [FeCl3(1-H)], 360 nm; 7 [FeCl3(1-NO2)] 0.0050 > 99
[FeCl3(1-OMe)], 360 nm; [FeCl3(1-CF3)], 360 nm; [FeCl3(1-NO2)], 363 nm. Imine IR 8 [FeCl3(1-NO2)] 0.0010 94 ± 1
stretching (in nujol): [FeCl3(1-H)], 1626 cm− 1; [FeCl3(1-OMe)], 1623 cm− 1; [FeCl3(1- 9 [FeCl3(1-NO2)] 0.00050 78 ± 1
CF3)], 1635 cm− 1; [FeCl3(1-NO2)], 1634 cm− 1. Anal. Calcd (found): [FeCl3(1-H)] 10 FeCl3·6H2O 0.00050 84 ± 1
(C12H10Cl3N2Fe): C, 41.85 (41.61); H, 2.93 (2.99); N, 8.13 (8.24). [FeCl3(1-OMe)] 11 Fe(ClO4)3 0.00050 87 ± 1
(C13H12Cl3N2OFe): C, 41.70 (41.52); H, 3.23 (3.35); N, 7.48 (7.32). [FeCl3(1-CF3)] 12 [FeCl3(1-H)] 0.00050 45 ± 1
(C13H9Cl3F3N2Fe): C, 37.86 (37.93); H, 2.20 (2.09); N, 6.79 (6.67). [FeCl3(1-NO2)] 13 [FeCl3(1-OMe)] 0.00050 46 ± 1
(C12H9Cl3N3O2Fe): C, 37.01 (37.26); H, 2.33 (2.40); N, 10.79 (10.61). 14 [FeCl3(1-CF3)] 0.00050 74 ± 1

a
Under reflux, 1.5 h, acetone purity grade 99.8% (2.3 g, 40 mmol), glycerol purity
conditions, very low catalyst loading (up to only 10 ppm), and easy grade 99.5% (0.92 g, 10 mmol).
isolation of the product. This set of conditions meets the constraints b
With respect to glycerol.
imposed by green chemistry, that imply use of economical and safe c
Selectivity ≥ 98%.
solvents, the absence of auxiliaries, the application of efficient and
cheap catalysts that do not require a demanding work-up for their se- mol), the same catalysts showed significant activities (entry 5 vs 9).
paration. Noteworthy the average TOFs were meaningful (up to 105 h− 1 for entry
5 case), and much higher than those ever reported for this reaction in
similar experimental conditions: ca. 500 h− 1 for expensive iridium
2. Results and discussion catalysts [7], and ca. 50 h− 1 for SnCl2 and common Brønsted acids,
such as sulfuric or p-toluenesulfonic acid [11]. This evidence also de-
The catalysts [FeCl3(1-R)] [29] were prepared in diethyl ether monstrates that the iron complexes act as true Lewis acids, plausibly
through a simple template procedure [30], which involved the reaction through the mechanism described in literature [20]: the key-step is the
in situ between pyridine-2-carboxaldehyde and the appropriate aniline attack of the alcoholic –CH2OH function of glycerol to the carbonyl
in the presence of FeCl3. Immediate precipitation of the yellow-orange group, activated via coordination to the metal center. This step is fol-
complexes was observed. UV [31] and IR [32] spectroscopy confirmed lowed by cyclization and dehydration with formation of the product.
the presence of the C]N bonds. The same loading (0.00050%) was adopted for comparing the other
A first screening of catalysts was carried out using FeCl3 and selected iron(III) compounds (entries 10–14). The screening of the en-
[FeCl3(1-NO2)] in the range 0.050–0.0010% by moles with respect to tire panel revealed that the simple salts along with [FeCl3(1–CF3)] and
the alcohol (entries 2–4 vs 6–8 of Table 1). [FeCl3(1–NO2)] are considerably active. These results also point out
The catalytic runs were performed under reflux with a acetone/ that the likely mitigation of acidity due to the presence of the nitrogen
glycerol ratio 4/1 for 90 min. The vapors passed through a short donors is largely compensated by the electron-withdrawing sub-
column of 3A molecular sieves before reaching the condenser. In gen- stituents.
eral, the initial reaction mixture is biphasic due to the poor miscibility Therefore, successive experiments (Table 2) were carried out using
of glycerol in acetone. As the reaction proceeds, the system becomes [FeCl3(1–NO2)] because of its excellent activity even at low con-
homogeneous because the presence of solketal improves the mutual centrations and its insensitivity to air and moisture. These further ex-
miscibility of the components. periments were carried out varying the acetone/glycerol ratio while
The mixture was directly analyzed through 1H NMR spectroscopy, keeping the catalyst loading between 0.0010 and 0.0020% mol/mol.
and conversion and selectivity were assessed by integrating proper re- It is clear that the ratio acetone/glycerol plays a fundamental role.
gions (see example of Fig. 2). In fact a high value (6/1, entries 1 and 5) depresses the conversion,
Without any catalyst (entry 1 of Table 1) the conversion was poor probably due to the consequent dilution of the catalyst. Instead a low
and, in fact, the biphasic mixture disclosed the presence of substantial ratio (2/1, entries 4 and 8) does not give satisfactory results, perhaps
amounts of unreacted glycerol. given the lower availability of the reagent. Therefore, the best outcomes
Even upon further reduction of their concentration (0.00050% mol/

671
R. Esposito et al. Fuel Processing Technology 167 (2017) 670–673

4.4 4.3 4.2 4.1 4.0 3.9 3.8 3.7 3.6 3.5 3.4 3.3 3.2 3.1 3.0 2.9 2.8 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3

Fig. 2. 1H NMR spectrum of an exemplificative reaction mixture at high conversion (in D2O).

Table 2 with unprecedented TOFs values. The easy procedure allows the pro-
Optimization of the catalysis with [FeCl3(1-NO2)]a. duction of pure solketal with low iron content. The possibility to use
complexes with modular nitrogen ligands is particularly attractive for
Entry [FeCl3(1-NO2)], % (mol/ Acetone (mmol)/glycerol Conversion, %c
mol)b (mmol) their stability and handling, and also because their tunability makes the
methodology transferable to other carbonyl compounds wherein simple
1 0.0010 6:1 78 ± 1 iron salts may not be soluble. Furthermore, the introduction of chiral
2 0.0010 4:1 94 ± 1
ligands opens towards the enantioselective synthesis of chiral acetals,
3 0.0010 3:1 88 ± 1
4 0.0010 2:1 72 ± 1
which are important synthons for high value-added compounds. All of
5 0.0020 6:1 79 ± 1 these possibilities will be studied in the near future.
6 0.0020 4:1 95 ± 1
7 0.0020 3:1 97 ± 1 Acknowledgment
8 0.0020 2:1 65 ± 1

a
Under reflux, 1.5 h, acetone purity grade 99.8% (2.3 g, 40 mmol), glycerol purity Angela D'Amora thanks MIUR for a scholarship (Finanziamento
grade 99.5%. Progetti Competitivi).
b
With respect to glycerol.
c
Selectivity ≥ 98%. References

were observed with intermediate ratios of 3/1 and 4/1. [1] P.G. Jessop, The use of auxiliary substances (e.g. solvents, separation agents) should
On this basis, the reaction mixture of entry 7 was treated aiming at be made unnecessary wherever possible and innocuous when used, Green Chem. 18
isolating solketal. After the catalytic run, excess acetone was recovered (2016) 2577–2578.
[2] C.M. Alder, J.D. Hayler, R.K. Henderson, A.M. Redman, L. Shukla, L.E. Shuster,
by flash distillation. This simple work-up, which did not require any H.F. Sneddon, Updating and further expanding GSK's solvent sustainability guide,
additional solvents or severe conditions, returned solketal with an iron Green Chem. 18 (2016) 3879–3890.
content below 10 ppm, i.e. a purity grade compatible with bulk appli- [3] C.-H. Zhou, J.N. Beltramini, Y.-X. Fan, G.Q. Lu, Chemoselective catalytic conversion
of glycerol as a biorenewable source to valuable commodity chemicals, Chem. Soc.
cations. Rev. 37 (2008) 527–549.
As final remark, it is interesting to know that a preliminary test also [4] J. Deutsch, A. Martin, H. Lieske, Investigations on heterogeneously catalysed con-
successfully verified the compatibility of the catalysts with crude gly- densations of glycerol to cyclic acetals, J. Catal. 245 (2007) 428–435.
[5] C.X.A. da Silva, V.L.C. Gonçalves, C.J.A. Mota, Water-tolerant zeolite catalyst for
cerol from biodiesel production through our catalytic systems [27]. the acetalisation of glycerol, Green Chem. 11 (2009) 38–41.
Complete conversion of glycerol was obtained in the conditions of entry [6] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martín, Acetalisation of bio-
6 of Table 1, and this result strongly encourages further studies aimed glycerol with acetone to produce solketal over sulfonic mesostructured silicas,
Green Chem. 12 (2010) 899–907.
at bulky applications.
[7] C. Crotti, E. Farnetti, N. Guidolin, Alternative intermediates for glycerol valoriza-
tion: iridium-catalyzed formation of acetals and ketals, Green Chem. 12 (2010)
3. Conclusion 2225–2231.
[8] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Valorisation of
glycerol by condensation with acetone over silica-included heteropolyacids, Appl.
These preliminary results demonstrate that simple iron(III) com- Catal. B Environ. 98 (2010) 94–99.
pounds promote the formation of solketal from acetone and glycerol [9] N. Suriyaprapadilok, B. Kitiyanan, Synthesis of solketal from glycerol and its

672
R. Esposito et al. Fuel Processing Technology 167 (2017) 670–673

reaction with benzyl alcohol, Energy Procedia 9 (2011) 63–69. A. Riisager, E.J. Garcia-Suarez, Brønsted acid ionic liquids (BAILs) as efficient and
[10] L. Li, T.I. Korányi, B.F. Sels, P.P. Pescarmona, Highly-efficient conversion of gly- recyclable catalysts in the conversion of glycerol to solketal at room temperature,
cerol to solketal over heterogeneous Lewis acid catalysts, Green Chem. 14 (2012) ChemistrySelect 1 (2016) 5869–5873.
1611–1619. [22] M.N. Timofeeva, V.N. Panchenko, V.V. Krupskaya, A. Gil, M.A. Vicente, Effect of
[11] F.D.L. Menezes, M.D.O. Guimaraes, M.J. da Silva, Highly selective SnCl2-catalyzed nitric acid modification of montmorillonite clay on synthesis of solketal from gly-
solketal synthesis at room temperature, Ind. Eng. Chem. Res. 52 (2013) cerol and acetone, Catal. Commun. 90 (2017) 65–69.
16709–16713. [23] M.J. da Silva, M.J. de Ávila, A.A. Júlio, SnF2-catalyzed glycerol ketalization: a
[12] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaska, M.-A. Poirier, C.C. Xu, friendly environmentally process to synthesize solketal at room temperature over
Thermodynamic and kinetic studies of a catalytic process to convert glycerol into on solid and reusable Lewis acid, Chem. Eng. J. 307 (2017) 828–835.
solketal as an oxygenated fuel additive, Fuel 117 (2014) 470–477. [24] V. Rossa, Y.S.P. Pessanha, G.C. Díaz, L. Diógenes Tavares Câmara, S.B.C. Pergher,
[13] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.-A. Poirier, C.C. Xu, Catalytic D.A.G. Aranda, Reaction kinetic study of solketal production from glycerol ketali-
conversion of glycerol to oxygenated fuel additive in a continuous flow reactor: zation with acetone, Ind. Eng. Chem. Res. 56 (2017) 479–488.
process optimization, Fuel 128 (2014) 113–119. [25] M. Di Serio, G. Carotenuto, M.E. Cucciolito, M. Lega, F. Ruffo, R. Tesser,
[14] R. Rodrigues, M. Gonçalves, D. Mandelli, P.P. Pescarmona, W.A. Carvalho, Solvent- M. Trifuoggi, Shiff base complexes of zinc(II) as catalysts for biodiesel production,
free conversion of glycerol to solketal catalysed by activated carbons functionalised J. Mol. Catal. A 353–354 (2012) 106–110.
with acid groups, Cat. Sci. Technol. 4 (2014) 2293–2301. [26] V. Benessere, M.E. Cucciolito, G. Dal Poggetto, M. Di Serio, M. López Granados,
[15] P. Manjunathan, S.P. Maradur, A.B. Halgeri, G.V. Shanbhag, Room temperature F. Ruffo, A. Vitagliano, R. Vitiello, Strategies for immobilizing homogeneous zinc
synthesis of solketal from acetalization of glycerol with acetone: effect of crystallite catalysts in biodiesel production, Catal. Commun. 56 (2014) 81–85.
size and the role of acidity of beta zeolite, J. Mol. Catal. A Chem. 396 (2015) 47–54. [27] V. Benessere, M.E. Cucciolito, R. Esposito, M. Lega, R. Turco, F. Ruffo, M. Di Serio,
[16] S. Zhang, Z. Zhao, Y. Ao, Design of highly efficient Zn-, Cu-, Ni- and Co-promoted A novel and robust homogeneous supported catalyst for biodiesel production, Fuel
M-AlPO4 solid acids: the acetalization of glycerol with acetone, Appl. Catal. A Gen. 171 (2016) 1–4.
496 (2015) 32–39. [28] V. Benessere, M.E. Cucciolito, A. De Santis, M. Di Serio, R. Esposito, F. Ruffo,
[17] S. Gadamsetti, N.P. Rajan, G.S. Rao, K.V.R. Chary, Acetalization of glycerol with R. Turco, Sustainable process for production of azelaic acid through oxidative
acetone to bio fuel additives over supported molybdenum phosphate catalysts, J. cleavage of oleic acid, J. Am. Oil Chem. Soc. 92 (2015) 1701–1707.
Mol. Catal. A Chem. 410 (2015) 49–57. [29] K. Marjani, M. Mousavi, D.L. Hughes, Synthesis and crystal structure determination
[18] M. Gonçalves, R. Rodrigues, T.S. Galhardo, W.A. Carvalho, Highly selective acet- of copper(II) and iron(III) complexes of 2-(2-pyridyl)benzothiazole, Transit. Met.
alization of glycerol with acetone to solketal over acidic carbon-based catalysts Chem. 34 (2009) 85–89.
from biodiesel waste, Fuel 181 (2016) 46–54. [30] G. Olivo, O. Lanzalunga, S. Di Stefano, Non-heme imine-based iron complexes as
[19] R. Rodrigues, D. Mandelli, N.S. Gonçalves, P.P. Pescarmona, W.A. Carvalho, catalysts for oxidative processes, Adv. Synth. Catal. 358 (2016) 843–863.
Acetalization of acetone with glycerol catalyzed by niobium-aluminum mixed [31] G. Olivo, M. Nardi, D. Vìdal, A. Barbieri, A. Lapi, L. Gómez, O. Lanzalunga,
oxides synthesized by a sol–gel process, J. Mol. Catal. A Chem. 422 (2016) M. Costas, S. Di Stefano, CeH bond oxidation catalyzed by an imine-based iron
122–130. complex: a mechanistic insight, Inorg. Chem. 54 (2015) 10141–10152.
[20] M.R. Nanda, Y. Zhang, Z. Yuan, W. Qin, H.S. Ghaziaskar, C.C. Xu, Catalytic con- [32] M. Taghi Goldani, A. Mohammadi, R.J. Sandaroos, Green oxidation of alkenes in
version of glycerol for sustainable production of solketal as a fuel additive: a review, ionic liquid solvent by hydrogen peroxide over high performance Fe(III) Schiff base
Renew. Sust. Energ. Rev. 56 (2016) 1022–1031. complexes immobilized on MCM-41, J. Chem. Sci. 126 (2014) 801–805.
[21] Z. Gui, N. Zahrtmann, S. Saravanamurugan, I. Reyero, Z. Qi, M.A. Bañares,

673
 Chemical Engineering Journal 307 (2017) 828–835

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

SnF2-catalyzed glycerol ketalization: A friendly environmentally process

to synthesize solketal at room temperature over on solid and reusable
Lewis acid
Márcio José da Silva ⇑, Fabio de Ávila Rodrigues, Armanda Aparecida Júlio
Chemistry Department, Federal University of Viçosa, Viçosa, Minas Gerais 36590-000, Brazil

h i g h l i g h t s g r a p h i c a l a b s t r a c t

A solid SnF2-catalyzed process

produced solketal from glycerol at HO OH
O HO O
+OH
HO

room temperature. OH

SnF2 catalyst was easily recovered

and reused without loss activity. solid SnF 2 solid SnF2

A process flow diagram with the best OH

- H2 O
HO OH

operating conditions was described. glycerol-propanone solution O

_ HO
O
O

A mechanism for solid SnF2-catalyzed

RT solketal
magnetic stir

solketal synthesis was proposed.

SnF2 operate in biphasic conditions solid SnF2

with low proportion of glycerol to SnF2 catalyst in acon Simulaon of industrial process of SnF2 –
ketone. catalyzed solketal synthesis

a r t i c l e i n f o a b s t r a c t

Article history: Solid SnF2 was able to catalyze the glycerol ketalization with propanone giving solketal (i.e. 2,2-dimethyl-
Received 1 June 2016 1,3-dioxolane-4-methanol) in reactions with high conversions and regioselectivity (ca. 97%). Tin(II) fluo-
Received in revised form 30 August 2016 ride is commercially available, water tolerant and less corrosive solid Lewis acid catalyst. Herein, we have
Accepted 1 September 2016
described an alternative process to the Brønsted acid-catalyzed reactions in homogeneous phase, where
Available online 2 September 2016
the product neutralization steps are avoided, and the catalyst is easily recovered and reused without loss
activity. In addition to develop a catalytic system, the process engineering including the definition of
Keywords:
basic equipment for economic evaluation was too investigated. A process flow diagram (PFD) was devel-
Tin(II) fluoride
Glycerol
oped to generate the most favorable operating conditions for solketal production through SnF2-catalyzed
Solketal glycerol condensation with propanone. We used the UNISimTM software to perform the material and
Bioadditives energy balances. The plant could work with 432 t/y of glycerol and could produce 620.9 t/y of solketal.
Solvent-free process The solketal cost was 12.29 $ (USD) /kg. This analysis suggests that solketal production from glycerol
deserves serious consideration by the investors.
Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction it more competitive in the fuel market [1]. Among the reactions
that convert glycerol to more added value products, deserves be
Glycerin is a feedstock of commercial value in various industrial highlighted its hydrogenolysis, etherification, esterification, dehy-
applications. There is great interest in glycerin use, which would dration, and ketalization, which is the main goal of this work [2–
increase the viability of the biodiesel production process, making 5]. Recently, the conversion of glycerol to acetal or ketal has
attracted significant attention because these compounds have
⇑ Corresponding author at: Chemistry Department, Federal University of Viçosa, potential use as additive of diesel and biodiesel [6,7].
Avenida Peter Henry Rolfs, s/n, Viçosa, Minas Gerais, MG 36570-900, Brazil. Comparatively to the liquid Brønsted acid, Lewis acids Sn(II)
E-mail address: [email protected] (M.J. da Silva). catalysts have noteworthy advantages such as their lower corro-

http://dx.doi.org/10.1016/j.cej.2016.09.002
1385-8947/Ó 2016 Elsevier B.V. All rights reserved.
 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835 829

sion rates, easier handling, higher stability under reaction condi- curve built with the pure glycerol. To do it, when the reactions
tions and higher water tolerance [8]. Moreover, solid SnCl2 is more were stopped, ethanol was added to the solution to adjust the sam-
stable and cheaper than enzymatic catalysts and has been used in ples concentration to the calibration curve.
FFA esterification for biodiesel production [9–13].
There are scarce reports in the literature about the use of tin 2.3. Products identification
compounds in ketalization or esterification reactions of glycerol,
however, recently we have described homogeneous processes In general, reaction products were analyzed on a Shimadzu MS-
where tin halides were used as the catalysts [14,15]. Mainly when QP 2010 ultra mass spectrometer instrument operating on elec-
compared to the metal oxides, metal fluorides play just a minor tronic impact mode (70 eV), coupled with a Shimadzu 2010 GC
role in both homogeneous and heterogeneous catalysis [16]. instrument. Major product was isolated by column chromatogra-
Li et al. have assessed the glycerol ketalization over solid sup- phy using silica gel (60 G). The NMR spectra were obtained on
ported Lewis acid metal catalysts (i.e., M = Hf, Sn and Zr(IV) cations the Mercury-300 Varian Spectrometer 300 MHz for 1H, in CDCl3
metal) [17]. MCM-41 and mesoporous substituted silicates includ- solution using TMS as an internal standard. FT-IR spectroscopy
ing the novel TUD-1 material were the supports employed by them analyses were recorded in Varian 660 FT-IR spectrometer.
on these reactions. Those authors verified that when Lewis acids M
(IV) cations are supported on mesoporous materials they becomes 2.4. Recovery and reuse of SnF2 catalyst
active catalysts and do not suffer leaching, being thus efficiently
reused in consecutive catalytic cycles. Nonetheless, it is notewor- The reaction was carried out at room temperature for 2 h, with
thy that the catalytic runs reported were carried out at 353 K SnF2 catalyst (ca. 0.10 mol%) and 1:8 M ratio of glycerol to propa-
temperature. none. After the end of the reaction, the solution and the solid cat-
Herein, we wish to show a simple and effective solid SnF2- alyst were centrifuged three times with methanol, then taking off
catalyzed glycerol ketalization process at room temperature. We the liquid phase to total removing of unreacted glycerol and the
paid special attention to assessing factors driving the ketalization products. Afterwards, the solid was then dried in an oven at
selectivity and to optimizing the reaction conditions. Remarkably, 373 K, cooled, weighed, and then reused in another catalytic run.
under room conditions the SnF2-catalyzed glycerol condensation The leaching tests were carried out removing the catalyst after
with propanone achieved a high regioselectivity to solketal, and a 30 min of reaction and performing the run without the catalyst by
conversion ranging of 75–95% within a short reaction time (ca. 2 h. The reaction yielding was determined by GC analyses (see
2 h reaction). Section 2.2).
After we have adjusted reaction conditions, we build a Process
Flow Diagram (PFD). The thermal energy required for each piece 2.5. Procedure to determine the amount of soluble Sn in the reaction
of equipment, i.e., heater, reactor, and distillation column, it was solutions
estimated by performing material and energy balances for each
system, using the UNISimTM software [18]. Each piece of equipment The amount of soluble Sn in the reaction solutions before and
was nearly sized for economic analysis. after the catalyst removal was determined by AAS measurements
(Varian Spectra A-200 model, equipped with broker-absorbing
substance) [19]. From the non-polar layer collected after reaction
2. Experimental procedures
end or after recycle/reutilization protocol (see Section 2.4), three
portions were sampled and gently swirled for a complete homog-
2.1. Materials and physical methods
enization and then transferred to a volumetric flask.
The samples were pumped directly into flame: acetylene-
All the chemicals are commercially available. Tin(II) fluoride
nitrous oxide. The parameters used in determining were k
(99% wt., Sigma Aldrich), glycerol (99.5% wt., Vetec) and propanone
(235.5 nm), flame (C2H2/N2O) and detection limit (1 ppm). A cali-
(99% wt., Sigma Aldrich) were used without prior handling. The
bration curve was built based on a stock standard solution of Tin
solvent CH3CN (99% wt., Sigma Aldrich) was used as received.
(Merck, Germany) at a concentration of 1 ppm [19].

2.2. SnF2-catalyzed glycerol ketalization with propanone 2.6. Process simulation

Catalytic tests were carried out in a glass reactor (50 mL) fitted Data obtained in the catalytic tests were used in the simulation
with sampling septum, with a magnetic stirrer at room tempera- of the conditions to manufacturing solketal at industrial scale. The
ture for 2 h. To the mixture of glycerol and propanone in an ade- material and energy balances were solved using UNISimTM software
quate molar ratio (ca. 15 mL), was added the SnF2 solid catalyst [18].
(ca. 0.1–1 mol%). Effect of molar ratio glycerol: ketone was evalu-
ated using 1:1–1:10 proportions. 2.7. Techno-economic analysis
The catalytic runs were monitored by GC analyses (Shimadzu
GC 2010 instrument, FID, fitted with Carbowax 20M capillary col- Each piece of equipment was roughly sized and the approxi-
umn) of aliquots periodically collected at regular time intervals. mate cost determined. The total cost of equipment is then factored
Toluene was the internal standard. to give the estimated capital cost. The flow rates, the amount of
The reaction yielding was calculated from the corresponding raw materials and the utility needed annually was obtained from
chromatographic peak areas in comparison with the corresponding material and energy balances.
standard curve built with pure solketal. The dilution of aliquots The capital needed for the manufacturing and plant facilities
with CH3CN before the GC analyses adjusted the samples concen- was called the fixed-capital investment (FCI), while that necessary
tration to the calibration curve. It is important to note that these for the operation of the plant was termed the working capital
aliquots were collected from less polar phase (i.e. propanone). (WC). The sum of the fixed-capital investment and the working
The conversion of glycerol was too monitored by comparison of capital is known as the total capital investment (TCI). The Cost
area remaining of glycerol GC peak (i.e. in the aliquot under anal- Evaluation & Workbook spreadsheet was used in the economic
ysis) and corresponding GC peak area obtained of the standard evaluation of the process, where the main information to be pro-
830 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835

vided is the cost of the equipment, which allows the calculation of ketone molar proportion of 1:2 to 1:10 did not produce any signif-
the investment cost [20]. icant change on conversion. All the reactions carried out without
the catalyst achieved poor conversion within 2 h reaction (ca. max-
imum 5%) (Fig. 1a). Differently, in the presence of the SnF2 catalyst,
3. Results and discussion
an increase on the proportion of ketone to glycerol resulted in
higher conversions, which ranged from 58 to 97%.
3.1. General aspects
The highest conversion was achieved in the reactions with the
molar ratio equal or >1:8 (ca. 97%) (Fig. 1b). We have found that
Recently, we successfully described the use of tin halides as cat-
under these conditions (ca. 1:8 or 1:10 glycerol to propanone),
alysts in the esterification reactions of free fatty acids or glycerol in
the glycerol solubility is higher than that observed in the reactions
homogeneous phase [11–14]. In these works, SnCl2 was always the
with low proportion of glycerol to propanone (i.e. 1:2 to 1:6 M
most effective catalyst among Sn(II) compounds assessed. Inspired
ratio). However, in all these reactions, the SnF2 catalyst remained
by these findings, we investigated the catalytic activity of Sn(II)
insoluble.
salts in the glycerol ketalization with or without solvent, but in
The excess of propanone was determining to the initial rate of
homogeneous phase [15]. We realized that despites insoluble,
reaction. Indeed, within first 120 min of reaction, an increase of
SnF2 salt could be a potential heterogeneous catalyst for glycerol
ketone concentration resulted in an increase of initial rate. Never-
ketalization.
theless, we can to note that the reactions with glycerol: ketone
Hence, in this work we wish to show the results obtained in the
proportion equal to 1:8 or 1:10, equal conversions near to 97% con-
assessment of the solketal synthesis using SnF2 salt as a heteroge-
version were achieved.
neous catalyst. In addition, we carry out a study on the feasibility
On the other hand, the reactions with a low amount of ketone
and economic viability of this process.
(i.e. glycerol to ketone equal to 1:6 to 1:2), attained maximum con-
version near to 64%. The lowest conversion was obtained using
3.2. Effect of reactants stoichiometry in the SnF2-catalyzed glycerol glycerol and ketone in an equimolar amount (ca. 3%).
ketalization with propanone Two isomers solketal were obtained (i.e. 2,2-dimethyl-1,3-diox
olan-4-yl-methanol and 2,2-dimethyl-1-3-dioxane-5-ol) with
The reactions were performed at room temperature, with a selectivity close to 1:16 in all reactions (Scheme 1). The higher
molar ratio ranging of 1:2 to 1:10 and SnF2 load equal 1 mol%. selectivity to five-member ring can be explicated due to the fact
The literature describes that in excess of ketone (i.e. molar ratio the formation of five-member ring ketal (i.e., (1a) 1,3-dioxolane)
higher than 1:20), and at 333 K temperature, glycerol and propa- is kinetically favored [22].
none are completely miscible [15]. In addition, when there is a sol- Li et al. have assessed the glycerol ketalization with propanone
uble catalyst into the reaction medium, the ionic strength is over Sn(IV)/ MCM-41 (ca. 3 wt% relative to glycerol) using terty-
increased keeping the glycerol soluble in the propanone even in butanol as solvent [17]. After 6 h of reaction at 353 K, with molar
the low molar ratio of reactants [21]. proportion of glycerol to ketone equal to 1:1, the reactions
Herein, the reactions were performed at room temperature, achieved a maximum conversion of 42%. Herein, reactions carried
with a low proportion of ketone to glycerol and in the presence out over a simple and commercially affordable solid catalyst (i.e.,
of a solid catalyst. Consequently, the reaction system remained SnF2), under solvent free conditions, achieved conversions of 62%
biphasic throughout the whole process (i.e. a glycerol phase and after 2 h of reaction.
other propanone/solketal). So, to follow the reaction progress, ali-
quots of phase propanone were collected and the solketal yielding
determined by GC analyses. The solketal selectivity was higher
than 90% in almost reactions. The mass balance showed that glyc- 3.3. Effect of the catalyst load in the SnF2-catalyzed glycerol
erol was exclusively converted to solketal in all reactions with or ketalization with propanone
without the catalyst.
The glycerol ketalization is a reversible reaction, so a ketone When a reaction attains the equilibrium, an increasing on cata-
excess could shift the equilibrium towards products (Fig. 1). How- lyst load should not change the conversion rates. However, our
ever, when the catalyst was absent, the varying of the glycerol: interest was assessing what the minimum time and the lowest

6
100

80

4
conversion (%)

Conversion (%)

60

40
1:2
2 1:4
20 1:6
1:8
1:10
0
0 20 40 60 80 100 120
0 Time (min)
1:2 1:4 1:6 1:8 1:10
glycerol to propanone molar ratio

(a) Blanck-reactions (b) SnF2-catalysed reactions

Fig. 1. Effect of reactants molar ratio on kinetic curves of the glycerol ketalization. Reaction conditions: Glycerol/propanone (ca. 15 mL); SnF2 (1.0 mol%); temperature
(298 K).
 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835 831

H 3C CH 3 H 3C CH 3
OH O
SnF2 O O O O
2 HO OH + + 2 H2O
+
298 K

OH OH

2,2-dimethyl-1,3-dioxolan- 2,2-dimethyl-1,3-
methanol dioxolan-5-ol
(1a) (1b)
94 % 6%

Scheme 1. SnF2-catalyzed solketal synthesis.

load is required for the reaction achieving the maximum conver- 100
sion, regardless the equilibrium is reached. 0.10 mol %
In general, when catalyst load was increased from 0.10 to 0.25 mol %
80 0.5 mol %
1.00 mol%, a strong increase on conversion rate was observed
0.75 mol %
(Table 1). Conversely, the combined selectivity of two solketal iso- 1.00 mol %

conversion (%)
mers remained practically unchanged (ca. 94–97%). 60
The literature has described the use of solid supported Brønsted
acid catalysts in glycerol ketalization. The condensation of glycerol
40
with ketone over silica immobilized Keggin heteropolyacids
achieved conversion close to 90% after an 8 h reaction to 313–
343 K [23]. Amberlyst resins were used as heterogeneous catalysts 20
on solketal synthesis from glycerol and ketone, however, in almost
of times, temperatures higher than 313 K are always required, as
0
well as the removal of water from the reaction equilibrium [24]. 0 15 30 45 60 75 90 105 120
On the other hand, there is scarce data about the Lewis acid
time (min)
solids-catalyzed solketal synthesis. Zirconium sulfate-catalyzed
reactions achieved conversion and yield of 79 and 77%, respec- Fig. 2. Effect of catalyst load on glycerol ketalization with propanone. Reaction
tively, in reactions carried out at 343 K, pressure (600 psi) and conditions: glycerol (21.0 mmol), propanone (168.0 mmol), molar ratio (1:8),
WHSV (4 h1) [25]. Molybdenum phosphate were synthesized catalyst load (0.10–1.0 mol%), temperature (ca. 298 K), magnetic stir.
and used as solid catalyst on solketal synthesis, nonetheless, they
were active only when supported on SBA-15 mesoporous silica.
of the ketone, favoring its nucleophilic attack by less hindered
After an 2 h of reaction at room temperature (ca. 40% wt, 1:3 glyc-
hydroxyl of the glycerol (i.e., terminal hydroxyl) [27,28]. So, it is
erol ketone molar ratio), the reactions attained 100% of conversion
acceptable that after this step, a second nucleophilic attack of other
and selectivity [26].
nearest hydroxyl result in an intermediate protonated alkoxide,
It is important highlight that all the SnF2-catalyzed reactions
which undergo cyclization with consequent water elimination,
were highly regioselective for 1,3-dioxolane (five membered ring
providing the solketal [27]. For the Brønsted acid-catalyzed ketal-
solketal), regardless catalyst load used.
ization reactions, the five membered ring of solketal is preferen-
Kinetic curves in Fig. 2 show that the reactions performed with
tially obtained, possibly through a reaction pathway where a
higher load of catalyst attained a greater initial rate. It is suggestive
carbenium ion intermediate is generated [28].
that the presence of a higher amount of active sites (i.e. Sn(II)
Herein, there are no protons in the solution. In addition, the
cations present on crystal lattice of the solid SnF2) plays a key role
reaction occurs in a biphasic system wherein the solid Lewis acid
to activate the carbonyl group of propanone.
catalyst is in contact with the two liquid phases (i.e. glycerol and
propanone) and, therefore, it was expected that the reaction should
4. Mechanism proposal occur at the interface between the liquid and solid phases. There-
fore, another mechanism should is operating.
Usually, the role of H+ cations in the glycerol ketalization reac- The literature proposes that on solid supported SnCl2-catalyzed
tions catalyzed by Brønsted acids is protonating the carbonyl group FFA esterification reactions the carbonyl activation of fatty acid
occur through interaction with Sn(II) cations on catalyst surface
Table 1 [29,30]. Similarly, Nanda et al. have also proposed mechanism for
Effect of SnF2 catalyst load on conversion and selectivity of glycerol ketalization with
propanonea,.b.
Lewis acid ketalization glycerol [25,28]. They suggested that all
the other reaction steps (i.e. nucleophilic attack, proton migration,
Run SnF2 load (mol%) Conversionb (%) Solketal selectivityb (%) water elimination), involve intermediates that remain bonded to
1 0.10 3 94 the Sn(II) on catalyst surface.
2 0.25 8 93 Thus, based on these references and our results, we supposed
3 0.50 40 95
that at room temperature, both glycerol and ketone are in straight
4 0.75 76 96
5 1.00 97 97 contact with the catalyst (Scheme 2). This proximity may favor the
a
activation of the carbon atom of carbonyl group by the SnF2 cata-
Reaction conditions: glycerol (21.0 mmol): propanone (168.0 mmol) molar
lyst, increasing its electrophilicity. The terminal hydroxyl group
ratio (1:8); CH3CN (15 mL), temperature (ca. 298 K), magnetic stir.
b
Determined by GC and GC–MS analyses. In general, the molar ratio between the of glycerol attacks the carbon atom of the carbonyl group, resulting
dioxolane: dioxane isomers was 1:16 was obtained. in the intermediate depicted in the Scheme 2. While the proton is
832 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835

captured by the non-ligand electrons pair of secondary hydroxyl, 100

there is the formation of a bond between the carbonyl oxygen 90
atom and the secondary carbon atom of glycerol. These two simul-
80
taneous steps are followed by consecutive elimination of water to
form the 1,3 dioxolane five-membered ring (i.e., solketal) [31]. 70

glycerol conversion (%)

In according with Li et al., the role of the Lewis acid metal is 60
analogous action played in the Meewein-Pondndrof-Verley reduc-
tion mechanism or Oppenauer oxidation reactions, where it is 50

responsible by coordinating and activating the carbonyl group of 40

the acetone [17]. Corma et al. proposed that the framework atom
30
next to the Lewis acid metal site (i.e. in this case, fluoride is the
neighboring atom to the Lewis sites, Sn(II) cations) may play a vital 20 SnF2

role as a Lewis base to activating the glycerol [31]. without catalist aer 30 min
10
Alternatively, Nanda et al. have used a reaction framework for
0
the ketalization reaction proceeding via acidic catalytic mechanism 0 20 40 60 80 100 120
involving 3 steps [32]. They proposed that the first step comprises Time (min)
the surface reaction between the adsorbed ketone and glycerol
over the catalyst surface to form a hemi-acetal intermediate. The Fig. 3. Solubility test of SnF2 catalyst on glycerol ketalization with propanone.
Reaction conditions: glycerol (21.0 mmol); propanone (168.0 mmol); molar ratio
following step is the elimination of water resulting in the forma-
(1:8); catalyst load (4.76 mmol; 1.0 mol%); temperature (ca. 298 K); magnetic stir.
tion of a carbocation on the carbonyl carbon atom; the last step
is the removal of the proton to form the solketal. Nanda et al. sug-
gested that future studies aiming to determine these adsorbed spe- a very poor conversion was achieved in reactions with low load of
cies ‘‘in situ” may be useful to confirm these hypotheses [32]. SnF2 solid (ca. 0.10 mol%).
We also evaluated the reuse and recycle of tin(II) fluoride cata-
5. Reuse and recycling of SnF2 catalyst lyst (Fig. 4). The high rates of recovery obtained in successive
cycles of reuse are indicative that the procedure used was highly
Firstly, it is an imperative highlight that SnF2 was used as a solid efficient. In addition, any significant decrease of activity or selec-
catalyst without any previous treatment, therefore, seems inade- tivity was observed after for cycles of reuse. It was verified only
quate to use the word ‘‘leaching”, which is more appropriate for a little reduction on recovery rate, probably due to the partial
supported-solid catalysts. However, it is possible that the SnF2 cat-
alyst could be partially dissolved throughout the reaction. So, to
assessing this possibility, we removed the catalyst after 30 min
of reaction, and monitoring the conversion progress during the
time normally used (Fig. 3). Additionally, we measured the amount
100
of Sn(II) soluble after the 30 min studied.
It can be seen that the reaction had a minimum conversion after
80
the catalyst being removed, suggesting that no catalyst is solved
into reaction medium. Additionally, we also have quantified Sn
60
(II) cations possibly soluble into reaction medium. The AAS analy-

percent
ses of samples collected after recycle procedure revealed that only
40
a small amount of Sn was present in the solution; only 1.4% related
to initial mass of Sn(II) (initial mass before of reaction, 27.2 mg of
20
Sn(II); only 0.38 mg of Sn(II) was soluble). This result is in agree-
ment with literature, which reports the negligible solubility of 0
SnF2 in polar organic solvents. catalyst recovery
It is evidence that this is a predominantly heterogeneous pro- 4
3
cess. To prove this hypothesis, we carried out the reactions with solketal selectivity
re 2
an equal amount of soluble Sn(II) cations (i.e. soluble SnCl2 catalyst cy 1
cl e
was used in this test, with 0.38 mg of Sn(II) cations). We have fresh glycerol conversion
found that no reaction progress was observed (ca. conversion < 3%).
This result also agrees with that previously shown in Fig. 2, where Fig. 4. Recovery and reuse of SnF2 catalyst on glycerol ketalization.

δ
O OH
δ
δ+ H + H2
C CH2 δ C OH
C OH
H 3C CH 3 magnetic stirr O H
RT H3 C CH3
+ H2O
δ+ C O O
OH SnF2(s)
δO
HO CH OH H 3C CH 3

C C Sn Sn Sn
H2 H2
F F

Scheme 2. A mechanism proposal for SnF2-catalyzed glycerol ketalization with propanone in heterogeneous phase adapted from Li et al. (Ref. [17]).
 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835 833

Fig. 5. Process Flow Diagram (PFD) for solketal production from propanone and glycerol.

dissolution of catalyst and/or difficult in subsequent steps of performed to determine profitability [33]. The equipment costs
centrifugation/filtration. (Cp), using carbon steel, were estimated by Eq. (1) [33].
The high stability of SnF2 after reuse cycles is an aspect positive 2
if compared to the stability of other solid supported Lewis or log Cp0 ¼ K 1 þ K 2 logðAÞ þ K 3 ðlogðAÞÞ ð1Þ
Brønsted acid catalysts, which suffer with leaching due to high In Eq. (1), A is the capacity or size parameter for the equipment
polarity of reaction medium [24]. A significant decrease on conver- and K1, K2 and K3 are parameters. The cost of purchased equipment
sion was observed by Chary et al. in SBA-15 supported molybde- was corrected by the time factor (I), the material factor (FM) and
num phosphate-catalyzed reactions; although selectivity has the conditions factor (FP). So, the purchased equipment cost was
remained close to 100%, a conversion decreased from 100 to 60% expressed by Eq. (2):
after the catalyst be reused four times [26].
Cp ¼ Cp0  FM  FP  I ð2Þ
To account the inflation, the Chemical Engineering Plant Cost
6. Process simulation Index (2015) was used. After, the capital and operating costs of
the process were compared, evaluating the major contributions
Simulation of solketal production plant was based on experi- to the total cost (feedstock, catalyst, utilities, equipment, etc.). Dis-
mental data obtained in the catalytic tests already described. counted cash flow analysis was performed to obtain the minimum
Fig. 5 shows the Process Flow Diagram (PFD) developed for solketal sale price (MSP) of solketal (Eq. (3)), (i.e., the lowest price that
production from glycerol and propanone. leads the profit to zero).
The following unit operations compose the production plant:
Mixers (MIX-100, MIX-101), one reactor for conversion of glycerol MSP ¼ ðCRF  Total Capital Cost
and propanone to solketal (CRV-100), filter (SSSep-100), heat þ Total Operating CostÞ=Total production of solketal ð3Þ
exchanger (E-100) and two distillation columns (T-100 and T-
101). The conversion of glycerol was estimated at 80%. The reactor
(CRV-100) works at 298 K and 1 bar pressure. Unreacted glycerol
was recycled back into MIXER (MIX-101). Glycerol, propanone Annual operating costs
catalyst; 0.005% others*; 7.96%
and SnF2 were then separated from the water in a filter (SSSep- Plant Overhead;
Utilities; 2.65% 19.48%
100) and two distillation columns.
In general, when solid catalysts are used, they are easily recov- 6.93; Fixed
charges
ered from reaction medium by simple filtration steps. However,
the separation of unreacted glycerol from solid SnF2 it’s not easy
herein. Thus, ketone excess was removed via evaporation, and Glycerol;
7.50%
afterward we washed solid remaining with ethanol. This step dis-
solves the glycerol and keeps the solid the SnF2 catalyst, which is
insoluble in this solvent. Material and energy balances are shown Operating labor;
in the Supplementary material. Acetone; 16.19%
9.93%

7. Economic analysis
Maintenance;
13.86% General expense;
The raw materials used were propanone, glycerol and tin(II) flu-
15.49%
oride. With chemical and utility cost, the production cost of solke-
tal was obtained and a discounted cash flow analysis was Fig. 6. Annual operating costs.
834 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835

Fig. 7. Sensitivity analysis in solketal cost.

For this calculation, capital costs should be annualize, by use of 9. Conclusion

capital recovery factor (CRF) given by Eq. (4). In this equation, i is
the desired rate of return and n is the plant lifetime. Solid SnF2 is a simple, inexpensive and commercially affordable
compound able to catalyze the glycerol ketalization with propa-
n n
CRF ¼ ið1 þ iÞ =½ð1 þ iÞ  1 ð4Þ none at room temperature, giving solketal with high selectivity
To perform an economic analysis, some considerations should (ca. >94%) and converting almost completely the glycerol (ca.
be taken into account. >90%). A proposal mechanistic was done aiming clarifying the role
of Lewis acid sites on this reaction. In addition to be a water toler-

The plant was assumed to operate in a continuous steady state ant Lewis acid, easily handled and stable, the SnF2 solid catalyst
mode for 8766 h/y. Depreciation method was straight line, with was efficiently recovered and reused without loss activity. The
a tax rate of 40% and salvage value of 20% of initial capital cost. use of SnF2 as catalyst avoids the laborious work of synthesis of
The plant life was 20 years and the desire rate of return is set in solid supported catalysts. For these reasons, it is a promising can-
20%; didate to operate in industrial plants to convert the glycerol, an

The cost of propanone was assumed to be 1000 $/metric ton, the attractive and renewable raw material, in more value added pro-
cost of glycerol was estimated to 864 $/metric ton, the water duct (solketal). To verify these aspects, we also carried out an
cost was 0.1 $/metric ton and catalyst 70 $/kg [31,33]. assessment to show how computational design engineering and

The catalyst lifetime is assumed to be six months, and the experimental catalysis studies could be linked to analyze the prac-
recharge should be done each six months during the plant tical industry implementation. Specifically, the purpose was to
life. evaluate the solketal production process from glycerol using SnF2
catalyst. For this system, the solketal MSP estimated was 12.29 $/
Capital costs are showed in Table 3 (Supplementary material). kg. The process analyzed is novel, but extremely promising due
The major contributions for operating cost process can be seen in to its large market potential to produce fuel additives. However,
Fig. 6. The total operating cost was $5,467,000. Plant overhead rep- technological development and cost reduction is still necessary
resented the major cost of operation, followed by operating labor to economically competitive implementations.
and general expenses. The MSP calculated for solketal was 12.29
$/kg. Acknowledgments

The authors would like to thank CAPES, CNPq, Rede Mineira de

8. Sensitivity analysis Quimica and FAPEMIG for the financial support and HONEYWELL
for UNISIMTM software that made this research work possible.
The sensitivity analysis was performed to observe the effect of
some parameters (i.e., plant capacity, glycerol price, propanone Appendix A. Supplementary data
price) in the solketal price. There unto, the original value of each
parameter was decreased in 20% (Case A). Then, a new economic Supplementary data associated with this article can be found, in
analysis was separately performed for each new value and the the online version, at http://dx.doi.org/10.1016/j.cej.2016.09.002.
new MSP was calculated. The values increased in 20% in relation
to the original ones (Case B).
References
Sensitivity analysis for the process is showed in Fig. 7. Plant
capacity had a large effect, since the increase in 20% reduced the [1] M. Shirani, H.S. Ghaziaskar, C. Xu, Fuel Proces. Technol. 124 (2014) 206–211.
MSP in 14.9%. [2] R.R. Soares, D.A. Simonetti, J.A. Dumesic, Angew. Chem. Int. Ed. 45 (2006)
3982–3985.
The glycerol and propanone prices had a little effect in the
[3] J.A. Melero, R.V. Grieken, G. Morales, M. Paniagua, Energy Fuel 21 (2007) 1782–
solketal price. The supposed increase of 20% on glycerol price 1791.
leaded to a rise of nearly 1.96% on solketal MSP; conversely, a pro- [4] M.B. Güemez, J. Requies, I. Agirre, P.L. Arias, V.L. Barrio, J.F. Cambra, Chem. Eng.
panone price 20% higher leaded to an increasing of nearly 2.62% on J. 228 (2013) 300–307.
[5] C.H. Zhou, J.N. Beltramini, Y.X. Fan, G.Q. Lu, Chem. Soc. Rev. 37 (2008) 527–549.
MSP. The lowest MSP was around 11.43 $/kg assuming a 20% [6] R.P.V. Faria, C.S.M. Pereira, V.M.T.M. Silva, J.M. Loureiro, A.E. Rodrigues, Chem.
higher plant capacity. Eng. J. 233 (2013) 159–167.
 M.J. da Silva et al. / Chemical Engineering Journal 307 (2017) 828–835 835

[7] X. Hong, O. McGiveron, A.K. Kolah, A. Orjuela, L. Peereboom, C.T. Lira, D.J. [21] M.J. da Silva, A.A. Julio, F.C.S. Dorigetto, RSC Adv. 5 (2015) 44499–44506.
Miller, Chem. Eng. J. 222 (2013) 374–381. [22] H. Serafim, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Chem. Eng. J. 178
[8] C.S. Cho, D.T. Kim, H.-J. Choi, T.-J. Kim, S.C. Shim, Bull. Korean Chem. Soc. 23 (2011) 291–296.
(2002) 539–540. [23] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Appl. Catal. B 98
[9] A.L. Cardoso, S.C.G. Neves, M.J. Da Silva, Energy Fuels 23 (2009) 1718–1722. (2010) 94–99.
[10] A. Casas, M.J. Ramos, J.F. Rodríguez, Á. Pérez, Fuel Proces. Technol. 106 (2013) [24] J. Esteban, M. Ladero, F. García-Ochoa, Chem. Eng. J. 269 (2015) 194–202.
321–325. [25] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.-A. Poirier, C.C. Xu, Appl.
[11] A.B. Ferreira, A.L. Cardoso, M.J. Da Silva, ISRN Renewable Energy (2012) 1–13 Energy 123 (2014) 75–81.
142857. [26] Sailaja Gadamsetti, N. Pethan Rajan, G. Srinivasa Rao, Komandur V.R. Chary, J.
[12] M.L. Da Silva, A.P. Figueiredo, A.L. Cardoso, R. Natalino, M.J. Da Silva, J. Am. Oil Mol. Catal. A 410 (2015) 49–57.
Chem. Soc. 88 (9) (2011) 1431–1437. [27] S. Yan, S.O. Salley, K.Y.S. Ng, Appl. Catal. A 353 (2009) 203–212.
[13] A.B. Ferreira, A.L. Cardoso, M.J. da Silva, Catal. Lett. 143 (2013) 1240–1246. [28] M.R. Nanda, Y. Zhang, Z. Yuan, W. Qin, H.S. Ghaziaskar, C.C. Xu, Renew. Sust.
[14] C.E. Goncalves, L.O. Laier, M.J. da Silva, Catal. Lett. 141 (2011) 1111–1117. Energy Rev. 56 (2016) 1022–1031.
[15] F.D.L. Menezes, M.D.O. Guimaraes, M.J. da Silva, Ind. Eng. Chem. Res. 52 (2013) [29] W. Xie, H. Wang, H. Li, Ind. Eng. Chem. Res. 51 (2012) 225–231.
16709–16713. [30] V.M. Mello, G.P.A.G. Pousa, M.S.C. Pereira, I.M. Dias, P.A.Z. Suarez, Fuel Process.
[16] E. Kemnitz, Catal. Sci. Technol. 5 (2015) 786–806. Technol. 92 (2011) 53–57.
[17] L. Li, T.I. Korányi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012) 1611– [31] M.J. Climent, A. Corma, P. De Frutos, S. Iborra, M. Noy, A. Velty, P. Concepción, J.
1619. Catal. 269 (2010) 140–149.
[18] UNISim Honeywell, 2013. http://www.honeywell.com/. [32] M.R. Nanda, Z. Yuan, W. Kin, H.S. Ghaziaskar, M.-A. Poirier, C.C. Xu, Fuel 117
[19] M.J. Da Silva, A.L. Cardoso, R. Natalino, Int. J. Chem. React. Eng. 8 (A12) (2010) (2014) 470–477.
1–12. [33] R. Turton, R.C. Bailie, W.B. Whiting, J.A. Shaeiwitz, Analysis, Synthesis, and
[20] M.S. Peters, K.D. Timmerhaus, R.E. West, Plant design and economics for Design of Chemical Processes, 2nd ed., Prentice Hall, Englewood Cliffs, NJ,
chemical engineers, 5th ed., in: McGraw-Hill Chemical Engineering Series, 2003.
2003.
 Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212

Contents lists available at ScienceDirect

Journal of Molecular Catalysis A: Chemical

journal homepage: www.elsevier.com/locate/molcata

Solketal formation from glycerol and acetone over hierarchical

zeolites of different structure as catalysts
J. Kowalska-Kus ∗ , A. Held, M. Frankowski, K. Nowinska
Faculty of Chemistry, Adam Mickiewicz University, Poznan, Poland

a r t i c l e i n f o a b s t r a c t

Article history: Glycerol acetalization reaction with acetone was performed at 343 K and using a 1:1 molar ratio of
Received 6 September 2016 glycerol to acetone in the feed over hierarchical zeolites comprising pores of different diameters (MFI,
Received in revised form BEA, and MOR). The best catalytic performance for glycerol acetalization was exhibited by hierarchical
10 November 2016
(micro-mesoporous) MFI zeolites. Glycerol conversion over MFI zeolites exceeded 80%, with almost 100%
Accepted 11 November 2016
selectivity to the desired product (4-hydroxymethyl-2,2-dimethyl-1,3-dioxolane known as solketal). A
Available online 16 November 2016
significant increase in glycerol conversion and also selectivity to solketal in the studied reaction resulted
from the easier accessibility of the active sites to reagents due to the formation of mesopores by means
Keywords:
Acetalization
of desilication of the microporous zeolites.
Glycerol © 2016 Elsevier B.V. All rights reserved.
Desilication
Dealumination
Hierarchical zeolites

1. Introduction additionally as a fuel additive, increasing the octane number of

gasoline and reducing the emissions resulting from fuel combus-
Transesterification of vegetable oil and animal fats with tion.
methanol results not only in the production of biodiesel but also The reaction of glycerol and acetone is a reversible process
in the formation of a significant amount of waste glycerol phase resulting in 2,2-dimethyl-1,3-dioxolane-4-methanol (solketal) and
(about 100 kg for every tonne of biodiesel) which, besides glycerol, water according to Scheme 1. Isomeric 1,3-dioxan-5-ol may also be
is also comprised of spent catalyst and excess methanol [1]. Con- produced [11].
sidering that an increase in biodiesel production brings about an Our studies are concentrated on searching for the stable, easily
increasing surplus of glycerol, its efficient utilization is very chal- accessible, and active catalyst for the acetalization reaction. Tradi-
lenging. Glycerol may be transformed into different chemicals that tionally, liquid inorganic acids have been applied as catalysts for
are intermediate for the production of consumer goods. Among the mentioned process. Owing to the problems involved with both
others are acroleine and acrylic acid obtained by the oxydehydra- product separation and the utilization of spent catalysts, exten-
tion process [2,3], which are polymerized to acrylates; propylene sive research on the applications of heterogeneous catalysts has
glycol from glycerol hydrogenolysis [4], applied in the pharmaceu- been undertaken [12]. Among different solid acidic catalysts such
tical and food industry, and also dihydroxyacetone (produced by as heteropolyacids [13], mesostructure silica modified in a differ-
glycerol oxidation [5,6]), used for the production of cosmetics and ent manner (i.e. sulfonic acid-functionalized materials) [14], acidic
medicines. One promising procedure for glycerol utilization is the resins such as Amberlyst [15], and also various kinds of zeolites
reaction of glycerol with acetone or formaldehyde in the presence were tested [16–19]. So far, the best catalytic performance for glyc-
of acidic catalysts to produce acetals or ketals [7–9]. erol acetalization exhibited Amberlyst. However, its application on
A practical application of solketal, a product of glycerol acetal- a wide scale is limited due to its low thermal stability hindering the
ization with acetone, has been found in a number of industrial regeneration of Amberlyst by calcination after several using. There-
processes [10]. It may be used in the polymer industry as a plas- fore, more suitable catalysts of acetalization seem to be zeolites. The
ticizer, in pharmaceutical production as suspending agent, and highest activity in the studied reaction was found over BEA zeolite.
Despite significant activity of BEA zeolite, its high price makes the
process of glycerol utilization not profitable. On the other hand,
acetalization of glycerol performed over MFI zeolites, due to their
∗ Corresponding author.
relatively low price, wide availability and high thermal stability,
E-mail address: [email protected] (J. Kowalska-Kus).

http://dx.doi.org/10.1016/j.molcata.2016.11.018
1381-1169/© 2016 Elsevier B.V. All rights reserved.
206 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212

2.3. Characterization

2.3.1. XRD
X-ray diffraction measurements were carried out on a Bruker
AXS D8 Advance diffractometer using CuK␣ radiation in order to
confirm the structural correctness of the prepared samples. Lower-
Scheme 1. Acetalization reaction of glycerol with acetone.
ing of the crystallinity of alkali treated catalysts with comparison to
the parent ones was determined according to the method described
should be very attractive. The restriction that brings about lower in ref. [20,22,23]. The relative crystallinity was calculated according
activity of MFI catalysts for this process is related to the shape- to Eq. (1).
selectivity properties associated with the relatively narrow channel
size [20]. Our earlier calculation using the GaussView programme ␣ = ␣o × I/Io (1)
exploitation [21] indicated that the diameter of the acetalization
products is situated in the range of 0.43–0.51 nm, which could where we assume ␣o as equal to 100, I and Io are the areas of the
result in diffusion limitation. reflexes characteristic of parent and modified zeolites, respectively.
The previous studies have therefore become an inspiration to The I and Io values response to the sum of the areas of the most
improve the activity of zeolites for glycerol acetalization by lim- intense peaks at 2 Theta: 23.1◦ , 23.3◦ ,23.7◦ , 23.9◦ , 24.4◦ for MFI
itation of diffusion restriction. This goal has been attained by zeolite; 25.8◦ , 26.5◦ , 27.8◦ for MOR and 22.5◦ for BEA zeolite.
application of hierarchical (micro-mesoporous) MFI, BEA and MOR
zeolites as heterogeneous catalysts used for glycerol with acetone 2.3.2. Elemental analysis
acetalization. For this purpose, in the present paper, zeolites of Approximately 50 mg of the sample was weighed, placed in a
MFI structure (pore diameter 0.51–0.55 nm) as well as BEA, and Teflon bottle and 8 ml of supra pure nitric acid (Sigma-Aldrich) was
MOR, characterized with higher channel diameters (0.61–0.67 nm added. Then Teflon bottles were placed on the heating plate with
and 0.67–0.70 nm, respectively [20]), were prepared by desilica- magnetic stirring for 30 min. After this time 2 ml of hydrofluoric
tion procedure, which is relatively cheap and easy modification. acid (Sigma-Aldrich) was added and placed again on the heat-
It was expected that the application of micro-mesoporous zeolites ing plate with magnetic stirring until dissolution (around 30 min).
as catalysts for glycerol acetalization should demonstrate excellent After cooling, the samples were placed in polyethylene volumetric
glycerol conversion and selectivity to the desired product. flasks and filled up to 100 ml with deionised water. The content
of aluminium and silicon were determined by atomic absorp-
2. Experimental tion spectrometry with acetylene-nitrous oxide flame atomisation
(F-AAS) using the 3D double-beam Shimadzu AA7000 analytical
2.1. Materials technique (Shimadzu), equipped with an ASC-7000 autosampler
(Shimadzu). The hollow cathode lamps (HCL) employed were made
H-MFI (Süd Chemie, Si/Al = 12 and Alsi Penta, Si/Al = 27; denoted by Photron.
as MFI-1-P, MFI-2-P, respectively), H-Mordenite (PQ Corpora-
tion, Si/Al = 17; designed as MOR-P) and H-Beta (Sigma Aldrich,
2.3.3. BET
Si/Al = 12; denoted as BEA-P) were used as raw materials for alka-
N2 adsorption/desorption isotherms were recorded on a Quan-
line and subsequent acidic treatment.
tachrome Nova1000e analyzer at 77 K. The total surface area was
calculated according to the BET equation. The micropore and meso-
2.2. Catalyst modification
pore volume, and the external surface area were evaluated by the
t-plot method. Pore size was reported as the average pore diame-
2.2.1. Alkaline treatment
ter. Before measurements the samples were preheated at 573 K for
All the commercial zeolites were alkali-treated using the fol-
4 h in vacuum.
lowing procedure. 5.0 g of zeolite was added to a 0.2 M (150 ml)
aqueous NaOH solution and stirred at 353 K for 2 h. Afterwards, the
slurry was cooled, filtered, and then washed with distilled water 2.3.4. FT-IR spectra of adsorbed pyridine
until the pH of the filtration solution decreased to 7. After dry- The FT-IR spectroscopic self-supporting wafer technique on
ing at 363 K for 12 h, the alkali-treated zeolite was ion exchanged a Bruker-Vector 22 spectrometer was used to determine the
with 0.5 M NH4 NO3 aqueous solution by stirring at 308 K for 24 h, acidic properties of the parent and modified zeolites. The wafers
followed by filtering and rinsing with distilled water to remove (20 mg/cm2 ) were prepared from powdered zeolites. Initially, the
sodium ions. After drying at 363 K for 12 h, the ammonium form samples were heated in vacuum at 673 K for 2 h. The samples were
of zeolite was subsequently calcined in static air at 823 K for 3 h to then cooled to the room temperature (RT) after which the pyri-
form the H-form, denoted as MFI-number-AT, MOR-AT and BEA-AT, dine was adsorbed with following evacuation at 373 K to remove
respectively. physically adsorbed pyridine. The desorption temperature points
were 373 K, 473 K, 573 K, and 673 K. The spectrum was recorded
2.2.2. Acidic treatment after desorption in vacuum at a related temperature for 30 min.
Some of the alkaline treated samples were used for further acid The number of acidic centers was calculated with Eq. (2) [21,24]:
treatment with aqueous solutions of nitric or citric acids. The pro- AL Cd AB Cd
cedure of the acid treatment was as follows. One part of zeolite nT = + (2)
εL m εB m
was put into 0.5 M nitrate acid (NA) solution and the second one
was placed in 0.5 M citric acid (CA) solution. The resulting mixtures nT – total number of acidic sites expressed as micromoles of pyri-
were kept at 353 K for 3 h with stirring. The slurry was then cooled dine per 1 g of catalyst, AL and AB – integrated absorbance of IR
down, washed and dried at 363 K for 12 h and the resulting samples bands related to Lewis and Brønsted sites, Cd – cross section of
were designated as MFI-number-NA, MOR-NA, BEA-NA and MFI- wafer in cm2 and m – mass in g of pressed pellet, and ␧L and ␧B molar
number-CA, MOR-CA, BEA-CA, respectively. Finally, all the treated absorption coefficients for Lewis and Brønsted acid sites equal to
samples were calcined at 823 K for 3 h. 1.5 and 1.8 [␮mol−1 cm], respectively.
 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212 207

Fig. 1. XRD patterns of parent and modified MFI-1 (A) and MFI-2 (B) zeolites with Si/Al ratio equal to 12 and 27, respectively.

2.4. Catalytic activity measurement 3. Results and discussion

The activity of the parent and modified zeolites was tested for 3.1. Catalyst characterization
the acetalization reaction of glycerol with acetone. The reagents
used in the experiments, glycerol (99.5% purity) and acetone (99% 3.1.1. XRD
purity), were purchased from Chempur. All the experiments were The XRD patterns of parent and modified zeolites of MFI struc-
conducted at a temperature of 343 K in closed glass vials of 4 cm3 ture are presented in Fig. 1 and BEA and MOR structures are
capacity, under vigorous stirring (400 rpm) in a multiple-well par- demonstrated in Fig. 2. The patterns show that the proper struc-
allel reaction block in the presence of catalyst in the amount of 1% ture of the zeolites has been preserved in the samples after the
related to glycerol. A 1: 1 molar ratio of glycerol to acetone was alkaline treatment and after the ionic exchange with aqueous solu-
used in the feed. Catalysts were pre-activated in air at 623 K for tions of NH4 NO3 . No additional phase was noted in the modified
2 h followed by cooling down in a desiccator before the catalytic samples. However, a small decrease in the intensities of the char-
run. Liquid products were analysed by GC, using a VARIAN CP- acteristic reflexes was observed, which suggests a lowering of the
3800 chromatograph equipped with an FID detector with a VF-5 ms crystallinity. The related crystallinity of the modified samples with
column. Besides a predominant product, a 5-membered ring ketal reference to the initial materials is listed in Table 1. These results
2,2-dimethyl-1,3-dioxolane-4-methanol (solketal) and unreacted prove that alkali treatment induces some lowering in the crys-
glycerol and acetone, only a 6-membered ring ketal 2,2-dimethyl- tallinity. A decrease in crystallinity was visible for both MFI zeolites,
1,3-dioxolane-5-ol, a by-product, was identified by the Varian 4000 at a similar level (about 30% of the initial value). A similar reduction
GC–MS apparatus. The catalytic activity, reported as glycerol con- of crystallinity was noted for MOR (30% of the initial value), while
version (denoted as G C), selectivity to products: solketal (denoted the crystallinity of BEA zeolite was practically unchanged (2% of the
as S S) and isomer (denoted as S I), solketal yield (denoted as Y S), initial value).
was calculated on the basis of the equations presented in [25]. The following treatment of desilicated samples with HNO3 or
Conversion of glycerol (%) C6 H8 O7 solutions in the next step of catalyst modification was
motivated by the literature suggestions [26], indicating the pos-
sible occlusion of aluminum and hindrance of mesopores opening
mol of glycerol converted in alkaline treated zeolites. The data showed that acidic treatment
C= × 100 (3)
initial mol of glycerol does not influence the crystalline structure of the modified zeo-
lites significantly, however, it results in an increase in the Si/Al
ratio (Table 1). Only the treatment of desilicated BEA zeolite with
where mol of glycerol converted = initial mol of glycerol − final mol nitric acid leads to some lowering in the crystallinity up to 81%. This
of glycerol may be due to the partial extraction of framework aluminum from
Selectivity to solketal or isomer (%) BEA zeolite, considering that BEA zeolite is relatively easy to dealu-
minate [20]. A similar alteration of the crystallinity of BEA zeolite
modified by means of desilication with following treatment with
mol of X formed nitric acid solution was reported by Wang et al. [27].
S= × 100 (4)
mol of glycerol converted

3.1.2. Si/Al ratio

where X = solketal or isomer The alkali treatment of zeolite should lead to a preferential desil-
Yield of solketal (%) ication, which would then be reflected in the changing of Si/Al ratio.
The Si/Al ratios of the parent and modified zeolites are presented
in Table 1. The Si/Al ratio of all alkaline treated zeolites obtained
S×C via the F-AAS analysis was found to be lower than that of the par-
Y= (5)
100 ent catalysts. The presented results indicate that the higher Si/Al
208 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212

Fig. 2. XRD patterns of parent and modified BEA (A) and MOR (B) zeolites with Si/Al ratio equal to 12 and 17, respectively.

Fig. 3. N2 sorption isotherms of parent and modified MFI-1 (A) and MFI-2 (B) zeolites with Si/Al ratio equal to 12 and 27, respectively.

Table 1
Physicochemical and textural characterization of the parent and modified zeolites.

Zeolite Molar SBET b Sextc Mesporesc Average pore diameterb Vtotalb Vmesoc Relative
Si/Al ratioa (m2 /g) (m2 /g) (%) (nm) (cm3 /g) (cm3 /g) Crystallinityd

MFI-1-P 12 242 24 10 3.06 0.19 0.05 100

MFI-1-AT 9.2 292 135 46 4.08 0.30 0.20 72
MFI-1-CA 12.9 293 135 46 4.16 0.31 0.21 74
MFI-1-NA 13.6 308 151 49 4.17 0.32 0.22 72

MFI-2-P 27 252 29 12 3.04 0.19 0.05 100

MFI-2-AT 13.2 307 164 53 3.98 0.31 0.30 74
MFI-2-CA 21.5 310 167 54 3.94 0.32 0.22 79
MFI-2-NA 25.7 308 158 51 4.06 0.30 0.29 77

BEA-P 12 490 159 32 5.84 0.71 0.50 100

BEA-AT 9.3 525 252 48 9.19 1.21 1.03 98
BEA-CA 12.0 539 259 48 9.31 1.21 1.02 95
BEA-NA 18.3 538 282 52 9.00 1.25 1.08 81

MOR-P 17 398 74 32 3.54 0.35 0.14 100

MOR-AT 10.8 399 117 48 5.73 0.57 0.39 67
MOR-CA 20.2 410 129 48 5.87 0.63 0.45 70
MOR-NA 25.8 409 138 52 6.12 0.60 0.42 65
a
F-AAS.
b
BET-method.
c
T-plot method.
d
According to Refs. [20,22,23].

in applied MFI zeolites, the higher the degree of desilication was well as on the applied conditions. A high concentration of frame-
observed. According to literature [26,28,29], desilication of zeolite work aluminum (low Si/Al ratio) results in restriction of the silicon
depends significantly on the Si/Al ratio of the initial material as removal and in limitation of mesoporosity development. This is in
 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212 209

Fig. 4. N2 sorption isotherms of parent and modified BEA (A) and MOR (B) zeolites with Si/Al ratio equal to 12 and 17, respectively.

accordance with the lower degree of desilication in the MFI zeo- Although the isotherms of parent BEA and MOR (Fig. 4) show
lite of Si/Al equal to 12, when compared to the sample with Si/Al some hysteresis due to the presence of mesopores, the hysteresis
equal to 27 (Table 1). On the other hand, the relatively high meso- loop increases visibly as a result of the desilication process. The
porosity achieved in our experiment over MFI-1-AT may result from average pore diameter of modified samples was shifted towards
long duration of the treatment (2 h) at a relatively high temperature higher values, indicating the formation of macropores in the mod-
353 K. ified samples. A similarly high pore diameter was reported in the
The influence of zeolite structure on the desilication process literature [20,29].
has been indicated in the literature [30]. According to the results
presented in Table 1 silicon removal as an effect of alkaline
3.1.4. Acidic properties
treatment results in a decrease in the Si/Al ratio in the fol-
The acidity of the parent and modified zeolites was estimated on
lowing way: MFI(27) > MOR(17) ∼ MFI(12) > BEA(12). Subsequent
the basis of the FT-IR spectra of adsorbed pyridine after desorption
treatment with acids (HNO3 or C6 H8 O7 ) brings about an increase
at a temperature in the range of 373–673 K. Table 2 summarizes the
in the Si/Al ratio as a result of aluminum removal. Susceptibility
acidic properties of all applied zeolites. Desilication of MFI and also
to dealumination is especially visible for MOR and BEA zeolites. It
MOR and BEA zeolites with following ionic exchange with NH4 NO3
has been shown by Groen et al. [31] and also by Tian et al. [32] that
resulted in the increase of both total acidity estimated at 373 K
resistance of BEA zeolite to dealumination is lower when compared
(Figs. 5 and 6) as well as strong acidity measured after the evacua-
to MFI zeolites.
tion of pyridine at 673 K (Table 2). Both Brønsted acidity and Lewis
acidity were responsible for the increase in total acidity in MFI zeo-
3.1.3. Textural properties
lites. The high acidity after the desilication process indicates that
The influence of the alkali treatment on the porosity of the
aluminum is still in tetrahedral coordination. On the other hand, the
applied zeolites was studied by the N2 adsorption–desorption
lower Brønsted acidity and higher Lewis acidity observed on desil-
procedure at 77 K and the results are listed in Table 1. Alkaline treat-
icated MOR and BEA zeolites may evidence that the dealumination
ment of the zeolites with NaOH solution and the following ionic
process occurs with desilication (Tables 1 and 2).
exchange with aqueous solution of NH4 NO3 leads to spectacular
When the desilicated samples were treated with aqueous solu-
mesopore formation. However, the enlargement of the surface area
tions of HNO3 or C6 H8 O7 , the Lewis acidity was clearly lower in
and the mesoporosity depends on the structure of the initial sam-
comparison with the samples exchanged with NH4 NO3 . This sug-
ples. MFI zeolites with a different Si/Al ratio show a similar increase
gests that the treatment of desilicated zeolites with acidic solutions
in surface area (in the range of 20–25% of the initial values), while
results in dissolution of aluminum debris, as suggested in the lit-
the surface area of BEA zeolite increases by 10% and changes only
erature [26]. This is in accordance with alteration in the Si/Al ratio
slightly for MOR (3% of the initial value). Desilication also results in
(Table 1).
an increase of the surface of mesopores. The contribution of meso-
pores in total surface area increased from approximately 10% for
initial zeolites of MFI structure to about 50% for modified ones, 3.2. Catalytic activity for glycerol acetalization
while for the BEA zeolite and MOR the contribution of mesopores
increased from 32% to about 50% and from 19% to about 34%, respec- Acetalization of glycerol with acetone to obtain solketal was
tively. The following treatment with HNO3 or C6 H8 O7 acid solutions tested over parent zeolites of MFI, MOR and BEA structures and
influenced the textural parameters only slightly (Figs. 1 and 2). Even also on the materials subjected to desilication with following ionic
though, no significant changes in the modified zeolite texture were exchange with NH4 NO3 solution and subsequent acidic treatment.
recorded, the Si/Al ratio increased (Table 1) and Brønsted as well as Protonic forms of commercial zeolites displayed very different
Lewis acidity decreased (Table 2) after acidic treatment, confirming activity for glycerol acetalization depending on their pore arrange-
partial aluminum removal. ment. The initial MOR and BEA zeolites characterized by a relatively
As shown in Fig. 3, the isotherms of both MFI zeolites large pore opening (diameter of about 0.67 nm) and higher acces-
showed typical Langmuir Type-I adsorption patterns prior to alkali- sibility to active sites for the reactants show glycerol conversion
treatment. After alkaline treatment, all the isotherms exhibited in the range of 67–72% (Figs. 9 and 10). The desilication process
distinct hysteresis loops indicating Type-IV behavior as a result of MOR and BEA zeolites with generation of mesoporosity resulted
of mesopores formation from the desilication process. This is in in some increase in glycerol conversion up to 81%. The following
agreement with the observed increase in pore volume as well as in treatment with citric acid alters the glycerol conversion slightly,
enlargement of the contribution of mesopores surface in the total while treatment with nitric acid results in a reduction of glycerol
surface area (Table 1). A similar alteration in zeolite texture was conversion. A similar tendency in the lowering of glycerol conver-
described in the literature as a result of the desilication procedure sion over all zeolites treated with nitric acid, compared to values
[26]. obtained after citric acid modification, may arise from a signifi-
210 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212

Table 2
Acidic properties of the parent and modified zeolites.

Acidity (␮mol/g)

Totala Strongb

Zeolite B L B+L B L B+L

MFI-1-P 234 73 307 75 35 110

MFI-1-AT 327 272 599 133 143 276
MFI-1-CA 256 229 485 155 94 249
MFI-1-NA 262 199 461 151 97 248

MFI-2-P 279 63 342 187 28 215

MFI-2-AT 368 96 464 263 61 297
MFI-2-CA 170 199 369 86 73 159
MFI-2-NA 157 182 339 75 73 148

BEA-P 464 210 674 139 206 345

BEA-AT 282 470 752 155 303 458
BEA-CA 219 169 388 139 111 250
BEA-NA 94 186 280 69 55 124

MOR-P 378 78 456 133 99 232

MOR-AT 301 249 550 165 170 335
MOR-CA 235 102 337 130 121 251
MOR-NA 149 93 242 164 50 214
a
calculated after pyridine adsorption at RT and desorption at 373 K for 0.5 h.
b
calculated after pyridine adsorption at RT and desorption at 673 K for 0.5 h.

Fig. 5. FT-IR spectra of pyridine adsorbed at RT and desorbed at 373 K for 0.5 h, related to total acidity, recorded on parent and modified MFI-1 (A) and MFI-2 (B) zeolites
with Si/Al ratio equal to 12 and 27, respectively.

Fig. 6. FT-IR spectra of pyridine adsorbed at RT and desorbed at 373 K for 0.5 h, related to total acidity, recorded on parent and modified BEA (A) and MOR (B) zeolites with
Si/Al ratio equal to 12 and 17, respectively.
 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212 211

cant decrease in acidic sites resulting from partial dealumination

(Tables 1 and 2).
The activity of the parent zeolites of MFI structure, character-
ized by a different Si/Al ratio equal to 12 and 27, was definitely
lower (18% and 37%, respectively), because of the diffusion limita-
tion (Figs. 7 and 8). According to the literature [33] the Si/Al ratio
affects the desilication process and a value in the range of 25–50
has been indicated as optimal. The desilication process resulted in
an increase in glycerol conversion. The glycerol conversion over
MFI-1-AT reached 30%, whereas on the MFI-2-AT it increased up
to 80% (Figs. 7 and 8). The following treatment with citric acid
resulted in a significant increase in glycerol conversion up to 82%
and 85%, respectively. It seems that aluminum debris do not block
Fig. 7. Conversion of glycerol (G C), selectivity to products: solketal (S S) and isomer the mesopores for small nitrogen molecules (N2 applied for poros-
(S I), yield of solketal (Y S) over parent and modified MFI-1 zeolite with Si/Al ratio ity estimation), while they are not quite available for reagents.
equal to 12. Therefore, the removal of aluminum debris results in a significant
increase in glycerol conversion.
The desilication of zeolites also resulted in an increase in selec-
tivity to solketal, which achieved nearly 100%. Solketal, the main
product of the acetalization reaction, is kinetically favoured, while
the isomer 1,3 dioxane is thermodynamically more stable. Longer
contact time with catalytically active sites may result in isomeric
molecule formation (bringing about the selectivity to 1,3 dioxane
up to 25–27 on MFI-1P and MFI-2P – Figs. 7 and 8), therefore, faster
diffusion due to the presence of mesopores seems to be responsible
for limitation of its production up to 3%. The obtained data confirm
that an almost 80% yield of solketal, as the predominant product,
was obtained in the acetalization reaction. Interestingly, high glyc-
erol conversion (80% or more) and near 100% selectivity to solketal
was achieved for the stoichiometric acetone to glycerol ratio (ace-
Fig. 8. Conversion of glycerol (G C), selectivity to products: solketal (S S) and isomer tone: glycerol = 1:1). Considering that the acetalization of glycerol
(S I), yield of solketal (Y S) over parent and modified MFI-2 zeolite with Si/Al ratio
equal to 27.
with acetone is a reversible reaction, a higher acetone to glycerol
molar ratio shifts the reaction towards the products. The best solke-
tal yield reported in the literature was achieved for an acetone to
glycerol ratio equal to 4:1 [34]. The obtained results indicate that
the desilication of zeolites can create highly efficient catalysts for
glycerol acetalization.

4. Conclusions

Glycerol acetalization reaction with acetone performed for the

very first time over large pores, hierarchical zeolites (MFI, BEA and
MOR) as catalysts resulted in a significant increase in glycerol con-
version, even exceeding 80%. As a result of a dramatic shortening of
diffusion path length, selectivity to solketal reached almost 100%.
The most substantial increase in activity for the acetalization reac-
tion was recorded over hierarchical MFI zeolites which could enable
Fig. 9. Conversion of glycerol (G C), selectivity to products: solketal (S S) and isomer
(S I), yield of solketal (Y S) over parent and modified BEA zeolite with Si/Al ratio equal
the utilization of commonly used MFI zeolites in an industrially
to 12. relevant process of glycerol transformation.

References

[1] C.X.A. da Silva, C.J.A. Mota, Biomass Bioenergy 35 (2011) 3547–3551.

[2] C.F.M. Pestana, A.C.O. Guerra, G.B. Ferreira, C.C. Turci, C.J.A. Mota, J. Braz.
Chem. Soc. 24 (2013) 100–105.
[3] A. Alhanash, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. A: Gen. 378
(2010) 11–18.
[4] S. Wang, K. Yin, Y. Zhang, H. Liu, Catalysis 3 (2013) 2112–2121.
[5] N. Dimitratos, F. Porta, L. Prati, Appl. Catal. A: Gen. 291 (2005) 210–214.
[6] B. Katryniok, H. Kimura, E. Skrzyńska, J.S. Girardon, P. Fongarland, M. Capron,
R. Ducoulombier, N. Mimura, S. Paul, F. Dumeignil, Green Chem. 13 (2011)
1960–1979.
[7] R.R. Pawar, S.V. Jadhav, H.C. Bajaj, Chem. Eng. J. 235 (2014) 61–66.
[8] L. Li, T.I. Korányi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012)
1611–1619.
[9] C.N. Fan, C.H. Xu, C.Q. Liu, Z.Y. Huang, J.Y. Liu, Z.X. Ye, React. Kinet. Mech.
Catal. 107 (2012) 189–202.
Fig. 10. Conversion of glycerol (G C), selectivity to products: solketal (S S) and isomer [10] C.J.A. Mota, C.X.A. da Silva, N. Rosenbach, J. Costa Jr., F. da Silva, Energy Fuel 24
(S I), yield of solketal (Y S) over parent and modified MOR zeolite with Si/Al ratio (2010) 2733–2736.
equal to 17. [11] A.R. Trifoi, P.Ş. Agachi, T. Pap, Renew. Sustain. Energy Rev. 62 (2016) 804–814.
212 J. Kowalska-Kus et al. / Journal of Molecular Catalysis A: Chemical 426 (2017) 205–212

[12] P.S. Reddy, P. Sudarsanam, B. Mallesham, G. Raju, B.M. Reddy, J. Ind. Eng. [24] R. López-Medina, I. Sobczak, H. Golinska-Mazwa, M. Ziolek, M.A. Bañares,
Chem. 17 (2011) 377–381. M.O. Guerrero-Pérez, Catal. Today 187 (2012) 195–200.
[13] S. Sandesh, A.B. Halgeri, G.V. Shanbhag, J. Mol. Catal. A: Chem. 401 (2015) [25] L. Li, T.I. Korányi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012)
73–80. 1611–1619.
[14] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martín, Green Chem. 12 [26] D. Verboekend, S. Mitchell, M. Milina, J.C. Groen, J. Pérez-Ramírez, J. Phys.
(2010) 899–907. Chem. C 115 (2011) 14193–14203.
[15] J. Deutsch, A. Martin, H. Lieske, J. Catal. 245 (2007) 428–435. [27] Y. Wang, T. Yokoi, S. Namba, T. Tatsumi, Catalysis 6 (8) (2016) 1–19.
[16] C.X.A. da Silva, V.L.C. Gonçalves, C.J.A. Mota, Green Chem. 11 (2009) 38–41. [28] J.C. Groen, T. Sano, J.A. Moulijn, J. Pérez-Ramírez, J. Catal. 251 (2007) 21–27.
[17] P. Manjunathan, S.P. Maradur, A.B. Halgeri, G.V. Shanbhag, J. Mol. Catal. A: [29] A.N.C. van Laak, R.W. Gosselink, S.L. Sagala, J.D. Meeldijk, P.E. de Jongh, K.P. de
Chem. 396 (2015) 47–54. Jong, Appl. Catal. A: Gen. 382 (2010) 65–72.
[18] N. Narkhede, A. Patel, Appl. Catal. A-Gen. 515 (2016) 154–163. [30] J.C. Groen, J.A. Moulijn, J. Pérez-Ramírez, J. Mater. Chem. 16 (2006) 2121–2131.
[19] H. Serafim, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Chem. Eng. J. [31] J.C. Groen, Sònia Abelló, Luis A. Villaescusa, Javier Pérez-Ramírez,
178 (2011) 291–296. Microporous Mesoporous Mater. 114 (2008) 93–102.
[20] M.D. González, Y. Cesteros, P. Salagre, Microporous Mesoporous Mater. 144 [32] F. Tian, Y. Wu, Q. Shen, X. Li, Y. Chen, C. Meng, Microporous Mesoporous
(2011) 162–170. Mater. 173 (2013) 129–138.
[21] J. Kowalska-Kus, A. Held, K. Nowinska, React. Kinet. Mech. Catal. 117 (2016) [33] D. Verboekend, J. Pérez-Ramírez, Catal. Sci. Technol. 1 (2011) 879–890.
341–352. [34] F.D.L. Menezes, M.D.O. Guimaraes, M.J. da Silva, Ind. Eng. Chem. Res. 52 (2013)
[22] L. Zhao, J. Gao, C. Xu, B. Shen, Fuel Process. Technol. 92 (2011) 414–420. 16709–16713.
[23] M.M. Mohamed, T.M. Salama, I. Othman, I.A. Ellah, Microporous Mesoporous
Mater. 84 (2005) 84–96.
 Molecular Catalysis 434 (2017) 184–193

Contents lists available at ScienceDirect

Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat

Solvent-free microwave-assisted synthesis of solketal from glycerol

using transition metal ions promoted mordenite solid acid catalysts
Samudrala Shanthi Priya a,b,d , P.R. Selvakannan a , Komandur V.R. Chary b ,
Mannepalli Lakshmi Kantam c , Suresh K. Bhargava a,∗
a
Centre for Advanced Materials & Industrial Chemistry (CAMIC), School of Applied Sciences, RMIT University, GPO BOX 2476, Melbourne 3001, Australia
b
Inorganic and Physical Chemistry Division, CSIR – Indian Institute of Chemical Technology, Uppal Road, Hyderabad 500007, India
c
Department of Chemical Engineering, Institute of Chemical Technology, Matunga (E), Mumbai 400019, India
d
Department of Chemical Engineering, Faculty of Engineering, Monash University, Melbourne 3168, Australia

a r t i c l e i n f o a b s t r a c t

Article history: The present study demonstrates a clean and green approach for the production of bio-fuel additives
Received 14 November 2016 from renewable glycerol acetalization using microwave irradiation as heating method. Various transition
Received in revised form 20 January 2017 metal ion (M = Fe-, Co-, Ni-, Cu- & Zn-) promoted mordenite solid acid catalysts were prepared by wet
Accepted 1 March 2017
impregnation method and examined for the acetalization of glycerol with acetone. Investigation of the
Available online 24 March 2017
structural and chemical properties of the catalysts was achieved using XRD, FT-IR, TGA-TPD, SEM/EDX and
XPS techniques. Comparative studies using conventional heating were performed in order to study the
Keywords:
efficiency of microwave-assisted acetalization reactions. The effect of various operating parameters on the
Solketal
Glycerol
catalytic activity studies was examined. Amongst the metal-promoted mordenite catalysts tested, Cu-Mor
Mordenite had the best performance due to the presence of a large amount of acidic sites and the synergetic effects
Acetalization of metal particles interacting with mordenite. Microwave-assisted acetalization of glycerol into Solketal
Microwave irradiation was found to be an energy efficient route for glycerol valorization achieving 98% selectivity to solketal at
95% glycerol conversion during a reaction time of 15 min over Cu-Mor catalyst and acetone/glycerol at a
molar ratio of 3:1. Reuse of spent catalyst showed good repeatability for up to four reaction cycles with
a marginal decrease in conversion.
© 2017 Published by Elsevier B.V.

1. Introduction ing sustainable substitute to diesel fuels and its usage has already
been extended to various parts of the world. It can be used in its
The utilization of non-renewable energy sources for energy and purest form in existing diesel engines or can be easily blended
fuel production is not a sustainable approach in view of diminish- with petroleum diesel. Unlike petroleum-based diesel, biodiesel
ing fossil fuel resources and increased emission of greenhouse gases has certain unique advantages as it is biodegradable, nontoxic emit-
[1]. In order to achieve a sustainable biosphere, the use of renew- ting less air pollutants/greenhouse gases, relatively economical,
able resources as replacements for fossil fuels is more important enhanced lubricating property and combustible properties highly
and researchers have focussed their attention on the exploration of compatible to diesel engines [5]. Yet, biodiesel suffers from the
renewable energy sources for future energy needs. In this regard, problem of engine clogging and found to be less appropriate for its
Biomass was identified as an outstanding renewable source and a use in low temperature regions. On the other hand, the increased
potential alternative to fossil fuels for production of fuels, fuel addi- use of biodiesel is hindered by the large amount of glycerol obtained
tives and fine chemicals [2,3]. Biodiesel is such commonly used as a co-product in the biodiesel production which would greatly
renewable fuels produced by the transesterification of biomass affect the cost affordability of biodiesel processes [6,7]. Therefore,
derived plant oils or animal fats with methanol [4]. finding and developing feasible methods to dispose or effectively
As petro diesel is currently a major requisite for nearly all use the excess glycerol in value adding chemical processes is vital
kinds of transport means, Biodiesel stands as the most promis- and may possibly lessen both environmental contamination and
profitable losses (Scheme 1).
Glycerol (1,2,3-propanetriol) is a versatile bio feedstock and
∗ Corresponding author.
has been emerged as a fascinating platform molecule for the syn-
E-mail address: [email protected] (S.K. Bhargava).
thesis of biofuel additives and variety of chemical intermediates

http://dx.doi.org/10.1016/j.mcat.2017.03.001
2468-8231/© 2017 Published by Elsevier B.V.
 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193 185

Scheme 1. Biodiesel production through transesterification process.

Scheme 2. Acetalization of glycerol with acetone.

due to its non-toxic and multifunctional structure [8]. To this end, Several heterogeneous solid acid catalysts including sulphonic acid
extensive research has been developed for up gradation of glycerol mesostructured silicas [22], supported heteropolyacids [23], pro-
by various catalytic transformations to valuable chemicals. Sev- moted zirconia [24], supported metal oxides [25], zeolites [26], and
eral strategies including etherification [9], hydrogenolysis [10,11], promoted metal phosphates [27] have been tested in the glycerol
dehydration [12], acetalization [13], esterification [14], oxidation acetalization reaction. Although proven to be efficient, their activ-
[15] and steam reforming [16] have been explored. In the recent ity is still hampered by poor thermal stability, low surface area,
past, special interest was focussed on synthesis of acetals and ketals solubility in polar solvents, reduced regeneration ability and use of
by the acid catalyzed acetalization of glycerol with aldehydes and hazardous solvents and high temperatures. Therefore, to achieve
ketones. These are commercially important branched oxygenated a sustainable and economical process of glycerol acetalization, it
compounds finding numerous applications as disinfectants, fuel is indispensable to develop highly efficient, inexpensive and ther-
additives, flavouring agents, cosmetics, surfactants, pharmaceuti- mally stable catalysts.
cals, food and beverage industries [17]. In particular, the glycerol In this intellect, protonic zeolites seem to be the most promising
acetalization with acetone gives 5-membered and 6-membered and widely applied acid catalysts in the chemical industry. In addi-
cyclic isomeric compounds namely (2,2-dimethyl-[1,3] dioxane- tion to presenting the apparent benefits of heterogeneous catalysts
4-yl)-methanol (also known as Solketal) and 2,2-dimethyl-[1,3] like reusability and simple recovery, protonic zeolites exemplify
dioxane-5-ol respectively (Scheme 2). eco-friendly solid acid catalysts unlike conventional liquid acids.
Solketal is a potential fuel additive and an anti-knocking agent These are aluminosilicates characterized by uniform microporous
which aids in enhancing the octane number and leads to a signifi- crystalline structure, shape selectivity, hydrophobicity and high
cant decrease in harmful emissions when blended with diesel fuel thermal stability. It is ascertained that the stronger acidity of
[18]. It is a chief constituent used in the preparation of gasoline, and protonic zeolites is consistent with the bridging hydroxyl groups
biodiesel fuels that effectively reduces the viscosity, enhances the Al (OH) Si located in the cavities of zeolite [28]. H-Mordenite is
cold flow properties besides attaining the flash point and oxidation one such protonic zeolite found to hold excellent structural and tex-
stability of biodiesel during extended storage [19]. In a recent report tural properties that are not seen in other catalytic materials. Unlike
by Garcia et al. [20] it was established that the viscosity and other other zeolites, even though mordenite is a microporous material, it
properties of biodiesel improved significantly by the use of fuel possess a special multiple pore channel system where its elliptical
additives. In search of new bio-degradable, non-toxic and renew- pores are wide enough to allow many reactions to be attained with
able fuel additives for biodiesel, the synthesis of solketal from bio high selectivities.
derived glycerol and its usage finds great demand. Mordenite is a strong Brønsted acidic zeolite, chemically com-
Conventionally, the acetalization of glycerol with acetone is per- posed of Na8 Al8 Si40 O96 ·24H2 O with a Si/Al molar ratio ≥5 and
formed over homogenous liquid acids and Lewis acid catalysts highly stable towards chemical and thermal treatments. Morden-
[13]. However, the use of homogenous acid catalysts is not cost- ite is well considered for its use in chemical sensors, optics and
effective and presents environmental harms in addition to tough semiconductors in addition to the adsorptive separation of liquid
work-up processes. Thereby, heterogeneous catalysts are well cho- or gaseous mixtures. It is extensively used as a catalyst in various
sen in view of clear benefits over homogeneous acids for example reactions such as reforming, hydroisomerization, dewaxing, hydro-
their reusability, solvent less conditions, product selectivity and cracking and alkylation owing to its surface acidity and distinctive
above all, quite easy separation of products from catalysts [21]. pore system [29,30]. The formation of water during glycerol acetal-
186 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193

Zn-Mor

Cu-Mor
Intensity (a.u)

Ni-Mor

Co-Mor

Fe-Mor

Mor

10 20 30 40 50 60 70 80 90
2 Theta Fig. 3. TGA-TPD profiles of various metal-promoted mordenite catalysts.

Fig. 1. XRD patterns of pure and metal promoted mordenite catalysts.

become increasingly popular and is proven to be more efficient
than conventional heating method because of its high speed syn-
thesis involving rapid internal heating, with a direct transfer of
Zn-Mor
microwave energy from source to the molecules in the reaction.
Microwave processing brings about uniform heating that can regu-
late the temperature and brings about considerable energy saving
Cu-Mor [32,33]. Hence, the chemical syntheses by microwave irradiation as
heating method receive significant interest over traditional heating
methods offering high conversion and selectivity.
In an effort to develop an economic viable and a green chemical
protocol, the present work focussed on solvent-free microwave-
% Transmittance

Ni-Mor assisted synthesis of solketal from glycerol acetalization with

acetone over Mordenite (Mor) and transition metal promoted
mordenite (Fe-Mor, Co-Mor, Ni-Mor, Cu-Mor & Zn-Mor) catalysts.
Various transition metal-promoted mordenite catalysts were pre-
Co-Mor pared by wet impregnation method and thoroughly investigated
for its structural, chemical and acidic properties by using vari-
ous techniques such as XRD, FTIR, SEM/EDX, XPS and TGA-TPD.
Fe-Mor These catalysts with varied properties were tested in glycerol
acetalization via microwave irradiation method in order to screen
out the most efficient catalyst. Then, an optimization study was
performed for attaining maximum conversion and selectivity by
Mor
different reaction parameters including reaction temperature, glyc-
erol concentration, glycerol/acetone mole ratio and reaction time.
In addition, a comparative study of solketal synthesis by conven-
tional heating is also described.
4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1) 2. Experimental

Fig. 2. FTIR spectra of pure and metal promoted mordenite catalysts.

2.1. Catalyst preparation

ization reaction is problematic and has to be addressed in order to A series of transition metal promoted mordenite catalysts (M-
hinder the reverse reaction. The hydrophobic property of the zeolite Mor) were prepared by wet impregnation method. H-Mordenite
assists in inhibiting the reverse reaction of acetalization by diffusion with SiO2 /Al2 O3 ratio 20 and a surface area of ∼540 m2 /g was
of water in to the pores of zeolite. da Silva et al. [31] reported the supplied from Conteka, Netherlands. In a typical procedure, the
same phenomenon in beta zeolite catalyzed glycerol acetalization required amounts of 1 wt% metal nitrate precursors (M(NO3 )2 ;
reaction with acetone at 70 ◦ C. The modification of physiochemical M = Fe, Co, Ni, Cu & Zn) were dissolved in millipore water and
properties of zeolites to further improve their efficiency in reac- allowed to stir for 10 min to obtain uniform suspension. This was
tions can be achieved by the incorporation of metal atoms into the followed by the addition of finely powdered H-mordenite to the
framework or extra framework of zeolite structure. aqueous metal nitrate solution under stirring. The sample obtained
Most of the liquid phase reactions over heterogeneous cata- after evaporation of excess water was oven dried at 110 ◦ C for 15 h
lysts were conducted by direct heating or in oil bath which is and finally calcined in air at 550 ◦ C for 4 h. The as-prepared morden-
an inefficient and slow process in terms of energy transfer from ite supported metal catalysts are herein labelled as Fe-Mor, Co-Mor,
source to the reactants. Microwave-assisted chemical synthesis has Ni-Mor, Cu-Mor and Zn-Mor.
 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193 187

Fig. 4. O 1s, Al 2p and Cu 2p core level X-ray photoelectron spectra of metal promoted mordenite catalysts.

Fig. 5. SEM images of (a) Mor and (b) Cu-Mor catalysts.

2.2. Characterization 2.3. Reaction procedure

X-ray diffraction (XRD) studies were performed on a Bruker D8 2.3.1. Microwave-assisted reaction
X-ray diffraction system and the measurements were conducted in The glycerol acetalization reaction was carried out in
the range of 2 10.0◦ to 90.0◦ using a voltage of 40 kV and current CEM Discover-S (NP1009) high pressure and high temperature
of 40 mA with CuKa radiation. FTIR spectra of the catalysts were microwave reactor equipped with magnetic stirring and temper-
recorded on a Perkin-Elmer Spectrum 100 spectrophotometer, in ature cooling system. The mixture of glycerol (1 mmol), acetone
diffuse reflectance mode with the spectral resolution of 4 cm−1 . (3 mmol) and 40 mg of the catalyst were taken in a reactor ves-
SEM/EDX studies of the catalysts were obtained by a FEI nova nano sel fitted with a PTFE-lined cap and is irradiated under high power
SEM operated at 30 kV accelerating voltage. XPS data of M-Mor (100–600 W) to reach reaction temperature of 100 ◦ C at a heating
catalysts was collected using a Thermo k-Alpha XPS instrument rate of 10 ◦ C/min. The reactor vessel was run under continuous
with core levels aligned with C 1s binding energy of 284.6 eV and a stirring for 15 min in the microwave reactor. The reactor ves-
pressure of 1 × 10−9 torr. sel was then allowed to cooling and the reaction products were
TGA-TPD experiments for mordenite catalysts were conducted extracted and filtered by separating the used catalyst by centrifu-
using ammonia as a base molecule. Initially, the catalyst samples gation. The products were analyzed on HP5890 gas chromatograph
were pretreated in a flow of nitrogen atmosphere at 550 ◦ C at (GC) equipped with a flame ionization detector (FID). The products
10 ◦ C/min in order to remove adsorbed water and other volatiles were analyzed and confirmed by GC-MS.
completely. This is followed by the adsorption of ammonia on the
catalyst samples for about 30 min. The sample was then heated
to 700 ◦ C at a rate of 10 ◦ C/min under the flow of nitrogen, and 2.3.2. Room temperature reaction
the ammonia desorption was continuously recorded as a function A mixture of glycerol (1 mmol) and acetone (3 mmol) was sub-
of temperature. NH3 adsorbed catalysts Thermo gravimetric data jected to vigorous stirring in a RB flask at room temperature in
(TG) were collected using a Perkin Elmer TGA-7 from 35 ◦ C to 700 ◦ C the absence of any solvent. After 5 min, the catalyst (40 mg) was
with a heating rate of 10 ◦ C per minute under the flow of nitrogen. added and the reaction time started. The products were analyzed
and confirmed by GC-MS and quantitative analysis was analyzed
on a HP5890 gas chromatograph (GC) equipped with a flame ion-
188 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193

Fig. 6. SEM/EDX images and spectra of Cu-Mor catalyst.

ization detector (FID). The conversion of glycerol and selectivity of presented in Fig. 2. All the samples reveal the typical vibrational
solketal were calculated based on the following equations: bands of mordenite which agree with the previous reports [36]. The
moles of glycerol (ini) − moles of glycerol (Final)
FTIR spectra of pure mordenite exhibit two stretching vibrational
Glycerol Conversion(%) =
moles of glycerol (in)
× 100 peaks at 3744 cm−1 and 3605 cm−1 associated with the terminal
silanol groups and the bridging Si-OH-Al groups respectively. These
moles of solketal formed hydroxyl groups located on the inner surface of the mordenite
Solketal Selectivity(%) = × 100 structure can be attributed to stronger Brønsted acidity [37].
moles of glycerol converted
Furthermore, a weak band observed around 3655 cm−1 corre-
sponds to the extra framework aluminium (AlOH ) species. H OH
3. Results and discussion bending vibrational bands were observed at 1636 cm−1 and bands
due to Si O stretching vibrations were found at 1040 cm−1 and
3.1. Characterization results 1198 cm−1 . The pure mordenite revealed a vibrational band at
1480 cm−1 assigned to OH vibrations and this band appeared
3.1.1. X-ray diffraction studies to disappear gradually after metal impregnation suggesting the
XRD patterns of pure mordenite and transition metal promoted breaking of OH bond to form OM. Additionally, the bands
mordenite catalysts (Fe-Mor, Co-Mor, Ni-Mor, Cu-Mor and Zn-Mor) observed at 918 cm−1 , 585 cm−1 , 520 cm−1 and 468 cm−1 resulted
revealed the presence of characteristic peaks of mordenite zeolite at from Al O Al bending vibration, coupling vibration of Al O
2 = 9.76◦ , 13.52◦ , 19.70◦ , 22.42◦ , 25.73◦ , 26.44◦ , and 27.64◦ which and Si O, Al-O stretching vibration and Si O bending vibrations
is similar to the respective mordenite diffraction peaks found in the respectively [38]. Combined with XRD results, the FTIR spectra
literature [34,35]. As shown in Fig. 1, all of the samples presented below shows that the structure of mordenite did not appreciably
similar crystallinity and no significant differences were observed appear to be altered after metal impregnation.
in the mordenite structure after metal impregnation. Consequently
the absence of metal/metal oxide peaks in the diffraction patterns 3.1.3. TGA-TPD studies
suggests that these metal ions were highly dispersed on the mor- The acid site distribution and relative acidic strength of a zeo-
denite surface. This confirms that all of the transition metal ions lite catalyst is the characteristic feature used to obtain a correlation
promoted mordenite materials did not alter the structure of the with its catalytic activity. The temperature programmed desorp-
zeolite and retain the high crystallinity of mordenite. tion (TPD) of a base (ammonia) molecule from the surface of zeolite
catalyst is one efficient technique to measure the acidity of cata-
3.1.2. Fourier transform-infrared spectra lysts. Thermo gravimetric analysis (TGA) of NH3 adsorbed zeolites
The structure of mordenite and the presence of surface hydroxyl is a convenient method used for characterization of TPD profiles
groups has been studied by Fourier transform infrared (FTIR) spec- of zeolites [39]. TGA-TPD studies were performed to identify the
troscopy in the range of 4000–500 cm−1 and the spectral results desorption of ammonia from metal promoted mordenite catalysts
of pure mordenite and metal promoted mordenite catalysts are as a function of temperature and the resulting data was used to
 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193 189

HO OH O O
+ Cu-Mor
OH
100 oC, ~500W OH
Glycerol Acetone 15 min Solketal
95% conversion 98 % selectivity

Scheme 3. Microwave-assisted glycerol acetalization over Cu-Mor catalyst.

Table 1 tively [41]. These three O 1s components were assigned to the

TGA-TPD acidity measurements of various mordenite catalysts.
surface oxides, hydroxides and bulk oxide respectively.
Catalyst Total acidity (NH3 mmol/g) In all cases, the O 1s spectral profile showed significant broad-
HM 7.3 ening as compared to that of pure mordenite. However, in the case
Fe-Mor 5.1 of Cu-Mor catalyst, it was observed that the O 1s component cor-
Co-Mor 5.5 responding to the surface oxide species has been shifted to lower
Ni-Mor 6.2 binding energies. It is known that the surface oxides and hydroxides
Cu-Mor 6.7
significantly contribute to the total acidity of the catalysts, and the
Zn-Mor 6.4
O 1s spectra provide a clear indication of the modification of mor-
denite by transition metal ions. Another major change observed
depict the acid site distribution and strength of acid sites present was related to Al 2p core levels after Mor was modified with tran-
in various mordenite catalysts. sition metal ions. Pure mordenite exhibited Al 2p binding energy
The TGA-TPD profiles of all the samples are presented in Fig. 3 at 74.3 eV, however it was shifted to 74.8 eV in the case of Cu-Mor
and the amount of acid sites in millimoles per gram of catalyst catalyst. All the other metal-modified mordenite catalysts showed
obtained from the peak minimum of TGA desorption curves are significantly broader Al 2p spectra without much change in Al 2p
listed in Table 1. As can be seen from Fig. 3, the TGA-TPD curves binding energies.
showed a gradual weight loss in two different regions. The first Cu 2p core level binding energy in Cu-Mor catalyst was observed
weight loss region at lower temperatures (I) corresponds to the at 932.8 eV without any satellite features, suggesting the formation
presence of weak acid sites due to the desorption of weakly bound of Cu2 O species formation on the surface of mordenite [42]. Inter-
ammonia molecules and the second weight loss region found at estingly, both O 1s and Al 2p core level spectra were broadened
higher temperatures (II) is attributed to the desorption of strongly in all the metal modified mordenite catalysts whereas the Cu-Mor
bound ammonia molecules [40]. catalyst showed a shift in binding energy. This is a clear indication
The total acidity values corresponding to pure mordenite (HM) of a metal interaction with the acidic sites of mordenite particularly
and various metal-promoted mordenite (Fe-Mor, Co-Mor, Ni-Mor, with the Lewis acidic site of Al3+ . The XPS analysis clearly revealed
Cu-Mor, Zn-Mor) catalysts are presented in Table 1, calculated by that Cu-Mor catalyst demonstrated a strong interaction with mor-
using the following equation: denite support and the results are in corroboration with TGA-TPD
studies.
 % Weight loss   1
  100 
mg
Acidity =
100 MW of NH3 1g
3.1.5. SEM/EDX studies
The results from Fig. 3 and Table 1 illustrates that the total Scanning electron microscopic (SEM) analysis was used to study
acidity of pure mordenite is comparatively higher than the metal- the morphological changes of Mor before and after Cu ions modifi-
promoted mordenite catalysts which is probably due to the cation. The pure mordenite sample (Fig. 5a) exhibited needle-like
blockage of acid sites by metal oxides impregnation on the morden- crystalline morphology, typical characteristic of mordenite zeolites
ite surface. This demonstrates that the introduction of transition [43]. As seen in Fig. 5b, the Cu-Mor catalyst occur in clusters of
metal ions substantially modified the inherent acidity of the closely spaced crystals in the form of fibrous aggregates, and the
catalysts. However, among various metal promoted mordenite filaments commonly intertwine into packed mesh like structures.
catalysts, Cu-Mor catalyst exhibited higher affinity towards NH3 EDX studies were performed to determine the elemental com-
and showed higher total acidity. This indicates that Cu-Mor cat- positions of the Cu-Mor catalyst. The images and spectra generated
alyst possess large number of acidic sites compared to other as a result of elements in the catalyst are presented in Fig. 6. EDX
metal-promoted catalysts and can demonstrate a good correlation analysis showed the presence of major elements such as Si, Al, O,
between the acidity and catalytic activity. and C in the catalyst [44]. In addition, copper is also identified in the
Cu-Mor sample which clearly revealed the distribution of copper
on the surface of mordenite.
3.1.4. XPS measurements
TGA results clearly demonstrated that addition of transition
metal ions significantly altered the total acidic sites, therefore it 3.2. Catalytic activity results
is essential to investigate on how the metal oxides can modify the
acidic sites. In metal ions-promoted mordenite catalysts, the tran- 3.2.1. Screening of catalysts
sition metal ions either interact with the Brønsted or Lewis acidic Initially, the screening of different transition metal ions pro-
sites. Therefore XPS analysis of these catalysts was carried out to moted mordenite solid acid catalysts (Fe-Mor, Co-Mor, Ni-Mor,
understand the nature of interaction (Fig. 4). Pure Mordenite has Cu-Mor and Zn-Mor) was performed in order to investigate their
shown three chemically distinct O 1s components of having the efficiency towards glycerol acetalization with acetone to produce
peak binding energies of 532.6 eV, 533.2 eV and 531.6 eV respec- solketal. The reactions were carried out using microwave irradi-
190 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193

120

Conversion & Selecvity (%)

100

80

60

40
Gly Conv (%) Solketal Sel (%)
20

0
40 60 80 100 120
Reacon Temperature (oC)

Fig. 7. Glycerol acetalization over various mordenite catalysts. Fig. 8. Effect of temperature on glycerol acetalization over Cu-Mor catalyst.

Table 2
ation as a heating method under the same reaction conditions as Effect of glycerol/acetone molar ratio on glycerol acetalization over Cu-Mor catalyst.
follows: 40 mg catalyst, glycerol/acetone ratio is 1:3, reaction tem-
Glycerol/acetone Glycerol Solketal
perature 100 ◦ C and a reaction time of 15 min. The influence of mole ratio conversion (%) selectivity (%)
different catalysts on the conversion of glycerol to solketal was
1:1 81 79
comparatively analyzed and correlated with catalytic properties.
1:2 86 88
Castanheiro et al. [45] investigated glycerol acetalization over dif- 1:3 95 98
ferent zeolites (USY, BEA and ZSM-5) and found that the catalytic 1:4 90 98
activity was mostly influenced by the pore structure and acidity
of catalysts. All zeolites tested presented high selectivity to five-
membered ring solketal at about 77–82% (Scheme 3). power and reaction time were varied to improve the glycerol con-
Fig. 7 shows the results of glycerol conversion and selectivity version and solketal selectivity.
to solketal in acetalization reaction attained by all catalysts. 1,3-
dioxalane (5-membered, solketal) was obtained as a major product 3.2.2.1. Reaction temperature. Different experiments were carried
with high selectivity than the 1,3-dioxane (6-membered) over pure out by varying the reaction temperature in the microwave from
mordenite as well as metal promoted mordenite catalysts. This is 40 ◦ C to 120 ◦ C, in order to evaluate its effect on the acetalization of
because, the five membered ring 1,3-dioaxalne product is kinet- glycerol with acetone over Cu-Mor catalyst. The results of effect of
ically more favourable than the six membered ring 1,3-dioaxane reaction temperature on glycerol conversion and on the selectivity
product. It is to be noted that the six membered product although to solketal are shown in Fig. 8. As anticipated, the glycerol conver-
thermodynamically more stable, it is kinetically less favourable as sion increased with the temperature and reached maximum (95%)
one of the methyl groups occupy axial position. Hence glycerol at 100 ◦ C (Fig. 8). It was observed that when the acetalization of
acetalization with acetone favours the formation of five-membered glycerol with acetone was carried out at 40 ◦ C the selectivity to
ring transition state which leads to the formation of solketal. Similar solketal was found to be 88% and increased to 98% at 100 ◦ C. When
findings were reported in the previous studies [13,45,46]. the reaction temperature is further increased to 120 ◦ C, there was
The glycerol acetalization over pure mordenite resulted in 66% a slight decrease in the solketal selectivity although the glycerol
glycerol conversion with 64% and 36% selectivity to 5-M and 6- conversion remained constant. Therefore 100 ◦ C was chosen as the
M products respectively. Interestingly, over metal ions promoted optimum reaction temperature in microwave assisted synthesis of
mordenite catalysts, the glycerol conversion and solketal selec- glycerol acetalization.
tivity increased significantly with a decline in the selectivity of
1,3-dioxane (6-membered). All the metal promoted mordenite (Fe- 3.2.2.2. Molar ratio of glycerol to acetone. The ratio or amount of the
Mor, Co-Mor, Ni-Mor, Cu-Mor and Zn-Mor) catalysts exhibited reactants in a chemical reaction plays a crucial role in controlling
appreciable activities of glycerol conversion (86–95%) and solketal the product formation and hence it is imperative to optimize the
selectivity (81–98%). A blank reaction without using any catalyst reaction by using the precise molar ratio of reactants to achieve bet-
was also performed which showed about 8% glycerol conversion ter conversion and selectivity of products. In order to investigate
and no selectivity towards solketal. This indicates that catalyst the optimum glycerol:acetone molar ratio required for glycerol
mediated glycerol acetalization reaction provides high efficiency acetalization, reactions were carried out by using various propor-
glycerol conversion and product selectivity. The highest glycerol tions of glycerol:acetone such as 1:1, 1:2, 1:3 and 1:4 at 100 ◦ C.
conversion of 95% was obtained over Cu-Mor catalyst with a solke- Table 2 presents results of the effect of molar ratio of glycerol to
tal selectivity of 98%. This result can be explained based on the acetone on glycerol conversion and solketal selectivity. The results
greater acidity of the Cu-Mor catalyst and better interaction of Cu clearly show that glycerol conversion and selectivity to solketal
with the mordenite support which influenced the progression of gradually raised up by increasing the glycerol:acetone molar ratio
the reaction. from 1:1 to 1:3. However, when the molar ratio of glycerol to ace-
tone increases from 1:3 to 1:4, there was no further increase in
3.2.2. Reaction optimization solketal selectivity but a slight fall in the glycerol conversion was
Several reaction parameters were surveyed to achieve the opti- noticed. This can be ascribed to the saturation of acetone over the
mized conditions in microwave assisted synthesis of solketal from active sites available for glycerol to react. Similar outcomes were
glycerol acetalization over the best performed Cu-Mor catalyst. observed in the literature [22,47].
The reaction temperature, glycerol/acetone molar ratio, microwave
 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193 191

Fig. 11. Effect of reaction time on glycerol acetalization over Cu-Mor catalyst.

Fig. 9. Effect of microwave power on glycerol acetalization over Cu-Mor catalyst.

Fig. 12. Glycerol acetalization at room temperature over Cu-Mor catalyst.

3.2.2.5. Reaction time. The reaction was monitored at different

reaction times from 5 min to 20 min and Fig. 11 shows the results
of effect of reaction time over Cu-Mor catalyst at 100 ◦ C and 500 W
Fig. 10. Effect of catalyst amount on glycerol acetalization over Cu-Mor catalyst.
microwave power with a molar ratio of glycerol to acetone 1:3.
After 5 min of reaction time, the reaction resulted in 80% conversion
of glycerol with 70% selectivity to solketal. Gradually, with increase
3.2.2.3. Microwave power. The glycerol acetalization was system-
in reaction time to 15 min, the maximum glycerol conversion (95%)
atically studied using microwave power ranging from 100 to 600 W
and solketal selectivity (98%) was achieved. Further increase in the
at 100 ◦ C and with 1:3 glycerol/acetone molar ratio. As shown in
reaction time to 20 min did not result in significant increase in the
Fig. 9, a noticeable increase in the glycerol conversion was observed
activity, hence 15 min was considered to be the optimum reaction
with increase in the microwave power from 100 to 600 W. The
time in the present study.
selectivity to solketal was found to be considerably high in all
experiments and attained 98% at 500 W microwave power. This
is ascribed to the highly accelerated reaction by the microwave 3.2.3. Glycerol acetalization at room temperature
irradiation. The glycerol acetalization with acetone was carried out at room
temperature using 1:3 glycerol/acetone ratio and 50 mg catalyst.
The reaction proceeded very slowly and reached 75% glycerol con-
3.2.2.4. Effect of catalyst amount. The effect of catalyst amount on
version and 80% solketal selectivity after 2 h reaction time. This is
glycerol conversion and solketal selectivity during acetalization of
possibly due to the poor interaction of catalyst active sites with
glycerol was studied by varying the amount of catalyst from 20 mg
the reactant molecules in the reaction to accelerate the reaction.
to 60 mg at 100 ◦ C, 500 W and 1:3 glycerol/acetone molar ratio. As
The results of glycerol acetalization at room temperature are pre-
presented in Fig. 10, the glycerol conversion and solketal selectiv-
sented in Fig. 12. In contrast, microwave irradiation enhanced the
ity increased from 90% to 95% and 92% to 98% respectively, with
activity to a greater extent and resulted in high glycerol conversion
increase in the catalyst amount from 20 mg to 40 mg, which is due
(95%) and solketal selectivity (98%) within 15 min of reaction. The
to the availability of greater number of active sites in the catalyst.
rapid internal heating of reactants by microwave energy was highly
It is observed that there is no further variation in the conversion
efficient to promote the active interactions between the reactant
and selectivity beyond that weight of catalyst amount and hence
molecules and the active sites on the catalyst surface. Microwave
40 mg of catalyst is considered to be the optimum amount for the
irradiation process was highly efficient and accelerated the reac-
reaction.
tion rate by remarkably decreasing the reaction time from hours to
minutes [48]. Therefore microwave-assisted synthesis of solketal
192 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193

Scheme 4. Plausible reaction mechanism of glycerol acetalization over metal promoted mordenite catalysts.

from glycerol acetalization over Cu-Mor catalyst was found to be a 3.2.5. Recyclability studies
promising route over conventional methods. In order to investigate the catalyst stability, the recyclability
studies of glycerol acetalization were performed over Cu-Mor cat-
alyst. The spent catalyst was recovered by simple centrifugation
and washed with water several times to remove the impurities
3.2.4. Proposed reaction mechanism
adsorbed on the catalyst surface. It is then oven dried at 100 ◦ C
Previous literature shows that the majority of glycerol acetal-
for 2–3 h and the catalyst was reused directly for glycerol acetal-
ization reactions with different carbonyl compounds have been
ization reaction in microwave under the same optimized reaction
promoted by solid acid catalysts [21,22,49]. Accordingly, the
conditions used for the fresh catalyst [50]. The reused catalyst sam-
present study demonstrates the glycerol acetalization reaction over
ples were also subjected to XRD, FT-IR, and TGA-TPD analyses. XRD
acidic zeolite catalysts. The possible reaction pathway for solketal
and FTIR results did not show any significant changes in features
formation through acetalization of glycerol with acetone over metal
of used catalyst compared to fresh catalyst. However, there was a
promoted mordenite catalysts is shown in Scheme 4. The acidic
slight decline in the acidity of spent sample after the fourth reaction
sites of metal promoted mordenite catalysts aided in the effective
cycle indicating that there could be slight deactivation of Cu-Mor
acetalization of glycerol with acetone. Initially, the oxygen atom of
catalyst. Fig. 12 shows the catalytic activity results of glycerol con-
the carbonyl group of acetone strongly coordinates with the cata-
version and solketal selectivity for five reaction cycles. The results
lyst generating a positively charged carbonyl carbon which is then
show that the catalyst remained stable with no significant change
attacked by a lone pair on oxygen atoms of glycerol and results in
in the glycerol conversion and selectivity of solketal up to four reac-
the formation of tertiary alcohol intermediate. In the next step, the
tion cycles whereas a slight decline in activity was observed after
proton released from acidic sites of the catalyst facilitates the proto-
fourth reaction cycle of glycerol acetalization. The observed drop in
nation of the intermediate. Finally, the other nucleophilic hydroxyl
the activity could be attributed to the partial deactivation of catalyst
group of glycerol attacks the tertiary carbon of the intermediate
due to the blockage of acidic sites (Fig. 13).
with simultaneous elimination of water molecule. This affords 5-
membered and 6-membered isomeric cyclic products namely (2,
2-dimethyl-[1,3] dioxane-4-yl)-methanol (also known as Solketal) 4. Conclusions
and 2, 2-dimethyl-[1,3] dioxane-5-ol (acetal) respectively. When
secondary hydroxyl group is involved, a 5 membered solketal is In summary, a sustainable method of solketal synthesis from
formed and a 6 membered acetal product is obtained when pri- acetalization of glycerol using microwave irradiation is presented.
mary hydroxyl group of glycerol is involved in the attack. The Mordenite and transition metal-promoted mordenite catalysts
5-membered cyclic product (solketal) was found to be the major were found to be highly efficient solid acid catalysts for glycerol
product in the present study which is in accordance with earlier acetalization. Cu-Mor catalyst exhibited excellent activity among
reports. all the tested catalysts and achieved 98% selectivity to solketal at
 S.S. Priya et al. / Molecular Catalysis 434 (2017) 184–193 193

Glyc Conv (%) Solk Sel (%) Acidity (mmol/g) [13] S.B. Umbarkar, T.V. Kotbagi, A.V. Biradar, R. Pasricha, J. Chanale, M.K. Dongare,
99 6.75 A.S. Mamede, C. Lancelot, E. Payen, J. Mol. Catal. A 310 (2009) 150–158.
[14] W.D. Bossaert, D.E. De Vos, W.M. Van Rhijn, J. Bullen, P.J. Grobet, P.A. Jacobs, J.
98 6.7 Catal. 182 (1999) 156–164.
Conversion & Selecvity (%)

[15] D.L. King, L. Zhang, G. Xia, A.M. Karim, D.J. Heldebrant, X. Wang, T. Peterson, Y.
97 6.65 Wang, Appl. Catal. B 99 (2010) 206–213.

Acidity (mmol/g)
[16] S. Demirel, M. Lucas, J. Warna, T. Salmi, D. Murzin, P. Claus, Top. Catal. 44
96 6.6 (2007) 299–305.
[17] J. Delgado, Spanish Patent 1 331 260, 2002.
[18] C.J.A. Mota, C.X.A. da Silva, N. Rosenbach Jr., J. Costa, F. da Silva, Energy Fuels
95 6.55
24 (2010) 2733–2736.
[19] B. Delfort, I. Durand, A. Jaecker, T. Lacome, X. Montagne, F. Paille, U.S. Patent 0
94 6.5 163 949, 2003.
[20] E. Garcia, M. Laca, E. Perez, Á. Garrido, J. Peinado, Energy Fuels 22 (2008)
93 6.45 4274–4280.
[21] L.W. Benjamin, S. Liu, M. Thom, D. Stanley, M.M. Abu-Omar, ACS Catal. 2
92 6.4 (2012) 2524–2530.
Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 [22] G. Vicente, J.A. Melero, G. Morales, M. Paniagu, E. Martin, Green Chem. 12
(2010) 899–907.
No. of Reacon Cycles [23] M.J. da Silva, A.A. Julio, F.C.S. Dorigetto, RSC Adv. 5 (2015) 44499–44506.
[24] P.S. Reddy, P. Sudarsanam, B. Mallesham, G. Raju, B.M. Reddy, J. Ind. Eng.
Fig. 13. Reusability of Cu-Mor catalyst for glycerol acetalization. Chem. 17 (2011) 377–381.
[25] B. Mallesham, P. Sudarsanam, G. Raju, B.M. Reddy, Green Chem. 15 (2013)
478–489.
95% glycerol conversion under the optimized reaction conditions of [26] H. Serafima, I.M. Fonsecab, A.M. Ramosb, J. Vital, J.E. Castanheiro, Chem. Eng. J.
1:3 glycerol/acetone molar ratio, 100 ◦ C, 500 W microwave power 178 (2011) 291–296.
[27] S. Zhang, Z. Zhao, Y. Ao, Appl. Catal. A 496 (2015) 32–39.
and 15 min reaction time. Compared to the room temperature syn-
[28] A. Corma, Chem. Rev. 97 (1997) 2373–2420.
thesis of solketal from glycerol acetalization, microwave-assisted [29] R.A. Asuquo, G. Eder-Mirth, A.J. Lercher, J. Catal. 155 (1995) 376–382.
solketal synthesis showed great efficacy due to the high energy [30] S.T. Sie, Stud. Surf. Sci. Catal. 85 (1994) 587–631.
[31] C.X.A. da Silva, V.L.C. Goncalves, C.J.A. Mota, Green Chem. 11 (2009) 38–41.
microwave process. The catalytic activity is well correlated with
[32] R.R. Pawar, S.V. Jadhav, C. Hari Bajaj, Chem. Eng. J. 235 (2014) 61–66.
acidity of the catalyst and Cu-Mor catalyst was found to be effi- [33] J.D. Moseleyand, C.O. Kappe, Green Chem. 13 (2011) 794–806.
ciently reusable for consecutive catalytic reaction cycles. [34] H. González, O.S. Castillo, J.L. Rico, A. Gutiérrez-Alejandre, J. Ramírez, Ind. Eng.
Chem. Res. 52 (2013) 2510–2519.
[35] F. Lonyi, A. Kovács, A. Szegedi, J. Valyon, J. Phys. Chem. C 113 (2009)
Acknowledgements 10527–10540.
[36] I.I. Ivanova, E.E. Knyazeva, Chem. Soc. Rev. 42 (2013) 3671–3688.
S.S.P. duly acknowledges RMIT-IICT joint research centre for [37] Y.P. Khitev, I.I. Ivanova, Y.G. Kolyagin, O.A. Ponomareva, Appl. Catal. 441–442
(2012) 124–135.
providing the research facilities and fellowship. The authors thank [38] M. Bevilacqua, G. Busca, Catal. Commun. 3 (2002) 497–502.
the RMIT Microscopy and Microanalysis facility staff members for [39] S. Bhatia, J. Beltramini, D.D. Do, Catal. Today 7 (1990) 309–435.
their scientific and technical assistance. [40] Thermal Analysis Applications Brief TGA Evaluation of Zeolite Catalysts,
Number TA-231, http://www.tainstruments.com/pdf/literature/TA231.pdf.
[41] S.S. Priya, P. Bhanuchander, V.P. Kumar, D. Deepa, P.R. Selvakannan, M.L.
References Kantam, S.K. Bhargava, K.V.R. Chary, ACS Sustain. Chem. Eng. 4 (2016)
1212–1222.
[1] D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Chem. Soc. Rev. 41 (2012) [42] S.S. Priya, P. Bhanuchander, V.P. Kumar, S.K. Bhargava, K.V.R. Chary, Ind. Eng.
8075–8098. Chem. Res. 55 (2016) 4461–4472.
[2] A.M. Ruppert, K. Weinberg, R. Palkovits, Angew. Chem. Int. Ed. 51 (2012) [43] M. Sakizci, L. Ozg, U.L. Tanriverdi, Turk. J. Chem. 39 (2015) 970–983.
2564–2601. [44] H.M. Aly, M.E. Moustafa, E.A. Abdelrahman, Adv. Powder Technol. 23 (2012)
[3] C.H. Zhou, X. Xia, C.X. Lin, D.S. Tong, J. Beltramini, Chem. Soc. Rev. 40 (2011) 757–760.
5588–5617. [45] H. Serafim, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Chem. Eng. J.
[4] F. Ma, M. Hanna, Bioresour. Technol. 70 (1999) 1–15. 178 (2011) 291–296.
[5] A.P. Vyas, J.L. Verma, N. Subrahmanyam, Fuel 89 (2010) 1–9. [46] P.H.R. Silva, V.L.C. Goncalves, C.J.A. Mota, Bioresour. Technol. 101 (2010)
[6] B. Katryniok, S. Paul, V. Belliere-Baca, P. Rey, F. Dumeignil, Green Chem. 12 6225–6229.
(2010) 2079–2098. [47] G. Sailaja, N.P. Rajan, G.S. Rao, K.V.R. Chary, J. Mol. Catal. A 410 (2015) 49–57.
[7] P.C. Smith, Y. Ngothai, Q.D. Nguyen, B.K. O’Neill, Renew. Energy 35 (2010) [48] R.R. Pawar, K.A. Gosai, A.S. Bhatt, S. Kumaresan, S.M. Lee, H.C. Bajaj, RSC Adv. 5
1145–1151. (2015) 83985–83996.
[8] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539–554. [49] P. Manjunathan, S.P. Maradur, A.B. Halgeri, G.V. Shanbhag, J. Mol. Catal. A 396
[9] R.S. Karinen, A.O.I. Krause, Appl. Catal. A 306 (2006) 128–133. (2015) 47–54.
[10] Y. Wang, J. Zhou, X. Guo, RSC Adv. 5 (2015) 74611–74628. [50] H.R. Prakruthi, B.M. Chandrashekara, B.S. Jai Prakash, Y.S. Bhat, Catal. Sci.
[11] S.S. Priya, V.P. Kumar, M.L. Kantam, S.K. Bhargava, A. Srikanth, K.V.R. Chary, Technol. 5 (2015) 3667–3674.
Ind. Eng. Chem. Res. 54 (2015) 9104–9115.
[12] Y.T. Kim, K.D. Jung, E.D. Park, Microporous Mesoporous Mater. 131 (2010)
28–36.
 Journal of Cleaner Production 165 (2017) 1090e1096

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Valorization of waste “date seeds” bio-glycerol for synthesizing

oxidative green fuel additive
Farrukh Jamil a, Sandeep K. Saxena b, Ala'a H. Al-Muhtaseb a, *, Mahad Baawain c,
Mohammed Al-Abri a, Nagabhatla Viswanadham b, Gopalakrishnan Kumar d,
Ahmad M. Abu-Jrai e
a
Department of Petroleum and Chemical Engineering, College of Engineering, Sultan Qaboos University, Muscat, Oman
b
Refining Technology Division, CSIR-Indian Institute of Petroleum, Dehradun, 248005, India
c
Department of Civil and Architecture Engineering, College of Engineering, Sultan Qaboos University, Muscat, Oman
d
Department of Environmental Engineering, Daegu University, South Korea
e
Environmental Engineering Department, Faculty of Engineering, Al-Hussein Bin Talal University, Ma'an, Jordan

a r t i c l e i n f o a b s t r a c t

Article history: Bio-glycerol is a polar triol with a high boiling point and obtained as a by-product from trans-
Received 18 February 2017 esterification of plant based oils. In this study, a fuel additive (Solketal, 2,2-dimethyl-1,3-dioxolane-4-
Received in revised form methanol) was synthesized from the condensation of bio-glycerol with acetone in the presence of
25 July 2017
different tailored forms of beta zeolite. The bio-glycerol was obtained from the transesterification of
Accepted 28 July 2017
Available online 29 July 2017
waste “date seed” oil. Beta zeolite catalysts treated with acids (hydrochloric acid, nitric acid and oxalic
acid) exhibited enhanced catalytic activity, irrespective of the nature of the acid used for the deal-
umination. Modified beta zeolite was characterized using the following techniques: X-Ray Diffraction,
Keywords:
Date seeds
Scanning Electron Microscopy, Transmission Electron Microscopy, Energy Dispersive Spectroscopy,
Bio-glycerol Ammonia Temperature Programmed Desorption and Brunauer Emmett and Teller analysis. The nitric-
Solketal acid-treated beta zeolite sample exhibited a higher conversion than the other acid-treated samples. At
Bio-acetone optimum conditions, the bio-glycerol conversion and solketal yield were 94.26% and 94.21 wt%,
Beta zeolite respectively. The best catalyst formulation was reusable without any significant loss in its activity. The
Acetylation process described in this study offers an attractive route for converting bio-glycerol to green fuel additive
(solketal), which has a wide range of commercial applications.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction biodegradable, non-toxic and renewable nature; these character-

istics are desirable in the present search for alternatives to the
Globally, the diversification of fuel sources for energy produc- conventional fossil fuels (Climent et al., 2014; Leung et al., 2010).
tion and transportation is a major focus because of the environ- However, large-scale production of biodiesel gives glycerol as a
mental and availability issues associated with fossil fuels (Eguchi by-product, usually in excess, and this accounts for almost 10% of
et al., 2015; Macombe et al., 2013). Currently, most of the energy the biodiesel yield (Lu et al., 2017). The properties of glycerol such
production relies on fossil fuel sources which include oil, natural as diesel-immiscibility make it unsuitable for fuel blending; be-
gas and coal. However, combustion of fossil-based fuels results in a sides, at high temperatures, glycerol polymerizes and a fraction of it
dramatic increase in CO2 emissions. The rapid depletion of fossil is given off as toxic acrolein by oxidation (Siew et al., 2015). These
fuel sources indicates the emergent need for new and ongoing drawbacks have led to the search for new alternative ways, for
research on the use of biomass-based feedstock as an alternative adding value to this inexpensive and abundant raw material
source of fuel for energy production and transportation. Biodiesel (glycerol).
synthesis is gaining increasing importance because of its There are several routes for the transformation of glycerol such
as trans-esterification, etherification, oxidation, acetylation,
hydrogenolysis and dehydration (Bagheri et al., 2015; Souza et al.,
* Corresponding author. 2015). The oxygenates produced from glycerol can potentially
E-mail address: [email protected] (A.H. Al-Muhtaseb).

http://dx.doi.org/10.1016/j.jclepro.2017.07.216
0959-6526/© 2017 Elsevier Ltd. All rights reserved.
 F. Jamil et al. / Journal of Cleaner Production 165 (2017) 1090e1096 1091

enhance the properties of diesel in blended form; in addition, it can successfully modified using various acids for framework deal-
help to increase the biodiesel yield, which can decrease the pro- umination (without affecting the catalyst structure) to improve the
duction cost of biodiesel from different feedstocks (Bozkurt et al., catalytic activity and stability towards solketal production. To our
2015). Alcohols and olefins are the most common agents used for knowledge, this is the first study utilizing bio-glycerol produced
etherification of glycerol, although they cause the formation of side from the transesterification of waste “date seeds” oil for synthesis
products such as di-olefins in case of olefins and the formation of of fuel additives (solketal).
water in large quantity when alcohol is used. There are several
possible routes for glycerol conversion such as esterification in the 2. Materials and methodology
presence of low molecular weight acids, trans-esterification with
small molecular weight esters and acetylation in the presence of 2.1. Materials
aldehydes or ketones (Okoye and Hameed, 2016; Zhou et al., 2013).
Due to the sustainable nature of acetone produced from biomass Bio-glycerol was collected as a by-product from the trans-
(ABE fermentation process and Mix Alco process), it has gained esterification of waste “date seeds” oil. Commercial beta zeolite (Si/
importance in acetylation as well as in the chemical process of Al ¼ 12.5) in powdered Hþ form was purchased from Sud-Chemie
cumene cracking (Gallezot, 2012; Trejo et al., 2011). India Ltd. All other chemicals and reagents, which were of analyt-
The reaction between both biomass-derived feedstocks, that is, ical grade, were purchased from Sigma Aldrich (Germany) and
glycerol and acetone, is advantageous because these compounds Merck (Germany).
together make up a tremendous component for the formulation of
gasoline, diesel and biodiesel fuels. The five-membered ring com- 2.2. Catalysts preparation
pound that is synthesized, is known as solketal (2,2-dimethyl-1,3-
dioxolane-4-methanol), and has direct applications as a fuel addi- Commercial beta zeolite was used as parent sample, which was
tive. This compound tends to improve diesel properties such as termed as AB catalyst. Dealumination of beta zeolite catalyst was
reducing emissions (by decreasing viscosity), and improving cloud carried out using low molecular weight acids: nitric acid (HNO3),
point. In addition, less undesired particles, unburned hydrocarbons, hydrochloric acid (HCl) and oxalic acid analytical reagent (AR)
carbon monoxide and free aldehydes were observed in emissions grade. Initially, 25 g of beta zeolite was measured into a round-
with enhanced cold flow and flash point properties. These benefits bottom flask and 250 mL of acid solution (2 N) was added. The
of solketal are highly significant because of the increased demand mixture was refluxed for 2 h at 100  C. After refluxing, the solution
for biodiesel additives that are biodegradable, non-toxic and from the round-bottom flask was transferred to a Buckner funnel
renewable, and which are also derived from renewable sources lined with a paper filter and washed thoroughly with deionized
(Herna
ndez et al., 2012). water until it was neutralized (pH z 7). However, the process was
Acetylation of glycerol has been used as an acid-catalyzed re- fastened by attaching a vacuum pump to the filtration flask.
action for the production of solketal using homogeneous and het- Washing was done to remove aluminium (which may occupy the
erogeneous processes. Mineral acids like hydrochloric acid (HCl), pores of parent beta zeolite) from the solution, as well as to remove
hydrofluoric acid, and sulfuric acid have been employed in homo- traces of acidic solution from the mixture. The washed zeolite was
geneous catalytic process (García et al., 2008). However, they have dried at room temperature 25  C for 6 h followed by oven heating
several issues which include corrosion and increasing environ- 100  C for 8 h and finally calcined after drying at 500  C for 4 h. The
mental and economic distress due to effluent disposal. Heteroge- synthesized catalysts were named AB-1, AB-2 and AB-3 based on
neous catalysts are better than homogeneous catalysts for a the acid used for treatment: hydrochloric acid, nitric acid and oxalic
number of reasons, such as easy recovery, reusability, excellent acid, respectively.
thermal stability and provision of multiple active sites for reaction.
The main hindrance for acetylation of glycerol to produce solketal is 2.3. Catalysts characterization
the formation of water, which must be removed to avoid revers-
ibility of reaction and deactivation of heterogeneous catalysts The synthesized zeolite catalysts were characterized following
(Fechete et al., 2012; Heitbaum et al., 2006). The presence of sol- X-ray diffraction (XRD), Scanning Electron Microscope (SEM) and
vents such as benzene, toluene, petroleum ether and chloroform Transmission Electron Microscopy (TEM) techniques for particle
has been observed to increase the conversion of glycerol but the morphology. PANalytical, Xpert PRO instrument (USA), containing
involvement of solvents causes handling and environmental rotatable anode with CuKa radiations, was used for measuring XRD
problems (Deutsch et al., 2007). In heterogeneous catalysis, patterns of the samples. Surface morphology was done by using
different types of catalysts, such as ion exchange resins (Amberlyst- SEM on JEOL JSM-7900F instrument (Japan). Internal morphology
15, Amberlyst-36, Amberlyst-35), are used in the acetylation re- of synthesized catalysts was observed using TEM images on JEOL,
actions of glycerol (Wang et al., 2016; Zhou et al., 2012). These JSM-2100F instrument (Japan). Properties such as pore size, specific
catalysts show good initial conversion but are prone to fast catalyst surface area, and pore volume were measured using Brunauer,
deactivation and production of undesirable side products. Alter- Emmett and Teller (BET) on ASAP 2020, Micromeritics Instruments
natively, sulfated zirconia, SBA-15, ZSM-5, heteropoly acid and Inc., Norcross, GA (USA) using nitrogen gas (99.995% pure). For
carbon silica composite materials have been recently used for sol- determining the functional groups present in the synthesized cat-
ketal production (da Silva et al., 2015; Gonçalves et al., 2008; Testa alysts, Fourier Transform Infrared Spectroscopy (FTIR) was con-
et al., 2013), but these systems also has a disadvantagedthe cata- ducted using Perkin Elmer- spectrum etwo (USA) in accordance
lyst can be deactivated by the water formed during the reaction. with the KBr pellet technique. NH3-TPD was used to determine the
Silicious zeolites possessing higher hydrophobic nature have been acidic properties of synthesized catalysts.
observed to give better water resistance and catalytic performance
(da Silva et al., 2009). 2.4. Catalyst evaluation studies
This study presents a low temperature synthesis of solketal from
bio-glycerol (produced from transesterification of date seeds oil) The reaction of bio-glycerol with acetone was carried out in a
with acetone in the presence of modified solid acid catalysts. Beta magnetically-stirred two-necked round-bottom flask. In a typical
zeolite was used as catalyst and its catalytic properties were experiment, catalyst (10% of bio-glycerol mass) was taken in a
1092 F. Jamil et al. / Journal of Cleaner Production 165 (2017) 1090e1096

round-bottom flask with the required amount (molar ratio 1:2 to analysis of all samples confirmed the dealumination, because the
1:8) of bio-glycerol and acetone. The mixture (bio-glycerol to Si/Al ratio increased from 12.5 in the parent beta AB to 19.1, 20.4
acetone in desired molar ratio) was re-fluxed at the required and 18.0 in the acid-treated samples, AB-1, AB-2 and AB-3,
temperature (60e100  C) for the required time (1e5 h). Once the respectively. It was observed that the high ratio was obtained
reaction was complete, the catalyst was removed from the reaction when nitric was used as treating acid.
products by centrifugation at 5000 rpm. The used catalyst was TEM images of all samples (AB, AB-1, AB-2 and AB-3) are shown
washed with ethanol to completely remove traces of reaction in Supplementary Material, Fig. S2. The images show the expected
mixture and then used again. The kinetic experiments were done in porosity and crystalline nature of the samples and the inter-
autoclave Parr reactor and once the reaction was complete, the connected pores are clearly visible. However, the pores are not
same procedure as mentioned before was repeated for product uniform and this tends to give a sponge-like network. It was
separation. The reaction products were analyzed via a gas chro- observed that there was formation of mesopores on acid-treated
matograph (GC, PerkinElmer) equipped with HP5 capillary column beta zeolite for AB-2. But the formation in AB-1 and AB-3 was
(30 m length, 0.28 mm id, 0.25 mm film thickness) and a Flame less when compared to that on AB-2. This is due to acid leaching
Ionization Detector (FID). which alters framework material and improves the porous
properties.
The BET results of parent beta zeolite along with its modified
3. Results and discussion
forms are presented in Table 1. It was observed that the specific
surface area of the beta zeolite increased up to 607.22 from
3.1. Catalysts characterization
467.40 m2/g after nitric acid dealumination for AB-2 sample.
Similarly, AB-1 and AB-3 samples also exhibited increase in the BET
The XRD patterns of the various acid-treated beta zeolite sam-
specific surface area after acid dealumination. The increase in
ples (AB-1, AB-2 and AB-3) are shown in Fig. 1. For all the samples, it
specific surface area of the beta zeolite upon acid treatment can be
was observed that two intense peaks appeared at 8.2 and 22.5
ascribed to the formation of some additional pores. Removal of
which are characteristic peaks, representing the nature of beta
some amorphous material formed during the synthesis of beta
zeolite. Besides, a comparison between the patterns for parent beta
zeolite and the rearrangement of framework defect sites originally
zeolite and acid-treated beta zeolites showed that both character-
present in the parent beta zeolite are also responsible for
istic peaks were present in all, which confirmed that the framework
improvement in the specific surface area after acid treatment. The
structure of beta zeolite was preserved after the acid treatments.
porosity data of the parent and acid-treated samples also indicate a
Manjunathan et al. reported the valorization of glycerol with
significant increase in mesopore volume in AB-2 catalyst as
treated beta zeolite catalysts (Manjunathan et al., 2015). They also
compared to AB-1 and AB-3 catalyst. Thus, the trend in the total
reported that the crystalline nature of the parent material was
pore volume and mesopore volume is in the increasing order of
maintained after the acid treatment, and the XRD patterns showed
AB < AB-1<AB-3<AB-2, but the micropore volume shows a reverse
peaks that were similar to the two characteristic peaks of beta
trend as compared to mesopore volume. This clearly indicates that
zeolite. It can therefore be concluded that the acid treatment was
during acid leaching, the enhancement of total pore volume occurs
mild, hence it did not disturb the integrity of beta zeolite and the
because of the extra pore formation in the mesopore region.
crystalline nature of the parent material was maintained.
Overall, in comparison with the parent sample, specific surface area
The effect of acid treatment, using different kinds of mineral
and mesopore volume enhancement were observed in all the acid-
acids (HCl, HNO3 and C2H2O4), on the morphology of beta zeolite is
leached dealuminated samples. These results support the
shown in Supplementary Material, Fig. S1. It can be observed from
improvement in beta zeolite porosity after dealumination.
the surface images that the parent beta zeolite possesses irregular
Acidity of all samples was measured by temperature-
spherical shaped particles. The SEM images of the acid-treated
programmed desorption of ammonia and all trends with respect
samples indicated slightly rough surfaces of the crystals
to temperature are shown in Fig. 2. The trends indicate the presence
compared to those of the parent sample, suggesting the formation
of strong and weak acidic sites in all materials. The surface acidity of
of some mesopores. Treatment of beta zeolite with acids causes
catalyst is an important factor in assessing glycerol conversion in
dealumination of the parent material and this causes the ratio of
the presence of acetone as this reaction proceeds on surface acidic
silica to alumina (Si/Al) to increase (Manjunathan et al., 2015). Thus
sites. It was observed that the TPD spectra profiles are essentially
the results obtained from the Energy Dispersive Spectroscopy (EDS)
made of three non-symmetrical NH3 desorption peaks, which can
be classified based on the temperature of desorption of NH3 as
weak (120e300  C), moderate (300e500  C) and strong
(500e700  C). The TPD spectra clearly indicates that a slight in-
crease was observed in the case of AB-2 catalyst when compared to
the other acid-treated catalysts. The acidity of material is directly
related to dealumination, which occurred due to treatment with

Table 1
BET analysis of specific surface area, pore volume (micro and meso), and diameter of
beta zeolite (AB) catalyst with its modified forms.

Properties Catalysts

AB AB-1 AB-2 AB-3

SBET (m2/g) 467.40 597.15 607.22 527.09

Total Pore Volume (cm3/g) 0.55 0.58 0.60 0.59
Micro pore volume (cm3/g) 0.16 0.15 0.11 0.15
Meso pore volume (cm3/g) 0.39 0.43 0.49 0.44
Pore Diameter (nm) 0.57 0.55 0.58 0.53
Fig. 1. XRD patterns of parent beta zeolite (AB) along with its modified forms (AB 1e3).
 F. Jamil et al. / Journal of Cleaner Production 165 (2017) 1090e1096 1093

samples was retained even after modification. This was confirmed

by XRD analysis and microscopic (SEM, EDS and TEM) analyses
(based on images). It was further confirmed by the changes in ratio
of Si/Al which increased after treatment of base material with acids
up to their nature. Definite changes were observed in specific sur-
face area along with an increase in pore size after modification. This
could mean that the modification is good for reactions involving big
reactants and products such solketal formation from glycerol. Also,
the reaction was supported by acidic catalyst and ammonia TPD
confirmed the acidic nature of catalysts. Thus it can be concluded
that modified catalysts can be successfully utilized for obtaining the
main objective of valorizing the bio-glycerol by-product.

3.2. Solketal synthesis

The liquid-phase acetylation of bio-glycerol with acetone over

Fig. 2. NH3-TPD patterns for parent beta zeolite and its modified forms. various acid dealuminated catalysts was carried out at mild reac-
tion conditions of 60  C reaction temperature and 4 h run time in a
round-bottom flask. This reaction mainly yields the five-membered
acid as it was observed that silica to alumina ratio was minimum for
ring, ketal (solketal), and also forms the six-membered ring as a
beta zeolite while it was maximum for AB-2. This can be attributed
side product, whose relative formation depends on the acetylation
to more uniform channel space and active acidic sites which were
position within the glycerol molecule and catalytic properties
available for the reaction. Although dealumination occurred in
(Manjunathan et al., 2015; Vicente et al., 2010). The bio-glycerol
samples AB-1 and AB-3, it was less compared to that of AB-2. The
acetylation with acetone favors the formation of five-membered
silica to alumina ratio of both samples were less than that of sample
ring transition states, which leads to the production of solketal.
AB-2 and the presence of strong acidic and weak acidic sites, as
The overall performance of all the catalysts for solketal production
shown in Fig. 2, was lower in AB-1 and AB-3.
is given in Table 2. The bio-glycerol conversion and solketal yield
All samples were analyzed by FTIR to determine the functional
were observed to increase on the acid-dealuminated beta zeolite,
groups present. Spectra of all samples are shown in Fig. 3. The
but the increase was observed to be highest on the AB-2 catalyst.
presence of sharp adsorption peak in the range of 565e580 1/cm
The conversion increased from 70% to 94% and the solketal yield
confirmed the structural identity for of beta zeolite framework. The
increased from 64 wt% to 94 wt% with a reasonable decrease in six-
fingerprint bands of beta zeolite observed at 1390, 980, 810 1/cm
membered (2,2-dimethyl-1,3-dioxolane-5-ol) by-product on the
are usually assigned to nas [Al-OR(R ¼ H or OH)], n[Si-O] and ns (T-
AB-2 catalyst. The better performance of AB-2 when compared to
O-T,T ¼ Si or Al) respectively. The peak present at 1650 1/cm is
AB-1 and AB-3 catalyst also indicates the possible role of higher
attributed to bending vibration of adsorbed water molecules. The
specific surface area, mesoporosity and strong acidity possessed by
decreasing intensity of band at ~3650 1/cm indicates elimination of
AB-2 catalyst.
some Al atom from the zeolite framework after dealumination.
The selectivity of solketal is directly related to pore volume, pore
Thus, FTIR confirmed that the structural identity of beta zeolite
size and specific surface area of catalysts. Catalysts with larger pore
after modification with acids, remains unchanged.
size will allow the reactants to convert more and cause no hin-
Modifications done on beta zeolite and characterization of all
drance in the formation of large molecules like solketal and in the
samples revealed that changing the acid for modifying the base
release of the product. The reverse happens for catalysts possessing
material causes several changes. The crystalline nature of all
more micro pores (AB, AB-1 and AB-3) than mesospores. Further-
more, high activity of AB-2 could be attributed to easy diffusivity of
molecules in large pores and short path length due to small crys-
talline size.
The bio-glycerol condensation with acetone can be directly
related to silica to alumina ratio of the catalyst. AB-2 has high ratio
Si/Al 20.4 which allows more conversion and high solketal selec-
tivity due to large specific surface area and pore volume. Mean-
while, AB-3 has low silica to alumina ratio of 18, compared to that of
AB-1 which was 19.1, but it gave high bio-glycerol conversion and
solketal selectivity due to the presence of a large mesopore volume
which allows more interaction of reactants with acidic sites inside
the channels and causes more solketal to be formed as a product.

Table 2
Conversion of glycerol using synthesized catalysts along with the yield of the desired
product.

Catalysts Temperature Glycerol Solketal Acetal

( C) conversion (wt%) yield (wt%) yield (wt%)

AB 60 70.13 64.65 5.48

AB-1 60 86.32 83.71 2.61
AB-2 60 94.26 94.21 0.05
AB-3 60 92.45 89.60 2.85
Fig. 3. FTIR spectra of parent beta zeolite and its modified forms.
1094 F. Jamil et al. / Journal of Cleaner Production 165 (2017) 1090e1096

Therefore, it can be concluded that among all parameters, pore

volume is more effective for bio-glycerol conversion. Hence,
modification of beta zeolite with nitric acid gave high pore volume
due to maximum silica to alumina ratio and large number of strong
and weak acidic sites.

3.3. Parametric effect

The effect of reactants mole ratio of bio-glycerol to acetone was

studied at 60  C using 10 wt% of catalyst with respect to bio-
glycerol. The mole ratio of bio-glycerol to acetone was varied
from 1:2 to 1:8. As shown in Fig. 4, the conversion of bio-glycerol
increased with the mole ratio, from 1:2 to 1:6, due to the in-
crease in accessibility of acetone with bio-glycerol. Bio-glycerol
conversion remained almost the same with further increase in
mole ratio to 1:8. The maximum solketal yield obtained was
Fig. 5. Effect of process temperature with respect to solketal yield and glycerol
94.21 wt% when the molar ratio was 1:6. Thus, the consistency in conversion.
solketal yield with molar ratio 1:6e1:8 may be attributed to the by-
product (water) formation, which causes conversion of solketal
back to glycerol and causes the equilibrium in solketal yield (Mota 4 h, there was no more change observed in glycerol conversion and
et al., 2010). Based on the results, the optimum bio-glycerol to solketal selectivity. Thus, it can be inferred that equilibrium had
acetone molar ratio for catalytic transformation of bio-glycerol to been achieved in 4 h, and the obtained results were in good
solketal, was found to be 1:6 on AB-2 catalyst (Fig. 4). agreement with those reported earlier (Khayoon and Hameed,
The reaction temperature also influenced the bio-glycerol con- 2013).
version and solketal yields. The trend for solketal formation from
glycerol conversion, shown in Fig. 5, indicates that, at the initial
3.4. Solketal characterization
stages of the reaction, formation of solketal increased with increase
in temperature. The high initial yield with increase in temperature
Some basic fuel properties of solketal and its blends with Date
shows that at higher temperatures, both reactants possess enough
Pits Oil Biodiesel (DPOB) (Jamil et al., 2016) were determined as
energy to easily react with each other. But, above 60  C, the solketal
mentioned in Table 3. All the determined properties were
yield was observed to decrease. Thus, 60  C was recorded as the
compared with international standards such as American Society
optimum reaction temperature. The decrease in solketal yield, at
for Testing and Materials (ASTM) and European (EN) to confirm the
temperatures above 60  C, may be attributed to the fact that at
commercial and industrial viability of solketal in blended form with
higher temperatures, acetone vaporization increases rapidly and
biodiesel. It can be observed from the properties of DPOB-solketal
this reduces the availability of one reactant (acetone) for the re-
(DPOBS) blends that they satisfy the international standards
action. Therefore, it can be stated that glycerol-acetone reaction
(ASTM and EN) limits. Moreover, as the amount of solketal was
proceeds preferably at low temperature.
increased the density as well as viscosity increased, but when
Fig. 6 shows the effect of process time on glycerol conversion
solketal was 10 wt%, the density of DPOBS-3 was a little higher than
and solketal selectivity. It can be observed that initially, the glycerol
its limit defined by EN. Meanwhile, the viscosity still satisfies the
conversion along with glycerol selectivity was low but as the time
ASTM and EN standards for all blends (DPOBS-1 to 3). Thus, vis-
was increased, glycerol conversion along with solketal selectivity
cosity of DPOBS blends confirms that it will not affect its atomi-
was enhanced. The glycerol conversion increased rapidly till the
zation in engine to have adverse affect on the engine efficiency.
process time was 3 h and a small change was observed when
Flash point is one of the major properties of fuels, and it can be
increased to 4 h, but when the process time was increased beyond
observed that by increasing the solketal amount in blends it was
decreased and, for DPOBS-3, it was 107.98  C but it was higher than

Fig. 4. Effect of reactants molar ratio with respect to solketal yield and glycerol
conversion. Fig. 6. Effect of process time with respect to solketal yield and glycerol conversion.
 F. Jamil et al. / Journal of Cleaner Production 165 (2017) 1090e1096 1095

Table 3
Properties of solketal and blends of solketal with Date Pits Oil Biodiesel (DPOB).

Property Solketal DPOB DPOBS-1a DPOSB-2b DPOB-3c EN 14214 ASTM 6751

Density (g/cm3) 1.049 0.870 0.881 0.894 0.911 0.86e0.90 d

Viscosity (mm2/s) 4.461 3.970 4.011 4.097 4.209 3.5e5.0 1.9e6.0

d d
Gross calorific value (MJ/kg) 31.15 44.10 43.65 41.73 40.15
Flash point ( C) 88.21 137.00 129.35 118.87 107.98 120 min 93 min
Acidity (mg KOH/g) 0.31 0.21 0.22 0.24 0.28 0.5 max 0.5 max
a
DPOB with 1 wt% of solketal.
b
DPOB with 5 wt% of solketal.
c
DPOB with 10 wt% of solketal.
d
Not specified.

the limit defined by the EN standard. So, a higher flash point lead to used for next two runs, the bio-glycerol conversion as well as sol-
observation that produced biodiesel blends have better stability ketal yield decreased significantly, which indicates that the catalyst
and storability. Similarly, acid values were also consistent with was active up to the third run. Thus, the decrement in bio-glycerol
standard values. Hence, it can be attributed that solketal does not conversion can be attributed to the fact that it might be due some
have a significant negative impact on the basic fuel properties of reactants deposition or due to decrement in active sites of catalyst
biodiesel. A similar kind of work has been reported earlier by Alp- (Manjunathan et al., 2015).
tekin E. in which he produced blends of solketal (produced from
glycerol) with biodiesel, measured the basic fuel properties of 4. Conclusions
blends, tested the blends in an engine and observed that the effi-
ciency had improved and the emission analysis showed a reduction The present study identifies an efficient method for the syn-
in CO, CO2 and hydrocarbons (Alptekin, 2017). Thus, based on this, thesis of solketal from low-value bio-source bio-glycerol in the
and in comparison to the determined properties of produced sol- presence of high-silica mesoporous beta zeolite at solvent-free low-
ketal and its blends with biodiesel, it can be concluded that pro- temperature conditions. It was observed that the catalytic activity
duced solketal can be commercially used in blended form with of beta zeolite increased with the acid treatment, and nitric-acid-
biodiesel without affecting the basic fuel properties and can help on treated catalyst AB-2 showed the highest catalytic activity with a
an industrial scale for achieving better engine efficiency and 94.21 wt% solketal yield as compared to the hydrochloric-acid-
reducing the emission of unwanted gases (CO and CO2) and hy- treated and oxalic-acid-treated zeolite catalysts. The high activity
drocarbons. Moreover, valorization of bio-glycerol to solketal will of AB-2 catalyst could be attributed to easy diffusivity of molecules
also make biodiesel production economical. in the large pores and short molecular path length in the pores. The
parametric study revealed optimum glycerol conversion 94.26%
was obtained following operating conditions such as process
3.5. Reusability
temperature 60  C with molar ratio of bio-glycerol to acetone 1:6
and process time 4 h. The oxygenated product, named solketal, is an
Considering the reusability of catalyst as an important factor,
odorless and colorless liquid which is miscible with water. Thus,
AB-2 was used for five consecutive experimental runs of bio-
based on this, it can be concluded that bio-glycerol obtained as a
glycerol acetalization and the catalyst reusability test was con-
by-product from transesterification can be valorised to economize
ducted after the reaction was completed. The catalyst was sepa-
the biodiesel production.
rated from the product mixture, washed with ethanol and dried at
120  C for 4 h, and reused for the fresh bio-glycerol acetylation
reaction under optimized reaction conditions. As shown in Fig. 7, it Acknowledgement
was observed that up to the third run, the solketal yield along with
bio-glycerol conversion was not significantly affected. But when The authors would like to thank The Research Council of Oman
for its financial support under grant number ORG/EI/13/013.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://

dx.doi.org/10.1016/j.jclepro.2017.07.216.

References

Alptekin, E., 2017. Emission, injection and combustion characteristics of biodiesel

and oxygenated fuel blends in a common rail diesel engine. Energy 119, 44e52.
Bagheri, S., Julkapli, N.M., Yehye, W.A., 2015. Catalytic conversion of biodiesel
derived raw glycerol to value added products. Renew. Sustain. Energy Rev. 41,
113e127.
€
Bozkurt, O.D., Tunç, F.M., Bag _
lar, N., Çelebi, S., Günbaş, I.D., Uzun, A., 2015. Alter-
native fuel additives from glycerol by etherification with isobutene: structur-
eeperformance relationships in solid catalysts. Fuel Process. Technol. 138,
780e804.
Climent, M.J., Corma, A., Iborra, S., 2014. Conversion of biomass platform molecules
into fuel additives and liquid hydrocarbon fuels. Green Chem. 16, 516e547.
da Silva, C.X.A., Goncalves, V.L.C., Mota, C.J.A., 2009. Water-tolerant zeolite catalyst
for the acetalisation of glycerol. Green Chem. 11, 38e41.
da Silva, M.J., Julio, A.A., Dorigetto, F.C.S., 2015. Solvent-free heteropolyacid-
Fig. 7. Reusability of AB-2 with respect to solketal yield and glycerol conversion. catalyzed glycerol ketalization at room temperature. RSC Adv. 5, 44499e44506.
1096 F. Jamil et al. / Journal of Cleaner Production 165 (2017) 1090e1096

Deutsch, J., Martin, A., Lieske, H., 2007. Investigations on heterogeneously catalysed Manjunathan, P., Maradur, S.P., Halgeri, A.B., Shanbhag, G.V., 2015. Room temper-
condensations of glycerol to cyclic acetals. J. Catal. 245, 428e435. ature synthesis of solketal from acetalization of glycerol with acetone: effect of
Eguchi, S., Kagawa, S., Okamoto, S., 2015. Environmental and economic performance crystallite size and the role of acidity of beta zeolite. J. Mol. Catal. A Chem. 396,
of a biodiesel plant using waste cooking oil. J. Clean. Prod. 101, 245e250. 47e54.
Fechete, I., Wang, Y., Ve
drine, J.C., 2012. The past, present and future of heteroge- Mota, C.J.A., da Silva, C.X.A., Rosenbach, N., Costa, J., da Silva, F., 2010. Glycerin
neous catalysis. Catal. Today 189, 2e27. derivatives as fuel additives: the addition of glycerol/acetone ketal (solketal) in
Gallezot, P., 2012. Conversion of biomass to selected chemical products. Chem. Soc. gasolines. Energy & Fuels 24, 2733e2736.
Rev. 41, 1538e1558. Okoye, P.U., Hameed, B.H., 2016. Review on recent progress in catalytic carboxyla-
García, E., Laca, M., Pe
rez, E., Garrido, A., Peinado, J., 2008. New class of acetal tion and acetylation of glycerol as a byproduct of biodiesel production. Renew.
derived from glycerin as a biodiesel fuel component. Energy & Fuels 22, Sustain. Energy Rev. 53, 558e574.
4274e4280. Siew, K.W., Lee, H.C., Gimbun, J., Chin, S.Y., Khan, M.R., Taufiq-Yap, Y.H., Cheng, C.K.,
Gonçalves, V.L.C., Pinto, B.P., Silva, J.C., Mota, C.J.A., 2008. Acetylation of glycerol 2015. Syngas production from glycerol-dry (CO2) reforming over La-promoted
catalyzed by different solid acids. Catal. Today 133e135, 673e677. Ni/Al2O3 catalyst. Renew. Energy 74, 441e447.
Heitbaum, M., Glorius, F., Escher, I., 2006. Asymmetric heterogeneous catalysis. Souza, T.E., Padula, I.D., Teodoro, M.M.G., Chagas, P., Resende, J.M., Souza, P.P.,
Angew. Chem. Int. Ed. 45, 4732e4762. Oliveira, L.C.A., 2015. Amphiphilic property of niobium oxyhydroxide for waste
Herna
ndez, D., Ferna
ndez, J.J., Mondrago
n, F., Lo

pez, D., 2012. Production and uti- glycerol conversion to produce solketal. Catal. Today 254, 83e89.
lization performance of a glycerol derived additive for diesel engines. Fuel 92, Testa, M.L., La Parola, V., Liotta, L.F., Venezia, A.M., 2013. Screening of different solid
130e136. acid catalysts for glycerol acetylation. J. Mol. Catal. A Chem. 367, 69e76.
Jamil, F., Al-Muhtaseb, A.A.H., Al-Haj, L., Al-Hinai, M.A., Hellier, P., Rashid, U., 2016. Trejo, F., Rana, M.S., Ancheyta, J., 2011. Genesis of AcidBase support properties
Optimization of oil extraction from waste “Date pits” for biodiesel production. with variations of preparation conditions: cumene cracking and its kinetics. Ind.
Energy Convers. Manag. 117, 264e272. Eng. Chem. Res. 50, 2715e2725.
Khayoon, M.S., Hameed, B.H., 2013. Solventless acetalization of glycerol with Vicente, G., Melero, J.A., Morales, G., Paniaguab, M., Mart
ın, E., 2010. Acetalisation of
acetone to fuel oxygenates over NieZr supported on mesoporous activated bio-glycerol with acetone to produce solketal over sulfonic mesostructured
carbon catalyst. Appl. Catal. A General 464e465, 191e199. silicas. R. Soc. Chem. 12, 899e907.
Leung, D.Y.C., Wu, X., Leung, M.K.H., 2010. A review on biodiesel production using Wang, Z.-Q., Zhang, Z., Yu, W.-J., Li, L.-D., Zhang, M.-H., Zhang, Z.-B., 2016.
catalyzed transesterification. Appl. Energy 87, 1083e1095. A swelling-changeful catalyst for glycerol acetylation with controlled acid
Lu, J., Liu, Z., Zhang, Y., Li, B., Lu, Q., Ma, Y., Shen, R., Zhu, Z., 2017. Improved pro- concentration. Fuel Process. Technol. 142, 228e234.
duction and quality of biocrude oil from low-lipid high-ash macroalgae Zhou, C.H., Zhao, H., Tong, D.S., Wu, L.M., Yu, W.H., 2013. Recent advances in cata-
Enteromorpha prolifera via addition of crude glycerol. J. Clean. Prod. 142 (Part lytic conversion of glycerol. Catal. Rev. 55, 369e453.
2), 749e757. Zhou, L., Nguyen, T.-H., Adesina, A.A., 2012. The acetylation of glycerol over
Macombe, C., Leskinen, P., Feschet, P., Antikainen, R., 2013. Social life cycle assess- amberlyst-15: kinetic and product distribution. Fuel Process. Technol. 104,
ment of biodiesel production at three levels: a literature review and develop- 310e318.
ment needs. J. Clean. Prod. 52, 205e216.
 Chemical Engineering Journal 347 (2018) 41–51

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Continuous reactive coupling of glycerol and acetone – A strategy for T

triglyceride transesterification and in-situ valorisation of glycerol
by-product
⁎
Valentine C. Eze , Adam P. Harvey
School of Engineering, Newcastle University, NE1 7RU, UK

H I GH L IG H T S

• Clear steady states achieved for the triacetin reaction at short induction times.
• Higher rate of triacetin conversion was obtained using the Amberlyst A26-OH. TM

• The TM
Amberlyst resins are stable for use in heterogeneous catalysis.
• One-stage process achieved 48.5 ± 2.7% glycerol by-product to solketal conversion.
• Higher solketal conversions (76.5 ± 2.8%) occurred using the two-stage process.

A R T I C LE I N FO A B S T R A C T

Keywords: Methyl esters of fatty acids are widely used as biodiesel, a sustainable replacement for petro-diesel. The con-
Reactive coupling ventional biodiesel process produces crude glycerol, which constitutes about 10 wt% of the total products. This
Methyl esters has led to a surplus of crude glycerol due to global increase in biodiesel use, necessitating increased research into
Solketal sustainable processes that could convert the crude glycerol into higher value-added products. This study in-
AmberlystTM resin catalysts
vestigates biodiesel processes for continuous transesterification of triglycerides to methyl esters, coupled to
Acetalisation
Mesoscale oscillatory baffled reactors (meso-
conversion of the glycerol by-product into solketal, a value-added product, via reaction with acetone in situ. The
OBRs) study was carried out using one-stage and two-stage catalytic transesterification of triacetin and methanol in
mesoscale oscillatory baffled reactors (meso-OBRs). The two-stage process involved two meso-OBRs in series
packed with AmberlystTM resin catalysts: a basic AmberlystTM A26-OH in the first stage to catalyse transes-
terfication of triacetin with methanol, and an acidic AmberlystTM 70-SO3H in the second stage to catalyse the
coupling of glycerol and acetone to form solketal. One-stage triacetin transesterification and glycerol coupling
with acetone was carried out in a meso-OBR packed with the acidic AmberlystTM 70-SO3H resin. In the two-stage
process, the triacetin was converted to 99.1 ± 2.0% methyl acetate and 98.0 ± 1.3% glycerol after 25 min
residence time in the first reactor and the glycerol was reacted with acetone in the second reactor to achieve
76.5 ± 2.8% solketal conversions after 35 min. The single-stage process achieved 48.5 ± 2.7% solketal con-
version after 30 min. The meso-OBR was operated continuously to achieve high quality steady states and con-
sistent triacetin conversions. The triglyceride transesterification with reactive coupling of glycerol with acetone
produces less crude glycerol by-product. This process strategy could be optimised for future biodiesel produc-
tion.

1. Introduction % crude glycerol as a by-product (Fig. 1). The most commonly applied
method for biodiesel production is the base-catalysed homogeneous
Fatty acid methyl esters are the main constituents of biodiesel, a process using alkali-metal hydroxides and methoxides [1–3], particu-
product that is widely used as alternative to petro-diesel. These esters larly sodium methoxide [3]. The biodiesel produced by this method is
are mostly produced via transesterification of triglycerides with me- majorly a mixture of fatty acid methyl esters (FAME). These alkali
thanol using acid and base catalysts, in a process which produces 10 wt catalysts are commonly used because of their high catalytic activity

⁎
Corresponding author.
E-mail address: [email protected] (V.C. Eze).

https://doi.org/10.1016/j.cej.2018.04.078
Received 18 May 2017; Received in revised form 6 April 2018; Accepted 13 April 2018
Available online 14 April 2018
1385-8947/ Crown Copyright © 2018 Published by Elsevier B.V. All rights reserved.
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

Fig. 2. The acid-catalysed reaction of glycerol and acetone to form solketal.

Fig. 1. Transesterification reaction for biodiesel production.

tonnes in 2025. The negative impact of glycerol oversupply on the
glycerol market has been enormous. A study [15] has shown that crude
which leads to high biodiesel yield at a short reaction time compared to glycerol has effectively become a waste product, as the price has cra-
acid-catalysed processes [4]. shed to ∼ €0 per tonne, whereas the market price for refined glycerol
However, the use of homogeneous catalysts comes at a great cost. It (99.5% Kosher grade) has decreased from €4000 per tonne in the year
necessitates various operating and capital costs downstream, arising 2000, to about €450 per tonne in 2010. These trends demonstrate that a
from the inability to recover the catalyst for further use and the large paradigm shift in biodiesel processing technology is necessary to ensure
amount of water and energy required for the washing and drying of the sustainability of the biodiesel industry.
biodiesel. The crude glycerol by-product from the conventional process Various studies have been carried out on transformation of glycerol
commands low market prices because of contamination by salts formed into valuable products, such as glycerol carbonate [16–18] and solketal
from the catalyst neutralisation. Heterogeneous catalysis of biodiesel [19–22]. Glycerol carbonate has applications as a curing agent in ce-
production [5,6] would address many of these problems. It would ment and concrete building, solvent for active medical ingredients in
eliminate or minimise the cost of the downstream processing steps as- various pharmaceutical formulations, an electrolyte in lithium and li-
sociated with the use of homogeneous catalysts. The crude glycerol thium-ion batteries, and in production of various polymers and plastics
produced through heterogeneous catalysis would be of higher quality [23,24]. Solketal is one of most useful derivatives of glycerol, produced
and command a higher market price, adding to the profitability of via acid-catalysed condensation of glycerol with acetone (Fig. 2).
commercial biodiesel. Another huge advantage of heterogeneous cata- Solketal has applications as a fuel additive [25,26], surfactant and
lysts is the removal of the costs of continual catalyst replacement. This flavouring agent [19]. As a fuel additive, solketal can be utilised by
is a “greener” strategy for productions of valuable chemicals, including blending in biodiesel [25], or gasoline [26]. The use of solketal as an
biodiesel. additive for biodiesel allows for modifications in the current biodiesel
Studies on heterogeneous catalysis of biodiesel production reactions processing technology, by producing biodiesel that already contains a
have shown that anion (OH−) exchange resins [7–9] and cation solketal fuel additive.
(−SO3H) exchange resins [10–12] are promising catalysts. Catalysis of This study was carried out using oscillatory baffled reactors (OBRs)
batch transesterification of triolein with ethanol at 20:1 M ratio and as a process development reactor platform, partly to ensure that mass
50 °C using 4 g of OH− ion exchange resins (PA306) was used to transfer effects are negated. The liquid–liquid reaction in homogeneous
achieve over 98% yield of fatty acid ethyl esters after 5 h reaction time biodiesel processing becomes instead a three-phase liquid–liquid-solid
without catalyst deactivation[7]. Another study also showed that 70% reaction when using heterogeneous catalysts, and this requires strong
FAME yield was achieved after 0.18 h residence time for continuous mixing to ensure mass transfer independence. Process intensification
methanolysis of coconut oil with OH− anion exchange resin at 6:1 reactors are at the heart of current drive for Green Chemistry processes,
methanol to oil molar ratio and 60 °C, without catalyst deactivation allowing for a more efficient utilisation of raw materials, energy effi-
after 10 cycles [8]. No catalyst deactivation was reported for transes- ciency and safer operation through smaller volume of reactors and
terification of soybean oil with methanol with 3 g of OH− anion ex- lower system footprints [27]. Other “intensified” reactors that have
change resin (BRI), where reactions at 150:1 methanol to soybean oil been studied for biodiesel production reactions include microwave re-
molar ratio and 80 °C was used to achieve 97.3% FAME yield after 4 h actors [28,29], ultrasonic reactors [30–32] spinning disc reactors [33]
reaction time [9]. One of the cation-exchange resins that have been and micro-reactors [34–36].The advantages of OBRs [6,37–39] are ease
reported for catalysis of triglyceride transesterification is PTSA–SO3H. of scale up, reproducibility of experiments, compatibility with homo-
The PTSA–SO3H resin was used to catalyse batch transesterification of geneous and heterogeneous catalysts, and potentials for operations in
babaçu oil to obtain 98% FAME yield after 8 h reaction time at 30:1 batch and continuously as a plug flow reactor. Mixing in the OBR is
methanol to oil molar ratio, ∼65 °C and 7 wt% of the catalyst [12]. independent of net flow and controlled by an oscillatory flow, allowing
These studies on ion-exchange resins catalysts clearly demonstrate reactions that require residence times up to hours to be operated in a
that they potentially have significant advantages in terms of operational continuous mode with a greatly reduced ratio of length to diameter.
stability over other catalysts, notably doped and undoped metal oxides The effectiveness of mixing and radial transport in OBRs are controlled
[13], and could be tailored towards heterogeneous catalysis of biodiesel by the net flow Reynolds number (Ren) and the oscillatory Reynolds
production. Therefore, it is envisaged that the use of hydroxide and number (Reo) [38].
sulphonic acid functionalised resin catalysts for biodiesel production
will produce glycerol of high market value, increasing the commercial ρuD
Ren =
viability of biodiesel production. μ (1)
Although glycerol produced using heterogeneous catalysts is of
higher quality than the crude glycerol from the homogeneously cata- 2πfx 0 ρD
Reo =
lysed process, currently it seems that co-production of glycerol of high μ (2)
purity has little economic advantage on biodiesel plants. This is because
where: f: oscillatory frequency (Hz), xo: centre-to-peak amplitude of
glycerol oversupply has significantly reduced the glycerol price [14].
oscillation (m), µ: dynamic viscosity of the reactant mixture (Pa.s), ρ:
The huge rise in global glycerol production has caused surplus of gly-
density of the reactant mixture (kg/m3), D: internal diameter of the
cerol, from only 200,000 tonnes in the year 2003, to over 2 million
baffled tube (m).
tonnes in 2011; and this is predicted [15] to rise to over 6 million
This study investigates biodiesel processes for integrated

42
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

Fig. 3. Reaction scheme for integrated triglyceride transesterification and acetalisation of glycerol.

triglyceride transesterification to form methyl esters, and in-situ con- the AmberlystTM 70-SO3H resin, and the second meso-OBR with the
versions of the glycerol by-product into solketal via reactive coupling ∼600 µm beads of the AmberlystTM A26-OH resin. The catalysts had
with acetone (Fig. 3). been conditioned by washing 5 times with 20 mL of anhydrous me-
This new process produces biodiesel and solketal which are both thanol before packing. The base and top of the packed meso-OBRs were
combustible fuels and would require no downstream separations after sealed with 5 mm discs of stainless steel wire mesh (#60) of 160 µm
recovery of excess methanol and acetone. Triacetin was used as a re- wire diameter and 263 µm apertures to constrain the catalyst beads
presentative triglyceride due to its fast reaction rate which allows for from leaving the reactor. Each reactor was then assembled with the
rapid screening of the proposed reactions. The study investigates pos- base, connected through Swagelok fittings to three Confluent syringe
sible routes for achieving simultaneous conversions of glycerol to pumps (Eurodyne Ltd). One of these pumps was used to provide the
solketal in biodiesel plants using basic (eOH−) and acidic (eSO3H) ion oscillations at 4.5 Hz frequency and 4 mm amplitude, and the other two
exchange resin catalysts. The ion-exchange catalysts were selected for to provide the net flows of triacetin and methanol as shown in Fig. 5.
these reactions because of their reported catalytic stability as hetero- The syringe pump used for oscillation was connected to the base of
geneous catalysts. Small scale OBRs (meso-OBRs) were used because of the reactor, and the fluid mixing inside the reactor provided by ad-
their small volume (a few millilitres), which allows significantly smaller justing the speed of the piston movement (frequency) and the oscilla-
amounts of reagents to be used and minimises the amounts of waste tion amplitude (center-to-peak) of the pump [39]. The syringe pumps
generated. were controlled via a PC interface. The pumps were initialised and set
at the required mixing intensity (amplitude and frequency) and re-
2. Materials and methods actants net flow rates prior to each experiment.
The volumes of the reactors after packing (with 9 g of AmberlystTM
2.1. Materials 70-SO3H, or 8 g of AmberlystTM A26-OH) were 8.3 mL and 9.2 mL, re-
spectively, and the volume of the reactor without packing was 15.2 mL.
Materials used in the experiments were triacetin (99%, Sigma- The reaction temperature was maintained by the circulation of heated
Aldrich), anhydrous methanol (99.8%, Sigma-Aldrich), methyl acetate water through the jackets of the meso-OBRs using a temperature-con-
(99%, Sigma-Aldrich), glycerol (99%, Sigma-Aldrich), solketal (98%, trolled water bath (Ecoline, LUADA E100). All the reactions were per-
Sigma-Aldrich), toluene (99.9%, Sigma-Aldrich) and 2-propanol formed at 50 °C to reduce formation of vapour/gas phase inside the
(99.9%, Sigma-Aldrich). The hydroxide (AmberlystTM A26-OH) and reactor due to the low boiling points of acetone (56 °C) and methyl
sulphonic acid (AmberlystTM 70-SO3H) functionalised styrene- divi- acetate (57 °C).
nylbenzene copolymer catalysts were obtained from the Dow Chemical Triacetin and methanol (and acetone) feeds were dispensed from
Company, Netherland. Physical and chemical properties of the reservoirs maintained at the reaction temperature of 50 °C inside the
AmberlystTM resin catalysts used could be found in the product data- water bath. Steady state performances of the packed meso-OBRs were
sheets and have been reported elsewhere [40,41]. evaluated for the triacetin transesterification using methanol to tria-
cetin molar ratio of 6:1 and residence times of 0.5 min–30 min for the
2.2. Experimental procedures AmberlystTM A26-OH, and methanol to triacetin molar ratio of 30:1 and
residence times of 2.5 min–60 min for the AmberlystTM 70-SO3H. Prior
The AmberlystTM resins were packed in a variety of meso-OBRs and to each experiment, continuous circulation of methanol over the
screened individually for catalysis of triacetin transesterification AmberlystTM resin packed bed was carried out to swell catalysts and
(AmberlystTM 70-SO3H and AmberlystTM A26-OH) and glycerol reac- make the pores network accessible to the reactants [43,44]. All the
tion with acetone (AmberlystTM 70-SO3H). The meso-OBRs used has experiments were carried out at oscillation conditions of ≥4 mm am-
been previously reported [42]. The reactors were jacketed, integrally plitude and ≥4.5 Hz (Reo ≥ 160) where the reaction was clearly
baffled glass tubes of about 770 mm length, 8 mm outer diameter, 5 mm mixing independent (Fig. 6). Incidentally, the Fig. 6 below illustrates
inner diameter and periodic constrictions of 2.5 mm diameter along the well the fact that the interaction of baffles and fluid oscillation leads to
length of the tube at 7.5 mm spacing (Fig. 4). an earlier onset of mixing independence.
One of the meso-OBRs was fully packed with the ∼500 µm beads of

43
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

Fig. 4. Configuration of the reactors [42]. (a) Integrally baffled meso-OBR and the schematics of the internal configuration, (b) diagrammatic view of the meso-OBR
used in the reaction: jacketed meso-OBR (1), oscillation line (2), Feed lines (3 & 4), product/sampling point (5), hot water in (6), and hot water out (7).

Fig. 5. Set-up for the screening: (1) packed bed of AmberlystTM resin, (2) sampling point,

2.3. Simultaneous triacetin transesterification and conversion of glycerol to acetone-triacetin molar ratio of 6:4:1, 50 °C temperature and residence
solketal (one-stage) times of 2.5 min–30 min. An acetone to glycerol molar ratio of 4:1 was
used, based on 1:1 stoichiometry of triacetin and glycerol in a complete
Triacetin transesterification over the meso-OBR packed with 9 g transesterification. A set of experiments was conducted to investigate
AmberlystTM 70-SO3H (as shown in Fig. 5) and simultaneous conversion the direct reactions of glycerol with acetone using the meso-OBR
of glycerol by-product to solketal through reactive coupling with packed with AmberlystTM 70-SO3H. The glycerol used was diluted in
acetone was investigated. The experimental conditions were methanol- methanol at 1:1 vol ratio to reduce its viscosity and ease the flow

44
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

residence times of 5–35 min. About 0.5 mL of sample was collected

through the sampling units at the outputs of the 1st and 2nd reactors for
each residence time

2.5. Sample collections and analysis

Approximately 0.5 mL of sample was collected each time. All the

samples were taken in 2 mL vials and stored at −20 °C for GC analysis.
The gas chromatography method used in the analysis of the triacetin
reaction was adopted from an existing study [46], with some mod-
ifications in the GC conditions. The modified GC method was applied
for determinations of triacetin, methyl acetate, glycerol and solketal in
samples, using 2-propanol as a solvent for all the preparations. Cali-
bration curves for the analysis were prepared with known concentra-
Fig. 6. Mixing independent regions for transesterification at 50 °C, 6:1 me- tions of triacetin, methanol, methyl acetate, glycerol and solketal, using
thanol to triacetin molar ratio and 1 min residence time using 8 g AmberlystTM toluene as the GC internal standard. Stock solutions for these analytes
A26-OH resin catalyst packed in meso-OBR and un-baffled tubular reactor were prepared to the concentrations of 500 mg per millilitre of 2-pro-
(Ren = 24). panol for each component. The toluene internal standard was mixed
with 2-propanol to the concentration of 1:16 (v/v) toluene to 2-pro-
through the PVM syringe pumped line, and to examine any interference panol volume ratio, corresponding to about 48 mg/mL. Various vo-
of methanol in the proposed integrated biodiesel process. The experi- lumes of the stock solutions were measured into 5 mL vials using mi-
ments were performed at a 4:1 acetone to glycerol molar ratio as re- cropipettes, followed by the additions of a 0.5 mL of the 1:16 (v/v)
ported elsewhere [45] using residence times of 2.5 min–30mins and toluene internal standard solution, to achieve analyte/toluene weight
50 °C reaction temperature. fractions in the range of 0–8wt/wt. 1 mL of 2-propanol was added to
each vial to keep the GC signals within the detection limit. About 0.5µL
of each mixture was manually injected into the GC using 5 µL SGE GC
2.4. Two-stage triacetin transesterification and in-situ conversion of glycerol syringe.
to solketal Separations of various components in the calibration mixture and
the triacetin reaction samples were achieved using Varian CP Wax ca-
A two-stage process for triacetin transesterification and in situ pillary (BPX70) column and ramped oven temperature programme. The
conversion of the glycerol by-product to solketal was investigated using oven initial temperature of 50 °C was held for 10 min, ramped at 30 °C/
two interconnected meso-OBRs packed with AmberlystTM A26-OH min to 180 °C and held for 10 min, and then ramped at 30 °C/min to
catalyst (1st reactor) and AmberlystTM 70-SO3H catalyst (2nd reactor), 210 °C and held for 14.67 min. Calibration curves for various compo-
as shown in the Fig. 7. Continuous transesterification at a methanol to nents (R2 > 0.99) were generated by the correlations of the peak areas
triacetin molar ratio of 6:1 was carried out in the presence of the Am- of the FID signals (mV.s) with the known concentrations of the chemical
berlystTM A26-OH catalyst in the 1st reactor to form methyl ester and species in the mixtures.
glycerol, followed by an online reactive coupling of the glycerol with The reaction samples were analysed using the calibration data and
acetone over the AmberlystTM 70-SO3H catalyst (2nd reactor) to form the above oven temperature programme. Approximately 50.0–100.0 mg
solketal. The net flows of the feed correspond to 6:4:1 methanol- of sample was carefully weighed in a 2 mL vial using A&D HR-200
acetone-triacetin molar ratio. This process was screened at 50 °C and weighing balance ( ± 1mg). This was followed by addition of 0.5 mL of

Fig. 7. Experimental set-up for catalytic dual bed: (1) AmberlystTM A26-OH bed; (2) AmberlystTM 70-SO3H bed; (3) syringe pumps: oscillation (3), methanol (3a),
triacetin (3b), acetone (3c), hot water - inlet (4); outlet (5), (6) sampling unit.

45
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

the 1:16 (v/v) toluene internal standard solution, dilution of the sample
mixture with 1 mL of 2-propanol and shaking to obtain a homogeneous
solution. About 0.5µL of the prepared sample was manually injected
into the GC using the 5µL SGE GC syringe. Amounts of triacetin, methyl
acetate, glycerol and solketal in the samples were quantified based on
the calibration curves using Eq. (3). Conversions of the triacetin, methyl
acetate, glycerol and solketal were calculated based on the Eqs. (4)–(7).

A V ·C
Wa = R a·⎛ a ⎞·⎛ is is ⎞
⎜ ⎟ ⎜ ⎟

⎝ Ais ⎠ ⎝ Ms ⎠ (3)

[Triacetin]0 −[Triacetin]t ⎞
Triacetinconversion(%) = 100∗ ⎛ ⎜ ⎟

⎝ [Triacetin]0 ⎠ (4)

[Methylacetate]t ⎞
Methylacetateconversion(%) = 100∗ ⎛ ⎜ ⎟

⎝ 3∗ [Triacetin]0 ⎠ (5)

[Glycerol]t ⎞
Glycerolconversion(%) = 100∗ ⎛ ⎜ ⎟

⎝ Triacetin]0 ⎠
[ (6)

[Solketal]t ⎞ [Solketal]t ⎞
Solketalconversion(%) = 100∗ ⎛ ⎜ = 100∗ ⎛
⎟ ⎜ ⎟

⎝ [Glycerol ]0 ⎠ ⎝ Triacetin]0 ⎠
[ (7)
where: Ra: Response factor (the inverse of the slope of the calibration
curve) for the analyte, Wa: weight fraction of the analyte in the sample;
Ms: mass of the sample (mg); Aa: Peak area of the analyte (mV.s); Cis:
concentration of the internal standard solution (mg/mL); Ais: Peak area Fig. 8. Multi-steady states triacetin conversions for transesterification in meso-
of the internal standard (mV.s); Vis: volume of the internal standard OBR at 50 °C using: (a) 6:1 methanol to triacetin molar ratio, AmberlsytTM A26-
used (mL). OH resin and ramped residence times of 0.5 min–30 min; (b) 30:1 methanol to
triacetin molar ratio, AmberlsytTM 70-SO3H resin and ramped residence times
3. Results and discussions of 2.5 min–60 min.

3.1. Steady states performance of the packed Meso-OBRs triacetin conversions were 28.5 ± 0.4% at 2.5 min residence time,
36.9 ± 0.5% at 5 min, 59.6 ± 0.4% at 10 min, 82.8 ± 0.6% at
Fig. 8 shows the steady state triacetin conversions for transester- 20 min, 95.2 ± 0.1% at 30 min and 99.8 ± 0.1% at 60 min, for the
ifications in meso-OBR at 50 °C using 6:1 methanol to triacetin molar continuous multi-steady states ramp at 30:1 methanol to triacetin molar
ratio with AmberlystTM A26-OH catalyst at ramped residence times of ratio and 50 °C reaction temperature. The small variance in the triacetin
0.5 min–30 min (Fig. 8(a)), and 30:1 methanol to triacetin molar ratio conversions at the steady states is an indication of plug flow behaviour
with AmberlsytTM 70-SO3H catalyst at ramped residence times of and good mixing in the meso-OBR. There was no reduction in the
2.5 min–60 min (Fig. 8(b)). The results clearly showed that a step- triacetin conversions in any of the AmberlystTM packed meso-OBRs on
change occurred between residence times, and that steady states were prolonged use, showing that these AmberlystTM resin catalysts have
achieved for all the ramped residence times investigated. This agrees high degree of catalytic stability and reusability, in agreement with
with an existing study on triglyceride transesterification in meso-OBR existing studies [7–9,12].
using homogeneous base-catalysed [37,39], and this demonstrates that Catalytic activities of the AmberlystTM resins used in this study were
the flows inside the AmberlystTM catalyst packed meso-OBR system compared with other ion-exchange resin catalysts in terms of their
were in the plug flow regime at the experimental conditions, resulting turnover frequency (TOF). TOF is defined as turnover number (TON)
in a tight control of residence time and effective mixing. per unit time, while TON is the mole of a reactant converted to product
Steady states were achieved quickly in the packed meso-OBRs with per mole of active sites of the catalyst. Due to the cumulative nature of
little variation in triacetin conversions at induction times in the range of TON, a better comparison of catalytic activities for heterogeneous cat-
1.2–1.4 times residence time. The time spent by a continuous flow re- alysts can only be obtained by evaluations of their TOF, usually cal-
actor in start-up stage is an important factor, as short start-up time culated at the initial reaction rates. TOF of AmberlystTM A26-OH cat-
reduces the amount of feedstock required, and the wastes generated at alyst for the triacetin transesterification was 61.2 × 10−3 s−1, more
the induction stage. The induction times obtained for the AmberlystTM than a factor of 12 higher than 5.0 × 10−3 s−1 for the AmberlystTM 70-
packed meso-OBR were consistent with 1.5 times residence reported for SO3H catalyst. As expected, the triacetin reaction proceeds slower with
a meso-OBR in homogeneous base-catalysed transesterification of ra- the AmberlystTM 70-SO3H catalyst (Fig. 9), a trend that has been ob-
peseed oil [39]. served for homogeneous acid and base catalysts. According to previous
Generally, the AmberlystTM A26-OH packed meso-OBR achieved studies [47,48], homogeneous base-catalysed triglyceride transester-
higher rates of triacetin conversion as shown in Fig. 8(a), than the ification proceed at higher reaction rates than acid catalysts, about
AmberlystTM 70-SO3H in Fig. 8(b). The triacetin conversions for the 4000 times faster [47].
AmberlystTM A26-OH increased from 32.0 ± 0.9% at 0.5 min re- As shown Table 1, the TOF for the AmberlystTM 70-SO3H catalyst in
sidence time, 80.8 ± 1.4% at 2.5 min, 92.7 ± 0.8% at 5 min, this work compares well with that of other −SO3H functionalised
97.9 ± 0.5% at 10 min, to 98.8 ± 1.4% at 20 min, for the continuous polymers. For instance, TOF in the range of 1.9 − 6.4 × 10−3 s−1 for
multi-steady states ramp at 6:1 methanol to triacetin molar ratio and Amberlys-15 and 2.4 − 58.6 × 10−3 s−1 for Nafion NR50 have been
50 °C reaction temperature. For the AmberlystTM 70-SO3H system, the reported [46]. The active sites of the AmberlystTM A26-OH and

46
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

3.2. Effects of water on AmberlystTM A26-OH and AmberlystTM 70-SO3H

packed Meso-OBRs

Although water is not a by-product of transesterification, a robust

catalytic system that could tolerate small amounts of water in the
feedstock is desirable. Most importantly, water is a by-product of the
reaction of glycerol and acetone. Therefore, it is necessary to ensure
that the reactively-formed water does not adsorb permanently on the
active sites of the heterogeneous catalyst. Fig. 10 shows the results of
the investigation on the effects of water on the AmberlystTM catalysts
for triacetin transesterification with anhydrous methanol, water-laden
methanol and controlled water spiking. Fig. 10(a) shows that the Am-
berlystTM A26-OH system was not water-tolerant, as any addition of
Fig. 9. Average steady states triacetin conversions for transesterification at water permanently deactivated its basic active sites. For instance, the
50 °C and various residence times, for a meso-OBRs packed with AmberlystTM steady state triacetin conversion decreased from 98.3 ± 1.1% with
A26-OH at 6:1 methanol to triacetin molar ratio, and a meso-OBR with anhydrous methanol, to 92.7 ± 1.0% after spiking with 0.5 mL of
AmberlystTM 70-SO3H at 30:1 methanol to triacetin molar ratio. distilled water, in a continuous triacetin transesterification using a
meso-OBR packed with AmberlystTM A26-OH at 50 °C, 10 min residence
and 6:1 methanol to triacetin molar ratio. When the anhydrous me-
AmberlystTM 70-SO3H used in this study were obtained from nitrogen thanol flow was switched to water-laden methanol (2.5 vol% water),
and sulphur contents analysis, respectively. The high TOF of the Nafion the triacetin conversion drastically reduced to 22.8 ± 1.3%, from the
NR50 compared to other −SO3H ion-exchange could be attributed to “new” steady state conversion of 92.7 ± 1.0% achieved after spiking
strong sulphonic acid sites present in the perfluorinated Nafion NR50 with 0.5 mL. The AmberlystTM A26-OH never recovered its full catalytic
resin polymer support. Currently, no TOF has been reported for Am- activity even after the methanol flow was switched back to anhydrous
berlystTM A26-OH in triglyceride (triacetin) transesterification, hence methanol, as only 27.4 ± 1.0% triacetin conversion was achieved.
the findings here cannot be compared with any existing study. In sharp contrast to the AmberlystTM A26-OH system, investigations
of effects of water on the AmberlystTM 70-SO3H showed that this cat-
alyst was tolerant to water (Fig. 10(b)). Steady state triacetin conver-
Table 1 sion of 62.8 ± 1.8% was obtained before spiking with 0.5 mL of water,
Turn over frequency of some catalysts. and 61.5 ± 0.7% after the water spiking, for transesterification over
Catalysts TOF (10−3 s−1) Reference
the AmberlystTM 70-SO3H packed meso-OBR at 50 °C, 10 min residence
and 30:1 methanol to triacetin molar ratio. The largest effects of the
AmberlystTM A26-OH 61.2a This study water spiking were observed at about 60 min running time, corre-
AmberlystTM 70-SO3H 5.0a This study sponding to 10 min (one residence time) after the water spiking with
Amberlyst-15 1.9b–6.4c [46]
triacetin conversion of 26.9%. The AmberlystTM 70-SO3H system re-
Nafion NR50 2.4b–58.6c [46]
covered its catalytic activity at 2 residence times after water spiking,
a
calculated using the active sites from elemental analysis (nitrogen and which was slightly longer than the induction times of 1.2–1.4 residence
sulphur), 2.59 mmol/g for AmberlystTM 70-SO3H and 4.21 mmol/g for times for the AmberlystTM 70-SO3H packed reactor. This suggests a
AmberlystTM A26-OH. competitive but reversible adsorption of water on the -SO3H active sites
b
obtained using active sites in swollen states. of the catalyst.
c
obtained using active sites from NH3 chemisorption. The triacetin conversions for the AmberlystTM 70-SO3H system
dropped from the average values of 62.8 ± 1.8% to 38.4 ± 0.6%

Fig. 10. Robustness of the meso-OBR platform for rapid screening of effects of water on continuous triacetin transesterification in meso-OBR at 50 °C and 10 min
residence using, (a) AmberlystTM A26 and 6:1 methanol/triacetin molar ratio, (b) AmberlystTM 70 and 30:1 methanol/triacetin molar ratio.

47
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

when the anhydrous methanol flow was switched to methanol con- 3.3. Reactive coupling of glycerol and acetone to produce solketal during
taining 2.5 vol% water. However, the AmberlystTM 70-SO3H catalyst transesterification
entirely recovered its catalytic activity after the water-laden methanol
flow was switched back to anhydrous methanol, achieving steady state Fig. 11 shows the glycerol-acetone reaction at 4:1 acetone to gly-
triacetin conversion of 61.6 ± 1.0%. This demonstrates that water cerol molar ratio (Fig. 11(a)), and the one-stage process for continuous
does not permanently deactivate the AmberlystTM 70-SO3H catalyst in triacetin transesterification at methanol-acetone-triacetin molar ratio of
triglyceride transesterification, consistent with previous study by the 6:4:1 (Fig. 11(b)), using a meso-OBR packed with AmberlystTM 70-
authors on the effects of water on the catalyst during carboxylic acid SO3H catalyst at 50 °C temperature and residence times of
esterification [42]. This catalyst could be used in reactions where wet 2.5 min–30 min. The direct glycerol condensation was used to evaluate
feedstock or reactively-formed water could be encountered, such as in the activity of the AmberlystTM 70-SO3H for acetalisation reaction. As
esterification process and acetalisation reaction of glycerol and acetone. shown in the Fig. 11(a), the glycerol conversion to solketal increased
Ideally, such reactions require water-tolerant acid catalysts to prevent rapidly from 46.9 ± 2.7% at 5 min to 74.0 ± 1.8% at 20 min re-
catalyst deactivation. One of the advantages of continuous screening in sidence time, and afterwards improved gradually. For instance, the
transesterification reactions where water could be encountered is that solketal formation was 80.6 ± 2.1% at 30 min residence time. The
in-flow of fresh methanol into the meso-OBR helps to prevent any ac- solketal conversion in this study compares well, with 86% yield re-
cumulation of water on the catalyst surface [6,42]. This is unlike batch ported for glycerol-acetone reaction at 6:1 M ratio using Amberlyst-35
experiments, where any water added into the system remained in the after 15 min reaction time [49], and solketal conversions of
reaction mixture and in contact with the catalyst. 75.5%–80.7% that were achieved after 30 min reaction time using Ar-
The observed permanent deactivation of the AmberlystTM A26-OH SBA-15 catalyst [50]. Findings in this study showed that reactive cou-
catalyst by water was unexpected, as it was believed that continuous in- pling of glycerol and acetone can be carried out in the presence of
flow of fresh methanol into the packed bed of AmberlystTM A26-OH methanol, consistent with reports elsewhere [51,52]. Therefore, an
would remove water adsorbed on the catalyst surface. Mechanism for integrated biodiesel process – producing fatty acid methyl esters and
the AmberlystTM A26-OH deactivation by water has not been reported. solketal through reactive coupling of glycerol with acetone is feasible in
Elemental analysis for carbon, nitrogen and sulphur (CNS) was per- biodiesel plants. Removal of the reactively-formed water from the
formed on the fresh and spent AmberlystTM resins to investigate any system may be required to overcome equilibrium limitations and shift
change in their compositions during reactions. Nitrogen content of the the reaction towards more solketal formation.
AmberlystTM A26-OH resin decreased from 5.9 wt% (∼4.21 mmol/g) in The one-stage process for triacetin transesterification at methanol-
fresh catalyst to 5.5 wt% (∼3.93 mmol/g) after deactivation by me- triacetin-acetone molar ratio of 6:1:4 over the AmberlystTM 70-SO3H
thanol containing 2.5 vol% water, while the carbon contents increased packed meso-OBR at 50 °C achieved 95.3 ± 2.0% triacetin to methyl
from 64.2 wt% to 66.1 wt%. On the contrary, the AmberlystTM 70-SO3H esters and 48.5 ± 2.7% solketal conversions after 30 min reaction time
showed no change in composition for the fresh and spent catalysts, as (Fig. 11(b)). The conversions of triacetin to glycerol at 30 min residence
the carbon and sulphur contents remained consistent at 43.2 wt% and time was 88.0 ± 2.8% without acetone, and this was reduced to
8.3 wt% (∼2.59 mmol/g), respectively. There was no substantial 32.1 ± 2.6% through reactive coupling with acetone. Lower solketal
change in the nitrogen content of the AmberlystTM A26-OH to account conversion in the one-stage process compared to the direct glycerol-
for the degree of deactivation observed, therefore, the deactivation acetone reaction could be attributed to the use of large excess of me-
process was not due to loss of active sites by leaching. The increase in thanol (methanol-triacetin-acetone of 30:1:4) which reduces the con-
the carbon content of the deactivated AmberlystTM A26-OH resin sug- centrations of the acetone and glycerol, and consequently their lower
gest that the catalyst was being poisoned through irreversible reactions reaction rates in the system. The one-stage process has an advantage
with an organic intermediate formed during water spiking. A possible that both the acetalisation and transesterification reactions are carried
mechanism for the permanent deactivation could be poisoning by acetic out simultaneously in one reactor. However, this process requires large
acid formed via aqueous hydrolysis of triacetin and methyl acetate. This excess of methanol and would take longer time due to the slower rate of
leads to replacement of the OH− ions on the AmberlystTM A26-OH acid-catalysed transesterification.
resin with acetate ions (CH3COO−), leading to an increased carbon Fig. 12 shows the results of the two-stage process for continuous
content of the deactivated catalyst. reactions of triacetin, methanol and acetone. The triacetin was reacted

Fig. 11. Plots of conversions versus re-

sidence time for reactions using
AmberlystTM 70-SO3H packed meso-OBR at
50 °C: (a) direct acetalisation of glycerol at
4:1 acetone to glycerol molar ratio to pro-
duce solketal; (b) triacetin transesterifica-
tion and reactive glycerol-acetone coupling
at methanol-triacetin-acetone molar ratio of
30:1:4, to form methyl ester and solketal.

48
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

Fig. 12. Two-stage transesterification

process: (a) 1st stage transesterification at
6:1 methanol to triacetin molar ratio and
60 °C using AmberlsytTM A26-OH packed
meso-OBR; (b) 2nd stage online reactive
coupling of glycerol from the 1st reactor
with acetone at a methanol-triacetin-
acetone molar ratio of 6:1:4 and 50 °C
using a meso-OBR packed with
AmberlsytTM 70-SO3H.

with methanol at 6:1 M ratio over AmberlystTM A26-OH catalyst in the glycerol-acetone reaction using the AmberlystTM 70-SO3H packed meso-
1st reactor, followed by online reactive coupling of the glycerol by- OBR.
product with acetone over AmberlystTM 70-SO3H catalyst in the 2nd The processes investigated in this study could fundamentally change
reactor. There was a rapid triacetin transesterification in the 1st reactor the conventional biodiesel production strategy. They have the potential
(Fig. 12(a)), achieving 97.7 ± 2.3% triacetin to methyl acetate and to reduce the cost of biodiesel processing through contemporaneous
91.7 ± 1.5% glycerol conversions at 10 min residence time. Further productions of solketal from glycerol. The methyl esters produced when
increase in the reaction time to 25 min resulted in near complete con- operating at high conversions would contain approximately 10 wt%
version of the triacetin to methyl ester (99.1 ± 2.0%) and glycerol solketal, which could be left as biodiesel fuel additive. Use of solketal as
(98.0 ± 1.3%). an additive reduces biodiesel’s viscosity [25], and leads to reduced gum
The output streams of the 2nd reactor showed rapid in-situ con- formation and 2.5 points enhancement in octane number when blended
versions of the glycerol to solketal (Fig. 12(b)), with solketal formation at 1–5 vol% in gasoline [26].
increasing linearly from 39.4 ± 1.4% at 2.5 min to 76.5 ± 2.8% at Considering the use of heterogeneous catalysts in both stages, this
35 min residence time. Glycerol co-production was reduced from process would therefore require very little downstream purification.
98.0 ± 1.3% conversion in the 1st reactor to 23.6 ± 2.1% in the 2nd
reactor (about 76% reduction in glycerol by-product) after 35 min 4. Conclusion
through reactive coupling with acetone. There was no indication at this
reaction time that the solketal formation was approaching equilibrium. A biodiesel process for contiunous transesterifcation of triglycerides
This strongly suggests that methanol could mitigate the effects of re- to form methyl esters, whilst simultaneously converting the glycerol by-
actively-formed water on the thermodynamic equilibrium of the solk- product into a useful product (solketal) using acetone is described. The
etal formation reactions. A study has shown that use of methanol could study considered two-stage and one-stage catalytic process routes. In
have the potential in reducing the cost of solketal production from the two-stage process, two interconnected meso-OBRs packed with
glycerol-acetone reaction [52]. The higher rates of solketal formation in amberlyst resin catalysts were used: a basic AmberlystTM A26-OH in the
the two-stage process (76.5%), as compared to the one-stage in 1st reactor to catalyse transesterfication of triacetin with methanol, and
(48.5%), are probably due to the application of lower methanol molar an acidic AmberlystTM 70-SO3H in the 2nd reactor to catalyse the online
ratio for the triacetin transesterification in the 1st reactor. Also, the use coupling of glycerol and acetone to form solketal. Triacetin was con-
of AmberlsytTM A26-OH catalyst in the 1st reactor achieved higher rates verted to 99.1 ± 2.0% methyl acetate and 98.0 ± 1.3% glycerol after
of glycerol formation compared to the AmberlsytTM 70-SO3H. These 25 min residence time in the 1st reactor. This was followed by online
resulted in higher concentration gradients for the glycerol and acetone reactive coupling of the glycerol with acetone in the 2nd reactor to
in the 2nd reactor, hence the increased rates of solketal production. As achieve solketal conversions of 76.5 ± 2.8% at 35 min residence times,
shown in Fig. 12(b), no substantial conversions of triacetin to methyl reducing the glycerol content of the output stream by about 76%. In the
ester occurred in the 2nd reactor, consequently, the triacetin and me- one-stage process, triacetin transesterification and glycerol condensa-
thyl ester conversions from the output streams of the 1st and 2nd re- tion with acetone were achieved in a meso-OBR packed with the
actors were similar. This was attributed to the lower rate of the acid- AmberlystTM 70-SO3H. The single-stage reaction of triacetin and me-
catalysed transesterification. The two-stage process is more resource- thanol in the presence of acetone achieved lower solketal conversion,
efficient, as it requires less methanol than the one-stage. Another ad- e.g. 48.5 ± 2.7% at 30 min residence time. This study demonstrates
vantage of the two-stage process is the possible reduction in reaction that triglyceride transesterification could be modified to produce me-
time due to the higher rates of base-catalysed triglyceride transester- thyl esters and solketal, minimising the co-production of glycerol. The
ification. The solketal conversion achieved in the two-stage process processes investigated here could be optimised to provide future bio-
(76.5 ± 2.8%) compares well with 80.6 ± 2.1% for the direct diesel production strategy.

49
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

The AmberlystTM resins are stable and can be reused as hetero- W.A. Carvalho, Effect of niobia and alumina as support for Pt catalysts in the hy-
geneous catalysts for triglyceride transesterification. However, the drogenolysis of glycerol, Chem. Eng. J. 198–199 (2012) 457–467.
[15] R. Ciriminna, C.D. Pina, M. Rossi, M. Pagliaro, Understanding the glycerol market,
AmberlystTM A26-OH system was not water-tolerant, as water caused Eur. J. Lipid Sci. Technol. 116 (2014) 1432–1439.
permanent deactivation of its basic active sites. The steady state tria- [16] J. George, Y. Patel, S.M. Pillai, P. Munshi, Methanol assisted selective formation of
cetin conversion decreased from 98.3 ± 1.1% with anhydrous me- 1,2-glycerol carbonate from glycerol and carbon dioxide using nBu2SnO as a cat-
alyst, J. Mol. Catal. Chem. 304 (2009) 1–7.
thanol, to 27.4 ± 0.6% after reactions with methanol containing [17] J.R. Ochoa-Gómez, O. Gómez-Jiménez-Aberasturi, B. Maestro-Madurga,
2.5 vol% water, for continuous triacetin transesterification using a A. Pesquera-Rodríguez, C. Ramírez-López, L. Lorenzo-Ibarreta, et al., Synthesis of
meso-OBR packed with AmberlystTM A26-OH at 50 °C, 10 min residence glycerol carbonate from glycerol and dimethyl carbonate by transesterification:
Catalyst screening and reaction optimization, Appl. Catal. A Gen. 366 (2009)
and 6:1 methanol to triacetin molar ratio. The AmberlystTM 70-SO3H 315–324.
catalyst was found to be tolerant to water. The steady state triacetin [18] M. Malyaadri, K. Jagadeeswaraiah, P.S. Sai Prasad, N. Lingaiah, Synthesis of gly-
conversions for this catalyst remained consistent before (62.8 ± 1.8%) cerol carbonate by transesterification of glycerol with dimethyl carbonate over Mg/
Al/Zr catalysts, Appl. Catal. A Gen. 401 (2011) 153–157.
and after (61.6 ± 1.0%) water spiking, for transesterification over the
[19] L. Li, T.I. Koranyi, B.F. Sels, P.P. Pescarmona, Highly-efficient conversion of gly-
AmberlystTM 70-SO3H packed meso-OBR at 50 °C, 10 min residence and cerol to solketal over heterogeneous Lewis acid catalysts, Green Chem. 14 (2012)
30:1 methanol to triacetin molar ratio. 1611–1619.
The meso-OBR platform allowed a wide range of experimental [20] N. Suriyaprapadilok, B. Kitiyanan, Synthesis of solketal from glycerol and Its re-
action with benzyl alcohol, Energy Procedia 9 (2011) 63–69.
conditions to be investigated continuously in a single experiment, [21] J.R. Dodson, T.d.C.M. Leite, N.S. Pontes, B. Peres Pinto, C.J.A. Mota, Green acet-
through multi-steady state screening. This rendered process develop- ylation of solketal and glycerol formal by heterogeneous acid catalysts to form a
ment faster than would be the case when using batch reactions where biodiesel fuel additive, ChemSusChem. 7 (2014) 2728–2734.
[22] R. Rodrigues, M. Goncalves, D. Mandelli, P.P. Pescarmona, W.A. Carvalho, Solvent-
separate experiments must be conducted to compare initial rates under free conversion of glycerol to solketal catalysed by activated carbons functionalised
different conditions. Most other continuous screening platforms would with acid groups, Catal. Sci. Technol. 4 (2014) 2293–2301.
not be able to accommodate the relatively large AmberlystTM resin [23] W.K. Teng, G.C. Ngoh, R. Yusoff, M.K. Aroua, A review on the performance of
glycerol carbonate production via catalytic transesterification: effects of influencing
beads. Use of the meso-OBRs in process development saved time and parameters, Energy Convers. Manage. 88 (2014) 484–497.
substantially reduced the amount of feedstock required and the waste [24] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Improved utilisation of
generated. renewable resources: new important derivatives of glycerol, Green Chem. 10 (2008)
13–30.
[25] J.A. Melero, G. Vicente, G. Morales, M. Paniagua, J. Bustamante, Oxygenated
Acknowledgements compounds derived from glycerol for biodiesel formulation: influence on EN 14214
quality parameters, Fuel 89 (2010) 2011–2018.
[26] C.J.A. Mota, C.X.A. da Silva, N. Rosenbach, J. Costa, F. da Silva, Glycerin deriva-
The authors would like thank Dow Chemical Company, Netherland,
tives as fuel additives: the addition of glycerol/acetone ketal (solketal) in gasolines,
for supplying the AmberlystTM A26-OH and AmberlystTM 70 resins used Energy Fuels 24 (2010) 2733–2736.
in this study. This work was supported by the UK Engineering and [27] C. Ramshaw, Process Intensification and green chemistry, Green Chem. 1 (1999)
Physical Sciences Research Council (EPSRC) through grant number EP/ G15–G17.
[28] Y.C. Lin, K.H. Hsu, J.F. Lin, Rapid palm-biodiesel production assisted by a micro-
K026216/1 (Cleaning Land for Wealth). wave system and sodium methoxide catalyst, Fuel 115 (2014) 306–311.
[29] C.C. Liao, T.W. Chung, Optimization of process conditions using response surface
References methodology for the microwave-assisted transesterification of Jatropha oil with
KOH impregnated CaO as catalyst, Chem. Eng. Res. Des. 91 (2013) 2457–2464.
[30] D. Kumar, G. Kumar, C.P. Singh Poonam, Fast, easy ethanolysis of coconut oil for
[1] G. Vicente, M. Martínez, J. Aracil, Integrated biodiesel production: a comparison of biodiesel production assisted by ultrasonication, Ultrason. Sonochem. 17 (2010)
different homogeneous catalysts systems, Bioresour. Technol. 92 (2004) 297–305. 555–559.
[2] M.E. Bambase, N. Nakamura, J. Tanaka, M. Matsumura, Kinetics of hydroxide- [31] L.T. Thanh, K. Okitsu, Y. Sadanaga, N. Takenaka, Y. Maeda, H. Bandow,
catalyzed methanolysis of crude sunflower oil for the production of fuel-grade Ultrasound-assisted production of biodiesel fuel from vegetable oils in a small scale
methyl esters, J. Chem. Technol. Biotechnol. 82 (2007) 273–280. circulation process, Bioresour. Technol. 101 (2010) 639–645.
[3] G.W. Huber, S. Iborra, A. Corma, Synthesis of transportation fuels from biomass: [32] J.A. Colucci, E. Borrero, F. Alape, Biodiesel from an alkaline transesterification
Chemistry, catalysts, and engineering, Chem. Rev. 106 (2006) 4044–4098. reaction of soybean oil using ultrasonic mixing, J. Am. Oil Chem. Soc. 82 (2005)
[4] B. Freedman, E.H. Pryde, T.L. Mounts, Variables affecting the yields of fatty esters 525–530.
from transesterified vegetable oils, J. Am. Oil Chem. Soc. 61 (1984) 1638–1643. [33] Z. Qiu, J. Petera, L.R. Weatherley, Biodiesel synthesis in an intensified spinning disk
[5] P.T. Anastas, M.M. Kirchhoff, T.C. Williamson, Catalysis as a foundational pillar of reactor, Chem. Eng. J. 210 (2012) 597–609.
green chemistry, Appl. Catal. A: Gen. 221 (2001) 3–13. [34] E.E. Kalu, K.S. Chen, T. Gedris, Continuous-flow biodiesel production using slit-
[6] V.C. Eze, A.N. Phan, C. Pirez, A.P. Harvey, A.F. Lee, K. Wilson, Heterogeneous channel reactors, Bioresour. Technol. 102 (2011) 4456–4461.
catalysis in an oscillatory baffled flow reactor, Catal. Sci. Technol. 3 (2013) [35] Z. Wen, X. Yu, S.-T. Tu, J. Yan, E. Dahlquist, Intensification of biodiesel synthesis
2373–2379. using zigzag micro-channel reactors, Bioresour. Technol. 100 (2009) 3054–3060.
[7] N. Shibasaki-Kitakawa, H. Honda, H. Kuribayashi, T. Toda, T. Fukumura, [36] J. Sun, J. Ju, L. Ji, L. Zhang, N. Xu, Synthesis of biodiesel in capillary microreactors,
T. Yonemoto, Biodiesel production using anionic ion-exchange resin as hetero- Ind. Eng. Chem. Res. 47 (2008) 1398–1403.
geneous catalyst, Bioresour. Technol. 98 (2007) 416–421. [37] A.N. Phan, A.P. Harvey, M. Rawcliffe, Continuous screening of base-catalysed
[8] C.E.T. Co, M.C. Tan, J.A.R. Diamante, L.R.C. Yan, R.R. Tan, L.F. Razon, Internal biodiesel production using new designs of mesoscale oscillatory baffled reactors,
mass-transfer limitations on the transesterification of coconut oil using an anionic Fuel Process. Technol. 92 (2011) 1560–1567.
ion exchange resin in a packed bed reactor, Catal. Today 174 (2011) 54–58. [38] A.P. Harvey, M.R. Mackley, T. Seliger, Process intensification of biodiesel produc-
[9] M.G. Falco, C.D. Córdoba, M.R. Capeletti, U. Sedran, Basic ion exchange resins as tion using a continuous oscillatory flow reactor, J. Chem. Technol. Biotechnol. 78
heterogeneous catalysts for biodiesel synthesis, Adv. Mater. Res. 132 (2010) (2003) 338–341.
220–227. [39] A.N. Phan, A.P. Harvey, V. Eze, Rapid production of biodiesel in mesoscale oscil-
[10] Y. Feng, A. Zhang, J. Li, B. He, A continuous process for biodiesel production in a latory baffled reactors, Chem. Eng. Technol. 35 (2012) 1214–1220.
fixed bed reactor packed with cation-exchange resin as heterogeneous catalyst, [40] A. Orjuela, A.J. Yanez, A. Santhanakrishnan, C.T. Lira, D.J. Miller, Kinetics of mixed
Bioresour. Technol. 102 (2011) 3607–3609. succinic acid/acetic acid esterification with Amberlyst 70 ion exchange resin as
[11] B.M.E. Russbueldt, W.F. Hoelderich, New sulfonic acid ion-exchange resins for the catalyst, Chem. Eng. J. 188 (2012) 98–107.
preesterification of different oils and fats with high content of free fatty acids, Appl. [41] P.F. Siril, H.E. Cross, D.R. Brown, New polystyrene sulfonic acid resin catalysts with
Catal. A Gen. 362 (2009) 47–57. enhanced acidic and catalytic properties, J. Mol. Catal. A Chem. 279 (2008) 63–68.
[12] S.M. de Rezende, M. de Castro Reis, M.G. Reid, P. Lúcio Silva Jr., F.M.B. Coutinho, [42] V.C. Eze, J.C. Fisher, A.N. Phan, A.P. Harvey, Intensification of carboxylic acid
R.A. da Silva San Gil, E.R. Lachter, Transesterification of vegetable oils promoted by esterification using a solid catalyst in a mesoscale oscillatory baffled reactor plat-
poly(styrene-divinylbenzene) and poly(divinylbenzene), Appl. Catal. A Gen. 349 form, Chem. Eng. J. 322 (2017) 205–214.
(2008) 198–203. [43] R. Tesser, M. Di Serio, L. Casale, G. Carotenuto, E. Santacesaria, Absorption of
[13] C.S. MacLeod, A.P. Harvey, A.F. Lee, K. Wilson, Evaluation of the activity and water/methanol binary system on ion-exchange resins, Can. J. Chem. Eng. 88
stability of alkali-doped metal oxide catalysts for application to an intensified (2010) 1044–1053.
method of biodiesel production, Chem. Eng. J. 135 (2008) 63–70. [44] E. Van de Steene, J. De Clercq, J.W. Thybaut, Ion-exchange resin catalyzed trans-
[14] R. Rodrigues, N. Isoda, M. Gonçalves, F.C.A. Figueiredo, D. Mandelli, esterification of ethyl acetate with methanol: gel versus macroporous resins, Chem.

50
V.C. Eze, A.P. Harvey Chemical Engineering Journal 347 (2018) 41–51

Eng. J. 242 (2014) 170–179. [49] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.-A. Poirier, C. Xu, A new con-
[45] C.N. Fan, C.H. Xu, C.Q. Liu, Z.Y. Huang, J.Y. Liu, Z.X. Ye, Catalytic acetalization of tinuous-flow process for catalytic conversion of glycerol to oxygenated fuel ad-
biomass glycerol with acetone over TiO2-SiO2 mixed oxides, React. Kinet. Mech. ditive: catalyst screening, Appl. Energy 123 (2014) 75–81.
Catal. 107 (2012) 189–202. [50] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martin, Acetalisation of bio-
[46] D.E. López, J.G. Goodwin Jr, D.A. Bruce, E. Lotero, Transesterification of triacetin glycerol with acetone to produce solketal over sulfonic mesostructured silicas,
with methanol on solid acid and base catalysts, Appl. Catal. A. Gen. 295 (2005) Green Chem. 12 (2010) 899–907.
97–105. [51] C.X.A. da Silva, C.J.A. Mota, The influence of impurities on the acid-catalyzed re-
[47] J.M. Cervero, J. Coca, S. Luque, Production of biodiesel from vegetable oils, Grasas action of glycerol with acetone, Biomass Bioenergy 35 (2011) 3547–3551.
Aceites 59 (2008) 76–83. [52] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.A. Poirier, C. Xu, Catalytic con-
[48] B. Freedman, R.O. Butterfield, E.H. Pryde, Transesterification kinetics of soybean version of glycerol to oxygenated fuel additive in a continuous flow reactor: process
oil 1, J. Am. Oil Chem. Soc. 63 (1986) 1375–1380. optimization, Fuel 128 (2014) 113–119.

51
 Energy Conversion and Management 160 (2018) 251–261

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Multi-objective exergy-based optimization of continuous glycerol T

ketalization to synthesize solketal as a biodiesel additive in subcritical
acetone
⁎ ⁎ ⁎
Mortaza Aghbashloa, , Meisam Tabatabaeib,c, , Soleiman Hosseinpoura, , Hajar Rastegarid,
Hassan S. Ghaziaskard
a
Department of Mechanical Engineering of Agricultural Machinery, Faculty of Agricultural Engineering and Technology, College of Agriculture and Natural Resources,
University of Tehran, Karaj, Iran
b
Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), P.O. Box: 31535-1897, Agricultural Research, Education, and
Extension Organization (AREEO), Karaj, Iran
c
Biofuel Research Team (BRTeam), Karaj, Iran
d
Department of Chemistry, Isfahan University of Technology, Isfahan, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: This study was aimed at exergetically investigating and optimizing a continuous reactor applied to valorize
Adaptive neuro-fuzzy inference system glycerol into solketal as a biodiesel additive with subcritical acetone in the presence of Purolite PD206. The
Biodiesel additive effects of reaction temperature (20–100 °C), acetone to glycerol molar ratio (1–5), feed flow rate (0.1–0.5 mL/
Exergy analysis min), pressure (1–120 bar), and catalyst mass (0.5–2.5 g) were evaluated on the exergetic performance para-
Glycerol conversion
meters of the reactor. In order to optimize the operating conditions of the reactor, adaptive neuro-fuzzy in-
Non-dominated sorting genetic algorithm-II
ference system (ANFIS) was coupled with non-dominated sorting genetic algorithm-II (NSGA-II). The ANFIS was
Solketal synthesis
applied to develop objective functions on the basis of the process parameters. The developed objective functions
were then fed into the NSGA-II to find the optimum operating conditions of the process by simultaneously
maximizing universal and functional exergetic efficiencies and minimizing normalized exergy destruction.
Overall, the process parameters significantly affected the exergetic performance of the reactor. The ANFIS ap-
proach successfully modeled the objective functions with a correlation coefficient higher than 0.99. The optimal
ketalization conditions of glycerol were: reaction temperature = 40.66 °C, acetone to glycerol molar
ratio = 4.97, feed flow rate = 0.49 mL/min, pressure = 42.31 bar, and catalyst mass = 0.50 g. These conditions
could be applied in pilot- or industrial-scale reactors for converting glycerol into value-added solketal in a
resource-efficient, cost-effective, and environmentally-friendly manner.

1. Introduction up to 10-folds [4]. To improve the economic features of biodiesel in-

dustry, upgrading glycerol into different value-added chemicals and
Over the last few years, biofuels have drawn an increasing deal of making use of glycerol-containing waste streams are crucial.
attention as alternatives to fossil fuels due to their environmental There are already more than 1500 direct applications of glycerol,
benefits and the fact that they are produced from renewable feedstocks particularly in cosmetics, pharmaceuticals, and food industries [5].
[1]. Among the various biofuels commercialized to date, biodiesel has Nevertheless, there has been intense research interest towards finding
shown a great potential to supplement petrodiesel [2,3]. Consequently, methods to upgrade glycerol into more valuable renewable chemicals.
the production of biodiesel has increased substantially since the year For instance, several well-known value-added derivatives of glycerol
2000 especially to satisfy the stringent emission mandates for diesel are esters [6], acetals [7], and ketals [8]. These chemicals can poten-
engines and to partially meet the increasing global demand for diesel tially be used as oxygenated fuel additives to improve diesel/biodiesel
fuel. This in turn has also led to a glycerol surplus as a by-product of combustion and emissions. Among these glycerol derivatives, solketal
biodiesel synthesis, negatively affecting glycerol price dramatically by (4-hydroxymethyl-2,2-dimethyl-1,3-dioxolane) produced from glycerol

⁎
Corresponding authors at: Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), P.O. Box: 31535-1897, Agricultural Research,
Education, and Extension Organization (AREEO), Karaj, Iran (M. Tabatabaei).
E-mail addresses: [email protected] (M. Aghbashlo), [email protected] (M. Tabatabaei), [email protected] (S. Hosseinpour).

https://doi.org/10.1016/j.enconman.2018.01.044
Received 4 November 2017; Received in revised form 3 January 2018; Accepted 19 January 2018
0196-8904/ © 2018 Elsevier Ltd. All rights reserved.
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Nomenclature Greek letters

 ash percentage (%) ψ universal exergetic efficiency (%)

C catalyst mass (g) ϕ functional exergetic efficiency (%)
 carbon percentage (%) Ψ normalized exergy destruction (–)
Cp specific heat capacity (kJ/kg K)
ex specific chemical exergy (kJ/kg) Subscripts
Eẋ exergy flow rate (kJ/s)
F feed flow rate (mL/min) 0 dead state
 hydrogen percentage (%) AC acetone
M molar mass (g/mol) ch chemical
ṁ mass flow rate (kg/s) des destruction
ṅ molar flow rate (mol/s) ET ethanol
 nitrogen percentage (%) GL glycerol
 oxygen percentage (%) in inflow
P pressure (bar) i, j numerator
R universal gas constant (kJ/mol k) l loss
 sulfur percentage (%) out outflow
T temperature (°C or K) P product
x mass fraction (–) ph physical
X acetone to glycerol molar ratio (–) q heat transfer
y molar fraction (–) SK solketal
Ẇ work flow rate (kJ/s) WT water

and acetone has attracted more attention recently due to its unique [26–31]. According to these studies, exergy analysis could provide a
features in improving cold flow properties and flash point temperatures better understanding of the effect of process parameters on the ther-
of both diesel and biodiesel as well as their blends [9]. Solketal can also modynamic performance and can aid with the diagnosis of the most
be used as a solvent and plasticizer, suspension agent in pharmaceutical effective strategies to improve the process under investigation.
preparations, and anti-freezing agent [10]. A number of research works have been reported on the production
Solketal is conventionally produced through ketalization of glycerol of value-added chemicals from glycerol with their focus on technical
with acetone by using strong Brønsted acid catalysts like sulfuric acid, aspects and kinetics simulation of the reported processes [18,32,33].
hydrochloric acid, phosphoric acid, and para-toluene sulfonic acid in However, to the best of our knowledge, there is no report so far on the
solvents such as chloroform, ether, or oil [11]. This procedure suffers exergy analysis of a continuous or even a batch reactor applied for
from some serious drawbacks such as equipment corrosion, effluent synthesizing chemicals from glycerol. Therefore, solketal production
disposal, and difficulties in separation of the catalysts from the products from glycerol using a continuous reactor with subcritical acetone in the
which in turn increase the production cost and result in environmental presence of Purolite PD206 catalyst was exergetically analyzed in the
burdens. These issues can be addressed by using heterogeneous acidic present study. According to our previous report, this heterogeneous
catalysts like nafion [12], zeolite [13], Amberlyst [14], silica containing acidic catalyst could effectively upgrade glycerol to solketal [18].
heteropolyacids [15], and montmorillonite [16]. Batch process has Therefore, the main aim of the present survey was to enhance our un-
been used for solketal synthesis in the above-mentioned studies. How- derstanding of the ketalization step in the developed continuous reactor
ever, this mode of production has several major limitations such as long exergetically. More specifically, the effects of process parameters viz.
reaction time, high energy consumption, poor yield, and difficulty in reaction temperature (20–120 °C), reaction pressure (1–120 bar),
scale-up [17]. Therefore, in order to solve these issues, Shirani et al. acetone to glycerol molar ratio (1–5), feed flow rate (0.1–0.5), and
[18] developed an easy to scale-up continuous system for upgrading catalyst loading (0.5–2.5 g) on the exergetic performance parameters of
glycerol into solketal in the presence of Purolite PD206 as a hetero- the reactor were comprehensively assessed and discussed. Moreover, a
geneous catalyst. Even though the developed system efficiently syn- multi-objective exergy-based optimization was carried out to find the
thesized solketal from glycerol, advanced engineering paradigms such optimum operating conditions of the reactor using coupled adaptive
as exergy should still be used to assess the productivity and sustain- neuro-fuzzy inference system (ANFIS) and non-dominated sorting ge-
ability of such systems. netic algorithm-II (NSGA-II) approaches. The ANFIS approach was used
During the past two decades, there has been an increasing interest in to develop three objective functions, i.e., normalized exergy destruc-
applying exergy analysis for scrutinizing energy and material conver- tion, universal exergetic efficiency, and functional exergetic efficiency
sion systems from productivity and sustainability viewpoints [19–21]. as a function of the process parameters. The developed objective
Unlike energy analysis which is based on the first law of thermo- functions were then fed into the NSGA-II algorithm to simultaneously
dynamics, exergy analysis accounts for the degradation of the energy maximize universal and functional exergetic efficiencies and minimize
quality by irreversible processes [22]. In better words, by using the first normalized exergy destruction.
and second laws of thermodynamics simultaneously, exergy analysis
compensates for the shortcoming of energy analysis in revealing the
2. Materials and methods
energy quality loss because of thermodynamic imperfections [23].
Generally, exergy is not quantitatively conserved like energy but it is
2.1. Materials
destroyed due to irreversibilities within a system [24]. The quantity of
destroyed exergy is a measure of environmental pollution costs, pro-
Glycerol and acetone (both ≥99%) were purchased from Merck Co.
viding a quantitative comparison of environmental impacts [25].
(Germany). Solketal (≥97%) for GC calibration was obtained from
Hence, the exergy analysis has been widely used in recent years for
Sigma-Aldrich Co. (Germany). Absolute ethanol (ET) was supplied by
analyzing and optimizing various glycerol upgrading processes
Bidestan Co. (Iran). Toluene (≥99.9%) was also obtained from Merck.

252
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Co (Germany). Purolite PD206 was purchased from Purolite Co. (USA). ṁ GL,in + ṁ AC,in + ṁ ET ,in = ṁ SK ,out + ṁ GL,out + ṁ AC ,out + ṁ ET ,out
All the chemicals used herein were of analytical grade.
+ ṁ WT ,out (2)
The exergy balance for the system could be expressed as follows:
2.2. Experimental procedure and analytical method
Eẋin + Ẇ = Eẋ out + Eẋ q,l + Eẋ des (3)
Synthesis of solketal was carried out in a bench-scale continuous
The exergy flow rate of the feed and product streams could be
flow tubular reactor heated in an air oven. The reactor was made of a
computed on the basis of the summation of their chemical and physical
316-stainless steel tube with an 8 mm i.d., 30 cm length, filled with a
exergies. Accordingly, the above-mentioned exergetic balance equation
specific amount of catalyst. The catalyst was mixed with crushed Pyrex
could be rewritten as follows:
glass (CPG, mesh 10–20). The feed was pumped through an HPLC pump
(Shimadzu model LC-6A) at a specific flow rate into the reactor at a ch ph ch ph
Eẋ in + Eẋ in + Ẇ = Eẋ out + Eẋ out + Eẋ q,l + Eẋ des (4)
fixed temperature and pressure. The feed consisted of a homogenous
solution of reactants (glycerol and acetone) and ethanol. Since the so- The heat loss rate from the oven frame to the surroundings could be
lubility of glycerol in acetone is very low, ethanol was used as co-sol- neglected according to the outcomes of energy analysis. Therefore, the
vent mainly to improve its solubility in acetone. Otherwise, glycerol exergy balance equation can be measured as follows:
and acetone should have been separately fed into the reactor, in- ch ph ch ph
Eẋ in + Eẋ in + Ẇ = Eẋ out + Eẋ out + Eẋ des (5)
creasing the operating costs and leading to difficulties in the process
control. The controlling system of the air oven was used to adjust the The chemical and physical exergies of the feed and product streams
reaction temperature with an accuracy of ± 1.0. In addition, pressure could be determined using the following expressions, respectively:
was controlled through a back-pressure regulator (KPB, Swagelok,
G.B.). In each run, a pre-determined amount of Purolite PD206 was ⎛ ⎞
Eẋ ch = ṅ ⎜∑ yi εi + RT0 ∑ yi ln(yi )⎟
preloaded into the reactor. The effect of catalyst mass in the range of ⎝ i i ⎠ (6)
0.5–2.5 g was studied on the exergetic performance parameters of the
reactor. This range was considered on the basis of glycerol mass being T
̇ p ⎛T −T0−T0ln ⎞
Eẋ ph = mC ⎜ ⎟

accumulated in the reactor. It should be noted that the lower level of ⎝ T0 ⎠ (7)
catalyst mass, i.e., 0.5 g was considered based on our preliminary ex-
The following formula was used to determine the specific heat ca-
periments [18]. Accordingly, catalyst mass values below 0.5 g did not
lead to an acceptable yield due to the short length of the catalytic bed in pacity of the outflow stream.
the reactor. During each experiment, four samples were taken at 30 min Cp = ∑ x i Cp,i
intervals with the collection efficiency of more than 95%. The sche- i (8)
matic experimental set-up used for the continuous glycerol ketalization
The standard chemical exergy values of the organic components
is shown in Fig. 1. The chemical reaction occurring during the glycerol
involved in the process viz. glycerol, acetone, and solketal were cal-
ketalization with acetone is also shown in Fig. 2.
culated using the mathematical model developed by Song et al. [34].
Samples were analyzed using a GC-FID (Agilent technologies model
6890 N). The carrier gas was nitrogen and a polar capillary column of ex i = 363.439 + 1075.633 −86.308 + 4.14 + 190.798 −21.1 (9)
BP20 (30 m length, 0.25 mm i.d., and 0.25 μm film thickness) was used.
εi = Mi ex i (10)
The GC injection port and the detector temperature were set at 240 °C
and 260 °C, respectively. The temperature program used for the analysis The standard chemical exergy values of water and ethanol were
included an initial column temperature of 50 °C for 3.5 min, which was obtained from the published literature [35]. Table 1 summarizes the
subsequently increased from 50 °C to 260 °C at the rate of 40 °C/min. chemical formulas and standard chemical exergy values of the com-
Finally, the temperature was maintained at 260 °C for 3.5 min. Some ponents of feed and product streams.
standard solutions containing toluene as an internal standard were in- Solketal was the only product of the process. Accordingly, the ex-
jected and the peak areas were integrated to establish the calibration ergy rate of the product was equal to the solketal exergy rate.
curves. Identification of the products was carried out by gas chroma- ch
tography/mass spectrometry (GC/MS; model 6890N, Agilent Eẋ P = Eẋ SK ,out (11)
Technologies).
The universal exergetic efficiency of the glycerol ketalization

2.3. Theoretical considerations

The assumptions listed below were taken into consideration

throughout this study in developing the exergy model:

• The reactor operated at a steady state condition.

• The kinetic and potential exergies of the reactants and products
were neglected.
• The dead state temperature and pressure were taken into account to
be 25 °C and 1.01 bar, respectively.
• The dead state model did not vary with time.
According to Fig. 2, the mass flow rate balance for the system could
be written as follows:
Fig. 1. The experimental set-up used for ketalization of glycerol in subcritical acetone. (1)
ṁ in = ṁ out (1) Analytical balance, (2) Feed vessel, (3) HPLC pump, (4) On-off valve, (5) Air oven, (6)
Tubular reactor, (7) Cooling jacket, (8) Back-pressure regulator, and (9) Sample collection
vessel.
Or in more detailed form,

253
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Fig. 2. The chemical reaction occurring during

the glycerol ketalization with acetone.

Table 1 process was measured as:

Chemical formulas and standard chemical exergy values of the components of feed and
ch ph
product streams. Eẋ out Eẋ out + Eẋ out Eẋ des
ψ= = × 100 = 1− ch × 100
̇
Ex in ch ph
Eẋ in + Eẋ in + Ẇ
ph
Eẋ in + Eẋ in + Ẇ
Component Chemical formula Standard chemical exergy (kJ/mole)
(12)
Glycerol C3H8O3 1762.63
Acetone C3H6O 1821.96 The functional exergetic efficiency of the solketal synthesis was
Solketal C6H12O3 3505.85 determined as follows:
Ethanol C2H6O 1385.45 ch
Water H2O 0.95 Eẋ SK ,out
ϕ= ch ph
× 100
Eẋ in + Eẋ in + Ẇ (13)

Fig. 3. The first order Takagi-Sugeno neuro-fuzzy

system used throughout this study.

254
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

The normalized exergy destruction of the process was determined to Table 2

make decisions on the process parameters by taking into account both Specifications of NSGA-II optimization procedure.
the quantities of the exergy destroyed during the ketalization process
Parameter Value
and the exergy rate of the solketal synthesized as follows:
Objectives min normalized exergy destruction
Eẋ des max universal exergetic efficiency
Ψ= ch
Eẋ SK ,out (14) max functional exergetic efficiency
Variables X1 = Pressure, X2 = temperature, X3 = Molar ratio
X4 = Feed flow rate, X5 = Catalyst mass
Bounds 1 ≤ X1 ≤ 120, 20 ≤ X2 ≤ 100, 1 ≤ X3 ≤ 5, 0.1 ≤ X4 ≤ 0.5,
2.4. Developing objective functions and optimization procedure 0.5 ≤ X5 ≤ 2.5
Population type Double vector
Population size 50
In order to optimize the glycerol ketalization process, robust and
Selection function Tournament
accurate objective functions should first be built. Developing objective Crossover fraction 0.8
functions for multi-objective optimization of complex chemical reactors Mutation fraction 0.02
using the conventional approaches is very difficult. However, the ANFIS Stopping criteria Generations: 500/stall generations: 100
approach could satisfactorily establish the required objective functions
because of its capability to model the process under investigation
algorithm and its extensions have been introduced and applied during
without the need for any priori information. The theoretical background
the past decades [37]. Among the various extensions of the genetic
behind the ANFIS approach and its modeling procedure can be found in
algorithm, the NSGA-II has found many applications for efficiently
the published literature [36]. The first order Takagi-Sugeno neuro-fuzzy
solving the complex optimization problems. The implementation pro-
system was used throughout this study to build the objective functions
cedure of the NSGA-II algorithm can be found elsewhere [38]. In order
as manifested in Fig. 3. The best ANFIS models for estimating the
to optimize the glycerol ketalization process, the developed FIS models
universal exergetic efficiency (FIS1), functional exergetic efficiency
were fed into the NSGA-II as shown in Fig. 4. The optimization was
(FIS2), and normalized exergy destruction (FIS3) were determined on
carried out by maximizing the universal and functional exergy effi-
the basis of three statistical criteria viz. maximum coefficient of de-
ciencies and minimizing the normalized exergy destruction, simulta-
termination (R2), minimum mean absolute percentage error (MAPE),
neously.
and minimum mean-squared error (MSE) by trial-and-error method.
In order to model and optimize the glycerol ketalization process, a
Overall, multi-objective optimization of chemical processes is not
program was developed in MATLAB software (ver. R2008b). The Pareto
straightforward due to the existence of conflicting objectives. In order
optimal front set for the process was obtained as the best operating
to address this issue, evolutionary computation techniques like genetic

Fig. 4. Schematic representation of the developed approach for multi-objective exergy-based optimization of the glycerol ketalization process.

255
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

conditions. Table 2 summarizes the specifications of the developed could be attributed to the fact that small bead-size catalysts such as
NSGA-II algorithm achieved after some trials and errors. Purolite could aggregate at higher loadings, lowering the catalyst ac-
tivity and declining the solketal formation rate. Increasing reaction
3. Results and discussion temperature decreased the solketal exergy rate since the concentration
of acetone as a reactant was markedly decreased in the liquid phase at
Table 3 tabulates the mass balance of the glycerol ketalization with higher reaction temperatures. In better words, due to the low boiling
subcritical acetone in the continuous reactor under investigation. temperature of acetone (56 °C), increasing reaction temperature sig-
Table 4 also summarizes the exergetic performance parameters of the nificantly decreased the solketal evolution rate. It is worth mentioning
continuous reactor applied for glycerol ketalization at different reaction that the glycerol ketalization process with acetone is a reversible re-
temperatures, acetone to glycerol molar ratios, feed flow rates, pres- action [18]. The water evolved during the reaction imposed a ther-
sures, and catalyst loadings. The maximum solketal exergy rate was modynamic barrier, hindering the reaction in the forward direction.
found to be 55.27 W at pressure of 60 bar, reaction temperature of However, acetone concentration as a reactant in the liquid phase and,
60 °C, acetone to glycerol molar ratio of 3, feed flow rate of 0.5 mL/min, consequently on the catalyst surface, was increased at lower reaction
and catalyst loading of 1.5 g. The minimum solketal exergy rate was temperatures owing to its low vapor pressure. These in turn increased
determined as 10.61 W at pressure of 60 bar, acetone to glycerol molar the formation of solketal at lower reaction temperatures. Furthermore,
ratio of 3, feed flow rate of 0.3 mL/min, reaction temperature of 60 °C, increasing reaction pressure had a slight effect on the solketal exergy
and catalyst loading of 2.5 g. The importance of process parameters on rate since this process parameter could not significantly affect the
the solketal exergy rate could be ranked in the following order: feed viscosity and diffusion coefficients of the liquid reaction media.
flow rate > acetone to glycerol molar ratio > catalyst loading > According to the data presented in Table 4, exergy destruction rate
reaction temperature > pressure. Increasing feed flow increased of the glycerol ketalization process varied between 102.42 and
solketal exergy rate, while this parameter was decreased with in- 307.88 W. The order of the effect of process parameters on the exergy
creasing acetone to glycerol molar ratio, catalyst loading, reaction destruction rate was feed flow rate > reaction temperature >
temperature, and pressure. pressure > acetone to glycerol molar ratio > catalyst loading. In-
The rate of product formation increased at higher feed flow rates, creasing feed flow rate profoundly increased exergy destruction rate
resulting in an increase in the solketal exergy rate. In better words, because of an increase in the rate of ketalization reaction. The high-
there was a direct association between feed flow rate and solketal ex- quality electrical energy consumption was significantly increased with
ergy rate. By increasing acetone to glycerol molar ratio, the feeding rate increasing reaction temperature and pressure. In better words, a large
of glycerol into the reactor was decreased. This in turn lowered the rate amount of electrical energy was converted into low quality thermal and
of solketal formation and its exergetic rate. It was envisaged that the thermomechanical energies at higher reaction temperatures and pres-
solketal exergy rate would tend to increase by increasing catalyst sures. Accordingly, the rate of thermodynamic irreversibilities tended
loading because of promoting the ketalization process. In contrast, the to increase by elevating reaction temperature and pressure. Unlike the
solketal exergy rate was decreased by increasing catalyst loading. This other process parameters, increasing acetone to glycerol molar ratio

Table 3
Mass balance of the continuous glycerol ketalization process.

Run Pressure (bar) Temperature (°C) Molar ratio (–) Flow rate (g/min) Catalyst mass (g) Input (g/h) Output (g/h)

Glycerol Acetone Ethanol Glycerol Acetone Ethanol Solketal Water

1 120 60 3 0.3 1.5 6.12 11.52 0.36 3.42 9.90 0.36 3.78 0.54
2 30 40 2 0.2 2.0 5.16 6.60 0.24 3.00 5.16 0.24 3.12 0.48
3 60 60 3 0.3 1.5 6.12 11.52 0.36 2.52 9.18 0.36 5.22 0.72
4 60 100 3 0.3 1.5 6.12 11.52 0.36 4.32 10.44 0.36 2.52 0.36
5 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.18 0.36 5.40 0.72
6 90 80 4 0.2 1.0 3.36 8.52 0.12 1.80 7.44 0.12 2.28 0.36
7 60 60 3 0.3 0.5 6.12 11.52 0.36 4.32 10.44 0.36 2.52 0.36
8 30 80 2 0.2 1.0 5.16 6.60 0.24 3.12 5.28 0.24 2.88 0.48
9 60 60 3 0.1 1.5 2.04 3.84 0.12 0.60 3.00 0.12 1.98 0.30
10 30 40 4 0.2 1.0 3.36 8.52 0.12 1.44 7.20 0.12 2.88 0.36
11 90 80 2 0.4 1.0 10.32 13.20 0.48 6.00 10.32 0.48 6.24 0.96
12 30 80 4 0.2 2.0 3.36 8.52 0.12 1.68 7.44 0.12 2.40 0.36
13 30 80 2 0.4 2.0 10.32 13.20 0.48 5.76 10.08 0.48 6.72 0.96
14 60 20 3 0.3 1.5 6.12 11.52 0.36 2.88 9.72 0.36 4.32 0.72
15 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.36 0.36 5.22 0.72
16 90 80 2 0.4 2.0 10.32 13.20 0.48 9.36 12.24 0.48 1.68 0.24
17 30 40 2 0.4 1.0 10.32 13.20 0.48 6.72 10.80 0.48 5.28 0.72
18 60 60 3 0.3 2.5 6.12 11.52 0.36 5.04 10.98 0.36 1.44 0.18
19 30 80 4 0.4 1.0 6.72 17.04 0.24 4.80 15.84 0.24 2.64 0.48
20 1 60 3 0.3 1.5 6.12 11.52 0.36 1.80 9.00 0.36 5.94 0.90
21 60 60 5 0.3 1.5 4.32 13.50 0.18 1.80 11.88 0.18 3.60 0.54
22 30 40 4 0.4 2.0 6.72 17.04 0.24 3.12 14.40 0.24 5.52 0.72
23 90 40 4 0.2 2.0 3.36 8.52 0.12 1.20 7.20 0.12 3.00 0.48
24 60 60 3 0.5 1.5 10.20 19.20 0.60 4.80 15.90 0.60 7.50 1.20
25 90 80 4 0.4 2.0 6.72 17.04 0.24 3.84 15.12 0.24 4.08 0.72
26 90 40 4 0.4 1.0 6.72 17.04 0.24 2.64 14.40 0.24 5.76 0.96
27 90 80 2 0.2 2.0 5.16 6.60 0.24 3.72 5.52 0.24 2.16 0.36
28 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.18 0.36 5.40 0.72
29 60 60 3 0.3 1.5 6.12 11.52 0.36 2.52 9.18 0.36 5.22 0.72
30 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.18 0.36 5.40 0.72
31 60 60 1 0.3 1.5 10.80 6.66 0.54 5.94 3.96 0.54 6.48 1.08
32 90 40 2 0.2 1.0 5.16 6.60 0.24 2.52 4.92 0.24 3.72 0.60

256
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Table 4 decreased the exergy destruction rate because of a reduction in the

Exergetic performance parameters of the continuous reactor applied for glycerol ketali- solketal formation rate, as previously explained. Increasing catalyst
zation at different operating conditions.
loading insignificantly affected the exergy destruction rate since in-
Run Solketal Exergy Universal Functional Normalized creasing catalyst loading trivially decreased the solketal production
exergy destruction exergy exergy exergy rate. Even though useful information can be obtained from the above-
rate (W) rate (W) efficiency efficiency destruction (–) discussed exergetic parameters, meaningful dimensionless exergetic
(%) (%) parameters like normalized exergy destruction should be defined by
1 27.85 206.93 39.50 8.14 7.43 consolidating both parameters to elaborate on the process parameters
2 22.99 124.85 40.72 10.92 5.43 more comprehensively.
3 38.47 184.88 42.14 12.04 4.81 The normalized exergy destruction was found to be in the range of
4 18.57 307.88 30.55 4.19 16.58 3.22–22.50. The importance of process parameters on this index could
5 39.79 184.51 42.26 12.45 4.64
be ranked in the following order: reaction pressure > reaction tem-
6 16.80 255.92 26.47 4.83 15.23
7 18.57 184.21 42.35 5.81 9.92 perature > acetone to glycerol molar ratio > catalyst loading > feed
8 21.22 241.15 26.22 6.49 11.36 flow rate. Elevating feed flow rate profoundly lowered the normalized
9 14.59 169.16 20.96 6.82 11.59 exergy destruction, meaning that the quantity of resource destruction
10 21.22 124.29 42.67 9.79 5.86
was decreased at higher feed flow rates for synthesizing a given amount
11 45.98 279.64 38.03 10.19 6.08
12 17.69 240.68 27.73 5.31 13.61 of solketal. This finding could be explained by the fact that increasing
13 49.52 249.47 40.78 11.76 5.04 feed flow rate increased both the solketal exergy rate and exergy de-
14 31.83 102.42 56.80 13.43 3.22 struction rate, while this increment was less marked for the exergy
15 38.47 184.27 42.34 12.04 4.79 destruction rate. In better words, the provided chemical and electrical
16 12.38 278.58 38.26 2.74 22.50
exergies were efficiently utilized for the product formation. Unlike the
17 38.91 132.50 56.55 12.76 3.41
18 10.61 183.61 42.54 3.32 17.30 feed flow rate, increasing all the other process parameters increased the
19 19.45 248.73 42.63 4.49 12.79 normalized exergy destruction. Although loading more catalyst did not
20 43.77 162.85 45.25 14.72 3.72 affect the exergy destruction rate but it lowered the solketal exergy rate.
21 26.53 184.79 43.27 8.14 6.97
Therefore, the normalized exergy destruction was high at higher cata-
22 40.68 132.82 58.14 12.82 3.27
23 22.11 139.68 39.74 9.54 6.32
lyst loadings. Increasing acetone to glycerol molar ratio decreased both
24 55.27 201.06 52.70 13.00 3.64 the solketal exergy rate and exergy destruction rate while this decre-
25 30.07 279.51 39.70 6.49 9.30 ment was more pronounced for the solketal exergy rate. This in turn led
26 42.44 163.61 52.89 12.22 3.85 to the domination of exergy destruction rate over the solketal exergy
27 15.92 256.17 25.07 4.66 16.09
rate, elevating the normalized exergy destruction. In addition, in-
28 39.79 184.51 42.26 12.45 4.64
29 38.47 184.88 42.14 12.04 4.81 creasing both reaction temperature and pressure increased the exergy
30 39.79 184.51 42.26 12.45 4.64 destruction rate while decreased the solketal exergy rate, as discussed
31 47.75 185.45 38.91 15.73 3.88 previously. These in turn deteriorated the normalized exergy destruc-
32 27.41 140.08 37.91 12.15 5.11
tion at higher reaction temperatures and pressures.

Fig. 5. Developed fuzzy inference system (FIS1) for

modeling the universal exergetic efficiency of the
glycerol ketalization process.

257
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Fig. 6. Some parts of the obtained rules for predicting the universal exergetic efficiency.

Table 5
Statistical performance parameters of the selected ANFIS models.

FIS Training dataset Testing dataset

2
R MSE MAPE R2 MSE MAPE

−3 −4 −8
FIS1 0.9000 1.20 × 10 3.71 × 10 0.9999 2.34 × 10 1.97 × 10−6
FIS2 0.9991 1.12 × 10−2 4.40 × 10−3 1.0000 1.74 × 10−9 4.19 × 10−6
FIS3 0.9999 1.80 × 10−3 4.50 × 10−3 1.0000 1.62 × 10−9 2.23 × 10−6

The universal exergetic efficiency of the system ranged from a solketal exergy rate and the total input exergy rate. However, this in-
minimum value of 20.96% to a maximum value of 58.14%. Like the crement was more outstanding for the solketal exergy rate, resulting in
solketal exergy rate and exergy destruction rate, feed flow rate had the an increase in the functional exergetic efficiency. Increasing acetone to
highest impact on the universal exergetic efficiency. Acetone to glycerol glycerol molar ratio increased the total inlet exergy while it decreased
molar ratio, reaction temperature, catalyst loading, and reaction pres- the solketal exergy rate, unfavorably diminishing the functional ex-
sure had the next highest impact on the universal exergetic efficiency. ergetic efficiency. Even though loading more catalyst did not affect the
Increasing feed flow rate increased both total input exergy and exergy total inlet exergy, the solketal exergy rate was decreased at higher
destruction rate. However, the total input exergy rate was increased catalyst loadings, as explained previously. Furthermore, elevating re-
with a higher rate compared with exergy destruction rate which in turn action temperature and pressure enhanced the total inlet exergy be-
enhanced the universal exergetic efficiency. Similarly, increasing re- cause of an increase in the rate of electrical energy consumption. The
action temperature and pressure increased both electrical energy con- solketal exergy rate was slightly influenced by reaction temperature,
sumption and exergy destruction rate. However, the rate of exergy while reaction temperature negatively declined the solketal formation
destruction dominated the rate of electrical energy consumption, di- rate. Accordingly, increasing catalyst loading, reaction temperature,
minishing the universal exergetic efficiency. The universal exergetic and reaction pressure negatively affected the functional exergetic effi-
efficiency was positively affected by acetone to glycerol molar ratio. ciency.
The total input exergy feeding into the system was increased with in- Fig. 5 presents the developed FIS1 for modeling the universal ex-
creasing acetone to glycerol molar ratio due to the higher standard ergetic efficiency including its fuzzy inputs (i.e., fuzzy feed flow rate,
chemical exergy of acetone compared with glycerol (Table 1). On the fuzzy acetone to glycerol molar ratio, fuzzy catalyst loading, fuzzy re-
other hand, the exergy destruction rate was decreased with increasing action temperature, and fuzzy pressure). Obviously, the fuzzy system
acetone to glycerol molar ratio, as previously explained. These in turn with 2 Gaussian membership functions for each input and 32 inference
collectively increased the universal exergetic efficiency according to Eq. rules could successfully estimate the universal exergetic efficiency.
(12). This exergetic parameter was not influenced significantly by Additionally, some parts of the obtained rules for predicting the uni-
loading more catalyst into the system. This occurred because of the fact versal exergetic efficiency are shown in Fig. 6. The same findings were
that loading more catalyst did not markedly affect neither the total obtained for both the functional exergetic efficiency and the normalized
input exergy rate nor the exergy destruction rate. exergy destruction as well.
The functional exergetic efficiency of the ketalization process varied Statistical performance parameters of the selected ANFIS models in
from a minimum value of 2.74% to a maximum value of 15.73%. Unlike training and testing steps are tabulated in Table 5. The values of R2,
the feed flow rate, increasing all the other process parameters declined MSE, and MAPE obtained using the developed ANFIS models were
the functional exergetic efficiency. The order of the effect of process within acceptable ranges. In general, all of the ANFIS models had a
parameters on the functional exergetic efficiency was feed flow good generalization capability for estimating the exergetic performance
rate > acetone to glycerol molar ratio > catalyst loading > reaction parameters of the continuous reactor applied for solketal production
temperature > pressure. Increasing feed flow rate enhanced both the from glycerol.

258
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

60
Fig. 7 shows the experimental exergetic performance parameters of
Predicted universal exergetic efficiency (%)

the reactor against the predicted values using the developed ANFIS
55
models. The data points of all the considered exergetic parameters were
well accumulated around a 45° straight line, showing the suitability of
50
the developed ANFIS models for predicting the exergetic performance
parameters of the process. Accordingly, the developed ANFIS models
45
could be reliably used for the optimization of the process.
40
Table 6 tabulates Pareto optimal front set obtained by the NSGA-II.
It should be noted that these points demonstrate the optimal trade-off
35 among the three objectives. All the points proposed by the developed
approach could be used as the best operating conditions for glycerol
30 ketalization with acetone from the exergetic viewpoint. Overall, the
reaction temperature of 40.66 °C, acetone to glycerol molar ratio of
25 4.97, feed flow rate of 0.49 mL/min, pressure of 42.31 bar, and catalyst
mass of 0.50 g yielding universal exergetic efficiency of 90.36%, func-
20 tional exergetic efficiency of 17.33%, and normalized exergy destruc-
20 25 30 35 40 45 50 55 60
tion of 6.18 (bolded line in Table 6) could be suggested as the best
Experimental universal exergetic efficiency (%)
operating conditions. Indeed, these conditions were selected by con-
16
sidering the preferences of functional exergetic efficiency, normalized
Predicted functional exergetic efficiency (%)

exergy destruction, and universal exergetic efficiency, respectively. The

14
functional exergetic efficiency associated with this point was found to
be the highest, while the normalized exergy destruction was found to be
12 the lowest among the proposed conditions. In addition, the universal
exergetic efficiency was sufficiently high under these conditions. The
10 optimum operating conditions of the reactor on the basis of process
yield were reaction temperature of 20 °C, pressure of 120 bar, acetone
8
to glycerol mole ratio of 5, feed flow rate of 0.1 mL/min, and catalyst
mass of 0.77 g. Interestingly, there were profound differences in the
optimum operating conditions proposed by the exergy analysis com-
6
pared with those proposed according to the conventional process yield.
This in turn further confirmed the suitability of the exergy analysis and
4 its extension for analyzing and optimizing complex chemical systems.
Table 7 compares the optimum universal and functional exergy ef-
2 ficiencies obtained throughout this study with those reported in the
2 4 6 8 10 12 14 16
literature. Clearly, the universal exergy efficiency of the ketalization
Experimental functional exergetic efficiency (%)
process using subcritical acetone was comparable with those reported
25 previously [28]. This means that the exergy destruction rate of the
ketalization process was within an acceptable level. In addition to that,
Predicted normalized exergy destruction

the functional exergy efficiency of the glycerol ketalization process was

20 lower than the mean values obtained using the other methods
[26,27,39]. This could be attributed to the differences in the process
characteristics as well as the formulas used for measuring the functional
15 exergy efficiency of the process. On the other hand, this study was to
the best of our knowledge the only report conducted based on the ex-
perimental data including real-world constraints and uncertainties,
10 while the other report available are simulation-based databases at ideal
conditions [26–28,39].
Overall, the exergy concept could provide a holistic framework for
5
exploring the most thermodynamically, economically, and en-
vironmentally viable conditions for chemicals synthesis compared with
many available techniques like energy analysis and life cycle assess-
ment approach. Moreover, although the standard exergy analysis could
0
0 5 10 15 20 25 inherently reveal the most economical and sustainable operational
Experimental normalized exergy destruction conditions by measuring the thermodynamic irreversibilities, addi-
tional complementary information could be obtained from its elabo-
Fig. 7. Experimental exergetic performance parameters of the reactor against the pre-
dicted ones by the developed ANFIS models.
rated extensions like exergoeconomic and exergoenvironmental

Table 6
Pareto optimal front set. (Bolded line shows the best operating conditions and corresponding exergetic performance parameters.)

Pressure (bar) Temperature (°C) Molar ratio Feed flow rate Catalyst mass Universal exergetic Functional exergetic Normalized exergy
(–) (mL/min) (g) efficiency (%) efficiency (%) destruction (–)

42.87 40.75 4.99 0.49 0.50 90.39 17.31 6.22

42.31 40.66 4.97 0.49 0.50 90.36 17.33 6.18
42.69 41.20 4.99 0.49 0.50 90.39 17.27 6.29

259
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

approaches [40]. Therefore, these analyses must be applied in the fu-

Current study
References ture in order to optimize chemical systems to achieve the highest fea-
sible thermodynamic, economic, and environmental performance.
[26]

[27]

[28]

[39]
4. Conclusions

reactor obtained at reaction temperature of 40.66 °C, acetone to glycerol molar ratio of
at water/glycerol feed ratio of 5.5, reforming temperature of 900 K, and oxygen/glycerol
This value was the optimum functional exergy efficiency of the overall process obtained

This value was the optimum functional exergy efficiency of the overall process obtained

This value was the optimum functional exergy efficiency of the overall process obtained

4.97, feed flow rate of 0.49 mL/min, pressure of 42.31 bar, and catalyst mass of 0.50 g
This value was the universal exergetic efficiency of the glycerol steam reforming unit

These values were the optimum functional and universal exergy efficiencies of the
In this study, continuous solketal synthesis from glycerol through
ketalization with subcritical acetone in the presence of Purolite PD206
at reforming and preheating temperature of 800 °C and pressure of 240 atm

catalyst was exergetically analyzed and optimized. To this end, ex-

perimental data of glycerol ketalization process at different reaction
at water/glycerol feed ratio of 6 and reforming temperature of 1100 K

temperatures, reaction pressures, acetone to glycerol mole ratios, feed

flow rates, and catalyst loadings were used to compute the exergetic
performance parameters. The computed dimensionless exergetic para-
meters viz. normalized exergy destruction, universal exergetic effi-
ciency, and functional exergetic efficiency were then modeled as a
function of the process parameters using the ANFIS approach. The de-
veloped models were fed into the NSGA-II approach to find the optimal
operating conditions of the systems by maximizing universal and
functional efficiencies and minimizing normalized exergy destruction.
The main conclusions can be summarized as follows:

• Feed flow rate had the highest effect on the exergetic variables,
annexed to a biodiesel plant

while reaction pressure showed the lowest effect on these indicators.

• The developed ANFIS models could effectively estimate the ex-
feed ratio of 0.96

ergetic performance parameters of the process with an R2 higher

than 0.99.
•
Remarks

The optimal operating conditions of the reactor were found to be

reaction temperature of 40.66 °C, reaction pressure of 42.31 bar,
acetone to glycerol mole ratio of 4.97, feed flow rate of 0.49 mL/
min, and catalyst loading of 0.50 g.
Universal exergetic efficiency

• Under the selected conditions, normalized exergy destruction, uni-

versal exergetic efficiency, and functional exergetic efficiency were
found to be 6.18%, 90.36%, and 17.33%, respectively.
of the process (%)
Comparison of the optimum functional and universal exergetic efficiency values of various glycerol conversion methods.

The selected conditions could be applied in industrial-scale con-

tinuous systems for resource-efficient, cost-effective, and en-
90.36

vironmentally-friendly solketal synthesis from glycerol. Future studies

95
–

should aim at analyzing and optimizing the glycerol ketalization pro-

Functional exergetic efficiency

cess on the basis of exergoeconomic and exergoenvironmental ap-

proaches in order to complement the results achieved throughout this
study.
of the process (%)

Acknowledgements

The authors gratefully acknowledge financial support from the Iran

17.33

National Science Foundation (Grant no. 96005466).

33.8

66.1

67.8
–

References
Power and
Product(s)

[1] Odetoye TE, Onifade KR, AbuBakar MS, Titiloye JO. Thermochemical character-
Hydrogen

Hydrogen

Hydrogen
hydrogen

Solketal

isation of Parinari polyandra Benth fruit shell. Ind Crops Prod 2013;44:62–6.
[2] Ajala OE, Aberuagba F, Odetoye TE, Ajala AM. Biodiesel: sustainable energy re-
placement to petroleum-based diesel fuel – a review. ChemBioEng Rev
2015;2:145–56.
Data type used

[3] Hajjari M, Tabatabaei M, Aghbashlo M, Ghanavati H. A review on the prospects of

Experimental
Simulation

Simulation

Simulation

Simulation

sustainable biodiesel production: a global scenario with an emphasis on waste-oil

biodiesel utilization. Renew Sustain Energy Rev 2017;72:445–64.
[4] Yazdani SS, Gonzalez R. Anaerobic fermentation of glycerol: a path to economic
viability for the biofuels industry. Curr Opin Biotechnol 2007;18:213–9.
[5] Katryniok B, Paul S, Bellière-Baca V, Rey P, Dumeignil F. Glycerol dehydration to
Glycerol conversion method

Ketalization with subcritical

acrolein in the context of new uses of glycerol. Green Chem 2010;12:2079–98.

[6] Rastegari H, Ghaziaskar HS, Yalpani M, Shafiei A. Development of a continuous
Autothermal reforming

system based on azeotropic reactive distillation to enhance triacetin selectivity in

glycerol esterification with acetic acid. Energy Fuels 2017;31:8256–62.
Supercritical water

Steam reforming

Steam reforming

[7] Faria RPV, Pereira CSM, Silva VMTM, Loureiro JM, Rodrigues AE. Glycerol valor-
reforming

ization as biofuel: thermodynamic and kinetic study of the acetalization of glycerol

acetone

with acetaldehyde. Ind Eng Chem Res 2013;52:1538–47.

[8] Gorji YM, Ghaziaskar HS. Optimization of solketalacetin synthesis as a green fuel
Table 7

additive from ketalization of Monoacetin with acetone. Ind Eng Chem Res
(s)

2016;55:6904–10.
[9] Alptekin E. Emission, injection and combustion characteristics of biodiesel and

260
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

oxygenated fuel blends in a common rail diesel engine. Energy 2017;119:44–52. production from glycerol by supercritical water reforming. Chem Eng J
[10] Garcia E, Laca M, Pérez E, Garrido A, Peinado J. New class of acetal derived from 2013;218:309–18.
glycerin as a biodiesel fuel component. Energy Fuels 2008;22:4274–80. [27] Hajjaji N, Chahbani A, Khila Z, Pons M-N. A comprehensive energy–exergy-based
[11] Frusteri F, Spadaro L, Beatrice C, Guido C. Oxygenated additives production for assessment and parametric study of a hydrogen production process using steam
diesel engine emission improvement. Chem Eng J 2007;134:239–45. glycerol reforming. Energy 2014;64:473–83.
[12] Deutsch J, Martin A, Lieske H. Investigations on heterogeneously catalysed con- [28] Boloy RAM, Ferrán SJR, e Penalva D de CL, Corrêa C, Angulo JAP, de Castro Pereira
densations of glycerol to cyclic acetals. J Catal 2007;245:428–35. Filho R. Exergetic evaluation of incorporation of hydrogen production in a biodiesel
[13] Roldán L, Mallada R, Fraile JM, Mayoral JA, Menéndez M. Glycerol upgrading by plant. Int J Hydrogen Energy 2015;40:8797–805.
ketalization in a zeolite membrane reactor. Asia-Pacific J Chem Eng 2009;4:279–84. [29] Tippawan P, Thammasit T, Assabumrungrat S, Arpornwichanop A. Using glycerol
[14] Vicente G, Melero JA, Morales G, Paniagua M, Martín E. Acetalisation of bio-gly- for hydrogen production via sorption-enhanced chemical looping reforming: ther-
cerol with acetone to produce solketal over sulfonic mesostructured silicas. Green modynamic analysis. Energy Convers Manage 2016;124:325–32.
Chem 2010;12:899–907. [30] Presciutti A, Asdrubali F, Baldinelli G, Rotili A, Malavasi M, Di Salvia G. Energy and
[15] Ferreira P, Fonseca IM, Ramos AM, Vital J, Castanheiro JE. Valorisation of glycerol exergy analysis of glycerol combustion in an innovative flameless power plant. J
by condensation with acetone over silica-included heteropolyacids. Appl Catal B Clean Prod 2017;172:3817–24.
Environ 2010;98:94–9. [31] de Souza-Santos ML. Proposals for power generation based on processes consuming
[16] Maksimov AL, Nekhaev AI, Ramazanov DN, Arinicheva YA, Dzyubenko AA, biomass-glycerol slurries. Energy 2017;120:959–74.
Khadzhiev SN. Preparation of high-octane oxygenate fuel components from plant- [32] Samoilov VO, Ramazanov DN, Nekhaev AI, Maximov AL, Bagdasarov LN.
derived polyols. Pet Chem 2011;51:61–9. Heterogeneous catalytic conversion of glycerol to oxygenated fuel additives. Fuel
[17] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu CC. A new continuous- 2016;172:310–9.
flow process for catalytic conversion of glycerol to oxygenated fuel additive: cata- [33] Esteban J, Domínguez E, Ladero M, Garcia-Ochoa F. Kinetics of the production of
lyst screening. Appl Energy 2014;123:75–81. glycerol carbonate by transesterification of glycerol with dimethyl and ethylene
[18] Shirani M, Ghaziaskar HS, Xu CC. Optimization of glycerol ketalization to produce carbonate using potassium methoxide, a highly active catalyst. Fuel Process
solketal as biodiesel additive in a continuous reactor with subcritical acetone using Technol 2015;138:243–51.
Purolite® PD206 as catalyst. Fuel Process Technol 2014;124:206–11. [34] Song G, Xiao J, Zhao H, Shen L. A unified correlation for estimating specific che-
[19] Aghbashlo M, Tabatabaei M, Mohammadi P, Khoshnevisan B, Rajaeifar MA, Pakzad mical exergy of solid and liquid fuels. Energy 2012;40:164–73.
M. Neat diesel beats waste-oriented biodiesel from the exergoeconomic and ex- [35] Aghbashlo M, Tabatabaei M, Karimi K. Exergy-based sustainability assessment of
ergoenvironmental point of views. Energy Convers Manage 2017;148:1–15. ethanol production via Mucor indicus from fructose, glucose, sucrose, and molasses.
[20] Aghbashlo M, Mobli H, Rafiee S, Madadlou A. A review on exergy analysis of drying Energy 2016;98:240–52.
processes and systems. Renew Sustain Energy Rev 2013;22:1–22. [36] Aghbashlo M, Hosseinpour S, Tabatabaei M, Dadak A. Fuzzy modeling and opti-
[21] Aghbashlo M, Tabatabaei M, Hosseinpour S, Khounani Z, Hosseini SS. Exergy-based mization of the synthesis of biodiesel from waste cooking oil (WCO) by a low
sustainability analysis of a low power, high frequency piezo-based ultrasound re- power, high frequency piezo-ultrasonic reactor. Energy 2017;132:65–78.
actor for rapid biodiesel production. Energy Convers Manage 2017;148:759–69. [37] Mandegari MA, Pahlavanzadeh H. A study on the optimization of an air dehumi-
[22] Barati MR, Aghbashlo M, Ghanavati H, Tabatabaei M, Sharifi M, Javadirad G, et al. dification desiccant system. J Therm Sci Eng Appl 2013;5:41002.
Comprehensive exergy analysis of a gas engine-equipped anaerobic digestion plant [38] Aghbashlo M, Hosseinpour S, Tabatabaei M, Younesi H, Najafpour G. On the ex-
producing electricity and biofertilizer from organic fraction of municipal solid ergetic optimization of continuous photobiological hydrogen production using hy-
waste. Energy Convers Manage 2017;151:753–63. brid ANFIS–NSGA-II (adaptive neuro-fuzzy inference system–non-dominated
[23] Aghbashlo M, Tabatabaei M, Hosseini SS, Dashti BB, Mojarab Soufiyan M. sorting genetic algorithm-II). Energy 2016;96:507–20.
Performance assessment of a wind power plant using standard exergy and extended [39] Hajjaji N, Baccar I, Pons M-N. Energy and exergy analysis as tools for optimization
exergy accounting (EEA) approaches. J Clean Prod 2018;171:127–36. of hydrogen production by glycerol autothermal reforming. Renew Energy
[24] Mandegari MA, Farzad S, Pahlavanzadeh H. Exergy performance analysis and op- 2014;71:368–80.
timization of a desiccant wheel system. J Therm Sci Eng Appl 2015;7:31013. [40] Aghbashlo M, Rosen MA. Consolidating exergoeconomic and exergoenvironmental
[25] Seager TP, Theis TL. Exergetic pollution potential: estimating the revocability of analyses using the emergy concept for better understanding energy conversion
chemical pollution. Exergy, Int J 2002;2:273–82. systems. J Clean Prod 2018;172:696–708.
[26] Ortiz FJG, Ollero P, Serrera A, Galera S. Optimization of power and hydrogen

261
 Energy Conversion and Management 160 (2018) 251–261

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Multi-objective exergy-based optimization of continuous glycerol T

ketalization to synthesize solketal as a biodiesel additive in subcritical
acetone
⁎ ⁎ ⁎
Mortaza Aghbashloa, , Meisam Tabatabaeib,c, , Soleiman Hosseinpoura, , Hajar Rastegarid,
Hassan S. Ghaziaskard
a
Department of Mechanical Engineering of Agricultural Machinery, Faculty of Agricultural Engineering and Technology, College of Agriculture and Natural Resources,
University of Tehran, Karaj, Iran
b
Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), P.O. Box: 31535-1897, Agricultural Research, Education, and
Extension Organization (AREEO), Karaj, Iran
c
Biofuel Research Team (BRTeam), Karaj, Iran
d
Department of Chemistry, Isfahan University of Technology, Isfahan, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: This study was aimed at exergetically investigating and optimizing a continuous reactor applied to valorize
Adaptive neuro-fuzzy inference system glycerol into solketal as a biodiesel additive with subcritical acetone in the presence of Purolite PD206. The
Biodiesel additive effects of reaction temperature (20–100 °C), acetone to glycerol molar ratio (1–5), feed flow rate (0.1–0.5 mL/
Exergy analysis min), pressure (1–120 bar), and catalyst mass (0.5–2.5 g) were evaluated on the exergetic performance para-
Glycerol conversion
meters of the reactor. In order to optimize the operating conditions of the reactor, adaptive neuro-fuzzy in-
Non-dominated sorting genetic algorithm-II
ference system (ANFIS) was coupled with non-dominated sorting genetic algorithm-II (NSGA-II). The ANFIS was
Solketal synthesis
applied to develop objective functions on the basis of the process parameters. The developed objective functions
were then fed into the NSGA-II to find the optimum operating conditions of the process by simultaneously
maximizing universal and functional exergetic efficiencies and minimizing normalized exergy destruction.
Overall, the process parameters significantly affected the exergetic performance of the reactor. The ANFIS ap-
proach successfully modeled the objective functions with a correlation coefficient higher than 0.99. The optimal
ketalization conditions of glycerol were: reaction temperature = 40.66 °C, acetone to glycerol molar
ratio = 4.97, feed flow rate = 0.49 mL/min, pressure = 42.31 bar, and catalyst mass = 0.50 g. These conditions
could be applied in pilot- or industrial-scale reactors for converting glycerol into value-added solketal in a
resource-efficient, cost-effective, and environmentally-friendly manner.

1. Introduction up to 10-folds [4]. To improve the economic features of biodiesel in-

dustry, upgrading glycerol into different value-added chemicals and
Over the last few years, biofuels have drawn an increasing deal of making use of glycerol-containing waste streams are crucial.
attention as alternatives to fossil fuels due to their environmental There are already more than 1500 direct applications of glycerol,
benefits and the fact that they are produced from renewable feedstocks particularly in cosmetics, pharmaceuticals, and food industries [5].
[1]. Among the various biofuels commercialized to date, biodiesel has Nevertheless, there has been intense research interest towards finding
shown a great potential to supplement petrodiesel [2,3]. Consequently, methods to upgrade glycerol into more valuable renewable chemicals.
the production of biodiesel has increased substantially since the year For instance, several well-known value-added derivatives of glycerol
2000 especially to satisfy the stringent emission mandates for diesel are esters [6], acetals [7], and ketals [8]. These chemicals can poten-
engines and to partially meet the increasing global demand for diesel tially be used as oxygenated fuel additives to improve diesel/biodiesel
fuel. This in turn has also led to a glycerol surplus as a by-product of combustion and emissions. Among these glycerol derivatives, solketal
biodiesel synthesis, negatively affecting glycerol price dramatically by (4-hydroxymethyl-2,2-dimethyl-1,3-dioxolane) produced from glycerol

⁎
Corresponding authors at: Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), P.O. Box: 31535-1897, Agricultural Research,
Education, and Extension Organization (AREEO), Karaj, Iran (M. Tabatabaei).
E-mail addresses: [email protected] (M. Aghbashlo), [email protected] (M. Tabatabaei), [email protected] (S. Hosseinpour).

https://doi.org/10.1016/j.enconman.2018.01.044
Received 4 November 2017; Received in revised form 3 January 2018; Accepted 19 January 2018
0196-8904/ © 2018 Elsevier Ltd. All rights reserved.
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Nomenclature Greek letters

 ash percentage (%) ψ universal exergetic efficiency (%)

C catalyst mass (g) ϕ functional exergetic efficiency (%)
 carbon percentage (%) Ψ normalized exergy destruction (–)
Cp specific heat capacity (kJ/kg K)
ex specific chemical exergy (kJ/kg) Subscripts
Eẋ exergy flow rate (kJ/s)
F feed flow rate (mL/min) 0 dead state
 hydrogen percentage (%) AC acetone
M molar mass (g/mol) ch chemical
ṁ mass flow rate (kg/s) des destruction
ṅ molar flow rate (mol/s) ET ethanol
 nitrogen percentage (%) GL glycerol
 oxygen percentage (%) in inflow
P pressure (bar) i, j numerator
R universal gas constant (kJ/mol k) l loss
 sulfur percentage (%) out outflow
T temperature (°C or K) P product
x mass fraction (–) ph physical
X acetone to glycerol molar ratio (–) q heat transfer
y molar fraction (–) SK solketal
Ẇ work flow rate (kJ/s) WT water

and acetone has attracted more attention recently due to its unique [26–31]. According to these studies, exergy analysis could provide a
features in improving cold flow properties and flash point temperatures better understanding of the effect of process parameters on the ther-
of both diesel and biodiesel as well as their blends [9]. Solketal can also modynamic performance and can aid with the diagnosis of the most
be used as a solvent and plasticizer, suspension agent in pharmaceutical effective strategies to improve the process under investigation.
preparations, and anti-freezing agent [10]. A number of research works have been reported on the production
Solketal is conventionally produced through ketalization of glycerol of value-added chemicals from glycerol with their focus on technical
with acetone by using strong Brønsted acid catalysts like sulfuric acid, aspects and kinetics simulation of the reported processes [18,32,33].
hydrochloric acid, phosphoric acid, and para-toluene sulfonic acid in However, to the best of our knowledge, there is no report so far on the
solvents such as chloroform, ether, or oil [11]. This procedure suffers exergy analysis of a continuous or even a batch reactor applied for
from some serious drawbacks such as equipment corrosion, effluent synthesizing chemicals from glycerol. Therefore, solketal production
disposal, and difficulties in separation of the catalysts from the products from glycerol using a continuous reactor with subcritical acetone in the
which in turn increase the production cost and result in environmental presence of Purolite PD206 catalyst was exergetically analyzed in the
burdens. These issues can be addressed by using heterogeneous acidic present study. According to our previous report, this heterogeneous
catalysts like nafion [12], zeolite [13], Amberlyst [14], silica containing acidic catalyst could effectively upgrade glycerol to solketal [18].
heteropolyacids [15], and montmorillonite [16]. Batch process has Therefore, the main aim of the present survey was to enhance our un-
been used for solketal synthesis in the above-mentioned studies. How- derstanding of the ketalization step in the developed continuous reactor
ever, this mode of production has several major limitations such as long exergetically. More specifically, the effects of process parameters viz.
reaction time, high energy consumption, poor yield, and difficulty in reaction temperature (20–120 °C), reaction pressure (1–120 bar),
scale-up [17]. Therefore, in order to solve these issues, Shirani et al. acetone to glycerol molar ratio (1–5), feed flow rate (0.1–0.5), and
[18] developed an easy to scale-up continuous system for upgrading catalyst loading (0.5–2.5 g) on the exergetic performance parameters of
glycerol into solketal in the presence of Purolite PD206 as a hetero- the reactor were comprehensively assessed and discussed. Moreover, a
geneous catalyst. Even though the developed system efficiently syn- multi-objective exergy-based optimization was carried out to find the
thesized solketal from glycerol, advanced engineering paradigms such optimum operating conditions of the reactor using coupled adaptive
as exergy should still be used to assess the productivity and sustain- neuro-fuzzy inference system (ANFIS) and non-dominated sorting ge-
ability of such systems. netic algorithm-II (NSGA-II) approaches. The ANFIS approach was used
During the past two decades, there has been an increasing interest in to develop three objective functions, i.e., normalized exergy destruc-
applying exergy analysis for scrutinizing energy and material conver- tion, universal exergetic efficiency, and functional exergetic efficiency
sion systems from productivity and sustainability viewpoints [19–21]. as a function of the process parameters. The developed objective
Unlike energy analysis which is based on the first law of thermo- functions were then fed into the NSGA-II algorithm to simultaneously
dynamics, exergy analysis accounts for the degradation of the energy maximize universal and functional exergetic efficiencies and minimize
quality by irreversible processes [22]. In better words, by using the first normalized exergy destruction.
and second laws of thermodynamics simultaneously, exergy analysis
compensates for the shortcoming of energy analysis in revealing the
2. Materials and methods
energy quality loss because of thermodynamic imperfections [23].
Generally, exergy is not quantitatively conserved like energy but it is
2.1. Materials
destroyed due to irreversibilities within a system [24]. The quantity of
destroyed exergy is a measure of environmental pollution costs, pro-
Glycerol and acetone (both ≥99%) were purchased from Merck Co.
viding a quantitative comparison of environmental impacts [25].
(Germany). Solketal (≥97%) for GC calibration was obtained from
Hence, the exergy analysis has been widely used in recent years for
Sigma-Aldrich Co. (Germany). Absolute ethanol (ET) was supplied by
analyzing and optimizing various glycerol upgrading processes
Bidestan Co. (Iran). Toluene (≥99.9%) was also obtained from Merck.

252
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Co (Germany). Purolite PD206 was purchased from Purolite Co. (USA). ṁ GL,in + ṁ AC,in + ṁ ET ,in = ṁ SK ,out + ṁ GL,out + ṁ AC ,out + ṁ ET ,out
All the chemicals used herein were of analytical grade.
+ ṁ WT ,out (2)
The exergy balance for the system could be expressed as follows:
2.2. Experimental procedure and analytical method
Eẋin + Ẇ = Eẋ out + Eẋ q,l + Eẋ des (3)
Synthesis of solketal was carried out in a bench-scale continuous
The exergy flow rate of the feed and product streams could be
flow tubular reactor heated in an air oven. The reactor was made of a
computed on the basis of the summation of their chemical and physical
316-stainless steel tube with an 8 mm i.d., 30 cm length, filled with a
exergies. Accordingly, the above-mentioned exergetic balance equation
specific amount of catalyst. The catalyst was mixed with crushed Pyrex
could be rewritten as follows:
glass (CPG, mesh 10–20). The feed was pumped through an HPLC pump
(Shimadzu model LC-6A) at a specific flow rate into the reactor at a ch ph ch ph
Eẋ in + Eẋ in + Ẇ = Eẋ out + Eẋ out + Eẋ q,l + Eẋ des (4)
fixed temperature and pressure. The feed consisted of a homogenous
solution of reactants (glycerol and acetone) and ethanol. Since the so- The heat loss rate from the oven frame to the surroundings could be
lubility of glycerol in acetone is very low, ethanol was used as co-sol- neglected according to the outcomes of energy analysis. Therefore, the
vent mainly to improve its solubility in acetone. Otherwise, glycerol exergy balance equation can be measured as follows:
and acetone should have been separately fed into the reactor, in- ch ph ch ph
Eẋ in + Eẋ in + Ẇ = Eẋ out + Eẋ out + Eẋ des (5)
creasing the operating costs and leading to difficulties in the process
control. The controlling system of the air oven was used to adjust the The chemical and physical exergies of the feed and product streams
reaction temperature with an accuracy of ± 1.0. In addition, pressure could be determined using the following expressions, respectively:
was controlled through a back-pressure regulator (KPB, Swagelok,
G.B.). In each run, a pre-determined amount of Purolite PD206 was ⎛ ⎞
Eẋ ch = ṅ ⎜∑ yi εi + RT0 ∑ yi ln(yi )⎟
preloaded into the reactor. The effect of catalyst mass in the range of ⎝ i i ⎠ (6)
0.5–2.5 g was studied on the exergetic performance parameters of the
reactor. This range was considered on the basis of glycerol mass being T
̇ p ⎛T −T0−T0ln ⎞
Eẋ ph = mC ⎜ ⎟

accumulated in the reactor. It should be noted that the lower level of ⎝ T0 ⎠ (7)
catalyst mass, i.e., 0.5 g was considered based on our preliminary ex-
The following formula was used to determine the specific heat ca-
periments [18]. Accordingly, catalyst mass values below 0.5 g did not
lead to an acceptable yield due to the short length of the catalytic bed in pacity of the outflow stream.
the reactor. During each experiment, four samples were taken at 30 min Cp = ∑ x i Cp,i
intervals with the collection efficiency of more than 95%. The sche- i (8)
matic experimental set-up used for the continuous glycerol ketalization
The standard chemical exergy values of the organic components
is shown in Fig. 1. The chemical reaction occurring during the glycerol
involved in the process viz. glycerol, acetone, and solketal were cal-
ketalization with acetone is also shown in Fig. 2.
culated using the mathematical model developed by Song et al. [34].
Samples were analyzed using a GC-FID (Agilent technologies model
6890 N). The carrier gas was nitrogen and a polar capillary column of ex i = 363.439 + 1075.633 −86.308 + 4.14 + 190.798 −21.1 (9)
BP20 (30 m length, 0.25 mm i.d., and 0.25 μm film thickness) was used.
εi = Mi ex i (10)
The GC injection port and the detector temperature were set at 240 °C
and 260 °C, respectively. The temperature program used for the analysis The standard chemical exergy values of water and ethanol were
included an initial column temperature of 50 °C for 3.5 min, which was obtained from the published literature [35]. Table 1 summarizes the
subsequently increased from 50 °C to 260 °C at the rate of 40 °C/min. chemical formulas and standard chemical exergy values of the com-
Finally, the temperature was maintained at 260 °C for 3.5 min. Some ponents of feed and product streams.
standard solutions containing toluene as an internal standard were in- Solketal was the only product of the process. Accordingly, the ex-
jected and the peak areas were integrated to establish the calibration ergy rate of the product was equal to the solketal exergy rate.
curves. Identification of the products was carried out by gas chroma- ch
tography/mass spectrometry (GC/MS; model 6890N, Agilent Eẋ P = Eẋ SK ,out (11)
Technologies).
The universal exergetic efficiency of the glycerol ketalization

2.3. Theoretical considerations

The assumptions listed below were taken into consideration

throughout this study in developing the exergy model:

• The reactor operated at a steady state condition.

• The kinetic and potential exergies of the reactants and products
were neglected.
• The dead state temperature and pressure were taken into account to
be 25 °C and 1.01 bar, respectively.
• The dead state model did not vary with time.
According to Fig. 2, the mass flow rate balance for the system could
be written as follows:
Fig. 1. The experimental set-up used for ketalization of glycerol in subcritical acetone. (1)
ṁ in = ṁ out (1) Analytical balance, (2) Feed vessel, (3) HPLC pump, (4) On-off valve, (5) Air oven, (6)
Tubular reactor, (7) Cooling jacket, (8) Back-pressure regulator, and (9) Sample collection
vessel.
Or in more detailed form,

253
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Fig. 2. The chemical reaction occurring during

the glycerol ketalization with acetone.

Table 1 process was measured as:

Chemical formulas and standard chemical exergy values of the components of feed and
ch ph
product streams. Eẋ out Eẋ out + Eẋ out Eẋ des
ψ= = × 100 = 1− ch × 100
̇
Ex in ch ph
Eẋ in + Eẋ in + Ẇ
ph
Eẋ in + Eẋ in + Ẇ
Component Chemical formula Standard chemical exergy (kJ/mole)
(12)
Glycerol C3H8O3 1762.63
Acetone C3H6O 1821.96 The functional exergetic efficiency of the solketal synthesis was
Solketal C6H12O3 3505.85 determined as follows:
Ethanol C2H6O 1385.45 ch
Water H2O 0.95 Eẋ SK ,out
ϕ= ch ph
× 100
Eẋ in + Eẋ in + Ẇ (13)

Fig. 3. The first order Takagi-Sugeno neuro-fuzzy

system used throughout this study.

254
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

The normalized exergy destruction of the process was determined to Table 2

make decisions on the process parameters by taking into account both Specifications of NSGA-II optimization procedure.
the quantities of the exergy destroyed during the ketalization process
Parameter Value
and the exergy rate of the solketal synthesized as follows:
Objectives min normalized exergy destruction
Eẋ des max universal exergetic efficiency
Ψ= ch
Eẋ SK ,out (14) max functional exergetic efficiency
Variables X1 = Pressure, X2 = temperature, X3 = Molar ratio
X4 = Feed flow rate, X5 = Catalyst mass
Bounds 1 ≤ X1 ≤ 120, 20 ≤ X2 ≤ 100, 1 ≤ X3 ≤ 5, 0.1 ≤ X4 ≤ 0.5,
2.4. Developing objective functions and optimization procedure 0.5 ≤ X5 ≤ 2.5
Population type Double vector
Population size 50
In order to optimize the glycerol ketalization process, robust and
Selection function Tournament
accurate objective functions should first be built. Developing objective Crossover fraction 0.8
functions for multi-objective optimization of complex chemical reactors Mutation fraction 0.02
using the conventional approaches is very difficult. However, the ANFIS Stopping criteria Generations: 500/stall generations: 100
approach could satisfactorily establish the required objective functions
because of its capability to model the process under investigation
algorithm and its extensions have been introduced and applied during
without the need for any priori information. The theoretical background
the past decades [37]. Among the various extensions of the genetic
behind the ANFIS approach and its modeling procedure can be found in
algorithm, the NSGA-II has found many applications for efficiently
the published literature [36]. The first order Takagi-Sugeno neuro-fuzzy
solving the complex optimization problems. The implementation pro-
system was used throughout this study to build the objective functions
cedure of the NSGA-II algorithm can be found elsewhere [38]. In order
as manifested in Fig. 3. The best ANFIS models for estimating the
to optimize the glycerol ketalization process, the developed FIS models
universal exergetic efficiency (FIS1), functional exergetic efficiency
were fed into the NSGA-II as shown in Fig. 4. The optimization was
(FIS2), and normalized exergy destruction (FIS3) were determined on
carried out by maximizing the universal and functional exergy effi-
the basis of three statistical criteria viz. maximum coefficient of de-
ciencies and minimizing the normalized exergy destruction, simulta-
termination (R2), minimum mean absolute percentage error (MAPE),
neously.
and minimum mean-squared error (MSE) by trial-and-error method.
In order to model and optimize the glycerol ketalization process, a
Overall, multi-objective optimization of chemical processes is not
program was developed in MATLAB software (ver. R2008b). The Pareto
straightforward due to the existence of conflicting objectives. In order
optimal front set for the process was obtained as the best operating
to address this issue, evolutionary computation techniques like genetic

Fig. 4. Schematic representation of the developed approach for multi-objective exergy-based optimization of the glycerol ketalization process.

255
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

conditions. Table 2 summarizes the specifications of the developed could be attributed to the fact that small bead-size catalysts such as
NSGA-II algorithm achieved after some trials and errors. Purolite could aggregate at higher loadings, lowering the catalyst ac-
tivity and declining the solketal formation rate. Increasing reaction
3. Results and discussion temperature decreased the solketal exergy rate since the concentration
of acetone as a reactant was markedly decreased in the liquid phase at
Table 3 tabulates the mass balance of the glycerol ketalization with higher reaction temperatures. In better words, due to the low boiling
subcritical acetone in the continuous reactor under investigation. temperature of acetone (56 °C), increasing reaction temperature sig-
Table 4 also summarizes the exergetic performance parameters of the nificantly decreased the solketal evolution rate. It is worth mentioning
continuous reactor applied for glycerol ketalization at different reaction that the glycerol ketalization process with acetone is a reversible re-
temperatures, acetone to glycerol molar ratios, feed flow rates, pres- action [18]. The water evolved during the reaction imposed a ther-
sures, and catalyst loadings. The maximum solketal exergy rate was modynamic barrier, hindering the reaction in the forward direction.
found to be 55.27 W at pressure of 60 bar, reaction temperature of However, acetone concentration as a reactant in the liquid phase and,
60 °C, acetone to glycerol molar ratio of 3, feed flow rate of 0.5 mL/min, consequently on the catalyst surface, was increased at lower reaction
and catalyst loading of 1.5 g. The minimum solketal exergy rate was temperatures owing to its low vapor pressure. These in turn increased
determined as 10.61 W at pressure of 60 bar, acetone to glycerol molar the formation of solketal at lower reaction temperatures. Furthermore,
ratio of 3, feed flow rate of 0.3 mL/min, reaction temperature of 60 °C, increasing reaction pressure had a slight effect on the solketal exergy
and catalyst loading of 2.5 g. The importance of process parameters on rate since this process parameter could not significantly affect the
the solketal exergy rate could be ranked in the following order: feed viscosity and diffusion coefficients of the liquid reaction media.
flow rate > acetone to glycerol molar ratio > catalyst loading > According to the data presented in Table 4, exergy destruction rate
reaction temperature > pressure. Increasing feed flow increased of the glycerol ketalization process varied between 102.42 and
solketal exergy rate, while this parameter was decreased with in- 307.88 W. The order of the effect of process parameters on the exergy
creasing acetone to glycerol molar ratio, catalyst loading, reaction destruction rate was feed flow rate > reaction temperature >
temperature, and pressure. pressure > acetone to glycerol molar ratio > catalyst loading. In-
The rate of product formation increased at higher feed flow rates, creasing feed flow rate profoundly increased exergy destruction rate
resulting in an increase in the solketal exergy rate. In better words, because of an increase in the rate of ketalization reaction. The high-
there was a direct association between feed flow rate and solketal ex- quality electrical energy consumption was significantly increased with
ergy rate. By increasing acetone to glycerol molar ratio, the feeding rate increasing reaction temperature and pressure. In better words, a large
of glycerol into the reactor was decreased. This in turn lowered the rate amount of electrical energy was converted into low quality thermal and
of solketal formation and its exergetic rate. It was envisaged that the thermomechanical energies at higher reaction temperatures and pres-
solketal exergy rate would tend to increase by increasing catalyst sures. Accordingly, the rate of thermodynamic irreversibilities tended
loading because of promoting the ketalization process. In contrast, the to increase by elevating reaction temperature and pressure. Unlike the
solketal exergy rate was decreased by increasing catalyst loading. This other process parameters, increasing acetone to glycerol molar ratio

Table 3
Mass balance of the continuous glycerol ketalization process.

Run Pressure (bar) Temperature (°C) Molar ratio (–) Flow rate (g/min) Catalyst mass (g) Input (g/h) Output (g/h)

Glycerol Acetone Ethanol Glycerol Acetone Ethanol Solketal Water

1 120 60 3 0.3 1.5 6.12 11.52 0.36 3.42 9.90 0.36 3.78 0.54
2 30 40 2 0.2 2.0 5.16 6.60 0.24 3.00 5.16 0.24 3.12 0.48
3 60 60 3 0.3 1.5 6.12 11.52 0.36 2.52 9.18 0.36 5.22 0.72
4 60 100 3 0.3 1.5 6.12 11.52 0.36 4.32 10.44 0.36 2.52 0.36
5 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.18 0.36 5.40 0.72
6 90 80 4 0.2 1.0 3.36 8.52 0.12 1.80 7.44 0.12 2.28 0.36
7 60 60 3 0.3 0.5 6.12 11.52 0.36 4.32 10.44 0.36 2.52 0.36
8 30 80 2 0.2 1.0 5.16 6.60 0.24 3.12 5.28 0.24 2.88 0.48
9 60 60 3 0.1 1.5 2.04 3.84 0.12 0.60 3.00 0.12 1.98 0.30
10 30 40 4 0.2 1.0 3.36 8.52 0.12 1.44 7.20 0.12 2.88 0.36
11 90 80 2 0.4 1.0 10.32 13.20 0.48 6.00 10.32 0.48 6.24 0.96
12 30 80 4 0.2 2.0 3.36 8.52 0.12 1.68 7.44 0.12 2.40 0.36
13 30 80 2 0.4 2.0 10.32 13.20 0.48 5.76 10.08 0.48 6.72 0.96
14 60 20 3 0.3 1.5 6.12 11.52 0.36 2.88 9.72 0.36 4.32 0.72
15 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.36 0.36 5.22 0.72
16 90 80 2 0.4 2.0 10.32 13.20 0.48 9.36 12.24 0.48 1.68 0.24
17 30 40 2 0.4 1.0 10.32 13.20 0.48 6.72 10.80 0.48 5.28 0.72
18 60 60 3 0.3 2.5 6.12 11.52 0.36 5.04 10.98 0.36 1.44 0.18
19 30 80 4 0.4 1.0 6.72 17.04 0.24 4.80 15.84 0.24 2.64 0.48
20 1 60 3 0.3 1.5 6.12 11.52 0.36 1.80 9.00 0.36 5.94 0.90
21 60 60 5 0.3 1.5 4.32 13.50 0.18 1.80 11.88 0.18 3.60 0.54
22 30 40 4 0.4 2.0 6.72 17.04 0.24 3.12 14.40 0.24 5.52 0.72
23 90 40 4 0.2 2.0 3.36 8.52 0.12 1.20 7.20 0.12 3.00 0.48
24 60 60 3 0.5 1.5 10.20 19.20 0.60 4.80 15.90 0.60 7.50 1.20
25 90 80 4 0.4 2.0 6.72 17.04 0.24 3.84 15.12 0.24 4.08 0.72
26 90 40 4 0.4 1.0 6.72 17.04 0.24 2.64 14.40 0.24 5.76 0.96
27 90 80 2 0.2 2.0 5.16 6.60 0.24 3.72 5.52 0.24 2.16 0.36
28 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.18 0.36 5.40 0.72
29 60 60 3 0.3 1.5 6.12 11.52 0.36 2.52 9.18 0.36 5.22 0.72
30 60 60 3 0.3 1.5 6.12 11.52 0.36 2.34 9.18 0.36 5.40 0.72
31 60 60 1 0.3 1.5 10.80 6.66 0.54 5.94 3.96 0.54 6.48 1.08
32 90 40 2 0.2 1.0 5.16 6.60 0.24 2.52 4.92 0.24 3.72 0.60

256
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Table 4 decreased the exergy destruction rate because of a reduction in the

Exergetic performance parameters of the continuous reactor applied for glycerol ketali- solketal formation rate, as previously explained. Increasing catalyst
zation at different operating conditions.
loading insignificantly affected the exergy destruction rate since in-
Run Solketal Exergy Universal Functional Normalized creasing catalyst loading trivially decreased the solketal production
exergy destruction exergy exergy exergy rate. Even though useful information can be obtained from the above-
rate (W) rate (W) efficiency efficiency destruction (–) discussed exergetic parameters, meaningful dimensionless exergetic
(%) (%) parameters like normalized exergy destruction should be defined by
1 27.85 206.93 39.50 8.14 7.43 consolidating both parameters to elaborate on the process parameters
2 22.99 124.85 40.72 10.92 5.43 more comprehensively.
3 38.47 184.88 42.14 12.04 4.81 The normalized exergy destruction was found to be in the range of
4 18.57 307.88 30.55 4.19 16.58 3.22–22.50. The importance of process parameters on this index could
5 39.79 184.51 42.26 12.45 4.64
be ranked in the following order: reaction pressure > reaction tem-
6 16.80 255.92 26.47 4.83 15.23
7 18.57 184.21 42.35 5.81 9.92 perature > acetone to glycerol molar ratio > catalyst loading > feed
8 21.22 241.15 26.22 6.49 11.36 flow rate. Elevating feed flow rate profoundly lowered the normalized
9 14.59 169.16 20.96 6.82 11.59 exergy destruction, meaning that the quantity of resource destruction
10 21.22 124.29 42.67 9.79 5.86
was decreased at higher feed flow rates for synthesizing a given amount
11 45.98 279.64 38.03 10.19 6.08
12 17.69 240.68 27.73 5.31 13.61 of solketal. This finding could be explained by the fact that increasing
13 49.52 249.47 40.78 11.76 5.04 feed flow rate increased both the solketal exergy rate and exergy de-
14 31.83 102.42 56.80 13.43 3.22 struction rate, while this increment was less marked for the exergy
15 38.47 184.27 42.34 12.04 4.79 destruction rate. In better words, the provided chemical and electrical
16 12.38 278.58 38.26 2.74 22.50
exergies were efficiently utilized for the product formation. Unlike the
17 38.91 132.50 56.55 12.76 3.41
18 10.61 183.61 42.54 3.32 17.30 feed flow rate, increasing all the other process parameters increased the
19 19.45 248.73 42.63 4.49 12.79 normalized exergy destruction. Although loading more catalyst did not
20 43.77 162.85 45.25 14.72 3.72 affect the exergy destruction rate but it lowered the solketal exergy rate.
21 26.53 184.79 43.27 8.14 6.97
Therefore, the normalized exergy destruction was high at higher cata-
22 40.68 132.82 58.14 12.82 3.27
23 22.11 139.68 39.74 9.54 6.32
lyst loadings. Increasing acetone to glycerol molar ratio decreased both
24 55.27 201.06 52.70 13.00 3.64 the solketal exergy rate and exergy destruction rate while this decre-
25 30.07 279.51 39.70 6.49 9.30 ment was more pronounced for the solketal exergy rate. This in turn led
26 42.44 163.61 52.89 12.22 3.85 to the domination of exergy destruction rate over the solketal exergy
27 15.92 256.17 25.07 4.66 16.09
rate, elevating the normalized exergy destruction. In addition, in-
28 39.79 184.51 42.26 12.45 4.64
29 38.47 184.88 42.14 12.04 4.81 creasing both reaction temperature and pressure increased the exergy
30 39.79 184.51 42.26 12.45 4.64 destruction rate while decreased the solketal exergy rate, as discussed
31 47.75 185.45 38.91 15.73 3.88 previously. These in turn deteriorated the normalized exergy destruc-
32 27.41 140.08 37.91 12.15 5.11
tion at higher reaction temperatures and pressures.

Fig. 5. Developed fuzzy inference system (FIS1) for

modeling the universal exergetic efficiency of the
glycerol ketalization process.

257
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

Fig. 6. Some parts of the obtained rules for predicting the universal exergetic efficiency.

Table 5
Statistical performance parameters of the selected ANFIS models.

FIS Training dataset Testing dataset

2
R MSE MAPE R2 MSE MAPE

−3 −4 −8
FIS1 0.9000 1.20 × 10 3.71 × 10 0.9999 2.34 × 10 1.97 × 10−6
FIS2 0.9991 1.12 × 10−2 4.40 × 10−3 1.0000 1.74 × 10−9 4.19 × 10−6
FIS3 0.9999 1.80 × 10−3 4.50 × 10−3 1.0000 1.62 × 10−9 2.23 × 10−6

The universal exergetic efficiency of the system ranged from a solketal exergy rate and the total input exergy rate. However, this in-
minimum value of 20.96% to a maximum value of 58.14%. Like the crement was more outstanding for the solketal exergy rate, resulting in
solketal exergy rate and exergy destruction rate, feed flow rate had the an increase in the functional exergetic efficiency. Increasing acetone to
highest impact on the universal exergetic efficiency. Acetone to glycerol glycerol molar ratio increased the total inlet exergy while it decreased
molar ratio, reaction temperature, catalyst loading, and reaction pres- the solketal exergy rate, unfavorably diminishing the functional ex-
sure had the next highest impact on the universal exergetic efficiency. ergetic efficiency. Even though loading more catalyst did not affect the
Increasing feed flow rate increased both total input exergy and exergy total inlet exergy, the solketal exergy rate was decreased at higher
destruction rate. However, the total input exergy rate was increased catalyst loadings, as explained previously. Furthermore, elevating re-
with a higher rate compared with exergy destruction rate which in turn action temperature and pressure enhanced the total inlet exergy be-
enhanced the universal exergetic efficiency. Similarly, increasing re- cause of an increase in the rate of electrical energy consumption. The
action temperature and pressure increased both electrical energy con- solketal exergy rate was slightly influenced by reaction temperature,
sumption and exergy destruction rate. However, the rate of exergy while reaction temperature negatively declined the solketal formation
destruction dominated the rate of electrical energy consumption, di- rate. Accordingly, increasing catalyst loading, reaction temperature,
minishing the universal exergetic efficiency. The universal exergetic and reaction pressure negatively affected the functional exergetic effi-
efficiency was positively affected by acetone to glycerol molar ratio. ciency.
The total input exergy feeding into the system was increased with in- Fig. 5 presents the developed FIS1 for modeling the universal ex-
creasing acetone to glycerol molar ratio due to the higher standard ergetic efficiency including its fuzzy inputs (i.e., fuzzy feed flow rate,
chemical exergy of acetone compared with glycerol (Table 1). On the fuzzy acetone to glycerol molar ratio, fuzzy catalyst loading, fuzzy re-
other hand, the exergy destruction rate was decreased with increasing action temperature, and fuzzy pressure). Obviously, the fuzzy system
acetone to glycerol molar ratio, as previously explained. These in turn with 2 Gaussian membership functions for each input and 32 inference
collectively increased the universal exergetic efficiency according to Eq. rules could successfully estimate the universal exergetic efficiency.
(12). This exergetic parameter was not influenced significantly by Additionally, some parts of the obtained rules for predicting the uni-
loading more catalyst into the system. This occurred because of the fact versal exergetic efficiency are shown in Fig. 6. The same findings were
that loading more catalyst did not markedly affect neither the total obtained for both the functional exergetic efficiency and the normalized
input exergy rate nor the exergy destruction rate. exergy destruction as well.
The functional exergetic efficiency of the ketalization process varied Statistical performance parameters of the selected ANFIS models in
from a minimum value of 2.74% to a maximum value of 15.73%. Unlike training and testing steps are tabulated in Table 5. The values of R2,
the feed flow rate, increasing all the other process parameters declined MSE, and MAPE obtained using the developed ANFIS models were
the functional exergetic efficiency. The order of the effect of process within acceptable ranges. In general, all of the ANFIS models had a
parameters on the functional exergetic efficiency was feed flow good generalization capability for estimating the exergetic performance
rate > acetone to glycerol molar ratio > catalyst loading > reaction parameters of the continuous reactor applied for solketal production
temperature > pressure. Increasing feed flow rate enhanced both the from glycerol.

258
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

60
Fig. 7 shows the experimental exergetic performance parameters of
Predicted universal exergetic efficiency (%)

the reactor against the predicted values using the developed ANFIS
55
models. The data points of all the considered exergetic parameters were
well accumulated around a 45° straight line, showing the suitability of
50
the developed ANFIS models for predicting the exergetic performance
parameters of the process. Accordingly, the developed ANFIS models
45
could be reliably used for the optimization of the process.
40
Table 6 tabulates Pareto optimal front set obtained by the NSGA-II.
It should be noted that these points demonstrate the optimal trade-off
35 among the three objectives. All the points proposed by the developed
approach could be used as the best operating conditions for glycerol
30 ketalization with acetone from the exergetic viewpoint. Overall, the
reaction temperature of 40.66 °C, acetone to glycerol molar ratio of
25 4.97, feed flow rate of 0.49 mL/min, pressure of 42.31 bar, and catalyst
mass of 0.50 g yielding universal exergetic efficiency of 90.36%, func-
20 tional exergetic efficiency of 17.33%, and normalized exergy destruc-
20 25 30 35 40 45 50 55 60
tion of 6.18 (bolded line in Table 6) could be suggested as the best
Experimental universal exergetic efficiency (%)
operating conditions. Indeed, these conditions were selected by con-
16
sidering the preferences of functional exergetic efficiency, normalized
Predicted functional exergetic efficiency (%)

exergy destruction, and universal exergetic efficiency, respectively. The

14
functional exergetic efficiency associated with this point was found to
be the highest, while the normalized exergy destruction was found to be
12 the lowest among the proposed conditions. In addition, the universal
exergetic efficiency was sufficiently high under these conditions. The
10 optimum operating conditions of the reactor on the basis of process
yield were reaction temperature of 20 °C, pressure of 120 bar, acetone
8
to glycerol mole ratio of 5, feed flow rate of 0.1 mL/min, and catalyst
mass of 0.77 g. Interestingly, there were profound differences in the
optimum operating conditions proposed by the exergy analysis com-
6
pared with those proposed according to the conventional process yield.
This in turn further confirmed the suitability of the exergy analysis and
4 its extension for analyzing and optimizing complex chemical systems.
Table 7 compares the optimum universal and functional exergy ef-
2 ficiencies obtained throughout this study with those reported in the
2 4 6 8 10 12 14 16
literature. Clearly, the universal exergy efficiency of the ketalization
Experimental functional exergetic efficiency (%)
process using subcritical acetone was comparable with those reported
25 previously [28]. This means that the exergy destruction rate of the
ketalization process was within an acceptable level. In addition to that,
Predicted normalized exergy destruction

the functional exergy efficiency of the glycerol ketalization process was

20 lower than the mean values obtained using the other methods
[26,27,39]. This could be attributed to the differences in the process
characteristics as well as the formulas used for measuring the functional
15 exergy efficiency of the process. On the other hand, this study was to
the best of our knowledge the only report conducted based on the ex-
perimental data including real-world constraints and uncertainties,
10 while the other report available are simulation-based databases at ideal
conditions [26–28,39].
Overall, the exergy concept could provide a holistic framework for
5
exploring the most thermodynamically, economically, and en-
vironmentally viable conditions for chemicals synthesis compared with
many available techniques like energy analysis and life cycle assess-
ment approach. Moreover, although the standard exergy analysis could
0
0 5 10 15 20 25 inherently reveal the most economical and sustainable operational
Experimental normalized exergy destruction conditions by measuring the thermodynamic irreversibilities, addi-
tional complementary information could be obtained from its elabo-
Fig. 7. Experimental exergetic performance parameters of the reactor against the pre-
dicted ones by the developed ANFIS models.
rated extensions like exergoeconomic and exergoenvironmental

Table 6
Pareto optimal front set. (Bolded line shows the best operating conditions and corresponding exergetic performance parameters.)

Pressure (bar) Temperature (°C) Molar ratio Feed flow rate Catalyst mass Universal exergetic Functional exergetic Normalized exergy
(–) (mL/min) (g) efficiency (%) efficiency (%) destruction (–)

42.87 40.75 4.99 0.49 0.50 90.39 17.31 6.22

42.31 40.66 4.97 0.49 0.50 90.36 17.33 6.18
42.69 41.20 4.99 0.49 0.50 90.39 17.27 6.29

259
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

approaches [40]. Therefore, these analyses must be applied in the fu-

Current study
References ture in order to optimize chemical systems to achieve the highest fea-
sible thermodynamic, economic, and environmental performance.
[26]

[27]

[28]

[39]
4. Conclusions

reactor obtained at reaction temperature of 40.66 °C, acetone to glycerol molar ratio of
at water/glycerol feed ratio of 5.5, reforming temperature of 900 K, and oxygen/glycerol
This value was the optimum functional exergy efficiency of the overall process obtained

This value was the optimum functional exergy efficiency of the overall process obtained

This value was the optimum functional exergy efficiency of the overall process obtained

4.97, feed flow rate of 0.49 mL/min, pressure of 42.31 bar, and catalyst mass of 0.50 g
This value was the universal exergetic efficiency of the glycerol steam reforming unit

These values were the optimum functional and universal exergy efficiencies of the
In this study, continuous solketal synthesis from glycerol through
ketalization with subcritical acetone in the presence of Purolite PD206
at reforming and preheating temperature of 800 °C and pressure of 240 atm

catalyst was exergetically analyzed and optimized. To this end, ex-

perimental data of glycerol ketalization process at different reaction
at water/glycerol feed ratio of 6 and reforming temperature of 1100 K

temperatures, reaction pressures, acetone to glycerol mole ratios, feed

flow rates, and catalyst loadings were used to compute the exergetic
performance parameters. The computed dimensionless exergetic para-
meters viz. normalized exergy destruction, universal exergetic effi-
ciency, and functional exergetic efficiency were then modeled as a
function of the process parameters using the ANFIS approach. The de-
veloped models were fed into the NSGA-II approach to find the optimal
operating conditions of the systems by maximizing universal and
functional efficiencies and minimizing normalized exergy destruction.
The main conclusions can be summarized as follows:

• Feed flow rate had the highest effect on the exergetic variables,
annexed to a biodiesel plant

while reaction pressure showed the lowest effect on these indicators.

• The developed ANFIS models could effectively estimate the ex-
feed ratio of 0.96

ergetic performance parameters of the process with an R2 higher

than 0.99.
•
Remarks

The optimal operating conditions of the reactor were found to be

reaction temperature of 40.66 °C, reaction pressure of 42.31 bar,
acetone to glycerol mole ratio of 4.97, feed flow rate of 0.49 mL/
min, and catalyst loading of 0.50 g.
Universal exergetic efficiency

• Under the selected conditions, normalized exergy destruction, uni-

versal exergetic efficiency, and functional exergetic efficiency were
found to be 6.18%, 90.36%, and 17.33%, respectively.
of the process (%)
Comparison of the optimum functional and universal exergetic efficiency values of various glycerol conversion methods.

The selected conditions could be applied in industrial-scale con-

tinuous systems for resource-efficient, cost-effective, and en-
90.36

vironmentally-friendly solketal synthesis from glycerol. Future studies

95
–

should aim at analyzing and optimizing the glycerol ketalization pro-

Functional exergetic efficiency

cess on the basis of exergoeconomic and exergoenvironmental ap-

proaches in order to complement the results achieved throughout this
study.
of the process (%)

Acknowledgements

The authors gratefully acknowledge financial support from the Iran

17.33

National Science Foundation (Grant no. 96005466).

33.8

66.1

67.8
–

References
Power and
Product(s)

[1] Odetoye TE, Onifade KR, AbuBakar MS, Titiloye JO. Thermochemical character-
Hydrogen

Hydrogen

Hydrogen
hydrogen

Solketal

isation of Parinari polyandra Benth fruit shell. Ind Crops Prod 2013;44:62–6.
[2] Ajala OE, Aberuagba F, Odetoye TE, Ajala AM. Biodiesel: sustainable energy re-
placement to petroleum-based diesel fuel – a review. ChemBioEng Rev
2015;2:145–56.
Data type used

[3] Hajjari M, Tabatabaei M, Aghbashlo M, Ghanavati H. A review on the prospects of

Experimental
Simulation

Simulation

Simulation

Simulation

sustainable biodiesel production: a global scenario with an emphasis on waste-oil

biodiesel utilization. Renew Sustain Energy Rev 2017;72:445–64.
[4] Yazdani SS, Gonzalez R. Anaerobic fermentation of glycerol: a path to economic
viability for the biofuels industry. Curr Opin Biotechnol 2007;18:213–9.
[5] Katryniok B, Paul S, Bellière-Baca V, Rey P, Dumeignil F. Glycerol dehydration to
Glycerol conversion method

Ketalization with subcritical

acrolein in the context of new uses of glycerol. Green Chem 2010;12:2079–98.

[6] Rastegari H, Ghaziaskar HS, Yalpani M, Shafiei A. Development of a continuous
Autothermal reforming

system based on azeotropic reactive distillation to enhance triacetin selectivity in

glycerol esterification with acetic acid. Energy Fuels 2017;31:8256–62.
Supercritical water

Steam reforming

Steam reforming

[7] Faria RPV, Pereira CSM, Silva VMTM, Loureiro JM, Rodrigues AE. Glycerol valor-
reforming

ization as biofuel: thermodynamic and kinetic study of the acetalization of glycerol

acetone

with acetaldehyde. Ind Eng Chem Res 2013;52:1538–47.

[8] Gorji YM, Ghaziaskar HS. Optimization of solketalacetin synthesis as a green fuel
Table 7

additive from ketalization of Monoacetin with acetone. Ind Eng Chem Res
(s)

2016;55:6904–10.
[9] Alptekin E. Emission, injection and combustion characteristics of biodiesel and

260
M. Aghbashlo et al. Energy Conversion and Management 160 (2018) 251–261

oxygenated fuel blends in a common rail diesel engine. Energy 2017;119:44–52. production from glycerol by supercritical water reforming. Chem Eng J
[10] Garcia E, Laca M, Pérez E, Garrido A, Peinado J. New class of acetal derived from 2013;218:309–18.
glycerin as a biodiesel fuel component. Energy Fuels 2008;22:4274–80. [27] Hajjaji N, Chahbani A, Khila Z, Pons M-N. A comprehensive energy–exergy-based
[11] Frusteri F, Spadaro L, Beatrice C, Guido C. Oxygenated additives production for assessment and parametric study of a hydrogen production process using steam
diesel engine emission improvement. Chem Eng J 2007;134:239–45. glycerol reforming. Energy 2014;64:473–83.
[12] Deutsch J, Martin A, Lieske H. Investigations on heterogeneously catalysed con- [28] Boloy RAM, Ferrán SJR, e Penalva D de CL, Corrêa C, Angulo JAP, de Castro Pereira
densations of glycerol to cyclic acetals. J Catal 2007;245:428–35. Filho R. Exergetic evaluation of incorporation of hydrogen production in a biodiesel
[13] Roldán L, Mallada R, Fraile JM, Mayoral JA, Menéndez M. Glycerol upgrading by plant. Int J Hydrogen Energy 2015;40:8797–805.
ketalization in a zeolite membrane reactor. Asia-Pacific J Chem Eng 2009;4:279–84. [29] Tippawan P, Thammasit T, Assabumrungrat S, Arpornwichanop A. Using glycerol
[14] Vicente G, Melero JA, Morales G, Paniagua M, Martín E. Acetalisation of bio-gly- for hydrogen production via sorption-enhanced chemical looping reforming: ther-
cerol with acetone to produce solketal over sulfonic mesostructured silicas. Green modynamic analysis. Energy Convers Manage 2016;124:325–32.
Chem 2010;12:899–907. [30] Presciutti A, Asdrubali F, Baldinelli G, Rotili A, Malavasi M, Di Salvia G. Energy and
[15] Ferreira P, Fonseca IM, Ramos AM, Vital J, Castanheiro JE. Valorisation of glycerol exergy analysis of glycerol combustion in an innovative flameless power plant. J
by condensation with acetone over silica-included heteropolyacids. Appl Catal B Clean Prod 2017;172:3817–24.
Environ 2010;98:94–9. [31] de Souza-Santos ML. Proposals for power generation based on processes consuming
[16] Maksimov AL, Nekhaev AI, Ramazanov DN, Arinicheva YA, Dzyubenko AA, biomass-glycerol slurries. Energy 2017;120:959–74.
Khadzhiev SN. Preparation of high-octane oxygenate fuel components from plant- [32] Samoilov VO, Ramazanov DN, Nekhaev AI, Maximov AL, Bagdasarov LN.
derived polyols. Pet Chem 2011;51:61–9. Heterogeneous catalytic conversion of glycerol to oxygenated fuel additives. Fuel
[17] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu CC. A new continuous- 2016;172:310–9.
flow process for catalytic conversion of glycerol to oxygenated fuel additive: cata- [33] Esteban J, Domínguez E, Ladero M, Garcia-Ochoa F. Kinetics of the production of
lyst screening. Appl Energy 2014;123:75–81. glycerol carbonate by transesterification of glycerol with dimethyl and ethylene
[18] Shirani M, Ghaziaskar HS, Xu CC. Optimization of glycerol ketalization to produce carbonate using potassium methoxide, a highly active catalyst. Fuel Process
solketal as biodiesel additive in a continuous reactor with subcritical acetone using Technol 2015;138:243–51.
Purolite® PD206 as catalyst. Fuel Process Technol 2014;124:206–11. [34] Song G, Xiao J, Zhao H, Shen L. A unified correlation for estimating specific che-
[19] Aghbashlo M, Tabatabaei M, Mohammadi P, Khoshnevisan B, Rajaeifar MA, Pakzad mical exergy of solid and liquid fuels. Energy 2012;40:164–73.
M. Neat diesel beats waste-oriented biodiesel from the exergoeconomic and ex- [35] Aghbashlo M, Tabatabaei M, Karimi K. Exergy-based sustainability assessment of
ergoenvironmental point of views. Energy Convers Manage 2017;148:1–15. ethanol production via Mucor indicus from fructose, glucose, sucrose, and molasses.
[20] Aghbashlo M, Mobli H, Rafiee S, Madadlou A. A review on exergy analysis of drying Energy 2016;98:240–52.
processes and systems. Renew Sustain Energy Rev 2013;22:1–22. [36] Aghbashlo M, Hosseinpour S, Tabatabaei M, Dadak A. Fuzzy modeling and opti-
[21] Aghbashlo M, Tabatabaei M, Hosseinpour S, Khounani Z, Hosseini SS. Exergy-based mization of the synthesis of biodiesel from waste cooking oil (WCO) by a low
sustainability analysis of a low power, high frequency piezo-based ultrasound re- power, high frequency piezo-ultrasonic reactor. Energy 2017;132:65–78.
actor for rapid biodiesel production. Energy Convers Manage 2017;148:759–69. [37] Mandegari MA, Pahlavanzadeh H. A study on the optimization of an air dehumi-
[22] Barati MR, Aghbashlo M, Ghanavati H, Tabatabaei M, Sharifi M, Javadirad G, et al. dification desiccant system. J Therm Sci Eng Appl 2013;5:41002.
Comprehensive exergy analysis of a gas engine-equipped anaerobic digestion plant [38] Aghbashlo M, Hosseinpour S, Tabatabaei M, Younesi H, Najafpour G. On the ex-
producing electricity and biofertilizer from organic fraction of municipal solid ergetic optimization of continuous photobiological hydrogen production using hy-
waste. Energy Convers Manage 2017;151:753–63. brid ANFIS–NSGA-II (adaptive neuro-fuzzy inference system–non-dominated
[23] Aghbashlo M, Tabatabaei M, Hosseini SS, Dashti BB, Mojarab Soufiyan M. sorting genetic algorithm-II). Energy 2016;96:507–20.
Performance assessment of a wind power plant using standard exergy and extended [39] Hajjaji N, Baccar I, Pons M-N. Energy and exergy analysis as tools for optimization
exergy accounting (EEA) approaches. J Clean Prod 2018;171:127–36. of hydrogen production by glycerol autothermal reforming. Renew Energy
[24] Mandegari MA, Farzad S, Pahlavanzadeh H. Exergy performance analysis and op- 2014;71:368–80.
timization of a desiccant wheel system. J Therm Sci Eng Appl 2015;7:31013. [40] Aghbashlo M, Rosen MA. Consolidating exergoeconomic and exergoenvironmental
[25] Seager TP, Theis TL. Exergetic pollution potential: estimating the revocability of analyses using the emergy concept for better understanding energy conversion
chemical pollution. Exergy, Int J 2002;2:273–82. systems. J Clean Prod 2018;172:696–708.
[26] Ortiz FJG, Ollero P, Serrera A, Galera S. Optimization of power and hydrogen

261
 Renewable Energy 126 (2018) 242e253

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

On the exergetic optimization of solketalacetin synthesis as a green

fuel additive through ketalization of glycerol-derived monoacetin with
acetone
Mortaza Aghbashlo a, *, Meisam Tabatabaei b, c, **, Hajar Rastegari d,
Hassan S. Ghaziaskar d, Taha Roodbar Shojaei a
a
Department of Mechanical Engineering of Agricultural Machinery, Faculty of Agricultural Engineering and Technology, College of Agriculture and Natural
Resources, University of Tehran, Karaj, Iran
b
Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), P.O. Box: 31535-1897, Agricultural Research,
Education, and Extension Organization (AREEO), Karaj, Iran
c
Biofuel Research Team (BRTeam), Karaj, Iran
d
Department of Chemistry, Isfahan University of Technology, Isfahan, Iran

a r t i c l e i n f o a b s t r a c t

Article history: This study was devoted to an exergetically investigation and optimization of the operating conditions of
Received 3 November 2017 an easy-to-scale-up continuous reactor applied for solketalacetin synthesis as a green fuel additive from
Received in revised form glycerol-derived monoacetin in the presence of Purolite PD 206 catalyst. The process consisted of two
2 February 2018
steps, i.e., monoacetin synthesis by glycerol esterification with acetic acid followed by solketalacetin
Accepted 20 March 2018
Available online 21 March 2018
synthesis through reaction of the produced monoacetin with acetone. The main goal of this research was
to assess the effects of reaction temperature (20e80
C), acetone to monoacetin molar ratio (1e5),
catalyst loading (0.5e2.5 g), feed flow rate (0.2e1 mL/min), and pressure (0e120 bar) on the exergetic
Keywords:
Exergy analysis
performance parameters of the second stage of the process. Response surface methodology (RSM) was
Green fuel additive also used to optimize the operating conditions of the reactor by maximizing functional exergetic effi-
Monoacetin ciency (FEE) and minimizing normalized exergy destruction (NED), simultaneously. Overall, feed flow
Solketalacetin synthesize rate had the highest impact on the exergetic performance parameters of the reactor while these in-
Response surface method dicators were not significantly influenced by pressure. RSM successfully modeled both exergetic pa-
Glycerol valorization rameters with an R2 higher than 0.99. Reaction temperature of 30.8
C, acetone to monoacetin molar
ratio of 2.7, catalyst loading of 1.6 g, feed flow rate of 1.0 mL/min, and pressure of 14.5 bar yielding FEE of
20.39% and NED of 0.90 were determined as the best operating conditions of the reactor. According to the
results archived, process yield alone could not stand as the primary objective for making decisions on the
optimal operating conditions of the chemical reactors, further highlighting the significance of taking
energetic parameters into account in parallel.
© 2018 Elsevier Ltd. All rights reserved.

1. Introduction economically-viable, and environmentally-sustainable alternative

fuel to petrodiesel [1]. This renewable biofuel is mostly produced by
It is well-documented that biodiesel is a technically-feasible, transesterification of vegetable- and animal-based long chain fatty
acids with a short-chain alcohol like methanol and ethanol [2,3].
The products of transesterification process are methyl/ethyl esters
* Corresponding author. Department of Mechanical Engineering of Agricultural and glycerol [4]. The glycerol as the co-product of biodiesel pro-
Machinery, Faculty of Agricultural Engineering and Technology, College of Agri- duction process accounts for about 10 wt% of the total products. In
culture and Natural Resources, University of Tehran, Karaj, Iran. recent years, the increasing demand for biodiesel has accumulated
** Corresponding author. Microbial Biotechnology Department, Agricultural large quantities of glycerol on the market, lowering its price and
Biotechnology Research Institute of Iran (ABRII), P.O. Box: 31535-1897, Agricultural
Research, Education, and Extension Organization (AREEO), Karaj, Iran.
jeopardizing the economic vitality of biodiesel industry. Therefore,
E-mail addresses: [email protected] (M. Aghbashlo), meisam_tabatabaei@ it is essential to add more value to low-cost glycerol using efficient,
abrii.ac.ir (M. Tabatabaei). productive, and sustainable conversion routes for boosting

https://doi.org/10.1016/j.renene.2018.03.047
0960-1481/© 2018 Elsevier Ltd. All rights reserved.
 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253 243

Nomenclature W_ Work flow rate (kJ/s)

A Ash percentage (%) Greek letters

C Catalyst loading (g) b0 ; bi ; bii ; bij Coefficients
ℂ Carbon percentage (%) ε Standard chemical exergy (kJ/mol)
Cp Specific heat capacity (kJ/kg K) J Normalized exergy destruction ()
ex Specific chemical exergy (kJ/kg) f Functional exergetic efficiency (%)
_
Ex Exergy flow rate (kJ/s)
F Feed flow rate (mL/min) Subscripts
ℍ Hydrogen percentage (%) 0 Dead state
M Molar mass (g/mol) AC Acetone
m_ Mass flow rate (kg/s) ch Chemical
n Number of the process variables DA Diacetin
n_ Molar flow rate (mol/s) des Destruction
ℕ Nitrogen percentage (%) GL Glycerol
O Oxygen percentage (%) in Inflow
P Pressure (bar) i; j Numerator
Q_ Heat transfer rate (kJ/s) l Loss
R Universal gas constant (kJ/mol k) MA Monoacetin
S Sulfur percentage (%) out Outflow
T Temperature (
C or K) P Product
v Specific volume (m3/kg) ph Physical
x Mass fraction () q Heat transfer
x Actual independent factor SK Solketal
X Acetic acid to glycerol molar ratio () SA Solketalacetin
y Molar fraction () TA Triacetin
y Response

biodiesel industry. solketalacetin synthesis. Briefly, in the first step, monoacetin was
Biodiesel-derived glycerol can also be used as a burner fuel and obtained by esterifying glycerol with acetic acid. In the second step,
combusted to produce heat and electricity. However, lower heating solketalacetin was produced by reacting monoacetin with acetone.
value and ineffective combustion of raw glycerol and generation of Despite the promising results achieved, it is yet crucial to investi-
highly-toxic acrolein hinder its application as a fuel [5]. In order to gate the most resource-efficient, cost-effective, and eco-benign
address these issues, biodiesel-derived glycerol can be converted solketalacetin synthesis conditions for application in the biodiesel
into oxygenated fuel additives. The obtained additives can not only industry. To this end, advanced engineering methodologies like
recover the energy content of glycerol but can also mitigate haz- exergy analysis must be applied to scrutinize efficiency, produc-
ardous emissions owing to the presence of structural oxygen. tivity, and sustainability of glycerol valorization systems.
Glycerol can be upgraded to oxygenated fuel additives using Simply speaking, exergy is the maximum obtainable work from
various pathways including etherification [6], esterification [7], an energy/material flow as it approaches equilibrium with its
acetalization [8], and ketalization [9]. Among various fuel additives environment through reversible processes [12e14]. Exergy analysis
obtained from glycerol, solketal or (2,2-dimethyl-1,3-dioxolan-4- can reliably quantify the thermodynamic non-idealities or irre-
yl)methanol produced through ketalization process can be used versibilities of the process under consideration [15,16]. The amount
as fuel viscosity and flash point improver. Garcia et al. [10] syn- of thermodynamic irreversibilities is closely related to the quantity
thesized a new oxygenated biodiesel additive, i.e., (2,2-dimethyl- of resources depleted during a process, determining its degree of
1,3-dioxolan-4-yl) methyl acetate, called “solketalacetin” by Gorji sustainability [17,18]. Furthermore, there is a direct association
and Ghaziaskar [11], through a slight modification of solketal. This between the economic value and the exergy quantity as presented
derivative is a valuable additive for increasing fuel flash point, by Bejan et al. [19]. Therefore, this concept can be used to analyze
complying with the standards required by EN 14214 and ASTM and optimize various energy and material conversion processes
D6751 for flash point [10]. Moreover, it can decrease fuel viscosity with regard to thermodynamic, economic, and environmental is-
significantly, resulting in increased atomization and more complete sues [20,21]. Accordingly, a tremendous number of papers dealing
combustion [10]. with exergy analysis of glycerol upgrading processes have been
Garcia et al. [10] synthesized solketalacetin in a two-step pro- published in highly-referred journals [22e27].
cess. In the first step, solketal was synthesized by mixing glycerol On the basis of the outcomes of the above-mentioned studies,
and acetone in the presence of para-toluene sulfonic acid as ho- exergy analysis could aid in developing strategies and guidelines to
mogenous catalyst. In the second step, after heating the reaction make thermodynamic systems more efficient, productive, and
mixture to reflux for 16 h, solketalacetin was obtained by reacting sustainable. Therefore, this study was aimed at exergetically opti-
solketal with acetic anhydride in trimethylamine solution at room mizing the operational conditions of an easy-to-scale-up contin-
temperature for 4 h. However, this method cannot be easily scaled- uous reactor applied for solketalacetin synthesis by reacting the
up due to the higher price of acetic anhydride if compared with glycerol-originated monoacetin with acetone in the presence of
acetic acid [9]. In order to solve this problem, Gorji and Ghaziaskar Purolite PD 206 catalyst. More specifically, the main goal of this
[11] introduced a new easy-to-scale-up two-step procedure for study was to investigate the effects of reaction temperature
244 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253

(20e80
C), acetone to monoacetin molar ratio (1e5), catalyst

loading (0.5e2.5 g), feed flow rate (0.2e1 mL/min), and pressure
(0e120 bar) on the exergetic performance parameters of the sec-
ond stage of the solketalacetin synthesis proposed by Gorji and
Ghaziaskar [11]. Response surface method (RSM) was also used to
correlate two more relevant exergetic indicators, i.e., functional
exergetic efficiency (FEE) and normalized exergy destruction (NED)
with the process variables. The optimal operating conditions of the
developed reactor were then obtained using the developed models
by maximizing the FEE and minimizing the NED. The findings of
this study are expected to be of paramount interest to engineers
involved with design and optimization of industrial-scale reactors
being employed for solketalacetin synthesis.

Fig. 1. The experimental set-up used for ketalization of monoacetin with acetone. 1)
2. Materials and methods Feed vessel, 2) HPLC pump, 3) On-off valve, 4) Air oven, 5) Preheater, 6) Tubular
reactor, 7) Back-pressure regulator, and 8) Sample collection vessel.
2.1. Materials

Acetic acid (purity > 99.85%), glycerol (purity > 99.9%), acetone 2.3. Theoretical considerations
(purity > 99%), triacetin (purity > 99%) and diacetin (purity ¼ 50%),
Purolite PD206, methanol (purity>99.8%), and 2-ethylhexanol In this study, the following assumptions were considered in
(purity > 99%) were purchased from Fanavaran Petrochemical Co. developing the exergy model for the applied continuous system:
(Iran), Emery Oleochemicals (Malaysia), Sasol Co (South Africa),
Fluka (Germany), Purolite Co. (USA), Merck Co. (Germany), and Tat  The reactor as a control volume operated at a steady state
Chemical Co. (Iran), respectively. Monoacetin (purity > 95%) and condition.
solketalacetin (purity >95%) were synthesized by methods  The kinetic and potential exergies of the inflow/outflow streams
described in the literature [10]. were neglected.
 The dead state temperature and pressure were taken into ac-
count to be 25
C and 1.01 bar, respectively.
 The dead state temperature and pressure did not vary with time.
2.2. Experimental procedure and analytical method
Fig. 2 shows the schematic illustration of the reactor with input
Monoacetin was synthesized in batch mode in a 500 mL three
and output terms. It should be noted that a proportion of the re-
neck flask at a temperature of 85
C and at atmospheric pressure.
actants was exhausted from the system without participating in the
The mixture of acetic acid and glycerol with a molar ratio of 6 was
acetalization and ketalization reactions.
stirred magnetically for 4 h. Then, the solution was vacuum distilled
According to Fig. 2, the mass flow rate balance for the developed
to remove water and acetic acid. At this stage, a conversion rate of
system could be expressed as follows:
95% was achieved for glycerol with product selectivities of 63%,
33%, and 4.0%, for monoacetin, diacetin, and triacetin, respectively.
In the second stage, solketalacetin was synthesized via monoacetin
ketalization with acetone in a continuous flow system shown in
Fig. 1. All the experiments were carried out in a stainless steel
tubular reactor with 8 mm internal diameter and 25 cm length
placed in an air oven. The reactor was filled with stainless steel
filings and specific amount of Purolite PD 206 as catalyst. At the
beginning of each experiment, the reactor was heated at a deter-
mined temperature. Then, the feedstock was pumped into the
reactor using an HPLC pump (model PU-980, JASCO Co). The pres-
sure was adjusted via a back pressure regulator (model BP 1580-81,
JASCO Co.). Several samples were collected at different time
intervals.
Qualitative identification of the products was carried out by a
GC-MS (model 6890 N, Agilent Technologies), while Quantitative
analyses of the samples were carried out using a GC-FID (model
3420, Beifen, China). The carrier gas was Argon and the capillary
column used was an HP-5 (30 m length, 0.25 mm internal diameter,
and 0.25 mm film thickness). The GC injection port and the detector
temperature were set at 280
C and 300
C, respectively. Methanol
was used as solvent while 2-ethyl-hexanol as internal standard. The
following temperature program was used: the initial column tem-
perature was set at 60
C for 4 min, increased to 85
C at a rate of
45
C/min, ramped to 270
C at the rate of 30
C/min, and was held
at 270
C for 5 min. The conversion and solketalacetin yield were Fig. 2. The schematic illustration of the reactor used throughout this study with input
calculated according to Gorji and Ghaziaskar [11]. and output terms.
 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253 245

m_ in ¼ m_ out (1) _ ch
_ P ¼ Ex _ ch
Ex SK;out þ ExSA;out (12)

m_ GL;in þ m_ AC;in þ m_ MA;out þ m_ DA;out þ m_ TA;out The FEE of the process was computed as follows:

¼ m_ SK;out þ m_ SA;out þ m_ GL;in þ m_ AC;in þ m_ MA;out þ m_ DA;out _ ch _ ch

Ex SK;out þ ExSKA;out
þ m_ TA;out f¼  100 (13)
_ ch þ Ex
Ex _ ph þ W_
in in
(2)
The NED of the system was also computed to make decisions on
The exergy balance for the system could be expressed as
the process variables by taking into account the quantity of exergy
follows:
destroyed during the process and the exergy rate of the products
_ þW_ ¼ Ex
_ out þ Ex
_ _ synthesized.
Exin q;l þ Exdes (3)
_
Ex
The exergy rate of the inflow/outflow streams to/from the sys- J¼ des
(14)
tem could be determined based on the summation of their chem- _ ch
Ex _ ch
þ Ex
SK;out SA;out
ical and physical exergies. Accordingly, the above-mentioned
exergetic balance equation can be reformulated as follows:

_ ch þ Ex
_ ph þ W _ ch þ Ex
_ ¼ Ex _ ph þ Ex
_ _ 2.4. Experimental design
Ex in in out out q;l þ Exdes (4)

The heat loss rate from the reactor frame to the surroundings The effects of process variables on the exergetic performance
was determined using the first law of thermodynamics as follows: parameters of the reactor, applied for the ketalization of glycerol-
derived monoacetin with acetone, were investigated using design
Q_ l ¼ m_ in Cp;in ðTin  T0 Þ þ W_  m_ out Cp;out ðTout  T Þ
0 (5) of experiment. The process parameters were: reaction temperature
(T), acetone to monoacetin molar ratio (X), catalyst loading (C), feed
According to the above-mentioned equation, the heat loss rate flow rate (F), and pressure (P). Central Composite Design (CCD) at
from the reactor frame to the surroundings was found to be three coded levels (1, 0, 1) was considered throughout this study.
negligible. Therefore, the exergy rate balance can be rewritten as The proposed experimental trials by the RSM approach were then
follows: performed according to the design matrix tabulated in Tables 2 and
3. Once all the experiments were carried out completely, the main
_ ch þ Ex
Ex _ ph þ W _ ch þ Ex
_ ¼ Ex _ ph þ Ex
_ (6) exergetic parameters, i.e., FEE and NED were analyzed by the RSM.
in in out out des
_ supplied to the system was measured by The first and main step in the RSM approach is to find an appro-
The work rate (W)
priate model between responses and process variables. This pro-
recording the electrical power consumed by heating element and
cess is often begun with a low order polynomial. If the responses
feeding pump. The chemical and physical exergies of the inflow/
are not well approximated by a linear function of process variables,
outflow streams to/from the system could be measured using the
a high order polynomial must be applied as indicated below:
following equations, respectively [28,29]:
! X
n X
n X
n X
n
ch X X y ¼ b0 þ bi xi þ bii x2i þ bij xi xj (15)
_
Ex ¼ n_ yi εi þ RT0 yi lnðyi Þ (7) i¼1 i¼1 i¼1 j > 1
i i
The Design Expert version 7.0.0 software (Stat-Ease, Inc., Min-
    neapolis, MN) was applied to analyze the obtained data.
_ph T InT
Ex ¼ m_ Cp T  T0  0 þ vðP  P0 Þ (8)
T0
3. Results and discussion
The following formula was used to determine the specific heat
capacity of the outflow stream. The proposed experiments by the RSM and their corresponding
X solketalacetin yield and exergetic parameters are summarized in
CP ¼ xi Cp;i (9) Table 2. The maximum solketalacetin yield was found to be 55% at
i acetone to monoacetin molar ratio of 4, catalyst loading of 2 g, re-
The standard chemical exergy values of the glycerol, acetic acid, action temperature of 35
C, feed flow rate of 0.4 mL/min, and
monoacetin, diacetin, and triacetin, acetone, solketal, and sol-
ketalacetin were estimated using the mathematical model pre- Table 1
sented by Song et al. [30]. Chemical formulas and standard chemical exergy values of the components involved
in the process.
exi ¼ 363:439C þ 1075:633H  86:308O þ 4:14N þ 190:798S
Component Chemical formula Standard chemical exergya (kJ/mol)
 21:1A Glycerol C3H8O3 1762.63
(10) Acetic acid C2H4O2 1030.53
Monoacetin C5H10O4 2714.42
Diacetin C7H12O5 3666.20
εi ¼ Mi exi (11) Triacetin C9H14O6 4617.98
Acetone C3H6O 1821.97
Table 1 summarizes the chemical formulas and the standard Solketal C6H12O3 3505.85
chemical exergy values of the components involved in the reaction. Solketalacetin C8H14O4 4457.63
The exergy rate of the outflow products was computed using the a
The standard chemical exergy values of the component were computed using
following equation: the equation presented by Song et al. [30].
246 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253

Table 2
Experiments proposed by RSM and their corresponding solketalacetin yield and exergetic parameters.

Run Molar ratio Catalyst mass Temperature Feed flow rate (mL/ Pressure Solketalacetin yield Products exergy rate Exergy destruction rate NED FEE
() (g) (
C) min) (bar) (%) (W) (W) () (%)

1 3 1.5 80 0.6 60 41 62.98 278.92 4.43 11.22

2 3 1.5 50 0.6 60 46 71.17 188.58 2.65 15.10
3 3 1.5 50 1.0 60 48 127.02 220.71 1.74 17.94
4 3 1.5 50 0.6 0 47 71.17 140.58 1.98 16.81
5 4 2.0 65 0.8 90 42 72.77 266.79 3.67 11.16
6 2 2.0 35 0.4 30 38 46.71 111.85 2.39 16.11
7 4 2.0 35 0.8 30 54 90.90 127.47 1.40 17.71
8 3 1.5 50 0.6 60 48 73.90 188.00 2.54 15.68
9 2 1.0 65 0.8 90 34 87.41 265.04 3.03 13.97
10 5 1.5 50 0.6 60 47 54.79 189.53 3.46 11.29
11 1 1.5 50 0.6 60 34 82.19 186.02 2.26 18.63
12 3 1.5 50 0.6 60 50 76.63 188.62 2.46 16.25
13 3 1.5 50 0.6 60 46 71.17 188.58 2.65 15.10
14 2 2.0 65 0.8 30 35 87.27 201.42 2.31 15.53
15 3 2.5 50 0.6 60 30 49.33 189.52 3.84 10.46
16 3 1.5 20 0.6 60 55 82.09 124.94 1.52 20.13
17 4 2.0 65 0.4 30 47 39.54 187.62 4.75 10.46
18 3 1.5 50 0.2 60 52 25.14 157.63 6.27 10.04
19 3 1.5 50 0.6 60 48 73.90 188.00 2.54 15.68
20 4 1.0 65 0.4 90 30 27.00 218.62 8.10 6.59
21 3 0.5 50 0.6 60 19 32.96 189.37 5.75 6.99
22 2 1.0 35 0.8 30 36 90.90 125.73 1.38 18.67
23 2 1.0 35 0.4 90 37 44.91 142.95 3.18 13.97
24 2 2.0 35 0.8 90 32 80.02 190.18 2.38 14.53
25 2 2.0 65 0.4 90 32 39.54 218.26 5.52 9.97
26 4 1.0 35 0.4 30 41 35.96 112.63 3.13 11.87
27 3 1.5 50 0.6 60 50 76.63 188.62 2.46 16.25
28 3 1.5 50 0.6 120 44 68.44 235.96 3.45 13.18
29 4 1.0 65 0.8 30 30 54.64 202.09 3.70 9.29
30 4 1.0 35 0.8 90 40 72.77 191.86 2.64 12.61
31 2 1.0 65 0.4 30 36 44.91 186.41 4.15 12.31
32 4 2.0 35 0.4 90 55 46.71 144.22 3.09 13.97

Table 3
The ANOVA table for the NED and FEE.

1/Sqrt (NED) 1/Sqrt (FEE)

Source Squares df Square Value Prob > F Source Squares df Square Value Prob > F
Model 0.46 21 0.022 132.02 <0.0001 Model 0.049 22 2.22E-03 70.96 <0.0001
X 0.018 1 0.018 109.35 <0.0001 X 6.72E-03 1 6.72E-03 215.02 <0.0001
C 7.76E-03 1 7.76E-03 46.51 <0.0001 C 2.39E-03 1 2.39E-03 76.33 <0.0001
T 0.16 1 0.16 968.89 <0.0001 T 0.011 1 0.011 350.54 <0.0001
F 0.065 1 0.065 386.94 <0.0001 F 3.16E-03 1 3.16E-03 101.11 <0.0001
P 0.052 1 0.052 309.15 <0.0001 P 2.22E-03 1 2.22E-03 70.83 <0.0001
XC 0.012 1 0.012 72.47 <0.0001 XC 2.65E-03 1 2.65E-03 84.77 <0.0001
XT 2.14E-03 1 2.14E-03 12.83 0.005 XT 1.16E-03 1 1.16E-03 37.19 0.0002
XF 1.58E-04 1 1.58E-04 0.95 0.353 XF 2.64E-06 1 2.64E-06 0.084 0.778
XP 7.10E-04 1 7.10E-04 4.26 0.066 XP 2.99E-05 1 2.99E-05 0.96 0.3536
CT 9.00E-06 1 9.00E-06 0.054 0.821 CT 8.61E-05 1 8.61E-05 2.75 0.1314
CF 2.23E-05 1 2.23E-05 0.13 0.7222 CF 7.23E-05 1 7.23E-05 2.31 0.1626
CP 8.74E-04 1 8.74E-04 5.24 0.0451 CP 3.90E-05 1 3.90E-05 1.25 0.2928
TF 3.12E-04 1 3.12E-04 1.87 0.2015 TF 5.06E-04 1 5.06E-04 16.19 0.003
TP 4.14E-03 1 4.14E-03 24.8 0.0006 TP 4.33E-05 1 4.33E-05 1.38 0.2695
FP 4.14E-03 1 4.14E-03 24.84 0.0006 FP 1.42E-04 1 1.42E-04 4.54 0.0619
X2 5.80E-04 1 5.80E-04 3.48 0.0918 X2 1.40E-04 1 1.40E-04 4.49 0.0632
C2 0.044 1 0.044 265.14 <0.0001 C2 0.014 1 0.014 451.7 <0.0001
T2 1.05E-03 1 1.05E-03 6.29 0.0311 T2 4.24E-05 1 4.24E-05 1.36 0.2741
F2 2.93E-03 1 2.93E-03 17.57 0.0019 F2 7.28E-04 1 7.28E-04 23.29 0.0009
P2 6.77E-05 1 6.77E-05 0.41 0.5384 P2 2.65E-05 1 2.65E-05 0.85 0.3811
X2 F 1.53E-03 1 1.53E-03 9.15 0.0128 X2C 3.58E-04 1 3.58E-04 11.44 0.0081
X2F 2.30E-04 1 2.30E-04 7.35 0.024
Residual 1.67E-03 10 1.67E-04 Residual 2.81E-04 9 3.13E-05
Lack of fit 1.14E-03 5 2.27E-04 2.13 0.2133 Lack of fit 1.94E-04 4 4.86E-05 2.78 0.1457
Pure error 5.33E-04 5 1.07E-04 Pure error 8.72E-05 5 1.74E-05
Cor total 0.46 31 Cor total 0.049 31
R-Squared 0.9964 R-Squared 0.9943
Adj. R-Squared 0.9889 Adj. R-Squared 0.9803
Adeq. precision 46.861 Adeq. precision 36.467
C.V. % 2.19 C.V. % 2.02
Press 0.053 Press 0.015
 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253 247

Fig. 3. Response surface plots for the NED during esterification of glycerol with acetic acid in the developed reactor.

pressure 90 bar. The minimum solketalacetin yield was determined Increasing catalyst loading provided numerous active sites and
as 19% at acetone to monoacetin molar ratio of 3, catalyst loading of sufficient contact, prompting the ketalization process of mono-
0.5 g, reaction temperature of 50
C, feed flow rate of 0.6 mL/min, acetin. With increasing feed flow rate, a decrement in the sol-
and pressure 60 bar. According to the data presented in Table 2, the ketalacetin yield was observed as a result of decreased residence
importance of process variables on the solketalacetin yield could be time. Furthermore, increasing acetone to monoacetin molar ratio
ranked as follows: catalyst loading > feed flow rate > acetone to positively enhanced the solketalacetin yield. Even though 1 mol of
monoacetin molar ratio > reaction temperature > pressure. acetone is stoichiometrically required to convert monoacetin to
248 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253

Fig. 3. (continued).

solketalacetin completely, higher quantities must be used to drive not be used alone to make decisions on optimum operating con-
the ketalization in a forward direction for obtaining higher yields. ditions. In better words, the process yield could optimize the
However, further increments of acetone to monoacetin molar ratio monoacetin ketalization process based on the solketalacetin yield
not only could reduce the reaction rate due to dilution of the other while such conditions could not necessarily lead to the most
reagents but also could cause difficulties in downstream processes, resource-efficient, cost-effective, and eco-benign conditions as
i.e., products purification and separation. well. Accordingly, exergy analysis was applied throughout this
On the basis of the data reported in Table 2, increasing reaction study for investigating the ketalization process of monoacetin at
temperature negatively impacted the solketalacetin yield due to different acetone to monoacetin molar ratios, catalyst loadings,
the exothermic nature of monoacetin ketalization. On the other reaction temperatures, feed flow rates, and reaction pressures.
hand, enhancing reaction temperature beyond 56
C (at ambient Based on the data summarized in Table 2, the products exergy
pressure of 1 bar) could evaporate a portion of the utilized acetone, rate varied between 25.14 W and 127.02 W. Feed flow rate had the
diminishing the probability of successful collision of monoacetin highest impact on the products exergy rate, followed by catalyst
molecules with acetone. Increasing reaction pressure slightly loading, acetone to monoacetin molar ratio, reaction temperature,
changed the ketalization of monoacetin. A steady increase in the and reaction pressure. Increasing feed flow rate markedly increased
solketalacetin yield could be observed with increasing reaction the products exergy rate due to an increase in the rate of reactants
pressure up to 45 bar. Increasing reaction pressure increased the fed into the systems. Similar to solketalacetin yield, loading more
boiling temperature of acetone, thereby decreasing its evaporation catalyst into the reactor improved the products exergy rate since
rate. This in turn increased the possibility of successful collision of increasing catalyst concentration provided a large surface area for
monoacetin with acetone. However, the process yield was nega- the monoacetin ketalization. In addition, increasing acetone to
tively impacted beyond 45 bar because of catalyst agglomeration monoacetin molar ratio significantly lowered the products exergy
and, subsequent reaction deceleration. It should be noted that rate because of a decrease in the amount of monoacetin supplied to
increasing pressure drop across the reactor beyond a certain pres- the reactor. This in turn decreased the rate of products evolved and,
sure could result in adhesion, fouling, and accumulation of small consequently, decreased the products exergy rate. Increasing re-
beads of catalyst, thereby decreasing its catalytic activity [31]. action temperature decreased the products exergy rate owing to its
Although solketalacetin yield was used by Gorji and Ghaziaskar [11] negative effect on the solketalacetin yield, as previously discussed.
for optimizing the performance of the developed reactor, it could The products exergy rate was trivially decreased with increasing
 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253 249

Fig. 4. Response surface plots for the FEE during esterification of glycerol with acetic acid in the developed reactor.

reaction temperature, attributing to trivial changes in the viscosity general, increasing all five process variables increased the exergy
and diffusion coefficients of the liquid reaction media. destruction rate. Overall, exergy destruction in the ketalization
The exergy destruction rate was found to be ranging from process could be attributed to the conversion of the high-quality
111.85 W to 278.92 W (Table 2). The importance of process variables electrical power to low-quality thermomechanical energy. It
on the exergy destruction rate of the process could be ordered as should be noted that the electrical power was used by the heating
follows: feed flow rate > reaction temperature > reaction pres- element for heating the reaction media and by the HPLC pump for
sure > acetone to monoacetin molar ratio > catalyst loading. In feeding the reactants into the reactor. Severe chemical reactions
250 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253

Fig. 4. (continued).

and intensive heat transfer could also be mentioned as reasons for process.
the thermodynamic irreversibilities occurring in the ketalization The NED ranged from a minimum value of 1.38 to a maximum
process. The exergy destruction rate of the process was markedly value of 8.10 (Table 2). According to the experimental data ob-
increased with increasing feed flow rate due to an increase in the tained, the order of the effect of process variables on the NED was
rate of thermodynamically irreversible acetalization and ketaliza- feed flow rate > catalyst loading > acetone to monoacetin molar
tion reactions. Increasing both reaction temperature and pressure ratio > reaction temperature > reaction pressure. The NED
profoundly elevated the rate of electrical energy required by the decreased with increasing feed flow rate and catalyst loading, while
system and, consequently, increased the exergy destruction rate. increased with increasing acetone to monoacetin molar ratio, re-
Despite the fact that increasing acetone to monoacetin molar ratio action temperature, and reaction pressure. Increasing feed flow rate
decreased the monoacetin content being fed into the reactor, the and catalyst loading increased both the products exergy rate and
solketalacetin yield increased. This meant that the majority of the exergy destruction rate, while the rate of products exergy incre-
monoacetin fed into the reactor was successfully converted to ment dominated over that of exergy destruction rate. Accordingly,
solketalacetin, leading to an increase in the rate of exergy the NED increased with increasing both feed flow rate and catalyst
destruction. There was a direct association between solketalacetin loading. Increasing acetone to monoacetin molar ratio, reaction
yield and exergy destruction rate due to an increase in the rate of temperature, and reaction pressure increased the exergy destruc-
ketalization reaction. It could be concluded that maintaining the tion rate, while decreased the products exergy, as previously
rate of ketalization process high enough at lowest possible elec- elucidated. These in turn collectively increased the NED with
trical energy consumption could be an efficient strategy for miti- increasing acetone to monoacetin molar ratio, reaction tempera-
gating the rate of exergy destruction. ture, and reaction pressure. In better words, increasing these pro-
Despite the fact that both products exergy rate and exergy cess parameters increased resource depletion in producing a given
destruction rate could provide useful dimensioned information on amount of solketalacetin and solketal.
the thermodynamic performance of the reactor, more informative According to the data tabulated in Table 2, the FEE was found to
dimensionless indices should be defined for assessing and opti- be in the range of 6.59%e20.13%. The order of the effect of process
mizing the process. In this regards, NED and FEE of the process were variables on the FEE was feed flow rate > acetone to monoacetin
computed for all the experiments proposed by the RSM and the molar ratio > catalyst loading > reaction temperature > reaction
developed models were then used for optimizing the ketalization pressure. Increasing feed flow rate increased the total input exergy
 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253 251

Table 4
Developed models for the NED and FEE in terms of actual factors.

1/Sqrt (NED) 1/Sqrt (FEE)

Term Coefficient Term Coefficient

Constant 0.86229 Constant 4.79E-03
X 0.35321 X 0.28334
C 0.36857 C 0.07678
T 7.22E-03 T 5.55E-04
F 0.28685 F 0.14428
P 1.85E-03 P 2.89E-04
XC 0.054965 XC 0.12401
XT 7.71E-04 XT 5.68E-04
XF 0.49183 XF 1.99E-01
XP 2.22E-04 XP 4.56E-05
CT 1.00E-04 CT 3.09E-04
CF 0.01181 CF 2.13E-02
CP 4.93E-04 CP 1.04E-04
TF 1.47E-03 TF 1.87E-03
TP 3.57E-05 TP 3.66E-06
FP 2.68E-03 FP 4.97E-04
X2 0.046309 X2 4.21E-02
C2 0.15529 C2 0.087776
T2 2.66E-05 T2 5.34E-06
F2 0.24985 F2 1.25E-01
P2 1.69E-06 P2 1.06E-06
X2 F 0.08459 X2C 1.64E-02
X2 F 3.28E-02

Fig. 5. Plots of the predicted data against actual values for the NED and FEE.

Table 5 hand, increasing feed flow rate improved the products exergy rate,
Optimal operating conditions of the reactor proposed by RSM. as previously explained. Overall, the products exergy rate was the
Number X () C (g) T (
C) F (mL/min) P (bar) NED FEE most influential factor on the FEE compared with the total input
exergy rate, which in turn improved the FEE of the process with
1 2.7 1.6 30.8 1.0 14.5 0.90 20.39
2 3.0 1.7 36.9 0.8 20.1 1.26 21.09 increasing feed flow rate. Increasing acetone to monoacetin molar
3 1.0 0.7 36.0 0.7 0.2 1.37 21.63 ratio diminished concurrently both the total input exergy rate and
4 2.3 1.7 37.0 0.8 6.6 1.20 21.24 the products exergy rate. Notably, the standard chemical exergy of
5 2.2 1.3 28.4 0.8 25.9 1.16 21.03 acetone was very lower than that of monoacetin. Therefore,
6 3.5 1.8 25.6 0.8 28.9 1.14 21.96
7 1.6 1.3 34.2 0.7 9.3 1.24 22.35
increasing acetone to monoacetin molar ratio markedly decreased
8 1.3 1.1 24.9 0.8 47.0 1.36 20.81 the total input exergy rate. Unlike the feed flow rate, the total input
9 2.1 1.5 23.9 0.6 14.3 1.23 22.68 exergy rate dominated over the products exergy, lowering the FEE
10 1.9 1.4 31.3 0.8 4.0 1.08 21.75 with elevating acetone to monoacetin molar ratio. Increasing
catalyst loading profoundly improved the products exergy rate,
rate due to an increase in the rate of electrical energy consumption while did not affect the total input exergy rate. Accordingly, the FEE
as well as in the rate of reactants fed into the reactor. On the other of the process was significantly boosted by loading more catalyst.
252 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253

Increasing both reaction temperature and pressure remarkably such complementary studies could minimize the thermodynamic
increased the rate of electrical energy consumption, while this irreversibilities, the overall costs, and the overall environmental
increment did not improve the products exergy rate. Therefore, the impacts of solketalacetin synthesis.
FEE was decreased with increasing both reaction temperature and
pressure. Overall, the association between the process variables 4. Conclusions
and the exergetic performance parameters of the reactor was very
complex, emphasizing on the fact that multi-objective optimization This work was aimed at exergetically investigating and opti-
should be carried out to lower the NED and elevate the FEE, mizing an easy-to-scale-up continuous reactor applied for sol-
simultaneously. ketalacetin synthesis from glycerol-derived monoacetin in the
Both the NED and FEE of the process were modeled via the presence of Purolite PD 206 catalyst. Two quadratic models were
second order quadratic polynomial model. The effects of the pro- developed to correlate the NED and FEE with five input variables,
cess variables on the exergetic performance parameters were i.e., feed flow rate, acetone to monoacetin molar ratio, catalyst
evaluated by using Analysis of Variance (ANOVA). The lack-of-fit loading, reaction temperature, and reaction pressure. Unlike the
values of both models were insignificant (P > 0.05). It is worth FEE, increasing feed flow rate and catalyst loading decreased the
mentioning that a given term was used in the model if its incor- NED. In contrast with the NED, elevating acetone to monoacetin
poration elevated the lack of fit. Furthermore, R2, Adj-R2, Pred-R2 molar ratio, reaction temperature, and reaction pressure decreased
and Adq. precision were high enough, while C.V.% and PRESS were the FEE. Overall, the developed models could excellently predict the
low enough for both models, showing their accuracy and reliability exergetic performance parameters of the reactor with a coefficient
for navigating the design space. Figs. 3 and 4 indicate the combined of determination higher than 0.99. The optimum operating condi-
effects of two terms on each exergetic performance parameters of tions of the reactor were determined as feed flow rate of 1.0 mL/
the process. In these figures, the other remaining variables were min, acetone to monoacetin molar ratio of 2.7, catalyst loading of
considered fixed at the central value. In fact, these figures were 1.6 g, reaction temperature of 30.8
C, and reaction pressure of
used to visually show the effects of process parameters on the NED 14.5 bar. These conditions yielded the NED of 0.90 and FEE of
and FEE as two main exergetic parameters. The explanations 20.39%. The present study confirmed the capability of the exergy
related to these figures were comprehensively presented in the concept for finding the most thermodynamically, economically, and
previous paragraphs. environmentally favorable conditions for valorization of glycerol
The developed models for the NED and FEE in terms of actual and its derivatives. Finally, future studies should be devoted to
factors are tabulated in Table 4. allocating economic and environmental costs of the solketalacetin
The validation of the developed models was also carried out by synthesis by means of exergoeconomic and exergoenvironmental
comparing the actual data with the estimated values (Fig. 5). analyses.
Obviously, the data points were well located around a straight line
with a slope equal to 1, showing the adequacy and reliability of both Acknowledgements
models in approximating the exergetic performance parameters of
the reactor as a function of its process variables. Therefore, the The authors gratefully acknowledge financial support from the
developed models could be reliably applied to optimize the oper- Iran National Science Foundation (Grant no. 96005466).
ating conditions of the reactor from exergetic viewpoint.
Among the optimal operating conditions proposed by the RSM References
approach, feed flow rate of 1.0 mL/min, acetone to monoacetin
molar ratio of 2.7, catalyst loading of 1.6 g, reaction temperature of [1] E. Khalife, M. Tabatabaei, A. Demirbas, M. Aghbashlo, Impacts of additives on
30.8
C, and reaction pressure of 14.5 bar could be suggested as the performance and emission characteristics of diesel engines during steady
state operation, Prog. Energy Combust. Sci. 59 (2017) 32e78.
best operating conditions of the reactor (bold line in Table 5). The
[2] S. Hosseinpour, M. Aghbashlo, M. Tabatabaei, E. Khalife, Exact estimation of
selected operating conditions could yield an NED of 0.90 and an FEE biodiesel cetane number (CN) from its fatty acid methyl esters (FAMEs) profile
of 20.39%. These conditions were chosen by taking into account a using partial least square (PLS) adapted by artificial neural network (ANN),
Energy Convers. Manag. 124 (2016) 389e398.
similar level of importance for both objectives considered herein.
[3] M. Hajjari, M. Tabatabaei, M. Aghbashlo, H. Ghanavati, A review on the
The NED of the selected condition was very close to the lowest prospects of sustainable biodiesel production: a global scenario with an
value (0.90), resulting in minimized resource depletion for syn- emphasis on waste-oil biodiesel utilization, Renew. Sustain. Energy Rev. 72
thesizing a given quantity of solketalacetin and solketal. Further- (2017) 445e464.
[4] M. Aghbashlo, S. Hosseinpour, M. Tabatabaei, A. Dadak, Fuzzy modeling and
more, the FEE of the selected condition was high enough compared optimization of the synthesis of biodiesel from waste cooking oil (WCO) by a
with the other proposed conditions, leading to efficient synthesis of low power, high frequency piezo-ultrasonic reactor, Energy 132 (2017)
solketalacetin and solketal. Notably, for an ideal reversible ther- 65e78.
[5] A. Presciutti, F. Asdrubali, G. Baldinelli, A. Rotili, M. Malavasi, G. Di Salvia,
modynamic process, the best NED is zero while the best FEE is Energy and Exergy analysis of glycerol combustion in an innovative flameless
100%. The optimal operating conditions proposed by the RSM power plant, J. Clean. Prod. 172 (2017) 3817e3824.
approach were experimentally investigated in order to confirm the [6] J.F. Izquierdo, M. Montiel, I. Pale s, P.R. Outo n, M. Galan, L. Jutglar,
M. Villarrubia, M. Izquierdo, M.P. Hermo, X. Ariza, Fuel additives from glycerol
data obtained through the software solution. The experimentally etherification with light olefins: state of the art, Renew. Sustain. Energy Rev.
obtained NED and FEE values were very close to the predicted 16 (2012) 6717e6724.
values with a deviation lower than 2%. [7] H. Rastegari, H.S. Ghaziaskar, M. Yalpani, A. Shafiei, Development of a
continuous system based on azeotropic reactive distillation to enhance tri-
As conclusion, the outcomes of this study ascertained the use- acetin selectivity in glycerol esterification with acetic acid, Energy Fuels 31
fulness of exergy analysis in exploring the optimum operating (2017) 8256e8262.
conditions of chemical reactors applied to valorize glycerol and its [8] P.H.R. Silva, V.L.C. Gonçalves, C.J.A. Mota, Glycerol acetals as anti-freezing
additives for biodiesel, Bioresour. Technol. 101 (2010) 6225e6229.
derivatives with respect to the efficiency, productivity, sustain-
[9] S.S. Priya, P.R. Selvakannan, K.V.R. Chary, M.L. Kantam, S.K. Bhargava, Solvent-
ability issues. Moreover, elaborated extensions of exergy concept free microwave-assisted synthesis of solketal from glycerol using transition
like exergoeconomic and exergoenvironmental approaches should metal ions promoted mordenite solid acid catalysts, Mol. Catal. 434 (2017)
be applied in future investigations in order to make decisions on 184e193.
rez, A. Garrido, J. Peinado, New class of acetal derived
[10] E. Garcia, M. Laca, E. Pe
operating conditions of the developed reactor from monetary and from glycerin as a biodiesel fuel component, Energy Fuels 22 (2008)
environmental viewpoints. The implementation of the findings of 4274e4280.
 M. Aghbashlo et al. / Renewable Energy 126 (2018) 242e253 253

[11] Y.M. Gorji, H.S. Ghaziaskar, Optimization of solketalacetin synthesis as a Green [21] M. Aghbashlo, M. Tabatabaei, P. Mohammadi, B. Khoshnevisan, M.A. Rajaeifar,
fuel additive from ketalization of monoacetin with acetone, Ind. Eng. Chem. M. Pakzad, Neat diesel beats waste-oriented biodiesel from the exer-
Res. 55 (2016) 6904e6910. goeconomic and exergoenvironmental point of views, Energy Convers.
[12] L.J. Konwar, P. M€ aki-Arvela, P. Begum, N. Kumar, A.J. Thakur, J.-P. Mikkola, Manag. 148 (2017) 1e15.
R.C. Deka, D. Deka, Shape selectivity and acidity effects in glycerol acetylation [22] F.J.G. Ortiz, P. Ollero, A. Serrera, S. Galera, Process integration and exergy
with acetic anhydride: selective synthesis of triacetin over Y-zeolite and analysis of the autothermal reforming of glycerol using supercritical water,
sulfonated mesoporous carbons, J. Catal. 329 (2015) 237e247. Energy 42 (2012) 192e203.
[13] S.O. Oyedepo, R.O. Fagbenle, S.S. Adefila, M. Alam, Thermoeconomic and [23] F.J.G. Ortiz, P. Ollero, A. Serrera, S. Galera, Optimization of power and
thermoenvironomic modeling and analysis of selected gas turbine power hydrogen production from glycerol by supercritical water reforming, Chem.
plants in Nigeria, Energy Sci. Eng. 3 (2015) 423e442. Eng. J. 218 (2013) 309e318.
[14] M.R. Barati, M. Aghbashlo, H. Ghanavati, M. Tabatabaei, M. Sharifi, [24] N. Hajjaji, A. Chahbani, Z. Khila, M.-N. Pons, A comprehensive energyeexergy-
G. Javadirad, A. Dadak, M. Mojarab Soufiyan, Comprehensive exergy analysis based assessment and parametric study of a hydrogen production process
of a gas engine-equipped anaerobic digestion plant producing electricity and using steam glycerol reforming, Energy 64 (2014) 473e483.
biofertilizer from organic fraction of municipal solid waste, Energy Convers. [25] N. Hajjaji, I. Baccar, M.-N. Pons, Energy and exergy analysis as tools for opti-
Manag. 151 (2017) 753e763. mization of hydrogen production by glycerol autothermal reforming, Renew.
[15] S.O. Oyedepo, R.O. Fagbenle, S.S. Adefila, M.M. Alam, Exergy costing analysis Energy 71 (2014) 368e380.
and performance evaluation of selected gas turbine power plants, Cogent Eng. [26] R.A.M. Boloy, S.J.R. Ferr ^a, J.A.P. Angulo, R. de
an, D.de.C.L. e Penalva, C. Corre
2 (2015) 1101048. Castro Pereira Filho, Exergetic evaluation of incorporation of hydrogen pro-
[16] M.A. Mandegari, S. Farzad, H. Pahlavanzadeh, Exergy performance analysis duction in a biodiesel plant, Int. J. Hydrogen Energy 40 (2015) 8797e8805.
and optimization of a desiccant wheel system, J. Therm. Sci. Eng. Appl. 7 [27] P. Tippawan, T. Thammasit, S. Assabumrungrat, A. Arpornwichanop, Using
(2015) 31013. glycerol for hydrogen production via sorption-enhanced chemical looping
[17] M. Aghbashlo, M. Tabatabaei, S.S. Hosseini, B.B. Dashti, M. Mojarab Soufiyan, reforming: thermodynamic analysis, Energy Convers. Manag. 124 (2016)
Performance assessment of a wind power plant using standard exergy and 325e332.
extended exergy accounting (EEA) approaches, J. Clean. Prod. 171 (2018) [28] M. Aghbashlo, M. Tabatabaei, S. Hosseinpour, Z. Khounani, S.S. Hosseini,
127e136. Exergy-based sustainability analysis of a low power, high frequency piezo-
[18] M. Aghbashlo, M.A. Rosen, Consolidating exergoeconomic and exergoenvir- based ultrasound reactor for rapid biodiesel production, Energy Convers.
onmental analyses using the emergy concept for better understanding energy Manag. 148 (2017) 759e769.
conversion systems, J. Clean. Prod. 172 (2018) 696e708. [29] I. Dincer, M.A. Rosen, Exergy: Energy, Environment and Sustainable Devel-
[19] A. Bejan, G. Tsatsaronis, M.J. Moran, Thermal Design and Optimization, John opment, Newnes, 2012.
Wiley & Sons, New Jersey, USA, 1996. [30] G. Song, J. Xiao, H. Zhao, L. Shen, A unified correlation for estimating specific
[20] M. Aghbashlo, S. Hosseinpour, M. Tabatabaei, S.S. Hosseini, G. Najafpour, chemical exergy of solid and liquid fuels, Energy 40 (2012) 164e173.
H. Younesi, An exergetically-sustainable operational condition of a photo- [31] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Industrial Catalytic Pro-
biohydrogen production system optimized using conventional and innova- cesses, John Wiley & Sons, 2011.
tive fuzzy techniques, Renew. Energy 94 (2016) 605e618.
 Fuel 233 (2018) 377–387

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

One-pot synthesis of ZrMo-KIT-6 solid acid catalyst for solvent-free T

conversion of glycerol to solketal
⁎ ⁎
Zhenbin Li, Zhichao Miao , Xuan Wang, Jinping Zhao, Jin Zhou, Weijiang Si, Shuping Zhuo
School of Chemistry and Chemical Engineering, Shandong University of Technology, Zibo 255049, PR China

G R A P H I C A L A B S T R A C T

A R T I C LE I N FO A B S T R A C T

Keywords: A series of zirconium and molybdenum species incorporated into ordered mesoporous silicate KIT-6 (ZrMo-KIT-
KIT-6 6) materials were designed and synthesized from a one-pot hydrothermal method. The influences of Zr and Mo
Zirconium contents and calcination temperature in ordered mesoporous structure and existing states of introduced Zr and
Molybdenum Mo species were systematically researched by SXRD, N2-physisorption, TEM, elemental mapping, TG-DSC,
Solid acid
WXRD, Raman, UV and XPS techniques. The results indicated that Zr and Mo species were incorporated into the
Acetalization
Glycerol
skeleton of material as designed and existed as a highly dispersed state when the content was below 7%. Also,
the ZrMo-KIT-6 material had ordered mesostructure with excellent textural properties, and the ordered meso-
porous pores and highly dispersed Zr and Mo species could be preserved even treated at 700 °C. The acidic
properties of ZrMo-KIT-6 materials were tested by NH3-TPD and FT-IR spectra of adsorbed pyridine techniques.
Owing to the excellent acidic property, the ZrMo-KIT-6 material was employed as a solid acid catalyst for
acetalization of glycerol with acetone. The ZrMo-KIT-6(5–700) material showed optimal catalytic performance
(conversion of glycerol was 85.8% and selectivity of solketal was 97.8%), and there was no obvious decline in
catalytic performance even after five cycles.

1. Introduction source, exhibits enormous advantages compared with the traditional

diesel fuel based on petroleum. Principally, biodiesel is nontoxic and
These years, with the rapid development of global economy, the renewable, and it also can reduce the emission of CO2 [1,2]. Besides,
non-renewable petroleum-based fossil fuels have been largely con- the emission of pollutants, such as sulfur, CO, poly aromatics, hydro-
sumed, and the problems about energy and environment have become carbons, particulate matter and smoke can be largely decreased.
the research focus. Biodiesel, as an attractive new and clean energy Therefore, the substitution of biofuel has been encouraged in many

⁎
Corresponding authors.
E-mail addresses: [email protected] (Z. Miao), [email protected] (S. Zhuo).

https://doi.org/10.1016/j.fuel.2018.06.081
Received 22 January 2018; Received in revised form 18 June 2018; Accepted 19 June 2018
Available online 22 June 2018
0016-2361/ © 2018 Elsevier Ltd. All rights reserved.
Z. Li et al. Fuel 233 (2018) 377–387

Scheme 1. Acetalization reaction of glycerol with acetone.

countries [3–6]. Ordered mesoporous Si-based materials have drawn many atten-
Typically, biodiesel is obtained from the transesterification of ve- tions in the domain of catalyst, owing to their superior textural prop-
getable oils or fats with methanol. In this reaction, approximately 10% erties and thermal stability [34,35]. For examples, the mesoporous si-
of glycerol is gained as a by-product [7]. Therefore, how to efficiently lica-ceria-zirconia composite was employed as a support for the
use the glycerol is valuable to be investigated [8–10]. Several reactions methane dry reforming of carbon dioxide at high temperature (700 °C).
for glycerol conversion into value-added products have been reported, The Si-based solid acid catalysts, such as Zr-KIT-5, SO42−/Zr-KIT-6 and
such as oxidation of glycerol to dihydroxyacetone and hydroxypyruvic WOx/SBA-15 have been reported and used in many reactions, such as
acid [11,12], dehydration of glycerol to acrolein [13,14], acetylation of Friedel-Crafts alkylation, green diesel production from esterification of
glycerol to triacetin [15,16], esterification of glycerol to glycerol car- oleic acid and cellobiose hydrolysis reaction [36–38]. The abundant
bonate [17,18] and acetalization of glycerol to cyclic products [19–21]. pore structure could provide plentiful accessible active sites for the
In these reactions, the acetalization of glycerol with acetone to solketal molecules.
(five-membered cyclic product (2,2-dimethyl-1,3-dioxolan-4-yl)-me- In this paper, we design to combine the MoOx-ZrO2 active sites and
thanol) and acetal (six-membered cyclic product 2,2-dimethyl-1,3-di- KIT-6 materials, and get an excellent solid acid catalyst for acetalization
oxan-5-ol) (as shown in Scheme 1) has attracted many attentions for of glycerol with acetone. The Zr and Mo elements are directly in-
petrochemical, pharmaceutical and polymer industry because the pro- corporated into the KIT-6 framework through a one-pot method. In
ducts can be used as fuel additive, solvent and plasticizer [22–26]. addition, the influence of Zr and Mo contents and calcination tem-
For the acetalization of glycerol and acetone, typical homogeneous perature in mesoporous structure are investigated systematically by
acids such as HCl, H2SO4 and H3PO4 have been generally used. small-angle X-ray diffraction (SXRD), N2-physisorption, transmission
However, from the viewpoint of environmental problems, solid acid electron microscopy (TEM) and thermogravimetric-differential scan-
catalysts, which are easy separation, little corrosion and en- ning calorimetry (TG-DSC) characterizations, and the existing states of
vironmentally friendly should be encouraged. Different kinds of solid introduced Zr and Mo species are carefully researched by wide-angle X-
acid catalysts, such as heteropolyacids, zeolites, modified carbon ma- ray diffraction (WXRD), elemental mapping, Raman, UV and X-ray
terials and complex metal oxides are explored in the acetalization of photoelectron spectroscopy (XPS) techniques. The acidic properties of
glycerol. For instance, Chen et al. employed Cs2.5H0.5PW12O40 hetero- obtained materials are studied by the temperature programmed deso-
polyacids supported on mesoporous silica for acetalization under mild rption of ammonia (NH3-TPD) and FT-IR spectra of adsorbed pyridine
conditions [27]. Venkatesha et al. used dealuminated BEA zeolite for methods. Finally, the gotten ZrMo-KIT-6 materials are taken as a solid
selective synthesis of five-membered cyclic acetal from glycerol under acid catalyst in acetalization of glycerol with acetone, and the detailed
ambient conditions [28]. Gonçalves et al. synthesized acidic carbon- results and discussion are given as follows.
based catalysts from biodiesel waste, and about 80% of glycerol con-
version and 95% of solketal selectivity were reached [29]. Titanate 2. Experimental section
nanotubes and Mo and W-promoted SnO2 solid acids were investigated
as heterogeneous catalysts for the transformation of glycerol to value- 2.1. Synthesis of ZrMo-KIT-6
added products in an eco-friendly manner [30,31].
In the field of solid acid catalyst, WOx-ZrO2 and MoOx-ZrO2 mate- ZrMo-KIT-6 materials with different ZrMo/Si ratios were designed
rials have been widely studied, due to their excellent catalytic property and synthesized following the typical synthesis procedure described for
and stability. However, the formation of acid site in these two materials KIT-6 material [39,40]. In a general synthesis, 2.0 g of triblock copo-
usually needs a high temperature calcination process, which might lymer Pluronic P123 ((EO)20(PO)70(EO)20, Aldrich) was dissolved in
largely decrease the surface area of catalyst and number of active sites 70 mL of 0.5 M HCl solution at 35 °C. Following the complete dissolu-
[32,33]. Therefore, it is important to assure the amount of active sites tion of P123, 2.0 g of n-butanol was introduced and stirred for another
even treated at high temperature. 1 h. Finally, 8.4 g of tetraethyl orthosilicate (TEOS, Sinopharm

378
Z. Li et al. Fuel 233 (2018) 377–387

Chemical Reagent Co. Ltd.) and required amounts of zirconyl chloride 5 × 10−3 Pa, followed by adsorption of pyridine at room temperature
octahydrate (ZrOCl2·8H2O, Sinopharm Chemical Reagent Co. Ltd.), for 30 min. The sample was recorded at targeted temperature after
ammonium molybdate ((NH4)6Mo7O24·4H2O, Sinopharm Chemical desorption for 30 min under 5 × 10−3 Pa. The blank control experi-
Reagent Co. Ltd.) were successively added into above solution, and the ments were performed as conditions aforementioned above and em-
molar ratio of Zr and Mo was kept the same. Then, the mixture was ployed as background.
stirred at 35 °C for 18 h and followed by hydrothermal treated at 100 °C
for 48 h in a Teflon autoclave under static conditions. Finally, the solid 2.3. Acetalization of glycerol with acetone
product was washed, dried and calcined in air at 700 °C for 4 h. The
gained material was designated as ZrMo-KIT-6(X), and the X (X = 1, 3, A typical acetalization reaction of glycerol with acetone (as shown
5, 7, 9%) was the ratio of Zr or Mo/Si. The ZrMo-KIT-6 material treated in Scheme 1) was carried out in a round bottom flask fitted with a reflux
at different temperatures was designated as ZrMo-KIT-6(5–Y) (Y = 500, condenser and placed in an oil bath. Firstly, ZrMo-KIT-6 solid acid
600, 800 and 900 °C). catalyst (0.05 g); glycerol, 1 g; acetone, 5 g (the molar ratio of glycerol
to acetone was 1:8) were added into a reactor. Then, the mixture was
stirred at 50 °C for 4 h. At the end of reaction, 5 mL of ethanol was
2.2. Characterization introduced to form a homogeneous phase for all reactants and products.
Also, 0.4 mL of n-heptane was added into the mixture as an internal
Powder XRD was carried out on X’Pert Pro Multipurpose dif- standard substance. The mixture was separated by centrifugalization,
fractometer (PANalytical, Inc.) using Cu Kα radiation (0.15406 nm) and the liquid phase products were analyzed by a GC (Agilent-7890B)
source in 2θ range from 0.6 to 5.0° (small-angle) and 10.0 to 80.0° equipped with a PEG-20 M capillary column
(wide-angle). The N2 adsorption-desorption isotherms were obtained on (50 m × 0.25 mm × 0.5 μm) and a FID detector. In addition, the re-
ASAP 2020 (Micromeritics Instrument) static volumetric analyzer. cyclability of ZrMo-KIT-6 catalyst was tested and reused for five cycles.
Before measurements, samples were treated at 200 °C for 2 h under After each cycle, ZrMo-KIT-6 was recovered from the reaction solution
vacuum. The specific surface area was calculated via Brunauer-Emmett- and calcined at 500 °C for 2 h to remove the organic species adsorbed on
Teller (BET) method; the pore volume was estimated from the amount the catalyst. In the GC-chromatogram, the five- and six-membered
of nitrogen adsorbed at P/P0 of 0.990; pore size distribution (PSD) was cyclic products solketal and acetal were detected and no other products
obtained by analyzing the adsorption branches of isotherms using were observed in this reaction. Furthermore, the peak associated with
Barrett-Joyner-Halenda (BJH) method. High-resolution transmission glycerol exhibited low sensitivity and distinct tailing causing less ac-
electron microscopy (FEI TECNAI G2 F20) with accelerating voltage of curate quantification. Therefore, the conversion of glycerol and se-
200 kV was taken for TEM images, selected area electron diffraction lectivity of solketal were calculated according to the following equa-
(SAED) and elemental mapping tests. TG-DSC was taken on NETZSCH tions:
STA 449C analyzer with 10 °C·min−1 to 1000 °C under air atmosphere.
Raman spectra were tested at ambient condition on LabRam HR mole of solketal detected
Yield of solketal (mol%) = × 100 mol%
system equipped with a CCD detector and laser beam (λ = 532 nm) for mole of solketal theoretical
excitation. Diffuse reflectance UV–visible (UV–vis) spectra were col- Yield of solketal
lected in the range of 200–800 nm on PE Lambda 650S. The XPS Selectivity of solketal (mol%) =
Yield of solketal + Yiled of acetal
measurements were conducted on a Thermon Scientific ESCALAB250xi
× 100 mol%
spectrometer and calibrated to C1s line (284.8 eV). X-ray fluorescence
(XRF) spectrum analyzer was performed on Magix PW2403 Yield of solketal
(PANalytical, Inc.). Conversion of glycerol (mol%) = × 100 mol%
Selectivity of solketal
NH3-TPD was carried out with Finesorb 3010 (Finetec Instruments).
Samples were heated from room temperature to 500 °C to remove ad-
sorbed water and subsequently cooled down to 100 °C in flowing He 3. Results and discussion
gas. Ammonia (10 mol% NH3-He gas (50 mL·min−1)) was adsorbed at
100 °C for 30 min. Physically absorbed ammonia was then removed by 3.1. Ordered mesoporous structure
desorbing in He gas at 100 °C. Following this step, the temperature was
raised to 600 °C with a ramp of 10 °C/min and the desorbed ammonia Ordered mesoporous structure of ZrMo-KIT-6 materials with dif-
was recorded by TCD detector. FT-IR spectra of adsorbed pyridine were ferent ZrMo/Si ratios and calcination temperatures is successively re-
recorded by PE Frontier FT-IR spectrometer. In each measurement, a searched by SXRD, N2-physisorption and TEM characterizations. As
self-supporting wafer was first evacuated in situ 400 °C for 1 h under displayed in Fig. 1(1), the SXRD patterns of ZrMo-KIT-6 materials show

Fig. 1. Small-angle X-ray diffraction patterns of (1) ZrMo-KIT-6(X) with different ZrMo contents and (2) ZrMo-KIT-6(5) treated at different temperatures.

379
Z. Li et al. Fuel 233 (2018) 377–387

a well-resolved intense peak around 0.9° assigning to a cubic Ia3d indicates the existence of mesoporous pore with uniform size. This also
symmetry with d211 spacing, and another two peaks at 1.0 and 1.5°, can be confirmed by the corresponding PSD pattern (Fig. 2(2)), which
corresponding to (2 2 0) and (4 2 0) reflections, showing that these exhibits a narrow distribution of pore size. With the Zr and Mo contents
materials possess well-ordered pore arrangement [41,42]. Moreover, increasing to 9%, the H1-type hysteresis loop deforms, and the PSD
the intensity of peak varies little from ZrMo-KIT-6(0) to ZrMo-KIT-6(5), becomes broader, indicating the deformation of ordered mesoporous
suggesting that the introduced Zr and Mo species have little detrimental structure. The detailed textural parameters of obtained materials are
effect on the framework of KIT-6. Meanwhile, the peak position of these given in Table 1. The specific surface area and pore volume gradually
samples changes to the lower position (from 1.00 to 0.91°), implying decrease with the increasing of Zr and Mo species. Moreover, the pore
the increase of unit parameter (from 21.60 to 24.03 nm), and further size increases from 5.90 to 8.80 nm with the introduction of Zr and Mo
indicating that the Zr and Mo species have been successfully introduced species. Also, the wall thickness of ZrMo-KIT-6 material is calculated
into the framework of KIT-6. and given in Table 1. With the Zr and Mo contents increased from 1 to
The SXRD patterns of ZrMo-KIT-6 materials with different calcina- 7%, the wall thickness increases from 2.83 to 3.68 nm, further de-
tion temperatures are provided in Fig. 1(2). Apparent diffraction peaks monstrating that the Zr and Mo species have been introduced into the
can be clearly observed for the samples under 800 °C, showing the ex- framework of ZrMo-KIT-6 materials.
cellent thermal stability of ZrMo-KIT-6(5) material. Further enhancing The N2 adsorption-desorption isotherms and corresponding PSD
the calcination temperature to 900 °C, no diffraction peaks are found, patterns of ZrMo-KIT-6(5-Y) materials are given in Fig. 2(3, 4). Ordered
demonstrating the disappearance of ordered mesoporous structure. mesoporous structure with narrow PSD can be found even treated at
Besides, the position of diffraction peak shifts from 0.84 to 0.97° with 800 °C, implying the excellent thermal stability of ZrMo-KIT-6(5). When
calcination temperature increasing from 500 to 800 °C, proving the the calcination temperature reaches 900 °C, no H1-type hysteresis loop
decrease of unit parameter at the higher temperature. or narrow PSD is found, showing the disappearance of mesoporous
The pore nature of ZrMo-KIT-6 materials with different Zr and Mo structure in ZrMo-KIT-6(5–900). In addition, as shown in Table 1, the
contents and calcination temperatures is evaluated by nitrogen physi- textural properties gradually decrease with the increasing of calcination
sorption measurement. The N2 adsorption-desorption isotherms of temperature. Moreover, it is interesting to be found that the wall
ZrMo-KIT-6(X) materials (Fig. 2(1)) display type IV isotherm with an thickness increases from 2.85 to 3.75 nm with calcination temperature
H1-type hysteresis loop in the P/P0 range of 0.6–0.9, a typical indica- from 500 to 800 °C, and this might be attributed to the further crys-
tion of ordered mesoporous structure. Meanwhile, the position of hys- tallization of Zr and Mo species in the skeleton of ZrMo-KIT-6 materials.
teresis loop shifts to higher P/P0 region, showing that the Zr and Mo The TEM characterization (TEM, HRTEM images and elemental
species have been successfully introduced into the skeleton and en- mapping) is taken, and ZrMo-KIT-6(5), ZrMo-KIT-6(7), ZrMo-KIT-6(9),
hance the pore size of ZrMo-KIT-6(X) materials [43]. The well-defined ZrMo-KIT-6(5–900) materials are chosen as representative. As provided
and steep hysteresis loop with parallel adsorption-desorption branches in Fig. 3(a, b), ordered mesoporous pores with diameter about 10 nm

Fig. 2. Isotherms and pore size distributions of (1, 2) ZrMo-KIT-6(X) with different ZrMo contents and (3, 4) ZrMo-KIT-6(5) treated at different temperatures.

380
Z. Li et al. Fuel 233 (2018) 377–387

Table 1
Textural, acidic properties and TOF values of the ZrMo-KIT-6 samples derived from SXRD, N2 adsorption and desorption data, NH3-TPD and reactions.
Samples d211 (nm)a ao (nm)b Specific surface area (m2·g−1) Pore size (nm) Pore volume (cm3·g−1) Wall thickness (nm)c Acidity (mmol·g−1)d TOF(h−1)e

ZrMo-KIT-6(0) 8.82 21.60 697 5.90 1.10 4.90 0.08 0

ZrMo-KIT-6(1) 9.49 23.25 666 8.80 1.32 2.83 0.14 163
ZrMo-KIT-6(3) 9.70 23.76 528 8.20 1.15 3.68 0.38 159
ZrMo-KIT-6(5) 9.81 24.03 389 8.55 0.88 3.46 0.64 147
ZrMo-KIT-6(7) 10.15 24.86 355 8.75 0.81 3.68 0.48 90
ZrMo-KIT-6(9) 9.92 24.30 218 14.30 0.81 – 0.25 56
ZrMo-KIT-6(5–500) 10.51 25.74 625 10.02 1.23 2.85 0.53 103
ZrMo-KIT-6(5–600) 10.26 25.14 509 9.99 1.04 2.58 0.60 115
ZrMo-KIT-6(5–800) 9.10 22.29 298 7.40 0.62 3.75 0.36 29
ZrMo-KIT-6(5–900) – – 40 32.3 0.29 – 0.02 3

a
Acquired from XRD patterns using Bragg’s law.
b
ao = d211/(h2 + k2 + l2)1/2.
c
Wall thickness (nm) = ao/2-Pore size (nm).
d
The acidity gotten from NH3-TPD measurements.
e
The TOF value was defined as number of glycerol converted per Zr and Mo atom per hour at reaction time of 1 h.

are clearly observed, and no particles are found in ZrMo-KIT-6(5). 3.2. Existing states of Zr and Mo species
Further enhancing the Zr and Mo contents to 7% (Fig. 3(d, e)), ordered
mesoporous structure still exists. However, obvious particles appear in The existing states of introduced Zr and Mo species in the materials
ZrMo-KIT-6(7). For the ZrMo-KIT-6(9) materials (Fig. 3(g)), ordered are studied by WXRD, Raman, UV–vis and XPS characterizations. The
mesoporous pores transform to worm-like pores. The TEM images of WXRD patterns of ZrMo-KIT-6(X) samples are obtained from 10 to 80°
ZrMo-KIT-6(5) material treated at 900 °C are given in Fig. 3(h), and no and depicted in Fig. 6(1). For the samples with Zr and Mo contents
distinct pore structure is observed, implying the disappearance of me- below 5%, the XRD patterns show amorphous feature and no peaks of
soporous structure after treated at 900 °C. In consequence, the TEM crystalline phases [44]. With the increase of Zr and Mo species to 5 and
images keep consistent with the characterizations of SXRD and N2- 7%, weak peaks at about 30, 35, 50 and 60°, which are attributed to
physisorption. (1 0 1), (1 1 0), (1 1 2) and (2 1 1) of t-ZrO2 (JCPDF card No. 80-0784),
The calcination process of as-synthesized ZrMo-KIT-6(5) material is are observed, indicating the appearance of crystalline t-ZrO2. However,
investigated by the TG-DSC characterization. As shown in Fig. 5(1), the these peaks are quite weak, implying that the crystalline ZrO2 particles
main weight loss stages with apparent exothermic process, which are are small, and the introduced Zr and Mo species exhibit as a highly
ascribed to the removal of P123, happen at 200–500 °C, and there is no dispersed state. However, further increasing the Zr and Mo contents to
weight loss after 500 °C, implying the removal of P123 after treated at 7 and 9%, distinct diffraction peaks, which are the characteristic of
500 °C. ZrO2 and MoO3 (JCPDF card No. 05-0508), are detected, meaning that
highly dispersed Zr and Mo species aggregate to large crystalline par-
ticles. This also could be verified by the HRTEM images. As shown in
Fig. 3(c), there exists no crystalline structure in the ZrMo-KIT-6(5)

Fig. 3. TEM and HRTEM images of (a–c) ZrMo-KIT-6(5), (d–f) ZrMo-KIT-6(7), (g) ZrMo-KIT-6(9), (h, i) ZrMo-KIT-6(5–900).

381
Z. Li et al. Fuel 233 (2018) 377–387

Fig. 4. Elemental mapping of ZrMo-KIT-6(5).

Fig. 5. TG-DSC curves of (1) the as-synthesized ZrMo-KIT-6(5) and (2) ZrMo-KIT-6(5) treated at 500 °C.

material. However, further increasing the Zr and Mo content to 7%, highly dispersed Zr and Mo species. This conclusion also can be de-
particles with d-spacing 0.293 nm (Fig. 3(f)) which corresponding to monstrated by the HRTEM, elemental mapping and TG-DSC techniques.
the (1 0 1) lattice plane of t-ZrO2 are observed, demonstrating the ap- Fig. 3(i) shows that the interplanar spacing is 0.296 nm, which agrees
pearance of crystalline structure. Meanwhile, the elemental mappings well with the (1 0 1) crystal face of t-ZrO2. The elemental mapping (Fig.
are taken to research the dispersion of introduced Zr and Mo species. As S3) exhibits obvious aggregation of Zr or Mo species in ZrMo-KIT-
displayed in Fig. 4, it is found that the Zr and Mo species are success- 6(5–900). In addition, in the TG-DSC patterns (Fig. 5(2)), a broad
fully introduced and highly dispersed in the skeleton of ZrMo-KIT-6(5) exothermic peak which caused by the crystallization process is found at
material. However, as shown in the elemental mappings of ZrMo-KIT- 600–1000 °C, especially at 850 °C. We deduce that it is precisely due to
6(7) (Fig. S1) and ZrMo-KIT-6(9) (Fig. S2), the aggregation of Zr and the aggregation and crystallization of highly dispersed Zr and Mo spe-
Mo species happens, and obvious particles appear in these two samples. cies, and the ordered mesoporous structure completely vanishes after
The existing states of Zr and Mo species in ZrMo-KIT-6(5) material treated at 900 °C.
with different calcination temperatures are also explored. As given in Because of the high sensitivity of Raman spectra, it is employed to
Fig. 6(2), there is no diffraction peak in WXRD patterns with calcination research the existing states of Zr and Mo species in these materials
temperature under 800 °C, illustrating that the Zr and Mo species [33,45,46]. As given in Fig. 7(1), the ZrMo-KIT-6(0) material shows no
homogeneously disperse in the skeleton of materials. Further enhancing apparent peak, demonstrating the existence of amorphous SiO2 struc-
the temperature to 800 and 900 °C, distinct diffraction peaks, which are ture. With the Zr and Mo contents increasing from 1 to 5%, two broad
assigned to crystalline t-ZrO2 and m-ZrO2 (JCPDF card No. 83-0944) peaks at 880 and 950 cm−1 appear, and these peaks are ascribed to the
and MoO3, can be found. This might be attributed to the aggregation of Mo–O–Mo stretching mode and the terminal Mo]O stretching

382
Z. Li et al. Fuel 233 (2018) 377–387

Fig. 6. Wide-angle X-ray diffraction patterns of (1) ZrMo-KIT-6(X) with different ZrMo contents and (2) ZrMo-KIT-6(5) treated at different temperatures.

vibration in the highly dispersed Mo species [47]. Further increasing stronger when the calcination temperature reaches 600–800 °C, in-
the Zr and Mo contents to 7 and 9%, these two peaks weaken and new dicating the further crystallization of Zr and Mo species. However, the
peaks which corresponding to crystalline MoO3 appear at 820 and peak becomes weaker for ZrMo-KIT-6(5–900) sample, implying the
990 cm−1 [48,49]. In addition, a new group of peaks at 160, 280, 340 aggregation and formation of large crystalline ZrO2 and MoO3 particles.
and 650 cm−1 can be found, implying the appearance of crystalline t- To clarify the surface composition and valence states of surface Zr
ZrO2 [50]. The Raman spectra of ZrMo-KIT-6(5) material treated at and Mo species, the XPS spectra of ZrMo-KIT-6(5) are acquired. As
500–900 °C are given in Fig. 7(2). When the calcination temperature is shown in Fig. 9, the signals of Zr 3d, Mo 3d, Si 2p and O 1s are detected,
under 700 °C, highly dispersed Zr and Mo species exist in the materials. manifesting the successful introduction of Zr and Mo species. Specifi-
Further enhancing the calcination temperature to 800 and 900 °C, cally, for the Zr 3d signal (Fig. 9(1)), two obvious peaks are observed at
crystalline ZrO2 (t-ZrO2 and m-ZrO2) and MoO3 are detected [49,50]. the binding energy of 184.7 (Zr 3d3/2) and 182.4 eV (Zr 3d5/2), which
These conclusions gotten from Raman spectra keep consistent with the are assigned to the Zr4+. In addition, the Mo 3d (Fig. 9(2)) spectrum
WXRD and TEM characterizations. exhibits two contributions, and 3d3/2 and 3d5/2 located at 235.5 and
As a sensitive probe for investigating transition metal in the silica 232.6 eV are ascribed to the Mo6+ [54]. The Si 2p signal (Fig. 9(3)) has
matrix, the UV–Vis spectra of ZrMo-KIT-6 materials with different Zr a peak at 103.5 eV, showing the existence of Si4+ species. For the O 1s
and Mo contents are taken and given in Fig. 8(1) [51,52]. For the ZrMo- signal (Fig. 9(4)), a major peak at 533.2 eV is found, and this is due to
KIT-6(0) material, only a weak peak at 260 nm is detected, and this is the Si-O bond in the material. Furthermore, the weak peak at 530.0 eV
ascribed to the Si-O bond. With the introduction of Zr and Mo species, a is assigned to the Zr–O and Mo–O bond [55]. All these results show the
broad peak at 280–330 nm, which is assigned to the ligand to metal existence of ZrO2, MoO3 and SiO2 species in the skeleton of ZrMo-KIT-
charge transfer from O2− to Mo6+ and O2− to Zr4+ [39,53], appears 6(5).
and becomes stronger with Zr and Mo contents increasing from 1 to 7%. The elemental compositions of mesoporous skeleton are researched
This might be owing to that the Mo species exist as highly dispersed by XRF technique. As shown in Table 2, the Zr and Mo contents in these
monomeric or oligomeric states, and Zr species exist as small nano- materials keep consistent with the theoretical values. The Zr and Mo
particles of ZrO2. This agrees quite well with the WXRD characteriza- species are successfully introduced into the skeleton of ZrMo-KIT-6 and
tion, which only provides weak diffraction peaks of t-ZrO2 for these suffer little loss in the synthesis and calcination process. Moreover, the
samples. Further increasing the Zr and Mo content to 9%, the peak elemental contents in the surface of ZrMo-KIT-6(5) are analyzed by the
becomes weaker, and this might be due to the aggregation of highly XPS spectra. The ratios of Zr/Si and Mo/Si are 6.20% and 4.07%. It is
dispersed Zr and Mo species. Moreover, as displayed in Fig. 8(2), the interesting to observe that the Zr/Si ratio (6.20%) is slightly higher than
ZrMo-KIT-6(5) treated at 500 °C has a weak peak at 230 nm, which is the theoretical value (5%) and value gotten from XRF (4.72%). This
caused by the Zr-O-Zr linkages. The peak at 280–330 nm becomes might be owing to the appearance of small crystalline t-ZrO2 particles,

Fig. 7. Raman spectra of (1) ZrMo-KIT-6(X) with different ZrMo contents and (2) ZrMo-KIT-6(5) treated at different temperatures.

383
Z. Li et al. Fuel 233 (2018) 377–387

Fig. 8. UV–vis spectra of (1) ZrMo-KIT-6(X) with different ZrMo contents and (2) ZrMo-KIT-6(5) treated at different temperatures.

which has been proved by the XRD patterns. and reaches the maximum for ZrMo-KIT-6(5). This indicates that the
acidity (as shown in Table 1) is successfully improved by the introduced
Zr and Mo species. However, further enhancing the Zr and Mo contents
3.3. Acidic properties to 7 and 9%, the acidity begins to decrease, and this might be due to the
aggregation of Zr and Mo species to crystalline ZrO2 and MoO3. The
The acidic properties of ZrMo-KIT-6 materials with different Zr and NH3-TPD patterns of ZrMo-KIT-6(5) treated at 500–900 °C are depicted
Mo contents and calcination temperatures are studied by NH3-TPD in Fig. 10(2). It is found that the intensity of desorption peak slightly
characterization. As provided in Fig. 10(1), the ZrMo-KIT-6(0) material increases from 500 to 700 °C. This might be attributed to that the in-
without Zr and Mo species shows no obvious peak. With the increase of crease of treating temperature is beneficial to the formation of acid
Zr and Mo content from 1 to 5%, the peak caused by the desorbed NH3 sites. However, the peak becomes weaker from 700 to 900 °C, implying
appears at 150–400 °C. Meanwhile, the peak intensity becomes stronger

Fig. 9. High-resolution XPS spectra of ZrMo-KIT-6(5): (1) Zr 3d, (2) Mo 3d, (3) Si 2p and (4) O 1s.

384
Z. Li et al. Fuel 233 (2018) 377–387

Table 2
The Mo/Si ratios in the ZrMo-KIT-6 materials.
Samples Theoretical Zr/Si ratio (%) Zr/Si ratios (%) gotten from XRF Theoretical Mo/Si ratio (%) Mo/Si ratios (%) gotten from XRF

ZrMo-KIT-6(0) 0 0 0 0
ZrMo-KIT-6(1) 1 1.09 1 0.95
ZrMo-KIT-6(3) 3 2.38 3 1.86
ZrMo-KIT-6(5) 5 4.72 5 4.24
ZrMo-KIT-6(7) 7 7.25 7 6.70
ZrMo-KIT-6(9) 9 9.20 9 7.80

that the acidity decreases after treated at 800 and 900 °C.
Moreover, the types of acid sites in ZrMo-KIT-6(5) material are in-
vestigated by the FT-IR spectra of adsorbed pyridine [56]. As shown in
Fig. 11, two bands at 1450 and 1610 cm−1 are observed, meaning the
presence of Lewis acid sites in the material. The bands at 1550 and
1640 cm−1 are found, which caused by the pyridine adsorbed on the
Brønsted acid sites. These imply that both Lewis and Brønsted acid sites
exist in the ZrMo-KIT-6(5) material. Moreover, all these bands still can
be detected even evacuating at 400 °C, showing that the intensity of
acid site is strong.

3.4. Catalytic performances

In consideration of the excellent acidic properties of ZrMo-KIT-6

material, it is employed as a solid acid catalyst for the acetalization of
glycerol with acetone and the possible mechanism is shown in Scheme
1. As the first step, the glycerol and acetone react to form hemiketal.
Afterwards, the hemiketal reacts on the Brønsted acid sites, and the
Fig. 11. IR spectra for adsorbed pyridine of ZrMo-KIT-6(5) evacuated at dif-
water molecule is removed. Then, a tertiary carbonium ion forms and a ferent temperatures: (a) 150 °C, (b) 250 °C, (c) 350 °C, (d) 400 °C.
quick nucleophilic attack of the secondary hydroxyl group occurs to
yield solketal and another nucleophilic attack of the terminal hydroxyl
group occurs to yield acetal [57–60]. In addition, as reported in pre- to 7 and 9%, the TOF values sharply decrease to 90 and 56 h−1. This
vious researches, the solketal is the principal product, and the se- might be ascribed to that the highly dispersed Zr and Mo species ag-
lectivity of solketal usually exceeds 90% [61,62]. gregate and form crystalline ZrO2 and MoO3. Therefore, the amounts of
The influence of Zr and Mo contents in catalytic performance is active sites decrease. Moreover, the catalytic performance of sample
investigated and given in Fig. 12(1). The ZrMo-KIT-6(0) material only with Zr (Zr-KIT-6(5)) or Mo (Mo-KIT-6(5)) species is also studied.
without adding Zr and Mo species shows poor activity. With the in- These two samples exhibit poor catalytic performance, and the yield of
troduction of Zr and Mo species, the catalytic performance is gradually solketal is 6.5% for Zr-KIT-6(5) and 8.3% for Mo-KIT-6(5). Therefore,
improved and ZrMo-KIT-6(5) material exhibits the best performance we deduce that the interaction between Zr and Mo species generates the
among all the samples tested in this work, showing the highest con- acid sites, and catalyzes the acetalization reaction.
version of glycerol (85.8%) and selectivity of solketal (97.8%). How- The catalytic performances of ZrMo-KIT-6(5) material treated at
ever, further increasing the Zr and Mo contents to 7 and 9%, the cat- different temperatures are given in Fig. 12(2). With the calcination
alytic activity begins to decrease. The TOF values of different catalysts temperature increases from 500 to 700 °C, the catalytic performance is
are calculated and given in Table 1. The TOF values of catalysts with Zr gradually improved. Further enhancing the calcination temperature to
and Mo content below 5% show the similar value (163 h−1 for 1%, 800 and 900 °C, catalytic activity sharply decreases. The TOF values
159 h−1 for 3% and 147 h−1 for 5%). This confirms that the highly (Table 1) of ZrMo-KIT-6(5) treated at different calcination temperatures
dispersed Zr and Mo species are the active sites for the acetalization of have the same regularity, in which the catalyst treated at 700 °C shows
glycerol with acetone. However, with the Zr and Mo contents increase the maximal value. All these indicate that the highly dispersed Zr and

Fig. 10. NH3-TPD profiles of (1) ZrMo-KIT-6(X) with different ZrMo contents and (2) ZrMo-KIT-6(5) treated at different temperatures.

385
Z. Li et al. Fuel 233 (2018) 377–387

Fig. 12. Acetalization reaction of glycerol with acetone catalyzed by (1) ZrMo-KIT-6(X) with different ZrMo contents and (2) ZrMo-KIT-6(5) treated at different
temperatures.

Mo species and the increase of calcination temperature are beneficial present study. To research the heterogeneity of ZrMo-KIT-6 solid acid
for enhancing the catalytic performance with the temperature under catalyst, leaching test is carried out by removing the catalyst from the
700 °C. Further increasing the calcination temperature to 800 and reaction medium after 1 h, and the liquid mixture reacts in the absence
900 °C, the highly dispersed Zr and Mo species transform to crystalline of heterogeneous catalyst for another 5 h. It is observed that the reac-
ZrO2 and MoO3 species. Therefore, the catalytic performance sharply tion is interrupted after the removal of catalyst, indicating that there
decreases. are no active sites in the liquid phase. Therefore, the ZrMo-KIT-6 ma-
Moreover, different reaction parameters, such as reaction time, terial is a heterogeneous catalyst.
leaching test, catalyst amount and ratios of acetone to glycerol are in- The effect of catalyst amount on glycerol conversion and solketal
vestigated [25,58]. As displayed in Fig. 13(1), the conversion of gly- selectivity is investigated by varying the amount of catalyst from 0 to
cerol, selectivity of solketal and yield of solketal increase with the re- 0.07 g. As provided in Fig. 13(2), no product is detected without adding
action time and reach the maximum at 4 h. Further prolonging the catalyst and the catalytic performance is gradually improved with the
reaction time to 6 h, there is no significant increase in catalytic activity. increase of catalyst amount from 0.01 to 0.05 g, which is owing to the
Therefore, 4 h is considered to be the optimum reaction time in the increase of active sites in the reaction system. Further increasing the

Fig. 13. Acetalization reaction of glycerol with acetone catalyzed by ZrMo-KIT-6(5): (1) reaction time, (2) catalyst amount, (3) ratio of acetone and glycerol and (4)
reused for five cycles.

386
Z. Li et al. Fuel 233 (2018) 377–387

catalyst amount to 0.07 g, no variation happens. Therefore, 0.05 g of [12] Ning X, Li Y, Yu H, Peng F, Wang H, Yang Y. J Catal 2016;335:95–104.
catalyst is considered to be the optimal amount for the reaction. [13] Katryniok B, Paul S, Bellière-Baca V, Rey P, Dumeignil F. Green Chem
2010;12:2079–98.
In the chemical reaction, the ratio of the reactants plays an im- [14] Galadima A, Muraza O. J Taiwan Inst Chem E 2016;67:29–44.
portant role in catalytic performance. To optimize the reaction by using [15] Sun J, Tong X, Yu L, Wan J. Catal Today 2016;264:115–22.
the precise molar ratio of reactants to achieve better conversion and [16] Okoye PU, Abdullah AZ, Hameed BH. Renewable Sustainable Energy Rev
2017;74:387–401.
selectivity of products, we investigate the appropriate molar ratio of [17] Popova M, Lazarova H, Kalvachev Y, Todorova T, Szegedi Á, Shestakova P, et al.
acetone to glycerol to get the optimal condition. As shown in Fig. 13(3), Catal Commun 2017;100:10–4.
as the increase of ratio from 1:1 to 8:1, the catalytic performance is [18] Okoye PU, Abdullah AZ, Hameed BH. Fuel 2017;209:538–44.
[19] Ruiz VR, Velty A, Santos LL, Leyva-Pérez A, Sabater MJ, Iborra S, et al. J Catal
improved, and the maximum values (conversion of glycerol is 85.8% 2010;271:351–7.
and selectivity of solketal is 97.8%) are gotten. Further enhancing the [20] Sun S, He M, Dai Y, Li X, Liu Z, Yao L. Catalysts 2017;7:184.
ratio to 10:1, the catalytic performance changes little. Similar outcomes [21] Konwar LJ, Samikannu A, Mäki-Arvela P, Boström D, Mikkola J-P. Appl Catal B
2018;220:314–23.
were observed in the literature [29,57].
[22] Gadamsetti S, Rajan NP, Rao GS, Chary KVR. J Mol Catal A Chem 2015;410:49–57.
Under the optimal reaction conditions (reaction time: 4 h, catalyst [23] Rodrigues R, Mandelli D, Gonçalves NS, Pescarmona PP, Carvalho WA. J Mol Catal
amount: 0.05 g, ratio of acetone to glycerol: 8:1), the ZrMo-KIT-6(5) A Chem 2016;422:122–30.
catalyst is reused for five cycles to test the stability of catalyst. As shown [24] Priya SS, Selvakannan PR, Chary KVR, Kantam ML, Bhargava SK. Mol Catal
2017;434:184–93.
in Fig. 13(4), there is no obvious decline in catalytic activity even after [25] Samoilov VO, Ramazanov DN, Nekhaev AI, Maximov AL, Bagdasarov LN. Fuel
five cycles, indicating the excellent reusability of ZrMo-KIT-6. In ad- 2016;172:310–9.
dition, the used catalyst is studied by the TEM characterization (Fig. S4) [26] Samoilov VO, Onishchenko MO, Ramazanov DN, Maximov AL. ChemCatChem
2017;9:2839–49.
and ordered mesoporous and highly dispersed Zr and Mo species still [27] Chen L, Nohair B, Zhao D, Kaliaguine S. ChemCatChem 2018;10:1918–25.
could be observed. Therefore, ZrMo-KIT-6 is a promise solid acid cat- [28] Venkatesha NJ, Bhat YS, Jai Prakash BS. RSC Adv 2016;6:18824–33.
alyst for acetalization of glycerol with acetone. [29] Gonçalves M, Rodrigues R, Galhardo TS, Carvalho WA. Fuel 2016;181:46–54.
[30] de Carvalho DC, Oliveira AC, Ferreira OP, Filho JM, Tehuacanero-Cuapa S, Oliveira
AC. Chem Eng J 2017;313:1454–67.
4. Conclusion [31] Mallesham B, Sudarsanam P, Raju G, Reddy BM. Green Chem 2013;15:478–89.
[32] Li S, Zhou H, Jin C, Feng N, Liu F, Deng F, et al. J Phys Chem C 2014;118:6283–90.
[33] Ciptonugroho W, Al-Shaal MG, Mensah JB, Palkovits R. J Catal 2016;340:17–29.
In this paper, the ZrMo solid acid sites were successfully in- [34] Xiang X, Zhao H, Yang J, Zhao J, Yan L, Song H, et al. Appl Catal A
corporated into the skeleton through a one-pot hydrothermal synthesis. 2016;520:140–50.
The ZrMo-KIT-6 materials exhibited excellent textural properties and [35] Liu Q, Li J, Zhao Z, Gao M, Kong L, Liu J, et al. J Catal 2016;344:38–52.
[36] Gopinath S, Kumar PSM, Arafath KAY, Thiruvengadaravi KV, Sivanesan S,
thermal stability. The relationship between the existing states of Zr and
Baskaralingam P. Fuel 2017;203:488–500.
Mo species and acidic properties was detailedly investigated, and the [37] Wang H, Guo Y, Chang C, Zhu X, Liu X, Han J, et al. Appl Catal A 2016;523:182–92.
result showed that the highly dispersed Zr and Mo species were in favor [38] Ramanathan A, Zhu H, Maheswari R, Subramaniam B. Chem Eng J
of improving the acidic properties of ZrMo-KIT-6. Meanwhile, both 2015;278:113–21.
[39] Ramanathan A, Subramaniam B, Maheswari R, Hanefeld U. Microporous
Brønsted and Lewis acid sites existed in the ZrMo-KIT-6 material. The Mesoporous Mater 2013;167:207–12.
best catalytic performance in conversion of glycerol to solketal was [40] Pan Q, Ramanathan A, Snavely WK, Chaudhari RV, Subramaniam B. Ind Eng Chem
achieved for ZrMo-KIT-6(5–700), whose conversion of glycerol is Res 2013;52:15481–7.
[41] Li B, Luo X, Huang J, Wang X, Liang Z. Chin J Catal 2017;38:518–28.
85.8% and selectivity of solketal is 97.8%. In addition, there existed no [42] Pirez C, Caderon J-M, Dacquin J-P, Lee AF, Wilson K. ACS Catal 2012;2:1607–14.
noticeable decline in the catalytic performance even after five cycles. [43] Cai W, Yu J, Anand C, Vinu A, Jaroniec M. Chem Mater 2011;23:1147–57.
Therefore, the ZrMo-KIT-6 material was a promise solid acid catalyst for [44] Xu L, Wang F, Chen M, Zhang J, Yuan K, Wang L, et al. ChemCatChem
2016;8:2536–48.
solvent-free acetalization of glycerol to solketal. [45] Sarkar A, Pramanik S, Achariya A, Pramanik P. Microporous Mesoporous Mater
2008;115:426–31.
Acknowledgements [46] Dou J, Zeng HC. J Am Chem Soc 2012;134:16,235–46.
[47] Shen K, Liu X, Lu G, Miao Y, Guo Y, Wang Y, et al. J Mol Catal A Chem
2013;373:78–84.
The authors sincerely acknowledge the financial supported by the [48] Kotbagi TV, Biradar AV, Umbarkar SB, Dongare MK. ChemCatChem
Natural Science Foundation of China (21576158 and 21576159), 2013;5:1531–7.
[49] Umbarkar SB, Kotbagi TV, Biradar AV, Pasricha R, Chanale J, Dongare MK, et al. J
Shandong Provincial Natural Science Foundation (ZR2017JL014).
Mol Catal A Chem 2009;310:150–8.
[50] Shetty M, Murugappan K, Green WH, Román-Leshkov Y. ACS Sustainable Chem Eng
Appendix A. Supplementary data 2017;5:5293–301.
[51] Imran G, Srinivasan VV, Maheswari R, Ramanathan A, Subramaniam B. J Porous
Mater 2015;23:57–65.
Supplementary data associated with this article can be found, in the [52] Rivoira L, Martínez ML, Anunziata O, Beltramone A. Microporous Mesoporous
online version, at http://dx.doi.org/10.1016/j.fuel.2018.06.081. Mater 2017;254:96–113.
[53] Hamdy MS, Mul G. Catal Sci Technol 2012;2:1894–900.
[54] Sarkar B, Goyal R, Sivakumar Konathala LN, Pendem C, Sasaki T, Bal R. Appl Catal
References B 2017;217:637–49.
[55] Iglesias J, Melero JA, Bautista LF, Morales G, Sánchez-Vázquez R, Andreola MT,
[1] Aransiola EF, Ojumu TV, Oyekola OO, Madzimbamuto TF, Ikhu-Omoregbe DIO. et al. Catal Today 2011;167:46–55.
Biomass Bioenergy 2014;61:276–97. [56] Zhu H, Maheswari R, Ramanathan A, Subramaniam B. Microporous Mesoporous
[2] Lee AF, Bennett JA, Manayil JC, Wilson K. Chem Soc Rev 2014;43:7887–916. Mater 2016;223:46–52.
[3] Maksimov AL, Nekhaev AI, Ramazanov DN, Arinicheva YA, Dzyubenko AA, [57] Khayoon MS, Hameed BH. Appl Catal A 2013;464–465:191–9.
Khadzhiev SN. Petrol Chem 2011;51:61–9. [58] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu CC. Fuel
[4] Nanda MR, Zhang Y, Yuan Z, Qin W, Ghaziaskar HS, Xu C. Renewable Sustainable 2014;117:470–7.
Energy Rev 2016;56:1022–31. [59] Ramazanov DN, Dzhumbe A, Nekhaev AI, Samoilov VO, Maximov AL, Egorova EV.
[5] Anuar MR, Abdullah AZ. Renewable Sustainable Energy Rev 2016;58:208–23. Petrol Chem 2015;55:140–5.
[6] Knothe G, Razon LF. Prog Energy Combust 2017;58:36–59. [60] Stawicka K, Díaz-Álvarez AE, Calvino-Casilda V, Trejda M, Bañares MA, Ziolek M. J
[7] Knothe G, Cermak SC, Evangelista RL. Fuel 2012;96:535–40. Phys Chem C 2016;120:16699–711.
[8] Ayoub M, Abdullah AZ. Renewable Sustainable Energy Rev 2012;16:2671–86. [61] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu C. Appl Energy
[9] Zhou CH, Zhao H, Tong DS, Wu LM, Yu WH. Catal Rev 2013;55:369–453. 2014;123:75–81.
[10] Bagheri S, Julkapli NM, Yehye WA. Renewable Sustainable Energy Rev [62] Pierpont AW, Batista ER, Martin RL, Chen W, Kim JK, Hoyt CB, et al. ACS Catal
2015;41:113–27. 2015;5:1013–9.
[11] Lari GM, Mondelli C, Pérez-Ramı́rez J. ACS Catal 2015;5:1453–61.

387
 Catalysis Today 254 (2015) 83–89

Contents lists available at ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

Amphiphilic property of niobium oxyhydroxide for waste glycerol

conversion to produce solketal
Talita E. Souza a , Izabela D. Padula a , Mariana M.G. Teodoro b , Poliane Chagas a ,
Jarbas M. Resende a , Patterson P. Souza b , Luiz C.A. Oliveira a,∗
a
Department of Chemistry, Federal University of Minas Gerais, Belo Horizonte CEP 31270-901, Brazil
b
Federal Center of Technological Education of Minas Gerais, Belo Horizonte CEP 30421-169, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Amphiphilic catalysts are synthesized using NbCl5 in the presence of the CTAB (cetyltrimethylammonium
Received 4 September 2014 bromide) to generate partial hydrophobicity to the catalysts. The partial hydrophobicity of the niobium
Received in revised form 3 December 2014 oxyhydroxides improved the acetalization reaction of a residual glycerol from biodiesel production by
Accepted 4 December 2014
decreasing the interaction between the water molecules and the acid sites of the catalyst. A waste glycerol
Available online 12 January 2015
conversion of 73% with a selectivity to solketal (2,2-dimethyl-[1,3]dioxolan-4-yl)methanol of 95% was
obtained. Many reuses of the catalysts showed glycerol conversions between 70 and 80%.
Keywords:
© 2015 Elsevier B.V. All rights reserved.
Niobium
Amphiphilic properties
Acetalization
Waste glycerol

1. Introduction In heterogeneous catalysis, different types of catalysts are used

in the acetalization reactions of glycerol, mesoporous silicates [13],
Many possibilities for use of the residual glycerin have been meso-structured silica functionalized with sulfonic acid [14], ZrO2
investigated, especially employing catalytic processes [1,2]. Cat- [15], Amberlyst resins [14] and oxides [8,15], among others.
alytic conversion reactions of glycerol, which is a highly reactive Modified catalysts based on niobium compounds, have received
molecule, favor obtaining chemical products with high added value, special attention because of their catalytic activity in several impor-
which can be obtained by reactions such as selective oxidation and tant chemical processes [16–20], as a promoter effect in oxidation
hydrogenolysis, dehydration, steam reforming, thermal reduction reactions [21,22], supporting effect in hydrogenation reactions
in synthesis gas, oligomerization/polymerization, selective trans- [23], metathesis [24], hydrotreating [25] and acid catalysis [26].
esterification, etherification and epichlorohydrin synthesis [3–6]. The acetalization reaction occurs through the condensation of a
Among these, we highlight the acetalization and etherification of glycerol molecule with one molecule of acetone, resulting in the lib-
glycerol to obtain fuel additives, which have a lower polarity and eration of water. The formed water can favor the reverse reaction,
viscosity, and therefore greater volatility, which can improve the reducing the formation of solketal [14,27]. Alternatives for these
combustion efficiency [2,7]. reactions have been the use of distillation processes [27] after the
In acetalization reactions, through the reaction of glycerol with reaction, or hydrophobic catalysts [28], which decrease the inter-
acetone, the products are cyclic acetals with 5 or 6 ring members, action between water molecules and the acid sites, without losing
with the solketal ((2,2-dimethyl-1.3-dioxolan-4-yl)methanol) as catalytic properties.
major product [8,9]. These ketals can act as additives in order to Niobium oxide shows a good water tolerance property as
increase the octane level of gasoline. reported by Okuhara [28], and exhibited the highest activity among
To obtain solketal, glycerol and acetone catalytic reactions have the solid acids. Nair et al. [8] shows the water tolerance of niobia
been studied using homogeneous and heterogeneous processes. In acid sites in acetalization reactions. Studies reported by Okuhara
homogeneous catalysis p-toluene sulphonic acid [10], hydrochloric [28] and Nair et al. [8], employ niobium oxide calcinated until
acid and sulfuric acid [11,12] have been employed. 773 K. Above 773 K this activity decreases the water tolerance,
thus it can be suggest that structural changes in the niobium
oxide occur [28]. In this work, we report the synthesis of a new
∗ Corresponding author. Tel.: +55 31 34096384; fax: +55 31 34095700. series of catalysts based on niobium oxyhydroxide obtained from
E-mail address: [email protected] (L.C.A. Oliveira). the precursor of niobium chloride and surface modification by

http://dx.doi.org/10.1016/j.cattod.2014.12.027
0920-5861/© 2015 Elsevier B.V. All rights reserved.
84 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89

partial hydrophobization with surfactants groups. Preliminary 100

S4
results from our research group have shown success in hydropho- 4.79%
bization processes of niobium-based catalysts with excellent S4-2D
95
5.87%
results in many types of reactions [29–32]. S4-SS
7.66% 7.14%
90
12.79%

Weight Loss (%)

2. Materials and methods
85
2.34%
2.1. Synthesis and characterization of the catalysts
80
The catalysts were synthesized, starting from the precursor of 13.94% 2.95%
niobium chloride (NbCl5 99%, purchased from Sigma–Aldrich). A 75
water:oil:surfactant system was prepared, consisting of a mix-
ture of Milli-Q water, 1-butanol (99%, Vetec), and 1-hexanol (99%, 70
Vetec) in proportions of 17, 22 and 61%, respectively. 41.2 mmol
CTAB (cetyltrimethylammonium bromide) surfactant was added 65
and after solubilization 46.2 mmol of (NbCl5 ) was added at 60 ◦ C 0 100 200 300 400 500 600
and constant stirring. A solution of NH4 OH (5 mol L−1 ) was then o
Temperature ( C)
added, dropwise, under constant stirring for 12 h (S4 catalyst) or
48 h (S4-2D catalyst). Finally, the resulting white solid is filtered, Fig. 1. Thermogravimetric analyses for catalysts based on niobium oxyhydroxide.
washed and dried at 60 ◦ C for 12 h.
A catalyst without surfactant was also synthesized by dissolv-
ing NbCl5 in 50% alcoholic solutions at 60 ◦ C, after the addition of oven temperature profile was 80 ◦ C for 5.0 min increasing to 100 ◦ C
NH4 OH until it reached pH 7, a solid white is precipitated. This at 3 ◦ C min−1 , maintaining the final temperature for 1 min.
is kept under stirring for 12 h, thus this is filtered, washed, and The calculation of the conversion, selectivity and products yield
oven-dried at 60 ◦ C for 12 h. This procedure is from catalyst S4-SS. were determined by the calibration curve with standard glycerol PA
The structural properties of materials were investigated (Modern Chemical, 99.5%) and solketal PA (provided by Petrobras).
by different characterization techniques. The thermal analysis
measurements were performed on a Shimadzu-TGA50H ther-
3. Results and discussions
mobalance at 700 ◦ C in air atmosphere and flow of 100 mL min−1 ,
with a heating rate of 10 ◦ C min−1 . Analyses of infrared spec-
3.1. Characterizations of the catalysts
troscopy (FTIR) were obtained using a Perkin Elmer RXI FTIR
spectrophotometer. The spectra of the samples were obtained in
Thermogravimetric analyses of the catalysts, Fig. 1, showed
the 4000–500 cm−1 region with resolution of 4 cm−1 and an aver-
three mass loss events, which are associated with the surface
age of 64 scan signal. Analysis of X-ray diffraction (XRD) was
hydroxyl groups above 110 ◦ C and after 200 ◦ C mass losses asso-
performed in the range between 2 = 20 and 70 in continuous
ciated with the surfactant group.
scan mode of 2.0◦ min−1 using k␣Cu radiation ( = 1.542 Å) in a
A mass loss of 14 and 3% was observed related to surfactant
Siemens D5000 diffractometer. Acidity of the materials was char-
group decomposition for S4 and S4-2D, respectively. These results
acterized by NH3 -TPD in a Quantachrome ChemBET-3000, TCD
indicate a lower degree of hydrophobization of the catalyst S4-2D
detector with 150 mA current and attenuation value of 32. N2
compared to S4. Above 450 ◦ C, the event observed in three sam-
adsorption/desorption isotherms were measured at 77 K using a
ples is associated with the loss of hydroxyl groups from the bulk,
Quantachrome Autosorb IQ2 equipment. Data were analyzed by
as suggested by some authors [31,33]. The mass loss for the cata-
BET method and pore size distribution using the BJH model. Scan-
lyst SS-S4 suggests that initially, the event with 5.87% loss is due
ning electron microscopy (SEM) obtained the morphology of the
to the humidity and the two following events are related to OH
catalysts in a Quanta 200 FEG-FEI scanning electron microscope,
loss (11.74%), from surface and bulk. These data suggests that the
operating at 30 kV with a resolution of 1.6 nm.
molecular formula of the S4-SS catalyst is NbO2 (OH), referred to as
niobium oxyhydroxide [32,34].
2.2. Catalytic reactions Analyses by infrared spectroscopy are shown in Fig. 2. The cat-
alyst S4 presented bands related to ␯C-H at 2922 and 2852 cm−1 ,
Typical acetalization reactions were performed at 70 ◦ C with which is a strong evidence of the presence of surfactant groups on
constant stirring in a batch reactor with an acetone/glycerol molar the surface of the catalyst.
ratio of 2/1, 0.2 g catalyst and 1 h of reaction. Residual glycerol from The S4-2D catalyst presented similar profile compared to S4-SS.
biodiesel production and a desalinated one were used in the acetal- The inset of Fig. 2, shows OH bond stretching observed at 3405 cm−1
ization reactions, both provided by Petrobras. NMR and GC-MS related to surface hydroxyls and at 3123 cm−1 due to the OH from
analyzed the conversion of glycerol and product identification. bulk, as reported by Schwertman and Cornell [32,34]. This fact gives
1 H, 13 C and DEPT, experiments, were performed at 20 ◦ C on an evidence of the formation of niobium oxyhydroxide. The presence
RMN spectrometer Avance DRX 200 at the High Resolution Mag- of the surfactant over the catalysts gives them a hydrophobic char-
netic Resonance Laboratory of UFMG. With a 5 nm multinuclear acter, which can contribute to the glycerol acetalization reactions.
probe and procedure to optimize the field homogeneity (shim- X-ray diffraction analyses not show diffraction patterns for cata-
ming), as well as probe tuning in the frequency of 1 H and 13 C, in lysts S4, S4-2D and S4-SS. The amorphous character of the catalysts
order to have optimal resolution and signal detection. are expected, because these were not calcinated during the synthe-
Analyses by GC–MS in an Agilent TechnologiesTM Model 7890A, sis to preserve the surfactant group, as surface and bulk hydroxyl
coupled to a mass detector model 5975C, were performed on HP- groups, which are characteristic of oxyhydroxides.
5MS (5% polymethylphenylsiloxane), 30 m × 250 ␮m × 0.25 ␮m The strength of the acid sites and total acidity were determined
column. Helium gas was used as carrier gas at a flow of by NH3 -TPD. The catalysts showed similar NH3 desorption profiles,
1.4 mL min−1 , split ratio of 130/1, injector temperature of 25 ◦ C and as shown in Fig. 3.
 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89 85

The catalyst without hydrophobization (S4-SS) showed a greater

area under the ammonia desorption curve. This result was
expected, since the hydrophobic group is attached to the catalyst
surface, replacing the Brønsted acid groups [35]. Using a zeolite as
a standard material, the number of acid sites could be calculated:
ν O-H
90.1, 63.3 and 58.5 ␮mol g−1 for S4-SS, S4-2D and S4, respectively.
Absorbance (arb. unit.)

ν O-H

Surface Bulk ν O-H The low acidity of the catalysts with highest degree of hydropho-
3600 3300 3000 2700 bization confirms the thermal analysis data.
The N2 adsorption and desorption isotherms and pore size dis-
tribution of the S4-SS, S4-2D and S4 catalysts are shown in Fig. 4.
The N2 adsorption isotherms are classified as Type IV according
to IUPAC, as reported by Kruk and Jaroniec [36], for materials
S4 with hydrophobic surfaces (surfactant-containing), however, type
S4-2D II isotherm is better to classified the S4 catalyst. The S4-2D and S3-
S4-SS SS catalysts have isotherm profiles very similar, this is due a similar
structure, since TG analyses show that hydrophobization was 3% for
4000 3500 3000 2500 2000 1500 1000 500
-1 S4-2D catalyst.
Wavenumber (cm )
The hysteresis may be classified as H3, for S4 catalysts, which
Fig. 2. Vibrational spectroscopy in the infrared region for the catalysts S4, S4-2D does not level off at relative pressures close to the saturation
and S4-SS. vapor pressure and is reported for materials comprised of aggre-
gated (loose assemblages) platelike particles forming slitlike pores.
The hysteresis in S4-2D and S4-SS catalyst is like H2 hysteresis,
S4 A = 21040 these achieve the level off at relative pressures close to the sat-
100 uration vapor pressure, the shape triangular, such as observed for
many porous inorganics oxides and assigned to porous connectivity
effect, as reported by Liu et al. [37].
The catalyst without CTAB (S4-SS) showed low BET-specific
Sinal TPD-NH3

S4-2D A= 22782
area value (135 m2 g−1 ). In the case of S4 and S4-2D, the presence
of surfactant leads to the formation of a material with relatively
100 high specific surface area (167 and 198 m2 g−1 for S4 and S4-2D,
respectively). The isotherm profile shows a continuous adsorption
of N2 throughout the range of relative pressures, and hysteresis are
observed during desorption. This result suggests the presence of
S4-SS
A= 32433 micro- and mesopores, as shown by the pore size distribution in
inset of Fig. 4.
100
The morphology of the materials was investigated using scan-
100 200 300 400 500 600 700
ning electron microscopy (Fig. 5). The S4 material synthesized using
NbCl5 as a precursor, exhibited a morphology distinct from that of
Temperatura ( °C)
the other materials with square openings that form a well-defined
porous structure, possibly generating a material with a high specific
Fig. 3. NH3 -TPD profile of the catalysts: S4, S4-2D and S4-SS.

200
1.5
3 -1

S4
Dv (log d)/cm g

1.0

160 S4-2D
0.5
S4-SS
Volume of N2/cm3 g-1

0.0
120 10 20 40 60 80100 200 400 600
Pore Diameter / Å

80

Área = 167 m²/g

40 2
198 m /g
2
135 m /g

0
0.0 0.2 0.4 0.6 0.8 1.0
P/Po

Fig. 4. N2 adsorption/desorption isotherms and pore size distribution of catalysts.

86 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89

Fig. 5. Micrographs obtained for the catalysts.

surface area. The S4-SS and S4-2D showed a change in morphology 100
9%
as agglomerated particles. Water Loss Crude Glycerin
16%
80
Desalinated Glycerin
3.2. Catalytic studies
Weight Loss (%)

Glycerol acetalization reactions were performed using different 60

types of glycerine from biodiesel production (crude and desali-
nated glycerol) and commercial glycerol. Usually the crude glycerol
contains between 2–3% fatty acid residues derived from the trans- 40
esterification reaction for biodiesel production, 6–8% of NaCl and
10–15% water. The glycerol amount in the crude is between 75 and
80%. 20

Thermogravimetric analyzes were performed to determine the <1%

6% Salt
salt and water amounts present in the waste glycerol utilized in
0
this work, in order to estimate the concentration of glycerol for the
0 100 200 300 400 500 600 700 800
calculation of conversion and yield, Fig. 6. o
Temperature ( C)
As expected, the amount of salt and water present in crude glyc-
erin was 6% and 16% of water, respectively. In the case of desalinated Fig. 6. Thermogravimetric analysis for crude and desalinated glycerin.
glycerin has 1% of salt and 9% of water. The amount of glycerol were
considered 78% and 90% for crude and desalinated glycerin, respec-
tively, which were used to calculate the conversion and yield of
the reactions. Fig. 7 shows data of the conversion of glycerol using The best results were obtained using desalinated glycerin that pre-
catalysts S4-SS, S4 and S4-2D with different types of glycerin. sented conversion values of 22, 43, 40 and 55% for S4, S4-2D, S4-SS
Fig. 7 presents an overview of glycerin (crude, desalinated and and HY-340, respectively. Acetalization reactions using commercial
commercial glycerol) conversion using the synthesized catalysts and crude glycerin, showed a similar conversion for the different
S4, S4-SS and S4-2D compared to commercial niobic acid, HY-340. catalyst. The low conversion presented using commercial and crude
 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89 87

70 100
S4-SS
S4
S4-2D
60 S4-2D
S4
S4-SS 80
HY340
HY-340
50

Glycerol conversion (%)

Glycerol Conversion (%)

60
40

30
40

20

20
10

0 0
1:2
1 1:4 1:6 2:2
Commercial glycerol Crude glycerin Desalinated glycerin
Molar ratio glycerol/acetone
Fig. 7. Conversion of acetalization reactions using the catalysts S4-SS, S4, S4-2D and
HY-340 at 70 ◦ C for 60 min, molar ratio glycerol/acetone 1/2. Fig. 8. Conversion of glycerin at molar ratios equal to 1/2, 1/4, 1/6 and 2/2 of glyc-
erol/acetone.

Table 1
TOF values obtained from acetalization reaction using desalinated glycerin with a
2/1 molar ratio of acetone (70 ◦ C for 1 h).
number of moles of solketal formed and the number of acid sites
Catalysts TOF (h−1 )
determined by NH3 -TPD (1 h reaction). The high TOF value for the
HY-340 380 S4-2D catalyst explains the high catalytic activity with desalinated
S4 618
glycerin.
S4-2D 1106
S4-SS 716 Reactions with different glycerol/acetone molar ratios (Fig. 8)
were performed. The conversion of glycerol showed a significant
increase when the glycerol/acetone molar ratio was 1/4 (73% con-
glycerol is probably due to the presence of NaCl (crude glycerol) and version) for 1 h of reaction using the catalyst S4-2D. Other studies
the high viscosity of commercial glycerol [38]. have shown the influence of the 1//1, 1/2 and 1/4 glycerol/acetone
The S4-2D catalyst showed better conversion even under high molar ratio, and the results were the best with 1/4, with greater
viscosity conditions of the commercial glycerol and in the presence than 90% conversion after 5 h of reaction using catalysts based on
of large amounts of salt (crude glycerin). The high specific area and carbon functionalized with acid groups [39]. In the literature, it
number of acid sites can explain these results. Table 1 shows the can be found that the conversion with an acetone/glycerol molar
TOF (turnover frequency), which is the number of catalytic cycles ratio (1/1) using Amberlyst-15 resin and K-10 Montmorillonite was
in the active center of the catalyst per time unit, obtained from the approximately 60% after 60 min of reaction [2]. For mesoporous

Fig. 9. NMR spectra of reaction 1/4 molar ratio glycerol/acetone with S4-2D catalyst.
88 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89

Scheme 1. Acetalization reaction of glycerol with acetone under acid catalysis to form as the major product (2,2-dimethyl-[1,3]-dioxolan-4-yl)methanol.

100 interface, reducing the interference of water molecules on the cat-

alyst surface.
The different textural properties such as surface area and acid-
80 ity obtained in partially hydrophobized catalysts and with only
hydrophilic surface showed good conversion results, about 77%
Glycerol Conversion (%)

depending on the reaction conditions studied, and 95% selectivity

60 for solketal formation.
In reactions with a 1/4 glycerol/acetone molar ratio, the glycerol
conversion was still greater than 80% after some reaction cycles.
40 The results demonstrated sustainable use of desalinated residual
glycerin, production of solketal, and its application as fuel additives.

20 Acknowledgments

The authors thank Petrobras for funding, as well as the

0 Microscopy Center – UFMG, the infrastructure of the Chemistry
Reaction Resuse 1 Reuse 2 Reuse 3
Department of UFMG and CEFET. The authors also thank FAPEMIG,
S4-2D Yields (%) S4-2D HY-340 Yelds (%) HY-340 CNPQ and CAPES for financial support.

Fig. 10. Conversion and yield of the reuse of the catalysts S4-2D and HY-340 with
References
glycerol/acetone molar ratio of 1/4, at 70 ◦ C, 1 h.

[1] L.C.A. Oliveira, M.F. Portilho, A.C. da Silva, H.A. Taroco, P.P. Souza, Appl. Catal. B
silicates with a 1/1 molar ratio, there were conversions ranging 117–118 (2012) 29–35.
from 28 to 52% for reactions at 80 ◦ C 6 h [13]. [2] C.X.A. da Silva, V.L. Gonçalves, C.J.A. Mota, Green Chem. 11 (2009) 38–41.
[3] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C.D. Pina, Angew. Chem. Int. Ed.
In acetalization reactions the products formed are 5-membered 46 (2007) 4434–4440.
rings (2,2-dimethyl-[1,3]-dioxolan-4-yl)methanol, called solketals, [4] Y. Zheng, X. Chen, Y. Shen, Chem. Rev. 108 (2008) 5253–5277.
and a 6-membered ring (2,2-dimethyl-1,3-dioxan-5-ol). However, [5] J. Barrault, F. Jerome, Eur. J. Lipid Sci. Technol. 110 (2008) 825–830.
[6] A. Behr, J.P. Gomes, J. Eur, Lipid Sci. Technol. 112 (2010) 31–50.
the selectivity of formation of each of these products depends on [7] N. Viswanadham, S.K. Saxena, Fuel 103 (2013) 980–986.
the acid strength of the catalyst, the higher the acid strength, the [8] G.S. Nair, E. Adrijanto, A. Alsalme, I.V. Kozhevnikov, D.J. Cooke, D.R. Brown, N.R.
higher the selectivity for solketal [2,8]. The reaction for obtaining Shiju, Catal. Sci. Technol. 2 (2012) 1173–1179.
[9] M.J. Climent, A. Corma, A. Velty, Appl. Catal. A263 (2004) 155–161.
solketal is shown in Scheme 1. [10] Organic Syntheses III, Collective Volume, Wiley & Sons, New York, 1955.
The selectivity presented by S4-2D catalyst after 1 h of reaction [11] N. Suriyaprapadiloka, B. Kitiyanana, Energy Procedia 9 (2011) 63–69.
was 95%; as reported by Souza et al. [32], the other products were [12] A. Krief, L. Provins, A. Froidbise, Tetrahedron Lett. 39 (1998) 1437–1440.
[13] L. Li, T.I. Koranyi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012) 1611–1612.
not identified by NMR and GC–MS, like shown in Fig. 9, only solketal [14] G. Vicent, J. Melero, G. Morales, M. Paniagua, E. Martín, Green Chem. 12 (2010)
has formed. 899–907.
In order to study the stability of the catalysts the reuse of mate- [15] P.S. Reddy, P. Sudarsanam, B. Mallessham, G. Raju, B.M. Reddy, J. Ind. Eng. Chem.
17 (2011) 377–381.
rials using desalinated glycerin was studied. The results are shown
[16] M. Ziolek, Catal. Today 78 (2003) 47–64.
in Fig. 10. [17] K. Tanabe, Catal. Today 78 (2003) 65–77.
The reuse of the catalysts was tested in many subsequent reac- [18] M. Ziolek, I. Sobczak, I. Nowak, P. Decyk, A. Lewandowsk, J. Kujawa, Microporous
Mesoporous Mater. 35–36 (2000) 195–207.
tions using desalinated glycerin (Fig. 10). The conversion remained
[19] I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603–3624.
constantly at approximately 70%, indicating the high stability of [20] K. Tanabe, S. Okazaki, Appl. Catal. A 133 (1995) 191–218.
the hydrophobized catalyst S4-2D. Interestingly, in addition to the [21] N. Marin-Astorga, J.J. Martínez, D.N. Suarez, J. Cubillos, H. Rojas, C.A. Ortiz, Curr.
catalyst remaining active, only the formation of solketal occurs, Org. Chem. 16 (2012) 2797–2801.
[22] H.S. Oliveira, L.D. Almeida, V.A.A. de Freitas, F.C.C. Moura, P.P. de Souza, L.C.A.
showing that the material maintains its properties in the acetal- Oliveira, Catal. Today 240 (2015) 176–181.
ization process. The yield of the reaction is greater than 80% for [23] T. Iizuka, Y. Tanaka, K. Tanabe, J. Mol. Catal. 17 (1982) 381–389.
S4-2D catalyst in the first two cycles of reuse. On the other hand, [24] R. Buffon, et al., J. Mol. Catal. 76 (1992) 287–295.
[25] C. Geantet, J. Afonso, M. Breysse, N. Allali, M. Danot, Catal. Today 28 (1996)
HY-340 catalyst presented 40% of yield in the second reuse. 23–30.
[26] J. Souza, P.M.T.G. Souza, P.P. de Souza, D.L. Sangiorge, V.M.D. Pasa, L.C.A. Oliveira,
4. Conclusions Catal. Today 213 (2013) 65–72.
[27] J.S. Clarkson, A.J. Walker, M.A. Wood, Org. Process Res. Dev. 5 (2001) 630–635.
[28] T. Okuhara, Chem. Rev. 102 (2002) 3641–3666.
In this work niobium based catalysts with high catalytic activity [29] P. Chagas, H.S. Oliveira, R. Mambrini, M. Le Hyaric, M.V. de Almeida, L.C.A.
due to the specific surface area and acidic properties were syn- Oliveira, Appl. Catal. A 454 (2013) 88–92.
[30] L.C.A. Oliveira, H.S. Oliveira, G. Mayrink, H.S. Mansur, A.A.P. Mansur, R.L. Mor-
thesized and characterized. The higher hydrophobization decrease eira, Appl. Catal. B 147 (2014) 403–412.
the number of acid sites of the catalysts as shown by NH3- [31] N.T. Prado, F.G.E. Nogueira, A.E. Nogueira, C.A. Nunes, R. Diniz, L.C.A. Oliveira,
TPD. However, the hydrophobic groups act at the glycerol/acetone Energy Fuels 24 (2010) 4793–4796.
 T.E. Souza et al. / Catalysis Today 254 (2015) 83–89 89

[32] T.E. Souza, M.F. Portilho, P.M.T.G. Souza, P.P. Souza, L.C.A. Oliveira, Chem- [36] M. Kruk, M. Jaroniec, S.H. Ryoo, Chem. Mater. 12 (2000) 1414–1421.
CatChem 6 (2014) 2961–2969. [37] H. Liu, N.A.J.J. Seaton, Colloid Interface Sci. 156 (1993) 285–293.
[33] H. Landmesser, et al., Solid State Ion. 271 (1997) 101–103. [38] C.X.A. da Silva, C.J.A. Mota, Biomass Bioenergy 35 (2011) 3547–3551.
[34] R.M. Cornell, U. Schwertmann, The Iron Oxides, VCH, New York, 1996. [39] R. Rodrigues, M. Gonçalves, D. Mandelli, P.P. Pescarmona, W.A. de Carvalho,
[35] A. Corma, V. Fornes, M.T. Navarro, J. Perez-Pariente, J. Catal. 148 (1994) Catal. Sci. Technol. 4 (2014) 2293–2301.
574–769.
 energies
Review
Glycerol to Solketal for Fuel Additive: Recent
Progress in Heterogeneous Catalysts
Is Fatimah 1, * , Imam Sahroni 1 , Ganjar Fadillah 1 , Muhammad Miqdam Musawwa 1 ,
Teuku Meurah Indra Mahlia 2 and Oki Muraza 3, *
1 Department of Chemistry, Faculty of Mathematics and Natural Sciences, Universitas Islam Indonesia,
Jl. Kaliurang Km 14, Sleman, Yogyakarta 55584, Indonesia
2 School of Information, Systems and Modelling, Faculty of Engineering and Information Technology,
University of Technology Sydney, Sydney, NSW 2007, Australia
3 Center of Research Excellence in Nanotechnology and Chemical Engineering Department,
King Fahd University of Petroleum & Minerals, Dhahran 31261, Saudi Arabia
* Correspondence: [email protected] (I.F.); [email protected] (O.M.)




Received: 4 June 2019; Accepted: 10 July 2019; Published: 25 July 2019


Abstract: Biodiesel has been successfully commercialized in numerous countries. Glycerol, as

a byproduct in biodiesel production plant, has been explored recently for fuel additive production.
One of the most prospective fuel additives is solketal, which is produced from glycerol and acetone via
an acetalization reaction. This manuscript reviewed recent progress on heterogeneous catalysts used in
the exploratory stage of glycerol conversion to solketal. The effects of acidity strength, hydrophobicity,
confinement effect, and others are discussed to find the most critical parameters to design better
catalysts for solketal production. Among the heterogeneous catalysts, resins, hierarchical zeolites,
mesoporous silica materials, and clays have been explored as effective catalysts for acetalization of
glycerol. Challenges with each popular catalytic material are elaborated. Future works on glycerol
to solketal will be improved by considering the stability of the catalysts in the presence of water
as a byproduct. The presence of water and salt in the feed is certainly destructive to the activity
and the stability of the catalysts.

Keywords: fuel additives; biodiesel; glycerol; solketal; solid acid catalysts.

1. Introduction
The exploration of renewable energy to supplement limited fossil fuels in the next few years
is one the most concerned research topics. Among some renewable energy resources, biofuels are
receiving intensive attention, especially for some countries with a large production of vegetable oils
and bio-oils for biodiesel production [1–4]. Annual production and consumption of biodiesel is likely
to increase significantly in the coming few years. Numerous sources of abundant edible and potential
non-edible oils have been identified [5]. Regardless, this fact leads to increasing glycerol production as
the byproduct of biodiesel conversion [2]. Due to the chemical process of the biodiesel production,
the molar ratio of glycerol to the methyl ester is 3:1, or about 10% to 20% of the total volume of biodiesel
produced is made up of glycerol. The rapid growth of biodiesel production has contributed much to
the increasing glycerol production since it was reported that the worldwide production of glycerol
increased from 7.8 billion liters in 2006 to 36 billion liters in 2018 [6,7]. This fact revealed that glycerol
is an abundant renewable chemical feedstock in the world. The conversion of glycerol into more
valuable chemicals is the best option to create a new market for glycerol and improve the sustainability
of biodiesel production [7–14].

Energies 2019, 12, 2872; doi:10.3390/en12152872 www.mdpi.com/journal/energies

Energies 2019, 12, 2872 2 of 14

This mini review paper aims to emphasize the potential exploration of catalytic materials
for the conversion of glycerol to solketal by analyzing recent papers, especially open literature
from after 2010. Rahmat et al. (2010) [15] wrote an overview of glycerol conversion to fuel
additives, with an emphasis on reaction parameters (catalyst, reactant, temperature, and reaction
time). In the range of 2009 to 2018, Cornejo et al. [16] wrote a review in 2017 on glycerol valorization
to fuel additives over different co-reactants. These included second feeds, such as formaldehyde,
acetaldehyde, butanal, and acetone, and many others. Nanda et al. [17] published a review on solketal
as a fuel additive, with an emphasis on the historical and future context. This paper also summarized
the effect of acidity, reactor models, kinetics and reactor kinetics, and the daily procedure to use glycerol
to solketal.
Many scenarios were conducted for the conversion of glycerol to different value-added chemicals,
such as propane-acrolein, 1, 3-diol, propane-1,2-diol, acetal or ketal, polyols and polyurethane foams,
glycerol carbonate, etc. [10,11,18]. Table 1 shows that among these glycerol conversions, the conversion
of glycerol to solketal by acetalization is an interesting route. Solketal is one of the glycerol acetalization
products together with glycerol acetal and glycerol formal (GlyF). Similar to other acetalization products,
solketal can be used directly as a fuel additive for the reduction of soot and gum formation [19].
Solketal addition to a gasoline blend showed better fuel properties with a higher octane number [19].
Other applications of solketal are in solvents, inks, pharmaceuticals, and paints [20].

Table 1. Different conversion routes from glycerol to value-added products.

Conversion Catalyst Results Ref

Diglycerol isomer was also increased from 35% to 55% while
Etherification Li/clay [21]
the selectivity to aa isomer was decreased from 65% to 35%
K/Al2 O3 -ZrO2 -FeOX Improvement in conversion
Conversion glycerol
Alkali metals supported With the increase of the K content in the catalyst, allyl [22]
to allyl Alcohol
to ZrO2-FeOx alcohol yield increased up to 27%-C
The prepared catalyst performed better for allyl alcohol
ZSM-5-supported iron
Allyl alcohol production as compared to catalysts synthesized [23]
catalysts
by other methods
Conversion alcohol Transesterification of glycerol to glycerol carbonate (GC)
Mg3−x Al1 Cux [24]
to glycerol carbonate increased to 96% of yields
Acetalization glycerol MoPO supported A 40% MoPO/SBA-15 showed a conversion of 100%
[25]
with acetone to SBA-15 and selectivity of 98%
Acetalization of glycerol BEA zeolite with Showed conversion of 88% and selectivity 80% of five
[26]
with butanal the ratio Si/Al of 40 member rings acetal (2-propyl-1,3-dioxolan-4-yl)methanol
Glycerol etherification Achieving conversion of 93% and selectivity above 80% of
Cationic acidic resin [27]
with benzaldehyde 2-phenyl-1,3-di-oxan-5-ol
Acetalization of glycerol
with mono-substitude MoOx /TiO2 -ZrO2 Glycerol conversion to 1,3-dioxolane (74%) within 30 min [28]
benzaldehyde
Acetalization of glycerol MoO3 and WO3 A 71% glycerol conversion and a 96% solketal selectivity
[29]
with acetone supported to SnO2 were achieved.

As shown in Table 2 and Figure 1, different types of catalyst materials were reported for the solketal
production consisting of zeolites, clays, resins, heteropolyacids, and others. Each catalyst has both
advantages and drawbacks. A homogeneous catalyst, such as H2 SO4 , offers high activity, however,
these homogenous catalysts are corrosive, not recyclable, difficult to separate, and considerably more
expensive. Similarly, chloride, such as tin chloride (SnCl2 ), is also unwanted due to its corrosion
tendency [30]. Reusability is also an important part of studies. Reusability is a factor which is studied as
a typical sustainable principle. The basic mechanism of the metal salt catalysis is a nucleophilic attack
by the hydroxyl group of glycerol to the carbocation obtained from the protonation step, resulting in
the formation of the intermediate, followed by a water elimination step. The carbocation is produced
from the Lewis or Brønsted acid sites, which activates the ketone carbonyl group through a protonation
step (i.e., Brønsted acids) or polarization.
However, homogeneous catalysts are not considered as environmental-friendly for the reaction
system. Another challenge in the utilization of heterogeneous catalysts in solketal production is
 Energies 2017, 10, x FOR PEER REVIEW 3 of 14

However, homogeneous catalysts are not considered as environmental-friendly for the reaction
system. Another challenge in the utilization of heterogeneous catalysts in solketal production is the
Energies 2019, 12, 2872 3 of 14
byproduct (water) formed during the reaction, which induces a reversible reaction. Heterogeneous
catalysts are regenerated easily and are more easily handled. Many resin catalysts exhibited excellent
theconversion
byproductof glycerol
(water) to solketal
formed duringandtheselectivity, where
reaction, which the best
induces catalytic performance
a reversible was obtained
reaction. Heterogeneous
by amberlyst. However, it is not feasible for a higher scale of production
catalysts are regenerated easily and are more easily handled. Many resin catalysts exhibited due to the limitation
excellentof
thermal stability,
conversion of glycerolsotoitsolketal
is not andeasyselectivity,
to regenerate.
where The higher
the best thermal
catalytic stability was
performance can obtained
be found byin
hierarchical zeolite. The highest conversion of glycerol to solketal of 72% and the
amberlyst. However, it is not feasible for a higher scale of production due to the limitation of thermal selectivity of 72%
stability, so it is not easy to regenerate. The higher thermal stability can be found in hierarchical5%
are reached by using H-Beta (BEA framework) under the condition of 60 °C, stirring at 700 rpm,
of catalyst,
zeolite. and molar
The highest ratio ofof
conversion glycerol:acetone of 1:4offor
glycerol to solketal 72% H-BEA.
and theWithin the zeolite
selectivity of 72%materials,
are reached MFI
zeolite showed 80%, which is a lower catalytic activity in comparison
◦ with amberlyst,
by using H-Beta (BEA framework) under the condition of 60 C, stirring at 700 rpm, 5% of catalyst, but with almost
100%
and selectivity.
molar The lower conversion
ratio of glycerol:acetone is H-BEA.
of 1:4 for due to the relatively
Within narrow
the zeolite channel
materials, sizezeolite
MFI that affects
showed the
transport of the reactant carried out and the shape selectivity.
80%, which is a lower catalytic activity in comparison with amberlyst, but with almost 100% selectivity.
The lower conversion is due to the relatively narrow channel size that affects the transport of the reactant
Table 2. Classification of heterogeneous catalysts for solketal production.
carried out and the shape selectivity.
Double Layer
Others Heteropolyacid Resin Meso-SiO2 Zeolites
Table 2. Classification of heterogeneous catalysts for solketal production.
Hydroxide and Clay
Co/CNT Si-W (tungstosilisic) Amberlyst KIT-6 ZrO2 dolomite Zeolite X
Na- Double Layer
Others Heteropolyacid
HMQ-SJW Resin
Cat. Ex. Meso-SiO
Me-SBA-5J
2 Nb,AlOx ZeolitesMOR
VnOx/FER
lignosulfonate Hydroxide and Clay
Co/CNT SnF 2 H PW12040
Si-W (tungstosilisic)
3 Amberlyst-46
Amberlyst Hf-SBA-15
KIT-6 Nb oxy OH
ZrO2 dolomite BEAX
Zeolite
Na-lignosulfonate
Ionic liquid HMQ-SJW Cat. Ex.
Amberlyst-46 Me-SBA-5J
Mo-SBA-15 Nb, COK-S
AlOx VnOx/FER MOR
Hierarchical
SnFCarbon
2 H3 PW12040 Amberlyst-46
KU-2-8 Hf-SBA-15
Sn TUD-1 Nb oxy OH
MgLDH BEAMOR
BEA,
Ionic liquid Amberlyst-46
Lewatit Mo-SBA-15
Al-MCM- COK-S Hierarchical
Carbon KU-2-8 Sn TUD-1 Montmorillonite
MgLDH ZSM-5
BEA, MOR(MFI)
GF101 41
Lewatit Sulfonic
GF101 Al-MCM-41
Ga-MCM-4 Montmorillonite ZSM-5 (MFI)
DeAl BEA
Sulfonic Ga-MCM-4 DeAl BEA
Amberlyst-35 Acidity BEA
Amberlyst-35 Acidity BEA

Figure 1. Popularity of different types of catalytic materials for solketal production from 2014 to 2018.
(Source:
FigureWeb of Knowledge,
1. Popularity https://www.webofknowledge.com,
of different November
types of catalytic materials for solketal 2018). from 2014 to 2018.
production
(Source: Web of Knowledge, https://www.webofknowledge.com, November 2018).
2. Glycerol-to-Solketal Over Resin Catalysts
2. Glycerol-to-Solketal Over Resin
Overall, the most important Catalysts
properties of solid acid catalysts for the conversion glycerol to solketal
production was the Brønsted acidity of solid acids [31]. The conversion of glycerol to solketal with resin
Overall, the most important properties of solid acid catalysts for the conversion glycerol to
catalysts has been carried out [32–36]. Table 3 summarizes the conversion of glycerol to solketal over
solketal production was the Brønsted acidity of solid acids [31]. The conversion of glycerol to solketal
resin catalysts. A typical resin catalyst (i.e., amberlyst) catalyzed the reaction of glycerol with acetone
with resin catalysts has been carried out [32–36]. Table 3 summarizes the conversion of glycerol to
to produce above 80% of the glycerol conversion. Guidi et al. [36] reported that a resin, amberlyst-36,
solketal over resin catalysts. A typical resin catalyst (i.e., amberlyst)◦catalyzed the reaction of glycerol
which was applied at different reaction temperatures from 25 to 70 C, was an excellent catalyst to
with acetone to produce above 80% of the glycerol conversion. Guidi et al. [36] reported that a resin,
convert glycerol with a conversion of 85% to 97% to solketal with a selectivity of 99%. The catalyst is
amberlyst-36, which was applied at different reaction temperatures from 25 to 70 °C, was an excellent
also active at lower pressures with similar reaction parameters either in pure glycerol or in an equimolar
catalyst to convert glycerol with a conversion of 85% to 97% to solketal with a selectivity of 99%. The
reactant. According to some references, the high conversion was influenced not only by the surface
acidity but also by the resin structure. Moreover, the surface acidity was an important parameter that
played a crucial role in improving the selectivity and the conversion in the production of solketal.
Energies 2019, 12, 2872 4 of 14

Although amberlyst-46 and amberlyst-36 is a similar material, both types of resins have a different
acid capacity and structure morphology. Furthermore, all resins showed good selectivity to solketal
(>80%), and the important catalytic parameter of the resin to conversion glycerol is the acid capacity
(oversulfonated resin). With the highest acid capacity (sulfonic acid), these catalyst materials can
improve not only the selectivity to solketal production but also the conversion of raw glycerol to above
90%. Another important thing to be highlighted as a limitation of the catalyst activity is the presence
of NaCl as a poison for the surface acidity, which is possibly due to the impurities in glycerol.

Table 3. Glycerol-to-solketal over resin catalysts.

Selectivity
Source Catalyst Condition Conversion Remark Ref
to Solketal
Glycerol Glycerol:acetone = 1:2, 7.0 g of
Amberlyst-15 50 ◦ C 92% 96% [32]
and Acetone amberlyst-15 in 96 min
Glycerol
Amberlyst-46 60 ◦ C 84% 97% %1 (w/w) catalyst, 30 min [33]
and Acetone
Glycerol Amberlyst Glycerol:acetone = 1:2
70 ◦ C 97% 98% [34]
and Acetone DPT-1 at ambient pressure
Glycerol:acetone = 1:20, catalyst
Glycerol DT-851 sulfonic DT-851 sulfonic acid resin dosage
58 ◦ C 95% 99% [35]
and Acetone acid resin is 5% (wt., calculated by glycerol),
reaction time is 2 h.
At 10 barr and 25 ◦ C, A36 was
a highly active catalyst allowing
Glycerol good-to-excellent conversion
Amberlyst-36 25 ◦ C 85%–97% 99% [36]
and acetone (85%–97%) and selectivity (99%)
when either pure or wet glycerol
was used as a reagent.
Note: glycerol to the second reactant ratio was presented as molar ratio.

3. Glycerol-to-Solketal over Mesoporous Silica

Koranyi et al. [37] reported the superiority of hafnium and zirconium modified TUD-1 as superior
catalysts for the conversion of glycerol to solketal. These two catalysts (Hf-TUD-1 and Zr-TUD-1)
were more active than Sn-MCM-41 and Al-TUD-1. The Zr and Hf-TUD-1 are examples of active
metal-modified mesoporous silica in which Hf and Zr are in the framework. Their activity was higher
than FAU(USY) and Al(TUD-1). The highest conversion of glycerol to solketal was more than 50%.
The catalytic activity was a function of (i) the number of acid sites, (ii) the presence of mesopores,
(iii) the existence of a large surface area, and (iv) the hydrophobicity of the catalyst [38]. The later,
the hydrophobicity of the catalyst, was crucial to prevent the hydrolysis of solketal [37–41]. According to
Table 4, Cs 2.5/KIT-6 catalyst was one of the best catalysts for the conversion of glycero-to-solketal [42].
KIT-6 was selected because of its large surface area (600–1000 m2 /g), active sites, and accessible
pores [42].
Numerous references reported that mesoporous silica catalysts have the advantage of high
stability in the conversion of glycerol to solketal, resulting in products with a relatively large percentage
of conversion (95%) and selectivity to solketal (98%) [37,42–46]. The mesoporous structure with
an activated surface by sulfonic acid might be applied efficiently for the conversion of glycerol to fuel
additive [37,43,47]. A sulfonic acid-functionalized mesoporous polymer (MP-SO3 H) contains a high
acidity surface (1.88 mmol/g). The surface acidity of catalytic materials can accelerate the formation
products of solketal via ketalization reactions as shown in Figure 2.
Energies 2017,12,
Energies2019, 10,2872
x FOR PEER REVIEW 55 of
of 14
14

Table 4. Glycerol-to-solketal over mesoporous silica.

Table 4. Glycerol-to-solketal over mesoporous silica.
Selectivity
Source Catalyst Condition Conversion
Selectivity Remark Ref
Source Catalyst Condition Conversion to Solketal Remark Ref
to Solketal
Glycerol and Glycerol:acetone = 1:6,
Glycerol Cs 2.5/KIT-6 25 °C 95% in 15 min 98% Glycerol:acetone = 1:6, catalyst [42]
Acetone Cs 2.5/KIT-6 25 ◦ C 95% in 15 min 98% catalyst loading was 5 wt.%.[42]
and Acetone loading was 5 wt.%.
PropylsulfonicAcid
Propylsulfonic Acid Glycerol:formaldehyde =
Glycerol and
Glycerol and Functionalized
FunctionalizedSBA-15 90 °C 91.5 in 8 h 98% Glycerol:formaldehyde
1:1.5 with 4 wt.% = 1:1.5
catalyst [43]
formaldehyde 90 ◦ C 91.5 in 8 h 98% [43]
formaldehyde SBA-15 Mesoporous
Mesoporous Silica with 4 wt.% catalyst loading
loading
Glycerol and Silica
arenesulfonic acid-
arenesulfonic 70 °C 84% in 30 min 81% Glycerol:acetone = 1:6 [44]
Acetone
Glycerol functionalized silica
acid-functionalized 70 ◦ C 84% in 30 min 81% Glycerol:acetone = 1:6 [44]
and Acetone Sulfonic
silica acid- Glycerol:methyl acetate =
methyl acetate functionalized
Sulfonic
170 °C 99.5% in 4 h 74.2% Glycerol:methyl
1:50 andacetate
catalyst= 1:50
loading [45]
tomethyl acetate mesostructured
glycerol acid-functionalized
SBA- 170 ◦ C 99.5% in 4 h 74.2% and catalyst loading (7.5onwt.%
to glycerol mesostructured (7.5 wt.% based glycerol)[45]
15 silicas based on glycerol)
SBA-15 silicas
A sulfonic acid- The MP-SO3H catalyst
A sulfonic
Glycerol and functionalized The MP-SO 3 H catalyst
performed performed
better than other
Glycerol acid-functionalized 30 °C
30 ◦ C 94% 94% 98.5% 98.5% better than other conventional [46]
acetone
and acetone mesoporous
mesoporouspolymer conventional solid acid [46]
solid acid catalysts
polymer (MP-SO
(MP-SO 3H)3 H) catalysts
Glycerol:acetone = 1:2, 25 mg
Glycerol:acetone of 25
= 1:2,
Glycerol
Glycerol and Zr-TUD-1 80 ◦ C 64% catalyst, at room temperature, [37]
and acetone Zr-TUD-1 80 °C 64% mg of catalyst, at room [37]
acetone for 6 h.
temperature,
Glycerol:acetone = 1:2, 25for
mg6ofh.
Glycerol
Hf-TUD-1 80 ◦ C 65% catalyst,Glycerol:acetone
at room temperature,= 1:2, 25 [37]
Glycerol and
and acetone
Hf-TUD-1 80 °C 65% mgfor
of 6catalyst,
h. at room [37]
acetone
temperature, for 6 h.
Note: glycerol to the second reactant ratio was presented as molar ratio.
Note: glycerol to the second reactant ratio was presented as molar ratio.

Figure 2. Scheme
Figure 2. Scheme of
of mechanism
mechanism for
for the
the ketalization
ketalization reaction
reaction of
of glycerol
glycerol and
and acetone.
acetone.

4. Ketalization of Glycerol over Clay Minerals

4. Ketalization of Glycerol over Clay Minerals
Malaya et al. [17,48] studied different clay-based catalysts with different acid strengths ranging
Malaya et al. [17,48] studied different clay-based catalysts with different acid strengths ranging
from 0.12 to 5.7 meq/g [17]. The results show that a stronger acidity improved the conversion of
from 0.12 to 5.7 meq/g [17]. The results show that a stronger acidity improved the conversion of
glycerol up to ca. 80%. As shown in Table 5, solketal production from glycerol used two different
glycerol up to ca. 80%. As shown in Table 5, solketal production from glycerol used two different
sources, namely acetone or formaldehyde over solid acid catalysts [49–52]. Based on the conversion of
sources, namely acetone or formaldehyde over solid acid catalysts [49–52]. Based on the conversion
glycerol and selectivity to solketal, the clay catalyst which showed the optimum results was reported
of glycerol and selectivity to solketal, the clay catalyst which showed the optimum results was
by Timofeeva et al. in a batch reactor with activated catalyst by nitric acid of 0.5 M [53]. In the activated
reported by Timofeeva et al. in a batch reactor with activated catalyst by nitric acid of 0.5 M [53]. In
K10 montmorillonite by acid solution, this impact causes an increasing rate of reaction with the acid
the activated K10 montmorillonite by acid solution, this impact causes an increasing rate of reaction
site of the material. It is well-known that the acid activation of natural montmorillonite with nitric
with the acid site of the material. It is well-known that the acid activation of natural montmorillonite
acid can change the structure of montmorillonite (leaching of Al3+ cations from the octahedral to
with nitric acid can change the structure of montmorillonite (leaching of Al3+ cations from the
increase the surface area and microporosity of catalyst materials) [54–56]. The reaction of solketal
octahedral to increase the surface area and microporosity of catalyst materials) [54–56]. The reaction
production is shown in Figure 3. The use of formaldehyde as the major source of solketal production
of solketal production is shown in Figure 3. The use of formaldehyde as the major source of solketal
has a lower conversion value (only 83% glycerol conversion), with the K10 montmorillonite used as
production has a lower conversion value (only 83% glycerol conversion), with the K10
a catalyst. It may be due to the formation of the hemiacetal or hemicetal via two different pathways.
montmorillonite used as a catalyst. It may be due to the formation of the hemiacetal or hemicetal via
The reaction between glycerol and acetone is preferred as it produces a more stable intermediate,
two different pathways. The reaction between glycerol and acetone is preferred as it produces a more
stable intermediate, hemicetal compound, with a tertiary carbenium ion [37]. While, in the reaction
Energies 2019, 12, 2872 6 of 14

hemicetal compound, with a tertiary carbenium ion [37]. While, in the reaction between glycerol with
formaldehyde, the produced hemiacetal formation is not a stable carbenium ion. Thus, the conversion
value for the glycerol-formaldehyde system is relatively small as compared to the reaction where
acetone is used as a co-reactant [57–59].

Table 5. Glycerol-to-solketal over clay minerals.

Selectivity
Source Catalyst Condition Conversion Remark Ref
to Solketal
Montmorilonite
Glycerol Glycerol:acetone = 1:4, 10 mg of
modified by T = 25 ◦ C 94% 95.4% [53]
and acetone catalyst, time at 10 min
HNO3
Glycerol and K10 Glycerol:benzaldehyde dimethyl
T = 40 ◦ C 83% 99% [17]
benzaldehyde Montmorillonite acetal = 1:1.1 at 6 h.
Glycerol:acetone = 1:6,
Glycerol
K10 clays T = 30 ◦ C 87% 85% catalyst loading was 3 wt.% of total [60]
and acetone
reactant weight, time at 120 min
Glycerol and K10
T = 70 ◦ C 80% - Glycerol: formaldehyde = 1:1.2 [61]
formaldehyde Montmorillonite
Glycerol:acetone = 2:6, P=600 psi,
The amount of catalyst in each run
Glycerol K10
PEER REVIEWT = 40 C
◦ 69% 68% was determined by the selected [48]
Energies 2017, 10, x FOR
and acetone Montmorillonite 6 of 14
weight hourly space velocity
(WHSV) at 4 h−1
between glycerol with formaldehyde, the produced hemiacetal formation is not a stable carbenium
Note: glycerol to the second reactant ratio was presented as a molar ratio.
ion. Thus, the conversion value for the glycerol-formaldehyde system is relatively small as compared
to the reaction where acetone is used as a co-reactant [57–59].

Figure 3. Synthesis scheme of glycerol to solketal.

Figure 3. Synthesis scheme of glycerol to solketal.
Koranyi et al. (2012) [37] reported the effect of water as an impurity in the acetalization of glycerol.
Table 5. Glycerol-to-solketal over clay minerals.
The presence of water reduced the activity ca. 50% lower than the one with the model compound
(pure glycerol). A high
Source number ofCondition
Catalyst Brønsted and Lewis sitesSelectivity
Conversion does not correspond directly to a high
Remark Ref
to Solketal
activity. Dealumination FAU and Al-TUD-1 with a high Brønsted and Lewis acidity were poor in
Montmorilonite
Glycerol
the and
acetalization ofmodified
glycerol [37]. Hydrophobic catalysts, such95.4%
as hafniumGlycerol:acetone
and TUD-1= zirconium
1:4, 10 mg on
by T = 25 °C 94% [53]
acetone of catalyst, time at 10 min
TUD-1, are very prospective
HNO3 for glycerol to solketal. Ammaji et al. (2017) [62] also reported a similar
Glycerol and as the Zr-SBA-15
observation, K10 was Tthe most active83% catalyst. Glycerol:benzaldehyde
and selective99%
= 40 °C [17]
benzaldehyde Montmorillonite dimethyl acetal = 1:1.1 at 6 h.
5. Perspective on Ketalization of Glycerol over Hierarchical Zeolites Glycerol:acetone = 1:6,
Glycerol and catalyst loading was 3 wt.%
K10 clays T = 30 °C 87% 85% [60]
Dmitriev et al. (2016) [63] reported that zeolite beta was the mostofactive
acetone solid weight,
total reactant acid catalyst
as compared to amberlist-35 and cation-exchange resin (KU-2-8) [62]. The time at 120 min
zeolite beta applied
Glycerol and K10 Glycerol: formaldehyde =
was a commercial
formaldehyde
one from zeolyst with SiO2 /Al2 O3 of 25 and a zeolite beta made
Montmorillonite
T = 70 °C 80% -
1:1.2
by Angarsk. [61]
Kowalska et al. [64,65] studied the effect of (i) different zeolite topologies (MFI, BEA,
Glycerol:acetone = 2:6,and MOR),
P=600
(ii) Si/Al ratio from 9.2 to 25.8, and (iii) mesoporosity. Two parent MFI zeolites with different
psi, The amount of catalyst Si/Al
Glycerol
were and
applied (Si/Al =K10
12 and Si/Al T==27) [64]. The69%
hierarchical 68% in each
zeolites were run was determined
obtained by desilication
40 °C [48]
acetone Montmorillonite by the selected weight
using 0.2 M NaOH and dealumination using citric acid (0.5 M) and nitric acid (0.5 M). The
hourly space velocity
diffusion
limitation of the parent zeolites was considered as the highest activity of the parent MFI was
(WHSV) at 4 hsignificantly
−1

lower than the one from

Note: the hierarchical
glycerol MFI.
to the second A high
reactant selectivity
ratio (up toas100%)
was presented to ratio.
a molar solketal was obtained
with an acetone:glycerol ratio of 1. A higher acetone to glycerol ratio was obtained over a higher acetone
Koranyi et al. (2012) [37] reported the effect of water as an impurity in the acetalization of
to glycerol ratio. Both desilication and dealumination are very effective in improving the catalyst
glycerol. The presence of water reduced the activity ca. 50% lower than the one with the model
stability of zeolite based catalyst [66–68].
compound (pure glycerol). A high number of Brønsted and Lewis sites does not correspond directly
to a high activity. Dealumination FAU and Al-TUD-1 with a high Brønsted and Lewis acidity were
poor in the acetalization of glycerol [37]. Hydrophobic catalysts, such as hafnium and TUD-1
zirconium on TUD-1, are very prospective for glycerol to solketal. Ammaji et al. (2017) [62] also
reported a similar observation, as the Zr-SBA-15 was the most active and selective catalyst.
Energies 2019, 12, 2872 7 of 14

Rossa et al. [69] conducted the kinetics study of acetalization of glycerol with acetone to produce
solketal with optimization of the kinetics parameters. Zeolite beta with an Si/Al of 19 was applied
to find the best parameters: (i) External mass transfer (stirring rate), (ii) temperature, (iii) catalyst
amount, and (iv) glycerol to acetone ratio. The targeted goals were glycerol conversion and solketal
selectivity. The experimental design for beta zeolite showed that the suggested reaction parameters
are: Temperature at 60 ◦ C, stirring rate of 700 rpm, catalyst loading of 5%, and glycerol to acetone ratio
of 1:3. A higher acetone content will increase the conversion of glycerol [24,70]. However, an increase
of the acetone to glycerol ratio will increase the exergy destruction rate due to a reduction in the rate
of formation toward the product and a higher consumption of electrical exergy to the acetalization
reactor [20,71–80].
Hierarchical zeolite shows excellent glycerol conversion and selectivity to solketal through
acetalization reactions. The catalytic materials show a higher glycerol conversion (until more than
an 80% glycerol conversion) as compared to other porous and non-porous catalysts due to a large pore
size and easy molecular diffusivity. The enhancement of the catalytic activity of zeolites in glycerol
acetalization, through the generation of a hierarchical porosity, has been applied by different authors
as shown in Table 6. Based on the literature, the crystallite size was one of the most determining
factors in the activity of hierarchical zeolite as a catalyst [64,81–85]. The smaller the crystal size
of zeolite, the easier the diffusion of the reactant and products though the zeolite pores [73,86,87].
The pore structure of the zeolite can be changed through the dealumination and desilication processes.
The process not only can change the mesopore materials but also can increase the catalytic activity
(improving the accessibility and mass transfer on the surface) [88]. Hierarchical zeolites with different
topologies, such as ZSM-5 (MFI) [67,89,90], beta (BEA) [81,91,92], and Y (FAU) [64], have also been
used in the acetalization of glycerol, and the results show that smaller pores can produce high glycerol
conversion and selectivity to selectivity (almost 100% selective for solketal formation). However,
overall, all materials displayed very good catalytic performance when reacting equimolar mixtures of
glycerol and acetone [37,39]. From the experiments on H-beta zeolite, it was found that dealumination
resulted in a decrease of strong acid sites, thus decreasing the catalytic activity.

Table 6. Glycerol-to-solketal over hierarchical zeolite catalysts.

Selectivity
Source Catalyst Condition Conversion Remark Ref
to Solketal
hierarchical
(micro-mesoporous) Glycerol:acetone = 1:1, catalyst in
Glycerol
MFI zeolites T = 70 ◦ C 80% 100% the amount of 1% related [64]
and acetone
(pore diameter to glycerol.
0.51–0.55 nm)
Glycerol:acetone = 1:2, catalyst
Glycerol T = 28 ◦ C
H-B-1 zeolites 86% 98.5% amount = 5 wt.% referred to [73]
and acetone (room temperature)
glycerol in 1 h.
Glycerol Dealumination of Glycerol:acetone = 1:1, t = 30 min,
T = 30 ◦ C 80% 100% [72]
and acetone BEA Zeolites catalyst loading was 0.5 g
Glycerol H-Zeolite (pore size Glycerol:acetone = 1:3 were used
T = 70 ◦C 75% 92% [65]
and acetone 4.10 nm) with 0.05 g of catalyst for 2 h
Glycerol:acetone = 1:4, catalyst
Glycerol
H-BEA Zeolite T = 60 ◦ C 70% 97.9% amount was loading at 5 wt.% [69]
and acetone
for 1 h.

6. Solketal Synthesis over Carbon/Activated Carbon-Based Catalyst

Considering the abundant source of biomass as carbon and activated-carbon precursor, activated
carbons were functionalized with acid groups for solketal synthesis [93,94]. Some papers showed
the excellent performance of activated carbon for catalyzing the conversion of glycerol to solketal
(Table 7) and some of these exhibited a high activity and selectivity under green conditions (solvent-free
conditions at a mild temperature). The high surface area of activated carbon preserves the higher
surface acid sites by some modification, including acid, metal, and composite modifications [24,95–97].
Therefore, they are promising candidates as heterogeneous catalysts for the acetalization of acetone with
Energies 2019, 12, 2872 8 of 14

glycerol. From the utilization of acid functionalized activated carbon, the superior catalytic activity of
the four acid-treated carbons was underlined as compared to the untreated activated carbon, confirming
the importance of the higher number and strength of acid sites generated by the acid treatments.
The catalysts were prepared by HNO3 and H2 SO4 treatment to activated carbon. The catalytic
activity of the catalyst showed excellent performance due to the high conversion and selectivity
at room temperature.

Table 7. Glycerol-to-solketal over carbon/activated carbon-based catalyst.

Selectivity
Source Catalyst Condition Conversion Remark Ref
to Solketal
Room The highest number and strength
acid
acetone temperature, of acid sites generated by the acid
functionalized 97% 96% [39]
and glycerol glycerol to acetone treatments onto activated carbon
activated carbon
molar ratio of 1:4 gave better yield and selectivity
Graphene catalyst produced 76%
glycerol with
Graphene 100 ◦ C and 120 ◦ C 97% yield at 100 ◦ C and 85% yield [98]
benzaldehyde at
at 120 ◦ C, selectivity 100%
sulfonated acetone
acetone carbon-silica-meso and glycerol
82% 99% [99]
and glycerol composite molar ratio of 1:6,
materials re- fluxed at 70 ◦ C
acidic
acetone
carbon-based 80% 95% [93]
and glycerol
catalysts
Room
Ni-Zr supported Conversion and selectivity are
acetone Temperature
on mesoporous 75% 100% affected by glycerol/acetone ratio [100]
and glycerol glycerol/acetone
activated carbon and temperature
ratio of 1:10

From the acid-modified carbon catalyst, it was found that the presence of acid groups, mainly
sulfonic groups, was the key factor for the improved catalytic performance. A similar pattern also
appeared from the Ni-Zr support on the activated carbon [100], in which the active metal contributes
by enhancing the catalyst acidity. Another factor affecting the catalytic activity was the higher total
acid density, the large mesopore of the carbon structure, and the activity of the metals.

7. Perspective and Conclusions

This mini review highlighted the recent development on solid catalysts for the conversion of
glycerol-to-solketal. The product is an additive for fuels, which are very useful to reduce GHGs
and to improve the economic viability of biodiesel business [6,8,16,20,34,101–105]. Tailor-made
heterogeneous catalyst for an optimal conversion of glycerol is developed and required. Five major
heterogeneous catalysts were emphasized in this study: Resins, mesoporous silica, zeolites, clays,
and activated carbons. The stability of catalysts is one of the main hurdles for the commercialization
of glycerol to solketal. Even though the reaction temperature was considered as mild, the stability
of most of the solid catalysts decayed in the presence of water as a byproduct and other impurities
(NaCl, methanol) from the glycerol source. The deactivation rate is even higher when the raw glycerol
(contaminated with water) was fed to the reactor [106–109]. Therefore, the viability of the commercial
plant depends on (i) the source of feeds [110], (ii) availability of glycerol and other feeds, and (iii) cost
of glycerol as the feed. In general, at least three main challenges were identified:

a. The presence of water and impurities in the feed.

b. The shift from the batch reactor to the fixed bed reactor.
c. The presence of equilibrium offers other difficulties as higher acetone demand is expected.
However, higher acetone to glycerol will lead to destructive instruments.

Acidity is agreed as an important properties of zeolite catalysts for glycerol to solketal.

Strong acidity and medium hydrophobicity were expected in the design of the reactor. Based on
some limitations of the catalyst performance, the utilization of raw glycerol directly will reduce
Energies 2019, 12, 2872 9 of 14

the stability of the catalyst. This review described how a better material should be designed for
the optimum conversion of glycerol (and generally polyol) to solketal. Hydrophobic catalysts, such as
hafnium/TUD-1 and zirconium/TUD-1, are very prospective for glycerol to solketal. Extended works
on low aluminum mesoporous silica materials are expected in the coming years.

Author Contributions: I.F. and O.M. contribute to design and conception, drafting the article, and final approval
of the article. I.S., M.M.M., and G.F. contribute to collect the references, drafting the article, preparing all figures
and all tables, and discussion. T.M.I.M. contributes to help data analysis and discussion.
Funding: This research was funded by Ministry of Research, Technology and Higher Education
(KEMENRISTEKDIKTI) Republic of Indonesia through World Class Professor program in 2018, grant number:
123.6/D2.3/KP/2018. The APC was funded by the University of Technology Sydney seed fund (Org Unit 321740)
with Account number (2232397).
Acknowledgments: Authors would like to express appreciation for the support from Ministry of Research,
Technology and Higher Education (KEMENRISTEKDIKTI) Republic of Indonesia through World Class Professor
program in 2018, grant number: 123.6/D2.3/KP/2018. The authors are thankful to Professor Paolo Pescarmona
from University of Groningen for his rich suggestions on prospective catalytic materials in glycerol to solketal.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Silitonga, A.S.; Atabani, A.E.; Mahlia, T.M.I.; Masjuki, H.H.; Badruddin, I.A.; Mekhilef, S. A review on
prospect of Jatropha curcas for biodiesel in Indonesia. Renew. Sustain. Energy Rev. 2011, 15, 3733–3756.
[CrossRef]
2. Ong, H.C.; Masjuki, H.H.; Mahlia, T.M.I.; Silitonga, A.S.; Chong, W.T.; Leong, K.Y. Optimization of biodiesel
production and engine performance from high free fatty acid Calophyllum inophyllum oil in CI diesel
engine. Energy Convers. Manag. 2014, 81, 30–40. [CrossRef]
3. Galadima, A.; Muraza, O. Hydrothermal liquefaction of algae and bio-oil upgrading into liquid fuels: Role of
heterogeneous catalysts. Renew. Sustain. Energy Rev. 2018, 81, 1037–1048. [CrossRef]
4. Galadima, A.; Muraza, O. Biodiesel production from algae by using heterogeneous catalysts: A critical
review. Energy 2014, 78, 72–83. [CrossRef]
5. Silitonga, A.S.; Masjuki, H.H.; Ong, H.C.; Sebayang, A.H.; Dharma, S.; Kusumo, F.; Siswantoro, J.;
Milano, J.; Daud, K.; Mahlia, T.M.I.; et al. Evaluation of the engine performance and exhaust emissions of
biodiesel-bioethanol-diesel blends using kernel-based extreme learning machine. Energy 2018, 159, 1075–1087.
[CrossRef]
6. Quispe, C.A.G.; Coronado, C.J.R.; Carvalho, J.A., Jr. Glycerol: Production, consumption, prices,
characterization and new trends in combustion. Renew. Sustain. Energy Rev. 2013, 27, 475–493. [CrossRef]
7. Monteiro, M.R.; Kugelmeier, C.L.; Pinheiro, R.S.; Batalha, M.O.; Da Silva César, A. Glycerol from biodiesel
production: Technological paths for sustainability. Renew. Sustain. Energy Rev. 2018, 88, 109–122. [CrossRef]
8. Dharma, S.; Masjuki, H.H.; Ong, H.C.; Sebayang, A.H.; Silitonga, A.S.; Kusumo, F.; Mahlia, T.M.I. Optimization
of biodiesel production process for mixed Jatropha curcas-Ceiba pentandra biodiesel using response surface
methodology. Energy Convers. Manage. 2016, 115, 178–190. [CrossRef]
9. Zakaria, Z.Y.; Linnekoski, J.; Amin, N.A.S. Catalyst screening for conversion of glycerol to light olefins.
Chem. Eng. J. 2012, 207, 803–813. [CrossRef]
10. Galadima, A.; Muraza, O. A review on glycerol valorization to acrolein over solid acid catalysts. J. Taiwan
Inst. Chem. Eng. 2016, 67, 29–44. [CrossRef]
11. Galadima, A.; Muraza, O. Sustainable Production of Glycerol Carbonate from By-product in Biodiesel Plant.
Waste Biomass Valorization 2016, 8, 141–152. [CrossRef]
12. Bagheri, S.; Julkapli, N.M.; Yehye, W.A. Catalytic conversion of biodiesel derived raw glycerol to value added
products. Renew. Sustain. Energy Rev. 2015, 41, 113–127. [CrossRef]
13. Rodrigues, A.; Bordado, J.C.; Santos, R.G.D. Upgrading the Glycerol from Biodiesel Production as a Source
of Energy Carriers and Chemicals—A Technological Review for Three Chemical Pathways. Energies 2017,
10, 1817. [CrossRef]
Energies 2019, 12, 2872 10 of 14

14. Smith, P.J.; Smith, L.; Dummer, N.F.; Douthwaite, M.; Willock, D.J.; Howard, M.; Knight, D.W.; Taylor, S.H.;
Hutchings, G.J. Investigating the Influence of Reaction Conditions and the Properties of Ceria for
the Valorisation of Glycerol. Energies 2019, 12, 1359. [CrossRef]
15. Rahmat, N.; Abdullah, A.Z.; Mohamed, A.R. Recent progress on innovative and potential technologies for
glycerol transformation into fuel additives: A critical review. Renew. Sustain. Energy Rev. 2010, 14, 987–1000.
[CrossRef]
16. Cornejo, A.; Barrio, I.; Campoy, M.; Lázaro, J.; Navarrete, B. Oxygenated fuel additives from glycerol
valorization. Main production pathways and effects on fuel properties and engine performance: A critical
review. Renew. Sustain. Energy Rev. 2017, 79, 1400–1413. [CrossRef]
17. Nanda, M.R.; Zhang, Y.; Yuan, Z.; Qin, W.; Ghaziaskar, H.S.; Xu, C. Catalytic conversion of glycerol for sustainable
production of solketal as a fuel additive: A review. Renew. Sustain. Energy Rev. 2016, 56, 1022–1031. [CrossRef]
18. Mahdi, H.I.; Irawan, E.; Nuryoto, N.; Jayanudin, J.; Sulistyo, H.; Sediawan, W.B.; Muraza, O. Glycerol
Carbonate Production from Biodiesel Waste Over Modified Natural Clinoptilolite. Waste Biomass Valorization
2016, 7, 1349–1356. [CrossRef]
19. Alptekin, E.; Canakci, M. Performance and emission characteristics of solketal-gasoline fuel blend in a vehicle
with spark ignition engine. Appl. Therm. Eng. 2017, 124, 504–509. [CrossRef]
20. Aghbashlo, M.; Hosseinpour, S.; Tabatabaei, M.; Rastegari, H.; Ghaziaskar, H.S. Multi-objective
exergoeconomic and exergoenvironmental optimization of continuous synthesis of solketal through glycerol
ketalization with acetone in the presence of ethanol as co-solvent. Renew. Energy 2019, 130, 735–748.
[CrossRef]
21. Ayoub, M.; Abdullah, A.Z. Diglycerol synthesis via solvent-free selective glycerol etherification process over
lithium-modified clay catalyst. Chem. Eng. J. 2013, 225, 784–789. [CrossRef]
22. Konaka, A.; Tago, T.; Yoshikawa, T.; Shitara, H.; Nakasaka, Y.; Masuda, T. Conversion of Biodiesel-Derived
Crude Glycerol into Useful Chemicals over a Zirconia–Iron Oxide Catalyst. Ind. Eng. Chem. Res. 2013,
52, 15509–15515. [CrossRef]
23. Sánchez, G.; Dlugogorski, B.Z.; Kennedy, E.M.; Stockenhuber, M. Zeolite-supported iron catalysts for allyl
alcohol synthesis from glycerol. Appl. Catal. A Gen. 2016, 509, 130–142. [CrossRef]
24. Manikandan, M.; Prabu, M.; Sk, A.K.; Sangeetha, P.; Vijayaraghavan, R. Tuning the basicity of Cu-based
mixed oxide catalysts towards the efficient conversion of glycerol to glycerol carbonate. Mol. Catal. 2018,
460, 53–62. [CrossRef]
25. Gadamsetti, S.; Rajan, N.P.; Rao, G.S.; Chary, K.V.R. Acetalization of glycerol with acetone to bio fuel additives
over supported molybdenum phosphate catalysts. J. Mol. Catal. A Chem. 2015, 410, 49–57. [CrossRef]
26. Serafim, H.; Fonseca, I.M.; Ramos, A.M.; Vital, J.; Castanheiro, J.E. Valorization of glycerol into fuel additives
over zeolites as catalysts. Chem. Eng. J. 2011, 178, 291–296. [CrossRef]
27. Yamamoto, K.; Kiyan, A.M.; Bagio, J.C.; Rossi, K.A.B.; Delabio Berezuk, F.; Berezuk, M.E. Green cyclic acetals
production by glycerol etherification reaction with benzaldehyde using cationic acidic resin. Green Process.
Synth. 2019, 8, 183–190. [CrossRef]
28. Sudarsanam, P.; Mallesham, B.; Prasad, A.N.; Reddy, P.S.; Reddy, B.M. Synthesis of bio–additive fuels
from acetalization of glycerol with benzaldehyde over molybdenum promoted green solid acid catalysts.
Fuel Process. Technol. 2013, 106, 539–545. [CrossRef]
29. Mallesham, B.; Sudarsanam, P.; Raju, G.; Reddy, B.M. Design of highly efficient Mo and W-promoted
SnO2 solid acids for heterogeneous catalysis: Acetalization of bio-glycerol. Green Chem. 2013, 15, 478–489.
[CrossRef]
30. Manjunathan, P.; Marakatti, V.S.; Chandra, P.; Kulal, A.B.; Umbarkar, S.B.; Ravishankar, R.; Shanbhag, G.V.
Mesoporous tin oxide: An efficient catalyst with versatile applications in acid and oxidation catalysis.
Catal. Today 2018, 309, 61–76. [CrossRef]
31. Stawicka, K.; Díaz-Álvarez, A.E.; Calvino-Casilda, V.; Trejda, M.; Bañares, M.A.; Ziolek, M. The Role of
Brønsted and Lewis Acid Sites in Acetalization of Glycerol over Modified Mesoporous Cellular Foams.
J. Phys. Chem. C 2016, 120, 16699–16711. [CrossRef]
32. Oliveira, P.A.; Souza, R.O.M.A.; Mota, C.J.A. Atmospheric Pressure Continuous Production of Solketal from
the Acid-Catalyzed Reaction of Glycerol with Acetone. J. Braz. Chem. Soc. 2016, 27, 1832–1837. [CrossRef]
33. Ilgen, O.; Yerlikaya, S.; Akyurek, F.O. Synthesis of Solketal from Glycerol and Acetone over Amberlyst-46 to
Produce an Oxygenated Fuel Additive. Period. Polytech. Chem. Eng. 2016, 61, 144–148. [CrossRef]
Energies 2019, 12, 2872 11 of 14

34. Wu, L.; Moteki, T.; Gokhale, A.A.; Flaherty, D.W.; Toste, F.D. Production of Fuels and Chemicals from
Biomass: Condensation Reactions and Beyond. Chem 2016, 1, 32–58. [CrossRef]
35. Yang, J.; Li, N.; Ma, W.J.; Zhou, J.H.; Sun, H.Z. Synthesis of Solketal with Catalyst Sulfonic Acid Resin.
Adv. Mater. Res. 2014, 830, 176–179. [CrossRef]
36. Guidi, S.; Noe, M.; Riello, P.; Perosa, A.; Selva, M. Towards a Rational Design of a Continuous-Flow Method
for the Acetalization of Crude Glycerol: Scope and Limitations of Commercial Amberlyst 36 and AlF3 .3H2 O
as Model Catalysts. Molecules 2016, 21, 657. [CrossRef]
37. Li, L.; Korányi, T.I.; Sels, B.F.; Pescarmona, P.P. Highly-efficient conversion of glycerol to solketal over
heterogeneous Lewis acid catalysts. Green Chem. 2012, 14, 1611–1619. [CrossRef]
38. Abreu, T.H.; Meyer, C.I.; Padró, C.; Martins, L. Acidic V-MCM-41 catalysts for the liquid-phase ketalization
of glycerol with acetone. Microporous Mesoporous Mater. 2019, 273, 219–225. [CrossRef]
39. Rodrigues, R.; Gonçalves, M.; Mandelli, D.; Pescarmona, P.P.; Carvalho, W.A. Solvent-free conversion of
glycerol to solketal catalysed by activated carbons functionalised with acid groups. Catal. Sci. Technol. 2014,
4, 2293–2301. [CrossRef]
40. Nasser, G.A.; Kurniawan, T.; Tago, T.; Bakare, I.A.; Taniguchi, T.; Nakasaka, Y.; Masuda, T.; Muraza, O.
Cracking of n-hexane over hierarchical MOR zeolites derived from natural minerals. J. Taiwan Inst. Chem. Eng.
2016, 61, 20–25. [CrossRef]
41. Ahmed, M.H.M.; Muraza, O.; Al-Amer, A.M.; Miyake, K.; Nishiyama, N. Development of hierarchical EU-1
zeolite by sequential alkaline and acid treatments for selective dimethyl ether to propylene (DTP). Appl. Catal.
A Gen. 2015, 497, 127–134. [CrossRef]
42. Chen, L.; Nohair, B.; Zhao, D.; Kaliaguine, S. Highly Efficient Glycerol Acetalization over Supported
Heteropoly Acid Catalysts. ChemCatChem 2018, 10, 1918–1925. [CrossRef]
43. Li, R.; Song, H.; Chen, J. Propylsulfonic Acid Functionalized SBA-15 Mesoporous Silica as Efficient Catalysts
for the Acetalization of Glycerol. Catalysts 2018, 8, 297. [CrossRef]
44. Vicente, G.; Melero, J.A.; Morales, G.; Paniagua, M.; Martín, E. Acetalisation of bio-glycerol with acetone to
produce solketal over sulfonic mesostructured silicas. Green Chem. 2010, 12, 899. [CrossRef]
45. Morales, G.; Paniagua, M.; Melero, J.A.; Vicente, G.; Ochoa, C. Sulfonic Acid-Functionalized Catalysts
for the Valorization of Glycerol via Transesterification with Methyl Acetate. Ind. Eng. Chem. Res. 2011,
50, 5898–5906. [CrossRef]
46. Churipard, S.R.; Manjunathan, P.; Chandra, P.; Shanbhag, G.V.; Ravishankar, R.; Rao, P.V.C.; Sri Ganesh, G.;
Halgeri, A.B.; Maradur, S.P. Remarkable catalytic activity of a sulfonated mesoporous polymer (MP-SO3 H)
for the synthesis of solketal at room temperature. New J. Chem. 2017, 41, 5745–5751. [CrossRef]
47. Konwar, L.J.; Samikannu, A.; Mäki-Arvela, P.; Boström, D.; Mikkola, J.-P. Lignosulfonate-based
macro/mesoporous solid protonic acids for acetalization of glycerol to bio-additives. Appl. Catal. B Environ.
2018, 220, 314–323. [CrossRef]
48. Nanda, M.R.; Yuan, Z.; Qin, W.; Ghaziaskar, H.S.; Poirier, M.-A.; Xu, C. A new continuous-flow process for
catalytic conversion of glycerol to oxygenated fuel additive: Catalyst screening. Appl. Energy 2014, 123, 75–81.
[CrossRef]
49. Mota, C.J.A.; Da Silva, C.X.A.; Rosenbach, N.; Costa, J.; Da Silva, F. Glycerin Derivatives as Fuel Additives:
The Addition of Glycerol/Acetone Ketal (Solketal) in Gasolines. Energy Fuels 2010, 24, 2733–2736. [CrossRef]
50. Lin, C.-Y.; Tsai, S.-M. Comparison of Fuel Properties of Nanoemulsions of Diesel Fuel Dispersed with Solketal
by Microwave Irradiation and Mechanical Homogenization Methods. Energy Fuels 2018, 32, 11814–11820.
[CrossRef]
51. Esteban, J.; Murasiewicz, H.; Simons, T.A.H.; Bakalis, S.; Fryer, P.J. Measuring the Density, Viscosity, Surface
Tension, and Refractive Index of Binary Mixtures of Cetane with Solketal, a Novel Fuel Additive. Energy Fuels
2016, 30, 7452–7459. [CrossRef]
52. Laskar, I.B.; Rajkumari, K.; Gupta, R.; Rokhum, L. Acid functionalized mesoporous polymer catalyzed
acetalization of glycerol to solketal, a potential fuel additive under solvent-free conditions. Energy Fuels 2018,
32, 12567–12576. [CrossRef]
53. Timofeeva, M.N.; Panchenko, V.N.; Krupskaya, V.V.; Gil, A.; Vicente, M.A. Effect of nitric acid modification
of montmorillonite clay on synthesis of solketal from glycerol and acetone. Catal. Commun. 2017, 90, 65–69.
[CrossRef]
Energies 2019, 12, 2872 12 of 14

54. Fatimah, I.; Wang, S.; Wulandari, D. ZnO/montmorillonite for photocatalytic and photochemical degradation
of methylene blue. Appl. Clay Sci. 2011, 53, 553–560. [CrossRef]
55. Fatimah, I.; Wang, S.; Narsito; Wijaya, K. Composites of TiO2 -aluminum pillared montmorillonite: Synthesis,
characterization and photocatalytic degradation of methylene blue. Appl. Clay Sci. 2010, 50, 588–593.
[CrossRef]
56. Fatimah, I.; Huda, T. Preparation of cetyltrimethylammonium intercalated Indonesian montmorillonite for
adsorption of toluene. Appl. Clay Sci. 2013, 74, 115–120. [CrossRef]
57. Sonar, S.K.; Shinde, A.S.; Asok, A.; Niphadkar, P.S.; Mayadevi, S.; Joshi, P.N.; Bokade, V.V. Solvent free
acetalization of glycerol with formaldehyde over hierarchical zeolite of BEA topology. Environ. Prog.
Sustain. Energy 2018, 37, 797–807. [CrossRef]
58. Mallesham, B.; Rao, B.G.; Reddy, B.M. Production of biofuel additives by esterification and acetalization of
bioglycerol. C. R. Chim. 2016, 19, 1194–1202. [CrossRef]
59. Da Silva, C.X.A.; Gonçalves, V.L.C.; Mota, C.J.A. Water-tolerant zeolitecatalyst for the acetalisation of glycerol.
Green Chem. 2009, 11, 38–41. [CrossRef]
60. Sandesh, S.; Halgeri, A.B.; Shanbhag, G.V. Utilization of renewable resources: Condensation of glycerol with
acetone at room temperature catalyzed by organic–inorganic hybrid catalyst. J. Mol. Catal. A Chem. 2015,
401, 73–80. [CrossRef]
61. Hasabnis, A.; Mahajani, S. Acetalization of Glycerol with Formaldehyde by Reactive Distillation. Ind. Eng.
Chem. Res. 2014, 53, 12279–12287. [CrossRef]
62. Ammaji, S.; Rao, G.S.; Chary, K.V.R. Acetalization of glycerol with acetone over various metal-modified
SBA-15 catalysts. Appl. Petrochem. Res. 2018, 8, 107–118. [CrossRef]
63. Dmitriev, G.S.; Terekhov, A.V.; Zanaveskin, L.N.; Khadzhiev, S.N.; Zanaveskin, K.L.; Maksimov, A.L. Choice
of a catalyst and technological scheme for synthesis of solketal. Rus. J. Appl. Chem. 2017, 89, 1619–1624.
[CrossRef]
64. Kowalska-Kus, J.; Held, A.; Frankowski, M.; Nowinska, K. Solketal formation from glycerol and acetone over
hierarchical zeolites of different structure as catalysts. J. Mol. Catal. A Chem. 2017, 426, 205–212. [CrossRef]
65. Kowalska-Kus, J.; Held, A.; Nowinska, K. Enhancement of the catalytic activity of H-ZSM-5 zeolites for
glycerol acetalization by mechanical grinding. React. Kinet. Mech. Catal. 2015, 117, 341–352. [CrossRef]
66. Ahmed, M.H.M.; Muraza, O.; Yoshioka, M.; Yokoi, T. Effect of multi-step desilication and dealumination
treatments on the performance of hierarchical EU-1 zeolite for converting methanol to olefins.
Microporous Mesoporous Mater. 2017, 241, 79–88. [CrossRef]
67. Muraza, O.; Bakare, I.A.; Tago, T.; Konno, H.; Taniguchi, T.; Al-Amer, A.M.; Yamani, Z.H.; Nakasaka, Y.;
Masuda, T. Selective catalytic cracking of n-hexane to propylene over hierarchical MTT zeolite. Fuel 2014,
135, 105–111. [CrossRef]
68. Ahmed, M.H.M.; Muraza, O.; Nakaoka, S.; Jamil, A.K.; Mayoral, A.; Sebastian, V.; Yamani, Z.H.; Masuda, T.
Stability Assessment of Regenerated Hierarchical ZSM-48 Zeolite Designed by Post-Synthesis Treatment for
Catalytic Cracking of Light Naphtha. Energy Fuels 2017, 31, 14097–14103. [CrossRef]
69. Rossa, V.; Pessanha, Y.D.S.P.; Díaz, G.C.; Câmara, L.D.T.; Pergher, S.B.C.; Aranda, D.A.G. Reaction Kinetic
Study of Solketal Production from Glycerol Ketalization with Acetone. Ind. Eng. Chem. Res. 2017, 56, 479–488.
[CrossRef]
70. Chen, Z.; Gao, L.; Han, W.; Zhang, L. Energy and exergy analyses of coal gasification with supercritical water
and O2 -H2 O. Appl. Therm. Eng. 2019, 148, 57–63. [CrossRef]
71. Aghbashlo, M.; Tabatabaei, M.; Hosseinpour, S.; Rastegari, H.; Ghaziaskar, H.S. Multi-objective exergy-based
optimization of continuous glycerol ketalization to synthesize solketal as a biodiesel additive in subcritical
acetone. Energy Convers. Manag. 2018, 160, 251–261. [CrossRef]
72. Venkatesha, N.J.; Bhat, Y.S.; Prakash, B.S.J. Dealuminated BEA zeolite for selective synthesis of five-membered
cyclic acetal from glycerol under ambient conditions. RSC Adv. 2016, 6, 18824–18833. [CrossRef]
73. Manjunathan, P.; Maradur, S.P.; Halgeri, A.B.; Shanbhag, G.V. Room temperature synthesis of solketal
from acetalization of glycerol with acetone: Effect of crystallite size and the role of acidity of beta zeolite.
J. Mol. Catal. A Chem. 2015, 396, 47–54. [CrossRef]
74. Aghbashlo, M.; Tabatabaei, M.; Rastegari, H.; Ghaziaskar, H.S.; Shojaei, T.R. On the exergetic optimization of
solketalacetin synthesis as a green fuel additive through ketalization of glycerol-derived monoacetin with
acetone. Renew. Energy 2018, 126, 242–253. [CrossRef]
Energies 2019, 12, 2872 13 of 14

75. Gholami, A.; Hajinezhad, A.; Pourfayaz, F.; Ahmadi, M.H. The effect of hydrodynamic and ultrasonic
cavitation on biodiesel production: An exergy analysis approach. Energy 2018, 160, 478–489. [CrossRef]
76. Ortiz, F.G.; Ollero, P.; Serrera, A.; Galera, S. Process integration and exergy analysis of the autothermal
reforming of glycerol using supercritical water. Energy 2012, 42, 192–203. [CrossRef]
77. Ortiz, F.G.; Ollero, P.; Serrera, A.; Galera, S. An energy and exergy analysis of the supercritical water reforming
of glycerol for power production. Int. J. Hydrogen Energy 2012, 37, 209–226. [CrossRef]
78. Hajjaji, N.; Baccar, I.; Pons, M.-N. Energy and exergy analysis as tools for optimization of hydrogen production
by glycerol autothermal reforming. Renew. Energy 2014, 71, 368–380. [CrossRef]
79. Presciutti, A.; Asdrubali, F.; Baldinelli, G.; Rotili, A.; Malavasi, M.; Di Salvia, G. Energy and exergy analysis of
glycerol combustion in an innovative flameless power plant. J. Clean. Prod. 2018, 172, 3817–3824. [CrossRef]
80. Antonova, Z.A.; Krouk, V.S.; Pilyuk, Y.E.; Maksimuk, Y.V.; Karpushenkava, L.S.; Krivova, M.G. Exergy
analysis of canola-based biodiesel production in Belarus. Fuel Process. Technol. 2015, 138, 397–403. [CrossRef]
81. Ahmed, M.H.M.; Muraza, O.; Galadima, A.; Yoshioka, M.; Yamani, Z.H.; Yokoi, T. Choreographing
boron-aluminum acidity and hierarchical porosity in *BEA zeolite by in-situ hydrothermal synthesis for
a highly selective methanol to propylene catalyst. Microporous Mesoporous Mater. 2019, 273, 249–255.
[CrossRef]
82. Galadima, A.; Muraza, O. Hydrocracking catalysts based on hierarchical zeolites: A recent progress. J. Ind.
Eng. Chem. 2018, 61, 265–280. [CrossRef]
83. Yang, X.; Wang, F.; Wei, R.; Li, S.; Wu, Y.; Shen, P.; Wang, H.; Gao, L.; Xiao, G. Synergy effect between
hierarchical structured and Sn-modified H[Sn, Al]ZSM-5 zeolites on the catalysts for glycerol aromatization.
Microporous Mesoporous Mater. 2018, 257, 154–161. [CrossRef]
84. Feliczak-Guzik, A. Hierarchical zeolites: Synthesis and catalytic properties. Microporous Mesoporous Mater.
2018, 259, 33–45. [CrossRef]
85. Possato, L.G.; Chaves, T.F.; Cassinelli, W.H.; Pulcinelli, S.H.; Santilli, C.V.; Martins, L. The multiple
benefits of glycerol conversion to acrolein and acrylic acid catalyzed by vanadium oxides supported on
micro-mesoporous MFI zeolites. Catal. Today 2017, 289, 20–28. [CrossRef]
86. Huang, G.; Ji, P.; Xu, H.; Jiang, J.-G.; Chen, L.; Wu, P. Fast synthesis of hierarchical Beta zeolites with uniform
nanocrystals from layered silicate precursor. Microporous Mesoporous Mater. 2017, 248, 30–39. [CrossRef]
87. Yu, Q.; Cui, C.; Zhang, Q.; Chen, J.; Li, Y.; Sun, J.; Li, C.; Cui, Q.; Yang, C.; Shan, H. Hierarchical ZSM-11 with
intergrowth structures: Synthesis, characterization and catalytic properties. J. Energy Chem. 2013, 22, 761–768.
[CrossRef]
88. Galadima, A.; Muraza, O. In situ fast pyrolysis of biomass with zeolite catalysts for bioaromatics/gasoline
production: A review. Energy Convers. Manag. 2015, 105, 338–354. [CrossRef]
89. Bawah, A.-R.; Malaibari, Z.O.; Muraza, O. Syngas production from CO2 reforming of methane over Ni
supported on hierarchical silicalite-1 fabricated by microwave-assisted hydrothermal synthesis. Int. J.
Hydrogen Energy 2018, 43, 13177–13189. [CrossRef]
90. Muraza, O.; Bakare, I.A.; Tago, T.; Konno, H.; Adedigba, A.-L.; Al-Amer, A.M.; Yamani, Z.H.; Masuda, T.
Controlled and rapid growth of MTT zeolite crystals with low-aspect-ratio in a microwave reactor. Chem. Eng. J.
2013, 226, 367–376. [CrossRef]
91. El Hanache, L.; Lebeau, B.; Nouali, H.; Toufaily, J.; Hamieh, T.; Daou, T.J. Performance of surfactant-modified
*BEA-type zeolite nanosponges for the removal of nitrate in contaminated water: Effect of the external
surface. J. Hazard. Mater. 2019, 364, 206–217. [CrossRef] [PubMed]
92. Sammoury, H.; Toufaily, J.; Cherry, K.; Hamieh, T.; Pouilloux, Y.; Pinard, L. Desilication of *BEA zeolites
using different alkaline media: Impact on catalytic cracking of n-hexane. Microporous Mesoporous Mater. 2018,
267, 150–163. [CrossRef]
93. Gonçalves, M.; Rodrigues, R.; Galhardo, T.S.; Carvalho, W.A. Highly selective acetalization of glycerol with
acetone to solketal over acidic carbon-based catalysts from biodiesel waste. Fuel 2016, 181, 46–54. [CrossRef]
94. Li, Z.; Miao, Z.; Wang, X.; Zhao, J.; Zhou, J.; Si, W.; Zhuo, S. One-pot synthesis of ZrMo-KIT-6 solid acid
catalyst for solvent-free conversion of glycerol to solketal. Fuel 2018, 233, 377–387. [CrossRef]
95. Liu, J.; Li, Y.; Liu, H.; He, D. Transformation of CO2 and glycerol to glycerol carbonate over CeO2 ZrO2 solid
solution—Effect of Zr doping. Biomass Bioenergy 2018, 118, 74–83. [CrossRef]
96. Malaika, A.; Kozłowski, M. Glycerol conversion towards valuable fuel blending compounds with the assistance of
SO3H-functionalized carbon xerogels and spheres. Fuel Process. Technol. 2019, 184, 19–26. [CrossRef]
Energies 2019, 12, 2872 14 of 14

97. Wan, Y.; Lei, Y.; Lan, G.; Liu, D.; Li, G.; Bai, R. Synthesis of glycerol carbonate from glycerol and dimethyl
carbonate over DABCO embedded porous organic polymer as a bifunctional and robust catalyst. Appl. Catal.
A Gen. 2018, 562, 267–275. [CrossRef]
98. Oger, N.; Lin, Y.F.; Le Grognec, E.; Rataboul, F.; Felpin, F.-X. Graphene-promoted acetalisation of glycerol
under acid-free conditions. Green Chem. 2016, 18, 1531–1537. [CrossRef]
99. Nandan, D.; Sreenivasulu, P.; Konathala, L.S.; Kumar, M.; Viswanadham, N. Acid functionalized carbon–silica
composite and its application for solketal production. Microporous Mesoporous Mater. 2013, 179, 182–190.
[CrossRef]
100. Khayoon, M.S.; Hameed, B.H. Solventless acetalization of glycerol with acetone to fuel oxygenates over
Ni–Zr supported on mesoporous activated carbon catalyst. Appl. Catal. A Gen. 2013, 464, 191–199. [CrossRef]
101. Talebian-Kiakalaieh, A.; Amin, N.A.S.; Rajaei, K.; Tarighi, S. Oxidation of bio-renewable glycerol to
value-added chemicals through catalytic and electro-chemical processes. Appl. Energy 2018, 230, 1347–1379.
[CrossRef]
102. Chol, C.G.; Dhabhai, R.; Dalai, A.K.; Reaney, M. Purification of crude glycerol derived from biodiesel
production process: Experimental studies and techno-economic analyses. Fuel Process. Technol. 2018,
178, 78–87. [CrossRef]
103. Ismail, M.S.; Moghavvemi, M.; Mahlia, T.M.I. Techno-economic analysis of an optimized photovoltaic
and diesel generator hybrid power system for remote houses in a tropical climate. Energy Convers. Manag.
2013, 69, 163–173. [CrossRef]
104. Fantozzi, F.; Frassoldati, A.; Bartocci, P.; Cinti, G.; Quagliarini, F.; Bidini, G.; Ranzi, E.M. An experimental
and kinetic modeling study of glycerol pyrolysis. Appl. Energy 2016, 184, 68–76. [CrossRef]
105. Norhasyima, R.S.; Mahlia, T.M.I. Advances in CO2 utilization technology: A patent landscape review. J. CO2 Util. 2018,
26, 323–335. [CrossRef]
106. Eze, V.C.; Harvey, A.P. Continuous reactive coupling of glycerol and acetone—A strategy for triglyceride
transesterification and in-situ valorisation of glycerol by-product. Chem. Eng. J. 2018, 347, 41–51. [CrossRef]
107. Jamil, F.; Saxena, S.K.; Al-Muhtaseb, A.A.H.; Baawain, M.; Al-Abri, M.; Viswanadham, N.; Kumar, G.;
Abu-Jrai, A.M. Valorization of waste “date seeds” bio-glycerol for synthesizing oxidative green fuel additive.
J. Clean. Prod. 2017, 165, 1090–1096. [CrossRef]
108. Priya, S.S.; Selvakannan, P.R.; Chary, K.V.R.; Kantam, M.L.; Bhargava, S.K. Solvent-free microwave-assisted
synthesis of solketal from glycerol using transition metal ions promoted mordenite solid acid catalysts.
Mol. Catal. 2017, 434, 184–193. [CrossRef]
109. Trifoi, A.R.; Agachi, P.Ş.; Pap, T. Glycerol acetals and ketals as possible diesel additives. A review of their
synthesis protocols. Renew. Sustain. Energy Rev. 2016, 62, 804–814. [CrossRef]
110. Remón, J.; Ruiz, J.; Oliva, M.; García, L.; Arauzo, J. Effect of biodiesel-derived impurities (acetic acid, methanol
and potassium hydroxide) on the aqueous phase reforming of glycerol. Chem. Eng. J. 2016, 299, 431–448.
[CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
 Applied Catalysis A, General 581 (2019) 1–10

Contents lists available at ScienceDirect

Applied Catalysis A, General

journal homepage: www.elsevier.com/locate/apcata

Modification of silicalite-1 with ammonium compounds aimed at T

preparation of acidic catalyst for acetalization of glycerol with acetone
⁎ ⁎
E. Janiszewskaa, , J. Kowalska-Kuśa, , K. Góra-Marekb, A. Szymochac,d, K. Nowińskaa,
S. Kowalaka
a
Adam Mickiewicz University in Poznan, Faculty of Chemistry, Umultowska 89b, Uniwersytetu Poznańskiego 8 61-614, Poznan, Poland
b
Jagiellonian University in Krakow, Faculty of Chemistry, Gronostajowa 2, 30-387, Krakow, Poland
c
University of Agriculture in Krakow, Al. Mickiewicza 21, 31-120, Krakow, Poland
d
Institute of Nuclear Physics PAN, Radzikowskiego 152, 31-342, Krakow, Poland

A R T I C LE I N FO A B S T R A C T

Keywords: Post-synthesis modification of silicalite-1 (MFI) with solutions of ammonium compounds (NH4F, NH4Cl,
Modified silicalite-1 NH4NO3, NH4OH) without any additional alkaline agent followed by further thermal treatment appeared a new,
Defective sites facile, and cheap method to generate the acidic sites in starting material. The number and strength of resulting
Surface acidity acid sites depend on the nature of anion in the applied ammonium salt. Used procedure resulted in partial
Glycerol acetalization
removal of external silanol groups and in the formation of acidic internal isolated as well as hydrogen bonded
OH groups (e.g. silanol nests). The novel acidic sites show a high catalytic activity for acetalization of glycerol
with acetone and high selectivity towards 2,2-dimethyl-1,3-dioxolane-4-methanol (solketal). The modified si-
licalite-1 indicates also high stability with time on stream. The proposed reaction mechanism considers a
combined contribution of both Lewis and very weak Brønsted acid sites.

1. Introduction modification with an aqueous solution of ammonium compounds with

different pH (NH4NO3/NH4OH or NH4Cl) influenced the nature,
Silicalite-1, the alumina-free MFI structure, known of its rather number and strength of acidic sites generated in silicalite-1. Kitamura
negligible acidity, recently gains an interest as a potential catalyst. Even et al. [11] has shown that treatment of silicalite of MFI structure, with
though the acidity of these materials is low, they show relatively high aqueous solution of ammonium salts and at least one basic compound
activity for the reactions requiring weak acidic sites such as toluene (e.g. aqueous ammonia solution, alkylamines, allylamines or alky-
alkylation [1], propionic acid ketonization [2], Beckmann rearrange- lammonium hydroxide) generates acid sites, efficient in the catalytic
ment [3] and some others [4,5]. Activity of silicalite-1 in acid catalyzed Beckmann rearrangement reaction. The increase in the activity was
reaction is associated with the formation of framework defects (silanol connected with the formation of H-bonded silanol defect groups of a
nests) created by removal of some silicon atom [2]. The remaining different kind indicating an acidic nature (Scheme 1) [10,12]. It was
oxygen atoms interact with hydrogen atoms forming the silanol nests also shown that the treatment of zeolites with a small amount of NH4F
(Scheme 1). According to some authors, the silanol nests behave as with subsequent calcination to introduce the fluorine atoms resulted in
weak Brønsted sites [6,7] in contrast to isolated external hydroxyl increased acid strength of existing acid sites, due to the inductive effects
groups, which are not acidic. Syntheses of silicalite-1 presented in the of highly electronegative fluorine, and influenced their catalytic ac-
literature were performed under various conditions [8] and in the tivity [13]. Recently, the treatment of zeolites with concentrated solu-
presence of diverse modifiers, which influenced the nature of resulting tion of NH4F allowed to increase the micropore volume in zeolite with
silicalite-1 surface active sites, including the acidic ones [9]. Prepara- small cages (sodalite) by opening these cages [14] or to obtain the
tion of silicalite-1 in the presence of fluoride medium (NH4F used as one hierarchical zeolite [15,16] with improved catalytic activity due to
of the reactants for silicalite-1 synthesis) resulted in negligible amounts increased accessibility to the active sites. The advantage of such method
of defects, whereas hydroxide route causes to obtain silicalite-1 with a of preparation of hierarchical zeolites is lack of change of framework
higher amount of defects [9]. On the other hand, Lanzafame et al. [6] composition during modification (due to similar rate of Si and Al ex-
and also Janiszewska et al. [10] indicated that post-synthesis traction) as well as preferential extraction of defect zones in the

⁎
Corresponding authors.
E-mail addresses: [email protected] (E. Janiszewska), [email protected] (J. Kowalska-Kuś).

https://doi.org/10.1016/j.apcata.2019.05.012
Received 31 January 2019; Received in revised form 10 May 2019; Accepted 11 May 2019
Available online 12 May 2019
0926-860X/ © 2019 Elsevier B.V. All rights reserved.
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

same portion of material. After the treatment, the samples were filtered,
washed with deionized water and then dried at room temperature and
calcined in the air at 550 °C for 3 h. The resulting samples were labelled
as Sil-1_x, where x stands for the anion of applied ammonium salt (e.g.
Sil-1_F was prepared with 1 M NH4F solution).

2.2. Characterization

The products were characterized by means of standard physico-

Scheme 1. Silanol defects created during the modification procedure of sili- chemical methods. Powder X-ray diffraction patterns were recorded on
calite-1. the Philips Bruker D8 Advance diffractometer using Cu Kɑ radiation
(λ = 1.54056 Å). The relative XRD crystallinity of the modified samples
crystals. was estimated as the ratio of the sum of the intensities of the most
Regarding the above, we have performed a modification of pristine intense reflections in the range of 23-25° and the corresponding sum of
silicalie-1, by means of post-synthesis treatment with solutions of var- reflections in the unmodified sample (Sil-1) [24,25]. Removal of rem-
ious ammonium agents (NH4NO3, NH4Cl, NH4F, NH4OH) at elevated nants of applied ammonium agents (i.e. fluorides and chlorides) as well
temperature in order to induce the generation of the acidic sites. as absence of other impurities (i.e. Al, Fe) was confirmed by means of
Contrary to earlier papers [6,10,11] the modification has been carried EDS spectrometer of Thermo Fisher Scientific with an Ultra Dry Silicon
out without any additional agents (basic compounds). The obtained Drift X-ray Detector (SDD), X-ray fluorescence (XRF) measurements
materials were examined as catalysts for acetalization of glycerol with using a Mini-Pal 2 spectrometer (Panalytical) and X-ray photoelectron
acetone to produce solketal (Scheme 2), that is considered as very im- spectra (XPS) on an Ultra High Vacuum (UHV) System (Specs, Ger-
portant oxygenate fuel additive enebling to reduce the emission of many). Scanning electron microscope (SEM) images were recorded on
particles and improve the cold flow properties of liquid transportation the Hitachi SU3500 microscope. Transmission electron microscopy
fuels. Solketal helps also to reduce the gum formation, improves the (TEM) images were recorded on a JEOL 2000 microscope operating at
oxidation stability, and enhances the octane number when added to accelerating voltage of 80 kV. Surface hydrophilicity of the catalysts
gasoline [17]. The role of acidity of the catalysts dedicated to acet- was measured by water adsorption. For this purpose, the samples were
alization of glycerol has been widely discussed in many papers [18–21]. moisturized in a desiccator containing demineralized water at a room
The number [22] and strength [23], as well as the nature of acidic sites temperature for 24 h. The amount of water adsorbed in the catalysts
(both Lewis and Brønsted) [20,21] were indicated as parameters in- were determined by using thermogravimetric analysis (SETARAM
fluencing both glycerol conversion and selectivity to solketal. SETSYS 12). The 29Si solid state MAS NMR spectra were measured on
the APOLLO console (Tecmag) at the magnetic field of 7.05 T provided
by the 300 MHz/89 mm superconducting magnet (Magnex). A Bruker
2. Experimental HP-WB high-speed MAS probe equipped with the 4 mm zirconia rotor
and KEL-F cap was used to spin the sample at 4 kHz. The resonance
2.1. Synthesis frequency was equal to 59.515 MHz, and a single 3 μs rf excitation pulse
corresponding to a π/2 flipping angle was applied. The number of scans
Silicalite-1 was prepared according to our earlier procedure [10] was 2048 or 7168, with the acquisition delay equal to 10 s. The fre-
with some modification. Tetrapropylammonium bromide (TPABr, 98% quency scale in ppm was referenced to TMS (tetra-methylo-silane). N2
Lancaster) and water glass (sodium metasilicate with weight percentage adsorption/desorption isotherms were measured at – 196 °C on the
ratio: Na2O/SiO2/H2O = 11.1/27.9/60.8, Chempur, Poland) were used Quantachrome Nova 1000 sorptometer. The samples were outgassed at
as a template and silica source, respectively. Phosphoric acid (85%, 300 °C prior to the measurement. The specific surface area was de-
POCh, Poland) was used for adjusting the pH of the initial gel. The termined using the BET method, whereas the external surface area and
molar composition of the synthesis gel was 0.08 TPABr/1 SiO2/20 H2O. micropore volume were calculated by the t-plot method. The total vo-
Template was dissolved in distilled water and then was merged with lume of pores was assessed using the single point model (at p/
water glass under stirring. The pH of the initial mixture was adjusted to p0 = 0.98). The BJH pore size distributions were derived from the ad-
value 11 with H3PO4. The obtained reaction mixture was stirred for sorption branch.
0.5 h and then transferred into Teflon lined stainless steel autoclave. The Fourier transform infrared spectra (KBr pellets) were recorded
The gel was crystallized at 170 °C under autogeneous pressure for 22 h. with Bruker Tensor 27 spectrophotometer. The spectra were also em-
The resulting product was filtered, washed with destilled water, dried ployed for estimation of crystallinity of the initial and modified samples
and calcined in air at 550 °C for 5 h to remove the template. according to the formula: (I550/I450)/0.72 ∙ 100% [25]. The values I550
The samples of calcined silicalite-1 (denoted as Sil-1) were modified and I450 stand for the intensities of the bands at 550 and 450 cm−1,
with 1 M solutions of various ammonium compounds (NH4NO3, NH4Cl, respectively.
NH4F, NH4OH). The sample (1 g) of silicalite-1 was mixed with 100 cm3 The recorded FTIR spectra (self-supported wafers) of adsorbed
of the respective NH4+ source solution. The mixture of silicalite-1 and pyridine (Py) and pivalonitrile (Pn) were used to estimate the number
aqueous solution of ammonium salt was stirred under reflux at 60 °C for and strength of acid sites. Considering the difference in kinetic dia-
1 h. The procedure was repeated twice, to increase the effect, with the meters of the applied probe molecules [26,27] (pyridine – 0.57 nm and

Scheme 2. Glycerol acetalization reaction with acetone.

2
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

pivalonitrile – 0.65 nm), the acidity resulted from pyridine adsorption equimolar mixture of glycerol and acetone. The equimolar mixture of
was treated as total acidity, while the acidity related to Pn adsorption glycerol and acetone, comprising 1 g of glycerol (11 mmol) and 0.8 cm3
was assigned to external acid sites. Additionally, on the base of mea- of acetone (11 mmol) was introduced into the glass vials with 0.01 g of
surements of Py and Pn adsorption the accessibility of acid sites has also the catalyst. The vials were tightly closed and heated on magnetic
been estimated. The presence of hydroxyl groups, localized in different stirrer (400 rpm) at 70 °C for 1 h. The reaction products (solketal, the 5-
sites of silicalite-1, depending on the salt used for modification, was membered ring ketal and isomer, the 6-membered ring ketal) as well as
noticed using FT-IR spectra in the range of 3200-3800 cm−1. The unreacted glycerol and acetone, were identified by mass spectroscopy
measurements were performed on a Vertex 70 Bruker spectrometer (Varian 4000 GC–MS). The results of reactions were measured by GC
equipped with a MCT detector. All spectra have been normalized to analyses on VARIAN CP-3800 chromatograph equipped with an FID
equal weight. Self-supporting wafers of the samples were prepared, detector and a VF-5 ms column. Toluene was used as an internal stan-
weighed and placed inside a bespoke quartz IR cell. Prior to the mea- dard. The conversion of glycerol (denoted as Glycerol conv.), selectivity
surements, the samples were outgassed at 400 °C for 1 h under high to products: solketal (denoted as S solketal) and isomer (denoted as S
vacuum. The concentrations of acid sites were determined on the basis isomer), solketal yield (denoted as Y solketal), were calculated ac-
of quantitative in-situ IR studies. Excess of pyridine vapors (≥99.8%, cording to the equations presented in [28].
Sigma-Aldrich) sufficient to neutralize all acid sites was adsorbed at To estimate the catalytic stability of the modified silicalite-1 in the
110 °C under static conditions, followed by an evacuation up to the studied reaction, the catalyst sample was also tested in a continuous
temperature of 150 °C to remove the gaseous and physisorbed pyridine flow system for 12 h. Regarding the low boiling point of acetone (56 °C)
molecules. This was tracked by evaluating the spectra on time (until the the reaction was conducted at 50 °C in the continuous flow system
disappearance of the bands attributed to physisorbed Py). Subse- under atmospheric pressure. Contrary to the batch conditions, the
quently, the FTIR spectrum at 150 °C was recorded. The band intensities acetone to glycerol ratio was 3 and additionally methanol was used as a
in those spectra were used to calculate the total concentration of solvent. Reaction products were analyzed and calculated as presented
Brønsted and Lewis sites. The total concentration of Lewis sites was above.
estimated using the intensities of the 1445-1450 cm−1 band of pyridine
coordinatively bonded to Lewis sites (PyL) assuming the extinction
coefficients of 0.11 cm2/μmol [26]. The band at 1545 cm−1, which 3. Results and discussion
appeared just after pyridine adsorption at 110 °C, was used to estimate
the number of very weak Brønsted acid sites. The total concentration of 3.1. Structural and textural characterization
Brønsted sites was calculated using the extinction coefficient of 0.07
cm2/μmol. The strength of Lewis acid sites has been determined by the Both pristine silicalite-1 (Sil-1) and the ammonium modified sam-
amount of pyridine desorbed under vacuum at given elevated tem- ples (Sil-1_OH; Sil-1_F; Sil-1_Cl and Sil-1_NO3) were characterized by
perature. The band maintained at 1445-1450 cm−1 (Lewis sites) upon XRD (Fig. 1) and showed correct MFI structure. High crystallinity was
evacuation at elevated temperature (300 °C) was assumed as a measure also confirmed by IR measurements (Fig. 2, Table 1). It indicates that
of the strength of acid sites. The weak acid sites release adsorbed Py the used modification procedure allows to maintain the MFI structure.
during desorption treatment at 300 °C for 10 min, whereas the strong The intensities of the reflections of the modified samples are compar-
sites still keep Py attached after this evacuation treatment. The ratio able with those of the unmodified sample, which indicates the similar
A300/A150, where A150 as well as A300 correspond to the intensities of IR crystallinity. The XRD crystallinities estimated by comparison of the
bands of Py coordinatively bonded to Lewis sites upon the evacuation at intensities of the major reflections in the range of 23-25° of modified
150 °C and at 300 °C, respectively, was taken as the measure of the acid samples to those in unmodified sample confirm also comparable crys-
strength of the Lewis acid sites. In order to estimate the concentration tallinity of the investigated samples (Table 1). Only the sample mod-
of the acid sites on the external crystal surface the pivalonitrile (98%, ified with NH4Cl or NH4OH show slightly lower crystallinity in com-
Sigma Aldrich) was adsorbed on the samples at room temperature fol- parison to initial sample (99% and 97%, respectively).
lowed by 20 min. evacuation at the same temperature to remove the The FTIR spectra (KBr) of parent and modified silicalite-1 samples
excess of physisorbed molecules. The concentration of the Lewis acid show the bands resulting from the basic vibrations characteristic for the
sites detected by the Pn was calculated from the maximum intensity framework zeolite MFI structure (Fig. 2). The bands at 1240 cm−1
(peak height) of the 2305 cm−1 band and its extinction coefficient
(0.15 cm2·μmol−1). The accessibility factor (AF) for the pivalonitrile
probe molecule was defined as the number of sites detected by ad-
sorption of the Pn (external sites) divided by the total amount of acid
sites in the studied materials quantified by pyridine sorption (external
and internal sites).
The NH3-TPD measurements of acidity were performed in a flow
reactor. In a typical experiment, about 40 mg of sample was heated in
He stream at the rate of 10°/min. up to 500 °C and kept at that tem-
perature for 0.5 h, then cooled down to 120 °C and afterwards saturated
with ammonia for 0.5 h. The physically adsorbed NH3 was removed by
purging with helium flow at 120 °C for 1 h. The TPD analysis was
conducted in a range of 100–600 °C with a heating rate of 10°/min. The
desorbed NH3 was recorded by a TCD analyzer. All TPD-NH3 profiles
presented in this work were normalized to the same sample weight.

2.3. Catalytic activity measurement

The catalytic activity of the samples was examined for glycerol

acetalization with acetone. The catalysts were pre-activated in air at
350 °C for 2 h and then cooled in a desiccator. The activated silicalite-1
catalysts (1 wt.% referred to glycerol mass) were added to the Fig. 1. XRD patterns of the indicated samples.

3
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

Fig. 2. FTIR (KBr) spectra of framework vibration of the indicated samples.

(external asymmetric stretch), 1120 cm−1 (internal asymmetric

stretch), 800 cm−1 (internal symmetric stretch) and at 450 cm−1 (Si-O
bending mode) are the bands characteristic of Si-O bonding in siliceous
materials that are insensitive to the structure. The band at 550 cm−1
originates from the vibrations of double five-membered rings and it is
characteristic for the materials of MFI structure [29]. The high intensity
of this band evidences a high crystallinity of the modified samples
Fig. 3. SEM images of the samples: A) Sil-1, B) Sil-1_NO3, C) Sil-1_OH, D) Sil-
which is in line with the XRD data. Comparison of this band intensity
1_Cl, and E) Sil-1_F.
with the intensity of that at 450 cm−1 was used to determine the IR
crystallinity of the modified materials (Table 1, Fig. 2). These results
also confirm the high crystallinity of the modified samples. The highest higher surface area (including the external surface) (Table 1) of the
crystallinity (IR as well XRD crystallinity) shows the sample modified above sample. However, for some modified samples (Sil-1_OH, Sil-1_Cl,
with NH4F solution (Sil-1_F) and it is even slightly higher than the and Sil-1_F), small amounts of amorphous material appeared as a result
crystallinity of the starting material. Similar observation was presented of etching [33]. Careful analysis of the image of Sil-1_F sample and its
by Valtchev et al. [16,30,31] for aluminosilicalite MFI treated with comparison with the picture of pristine silicalite-1 show some changes
NH4F-HF solutions. Higher crystallinity of ZSM-5 treated this way was in the crystallites due to the gentle etching. The above mentioned
explained by the removal of low crystalline domains from the treated changes for the Sil-1_F sample have no reflection in lowering its crys-
materials. Probably the fluorine anions used in our procedure act in a tallinity, as it was in the case of the Sil-1_OH and Sil-1_Cl samples.
similar way as the fluorine anions from the pair NH4F-HF. As presented in Table 1, the treatment of silicalite-1 with ammo-
Scanning electron micrographs (SEM) of the initial and modified nium salt solutions at an elevated temperature results in a change in its
samples are presented in Fig. 3. The morphologies of the Sil-1 and textural properties. Particularly an increase in specific surface area and
modified samples are quite similar. The samples exhibit well-formed the pore volume is seen. These changes are much less conspicuous for
crystals with a twinned structure typical for MFI materials [32]. The the sample modified with NH4NO3. The increase in the total pore vo-
surface of the crystals was smooth for the unmodified sample as well for lume (Vt) for modified samples at the unchanged volume of micropores
the sample modified with NH4NO3, whereas the other samples con- (Vmicro) compared to the initial sample indicates a generation of new
tained particles with a rough external surface. It was the most con- mesopores during the modification procedure. This is confirmed by the
spicuous for the sample modified with NH4OH and it might result from enhanced mesopore area (Sext). The highest increase of the external
removal of some framework Si atoms. It is consistent with slightly surface and total pore volume was observed for sample modified with

Table 1
Crystallinity, textural properties and the number of acid sites (from desorbed ammonia) of the indicated samples.
Sample Modifier CaIR [%] CbXRD [%] SBETc Vtd Vmicroe Sextf NH3des
[m2/g] [cm3/g] [cm3/g] [m2/g] [μmol/g]

Sil-1 – 100 100 318 0.19 0.11 95 44

Sil-1_NO3 NH4NO3 93 100 334 0.20 0.11 104 175
Sil-1_OH NH4OH 99 97 366 0.35 0.11 131 188
Sil-1_Cl NH4Cl 95 99 346 0.22 0.11 108 213
Sil-1_F NH4F 101 103 351 0.24 0.11 111 287

a
IR crystallinity defined as (I550/I450)/0.72·100% (I550 and I450 - the intensities of the bands at 550 and 450 cm−1), bXRD crystallinity defined as a ratio of the sum
of the areas of the four most intense reflections (in the range of 23-25°) of the modified sample and the corresponding sum for an unmodified sample,cBET specific
surface area, dtotal pore volume, emicopore volume, fexternal surface.

4
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

Fig. 4. N2 adsorption/desorption isotherms (A) and pore size distribution (B) of the indicated samples.

NH4OH and it was caused by a partial leaching of silicon from the applied catalysts is very important. According to [35], zeolites of ZSM-5
framework upon treatment with alkaline solution. Some changes of structure indicate the pores of diameter of 0.56 nm. On the other hand,
these values were noticed also for Sil-1_F due to gentle etching of silicon the diameters of the applied reagents and obtained products, calculated
from the structure by fluorine species [16,30,31]. Some mesopores re- with the GaussView program [28], falls within the range of
corded in the initial sample (0.08 cm3/g) might result from the inter- 0.43−0.51 nm. The modification of silicalite-1 by ammonia salts
crystalline voids, as implies the hysteresis loop recorded at higher re- treatment results in the increase of external surface (Table 1), which
lative pressures (Fig. 4A). The similar Vmicro value for all investigated indicates the mesopores formation. Analysis of pores distribution pre-
samples confirms again the similar crystallinity of the samples [26]. sented in Fig. 4B points at appearance of new pores with diameter of
The low temperature N2 sorption isotherms exhibit two hysteresis about 5 and 8 nm, especially marked for the samples obtained by
loops. The first one at a low partial pressure (˜ 0.2 p/p0) is not very treatment with NH4F, NH4Cl and NH4OH. The appearance of such
conspicuous for the unmodified sample, whilst it is very pronounced for mesopores results in a better availability of formed acidic sites, thus
the modified ones. Such hysteresis loop was detected for highly silica affecting the catalytic activity.
and pure silica MFI materials [34]. Both crystallites size (the materials
with large crystals, > 350 μm) and the generated defect sites are also 3.2. Acidity
responsible for the formation of hysteresis loop at low partial pressure
(˜ 0.2 p/p0) [34]. The starting material is a pure silica, thus the oc- The surface OH groups were characterized by means of IR spectra
currence of a hysteresis loop for it at p/p0 ˜ 0.2 is fully justified. Re- recorded in the range of 3200-3800 cm−1, with self-supporting discs
garding the SEM results showing no change in crystallite size after evacuated at 400 °C. The comparison of IR spectra of OH groups for Sil-
modification, the increasing volume of the hysteresis loop at p/p0 ˜ 0.2 1 and for the modified samples indicates an essential impact of post-
for modified materials can be explained by a generation of the frame- synthesis modification on properties of the silicalite-1 surface (Fig. 6).
work defects upon the modification. The creation of such defects in The IR spectrum of initial Sil-1 shows only the presence of the isolated
silicalite-1 as a result of modification with NH4Cl was already reported hydroxyl groups on the external surface (SieOHiso/ext at 3745 cm−1). A
[10]. The second hysteresis loop (p/p0 > 0.5) for the initial sample is treatment with solution of ammonium agents results in the formation of
attributed to the intercrystalline porosity. The growing second loop in framework defects represented by internal isolated OH groups (SieO-
the modified samples is associated with a generation of additional Hiso/int at 3730 cm−1), and a significant number of hydrogen-bonded
mesoporosity during the applied procedure. It is evidenced by the vicinal hydroxyl groups (SieOHHB/vic) as well as the silanol nests
course of pore distribution curve (Fig. 4B) as the additional pores with (SieOHHB/int), which consist of internal SieOH groups, that are located
larger diameters compared to those of the starting material. inside the pores. It is evidenced by revealing the bands at 3685 cm−1
TEM images of pristine silicalite-1 and of the modified samples and 3600−3300 cm−1, respectively [36,37]. IR spectrum of silicalite-1
characterized with the highest mesoporosity were recorded in order to modified with NH4OH still indicates a significant population of isolated
explore in details the formed pore structure. As shown in Fig. 5A, no SieOH groups on the external surface in comparison to the total
mesoporosity is seen in the initial Sil-1, whereas the modified samples amount of silanols, but the internal isolated and hydrogen-bonded si-
show the presence of well defined pores (Fig. 5 B and C). The mesopores lanol groups become predominant. In the case of sample modified with
look different, regarding the used modification agent. The sample NH4NO3, the contribution of external OH groups is much higher com-
treated with NH4OH indicates the irregular pores in the form of chan- pared to the others modified samples. The use of NH4Cl and NH4F as
nels as well as rectangulars, whereas the treatment with NH4F causes modifiers results in the removal of isolated external OH groups and
the formation only the rectangular mesopores. Similar rectangular subsequently in the formation of substantial amounts of isolated in-
shape of the pores where obtained by Valtchev et al. [16] by means of ternal hydroxyls (3730 cm−1) and the hydrogen-bonded, both vicinal
treatment of zeolite ZSM-5 with the concentrated (40%) aqueous so- hydroxyls (3685 cm−1) and those in silanol nests (defects, recorded as
lution of NH4F. IR bands in the range of 3600−3300 cm−1). However, the higher in-
Considering the diameters of reagent molecules and also that of the tensity of aforementioned bands for Sil-1_F indicates the greater
products resulting from glycerol ketalization, the pore diameter of the amounts of the silanol nests for this sample as a result of the higher

5
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

Fig. 5. TEM images of the pristine silicalite-1 (A) and the samples treated with NH4OH (B) and NH4F (C) solutions.

Scheme 3. Scheme of the formation of active sites in modified silicalite-1.

samples is observed (weight loss 1,95% for Sil-1; 3,34% for Sil-1_OH,
2,56% for Sil-1_F). The increase of hydrofilicity of modified samples in
comparison to the pristine one proves the formation of internal defects
during the modification procedure [43,44].
The 29Si MAS NMR spectrum of the Sil-1_F sample indicates the
presence of the main peak at -115 ppm assigned to the Q4 species [Si
(OSi)4] and hardly noticeable peak at -104 ppm corresponding to Q3
species [(HO)Si(OSi)3] (Fig. 7) [37]. The latter is seen only in the
measurement with a high number of scans and it indicates a small
contribution of the defects. Their concentrations estimated from the
relative peak areas was ˜3%. This value is near the detection limit of the
29
Fig. 6. IR spectra of the OH groups recorded at RT for the pristine and modified Si MAS NMR technique for such defects in silicalite-1 [37]. Probably
silicalite-1 samples. the using of 1H–29Si CP MAS NMR could give the more spectacular
result confirming the presence of new silanols groups [9,37,45].
The acidic nature and the strength of OH groups in the initial and
amount of formed internal defects. A disappearance (in the case of Sil-
modified samples have been estimated by means of NH3-TPD (Fig. 8)
1_F and Sil-1_Cl) or decrease (in the case of Si-l_OH) of the external
silanols can be caused by their transformation into hydrogen-bound OH
forms. According to Rimola et al. [38] and also Bordiga et al. [39] the
appearance of Lewis acid sites recorded by means of FTIR spectra of
adsorbed pyridine is one of the indicators of the formation of internal
defects called the hydroxyl nests. Their following partial dehydroxyla-
tion results in Lewis acid centers formation. The crystalline MFI struc-
ture without defects indicates only external OH groups recorded in FTIR
spectra at about 3745 cm−1.
The Lewis acidic sites are created as a result of thermal treatment of
silicalite, which enables the condensation of neighbouring hydroxyl OH
groups with the formation of “strained siloxane bridges”. According to
[38,40,41] the unsymmetrical siloxane bridge (labeled as S2R site),
characterized with electron deficit can act as Lewis acid. (Scheme 3).
On the other hand, the partial dehydroxylation still preserve some
hydroxyl groups with enhanced acidity [42], which may act as very
weak Brønsted acid sites (Scheme 3).
The formation of nests of H-bonded internal silanols caused the
increase of hydrofilicity of the modified samples. Thermal analysis of
the samples after adsorption of water indicates hydrophobic nature of
Fig. 7. 29Si MAS NMR spectra of the Sil-1_F sample recorded with indicated
their surface, hovewer slightly increase of hydrofilicity for modified
number of scans.

6
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

measurements and the IR spectra of adsorbed pyridine (Py) (Fig. 9).

The NH3-TPD profiles (Fig. 8) showed a low acidity of the starting
material, which may result from a small amount and very low acid
strength of the external hydroxyl groups. The TPD profiles of the
modified samples indicate creation of additional acid sites as a result of
the modification procedure. The number and the strength of generated
acid sites depend on the type of ammonium agents used for modifica-
tion (Fig. 8, Table 1). Samples modified with NH4Cl and NH4F indicate
the highest overall acidity and the higher contribution of stronger acid
centers in comparison to the other samples. The lowest number of
acidic sites among the studied modified samples indicates that modified
with NH4NO3 solution. The weakest acidic sites were noticed for the
sample Sil-1_OH (maximum of ammonia desorption at 300 °C).
The IR spectra of adsorbed pyridine were measured in order to es-
timate the number, strength, and the nature of acid sites in the in-
vestigated samples (Fig. 9). The spectrum of unmodified sample (Sil-1)
shows only very weak bands at 1440 and 1590 cm−1 attributed to
tentatively hydrogen-bonded pyridine [6,46] and to very weak electron
acceptor sites originating from the local framework defects [47]. The
modification with ammonium compounds causes a shift of the band at
1440 cm−1 towards higher wavenumber. The samples modified with
Fig. 8. NH3-TPD profiles of the initial and modified silicalite-1 samples. NH4NO3 and NH4OH indicate the band at 1445 and 1595 cm−1 at-
tributed to weak Lewis acid sites [6] of the strength higher than Lewis
sites in an initial material Sil-1. The modification with NH4Cl and NH4F
causes also a formation of strong Lewis acid sites (1450 and
1620 cm−1). The sample Sil-1_Cl indicates the presence of both weak
and strong Lewis sites with the predominance of weak ones, whereas
the sample modified with NH4F indicates the presence of almost only
strong Lewis sites.
The concentration of Lewis acid sites was determined by the
quantitative IR spectra of adsorbed pyridine. Negligibly low con-
centration of Lewis acid sites (30 μmol/g) was detected (Table 2) for the
initial sample (Sil-1). The applied post-synthesis treatment causes a
significant increase in the number of Lewis sites (7–9 times).
The analysis of the A300/A150 values from pyridine thermodesorp-
tion indicates the differences in the acid strength of the Lewis sites for
the modified materials. The samples modified with NH4NO3 and
NH4OH indicate the presence of almost exclusively weak acid sites,
whereas the samples obtained with NH4F show almost only strong
Lewis acid sites. The presence of small amount of very weak Brønsted
acid sites (recorded after evacuation at 110 °C – spectra not shown) has
also been recorded. Such weak Brønsted acid sites were observed for
Fig. 9. IR spectra of adsorbed pyridine of the initial and modified silicalite-1 silicalite materials (with MEL structure) also by other authors [4].
samples. The observed tendency of changes in acidity (i.e. total number of
acid sites) of the modified samples, estimated by means of IR spectra of
Py and NH3-TPD analysis, are similar, nevertheless, the acidity

Table 2
Acidity (the number and strength of acidic sites), the accessibility factor and catalytic characterization of initial and modified silicalites.
Concentration of acid sites [μmol/g] Accessibility factor AFe Strength of acid sitesf Glycerol conv. [%] Y solketal TON related to
[%] [%] L+B
a b c
LPy LPn Lint BPyd Total L + B acid sites

Sil-1 30 0 30 0 30 0 0 0 0 –
Sil-1_NO3 210 25 185 10 220 12 0.30 36.0 33.8 1682*
Sil-1_OH 220 90 130 5 225 41 0.25 37.6 36.1 1764*
Sil-1_Cl 220 35 185 35 255 16 0.65 50.0 47.5 2039*
Sil-1_F 265 70 195 40 305 26 0.85 65.8 63.9 2295*
Sil-1_F 265 70 195 40 305 26 0.85 85.5 84.2 98.4**
Glycerol conv. r2 0.3472 0.0317 0.8572 0.8456 0.9959 – – – – –
Y solketal r2 0.3232 0.0426 0.8762 0.8239 0.9980 – – – – –

a
total Lewis acid sites obtained from IR studies of pyridine recorded at 150 °C, bexternal Lewis acid sites obtained from IR studies of pivalonitrile recorded at RT,
c
internal Lewis acid sites = total Lewis (Py) – external Lewis (Pn), dprotonic acid sites obtained from IR studies of pyridine recorded at 110 °C (spectra not shown),
e
calculated as a ratio of the concentrations of acid sites accessible to pivalonitrile to those determined for adsorption of pyridine, facid strength obtained from IR
studies of pyridine thermodesorption expressed as A300/A150, * TON values calculated for 0.01g of the catalyst applied in a batch reactor, ** TON value calculated for
0.5g of the catalyst applied in continuous flow process, r2 linear correlation factor as a function of the total number acids sites (L + B) with glycerol conversion and
solketal yield.

7
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

calculated on the basis of ammonia TPD shows lower values. This dif-
ference is reasonable regarding the disparity in basicity and proton
affinities of both probe adsorbates. According to Hunter and Lias [48]
basicity of ammonia in gas phase was indicated as 819 kJ/mol, while
basicity of pyridine is equal to 8981 kJ/mol. Similarly, the proton af-
finity is lower for NH3 (8536 kJ/mol) than for pyridine (930 kJ/mol). It
indicates that pyridine may interact even with very weak acidic centers,
while ammonia may adsorb only on stronger acidic sites. It explains the
lower acidity measured by TPD of ammonia in comparison to acidity
estimated by FTIR with pyridine.
The number of generated acid centers does not correlate with the
contribution of formed amorphous phase (Table 1). The sample mod-
ified with NH4OH indicates the highest amount of amorphous phase (˜
3%), whereas the highest amount of generated acid sites was estimated
for the sample modified with NH4F solution. This suggests that the ef- Fig. 10. Catalytic activity of different silicalite-1 samples in acetalization re-
action of glycerol with acetone.
ficient active sites are mostly located in the crystalline MFI materials.
Even though, the presence of some acid sites in the amorphous phase
cannot be excluded, the amorphous silica applied as catalyst for gly-
cerol ketalization resulted in a lack of activity [49,50]. It allows us to
believe, that amorphous silica comprises mainly inactive silanol groups.
The number of formed acid sites related to the surface (expressed in
μmol/m2) increases with the growth in the surface area in the following
order Sil-1 < < Sil-1_OH < Sil-1_NO3 < Sil-1_Cl < Sil-1_F. In the case
of Sil-1_OH sample the increase in the number of acid sites related to
the surface area is the lowest, when compared to other modified sam-
ples, which is consistent with catalytic activity of modified samples
presented in the part concerning the catalytic activity.
The use of catalysts with wider pores and good accessibility of acid
sites should be beneficial for catalytic reactions, particularly, when
larger reactant molecules are involved and the diffusion constraints Fig. 11. The tendency of the changing of the number of acidic sites related to
could affect the reaction efficiency. According to the data presented in surface area (μmol/m2) and glycerol conversion of the studied catalysts.
Table 2 and also to textural parameters (Table 1), the modification of
silicalite-1 might improve the accessibility of acidic sites remarkably.

3.3. Catalytic performance

The role of acidity in acetalization of glycerol has been discussed in

many papers [18–22]. The number and strength [23] as well as the
nature [20], and avaibility [28] of acid sites were indicated as para-
meters influencing both glycerol conversion and selectivity to solketal.
Nair et al. [51] indicated that higher glycerol conversion and solektal
yield were obtained on catalysts with higher strength of Brønsted
acidity. On the other hand, Calvino-Casilda et al. [22] came to the
conclusion that the formation of solketal may be catalyzed by both
Lewis and Brønsted acid sites, however, higher glycerol conversion and
selectivity towards solketal was attributed to higher Brønsted acidity.
Similarly, Vieira et al. [52] investigating the role of Lewis and Brønsted Fig. 12. Glycerol conversion and selectivity towards solketal over Sil1_1F with
acid sites in the ketalization reaction of glycerol on vanadium oxides indicated time on stream.
deposited on lamellar ferrierite zeolites, indicated the possible two-step
mechanism involving Lewis or Brønsted acid sites. The importance of generated acidic sites. This catalyst indicates not only the largest
Lewis acid centers for glycerol acetalization has also been suggested in amount of Lewis sites but also some amount of weak Brønsted acid
the paper [53], even though, the applied catalysts also contained some centers (Table 2). According to the earlier paper [23] the combined
amount of Brønsted acidic sites. Zaher et al. [54] suggested the possible action of both Brønsted and Lewis sites was indicated as a favorable
important role of Lewis acidic sites of transition metal for glycerol ke- factor for glycerol acetalization. However, the rising glycerol conver-
talization over the iron catalyst (FeCl3 x 6H2O). sion was attributed to the nature of the acidic sites in a following order:
All the modified silicalites of MFI structure showed considerable L < L + B < B (L – Lewis sites and B - Brønsted sites). Considering that,
activity for glycerol acetalization with acetone (Fig. 10). Some differ- the increasing number of Brønsted acid sites (Table 2), even if very
ences in activity were caused by the number, nature, strength and po- weak ones, seems to support the observed increase in glycerol conver-
pulation (the number of acidic sites related to the total surface area) of sion (Fig. 10). Moreover, it has also been shown earlier [55,56], that the
acidic sites in modified samples. interaction of Lewis (medium) and Brønsted (weak) acid sites may
Sil-1_F with the highest number of isolated internal and hydrogen- generate higher activity for acid catalyzed reactions than that of each
bonded hydroxyl groups shows the best activity for glycerol acetaliza- type of centers alone. Isolated silanols or even hydrogen-bonded sila-
tion. Although the accessibility of internal acid sites is higher for Sil- nols, although being weakly Brønsted acidic are able to form quite
1_OH, the available centers are much weaker compared to these pre- strong, but easily reversible hydrogen bonds with polar molecules.
sented in Sil-1_F and Sil-1_Cl (Table 2) and it causes its lower activity. Above mentioned phenomena taking place in adsorption-desorption
High activity of Sil-1_F can be associated with the number and nature of cycles justifies the activity of Sil-1_F. Additionally, in “defective” form

8
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

Scheme 4. Proposed, combined Lewis-Brønsted me-

chanism for the glycerol acetalization with acetone (D –
defects acting as a Lewis acid center, OH – Brønsted acid
center).

of silicalite-1 the high population of H-bonds of silanol nests provides Brønsted sites (0.9959 and 0.9980, respectively). The glycerol conver-
the additional catalytic activity in an industrially important reaction as sion and solketal yield do not show as good correlation with the number
the vapor-phase Beckmann rearrangment of cyclohexanone oxime to of internal Lewis acid sites as well with Brønsted sites, analyzed sepa-
caprolactam conducted at 300 °C [57,58]. On the other hand, the en- rately. All the other correlations (Table 2) are meaningless.
hacement of the Bronsted acid sites strength by the neighberhooding Considering the above, it seems reasonable to conclude that the
Lewis moiety can provide an adittional positive impact on adsorption interaction of Lewis and weak Brønsted acidic sites generates the cat-
procces, and finally catalytic activity. alytic centers favorable for glycerol acetalization (Scheme 4).
Analysis of glycerol conversion, considering the number of acidic
sites related to surface area (μmol/m2), shows that the activity of the 4. Conclusions
studied catalysts increases simultaneously with the growth of the sur-
face-related acidity. Accordingly that, the highest glycerol conversion 1 Modification of silicalite-1 by treatment with ammonium salt solu-
shows the sample modified with NH4F, characterized with the highest tions at a higher temperature (without any additional alkaline
number and the strength of acid sites and modest enhancement of the agents) results in partial removal of external silanols and in a for-
surface and pore size (Fig. 11). It allows to conclude that the strong acid mation of acidic internal OH groups.
sites appear more decisive in high catalytic activity than the increased 2 Structure defects formed during the above transformation play a
surface area. role of Lewis sites.
Regardless of the different glycerol conversion over the samples 3 Silanols linked by hydrogen bonds or isolated silanols connected
modified with various agents, the selectivity to solketal was very high with defected silicon atoms are the source of Brønsted centers of
for all modified samples. According to [59,60] solketal, a 5-membered very low strength.
ketal is a kinetically favored product, while 6-membered ketal is a re- 4 A synergy of the Lewis and Brønsted acidic sites generated on
sult of solketal isomerization and it is characterized by higher ther- modified silicalites enhances the catalytic activity which results in
modynamic stability. Low strength of acid sites present in all applied high glycerol conversion and very high selectivity to solketal.
catalysts impedes the isomerization process and generation of un-
desirable product. Acknowledgement
To estimate the catalytic stability of modified silicalites for glycerol
acetalization reaction, the best catalyst (Sil-1_F) has been employed in a K. Góra-Marek acknowledges the Grant No. 2015/18/E/ST4/00191
continuous flow system (at 50 °C under atmospheric pressure) for 12 h. from the National Science Centre, Poland.
The activity of Sil-1_F was relatively stable and it showed only slight
deactivation within the time on stream over 12 h. Moreover, the se- References
lectivity to solketal was almost 100% (Fig. 12).
Activity of the applied catalysts has also been estimated by means of [1] J. Zhang, X. Zhu, G. Wang, P. Wang, Z. Meng, C. Li, Chem. Eng. J. 327 (2017)
TON values calculated with regards to the number of the acidic sites 278–285, https://doi.org/10.1016/j.cej.2017.06.114.
[2] H. Bayahia, E. Kozhevnikova, I. Kozhevnikov, Chem. Commun. 49 (2013)
participating in the reaction (Table 2). As mentioned above, the acidity 3842–3844, https://doi.org/10.1039/c3cc41161c.
was estimated on the base of pyridine (0.57 nm in molecule diameter), [3] Y.-Q. Deng, S.-F. Yin, C.-T. Au, Ind. Eng. Chem. Res. 51 (28) (2012) 9492–9499,
pivalonitrile (0.65 nm) and ammonia (0.19 nm) adsorption. The size of https://doi.org/10.1021/ie3001277.
[4] X. Meng, M. Zhang, C. Chen, C. Li, W. Xiong, M. Li, Appl. Catal. A Gen. 558 (2018)
ammonia molecule is much lower than the size of reagents and products 122–130, https://doi.org/10.1016/j.apcata.2018.04.001.
(0.43 – 0.51 nm), therefore reagents are not able to contact with all [5] Z. Honghan, Z. Shanhe, W. Haroen, L. Chunyi, China Petrol. Proc. Petrochem.
acidic sites indicated by ammonia. However, the use of TPD of am- Technol. 19 (4) (2017) 11–16.
[6] P. Lanzafame, K. Barbera, S. Perathoner, G. Centi, A. Aloise, M. Migliori,
monia is a good tool to characterize the strength of the acid sites pre- A. Macario, J.B. Nagy, G. Giordano, J. Catal. 330 (2015) 558–568, https://doi.org/
sent in the applied catalysts. Considering the diameter of adsorbates 10.1016/j.jcat.2015.07.028.
[26,61], reagents and also the products [28], as well the basicity of the [7] Y. Matsumura, K. Hashimoto, S. Yoshida, J. Chem. Soc. Faraday Trans. 1 (84)
(1988) 87–96, https://doi.org/10.1039/F19888400087.
adsorbates [48] we came to the conclusion that the number of acidic
[8] J.M. Chezeau, L. Delmotte, J.L. Guth, Z. Gabelica, Zeolites 11 (6) (1991) 598–606,
centers calculated with aid of the FTIR spectra of adsorbed pyridine https://doi.org/10.1016/S0144-2449(05)80011-9.
should be the most reliable and it was assumed to present the total [9] M. Trzpit, M. Soulard, J. Patarin, N. Desbiens, F. Cailliez, A. Boutin, I. Demachy,
acidity. TON was calculated considering the contribution of only Lewis A.H. Fuchs, Langmuir 23 (20) (2007) 10131–10139, https://doi.org/10.1021/
la7011205.
sites (Table 2, column 2) and also a sum of Lewis and Brønsted sites [10] E. Janiszewska, A. Macario, J. Wilk, A. Aloise, S. Kowalak, J.B. Nagy, G. Giordano,
(Table 2, column 6). The best correlation of the TON values alteration Microporous Mesoporous Mater. 182 (2013) 220–228, https://doi.org/10.1016/j.
with glycerol conversion and yield of solketal has been obtained taking micromeso.2012.12.013.
[11] M. Kitamura, H. Hichihashi, I. Tojima, US Patent 5 212 302, 1993.
into account the total acidity (sum of Lewis and Brønsted sites). Con- [12] B. Bonelli, L. Forni, A. Aloise, J.B. Nagy, G. Fornasari, E. Garrone, A. Gedeon,
sidering the above data and taking into account the literature in- G. Giordano, F. Trifirò, Microporous Mesoporous Mater. 101 (2007) 153–160,
formation mentioned earlier [20–23], it seems justified to assume that https://doi.org/10.1016/j.micromeso.2006.11.006.
[13] S. Kowalak, E. Szymkowiak, M. Łaniecki, J. Fluorine Chem. 93 (1999) 175–180,
both Lewis and Brønsted acid sites take part in the catalyzed reaction. https://doi.org/10.1016/S0022-1139(98)00312-1.
To answer the question about the role of the acidic sites nature in [14] Z. Qin, K.A. Cychosz, G. Melinte, H. El Siblani, J.-P. Gilson, M. Thommes,
ketalization reaction of glycerol we have performed the calculation of Ch. Fernandez, S. Mintova, O. Ersen, V. Valtchev, J. Am. Chem. Soc. 139 (48)
(2017) 17273–17276, https://doi.org/10.1021/jacs.7b10316.
the correlation of Lewis and Brønsted acid sites with the catalytic re- [15] J. Přech, K.N. Bozhilov, J. El Fallah, N. Barrier, V. Valtchev, Microporous
sults (Fig. 10) using linear regression function (Table 2). It is evident Mesoporous Mater. 280 (2019) 297–305, https://doi.org/10.1016/j.micromeso.
from the r2 analysis that both glycerol conversion as well as solketal 2019.02.023.
[16] Z. Qin, G. Melinte, J.‐P. Gilson, M. Jaber, K. Bozhilov, P. Boullay, S. Mintova,
yield show a good correlation with a total amount of Lewis and

9
E. Janiszewska, et al. Applied Catalysis A, General 581 (2019) 1–10

O. Ersen, V. Valtchev, Angew. Chem. Int. Ed. 55 (48) (2016) 15049–15052, https:// [39] S. Bordiga, C. Lamberti, F. Bonino, A. Travert, F. Thibault-Starzyk, Chem. Soc. Rev.
doi.org/10.1002/anie.201608417. 44 (2015) 7262–7341, https://doi.org/10.1039/C5CS00396B.
[17] A. Talebian-Kiakalaieh, N.A.S. Amin, N. Najaafi, S. Tarighi, Front. Chem. 6 (2018) [40] S.D. Fleischman, S.L. Scott, J. Am. Chem. Soc. 133 (2011) 4847–4855, https://doi.
573, https://doi.org/10.3389/fchem.2018.00573. org/10.1021/ja108905p.
[18] M. de Torres, G. Jiménez-osés, J.A. Mayoral, E. Pires, M. de los Santos, Fuel 94 [41] A. Morrow, I.A. Cody, J. Phys. Chem. 80 (1976) 1995–1998, https://doi.org/10.
(2012) 614–616, https://doi.org/10.1016/j.fuel.2011.11.062. 1021/j100559a009.
[19] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martín, Green Chem. 12 (2010) [42] S. Bordiga, P. Ugliengo, A. Damin, C. Lamberti, G. Spoto, A. Zecchina, G. Spanò,
899–907, https://doi.org/10.1039/B923681C. R. Buzzoni, L. Dalloro, F. Rivetti, Top. Catal. 15 (2001) 43–52 DOI:1022-5528/01/
[20] C.-N. Fan, C.-H. Xu, C.-Q. Liu, Z.-Y. Huang, J.-Y. Liu, Z.X. Ye, Reac. Kinet. Mech. 0100-0043.
Cat. 107 (2012) 189–202, https://doi.org/10.1007/s11144-012-0456-y. [43] G. Busca, Chem. Rev. 107 (2007) 5366–5410, https://doi.org/10.1021/cr068042e.
[21] V.R. Bakuru, S.R. Churipard, S.P. Maradur, S.B. Kalidindi, Dalton Trans. 48 (2019) [44] M.A. Camblor, A. Corma, H. García, V. Semmer-Herlédan, S. Valencia, J. Catal. 177
843–847, https://doi.org/10.1039/C8DT03512A. (1998) 267–272, https://doi.org/10.1006/jcat.1998.2110.
[22] V. Calvino-Casilda, K. Stawicka, M. Trejda, M. Ziolek, M.A. Bañares, J. Phys. Chem. [45] E.E. Mallon, M.Y. Jeon, M. Navarro, A. Bhan, M. Tsapatsis, Langmuir 29 (2013)
C 118 (20) (2014) 10780–10791, https://doi.org/10.1021/jp500651e. 6546–6555, https://doi.org/10.1021/la4001494.
[23] K. Stawicka, A.E. Díaz-Álvarez, V. Calvino-Casilda, M. Trejda, M.A. Bañares, [46] G.A.H. Mekhemer, A.K.H. Nohman, N.E. Fouad, H.A. Khalaf, Colloid Surf. A:
M. Ziolek, J. Phys. Chem. C 120 (30) (2016) 16699–16711, https://doi.org/10. Physiocochem. Eng. Aspects 161 (2000) 439–446, https://doi.org/10.1016/S0927-
1021/acs.jpcc.6b04229. 7757(99)00123-5.
[24] L. Zhao, J. Gao, C. Xu, B. Shen, Fuel Process. Technol. 92 (2011) 414–420, https:// [47] K. Góra-Marek, J. Datka, S. Dzwigaj, M. Che, J. Phys. Chem. B 110 (2006)
doi.org/10.1016/j.fuproc.2010.10.003. 6763–6767, https://doi.org/10.1021/jp0582890.
[25] A.J.J. Koekkoek, H. Xin, Q. Yang, C. Li, E.J.M. Hensen, Microporous Mesoporous [48] E.P.L. Hunter, S.G. Lias, J. Phys. Chem. Ref. Data 27 (1998) 413–656, https://doi.
Mater. 145 (2011) 172–181, https://doi.org/10.1016/j.micromeso.2011.05.013. org/10.1063/1.556018.
[26] K.A. Tarach, K. Góra-Marek, J. Martinez-Triguero, I. Melián-Cabrera, Catal. Sci. [49] M. Trejda, K. Stawicka, M. Ziolek, Appl. Catal. B: Environ. 103 (2011) 404–412,
Technol. 7 (2017) 858–873, https://doi.org/10.1039/c6cy02609e. https://doi.org/10.1016/j.apcatb.2011.02.003.
[27] K. Sadowska, K. Góra-Marek, J. Datka, J. Phys. Chem. C 117 (18) (2013) [50] L. Chen, B. Nohair, D. Zhao, S. Kaliaguine, Appl. Catal. A Gen. 549 (2018) 207–215,
9237–9244, https://doi.org/10.1021/jp400400t. https://doi.org/10.1016/j.apcata.2017.09.027.
[28] J. Kowalska-Kus, A. Held, M. Frankowski, K. Nowińska, J. Mol. Catal. A Chem. 426 [51] G.S. Nair, E. Adrijanto, A. Alsalme, I.V. Kozhevnikov, D.J. Cooke, D.R. Brown,
(2017) 205–212, https://doi.org/10.1016/j.molcata.2016.11.018. N.R. Shiju, Catal. Sci. Technol. 2 (2012) 1173–1179, https://doi.org/10.1039/
[29] Y.-P. Guo, H.-J. Wang, Y.-J. Guo, L.-H. Guo, L.-F. Chu, C.-X. Guo, Chem. Eng. J. 166 C2CY00335J.
(2011) 391–400, https://doi.org/10.1016/j.cej.2010.10.057. [52] L.H. Vieira, L.G. Possato, T.F. Chaves, S.H. Pulcinelli, C.V. Santilli, L. Martins, Mol.
[30] Z. Qin, L. Lakiss, J.-P. Gilson, K. Thomas, J.-M. Goupil, C. Fernandez, V. Valtchev, Catal. 458 (2018) 161–170, https://doi.org/10.1016/j.mcat.2017.11.027.
Chem. Mater. 25 (2013) 2759–2766, https://doi.org/10.1021/cm400719z. [53] L. Li, T.I. Korányi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012) 1611–1619,
[31] Z. Qin, J.-P. Gilson, V. Valtchev, Curr. Opin. Chem. Eng. 8 (2015) 1–6, https://doi. https://doi.org/10.1039/C2GC16619D.
org/10.1016/j.coche.2015.01.002. [54] S. Zaher, L. Christ, M. Abd El Rahim, A. Kanj, I. Karamé, Mol. Catal. 438 (2017)
[32] X. Yang, Ch. Gan, H. Xiong, L. Huang, X. Luo, RSC Adv. 6 (2016) 105737–105743, 204–213, https://doi.org/10.1016/j.mcat.2017.06.006.
https://doi.org/10.1039/C6RA23883A. [55] C. Morterra, G. Cerrato, V. Bolis, S. Di Ciero, M. Signoretto, J. Chem. Soc., Faraday
[33] L. Brabec, M. Kocirik, J. Phys. Chem. C 114 (32) (2010) 13685–13694, https://doi. Trans. 93 (1997) 1179–1184, https://doi.org/10.1039/A606608I.
org/10.1021/jp1027228. [56] C. Morterra, G. Cerrato, F. Pinna, M. Signoretto, G. Strukul, J. Catal. 149 (1994)
[34] Y. Tao, H. Kanoh, K. Kaneko, Adv. Mater. 17 (2005) 2789–2792, https://doi.org/ 181–188, https://doi.org/10.1006/jcat.1994.1283.
10.1002/adma.200401275. [57] D. Sanfilippo, I. Miracca, Catal. Today 111 (2006) 133–139, https://doi.org/10.
[35] D. Ohayon, R. Le Van Mao, D. Ciaravino, H. Hazel, A. Cochennec, N. Rolland, Appl. 1016/j.cattod.2005.10.012.
Catal. A: General 217 (2001) 241–251, https://doi.org/10.1016/S0926-860X(01) [58] D.T. Griffen, Silicate Crystal Chemistry, Oxford University Press, Oxford, U.K, 1992
00611-1. ISBN 0-19-504442-8.
[36] A.-B. Fernández, A. Marinas, T. Blasco, V. Fornés, A. Corma, J. Catal. 243 (2006) [59] C.X.A. da Silva, V.L.C. Gonçalves, C.J.A. Mota, Green Chem. 11 (2009) 38–41,
270–277, https://doi.org/10.1016/j.jcat.2006.06.029. https://doi.org/10.1039/B813564A.
[37] T. Karbowiak, M.-A. Saada, S. Rigolet, A. Ballandras, G. Weber, I. Bezverkhyy, [60] N. Narkhede, A. Patel, RSC Adv. 4 (2014) 19294–19301, https://doi.org/10.1039/
M. Soulard, J. Patarin, J.-P. Bellat, Phys. Chem. Chem. Phys. 12 (2010) C4RA01851F.
11454–11466, https://doi.org/10.1039/c000931h. [61] C.W. Kammeyer, D.R. Whitman, J. Chem. Phys. 56 (1972) 4419–4421, https://doi.
[38] A. Rimola, D. Costa, M. Sodupe, J.-F. Lambert, P. Ugliengo, Chem. Rev. 113 (2013) org/10.1063/1.1677883.
4216–4313, https://doi.org/10.1021/cr3003054.

10
 Fuel 281 (2020) 118724

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

An efficient process to synthesize solketal from glycerol over tin (II) T

silicotungstate catalyst
Márcio Jose da Silva , Milena Galdino Teixeira, Diego Morais Chaves, Lucas Siqueira
⁎

Chemistry Department, Federal University of Viçosa, Viçosa, Minas Gerais State Zip Code 36570-900, Brazil

GRAPHICAL ABSTRACT

ARTICLE INFO ABSTRACT

Keywords: In this work, Sn(II)-exchanged silicotungstic acid salt (i.e., Sn2SiW12O40) was synthesized and evaluated as the
Sn(II) silicotungstate catalysts catalyst on the acetalization of glycerol with acetone to produce solketal, a versatile bioadditive of fuel. The
Bioadditives Sn2SiW12O40 salt was compared to the other solid Keggin heteropoly salts (i.e., Sn3/4PMo12O40, Sn3/4PW12O40),
Glycerol and liquid (i.e., HCl, H2SO4 and p-toluenesulfonic acid) catalysts. Amongst the catalysts assessed, it was the most
Solketal
active, achieving a high conversion (ca. > 99%, after 1 h reaction at room temperature) and selective (ca. 97%)
toward the formation of solketal. Moreover, the Sn2SiW12O40 salt demonstrated to be more active than acid and
precursor Tin (II) chloride salt, as well as other heteropoly salts and solid supported catalysts. The effects of the
main reaction variables were assessed. The Sn(II) cation, as well as the silicotungstate anion, showed being
essential to convert glycerol to solketal. Insights on the reaction mechanism were performed. In a simple recycle
procedure, the solketal was purified, the acetone excess recycled, and the catalyst was reused without loss
activity.

⁎
Corresponding author at: Chemistry Department, Federal University of Viçosa, Campus Universitário, Avenue P.H. Rolfs, s/n, Viçosa, Minas Gerais State Zip Code
36570-000, Brazil.
E-mail address: [email protected] (M.J. da Silva).

https://doi.org/10.1016/j.fuel.2020.118724
Received 18 May 2020; Received in revised form 8 July 2020; Accepted 12 July 2020
Available online 21 July 2020
0016-2361/ © 2020 Elsevier Ltd. All rights reserved.
M.J. da Silva, et al. Fuel 281 (2020) 118724

1. Introduction exchanging by larger radium metal cations make HPA salts insoluble in
polar solvents [36]. Moreover, when other Lewis acid metal cations
The current increasing environmental awareness and the inevitable replace the protons, their catalytic properties can be significantly en-
depletion of fossil fuel reserves have prompted the growth of search for hanced toward goal-process [37,38]. Recently, Sn(II)-exchanged phos-
renewable energy sources, with a greater increment in the biodiesel photungstic acid salts demonstrated to be an efficient catalyst in gly-
industry [1,2]. Biodiesel has been typically produced by the alkaline cerol esterification reactions, achieving high ester yielding [39,40]. The
transesterification of triglycerides present in the vegetable oils, re- same was verified when it was used to synthesize tert-butyl ethers of
sulting in the co-production of glycerol in 10 wt% [3]. The actual glycerol [41].
market is still unable to consume this large surplus of glycerol, mainly In this work, the protons of silicotungstic acid were exchanged by
because the processes of purification are expensive to be performed at a Sn2+ cations, and the salt formed was used as a catalyst in reactions of
large scale [4,5]. Therefore, to develop processes to convert crude glycerol condensation with acetone to produce solketal. The catalytic
glycerol to high value-added products arising as an attractive option to activity of Sn2SiW12O40 was compared to the Lewis and Brønsted acid
consumption the glycerol generated by the biodiesel industry [6]. catalysts including other Keggin HPAs. The impacts of the main reac-
Considering this current trend, several routes to obtain valuable tion variables were investigated. Insights on the reaction mechanism
chemicals such as the bio-solvents, surfactants, polymers, and the more were performed. The reusability of the catalyst was successfully eval-
highlighted glycerol derivatives have been proposed [7–10]. Among uated.
them, glycerol derived bioadditives has attracted very attention, mainly
because these compounds can be blended to the diesel or gasoline, re- 2. Experimental section
ducing the emission of particulate material, and improving the physi-
cochemical properties of these liquid fuels [11–15]. 2.1. Chemicals
The reactions of glycerol with acetic acid or tert-butyl alcohol give
mono-, di- and tri-substituted glycerol derivatives, which are highly All chemicals were acquired from commercial sources and utilized
valuables compounds due to their properties as fuel bioadditives without prior handling as received. Glycerol (99.5 wt%), acetone
[16–19]. Recently, another glycerol derivative has also attracted at- (99 wt%), and dodecane (99 wt%) were purchased from Vetec. SnCl2
tention as a compound able to improve the properties of diesel fossil and H4SiW12O40∙n H2O, all 99.9 wt%) were purchased from Sigma
and gasoline; the solketal (i.e., 2,2-dimethyl-1,3-dioxolane-4-me- Aldrich.
thanol). It can be used to decrease gum formation, increase the octane
index, diminish viscosity, improve the flashpoint, and oxidation stabi- 2.2. Synthesis and characterization of Sn2SiW12O40 salt catalyst
lity [15,20–22].
Solketal, as well as ethers and esters of glycerol, are compounds Sn2SiW12O40 salt was prepared through a procedure adapted from
synthesized through the acid-catalyzed reactions, however, some of literature [41]. Usually, a SnCl2 solution at a stoichiometric amount
these processes involve the use of homogeneous catalysts which are was slowly added to an aqueous solution containing the solved
highly corrosive in nature and environmentally dangerous [23]. Some H4SiW12O40. The mixture obtained was magnetically stirred by 3 h at
works have demonstrated being possible use homogenous acid catalysts 333 K, followed by the evaporation to dryness to releasing the gaseous
as sulfuric acid to promote the hydrolysis of solketal to generate gly- HCl. The solid was dried in an oven at 423 K/ 3 h.
cerol with high level of purity, as the requirements of food and phar-
maceutical industry [24]. Nonetheless, these soluble catalysts have still 2.3. Characterization of the silicotungstic catalysts
serious drawbacks. The use of a homogeneous catalyst involves a te-
dious workup procedure to be separated from products, which result in Aiming a comparison, all the characterization data of pristine sili-
a large generation of effluent and neutralization residues, which should cotungstic acid were also analyzed. The infrared spectroscopy analyses
be disposed into the environment [25]. were performed in a Varian 660-IR spectrometer coupled to the atte-
Acidic solids can circumvent the drawbacks of the homogeneous nuated total reflectance accessory (FT-IR/ATR). The patterns of X-rays
catalysts, therefore, they have been used in newer cost-effective and diffraction (XRD) of the silicotungstic catalysts were recorded in an
selective processes to convert glycerol to solketal; sulfonic acid resins, XRD-rays diffraction system model D8-Discover Bruker using Ni filtered
metal oxides, zeolites, and solid supported catalysts are only some ex- Cu-kα radiation (λ = 1.5418 Å), at 40 kV and 40 mA, with time
amples [26–28]. The greater challenge of most of the heterogeneous counting 1.0 s, with diffraction angle (2θ) varying from 5 to 80°.
processes is to overcome the leaching of the active phase and the The H2 adsorption/ desorption isotherms were obtained in a NOVA
consequent deactivation of solid-supported catalysts. On this sense, 1200e High Speed, automated surface area and pore size analyzer
several works have assessed the tolerance of acidic catalysts to these Quantachrome instrument. Prior to the analysis, the sample was de-
challengers; while Amberlyst-36 resin was few effective and quickly gassed 5 h. The surface area was calculated by the Brunauer-Emmett-
deactivated, hydrate aluminum fluoride demonstrated to be a more cost Teller equation applied to the isotherms.
effective and efficient catalyst [29]. Similarly, zeolite Beta was sig- The strength of acidity of the silicotungstic catalysts was estimated
nificantly more resistant to the presence of water than Amberlyst-15 measuring the initial electrode potential (i.e. Bel, model W3B) of a
resin [30]. CH3CN solution containing the soluble or suspended sample. The acidic
Keggin heteropolyacids (HPAs) are attractive catalysts and are sites number was determined by potentiometric titration as follows;
highly active in oxidative processes or acid-catalyzed reactions [31]. typically, the sample (ca. 50 mg) was magnetically stirred in CH3CN by
Phosphotungstic acid is the strongest Brønsted acid among the Keggin 3 h and then titrated with an n-butylamine solution in toluene (ca.
HPAs; it is soluble in solvent polar and has been successfully used as a 0.025 mol L-1) [42].
homogeneous catalyst in reactions to converting glycerol to bioaddi- Thermal analyses (TG) were performed in a Perkin Elmer
tives [32]. Nonetheless, when solid, it has a low surface area, ham- Simultaneous Thermal Analyzer (STA) 6000. Typically, a sample (ca.
pering its use in conditions of heterogeneous catalysis. Therefore, they 10 mg) was heated at a rate of 10 Kmin−1 under nitrogen flow. The
have been used as solid supported catalysts in several acid-catalyzed temperature of the TG curves varied from 303 to 973 K.
reactions [33,34]. The elemental composition of silicotungstate salt was confirmed in
To circumvent the undesirable problem of leaching, the unique an energy dispersive spectrometry system (EDS). The scanning electron
chemical-physical properties of HPAs can be easily manipulated with microscopy (SEM) images were taken in a JEOL JSM-6010/LA micro-
suitable tailoring of their constitution [35]. For instance, the protons scope. A working distance of 10 mm and 20 KV acceleration voltage

2
M.J. da Silva, et al. Fuel 281 (2020) 118724

were used to acquire SEM images and EDS spectra.

2.4. Catalytic tests

Catalytic runs were carried out using glycerol, acetone, and an

adequate catalyst. In a typical reaction, glycerol and acetone at an
adequate proportion were magnetically stirred until complete solubi-
lization at room temperature at an adequate molar ratio. After adding
the catalyst, the reaction was started and carried out 4 h.
The reaction progress was periodically monitored collecting ali-
quots at regular intervals of time, and analyzing them by gas chroma-
tography (Shimadzu 2014, FID, Carbowax 20 M capillary column).
Prior to the analysis, the samples were diluted in methanol. The tem-
perature profile was as follows: 80 °C/ 3 min; 10 °C / min; final tem-
perature 210 °C/ 3 min. Injector and detector were both kept at 250 °C.
The glycerol conversion was calculated according to the Equation (1):

A solkcorrected Fig. 1. Effect of the catalyst on glycerol acetalization with acetonea aReation
Conversion(%) = x100
A glycerol + A solkcorrected (1) conditions: glycerol (38.0 mmol), acetone (152.0 mmol), catalyst (0.01 mol %),
298 K.
where Aglycerol, is the unreacted glycerol chromatographic peak area
and Asolketal corrected is the corrected chromatographic peak area of
almost untouched after the synthesis (Fig. 1SM).
•
solketal, obtained by the ratio of glycerol chromatographic peak area/
Powder XRD patterns analysis can be useful to verify if any changes
solketal chromatographic peak area, injected with same concentration.
happened on the secondary structure of HPA when the protons are
The reaction products were identified by mass spectrometry ana-
exchanged by other cations. X-rays diffractograms of Sn(II) silico-
lysis, performed in a Shimadzu GC 2010 gas chromatograph coupled
tungstate and its synthesis precursors (i.e. SnCl2 and H4SiW12O40)
with a MS-QP 2010 Ultra, with a carbowax capillary column
evidenced that secondary structure of HPA presented only a little bit
(0.25 μm × 0.25 mm × 30 m), and He as the carrier gas at 2mLmin−1.
changes (Fig. 2SM); although new diffraction lines were noticed at
The temperature program was equal to the GC analyses. The GC injector
low angle region (ca. 10°, 2 θ) of diffractogram of the salt, in gen-
and MS ions source temperatures were 523 and 533 K, respectively. The
eral, their profile was very similar to the acid. These changes are
MS detector operated in the EI mode at 70 eV, with a scanning range of
assigned the difference between the ionic radius of hydrate protons
m/z 50–400.
(i.e., H3O+, H2O5+) and the Sn2+ ions, that may affect the packa-
ging of the heteropolyanions on the secondary structure [41].
•
2.5. Reuse of the catalyst
The crystallite size was measured of silicotungstic catalysts was
measured applying the Scherrer to the most intense XRD peaks.
The Sn2SiW12O40 catalyst was reused after a simple procedure. At
Values varied from 24 to 42 nm for the salts, while the acid pre-
the end of the reaction, the solution was vapored under vacuum to
sented values of 26 nm.
•
remove the excess acetone, which was recovered to be used in another
The strength of acidic sites belonging to the H4SiW12O40 and
reuse cycle. The remaining liquid contains solketal, catalyst, and a
Sn2SiW12O40 was estimated measuring the initial electrode potential
small amount of unreacted glycerol (i.e., quantified by GC analysis of
of their acetonitrile solutions; values of 713 mV and 685 mV, were
aliquot when the reaction was stopped). After three steps of liquid–li-
obtained for the acid and silicotungstic salt, respectively (insert on
quid extraction with ethyl acetate, the solketal was removed. A simple
Fig. 3SM). It allowed us to classify the acidic sites of both catalysts
distillation provided the ethyl acetate, which was recycled, and the
as very strong [42]. Through the potentiometric titration curves was
pure solketal. To the reactor containing catalyst and the unreacted
possible to calculate the acid sites number of silicotungstic catalysts;
glycerol, recovered acetone and fresh glycerol were added, to start
1.3 and 1.2 meq g−1 were the values obtained for the acid and slat
another cycle of reuse. This procedure was four times repeated.
respectively (Fig. 3SM). Literature has explained the Brønsted
acidity is due to hydrolysis undergone by the metal cations
2.6. Results and discussion

2.6.1. Catalyst characterization

The characterization of the Sn(II) silicotungstate catalyst was pre-
viously discussed in another work, where they were used in an one-pot
synthesis of alkyl levulinates from biomass derivative carbohydrates
[43]. Notwithstanding, all the important data obtained in character-
ization (i.e., infrared spectra, powder XRD patterns, porosity properties,
BET surface area, EDS analyses, and measurements of the strength of
acidic sites are shown in the supplemental material (Fig. 1SM-6SM).
The main characterization data and the respective conclusions are
summarized as follow:

• The integrity of the primary structure of HPAs catalysts were con-

firmed by infrared spectroscopy analysis, after to check the finger-
print region that should contain the characteristic absorption bands
of Keggin anion [44]. A comparison of the infrared spectra of tin(II)
Fig. 2. Comparison of solketal yields of this study and other reported works.
silicotungstate salt and precursor HPA clearly showed that the pri- Reaction conditions: glycerol: acetone molar ratio (1:4), reaction time = 2 h,
mary structure of catalyst (i.e., Keggin heteropolyanion) was kept 298 K. areaction time = 1 h.

3
M.J. da Silva, et al. Fuel 281 (2020) 118724

Fig. 3. The impacts of variation on stoichiometry of reactants on Sn2SiW12O40-

catalyzed glycerol condensation with acetonea aReaction conditions: glycerol: Fig. 5. Effect of catalyst load on the reaction selectivity.
acetone (variable), volume (10 mL), catalyst (0.01 mol %), 298 K.

coordinated to the water molecules, which result in releasing of H+

cations [39–41].
• Larger metal cation salts with an ionic radius > 1.3 Å, like Cs+ ions,
make water-insoluble the HPA salts, in addition, they increase its
surface area [43]. Nonetheless, the Sn2+ ions included in the
H4SiW12O40 have a small radius (≈ 1.3 Å). The Sn(II) silicotung-
state salt almost insoluble in acetone. During the reaction, the water
formed favor the solubility of salt; however, at the end of the pro-
cess, the suspension should be centrifugate to give solid at side
bottom of the reactor.
• The MEV images reveled that silicotungstic acid has like rice grains,
while the salt presented particles with greater size (Fig. 4SM). The
EDS analyses confirmed the elemental composition of metal silico-
tungstate salts (Fig. 5SM).
•
Fig. 6. Comparison of solketal yields among this study and other work at 298 K
The hydration level of silicotungstic catalysts was determined by
and 1:6 glycerol to acetone aReaction temperature = 343 K. b1:12 glycerol to
TG/ DTG curves. While the silicotungstic acid had 14 mol of H2O/ acetone. RT = reaction time.
mol of catalyst, their salt presented 7 water mol per mol catalyst
(Fig. 6SM).
Additionally, liquid Brønsted acid catalysts (i.e., HCl, H2SO4 and p-to-
luenesulfonic acid (PTSA)) were also assessed.
2.7. Catalytic runs Regardless of acetone excess (ca. 1: 4 glycerol to acetone), without
catalyst, no conversion of glycerol was noticed. Conversely, despite the
2.7.1. Effect of catalyst nature low catalyst concentration used, in the presence of Lewis or Brønsted
Aiming to investigate the effect of Keggin anion on activity and acid catalysts the overall selectivity–defined as the percentage ratio of
selectivity of tin salt catalysts we carried out reactions shown in Fig. 1. the desired acetalization products (i.e., a total of isomers I and II) with
respect to the conversion–was always > 97% for solketal (Scheme 1).
It is important to note that the reaction conditions were not opti-
mized to achieve the maximum yield. Another important point is that
the Brønsted acid catalysts were used at the same H+ ions concentra-
tion; similarly, the metal salts were used with the same Sn2+ ions
concentration (i.e., 0.01 mol %). Fig. 1 shows that the Sn(II) heteropoly
salts had superior performance to that of Brønsted acids. Moreover,
amongst them, Sn2SiW12O40 was the most active catalyst.
Previously, we have exploited the activity of different catalysts (i.e.,
SnCl2, SnF2, H3PW12O40) in condensation reactions of glycerol with
acetone at the similar reaction conditions used herein (ca. 298 K, 1:4
glycerol to acetone). The main results are summarized in Fig. 2. It is
known that in general a catalyst does not affect the thermodynamic
equilibrium but only changes the kinetics. Nonetheless, our intention
was to compare the conversions achieved within a specific time interval
(see footnote of Fig. 2), independent of equilibrium has been reached or
not. After comparison, it is possible to realize that even used at the
Fig. 4. The effects of catalyst load on the kinetic curves of Sn2SiW12O40-cata- higher loads, the tin halides (i.e. (soluble) SnCl2, (solid) SnF2)), and the
lyzed glycerol condensation with acetonea aReaction conditions: glycerol: (soluble) phosphotungstate acid, achieved lower conversions (Fig. 2)
acetone (1:12), volume (10 mL), Sn2SiW12O40 catalyst (variable), 298 K. [45–47].

4
M.J. da Silva, et al. Fuel 281 (2020) 118724

Scheme 1. Lewis or Brønsted acid-catalyzed glycerol acetalization with acetone.

When the Brønsted acid catalysts are totally soluble, their activity Therefore, we suppose that both mechanisms may be operating (step I,
on glycerol acetalization can be linked to their strength of acidity, Scheme 2); the activation of the carbonyl group may be triggered by
which was estimated by pKa measurements in different solvents; protonation and or coordination to the Sn2+ ions of silicotungstic salt
H3PW12O40 > H4SiW12O40 > H3PMo12O40 > H2SO4 > HCl catalyst.
[48,49]. On this regard, we had evaluated the activity of soluble Keggin Another key aspect is the regioselectivity of process; although six-
HPAs and verified that a using 1:20 M ratio of glycerol to acetone and membered ring compounds are thermodynamically more stable than
3.0 mol % of catalyst load, the reactions in the presence of H3PW12O40, five-membered ones, the solketal (I) was always the most selectively
H4SiW12O40 or H3PMo12O40 achieved conversions of 91, 39 and 41% formed product herein. We assigned this preferential formation to an
after 2 h reaction [47]. Comparing the conversions of HPA-catalyzed easier attack of the secondary hydroxyl group on the charged positively
reactions and those in the presence of their salts we conclude that the carbonyl group (i.e., 1a intermediate, Scheme 2) if compared to the
presence of Sn(II) ions remarkably improves the performance of acid attack of the primary hydroxyl group (i.e., 1b intermediate, Scheme 2).
catalysts. Moreover, Mota et al., demonstrated that the methyl group in the axial
This synergic effect between the heteropolyanion and the tin cation position dioxane isomer repulsively interacts with two axial hydrogens
was previously reported in other acid-catalyzed reactions such as gly- of six members ring, make him less stable than dioxolane (i.e., solketal)
cerol and glycol etherification reactions [48,49]. The higher softness of [52]. Regardless of the catalyst, all the reactions in Fig. 2 provided
silicotungstic anion makes him more efficient to stabilize positively solketal with an average selectivity of 97%.
charged intermediates, an aspect that besides the high Lewis acid of The effect of the variation in the stoichiometry of the reactants was
Sn2+ ions may be useful to explain the highest activity of Sn2SiW12O40 also investigated and the main results are displayed in Fig. 3. It is im-
catalyst [41,48,49]. Moreover, the Sn2+ ions can undergo hydrolyzes, portant to note that the effects of diffusional limitations were also as-
reacting with residual water or generated along the process; conse- sessed, performing reactions with different molar ratios at different
quently, the H+ ions produced may itself catalyze the reaction. A stirring rates (Table 1SM). No significant changes in conversions of
possible reaction pathway is depicted in Scheme 2. reactions were observed using three distinct levels of the stirring speed
The literature has described that to be condensate with glycerol, the rate.
acetone should have their carbonyl group activated through the pro- Since that, the acetalization of glycerol is a reversible reaction, an
tonation step or polarization by the coordination to the metal cation. In increase of acetone load shifted the equilibrium toward the products,
this work, these two mechanisms (i.e., Brønsted and Lewis acid-cata- increasing both initial rate and final conversion of the reactions.
lyzed) may be operating. Recently, we demonstrated through pyridine Conversely, no significant changes in the selectivity were observed, and
adsorption measurements by infrared spectroscopy that the solketal was always the main product (ca. 89–97% selectivity), re-
Sn2SiW12O40 catalyst has these two types of acidic sites [48,49]. gardless of the excess of acetone.

Scheme 2. Possible pathway of Sn2SiW12O40-catalyzed reaction of glycerol with acetone (adapted from refs. [29,50,51]).

5
M.J. da Silva, et al. Fuel 281 (2020) 118724

Scheme 3. Procedure to recovery and reuse the catalyst, recycle acetone, and purify the solketal.

Nanda et al. performed a comparison of impacts of the catalyst one (Fig. 5). It means that the reaction selectivity was under kinetic control,
yield of the glycerol ketalization processes developed in both batch and which was impacted by the catalyst load. When a high catalyst load was
continuous reactors [28]. They conclude that continuous-flow processes used, the most stable product was always more selectively formed since
are more promising in large scale than batch processes. Those authors the reaction beginning. Whereas, when the catalyst was present at low
verified that since the equilibrium constant of this reaction is low, the loads, the reaction becomes slower, and the product kinetically favor-
best yields are achieved in continuous systems, where an excess of able although always the minority, had its formation enhanced, mainly
acetone was used, or the water generated is continuously removed. The in the initial period of the process.
best performance was achieved in the Amberlyst 36-catalyzed reac- In Fig. 6, a quick survey of literature highlights the main results
tions, at a proportion of 4:1 acetone to glycerol, 298 K; nonetheless, it achieved in HPA salts-catalyzed glycerol acetalization reactions with
was achieved at continuous-flow conditions (ca. WHSV of 2 h-) and high acetone [53-55]. The physical properties of heteropoly salt depend on
pressures (ca. 500 psi ≈ 34 atm) [28]. the cation nature used to replace the protons of Keggin acids. In special,
Due to slight enhancement obtained on the conversion with reac- the solubility of the salt impacts the workup needs to separate the
tions at higher proportions, 1:12 was the molar ratio selected to assess catalyst from the medium of reaction. For these reasons, although the
the other reaction variables. The effect of catalyst load was then eval- HPA salts containing cations with large ionic radius are almost in-
uated using this proportion at room temperature and the kinetic curves soluble in the reaction, sometimes it is difficult to separate or recovery
are displayed in Fig. 4. the catalyst. The cesium phosphotungstate is an example, and conse-
Different from the observed assessing the effect of the reactant quently, it has been used solid-supported [54].
stoichiometry, a variation on the catalyst load had a noticeable impact Although a better comparison requires to take into account the
on the reaction selectivity. Independent of the catalyst load, the con- amount of supported catalyst used in the reactions and its loading in the
version and the selectivity of reactions were almost similar after 1 h run support, it is possible realize that the Sn2SiW12O40 catalyst has a per-
(Fig. 4). formance equal or superior to the majority of the solid heteropoly salts
Regardless of the catalyst load, solketal was always obtained with or solid supported shown in Fig. 6.
selectivity equal to or higher than 82%. Nonetheless, during the reac- The reusability of Sn(II) silicotungstate was also assessed. The sus-
tions was verified that the less stable product (i.e., dioxane), even being pension formed by the catalyst in the reaction solution requires that it
always the minor product, had its formation more favored in the initial should be centrifugated to be recovered and reused. To circumvent this
period of reactions, mainly when the catalyst was present in lower load drawback, we envisaged a simple process to recover the catalyst. In this

6
M.J. da Silva, et al. Fuel 281 (2020) 118724

procedure, the excess acetone is removed and reused in another run. separation technologies. J Chem Technol Biotechnol 2012;87:861–79. https://doi.
The reactor contains unreacted glycerol, solketal, and catalyst. The org/10.1002/jctb.3785.
[4] Tan HW, Aziz ARA, Aroua MK. Glycerol production and its applications as a raw
addition of ethyl acetate solubilizes the solketal, which is then ex- material: a review. Renew Sust Energ Rev 2013;27:118–27. https://doi.org/10.
tracted. To the reactor containing unreacted glycerol and the catalyst, 1016/j.rser.2013.06.035.
recovered acetone is added and then a new run is carried out (Scheme [5] Kong PS, Aroua MK, Daud WMAW. Conversion of crude glycerol and pure glycerol
into derivatives: a feasibility evaluation. Renew Sust Energ Rev 2016;63:533–55.
3). https://doi.org/10.1016/j.rser.2016.05.054.
It is noteworthy that even if the catalyst had been recovered by [6] Ardi MS, Aroua MK, Hashim NA. Progress, prospect, and challenges in glycerol
filtration, the steps of recovery of the acetone and purification of the purification process: A review. Renew Sust Energ Rev 2015;42:1164–73. https://
doi.org/10.1016/j.rser.2014.10.091.
solketal should be also performed, therefore, they are not additional but [7] Leal-Duaso A, Pérez P, Mayoral JA, Garciá JI, Pires E. Glycerol-derived solvents:
obligatory steps of the process. synthesis and properties of symmetric glyceryl diethers. ACS Sustain Chem Eng
Therefore, following the procedure in Scheme 3, we successfully 2019;7:13004–14https://doi:10.1021/acssuschemeng.9b02105.
[8] Leal-Duaso A, Pérez P, Mayoral JA, Pires E, García JI. Glycerol as a source of de-
recovered and reused the Sn2SiW12O40 salt in 4 catalytic runs. No de-
signer solvents: Physicochemical properties of low melting mixtures containing
crease in the catalytic performance was noticed. Infrared analysis of glycerol ethers and ammonium salts. Phys Chem Chem Phys
catalyst after the last recycle showed that no significant change was 2017;19:28302–12https://doi:10.1039/c7cp04987k.
observed at the fingerprint region of the Keggin anion spectrum. [9] Velázquez D, Mayoral J, García-Peiro J, García J, Leal-Duaso A, Pires E.
Optimization of the synthesis of glycerol derived monoethers from glycidol by
means of heterogeneous acid catalysis. Molecules 2018;23:1–10https://doi:10.
3. Conclusions 3390/molecules23112887.
[10] Monteiro MR, Kugelmeier CL, Pinheiro RS, Batalha MO, César AS. Glycerol from
biodiesel production: Technological paths for sustainability. Renew Sust Energ Rev
In this work, the activity of Sn2SiW12O40 salt was assessed on the 2018;88:109–22. https://doi.org/10.1016/j.rser.2018.02.019.
acetalization of glycerol with acetone. The catalyst was spectro- [11] Samoilov VO, Maximov AL, Stolonogova TI, Chernysheva EA, Kapustin VM,
scopically characterized and demonstrated that after the exchange of Karpunina AO. Glycerol to renewable fuel oxygenates. Part I: Comparison between
solketal and its methyl ether. Fuel 2019;249:486–95. https://doi.org/10.1016/j.
the protons by Sn(II) cations, no modification in the primary structure fuel.2019.02.098.
(i.e., Keggin anion) was detected. In all the runs, glycerol was majority [12] Ballotin FC, da Silva MJ, Teixeira APC, Lago RM. Amphiphilic acid carbon catalysts
converted to solketal. The effects of the main reaction variables were produced by bio-oil sulfonation for solvent-free glycerol ketalization, Fuel
2020;274;117799-117805.https://doi.org/10.1016/j.fuel.2020.117799.
investigated. We have found that catalyst concentration affects the re- [13] Agirre I, Güemez MB, Ugarte A, Requies J, Barrio VL, Cambra JF, et al. Glycerol
action selectivity; when lower loads are used, the six-membered ring acetals as diesel additives: kinetic study of the reaction between glycerol and
dioxane has its formation favored, although the solketal remains even acetaldehyde. Fuel Process Technol 2013;116:182–8. https://doi.org/10.1016/j.
fuproc.2013.05.014.
as the main product. The Sn(II) cation showed to be a key constituent of
[14] Alptekin E, Canakci M. Performance and Emission Characteristics of Solketal-
catalyst. On the other hand, among the three catalysts with different Gasoline fuel blend in a vehicle with spark ignition engine. Appl Therm Eng
Keggin anions, that containing the silicotungstic anion was the more 2017;124:504–9https://doi:10.1016/j.applthermaleng.2017.06.064.
efficient. Finally, the catalyst was recovered and reused without loss of [15] Mota CJA, Da Silva CXA, Rosenbach N, Costa J, Da Silva F. Glycerin derivatives as
fuel additives: The addition of glycerol/acetone ketal (solketal) in gasolines. Energ
activity. Fuels 2010;24:2733–6https://doi:10.1021/ef9015735.
[16] Neto ABS, Oliveira AC, Rodriguez-Castellón E, Campos AF, Freire PTC, Sousa FFF,
CRediT authorship contribution statement et al. A comparative study on porous solid acid oxides as catalysts in the ester-
ification of glycerol with acetic acid. Catal Today 2018https://doi:10.1016/j.
cattod.2018.05.057.
Márcio Jose da Silva: Conceptualization, Methodology, Data [17] Costa BOD, Decolatti HP, Legnoverde MS, Querini CA. Influence of acidic properties
curation, Investigation, Writing - original draft, Project administration, of different solid acid catalysts for glycerol acetylation. Catal Today
2017;289:222–30https://doi:10.1016/j.cattod.2016.09.01.
Funding acquisition. Milena Galdino Teixeira: . Diego Morais [18] Sutter M, Silva ED, Duguet N, Raoul Y, Métay E, Lemaire M. Glycerol ether
Chaves: . Lucas Siqueira: . synthesis: a bench test for green chemistry concepts and technologies. Chem Rev
2015;115:8609–51. https://doi.org/10.1021/cr5004002.
[19] Magar S, Kamble S, Mohanraj GT, Jana SK, Rode C. Solid-acid-catalyzed ether-
Declaration of Competing Interest
ification of glycerol to potential fuel additives. Energ Fuels 2017;31:12272–7.
https://doi.org/10.1021/acs.energyfuels.7b02213.
[20] Rossa V, da Pessanha Y, SP, Díaz GC, Câmara LDT, Pergher SBC, Aranda DAG..
The authors declare that they have no known competing financial
Reaction kinetic study of solketal production from glycerol ketalization with
interests or personal relationships that could have appeared to influ- acetone. Ind Eng Chem Res 2017;56:479–88.
ence the work reported in this paper. [21] Dmitriev GS, Terekhov AV, Zanaveskin LN, Khadzhiev SN, Zanaveskin KL,
Maksimov AL. Choice of a catalyst and technological scheme for synthesis of
solketal. Russ J Appl Chem 2016;89:1619–24https://doi:10.1134/
Acknowledgements S1070427216100094.
[22] Gonçalves M, Rodrigues R, Galhardo TS, Carvalho WA. Highly selective acetaliza-
The authors are grateful for the financial support from CNPq and tion of glycerol with acetone to solketal over acidic carbon-based catalysts from
biodiesel waste. Fuel 2016;181:46–54https://doi:10.1016/J.FUEL.2016.04.083.
FAPEMIG (Brasil). This study was financed in part by the Coordenação [23] Fatimah I, Sahroni I, Fadillah G, Musawwa MM, Mahlia TMI, Muraza O. Glycerol to
de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - solketal for fuel additive: Recent progress in heterogeneous catalysts. Energies
Finance Code 001. 2019;12..
[24] Dmitriev GS, Zanaveskin LN, Terekhov AV, Samoilov VO, Kozlovskii IA, Maksimov
AL. Technologies for Processing of Crude Glycerol from Biodiesel Production:
Appendix A. Supplementary data Synthesis of Solketal and Its Hydrolysis to Obtain Pure Glycerol. Russ J Appl Chem
2018;91:1478–85. https://doi.org/10.1134/S1070427218090100.
[25] Ilgen Oguzhan, Yerlikaya Senol, Akyurek Funda Oguzkaya. Synthesis of Solketal
Supplementary data to this article can be found online at https:// from Glycerol and Acetone over Amberlyst-46 to Produce an Oxygenated Fuel
doi.org/10.1016/j.fuel.2020.118724. Additive. Period. Polytech. Chem. Eng. 2016. https://doi.org/10.3311/PPch.8895.
[26] Cornejo A, Barrio I, Campoy M, Lázaro J, Navarrete B. Oxygenated fuel additives
from glycerol valorization. Main production pathways and effects on fuel properties
References
and engine performance: A critical review. Renew Sustain Energy Rev
2017;79:1400–13.
[1] Anuar MR, Abdullah AZ. Challenges in biodiesel industry with regards to feedstock, [27] Rodrigues A, Bordado J, Galhano R. Upgrading the glycerol from biodiesel pro-
environmental, social and sustainability issues: a critical review. Renew Sust Energ duction as a source of energy carriers and chemicals–A technological review for
Rev 2016;58:208–23. https://doi.org/10.1016/j.rser.2015.12.296. three chemical pathways. Energies 2017;10:1817–52. https://doi.org/10.3390/
[2] Ambat I, Srivastava V, Sillanpää M. Recent advancement in biodiesel production en10111817.
methodologies using various feedstock: A review. Renew Sust Energ Rev [28] Nanda MR, Zhang Y, Yuan Z, Qin W, Ghaziaskar HS, Xu C. Catalytic conversion of
2018;90:356–69. https://doi.org/10.1016/j.rser.2018.03.069. glycerol for sustainable production of solketal as a fuel additive: a review. Renew
[3] Kiss AA, Bildea CS. A review of biodiesel production by integrated reactive Sustain Energy Rev 2016;56:1022–31. https://doi.org/10.1016/j.rser.2015.12.008.

7
M.J. da Silva, et al. Fuel 281 (2020) 118724

[29] Guidi S, Noè M, Riello P, Perosa A, Selva M. Towards a rational design of a con- doi.org/10.1007/s10971-017-4428-6.
tinuous-flow method for the acetalization of crude glycerol: Scope and limitations [43] Pinheiro PF, Chaves DM, da Silva MJ. One-pot synthesis of alkyl levulinates from
of commercial Amberlyst-36 and AlF3∙3H2O as model catalysts. Molecules biomass derivative carbohydrates in tin (II) exchanged silicotungstates-catalyzed
2016;21:657. https://doi.org/10.3390/molecules21050657. reactions. Cellulose 2019;26:7953–69. https://doi.org/10.1007/s10570-019-
[30] da Silva CX, Mota CJ. The influence of impurities on the acid-catalyzed reaction of 02665-w.
glycerol with acetone. Biomass Bioenerg 2011;35:3547–51. https://doi.org/10. [44] Popa A, Sasca V, Bajuk-Bogdanovi D, Holclajtner-Antunovic I. Synthesis, char-
1016/j.biombioe.2011.05.004. acterization, and thermal stability of cobalt salts of Keggin-type heteropolyacids
[31] Kozhevnikov IV. Sustainable heterogeneous acid catalysis by heteropoly acids. J supported on mesoporous silica. J Therm Anal Calorim 2016;126:1567–77. https://
Mol Catal A 2007;262:86–92. https://doi.org/10.1016/j.molcata.2006.08.072. doi.org/10.1007/s10973-016-5650-0.
[32] Gonçalves CE, Laier LO, Cardoso AL, da Silva MJ. Bioadditive synthesis from [45] da Silva MJ, Julio AA, Dorigetto FCS. Solvent-free heteropolyacid-catalyzed gly-
H3PW12O40-catalyzed glycerol esterification with HOAc under mild reaction con- cerol ketalization at room temperature. RSC Adv. 2015;5:44499–506. https://doi.
ditions. Fuel Process Technol 2012;102:46–52. https://doi.org/10.1016/j.fuproc. org/10.1039/C4RA17090C.
2012.04.027. [46] da Silva MJ, Rodrigues FA, Julio AA. SnF2-catalyzed glycerol ketalization: a friendly
[33] da Silva MJ, Liberto NA. Soluble and solid supported Keggin heteropolyacids as environmentally process to synthesize solketal at room temperature over on solid
catalysts in reactions for biodiesel production: challenges and recent advances. Curr and reusable Lewis acid. Chem Engin J 2017;307:828–35. https://doi.org/10.1016/
Org Chem 2016;20:1263–83. https://doi.org/10.2174/ j.cej.2016.09.002.
1385272819666150907193100. [47] da Silva MJ, Guimaraes MO, Julio AA. A Highly regioselective and solvent-free Sn
[34] Tan J, Lu T, Zhang J, Xie B, Chen M, hu X.. Highly efficient and recyclable catalysts (II)-catalyzed glycerol ketals synthesis at room temperature. Catal Lett
SnCl2–xH3PW12O40/AC with Brønsted and Lewis acid sites for terephthalic acid 2015;145:769–76. https://doi.org/10.1007/s10562-015-1477-8.
esterification. J Taiwan Inst Chem Engineer 2018;86:18–24. https://doi.org/10. [48] da Silva Márcio José, Chaves Diego Morais, Júlio Armanda Aparecida, Rodrigues
1016/j.jtice.2018.03.012. Fabio Avila, Bruziquesi Carlos Giovani Oliveira. Sn(II)-Exchanged Keggin
[35] da Silva MJ, de Oliveira CM. Catalysis by Keggin heteropolyacid salts. Curr Catal Silicotungstic Acid-Catalyzed Etherification of Glycerol and Ethylene Glycol with
2018;5:26–34. https://doi.org/10.2174/2211544707666171219161414. Alkyl Alcohols. Ind. Eng. Chem. Res. 2020;59(21):9858–68. https://doi.org/10.
[36] Batalha DC, Ferreira SO, da Silva RC, da Silva MJ. Cesium-exchanged lacunar 1021/acs.iecr.0c0022910.1021/acs.iecr.0c00229.s001.
Keggin heteropolyacid salts: efficient solid catalysts for the green oxidation of ter- [49] Timofeeva MN. Acid catalysis by heteropoly acids. Appl Catal A-Gen 2003;256:(1-
penic alcohols with hydrogen Peroxide. ChemistrySelect 2020;5:1976–86. https:// 2)19-35. http://dx.doi.org/10.1016/S0926-860X(03)00386-7.
doi.org/10.1002/slct.201903437. [50] Moreira MN, Faria, RPV, Ribeiro AM, Rodrigues AE. Solketal production from
[37] Coronel NC, da Silva MJ. Lacunar Keggin heteropolyacid salts: soluble, solid and glycerol ketalization with acetone: catalyst selection and thermodynamic and ki-
solid supported catalysts. J Clust Sci 2018;29:195–205. https://doi.org/10.1007/ netic reaction study. Ind Eng Chem Res 2019:58(38);17746-17759. https://doi.org/
s10876-018-1343-0.12. 10.1021/acs.iecr.9b03725.
[38] da Silva MJ, Liberto NA, Leles LCA, Pereira UA. Fe4(SiW12O40)3-catalyzed glycerol [51] Da Silva MJ, Teixeira MG. Assessment on the double role of the transition metal
acetylation: Synthesis of bioadditives by using highly active Lewis acid catalyst. J salts on the acetalization of furfural: Lewis and Brønsted acid catalysts. Mol Catal
Mol Catal A 2016;422:69–83. https://doi.org/10.1016/j.molcata.2016.03.003. 2018;461:40–7. https://doi.org/10.1016/j.mcat.2018.10.002.
[39] da Silva MJ, Vilanculo CB, Teixeira MG, Julio AA. Catalysis of vegetable oil [52] Braz J, Ozorio LP, Pianzolli R, Mota MBS, Mota CJA. Reactivity of glycerol/acetone
transesterification by Sn (II)-exchanged Keggin heteropolyacids: bifunctional solid ketal (solketal) and glycerol/formaldehyde acetals toward acid-catalyzed hydro-
acid catalysts. React Kinet Mech Catal 2017;122:1011–30. https://doi.org/10. lysis. Chem Soc 2012;23(5):931–7. https://doi.org/10.1590/S0103-
1007/s11144-017-1258-z. 50532012000500019.
[40] Chaves DM, Ferreira SO, Da Silva RC, Natalino R, Da Silva MJ. Glycerol ester- [53] Ferreira P, Fonseca IM, Ramos AM, Vital J, Castanheiro JE. Valorisation of glycerol
ification over Sn(II)-exchanged Keggin heteropoly salt catalysts: Effect of thermal by condensation with acetone over silica-included heteropolyacids. Appl Catal B:
treatment temperature. Energ Fuels 2019;33:7705–16. https://doi.org/10.1021/ Environ 2010;98:94–9. https://doi.org/10.1016/j.apcatb.2010.05.018.
acs.energyfuels.9b01583. [54] Sandesh S, Halgeri AB, Shanbhag GV. Utilization of renewable resources:
[41] da Silva MJ, Julio AA, Ferreira SO, da Silva RC, Chaves DM. Tin (II) phospho- Condensation of glycerol with acetone at room temperature catalyzed by organi-
tungstate heteropoly salt: An efficient solid catalyst to synthesize bioadditives c–inorganic hybrid catalyst. J Mol Catal A: Chem 2015;401:73–80. https://doi.org/
ethers from glycerol. Fuel 2019;254:115607–18. https://doi.org/10.1016/j.fuel. 10.1016/j.molcata.2015.02.015.
2019.06.015. [55] Chen L, Nohair B, Zhao D, Kaliaguine S. Highly efficient glycerol acetalization over
[42] Sosa AA, Gorsd MN, Blanco MN, Pizzio LR. Synthesis and characterization of supported heteropoly acid catalysts. ChemCatChem 2018;2018(10):1918–25.
tungstophosphoric acid-modified mesoporous silica nanoparticles with tunable https://doi.org/10.1002/cctc.201701656.
diameter and pore size distribution. J Sol-Gel Sci Technol 2017;83:355–64. https://

8
AJChE 2020, Vol. 20, No. 1, 67 – 76
Solketal Production by Glycerol Acetalization
Using Amberlyst-15 Catalyst
Hary Sulistyo*
Edwin Nur Huda
Tri Sarifah Utami
Wahyudi Budi Sediawan
Suprihastuti Sri Rahayu
Muhammad Mufti Azis
Department of Chemical Engineering, Faculty of Engineering, Universitas Gadjah Mada, Jl.
Grafika No. 2, Kampus UGM Yogyakarta 55281 INDONESIA
*e-mail: [email protected]

Glycerol, as a by-product of biodiesel production, has recently increased due to the rapid
growth of the biodiesel industry. Glycerol utilization is needed to increase the added value of
glycerol. Glycerol can be converted to solketal, which can be used as a green fuel additive to
enhance an octane or cetane number. Conversion of glycerol to solketal was conducted via
acetalization reaction with acetone using amberlyst-15 as the catalyst. The objective of present
study was to investigate the effect of some operation conditions on glycerol conversion.
Furthermore, it also aimed to develop a kinetic model of solketal synthesis with amberlyst-15
resins. The experiment was conducted in a batch reactor, equipped with cooling water,
thermometer, stirrer, and a water bath. The variables that have been investigated in the present
work were reaction temperature, reactants molar ratio, catalyst loading, and stirrer speed for 3
hours of reaction time. Temperatures, reactants molar ratio, and stirrer speed appeared to have
a significant impact on glycerol conversion, where the higher values led to higher conversion.
On the other hand, in the presence of catalyst, the increase of catalyst loading has a less
significant impact on glycerol conversion. The results showed that the highest glycerol
conversion was 68.75%, which was obtained at 333 K, the reactant’s molar ratio was 4, the
amount of catalyst was 1 wt%, and stirrer speed of 500 rpm. Based on the pseudo-
homogeneous kinetic model, the resulting kinetic model suitable for this glycerol acetalization.

The value of parameters k and Ea were 1.6135 10 8 min-1 and 62.226 kJ mol-1, respectively.
The simulation model generally fits the experimental data reasonably well in the temperature
range of 313-333 K.

Keywords: Acetone, Amberlyst-15, Fuel additive, Glycerol, Pseudo-homogenous model, Solketal

INTRODUCTION transportation fuel. As a by-product of

There has been renewed interest to biodiesel production, glycerol production
increase biodiesel production in Indonesia will also increase (Qadariyah et al. 2018,
following the plan to introduce B30 as a Sulistyo et al. 2020). Without proper

DOI: 10.22146/ajche.52455
68 Solketal Production by Glycerol Acetalization Using Amberlyst-15 Catalyst

treatment, glycerol may possess using mesoporous 5% Nickel-1%

environmental and economic burdens for Zirconium/Activated Carbon catalyst
the industry. Therefore, it is necessary to showed that glycerol conversion increased
valorize glycerol with sound production as the catalyst loading increased (Khayoon
costs. and Hameed 2013). By using ethanol as a
Solketal is a fuel additive that can be solvent and amberlyst-35 as the catalyst,
synthesized from glycerol and acetone Nanda et al. (2014) reported a 75%
using the catalyst. For solketal production, solketal yield.
a solid catalyst was selected because it is Aghbashlo et al. (2019) reported the
easy to separate and can be used several optimum ketalization condition for
times. There are numerous works in solketal synthesis, which corresponds to
literature mentioning the role of solid 308 K by using highly active ZrP-200 solid
catalyst in solketal production. According as the catalyst. The highly active ZrP-200
to da Silva et al. (2017), SnF2 appears as a showed a promising performance with
promising solid catalyst for glycerol 85.7% glycerol conversion and 98.6%
ketalization. The conversion rate was selectivity of solketal at 323 K. It was found
increased with increasing catalyst loading. that the ZrP-200 showed excellent stability
Moreover, Barbosa et al. (2019) proposed even under mild condition (Li et al. 2019).
the use of SiO2-SO3H as a catalyst for Glycerol ketalization with acetone
direct solketal formation by glycerol using mordenite solid acid could be
acetalization with acetone. Furthermore, to promoted by metal ion as a catalyst.
improve the catalyst activity, the addition Among the metal promoter, Cu-Mo
of quaternary ammonium agents could showed better performance than others. It
catalyze the di-solketal ether formation. was caused by large acid sites and the
The ketalization of waste glycerol as a side insertion effect of the metal particle in the
product of biodiesel production can also mordenite surface. The glycerol conversion
be conducted with acetone using an attained was 95% with solketal selectivity
amphiphilic catalyst. The conversion of of 98% for 15 min reaction time and
glycerol attained was 73%, with solketal reactant molar ratio of 3 (Priya et al. 2017).
selectivity of 95% (Souza et al. 2015). Manjunathan et al. (2015) investigated
Vicente et al. (2010) used acid an acetalization of glycerol at ambient
functionalized SBA-15 as a catalyst for temperature by using a modified beta
acetalization of crude glycerol. The zeolite catalyst of 7.5 wt %. It attained
optimum reaction condition was 343 K 87.1% of the glycerol conversion. They
and 5% loading of the Ar-SBA-15 catalyst also reported kinetics of solketal synthesis
to obtain 90% solketal yields. using a pseudo homogeneous model over
Unfortunately, the presence of sodium ion the zeolite H-BEA catalyst. They assumed
content in crude glycerol promoted the the reversible reaction has occurred.
deactivation of the sulfonic acid due to Aghbashlo et al. (2018) carried out an
reaction between sodium ion and optimization study of a continuous
hydrogen ion. Solventless acetalization process in solketal production from
 H. Sulistyo, E. Nur Huda, T. S. Utami, W. B. Sediawan, S. S. Rahayu, and M. M. Azis 69

glycerol and subcritical acetone using diameter of 0.5mm, the surface area of 53
Purolite PD 2006 catalyst. The optimal m2/g, the pore volume of 0.4 cm3/g, the
condition was attained at a temperature of average pore diameter of 300 Å.
313 K, a reactant ratio of 4.97 with 0.49 mL Amberlyst-15 was purchased from Sigma
-1
min of fresh feed flow rate. Based on this Aldrich. Acetone commercial-grade
result, Aghbashlo et al. (2018) proposed (99.5%) was used as reactant and
the condition for the conversion glycerol purchased from e-Merck.
for the industrial-scale reactor.
The kinetic for heterogeneous fluid- Experimental methods
solid catalytic reaction has been described The experimental setup that has been
by several models in the literature, such as used in the present investigation is
Pseudo-Homogeneous (PH), Eley Rideal generally similar to Sulistyo et al. (2020).
(ER), Langmuir Hinshelwood Hougen Briefly explained here, the studies were
Watson (LHHW) models. The simplest undertaken in a batch reactor equipped
model is the Pseudo-Homogeneous with a stirrer and water cooling system.
model. Nanda et al. (2014) investigated Glycerol was fed into the reactor, then
the LHHW model for ketalization glycerol stirrer and heat source was turned on to
with acetone using the Amberlyst-35 achieve a targeted temperature. Next,
catalyst Meanwhile, Sulistyo et al. (2020) acetone and Amberlyst-15 were fed to the
investigated the ER model for ketalization reactor. The investigations were carried
glycerol with acetone using the Indion out at the temperature range of 313 to
225Na catalyst. 333 K, reactants molar ratio from 3 to 7,
The present investigation aims to catalyst loading from 0 to 7 wt% of
observe glycerol utilization in solketal glycerol, stirrer speed of 400-700 rpm, and
production over Amberlyst-15 as a reaction time of 3h.
catalyst. The data of glycerol conversion Every 30 minutes, the sample was
were collected for several parameters. taken from the reactor. The liquid and
These are reaction temperature, catalyst solid catalysts were separated by filtration.
concentration, reactant molar ratio and The samples were analyzed by AOCS
mixing speed (Huda 2019, Utami 2019). official method Ca 14 – 56 to evaluate the
We also proposed a kinetic model based glycerol conversion (Sulistyo et al. 2020)
on a PH model to describe the reaction
over a specific temperature range. Kinetics modeling method
The reaction between glycerol and
EXPERIMENTAL AND MODELING acetone to form solketal can be expressed
METHODS in Eq. (1) as follows (Nanda et al., 2014):
Material O
O
CH3
CH3
Catalyst
Technical glycerol (99.5%) as raw
+ + H2O
OH O
CH3 CH3

material was purchased from e-Merck.

Acetone Solketal Water

Amberlyst-15 with cation exchange A+ G → S + W (1)

capacity of 5.0 meq/ml, the particle
70 Solketal Production by Glycerol Acetalization Using Amberlyst-15 Catalyst

A glycerol and acetone as reactant were parameter estimation was to evaluate the
mixed with amberlyst-15 as catalyst has value of k and 𝑎 by minimizing the sum
been studied as a catalytic heterogeneous of square error (SSE) of glycerol
reaction. For kinetics approach, it was conversion from the calculation and
proposed a pseudo-homogeneous batch experimental data (X calculation - X data).
reactor model for heterogeneous catalytic Parameter estimation was conducted
reaction. using lsqnonlin solver in Matlab combined
𝑑𝑋 with an ODE solver.
𝐶𝐺0
𝑑𝑡
𝑟𝐺 (2)

𝑟𝐺 1 𝐶𝐴
𝑚
𝐶𝐺 𝑛 (3) RESULTS AND DISCUSSION
From the stoichiometry, can be expressed:
The effect of reaction temperature on
𝐶𝐺 𝐶𝐺0 (4a)
glycerol conversion
𝐶𝐴 𝐶𝐴0 𝐶𝐺0 (4b) The impact of the reaction
𝐶𝑆 𝐶𝐺0 𝐶𝑆0 0 (4c) temperature on glycerol acetalization was
investigated between 313 to 333 K, as
𝐶𝑊 𝐶𝐺0 𝐶𝑊0 0 (4d)
shown in Figure 1. The catalyst loading of
Acetone is in excess, then 1 wt%, stirrer speed of 500 rpm, and
reactant molar of 4 were used, and those
𝐶𝐴0 4𝐶𝐺0 → 𝐶𝐴0 ≫ 𝐶𝐺0 (5a)
parameters were maintained. As presented
𝐶𝐴 𝐶𝐴0 𝐶𝐺0 (5b) in Figure 1, the glycerol conversion
𝐶𝐴 ≈ 𝐶𝐴0 (5c) increased steadily by increasing
temperature to attain the highest
Eq. (3) can be expressed conversion of 68.75 % at 333 K.
𝑟𝐺 1 𝐶𝐴0
𝑚
𝐶𝐺0 𝑛 𝑛
(6) Khayoon and Hameed (2013)
synthesized a solketal and found that 45°C
Eq (2) can be written in the following or 318 K was the optimal temperature to
equation. obtain glycerol conversion close to
𝐶𝐺0 𝑑𝑡
𝑑𝑋 𝑚
𝐶𝐺0 𝑛 𝑛
(7a) complete conversion. It can be explained
1 𝐶𝐴0
that the transition metal pasted in
𝑑𝑋
𝑑𝑡 1 𝐶𝐴0
𝑚
𝐶𝐺0 𝑛−1 𝑛
(7b) mordenite and promoted the acetalization
of glycerol. It was undertaken at 313 K the
Arrhenius equation
glycerol conversion was found to be 81%
𝐴𝑟 (8a) with solketal selectivity of 88% (Priya et al.
2017). Also, da Silva et al. (2017) presented
𝐴𝑟 𝐶𝐴0 𝑚 𝐶𝐺0 𝑛−1 (8b)
glycerol ketalization with glycerol
Finally, conversion close to 90% after 8h reaction
𝑑𝑋 𝐸𝑎 at 343 K. Similar result was found when
𝑛
(9)
𝑑𝑡 𝑅𝑇
glycerol acetalization with pentanal using
In the present study, the value of n has HPMo2@Y zeolite catalyst. After 5h, the
been determined as 3. The objective of solketal selectivity was 83% at 303K, 81%
 H. Sulistyo, E. Nur Huda, T. S. Utami, W. B. Sediawan, S. S. Rahayu, and M. M. Azis 71

at 323 K, and 80% at 343K (Castanhero et ratio of 7 gave the highest value, the lower
al. 2020). value of acetone to glycerol was favorable.
When using Indion 225Na as catalyst
and reactant molar ratio from 3 to 10,
Sulistyo et al. (2020) found a small
increase of glycerol conversion from
17.07% to 19.42%. In other studies, the use
of sulfate zirconia catalyst showed a rise of
glycerol conversion from 62% to 98%
when the reactants molar ratio was
increased from 1 to 6 (Castanheiro et al.
2020). As a result of the excess amount of
acetone, it would push the solketal
formation and improved the complete
Fig. 1: The effect of reaction temperatures to
the glycerol conversion conversion of glycerol. However, for
acetalization of glycerol without catalyst,
varying ketone to glycerol ratio did not
Although higher temperatures gave higher
affect the glycerol conversion (da Silva
conversion, it should be noted that higher
2017). By using reactants molar ratio of 1,
temperatures may also lower conversion,
2, and 3, Priya et al. (2017) observed the
typically for the exothermic reaction as in
glycerol conversion attained of 81%, 86%,
agreement to Nanda et al. (2014).
and 86 % with solketal selectivity 79%,
Aghbashlo et al. (2019) mentioned that
88%, and 95% respectively.
improving reaction temperature may
Gadamseti et al. (2015) expressed the
reduce the formation of solketal due to
effect of the reactant’s molar ratio on the
lower acetone concentration as a reactant
solketal selectivity. The reactant’s molar
in the liquid phase. The boiling
ratio was 1, 2, and 3, the selectivity of
temperature of acetone was of 333K.
solketal were fixed of 98%. It can be
concluded that the increase in reactants
The effect of acetone to glycerol ratio molar ratio does not affect the solketal
The impact of the reactants molar ratio selectivity. Manjunathan et al. (2015) also
on the glycerol conversion was shown in pointed out that the reactant’s molar ratio
Figure 2. For this section, the catalyst of 2 was observed to be the optimum for
loading of 3 wt%, the temperature of 333 the formation of solketal. The glycerol
K, and the stirrer speed of 500 rpm were conversion was nearly constant, with the
kept constant. The increase of reactants increase reactants molar ratio to 3. The
molar ratio gave an increase of glycerol solketal selectivity remained almost
conversion. The various reactant molar constant at 98.5%. Khayoon and Hameed
ratio from 3, 4, 5, and 7 gave a glycerol (2013) used the reactant molar ratio to be
conversion 44.7, 50.1, 58.8, and 71.06 %, changed from 4 to 8, and glycerol
respectively. Although acetone to glycerol conversion was observed to enhance from
72 Solketal Production by Glycerol Acetalization Using Amberlyst-15 Catalyst

52% to complete conversion. Furthermore, glycerol conversion was dropped to 59.02

an excess of acetone can also drive the and 57.46%. It was probably due to
formation of side reaction, such as a six- precipitation of catalyst in the bottom of
membered ring acetal. It may be due to the reactor.
the increase of reaction mixture polarity.

Fig. 3: The effect of catalyst loading to the

glycerol conversion
Fig. 2: The effect of acetone to glycerol ratio to
the glycerol conversion It was also found that the glycerol
acetalization reaction without catalyst
The effect of catalyst loading gave glycerol conversion of 36.56%.
The effect of catalyst loading Gadamseti et al. (2015) undertaken
experiments was investigated at various glycerol acetalization over MoPO/SBA-15
concentrations of Amberlyst-15 resin i.e., catalyst at ambient temperature. It was
1, 3, 5, and 7 wt% of glycerol and also observed that 40% MoPO/SBA-15 catalyst
without catalyst. For this variation, the gave a high conversion close to complete
temperature of 333 K, stirrer speed of 500 conversion at low catalyst concentration.
rpm and reactants molar ratio of 4 were The conversion of glycerol was found
maintained, as shown in Figure 3. 87.10% when acetalization by using 7.5
Interestingly, at 180 min (or the end of wt% H-Beta-1 catalyst (Manjunathan et al.
the experiment), the effect of catalyst 2015).
loading on glycerol conversion was not The glycerol acetalization with
significant. For instance, there was a pentanal using the HPMo2@Y zeolite
marginal increase in glycerol conversion catalyst was investigated. Castanheiro et
from 58.44 to 61.28%, with catalyst al. (2020) found that solketal formation
loading 1 and 3%, respectively. This increased with increasing catalyst loading.
increase could be associated with the rise The solketal selectivity decreased from
in the number of acid sites, which drives 84% to 80%, while glycerol conversion
glycerol acetalization reaction with increased from 35% to 75%. Reddy et al.
acetone to solketal formation. However, at (2011) carried out glycerol acetalization by
the catalyst loading of 5 and 7%, the using various catalyst loading to 5 wt%.
 H. Sulistyo, E. Nur Huda, T. S. Utami, W. B. Sediawan, S. S. Rahayu, and M. M. Azis 73

They used zirconia and promoted zirconia

catalysts. The higher result was observed
by using promoted zirconia was 98% of
glycerol conversion. As the catalyst
loading improved from 3% to 4%, the
glycerol conversion was changed from
87% to 100%. Meanwhile, the glycerol
conversion was found to be 10% when
zirconia catalyst was used.

The effect of mixing speed

The effect of mixing speeds was Fig. 4: The effect of mixing speed to the
investigated between 400 to 700 rpm, as glycerol conversion

presented in Figure 4. The glycerol

conversion was found higher as the stirrer The increase of conversion with the
speed increased from 400 to 700 rpm. increase of stirrer speed, as reported here,
Here, the glycerol conversions were 39.92, might be due to the combined effect of
50.08, 55.27, and 61.43%, respectively, at mass transfer and catalytic process that
180 minutes. The increase of stirrer speed occurs in the reactor. However, no further
from 400 to 700 rpm gave the investigation has been conducted to
improvement of glycerol conversion. This elaborate on this result.
result exhibits a different trend compared
to our previous study in glycerol The kinetics modeling result
acetalization with acetone using the Indion The result of kinetic modeling is
225Na catalyst (Sulistyo et al. 2020). presented in Figure 5 using a PH model, as
In other work, Rosa et al. (2016) carried in Eq. (9). As shown in Figure 5, it was
out the acetalization of glycerol by using observed that the model could capture the
the H-BEA catalyst. They mentioned that general trend of glycerol conversion over
the impact of external diffusion could be a temperature range of 313 – 333 K by
neglected by using higher agitation speed using catalyst loading of 1 wt.% and stirrer
higher than 400 rpm. Under a stirring speed of 500 rpm, reactants molar ratio of
speed of 750 rpm, there was no effect of 4. The model could predict the
mass transfer on the reaction kinetics. experimental data relatively well at 313
Similarly, Khayoon and Hameed (2013) and 323 K. A larger deviation was
showed that the reaction rate is observed at 318, 328, and 333 K as the
independent fof the stirring speed above experimental data appeared to be more
500 rpm. Therefore, in their experiments, scattered. The inspection of modeling
they carried out their experiments at a result from the time horizon shows that
constant stirring speed of 530 rpm. the deviation between simulation and
experimental data seems to be larger
above 60 min.
74 Solketal Production by Glycerol Acetalization Using Amberlyst-15 Catalyst

stirrer speed of 500 rpm, which gave

68.75%.
A pseudo-homogeneous kinetic model
for describing glycerol acetalization
reaction over Amberlyst-15 can be
expressed as:

The value of parameters k and Ea were

1.6135 108 min-1 and 62.226 kJ mol-1. The
simulation model results gave better
Fig. 5: The result of kinetics modeling as a
function of reaction temperature accuracy at 313 and 323 K. At other
temperatures, the model could still
From parameter estimation using capture the general trend with some
experimental data with temperature deviations.
variation, the estimated value of k and Ea
were 1.6135 108 min-1 and 62.226 kJ mol-1, NOMENCLATURE
respectively. Nanda et al. (2014) reported 𝐴 : acetone
overall activation energy for glycerol
𝐴𝑟 : Arrhenius constant [mol1-m-n
conversion to solketal as high as 55.6 ± 3.1 Lm+n-1min-1]
kJ/mol over Amberlyst 35 catalyst using
𝐶𝐴 : acetone concentration [mol L-3]
𝐶𝐴𝑂 : initial acetone concentration
two parameters of the LHHW model.
Hence, the value of Ea reported in the [mol L-3]
present study shows good agreement with 𝐶𝐺 : glycerol concentration [mol L-3]
the literature data. 𝐶𝐺𝑂 : initial glycerol concentration
-3
[mol L ]
CONCLUSION 𝐶𝑆 : solketal concentration [mol L-3]
The acetalization of glycerol has been 𝐶𝑆𝑂 : initiall solketal concentration
carried out using Amberlyst-15 as a [mol L-3]
catalyst. The influence of temperatures, 𝐶𝑊 : water concentration [mol L-3]
reactants molar ratio, and stirrer speed 𝐶𝑊𝑂 : initial water concentration
appeared to have a significant effect on [mol L-3]
glycerol conversion. Increasing those 𝑎 : Activation energy [kJ mol-1]
factors led to higher glycerol conversion. 𝐺 : glycerol
On the other hand, in the presence of the : overall reaction rate constant
catalyst, the increase of catalyst loading [min-1]
exhibits no significant impact on glycerol 1 : intrinsic reaction rate constant
conversion. The observed highest [mol1- m-n Lm+n-1min-1]
conversion was attained at a temperature m : acetone reaction order
of 333 K, reactants molar ratio of 4, 𝑛 : glycerol reaction order
catalyst loading of 3 wt% of glycerol and : gas constant [ 8.314 J mol-1 K-1]
 H. Sulistyo, E. Nur Huda, T. S. Utami, W. B. Sediawan, S. S. Rahayu, and M. M. Azis 75

𝑟𝐺 : reaction rate of glycerol solketal esters with a mixture of

-3 -1
[mol L min ] sulfonated silica and (Bu4N)(BF4)
𝑆 : solketal catalyst,” Catal. Commun., 120, 76–79.
: temperature [K] 4. Castanheiro, J. E., Vital, J., Fonseca, I.
: time [min] M., Ramos, A. M., (2020). “Glycerol
𝑊 : water conversion into biofuel additives by
: glycerol conversion acetalization with pentanal over
heteropolyacids immobilized on
ACKNOWLEDGEMENTS zeolites” , Cat. Today, 346, 76-80.
The authors would like to acknowledge 5. da Silva, M. J., de Ávila Rodrigues, F.
the financial support by the Ministry of and Júlio, A. A. (2017). “SnF2-catalyzed
Research, Technology, and Higher glycerol ketalization: A friendly
Education of Republic of Indonesia under environmentally process to synthesize
Penelitian Dasar Unggulan Perguruan solketal at room temperature over on
Tinggi (PDUPT) research grant 2019, under solid and reusable Lewis acid,” Chem.
contract number of 124/UN1/DITLIT/DIT- Eng. J., 307, 828–835.
LIT/LT/2019 at Universitas Gadjah Mada. 6. Gadamseti, S., Rajan, N. P., Rao, G. S.,
Chary, K. V. R., (2015). “Acetalization of
REFERENCES glycerol with acetone to bio fuel
1. Aghbashlo, M., Hosseinpour, S., additive over supported molyb-
Tabatabaei, M., Rastegari, H., and denum phosphate catalysts, “ J. Mol.
Ghaziaskar, H. S., (2019). “Multi- Cat. A: Chemical, 410, 49 – 57.
objective exergoeconomic and 7. Huda, E. N., (2019). “Pemanfaatan
exergoenvironmental optimization of Gliserol Hasil Samping Produksi
continuous synthesis of solketal Biodiesel untuk Pembuatan Solketal:
through glycerol ketalization with Pengaruh Suhu dan Konsentrasi
acetone in the presence of ethanol as Katalis,” Research Report of Chemical
co-solvent,” Renew. Energy, 130, 735– Reaction Engineering and Catalysis
748. Lab., Dep. Chem. Eng. Univ. Gadjah
2. Aghbashlo, M., Tabatabaei, M., Mada, Yogyakarta, Indonesia,.
Hosseinpour, S., Rastegari, H., and 8. Khayoon, M. S., and Hameed, B. H.,
Ghaziaskar, H. S., (2018). “Multi- (2013). “Solventless acetalization of
objective exergy-based optimization glycerol with acetone to fuel
of continuous glycerol ketalization to oxygenates over Ni–Zr supported on
synthesize solketal as a biodiesel mesoporous activated carbon
additive in subcritical acetone,” Energy catalyst,” Appl. Catal. A Gen., 464–
Convers. Manag., 160, 251–261. 465, 191–199.
3. Barbosa, S. L., Lima, P. C., dos Santos, 9. Li, X., Jiang, Y., Zhou, R., and Hou, Z.,
W. T. P., Klein, S. I., Clososki, G. C. and (2019). “Layered α-zirconium
Caires, F. J. (2019). “Oxygenated phosphate: An efficient catalyst for
biofuels: Synthesis of fatty acid the synthesis of solketal from
76 Solketal Production by Glycerol Acetalization Using Amberlyst-15 Catalyst

glycerol,” Appl. Clay Sci., 174, 120– from glycerol ketalization with
126. acetone,” Ind. Eng. Che. Res., vol 56,
10. Manjunathan, P., Maradur, S. P., pp 479 – 488.
Halgeri, A. B., and Shanbhag, G. V., 16. Souza, T. E., Padula, I. D., Teodoro, M.
(2015). “Room temperature synthesis M. G., Chagas, P., Resende, J. M.,
of solketal from acetalization of Souza, P. P., and Oliveira, L. C. A.,
glycerol with acetone: Effect of (2015). “Amphiphilic property of
crystallite size and the role of acidity niobium oxyhydroxide for waste
of beta zeolite,” J. Mol. Catal. A Chem., glycerol conversion to produce
396,. 47–54. solketal,” Catal. Today, 254, 83–89.
11. Nanda, M. R., Yuan, Z., Qin, W., 17. Sulistyo, H., Hapsari, I., Budhijanto,
Ghaziaskar, H. S., Poirier, M. A., and Sediawan, W. B., Rahayu, S. S., and
Xu, C.C., (2014). “Thermodynamic and Azis, M. M., (2019). “Heterogeneous
kinetic studies of a catalytic process to catalytic reaction of glycerol with
convert glycerol into solketal as an acetone for solketal production,”
oxygenated fuel additive,” Fuel, 117, MATEC Web Conf., 268, 7004.
470–477. 18. Sulistyo, H., Priadana, D. P., Fitriandini,
12. Priya, S. S., Selvakannan, P. R., Chary, Y. W., Ariyanto, T., Azis, M. M., (2020).
K. V. R., Kantam, M. L., and Bhargava, “Utilization of glycerol by ketalization
S. K., (2017). “Solvent-free microwave- reaction with acetone to produce
assisted synthesis of solketal from solketal using indion 225Na as
glycerol using transition metal ions catalyst,” Int. J. Tech, 11, 190 – 199.
promoted mordenite solid catalysts,” 19. Utami, T. S., (2019). “Pemanfaatan
Mol. Catal., 434, 184-193. Glicerol Hasil Samping Produksi
13. Qadariyah, L., Bhuana, D.S., Selaksa, R., Biodiesel untuk Pembuatan Solketal:
As Shodiq, J., and Mahfud, M.. Pengaruh Rasio Reaktan dan
"Biodiesel production from Kecepatan Pengadukan,” Research
Calophyllum inophyllum using base Report of Chemical Reaction
lewis catalyst”, ASEAN Journal of Engineering and Catalysis Lab., Dep.
Chemical Engingeering, 18, 53-59. Chem. Eng. Univ. Gadjah Mada,
14. Reddy, P. S., Sudarsanam P., Yogyakarta, Indonesia,
Mallesham, G., Raju, G., and Reddy, B. 20. Vicente, G., Melero, J. A., Morales, G.,
M., (2011). “Acetalization of glycerol Paniagua, M., and Martín, E., (2010).
with acetone over zirconia and “Acetalization of bio-glycerol with
promoted zirconia catalysts under acetone to produce solketal over
mild reaction conditions,” J. Ind. Eng. sulfonic mesostructured silicas,” Green
Chem., 17, 377–381. Chem., 12, 899–907.
15. Rossa, V., Pessanha, Y. S. P., Diaz, G. C.,
Camara, L. D. T., Pergher, S. B., C.
Aranda, D. A. G., (2017). “Reaction
kinetic study of solketal production
 Catalysis Today xxx (xxxx) xxx

Contents lists available at ScienceDirect

Catalysis Today
journal homepage: www.elsevier.com/locate/cattod

Solketal synthesis from ketalization of glycerol with acetone: A kinetic

study over a sulfated zirconia catalyst
Julián A. Vannucci a, b, Nora N. Nichio a, b, Francisco Pompeo a, b, *
a
Centro de Investigación y Desarrollo en Ciencias Aplicadas (CINDECA), Facultad de Ciencias Exactas, Universidad Nacional de La Plata (UNLP) – CONICET, Calle 47,
257, CP 1900, La Plata, Argentina
b
Facultad de Ingeniería, Universidad Nacional de La Plata (UNLP), Calle 1 esq. 47, CP 1900, La Plata, Argentina

A R T I C L E I N F O A B S T R A C T

Keywords: In this work, a series of acid catalysts were synthetized from a commercial zirconium oxide sulfated with a 0.5 M
Solketal H2SO4 solution by wet impregnation. The characterization results show a correlation between the calcination
Sulfated zirconia temperature and the acid sites generated on the materials. Among the catalysts prepared, the sulfated zirconia
Batch reactor
calcined in air at 400 ◦ C (Zr-S-400), with a molar ratio S/Zr = 0.23 was the most active one due to its larger acid
Kinetic model
density and greater acid strength caused by the generation of new Brönsted sites. The Zr-S-400 catalyst exhibited
an initial reaction rate of 0.0497 mol.min− 1. g− 1, and achieved a glycerol conversion of 80 % in 1 h of reaction at
40 ◦ C (glycerol:acetone molar ratio = 1:6). The Zr-S-400 material remained stable after four catalytic cycles,
demonstrating the stability of the superficial sulfate species (S/Zr ~ 0.2). In addition, the thermodynamics and
kinetics of the reaction were evaluated, as well as the influence of some operating conditions such as the molar
ratio of reactants and the water content in the reaction mixture. The following standard molar reaction properties
were obtained: ΔHº = -11.6 ± 1.1 kJ.mol− 1 and ΔGº = 4.0 ± 0.1 kJ.mol− 1. Taking into account that the
adsorption of water on this catalyst did not affect the number of acid sites available, a simple pseudo-
homogeneous kinetic expression was developed and successfully adjusted to the experimental data in the
range under study. Based on this model, the estimated activation energy of the reaction was 88.1 ± 8.9 kJ.mol− 1.

1. Introduction cosolvent, improving the solubility of glycerol in the acetone phase [5].
Thus, an excess of acetone is often used as a solubility enhancer [5] and
Glycerol is a promising biomass platform molecule having several to shift the chemical equilibrium toward the formation of solketal [2].
applications in many industries. Nowadays, the development of new Despite these problems, several reports have shown that the reaction
technological ways to add value to glycerol has become a research could be successfully carried out with heterogeneous catalysts such as
hotspot. Many studies have focused on the valorization of glycerol ion exchange resins [6], heteropolyacids [7], acid clays [8], meso­
through catalytic processes such as ketalization, dehydration, oxidation, structured silicas [9–11], and zeolites [12–14]. In comparison to the
and reforming. In particular, the condensation reaction of glycerol with aforesaid catalysts, promoted metal oxides offer several advantages,
carbonyl compounds to produce oxygenated compounds has received since they are stable, regenerable, and active. Li et al. synthesized
great attention [1]. Among the different glycerol ketals, solketal is layered crystalline α-zirconium phosphates and studied the effect of the
highly valuable for its potential applications as a green solvent, plasti­ calcination temperature over the acid properties of the material. The
cizer in the polymer industry [2], and solubilizing and suspending agent results showed that materials with a relation P/Zr ~ 2 possess higher
in biodegradable systems for the controlled release of medicinally active surface density of acid sites and stability when calcined at temperatures
substances [3]. In addition, it can be used as a fuel additive to increase below 300 ◦ C. The conversion of glycerol decreased from 86 to 45 %
the octane number and reduce gum formation [4]. Solketal is normally with the increasing calcination temperature from 200 to 600 ◦ C, which
synthesized by the ketalization of glycerol with acetone (Scheme 1). was attributed to the decomposition of the active species Zr
The miscibility of both reactants, glycerol and acetone, is poor. (HPO4)2∙H2O to ZrP2O7 [15].
However, as the reaction progresses, the produced solketal acts as a Miao et al. synthesized a series of ordered mesoporous titanium

* Corresponding author.
E-mail address: [email protected] (F. Pompeo).

https://doi.org/10.1016/j.cattod.2020.10.005
Received 18 May 2020; Received in revised form 19 August 2020; Accepted 1 October 2020
Available online 16 October 2020
0920-5861/© 2020 Elsevier B.V. All rights reserved.

Please cite this article as: Julián A. Vannucci, Catalysis Today, https://doi.org/10.1016/j.cattod.2020.10.005
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

were purchased from Cicarelli. Acetone (99.5 %), n-propanol (99.3 %),
and acetonitrile (99.8 %) were purchased from Anedra. Solketal (98.2
%) was purchased from TCI Chemicals, and powder zirconium oxide was
purchased from MEL Chemicals.

2.2. Catalyst preparation

Scheme 1. Ketalization of glycerol with acetone.

A commercial zirconium oxide powder, namely Zr, was subjected to
a thermal treatment at 600 ◦ C for 5 h, resulting in sample Zr600. Both
materials, Zr and Zr600, were promoted with a proper amount of 0.5 M
phosphate (M-TiPO) materials with controllable composition by a sim­
H2SO4 solution, following a wet impregnation technique. Excess water
ple one-pot method. With the incorporation of highly dispersed P species
was evaporated on a water-bath at 60 ◦ C assisted by ultrasound, oven-
in the structure of M-TiPO, the bonds in the material transferred from
dried, and calcined in air at 400 (Zr-S-400, Zr600-S-400) and 600 ◦ C
Ti–O–Ti to Ti–O–P bonds, thus improving the acid properties of the
(Zr-S-600) for 5 h.
material. The M–TiPO catalyst with a P/Ti ratio = 0.75 exhibited the
best catalytic effect, achieving a glycerol conversion of 91 % and a
2.3. Catalyst characterization
selectivity to solketal of 94 % [16].
Recently, Li et al. prepared a series of zirconium organophosphates
The XRD patterns in the 2θ angular range from 10 to 70◦ were ob­
with different quantities of fenil groups following a hydrothermal
tained with a Philips 3710 X’Pert using CuKα radiation, with a 0.04◦ /
method. The materials showed a hydrophobic character which was
min step. The identification of the XRD phases present in the sample was
enhanced with an increasing amount of phenyl groups. The best activity
aided by JCPDS data files.
was attributed to both abundant acidity and optimal hydrophilia/hy­
Textural properties were determined by nitrogen adsorption-
drophobicity balance [17].
desorption isotherms at the temperature of liquid nitrogen (− 196 ◦ C)
In reactions requiring acid materials, sulfate species represent one of
in a Micrometrics ASAP 2020 instrument. Before adsorption, the sam­
the most widely used promoters of metal oxides [18,19]. The sulfated
ples were evacuated by heating at 100 ◦ C in vacuum, with a pressure
zirconia catalyst has been employed in many industrially important
lower than 4 Pa for 12 h. The specific surface area was calculated ac­
vapor-phase reactions such as dehydration of ethanol [20], alkylation of
cording to the Brunauer-Emmett-Teller (BET) equation, in the relative
isobutane with 2-butene [21], and isomerization of hydrocarbons
pressure range 0.05− 0.35 [30]. The pore size distribution was obtained
[22–24]. Furthermore, it has been found to be very active in
by the Barret-Joyner-Halenda (BJH) method, using the adsorption
liquid-phase reactions, such as Friedel-Crafts monoalkylation [18],
branch and assuming slit-shape pore geometry [31].
synthesis of aromatics gem-dihalides [25], stereocontrolled glyco­
The X-ray photoelectron spectroscopy (XPS) measurements were
sidation [26], etc.
performed using a non-monochromatic Al Kα source (XR50, Specs
The application of sulfated zirconia to the ketalization of glycerol
GmbH) and a hemispherical electron energy analyzer (PHOIBOS 100,
with acetone has been previously explored [27–29]. Reddy et al.
Specs GmbH). O 1s, Zr 3d, S 2p, and C 1s lines were monitored. C 1s at
compared diverse promoters of zirconia, such as W, Mo, and sulfate ions,
284.6 eV was used as a charging reference. Spectra were analyzed with
concluding that the sulfated promoted zirconia showed the highest
CasaXPS software.
catalytic activity due to its higher surface area and number of acidic
The amount and strength of acid sites were estimated by potentio­
sites. Although their study demonstrated the efficacy of sulfated zirconia
metric titration. A known mass (0.05 g) of the solids was suspended in
in glycerol ketalization, the stability of the catalyst and the kinetics of
acetonitrile and kept under stirring for 3 h. The suspension was titrated
the reaction were not evaluated [27].
with a solution of n-butylamine in acetonitrile (0.05 M) at 0.05 mL
Nanda et al. compared a zirconium sulfate catalyst in the synthesis of
min− 1. The electrode potential variation was obtained on a digital pH
solketal in a continuous flow reactor with different commercial mate­
meter (Metrohm 794 Basic Titrino apparatus with a double junction
rials. According to the yield to solketal and glycerol conversion, the
electrode).
activity of these catalysts followed the sequence Amberlyst Wet ~
Pyridine FTIR was used to characterize the nature of the acid sites.
Zeolite ~ Amberlyst Dry > Zirconium Sulfate > Montmorillonite >
Spectra were recorded with a Thermo Nicolet iS10. To create a clean
Polymax. The stability of these catalysts was analyzed and a slight
surface for the analysis, the catalysts were heated to 400 ◦ C under
decrease in the activity was found, likely due to the loss of its acidity
vacuum conditions and maintained for 1 h. Thereafter, the catalyst was
after a long time on stream [29].
cooled to room temperature to collect the background spectra from the
The objectives of this work were to develop an efficient catalyst for
clean surface. After background collection, the adsorption was per­
the synthesis of solketal and determine the thermodynamic and kinetic
formed on the cell at 25 ◦ C. The excess pyridine was desorbed with a
parameters, which are essential tools for understanding the reactive
vacuum from room temperature to 400 ◦ C with 100 ◦ C steps. Brönsted
system, designing reactors, and scaling up the process. For this purpose,
and Lewis acid quantification was performed from 1545 cm− 1 and
commercial ZrO2 was promoted with H2SO4 following a wet impreg­
1450− 1460 cm− 1 bands, respectively, using the literature data on the
nation technique and employed as a catalyst to carry out the ketalization
integrated molar extinction coefficients [32].
of glycerol with acetone. Once the catalyst was proven efficient and
stable, several experimental data were collected in order to determine a
2.4. Catalytic activity measurements
kinetic expression that describes the reaction behavior accurately.
Characterization techniques, such as XRD, N2 adsorption-desorption,
The ketalization of glycerol was carried out in a 200 mL Büchi glass
potentiometric titration, XPS, and Pyridine FTIR, were employed to
batch reactor equipped with a manometer, magnetic stirrer, thermo­
relate the catalytic activity results to the presence of acid sites.
static bath, gas inlet, and release valves. Before the catalytic evaluation,
experiments were performed to verify the negligible contribution of the
2. Experimental
reaction in the absence of a catalyst and in the absence of external and
internal diffusion limitations.
2.1. Materials
In a typical experiment, acetone, glycerol, and the catalyst are loaded
in a reactor. Then, the reactor is pressurized to 0.2 MPa with N2 and
Glycerol (99.5 %), absolute ethanol (99.5 %) and sulfuric acid (98 %)
heated to the desired temperature. As previously mentioned, the use of

2
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

an excess of acetone improves the amount of solubilized glycerol. Molar presence of a methyl group in the axial position of the six-membered
ratios of glycerol to acetone in the range of 1:2 to 1:20 have been re­ ring of 2,2-dimethyl-1,3-dioxan-5-ol, making this molecule less ther­
ported in the literature [33]. However, molar ratios greater than 1:10 do modynamically stable than the solketal molecule [43].
not generate significant impact on catalytic performance [34–36]. The reaction rate of the glycerol ketalization could be expressed in
Concerning the catalyst, mass values between 0.5 % and 5 % wt. related the form of a Langmuir-Hinshelwood-Hougen-Watson (LHHW) model
to glycerol are commonly used [6,8,27,37,38]. In our experiments, we [44,45]. The mechanism is based on the adsorption of both reactants
employed a catalyst amount in the range of 0.3 %–2.5 % wt. of the total (glycerol and acetone) on the catalyst surface, followed by three
mass of glycerol and molar ratios glycerol:acetone in the range of 1:4 to consecutive reactions (Scheme 2) and the desorption of the products
1:8. (water and solketal).
Once the experiment was over, the reactor was cooled to 20 ◦ C and The kinetic parameters were estimated by implementing the
the catalyst was separated by centrifugation and filtration. The reactants Orthogonal Distance Regression algorithm for nonlinear curve fitting
and products were analyzed by gas chromatography with a Shimadzu using OriginLab software. The differences between the predicted values
GCMS-QP505A, equipped with a PE-Elite-Wax capillary column, and an for the variation of the glycerol concentration over time and the
FID detector, using n-propanol as an external standard. experimental data were minimized using the chi-squared criterion.
The conversion was determined with the following equation: The Mears criterion and Weisz-Prater criterion were used to evaluate
external and internal diffusion limitations. For this purpose, the diffu­
(initial glycerol moles − final glycerol moles)
X% = sivity coefficients for each component, the diffusivity coefficient of the
initial glycerol moles
multicomponent system, and the mass transfer coefficient were esti­
mated with the Scheibel correlation [46], Perkins and Geankoplis cor­
and the selectivity to solketal with
relation [47] and the Hixson and Baum correlation [48], respectively.
solketal moles
S% =
(initial glycerol moles − final glycerol moles) 3. Results and discussion

3.1. Catalytic activity

2.5. Thermodynamic and kinetic studies
Since the ketalization of glycerol with acetone requires the presence
The equilibrium thermodynamic constant, Keq, can be determined by
of strong acid sites, this reaction represents an additional characteriza­
the following equation
tion technique to evidence these active sites. The non-promoted mate­
aSolk .aW rials, samples Zr and Zr600, were not active in this reaction, which shows
Keq = (1)
aGly .aAc their lack of strong acid sites. By contrast, all sulfated samples were
active in the reaction. Table 1 presents the initial reaction rate and se­
ai = γ i .xi (2) lectivities for each catalyst. (determined in the first 10 min of the re­
The activity coefficients (γi ) at different temperatures and composi­ action). In all reactions, only solketal and 2,2-dimethyl-1,3-dioxan-5-ol
tions were calculated by UNIFAC [39] with the groups listed in Table S1 were identified as reaction products. Some authors have reported the
(Supplementary Information). presence of subproducts such as 4-hydroxy-4-methyl-2-pentanone [49],
The reaction mechanism proposed in the literature [40–42] is rep­ 1,2,3-propanetriol monoacetate, 2-(1-hydroxypropan-2-ylox­
resented in Scheme 2. In the first step, glycerol and acetone interact y)-propane-1,2-diol and 3-(1-hydroxypropan-2-yloxy)-propane-1,2-diol
forming a hemiketal (3-(2-hydroxypropan-2-yloxy)propane-1,2-diol) [50–52], in the presence of a catalyst with redox properties such as Nb
[40]. The second step differs depending on whether the reaction is [49], Co, Ni, or Pt [50,51]. Therefore, the absence of these properties in
conducted under Brönsted acid sites or Lewis acid sites. In the presence
of Brönsted acid sites, a carbenium ion is stabilized and activated for a Table 1
nucleophilic attack from one of the alcoholic groups of glycerol, leading Initial reaction rate. Experimental conditions: 10 min reaction, 40 ◦ C; 0.2 MPa
to the formation of solketal or the six-membered ring ketal (2, N2, glycerol: acetone molar ratio = 1:6, 0.6 wt.% (catalyst:glycerol).
2-dimethyl-1,3-dioxan-5-ol) [41]. The presence of Lewis acid sites co­
Sample Initial reaction rate (mol. SSolketal S2,2-dimethyl-1,3-dioxan-5-
ordinates and activates the tertiary alcohol of the hemiketal. Then, an min− 1. g− 1) (%) ol (%)
intramolecular reaction with one of the alcoholic groups leads to the
Zr-S-400 0.0497 86 14
formation of solketal and the isomer [42]. Zr-S-600 0.0075 88 12
Although both solketal and 2,2-dimethyl-1,3-dioxan-5-ol isomer are Zr600-S- 0.0361 81 19
produced in the reaction, selectivities toward solketal higher than 80 % 400
have been reported by several authors. This could be related to the

Scheme 2. Reaction mechanism proposed for the ketalization of glycerol with acetone over acid catalysts (adapted from Calvino-Casilda et al. [40]).

3
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

our catalyst improves the selectivity toward solketal. These results show Table 2
that the pre-calcined and sulfated Zr600-S-400 sample presents Textural and acidic properties.
two-thirds of the catalytic activity in comparison with sample Zr-S-400. BET Potentiometric titration
In addition, sample Zr-S-600 was the least active. The characterization
Sample BET Pore Pore E0 meq n-
of the results presented in Section 3.2 shows the relationship between Surface volume diameter (mV) butylamine.
this catalytic behavior and the acidic and textural properties of the (m2. g− 1) (cm3. g− 1) (nm) g− 1
sample. Zr 105 0.34 5.1 63 0.20
Sample Zr-S-400 was the most active and it was, therefore, selected Zr600 49 0.30 12.3 134 0.15
for the kinetic study. The study of the stability of the Zr-S-400 catalyst Zr-S- 58 0.23 5.8 560 0.75
represents an important factor in every kinetic determination. For this 400
Zr-S- 86 0.29 5.9 310 0.52
purpose, four activity cycles of the same catalyst batch were performed.
600
After each run, the reaction mixture was separated from the catalyst by Zr600-S- 24 0.13 11 510 0.52
centrifugation and filtration, and the catalyst was reintroduced into the 400
reactor (without any treatment). After these four runs (Fig. 1), the
catalyst showed a 16 % decrease over its initial activity. Since the sep­
aration process could lead to small losses of fine catalyst powders, the
result obtained in this study is remarkable, evidencing the stability of the
material. From these results, a kinetic study of the ketalization of glyc­
erol with an acetone reaction was performed using the Zr-S-400 catalyst.

3.2. Superficial characterization

The N2 adsorption-desorption isotherms obtained for the samples are

shown in the Supplementary Information (Figure S1). All samples show
type IV isotherms with a hysteresis loop of H3 type according to the
IUPAC classification. Table 2 summarizes the values of specific areas,
average pore size, and total pore volume, for all the samples prepared.
Fig. 2 presents the XRD patterns of commercial zirconia and sulfate
promoted zirconium oxides. It can be observed that all materials exhibit
both characteristic peaks from the monoclinic (2θ = 28.19◦ , 31.48◦ ,
34.17◦ , 50.13◦ ) and tetragonal zirconia phase (2θ = 30.18◦ ), the
monoclinic phase being dominant (JCPDS 86-1451, JCPDS 81-1544).
Fig. 2. X-ray diffraction patterns of commercial zirconia and sulfate promoted
Given that all diffractograms show the same characteristic peaks, it is
zirconium oxides. M: monoclinic phase, T: tetragonal phase.
possible to conclude that sulfation and subsequent heat treatment do not
substantially modify the crystal structure. However, some morpholog­
By means of potentiometric titration with n-butylamine, it is possible
ical changes are evidenced. Calcination of the commercial zirconium
to estimate the strength and number of acid sites present in a solid. It is
oxide at 600 ◦ C (sample Zr600) results in a drop in surface area and,
considered that the initial potential (E0) indicates the maximum strength
therefore, an increase in the pore diameter from 5.1 to 12.3 nm.
of the acid sites, and the value from which the plateau is reached is
Regarding the sulfated samples, Zr-S-400 and Zr600-S-400 present
indicative of the total number of acid sites present in the titrated solid
smaller surface areas than the unsulfated oxides, which would indicate
[53]. Potentiometric titration curves are shown in the Supplementary
the incorporation of sulfated species. By contrast, the larger surface area
Information (Figure S2), whereas the results are summarized in Table 2.
of the sulfated sample Zr-S-600 in comparison to sample Zr-S-400 could
It can be observed that all sulfated samples have higher acidity than the
be evidence of the loss of sulfate groups caused by an increase in the
Zr and Zr600 samples and that an increase in the calcination temperature
calcination temperature.
(from 400 to 600 ◦ C) leads to a decrease in the number of acid sites in the
sample. The latter behavior could be related to the loss of sulfate species
at temperatures above 600 ◦ C [54].
The superficial coverage of sulfate species was explored by XPS. The
results of fresh Zr-S-400 and that reused in 4 runs show an S/Zr atomic
ratio of 0.23 and 0.18, respectively (Figure S3). These values are similar,
within the error of the technique, demonstrating the stability of the
superficial sulfate species under the conditions studied.
An IR analysis of pyridine adsorbed on the Zr-S-400 sample was
performed to determine the type of active site of the catalyst. For
sulfated zirconia, traditional spectral interpretation attributes absorp­
tion peaks at 1607, 1574, and 1445 cm− 1 to Lewis sites [55–57], 1490
cm− 1 to Lewis and Brönsted sites [58], and 1640, 1610, and 1540 cm− 1
to Brönsted acid sites [59]. Fig. 3 shows the FTIR spectra of the Zr-S-400
sample. This spectrum produced peaks at 1445, 1490, 1540, 1574, 1610,
and 1640 cm− 1, showing the presence of both Lewis and Brönsted acid
sites in the material.
For comparison, the FTIR spectra of Zr is shown in the Supplemen­
tary Information (Figure S4). The concentrations of both sites were
Fig. 1. Stability of Zr-S-400. Experimental conditions: 40 ◦ C; 0.2 MPa N2, calculated from the intensities of the bands at 1540 cm− 1 (B) and 1440
glycerol:acetone molar ratio = 1:6, 0.3 wt.% (catalyst: glycerol) and an 80 min cm− 1 (L), in the spectra obtained at different temperatures for each
reaction in each run.

4
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

sulfation process, is partially lost when the material is calcined at 600

◦
C, leading to a decrease in the catalytic activity in sample Zr-S-600
[54].

3.3. Estimation of thermodynamic parameters

In order to determine the thermodynamic parameters, the reaction

was carried out with the Zr-S-400 catalyst and in sufficient time to
ensure no changes in the composition. Fig. 4 presents the linearization of
the experimental data. With these results, the dependence of the equi­
( )
librium constant with temperature could be expressed as ln Keq =
1397 T1 − 6.304.
The estimated values were ΔH◦ = -11.6 ± 1.1 kJ.mol− 1, and ΔG◦ =
4.0 ± 0.1 kJ.mol− 1, which show the exothermic nature of the reaction.
Cornejo et al. [60] conducted a thermodynamic study considering an
ideal behavior of the system and obtained a ΔH◦ = -6.6 ± 0.2 kJ.mol− 1
and ΔG◦ = -0.4 ± 0.1 kJ.mol− 1. While their study predicts an exergonic
character, ours predicts an endergonic character, highlighting the sig­
Fig. 3. FTIR spectra of pyridine adsorbed on Zr-S-400 and evacuated at
nificance of considering the non-ideality of the reactive mixture in these
different temperatures. L = Lewis; B = Brönsted.
studies.
Other studies have determined thermodynamic equilibrium param­
material (Table 3). It is observed that the sulfation process generates
eters using ethanol as a solvent. Nanda et al. [44] determined a ΔH◦ =
sites with Brönsted acidity and slightly increases the acidity of Lewis
-30.1 ± 1.6 kJ.mol− 1 and ΔG◦ = -2.1 ± 0.1 kJ.mol− 1, considering ideal
acid sites in Zr-S-400.
behavior. Moreira et al. [45] estimated the thermodynamic properties
In view of the characterization results obtained, it is evident that the
on the basis of activities, taking into account the non-ideal behavior of
catalytic activity of the materials is linked to its acid properties. From
the mixture. The obtained values were ΔH◦ = -20.1 ± 1.1 kJ.mol− 1 and
the results shown in Table 1, the turnover frequency (TOF) of materials
ΔG◦ = 1.4 ± 0.1 kJ.mol− 1. Although the thermodynamic parameters of
Zr-S-400, Zr-S-600 and Zr600-S-400 were estimated as 66.3, 14.4 and
Moreira et al. could not be easily compared with our results, they also
69.4 min− 1, respectively. Samples Zr-S-400 and Zr600-S-400 have prac­
determined an endergonic character.
tically the same initial potential in the titration experiments (E0 of 560
mV and 510 mV, respectively) indicating similar acid strength. This acid
strength, which is due to the generation of new Brönsted acid sites in the 3.4. Reaction kinetic study

3.4.1. Kinetic model

Table 3 The steps involved in the determination of the kinetic model are
Concentration (μmol. g− 1) of acid sites as calculated from adsorption/desorption presented in Table 4.
of pyridine followed by IR absorption spectroscopy. The * symbol represents an active site of the catalyst, and the * su­
T (◦ C) Concentration of acid sites (μmol. g− 1)
perscript indicates that the component is adsorbed on the catalyst
surface.
Zr Zr-S-400
The reaction starts with the adsorption of both reactants on the
Lewis Brönsted Lewis Brönsted catalyst surface. Then, a surface reaction between adsorbed glycerol
100 41 – 53 37 (Gly*) and acetone (Ac*) leads to the formation of the hemiketal (3-(2-
200 33 – 24 7 hydroxypropan-2-yloxy) propane-1,2-diol (HA*). In step 4, a short-lived
300 15 – 18 2 intermediate (Int*) and water are formed, followed by a subsequent
400 9
reaction, where the intramolecular reaction of the intermediate leads to
– – –

the formation of Solketal (Solk*). The final step of the reaction is the
desorption of solketal and water from the catalyst surface.
In our reaction experiments with Zr-S-400, we did not detect the
presence of the hemiketal even at low glycerol conversions. Using in situ
Raman studies, Calvino-Casilda et al. detected the presence of an
incipient formation of the hemiketal without a catalyst, and almost total
selectivity to solketal in the presence of a catalyst [40]. Therefore, we
consider that step 3 (formation of the hemiketal) is the controlling one.
The different equations employed (3–9) to develop the model are

Table 4
Steps involved in the kinetic model.
Step Reaction

1 Ac + ∗ ↔ Ac∗
2 Gly + ∗ ↔ Gly∗
3 Ac∗ + Gly∗ ↔ HA∗ + ∗
4 HA∗ + ∗ ↔ Int∗ + W∗
5 Int∗ ↔ Solk∗
6 Solk∗ ↔ Solk + ∗
7 W∗ ↔ W + ∗
Fig. 4. Linearization of ln (Keq) vs 1/T.

5
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

presented below:
θAc
K1 = (3)
θV .aAc

θGly
K2 = (4)
θV .aGly

r = K3+ θAc θGly − K3− θHA θV (5)

θInt .θW
K4 = (6)
θV .θHA

θSolk
K5 = (7)
θInt

θV .aSolk
K6 = (8)
θSolk

θV .aW
K7 = (9) Fig. 5. Effect of temperature on glycerol conversion. Experimental conditions:
θW
0.2 MPa N2, glycerol: acetone molar ratio = 1:6, 0.6 wt.% (catalyst: glycerol),
Zr-S-400 catalyst.
where θ represents the fraction of active sites occupied by each species
and θV the free active sites. From Eqs. 3 and 4, it is possible to obtain the
following expression: Table 5
Estimated kinetic constant at different temperatures.
θAc θGly = K1 .K2 .aAc .aGly θV 2 (10)
Temperature (◦ C) k (mol. g− 1. min− 1)
From Eqs. 6,7,8, and 9, the following expression for θHA is obtained:
30 0.04544 ± 0.0061
aSolk .aW .θV 35 0.07126 ± 0.0079
θHA = (11) 40 0.11516 ± 0.0093
K4 .K5 .K6 .K7
50 0.39720 ± 0.0460
Furthermore, taking into account the balance of active sites:
∑ Table 5 shows the estimated kinetic constant for each temperature.
n
1= θi (12)
i=1 The estimation was performed by fitting the experimental data with the
mathematical model (Eq. 15). It is worth mentioning that these values
The reaction rate could be expressed from Eqs. (3–12) as: were obtained assuming that none of the species is strongly adsorbed. To
aGly .aAc − aSolk .aW confirm this assumption, the experimental data were also adjusted
(13)
Keq
r = k( ) 2 considering water and solketal as strongly adsorbed species. Since the
∑n
1 + Ki .ai kinetic parameters obtained were inconsistent, the consideration that
i=1
none of the species involved are strongly adsorbed on the catalyst is
acceptable.
where k is the kinetic constant, Ki is the adsorption equilibrium constant
From the results shown in Table 5, the activation energy value (Ea)
of each component, and Keq is the equilibrium constant of the reaction.
was estimated as 88.1 ± 8.9 kJ.mol− 1, and the pre-exponential factor
In the literature, several articles consider that water is the most
(k0) as 6.55 × 1013 mol. g− 1. min− 1. These results are within the range of
strongly adsorbed component, and the adsorption of the remaining
values obtained by other authors. Esteban et al. [61] estimated a value of
species could be neglected [44,45,61]. If we apply this consideration,
Ea = 124.0 ± 12.9 kJ.mol− 1 assuming an Eley-Rideal mechanism on
the expression of the reaction rate is simplified to:
sulphonic ion-exchange resin in the absence of solvent. Nanda et al. [44]
and Moreira et al. [45] reported Ea = 55.6 ± 3.1 kJ.mol− 1 and Ea = 69.0
aSolk .aW
aGly .aAc −
(14)
Keq
r=k
(1 + KW .aW )2 ± 6.6 kJ/mol, respectively assuming an LHHW mechanism considering
water as the most adsorbed species and using Amberlyst-35 and ethanol
It is well known that the adsorption of water on sulfated zirconia as a solvent. Rossa et al. [63] estimated an Ea = 44.8 ± 1.2 kJ.mol-1
catalysts breaks the coordination of the Zr (IV) species bonded to sulfate assuming a pseudo-homogeneous model and using an H-BEA zeolite. In
species in order to bring Brönsted acid sites [62]. Therefore, the pres­ the matter of homogeneous catalysis, da Silva et al. [64] and Ji et al.
ence of water does not affect the number of active sites available, and the [65] reported the Ea of 26.0 kJ.mol-1 using salt Fe(NO3)3.9 H2O and the
reaction rate could be further simplified to: Ea of 28.2 kJ.mol-1 using ionic liquid [P(C4H9)3C14H29][TsO],
(
aSolk .aW
) respectively.
r = k aGly .aAc − (15) Fig. 6 A and B shows the effect of the initial glycerol to acetone molar
Keq
ratio on the kinetics and thermodynamics of the reaction. As shown in
these figures, the increase in acetone concentration not only increases
3.4.2. Estimation of kinetic parameters
the conversion reached in the equilibrium but also improves the reaction
The Mears criterion values were below 10− 5, and the Weisz-Prater
kinetics. This effect is more evident at lower catalyst content (Fig. 6 B).
criterion values below 10-8, confirming that both external and internal
Since one of the main impurities in the glycerol feedstock is the
diffusion limitations are negligible on the operation conditions. Tem­
presence of water, it is essential to study the influence of this component
perature, acetone, and water concentration were varied systematically.
on the reaction. In our study, water was added to the reaction mixture to
Fig. 5 shows the results of the kinetic model and the experimental ones.
simulate different glycerol feedstocks. Fig. 7 shows the effect of water on
It is observed that the behavior at different temperatures is linked to the
the reaction rate at 5 %, 10 %, and 20 % wt. of the total mass of glycerol.
exothermic character of the reaction.

6
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

Fig. 6. Effect of the glycerol to acetone molar ratio in the feed on glycerol conversion. Experimental conditions: 0.2 MPa N2, 40 ◦ C, Zr-S-400 catalyst. A: 0.6 wt.%
(catalyst: glycerol); B: 0.3 wt.% (catalyst: glycerol).

the glycerol ketalization in the liquid phase. The characterization results

showed a correlation between the calcination temperature and the acid
sites generated on the materials. Among the catalysts studied, the Zr-S-
400 material was the most active one and presented the largest acid
density and greatest acid strength caused by the generation of new
Brönsted sites, as determined by potentiometric titration and Pyridine
FTIR. This catalyst exhibited a glycerol conversion of 80 % in 1 h of
reaction at 40 ◦ C with a glycerol to acetone molar ratio = 1:6 and 0.6 wt
% catalyst:glycerol.
The chemical equilibrium was analyzed using the UNIFAC method to
determine the activity coefficients of each species. The thermodynamic
parameters, standard enthalpy and Gibbs free energies, were determined
to be -11.6 ± 1.1 kJ.mol− 1 and 4.0 ± 0.1 kJ.mol− 1, respectively.
Taking into account that the adsorption of water on this catalyst does
not affect the number of acid sites available, a simple pseudo-
homogeneous kinetic expression was developed and successfully
adjusted to the experimental data in the range studied. Based on this
model, the estimated activation energy of the reaction was 88.1 ± 8.9
Fig. 7. Effect of water composition in the feed on glycerol conversion. Exper­ kJ.mol− 1.
imental conditions: 0.2 MPa N2, 40 ◦ C, glycerol: acetone molar ratio = 1:6, 0.6 The study of the sensitivity of the reaction to the presence of water
wt.% (catalyst: glycerol), Zr-S-400 catalyst. showed that the presence of water in the reactive mixture generates
negative effects on the thermodynamics and kinetics of the reaction and,
more importantly, it was determined that the concentration of water in
the feedstock should be lower than 10 % to keep the efficiency of the
The presence of water in the reaction generates more extreme effects on
process.
the reaction rate and the conversion at equilibrium state. Feedstocks
presenting 20 % wt. of water significantly reduce the reaction rate, and
Credit author statement
the conversion at equilibrium decays from 80 % to 63 %. This highlights
the need to use purified glycerol feedstock, with water concentration
Julián A. Vannucci: Investigation, Catalyst preparation, Charac­
below 10 % for better efficiencies in the chemical process.
terization, Formal analysis. Nora Nichio: Conceptualization, Writing
original draft, Project administration. Francisco Pompeo: Conceptual­
4. Conclusions
ization, Methodology, Writing- Reviewing and Editing.
The sulfation of a commercial zirconium oxide, by wet impregnation
with a 0.5 M H2SO4 solution, allowed us to obtain an active catalyst for

7
J.A. Vannucci et al. Catalysis Today xxx (xxxx) xxx

Declaration of Competing Interest [25] A. Wolfson, O. Shokin, D. Tavor, J. Mol. Catal. A Chem. 226 (2005) 69–76.
[26] H. Nagai, K. Kawahara, S. Matsumura, K. Toshima, Tetrahedron Lett. 42 (2001)
4159–4162.
The authors declare that they have no known competing financial [27] P.S. Reddy, P. Sudarsanam, B. Mallesham, G. Raju, B.M. Reddy, J. Ind. Eng. Chem.
interests or personal relationships that could have appeared to influence 17 (2011) 377–381.
the work reported in this paper. [28] V.T. Vasantha, N.J. Venkatesha, S.Z.M. Shamshuddin, J.Q. D’Souza, B.G.V. Reddy,
ChemistrySelect 3 (2018) 602–608.
[29] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.A. Poirier, C. Xu, Appl. Energy
Acknowledgments 123 (2014) 75–81.
[30] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 60 (1938) 309–319.
[31] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380.
This research work was possible due to the funds received from [32] C. Emeis, J. Catal. 141 (1993) 347–354.
“Consejo Nacional de Investigaciones Científicas y Técnicas” (CONICET- [33] L. Roldán, R. Mallada, J.M. Fraile, J.A. Mayoral, M. Menéndez, Asia-Pacific J.
PIP 0065), “Universidad Nacional de La Plata” (UNLP-I248). The Chem. Eng. 4 (2009) 279–284.
[34] O. Ilgen, S. Yerlikaya, F.O. Akyurek, Period. Polytech. Chem. Eng. 61 (2017)
doctoral fellowship granted by CONICET to Julian A. Vannucci is 144–148.
greatfully acknowledged. The authors would like to thank Emilce D. [35] S. Sandesh, A.B. Halgeri, G.V. Shanbhag, J. Mol. Catal. A Chem. 401 (2015) 73–80.
Galarza for the FTIR measurements. [36] S. Zhang, S. Zhao, Y. Ao, Appl. Catal. A Gen. 496 (2015) 32–39.
[37] P. Manjunathan, S.P. Maradur, A.B. Halgeri, G.V. Shanbhag, J. Mol. Catal. A Chem.
396 (2015) 47–54.
Appendix A. Supplementary data [38] J. Kowalska-Kus, A. Held, M. Frankowski, K. Nowinska, J. Mol. Catal. A Chem. 426
(2017) 205–212.
[39] A. Fredenslund, J. Gmehling, P. Rasmussen, Vapor-liquid Equilibria Using Unifac:
Supplementary material related to this article can be found, in the
A Group-Contribution Method, Elsevier, 1977.
online version, at doi:https://doi.org/10.1016/j.cattod.2020.10.005. [40] V. Calvino-Casilda, K. Stawicka, M. Trejda, M. Ziolek, M.A. Bañares, J. Phys. Chem.
118 (2014) 10780–10791.
References [41] B. Mallesham, P. Sudarsanam, G. Raju, B.M. Reddy, Green Chem. 15 (2013)
478–489.
[42] L. Li, T.I. Korányi, B.F. Sels, P.P. Pescarmona, Green Chem. 14 (2012) 1611–1619.
[1] J.A. Melero, G. Vicente, G. Morales, M. Paniagua, J. Bustamante, Fuel 89 (2010) [43] L.P. Ozorio, R. Pianzolli, M.B.S. Mota, C.J.A. Mota, J. Braz. Chem. Soc. 23 (2012)
2011–2018. 931–937.
[2] M.R. Nanda, Y. Zhang, Z. Yuan, W. Qin, H.S. Ghaziaskar, C. Xu, Renew. Sustain. [44] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.A. Poirier, C.C. Xu, Fuel 117
Energy Rev. 56 (2016) 1022–1031. (2014) 470–477.
[3] K. Fraatz, D. Mertin, I. Heep, US Patent 8231903B2, 2012. [45] M.N. Moreira, R.P.V. Faria, A.M. Ribeiro, A.E. Rodrigues, Ind. Eng. Chem. Res. 58
[4] C.J.A. Mota, C.X.A. da Silva, N. Rosenbach Jr, J. Costa, F. da Silva, Energy Fuels 24 (2019) 17746–17759.
(2010) 2733–2736. [46] E.G. Scheibel, Ind. Eng. Chem. 46 (1954) 1569–1579.
[5] J. Esteban, A.J. Vorholt, A. Behr, M. Ladero, F. Garcia-Ochoa, J. Chem. Eng. Data [47] L.R. Perkins, C.J. Geankoplis, Chem. Eng. Sci. 24 (1969) 1035–1042.
59 (2014) 2850–2855. [48] A.W. Hixson, S.J. Baum, Ind. Eng. Chem. 33 (1941) 478–485.
[6] J. Esteban, F. García-Ochoa, M. Ladero, Green Process. Synth. 6 (2017) 79–89. [49] A. Feliczak-Guzik, I. Nowak, Microporous Mesoporous Mater. 277 (2019) 301–308.
[7] P. Ferreira, I.M. Fonseca, A.M. Ramos, J. Vital, J.E. Castanheiro, Appl. Catal. B 98 [50] I.S. Gomes, D.C. de Carvalho, A.C. Oliveira, E. Rodríguez-Castellón,
(2010) 94–99. S. Tehuacanero-Cuapa, P.T.C. Freire, J.M. Filho, G.D. Saraiva, F.F. de Sousa,
[8] M.N. Timofeeva, V.N. Panchenko, V.V. Krupskaya, A. Gil, M.A. Vicente, Catal. R. Lang, Chem. Eng. J. 334 (2018) 1927–1942.
Commun. 90 (2017) 65–69. [51] A.L.G. Pinheiro, J.V.C. do Carmo, D.C. Carvalho, A.C. Oliveira, E. Rodríguez-
[9] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martín, Green Chem. 12 Castellón, S. Tehuacanero-Cuapa, L. Otubo, R. Lang, Fuel Process. Technol. 184
(2010) 899–907. (2019) 45–56.
[10] L. Li, D. Cani, P.P. Pescarmona, Inorganica Chim. Acta 431 (2015) 289–296. [52] D.C. de Carvalho, A.C. Oliveira, O.P. Ferreira, J.M. Filho, S. Tehuacanero-Cuapa, A.
[11] Z. Li, Z. Miao, X. Wang, J. Zhao, J. Zhou, W. Si, S. Zhuo, Fuel 233 (2018) 377–387. C. Oliveira, Chem. Eng. J. 313 (2017) 1454–1467.
[12] H. Serafim, I.M. Fonseca, A.M. Ramos, J. Vital, J. Castanheiro, Chem. Eng. J. 178 [53] L.R. Pizzio, M.N. Blanco, Microporous Mesoporous Mater. 103 (2007) 40–47.
(2011) 291–296. [54] G.X. Yan, A. Wang, I.E. Wachs, J. Baltrusaitis, Appl. Catal. A Gen. 572 (2019)
[13] S.S. Priya, P. Selvakannan, K.V. Chary, M.L. Kantam, S.K. Bhargava, Mol. Catal. 210–225.
434 (2017) 184–193. [55] B.T. Loveless, A. Gyanani, D.S. Muggli, Appl. Catal. B 84 (2008) 591–597.
[14] L.H. Vieira, L.G. Possato, T.F. Chaves, S.H. Pulcinelli, C.V. Santilli, L. Martins, Mol. [56] C.D. Miranda M, A.E. Ramírez S, S.G. Jurado, C.R. Vera, J. Mol. Catal. A Chem. 398
Catal. 458 (2018) 161–170. (2015) 325–335.
[15] X. Li, Y. Jiang, R. Zhou, Z. Hou, Appl. Clay Sci. 174 (2019) 120–126. [57] C. Breitkopf, S. Matysik, H. Papp, Appl. Catal. A Gen. 301 (2006) 1–8.
[16] Z. Miao, Z. Li, M. Liang, J. Meng, Y. Zhao, L. Xu, J. Mu, J. Zhou, S. Zhuo, W. Si, [58] W.H. Chen, H.H. Ko, A. Sakthivel, S.J. Huang, S.H. Liu, A.Y. Lo, T.C. Tsai, S.B. Liu,
Chem. Eng. J. 381 (2020) 122594. Catal. Today 116 (2006) 111–120.
[17] X. Li, Y. Jiang, R. Zhou, Z. Hou, Appl. Clay Sci. 189 (2020) 105555. [59] K.T. Wan, C.B. Khouw, M.E. Davis, J. Catal. 158 (1996) 311–326.
[18] V.S. Marakatti, S. Marappa, E.M. Gaigneaux, New J. Chem. 43 (2019) 7733–7742. [60] A. Cornejo, M. Campoy, I. Barrio, B. Navarrete, J. Lázaro, React. Chem. Eng. 4
[19] J.R. Sohn, D.H. Seo, Catal. Today 87 (2003) 219–226. (2019) 1803–1813.
[20] A.I. Ahmed, S.A. El-Hakam, S.E. Samra, A.A. EL-Khouly, A.S. Khder, Colloids Surf. [61] J. Esteban, M. Ladero, F. García-Ochoa, Chem. Eng. J. 269 (2015) 194–202.
A Physicochem. Eng. Asp. 317 (2008) 62–70. [62] M. Hino, M. Kurashige, H. Matsuhashi, K. Arata, Thermochim. Acta 441 (2006)
[21] A. Corma, M.I. Juan-Rajadell, J.M. López-Nieto, A. Martinez, C. Martínez, Appl. 35–41.
Catal. A Gen. 111 (1994) 175–189. [63] V. Rossa, Y.S.P. Pessanha, G.C. Díaz, L.D.T. Câmara, S.B.C. Pergher, D.A.G. Aranda,
[22] X. Li, K. Nagaoka, J.A. Lercher, J. Catal. 227 (2004) 130–137. Ind. Eng. Chem. Res. 56 (2017) 479–488.
[23] N. Essayem, Y. Ben Taârit, C. Feche, P.Y. Gayraud, G. Sapaly, C. Naccache, J. Catal. [64] M.J. da Silva, A.A. Rodrigues, P.F. Pinheiro, Fuel 276 (2020) 118164.
219 (2003) 97–106. [65] Y. Ji, T. Zhang, X. Gui, H. Shi, Z. Yun, Chin. J. Chem. Eng. 28 (2020) 158–164.
[24] A. Corma, J.M. Serra, A. Chica, Catal. Today 81 (2003) 495–506.

8
 Applied Catalysis A, General 592 (2020) 117369

Contents lists available at ScienceDirect

Applied Catalysis A, General

journal homepage: www.elsevier.com/locate/apcata

Hydrophobic functionalization of HY zeolites for efficient conversion of T

glycerol to solketal
Mohammad Shahinur Rahamana, Thanh Khoa Phunga,b,c, Md. Anwar Hossaina,
Emtias Chowdhuryd, Sarttrawut Tulaphola,e, Shashi B. Lalvanif, Martin O’Tooled,
Gerald A. Willinga, Jacek B. Jasinskib, Mark Crockerg,h, Noppadon Sathitsuksanoha,*
a
Department of Chemical Engineering, University of Louisville, Louisville, KY, 40292, USA
b
Conn Center for Renewable Energy Research, University of Louisville, Louisville, KY, 40292, USA
c
School of Biotechnology, International University, Vietnam National University, HCMC, Quarter 6, Linh Trung Ward, Thu Duc District, Ho Chi Minh City, Vietnam
d
Department of Bioengineering, University of Louisville, Louisville, KY, 40292, USA
e
Department of Chemistry, King Mongkut’s University of Technology Thonburi, Bangkok, 10140, Thailand
f
Department of Chemical, Paper, and Biomedical Engineering, Miami University, Oxford, OH, 45056, USA
g
Center for Applied Energy Research, University of Kentucky, 2540 Research Park Drive, Lexington, KY, 40511, USA
h
Department of Chemistry, University of Kentucky, Lexington, KY, 40506, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Production of solketal by glycerol acetalization is a promising pathway to add values to glycerol. Our evidence
Solketal indicated that, initially, the reaction was limited by poor interfacial mass transfer. Acetone slowly diffused and
Hydrophobic solubilized in the glycerol phase, where it reached the catalyst’s active sites to form solketal. We describe a
HY zeolites strategy that overcomes this interfacial mass transfer limitation by grafting an organosilane surfactant (n-oc-
Emulsion
tadecyltrichlorosilane (OTS)) onto the surface of an HY catalyst. The OTS-grafted HY catalyst became hydro-
Glycerol acetalization
phobic and assisted in emulsion formation between the two immiscible reactants, minimizing the interfacial
Mass transfer
mass transfer limitation. As a result, the OTS-HY catalyst produced high catalytic activity (89% conversion)
compared with that of HY (28% conversion) at 30 °C after 60 min. The high catalytic activity of organosilane-
modified HY catalyst at low temperature makes it a promising candidate for other acid-catalysed two-phase
reactions.

1. Introduction gallons, and production is projected to reach 4 billion gallons by 2022

[9]. Biodiesel production generates ∼10 wt.% glycerol [10]. As our
The petroleum and animal farming industries are the leading con- society slowly moves to a sustainable future, the ability to convert
tributors to greenhouse gases, including carbon dioxide (CO2), me- glycerol to high-value products will accelerate its commercial use and
thane, and nitrous oxide [1,2]. Carbon dioxide emissions account for the fight against global warming.
65% of global greenhouse gases, causing a global temperature rise and Glycerol can be upgraded to valuable compounds such as propa-
climate change and negatively affecting agricultural/forest production, nediol, acrolein, glycerol carbonate, glyceric acid, tartronic acid,
livestock farming, and our standard of living [3–5]. In 2014, CO2 syngas, and solketal [11–19], thereby providing opportunities for ad-
emissions were estimated at ∼34.1 gigatonnes [6], and this quantity is ditional revenue for the biodiesel industry and the agricultural sector.
projected to rise [7]. CO2 emissions stem from fossil fuel processing and Glycerol acetalization is catalyzed by acid sites and produces cyclic
methane release from shale oil extraction and natural gas development acetals with 5- and 6-membered rings (4-hydroxymethyl-2,2-dimethyl-
[8]. Production of biofuels and bioproducts from renewable plant bio- 1,3-dioxane (solketal) and 5-hydroxy-2,2-dimethyl-1,3-dioxane
mass can reduce CO2 emissions, mitigating global warming and climate (acetal)) (Fig. 1). Solketal is of particular interest because (1) it is a
change. 100% bio-based chemical, produced from glycerol conversion with
Glycerol is a by-product of processing vegetable oils to produce acetone. Acetone can be derived from hemicellulose from biorefineries
biodiesel. In 2017, the biodiesel production volume was ∼ 1.6 billion [20]; (2) the reaction operates under mild conditions [21–23]; and (3)

⁎
Corresponding author.
E-mail address: [email protected] (N. Sathitsuksanoh).

https://doi.org/10.1016/j.apcata.2019.117369
Received 30 August 2019; Received in revised form 21 November 2019; Accepted 29 November 2019
Available online 07 January 2020
0926-860X/ © 2020 Elsevier B.V. All rights reserved.
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

radical and adds chemical characteristics. The organosilane solution

was added to the zeolite suspension, and the resulting mixture was
stirred for 24 h at 500 rpm at room temperature. The surface-modified
zeolite was then filtered and washed with ethanol five times and va-
cuum dried at 80 °C overnight. This functionalized zeolite was named
OTS-HY.
Fig. 1. Solketal production from glycerol acetalization by acid.
2.2. Characterization of catalytic materials

solketal can be used in many applications including fuel additives The degree of hydrophobicity of zeolites was determined by the
[24,25], solvents in paint and ink industries [26–28], cleaning products contact angle measurement using water droplets. Zeolite samples were
[29], additives for the pharmaceutical industry [30,31], and co-in- compressed into a 1 cm disc (OD) with a thickness of 2 mm. A 1 μL
itiators for polymerization [29,32]. water droplet was placed on the external surface of the disc using
Although homogeneous acid catalytic systems work for glycerol Optical contact angle measurement and drop contour analysis (OCA15,
acetalization, their use is complicated by the need to separate catalysts DataPhysics Instruments USA Corp., Charlotte, NC, USA). Infrared
from products. Thus, heterogeneous zeolite catalysts have been popular spectra of the zeolites were recorded on a JASCO Fourier transform
choices for glycerol acetalization [33,34]. However, the low glycerol infrared (FTIR) spectrometer (Easton, MD, USA), equipped with an
solubility in acetone (5 wt.% or ∼ 0.03 glycerol/acetone molar ratio) attenuated total reflection stage (ATR). High-resolution transmission
causes the two reactants to be immiscible, presenting a mass transfer electron microscopy (HRTEM) was performed on catalysts using a 200
limitation. Immiscibility greatly limits the joint contact between the kV-operated field emission gun FEI Tecanai F20 transmission electron
two reactants and catalysts’ active sites, and, therefore, impairs cata- microscope. Low-intensity illumination conditions were used to mini-
lysis [35–37]. Mass transfer limitation is unavoidable in multiphase mize the amorphization of zeolites.
reactions, including glycerol acetalization, but it can be minimized To confirm the changes in surface functionality after grafting OTS
[38,39]. onto HY catalysts, the diffuse reflectance infrared Fourier transforma-
The objective of this study was to develop a multifunctional, solid tion spectroscopy (DRIFT) was performed using the JASCO FTIR
acid catalyst modified with a surfactant to improve the contact between equipped with high-temperature DiffuseIR™ cell (PIKE Technology, WI,
glycerol and acetone. We hypothesized that grafting an organosilane USA). The sample treatment and temperature program were described
surfactant onto a solid HY zeolite catalyst would generate interface- elsewhere with a slight modification [41]. In short, all experiments
active materials that would create an emulsion in which catalysis would were performed after heating 5 mg catalyst sample in situ up to 230 °C
occur. The emulsion would enhance the interfacial mass transfer by under a flow of N2 (20 mL/min) with a heating rate of 10 °C/min. Then
increasing the contact surface between the two immiscible reactants. the temperature was maintained at 230 °C for 30 min. A background
We tested the foregoing conjecture by grafting an organosilane spectrum was recorded prior to each run and the 512 scans of spectra
surfactant, n-octadecyltrichlorosilane (OTS), onto an HY zeolite catalyst were collected in the range between 4000−1000 cm−1 at a 4 cm−1
and evaluated the performance of this OTS-grafted HY (OTS-HY) at resolution.
30 °C. We selected OTS because it has a long alkyl chain to provide The surface area and pore volume of zeolites were measured using
substantial hydrophobicity. Moreover, we selected HY zeolite (Si/ N2 adsorption/desorption by a Tristar Micromeritics (Norcross, GA,
Al = 2.6) as our catalyst because (1) it is widely used in the chemical USA) instrument. Prior to the measurement, the samples were pre-
industry, (2) it has a large pore dimension (7.3 nm), and (3) it has a low treated at 160 °C for 2 h using a Micromeritics FlowPrep with sample
Si/Al ratio, providing a high acidity (578 μmol NH3/g catalyst) [40] for degasser (Norcross, GA, USA). The surface area, SBET, was determined
glycerol acetalization. We found that the OTS-HY catalyst produced a by N2 isotherms using the Brunauer–Emmett–Teller equation (BET) on
higher glycerol conversion than did the HY catalyst. The OTS-HY cat- the basis of overall mass of the catalysts. The desorption cumulative
alyst enabled emulsion formation, increasing contact between the two pore volume was estimated according to the Barrett–Joyner–Halenda
reactants. This catalytic system addresses the fundamental limitation of (BJH) model. The moisture and organic compounds on the catalysts
the low contact between a catalyst’s active sites and the two immiscible were determined by thermogravimetric analysis (TGA) using a SDT
reactants in glycerol acetalization. Q600 TA instrument (New Castle, DE, USA). In short, ∼20 mg of the
sample was placed in a cylindrical alumina crucible and heated in the
2. Materials and methods air from room temperature to 500 °C with a heating rate of 10 °C/min
under N2 flow (100 ml/min). The moisture content of catalyst was
2.1. Functionalization of zeolites with surfactant calculated from the weight loss below 150 °C. X-ray diffraction (XRD)
analysis was performed on a Bruker D8 Discover diffractometer
HY zeolite (CBV600, Si/Al ratio = 2.6) was obtained from Zeolyst® (Bellerica, MA, USA) using CuKα radiation and 2θ ranging from 10° to
International (Conshohocken, PA, USA). We chose this HY zeolite with 60° with 0.2 s/step [41]. This 2θ range revealed the diffraction intensity
2.6 Si/Al ratio because it has a high acid site density with the hydro- of (220), (311), (331), (511), (440), (533), (642), and (555), respec-
philic characteristic, enabling us to observe (1) the negative effect of tively [42–44]. The suspension behavior of OTS-HY and HY catalysts
water formation on catalytic activity and (2) the benefit of adding was investigated by placing catalysts in the glycerol-acetone system (1/
surface hydrophobicity. The as-received HY zeolite was calcined at 1 v/v). Videos 1 and 2 show the catalyst behavior in both phases (see
500 °C for 1 h to remove residual impurities before use. The surface Supplementary information).
functionalization of the zeolite was performed as described [41]. In The total acidity of zeolites was determined by ammonia-tempera-
short, 1 g of zeolite was dispersed in 20 mL of toluene in a capped ture programmed desorption (NH3-TPD). The NH3-TPD experiments
250 mL flask using a sonicator at room temperature for 1 h. The orga- were performed using a Micromeritics ChemiSorb 2720 instrument
nosilane reagent (n-octadecyltrichlorosilane (OTS), 95%, Alfa Aesar) at equipped with a thermal conductivity detector (TCD) (Norcross, GA,
an organosilane/zeolite ratio of 0.5 mmol/g zeolite (theoretical OTS USA). The samples were dried in a vacuum oven overnight prior to
loading on zeolite) was mixed with 50 mL toluene at room temperature. NH3-TPD. About 20−40 mg of sample was pretreated at 250 °C for 1 h
The hydrolyzable Cl ions of the OTS underwent hydrolysis and formed a under flowing He gas to remove adsorbed water. The sample was then
stable condensation with silanol groups (-Si-OH) on the surface of HY. cooled to 100 °C and saturated with ammonia (10% NH3/He). Next, the
The organofunctional group (octadecyl) is a nonhydrolyzable organic samples were flushed with 40 mL/min He flow at 100 °C for 1 h to

2
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

remove physically adsorbed ammonia. TPD profiles were recorded by characteristic TeOeT band in the silane grafted SAPO-34 zeolites,
heating the samples to 700 °C at a rate of 10 °C/min in 40 mL/min He suggesting the formation of linkages between silane and zeolites [47].
flow. In the case of OTS-HY catalysts, we ran the TPD experiment We observed bands at 2919 and 2851 cm−1, associated with the CH
without pre-adsorption of NH3 to evaluate the OTS decomposition symmetric (υs(CH)) and asymmetric (υa(CH)) stretching vibrations for
temperature ranges and its contribution during the NH3-TPD. To de- CH2 groups of the OTS, respectively (Fig. S1A). The band at 1198 cm−1
termine the acid site density of OTS-HY catalyst, we subtracted the OTS was due to the formation of Si-O-Si linkages between OTS and HY
decomposition peak area from the NH3-TPD peak area. catalyst [48,49]. We did not observe the characteristic band of the self-
condensed OTS product at 1014 cm−1, suggesting that its formation
2.3. Study of glycerol acetalization was negligible (see Supplementary information, Fig. S1B).
We further confirmed the successful OTS grafting by DRIFT and
Reactions were performed in 15 mL glass pressure vials in an oil observed changes in the OH vibrational region (3500–3800 cm−1) of
bath. Typically, glycerol, acetone, and catalysts were added to the HY and OTS-HY catalysts. We detected a significant decrease in band
pressure vial which was sealed and stirred at the desired temperature. intensity at 3740 cm−1 after OTS grafting onto HY catalysts (Fig. 2C).
We ran this reaction for different times (10−60 min), at different Whereas the two bands at 3627 and 3562 cm−1 remained relatively
temperatures (30 and 50 °C), with different catalyst loading (5 and unchanged. The band at 3740 cm−1 was attributed to the free silanol
15 wt.%), and a fixed glycerol/acetone molar ratio of 1/12 (12 wt.% groups on the external surface of HY catalysts [50]. So the decrease in
glycerol), unless otherwise noted. The OTS-HY catalyst was loaded at this band intensity indicated a reduction in density of silanol groups
an amount that provided a number of active sites comparable to that of after OTS grafting. The bands at 3627 and 3562 cm−1 were attributed
HY. Dodecane was used as an internal standard. The glycerol conver- to the structural hydroxyl groups, the high-frequency (HF) OH groups
sion and product yield were calculated based on the internal standard. at the supercages and low frequency (LF) at the sodalite cages, re-
The reaction was stopped by quenching in a cold-water bath followed spectively [51]. These OH groups are responsible for the Brønsted
by adding ethanol (2 mL ethanol per 0.25 g glycerol) to dissolve the acidity of the zeolites [52,53]. After OTS grafting, intensities of these
remaining glycerol and acetone. The solution was centrifuged, and the bands remained relatively unchanged, suggesting that grafted OTS did
solid catalyst was removed. The liquid sample was then diluted with not block the active sites. Consistent with our results, Zapata et al. also
ethanol prior to analysis. observed negligible changes in intensities of these OH stretching bands
The reactants and products were analyzed using a gas chromato- after silylation of HY catalysts [41]. The retention of the active sites in
graph (7890B GC) (Agilent Technologies, Santa Clara, CA, USA) OTS-HY catalysts was helpful for the acetalization reaction. Hence, al-
equipped with a mass spectrometer and flame ionization detector (FID) terations of FTIR bands in the fingerprint region and the reduction of
for product identification and quantification, respectively. A DB-1701 the silanol density observed by DRIFT confirmed the formation of lin-
column (30 m x0.25 mm x0.25 μm, Agilent Technologies, Santa Clara, kages between OTS and HY catalysts.
CA, USA) was used for product separation with the following para- We used TGA to assess the stability of the OTS-HY and quantify the
meters: injection temperature 275 °C and FID detector temperature amount of OTS grafted onto the HY (Fig. 2A). Below 150 °C, HY in-
300 °C; split ratio 1:50. The temperature program started at 50 °C with a curred an initial weight loss (∼5%) that was attributed to breaking
heating rate of 8 °C/min to 200 °C. The glycerol conversion, product hydrogen-bond networks and desorption of water. OTS-HY’s weight loss
yield, and product selectivity were calculated as follows: was lower (∼2%). The sharper decrease in weight of HY indicated
higher water adsorption (∼5%) compared with that of OTS-HY (∼2%).
mole of glycerol reacted
Glycerol conversion (%) = x100 These results suggested that OTS-HY catalyst was more hydrophobic
initial mole of glycerol than HY catalyst because the surface of HY catalyst consists of portions
mole of product generated of free silanols (-Si-OH) that adsorb water molecules [54–56]. The
Product yield (%) = x100 weight loss of the OTS-HY occurred in three steps: (1) ambient to
initial mole of glycerol
150 °C, (2) 150−270 °C and (3) 400−500 °C. The weight losses in the
Product yield first and second steps were from the evaporation of moisture and re-
Product selectivity (%) = x100 sidue organic solvent during synthesis, respectively. The total weight
Glycerol conversion
loss for OTS-HY was higher than that of HY, which we attributed to the
slow decomposition of OTS at a higher temperature (> 400 °C). Zapata
3. Results and discussion et al. reported OTS decomposition from OTS-functionalized HY in the
range of 350–600 °C [59], corroborating our OTS decomposition find-
We first confirmed the successful grafting of the OTS surfactant onto ings. We estimated that the amount of grafted OTS on HY was ∼16%
the HY catalyst and characterized changes in the surface properties. (w/w).
These analyses were performed with Fourier-transform infrared spec- The crystallinity of HY was preserved after grafting with OTS. The
troscopy (FTIR), Thermal gravimetric analysis (TGA), High-resolution HRTEM image (Fig. 3A & D) of OTS-HY catalyst illustrated (1) the cubic
transmission electron microscope (HRTEM), X-ray diffraction (XRD), crystalline structure of the FAU zeolites [43,57], and (2) an unchanged
and N2-adsorption/desorption. Then we evaluated the effect of OTS crystalline structure after grafting OTS. Likewise, the XRD spectrum of
grafting onto the HY catalyst in the glycerol acetalization reaction with OTS-HY catalyst confirmed the presence of a highly crystalline HY
various conditions. zeolite [58] (Fig. 4). The HRTEM and XRD results suggested a negligible
loss of crystallinity after grafting OTS onto HY.
3.1. Characterization of the organosilane–grafted HY zeolites To evaluate how grafting OTS onto HY affected the catalyst’s acid
sites, we measured changes in surface area, pore-volume, and acidity by
The FTIR and DRIFT spectra of the OTS-HY catalyst confirmed the N2-adsorption/desorption and NH3-TPD. The N2 adsorption/desorption
successful grafting of OTS onto the HY catalyst. The skeletal FTIR isotherms of HY and OTS-HY catalysts exhibited the type IV isotherm
spectra of the unmodified HY catalyst presented an asymmetric [43], indicating that both catalysts were microporous (Fig. S2). The
stretching of the TeOeT bridges at 1056 and 1198 cm−1 (Fig. 2B), estimated surface area and pore volume of HY zeolite were 513 m2/g
where T is tetrahedrally coordinated Si or Al atoms [45,46]. Upon OTS and 0.36 cm3/g (Table 1), consistent with reported values [40]. Pre-
grafting onto HY catalyst, the characteristic TeOeT band at 1056 cm-1 vious investigators reported that grafting organosilanes onto zeolites
shifted to 1050 cm-1 because of the formation of the Si-O bond on the reduced the surface area and pore volume because some portion of the
tetrahedral TeOeT. We observed a similar down-shift of the zeolite’s pores was occupied by the organosilanes [41]. In addition, we

3
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

Fig. 2. TGA profiles (A), FTIR spectra (B), and DRIFT spectra (C) of OTS-HY and HY catalysts.

speculated that the decrease in the surface area and pore volume of our
OTS-HY catalysts might be partly because of the dilution contributed
from the grafted OTS. Moreover, our results revealed a slight decrease
in the total acidity of OTS-HY catalysts (552 μmol NH3/g catalyst),
measured by NH3-TPD, compared with that of HY catalyst (578 μmol
NH3/g catalyst), and this behavior is consistent with previous ob-
servations [41,59]. The slight change in the total acidity of the OTS-HY
catalysts agreed with our DRIFT results, showing comparable skeleton
OH groups (i.e., at 3627 and 3562 cm−1) and suggesting the retention
of acidity after the OTS grafting. We found that the grafting OTS onto
HY took place mostly on the external surface of the HY catalysts,
leaving the Brønsted and Lewis acid sites intact.

3.2. Emulsion formation by modified catalysts in two-phase systems

Typically, the HY catalyst contains many free silanols, making it

hydrophilic and limiting its suspension in the less polar acetone phase.
Thus, we investigated the catalyst suspension behaviour by placing HY
and OTS-HY in two liquid systems, (1) water-dodecane, and (2) gly-
cerol-acetone. We chose the water-dodecane system to represent the Fig. 4. XRD patterns of OTS-HY and HY catalysts.
polar-nonpolar solvent system. In the water-dodecane system, the OTS-
HY catalyst was suspended in dodecane, whereas the HY was suspended
in water (Fig. S3A & B). These visual observations were consistent with

Fig. 3. HRTEM images of HY and OTS-HY catalysts (A & D) and their contact angles (B & E). The suspension behaviour of HY and OTS-HY catalysts in the glycerol-
acetone system (C & F).

4
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

Table 1 phase reaction systems in which mass transfer limitation (limited con-
Surface properties and acidity of HY and OTS-HY catalysts. tact between reactants) inhibits reactivity. For example, the use of solid
Catalyst SBET (m2/g) Vpore Vmicro dpore Acidity particles, such as OTS-HY, to stabilize the emulsion enables easy
(cm³/g) (cm³/g) (Å) (μmol NH3/g catalyst) breaking of the emulsion and recovery of the two phases. Moreover,
this approach improves the contact between the two immiscible re-
Weak Strong Total actants without adding surfactants. Although adding surfactants in two-
HY 513 0.36 0.22 7.3 265 313 578 phase reactions can improve the contact between immiscible liquid
OTS-HY 347 0.23 0.14 6.5 286 267 552 reactants, adding surfactants complicates the downstream product
purification and adds to the carbon footprint of the process [66].
Note: SBET= surface area; Vpore= pore volume; Vmicro = micropore volume;
dpore = pore diameter.
3.3. Catalytic activity of OTS-HY catalysts in glycerol acetalization
the expectation that grafting OTS onto HY would add a hydrophobic
layer to the HY surface, enabling it to be suspended in the nonpolar Next, we demonstrated the benefit of the emulsion formation by
dodecane phase. These results agreed Zapata et al. who showed that. in OTS-HY in glycerol acetalization under reaction conditions. At 30 °C
a water-decalin system, a pristine hydrophilic catalyst settled in the upon agitation after 10 s, OTS-HY began dispersing in the glycerol
water phase [59], whereas an organosilane-grafted catalyst dispersed in (bottom phase) and forming an emulsion (See Video1, Supplementary
the decalin phase. Information). Conversely, the HY only stayed in the glycerol phase re-
To determine the degree of hydrophobicity of our OTS-HY catalyst, gardless of the reaction time (Fig. S6). An increase in reaction tem-
we measured its contact angle and observed its behaviour in the two- perature to 50 °C improved the glycerol solubility in acetone and en-
phase systems. The OTS-HY had a contact angle of ∼110°, suggesting abled the emulsion formation by OTS-HY after 10 s (Fig. S7). Moreover,
that OTS-HY was hydrophobic [60] (Fig. 3B). The water droplet for HY a 50 °C reaction temperature caused a larger HY suspension area com-
adsorbed into the disc, i.e., the contact angle was ∼0° (Fig. 3E). Our pared with the area at 30 °C, indicating that increasing reaction tem-
contact angle results agreed with another study by Zapata et al. [41] perature increased glycerol solubility in acetone and enabled more
that showed a high degree of hydrophobicity of OTS-grafted HY catalyst contact area between catalyst active sites and the two reactants. After
(126-135°). In addition, we conducted the water adsorption experi- 1 min, HY was well-dispersed in both phases (Fig. S7C). A further in-
ments using the OTS-HY and HY catalysts in a closed container with crease in reaction temperature to 70 °C enabled the higher glycerol
100 mL DI water in a 200 mL beaker and measured the water absorp- solubility in acetone caused HY to be well-dispersed in both phases after
tion during 12 h. The pristine HY catalyst had ∼3 times higher water 10 s (Fig. S8). These results suggested that the OTS-HY catalysts itself
adsorption capacity than the OTS-HY catalyst (Fig. S4), an observation assisted in emulsion formation, thus eliminating the need for higher
that corresponded with the TGA results. These results confirmed the reaction temperatures.
high degree of hydrophobicity of OTS-HY and the hydrophilicity of HY. The emulsion formation resulted in better contact between reactants
Nonetheless, these water adsorption results suggested that OTS-HY was with active sites of catalysts, leading to an improvement in the catalyst
not completely hydrophobic because water could still adsorb onto its activity (Fig. 5A & B). We applied OTS-HY catalyst to glycerol acet-
surface. alization at 30 °C and 12 wt.% glycerol for 600 min (10 h). We arbi-
The locations of the OTS-HY and HY catalysts in the reaction vessel trarily chose the 12 wt.% glycerol in acetone to provide excess acetone.
further confirmed the difference in the catalysts’ surface properties. The We performed similar experiments with HY catalyst as control. Solketal
glycerol-acetone system was a two-phase system due to low glycerol was the main reaction product with a trace amount of 6-MR (Table 2).
solubility in acetone. The OTS-HY catalyst was suspended at the in- Glycerol conversion increased with increasing reaction time and leveled
terface between glycerol and acetone (Fig. 3C) because, although
acetone is a polar solvent, it is less polar than glycerol. Hence, OTS-HY Table 2
The glycerol conversion and products selectivity of investigated catalysts as a
did not disperse in the acetone (Fig. 3F). Conversely, the pristine HY
function of temperature and time. Reaction condition: 5 wt.% catalyst loading,
catalyst was suspended only in the glycerol phase because of its
acetone/glycerol molar ratio of 12/1.
abundant silanol groups; the confinement of HY to the glycerol was
another indication of its hydrophilicity. Previous work has shown that
hydrophobic solid particles (contact angle > 90°) stabilize the forma-
tion of water-in-oil emulsions [41,61]. We hypothesized that the high
degree of hydrophobicity of OTS-HY catalysts would stabilize an
emulsion between glycerol and acetone.
Thus, we investigated emulsion formation by placing the OTS-HY Catalyst Temp. Time Glycerol Selectivity (%)
catalyst in the aforesaid same two liquid systems: (1) water-dodecane;
(°C) (min) conv. (%) Solketal Acetal
and (2) glycerol-acetone. First, we used a non-ionic surfactant, Tween
60, as a control. For the water-dodecane system, we added 1 wt.% HY 30 10 7 83 17
Tween 60 and vortexed for one min. The microscopic images, taken 20 10 87 13
after leaving this system for 30 min, showed water droplets dispersed in 60 28 88 12
480 85 92 8
the dodecane, indicating the formation of a stable emulsion layer (Fig. 600 89 98 2
S5C). Similarly, OTS-HY also stabilized the formation of the emulsion HY 50 10 60 93 7
layer in the water-dodecane system (Fig. S5A). These results suggested 20 65 94 6
that the OTS-HY catalyst assisted in the formation of Pickering emul- 60 88 98 2
480 88 98 2
sion in the water-dodecane system [62–65]. In contrast, in the glycerol-
600 88 98 2
acetone system, after adding either Tween 60 or OTS-HY and vortexing OTS-HY 30 10 73 88 12
for one min, we initially observed the formation of emulsion layer. 20 78 92 8
However, this emulsion layer broke and separated back into two phases 60 89 95 5
after stopping agitation (Fig. S5B & D) (See Video2, Supplementary In- 120 89 96 4
480 89 98 2
formation).
The formation of Pickering emulsion has many benefits in two- Note: Temp. = Temperature; Glycerol conv. = Glycerol conversion.

5
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

reaction temperature from 30 to 50 °C simply improved the reaction

rates. We observed the evolution of the phases during the course of the
reaction. At 12 wt.% glycerol in acetone, the system was a two-phase
system. At 30 °C, the two-phase glycerol-acetone system became one
phase after 60 min using OTS-HY catalysts, consistent with the 89%
glycerol conversion. In the case of HY catalyst at 30 °C, the reaction
system remained two phases even after 60 min, consistent with the 28%
glycerol conversion. Consistent with the glycerol conversion results,
these findings suggested the faster reaction rate was achieved by OTS-
HY catalysts.
An increase in reaction temperature from 30 to 50 °C raised the
catalytic activity of HY to nearly that of OTS-HY catalyst at 30 °C
(Fig. 5D). One reason for this increased activity of HY was that an in-
crease in temperature increased the glycerol solubility in acetone, en-
hancing the contact between HY’s active sites and both reactants. Yu
et al. reported a similar behaviour for SDS surfactants adsorbed on the
surface of multi-wall carbon nanotubes (MWCNTs) that enhanced sus-
pension of MWCNTs in aqueous solution [70].

3.4. Phase transition in glycerol acetalization due to the formation of

solketal
Fig. 5. Suspension behaviour of OTS-HY and HY catalysts in the glycerol-
acetone system (A & B). Glycerol conversion by OTS-HY and HY catalysts over Another important consideration was whether the formation of the
time (C). Changes in suspension behaviour of HY at 30 and 50 °C (D). product, solketal, affected the transition from two immiscible phases to
one single phase during the reaction. We performed a series of ex-
off at 89% after 60 min for OTS-HY catalyst at 30 °C (Fig. 5C). This periments to demonstrate the behavior of the reaction mixture at
plateau suggested that the reaction equilibrium was approached after varying reactants (glycerol and acetone) and products (solketal and
60 min. Our equilibrium conversion of 89% over the OTS-HY catalyst at water) concentration (i.e., simulated glycerol conversion between
30 °C was consistent with the previously reported value [67]. 0–100%) (Fig. 6). On the existence of the two immiscible phases, a slow
In the case of HY catalyst at 30 °C, the reaction conversion reached mass transfer of acetone to glycerol phase was indicated at the begin-
89% after 10 h. An increase in reaction temperature to 50 °C yielded the ning of the reaction because of a small supply of acetone. Upon the
glycerol conversion of 88% after 60 min. These results suggested that emulsion formation, the accumulation of the acetone in the glycerol
using OTS-HY catalysts reached equilibrium faster than HY catalysts at phase promoted an increase in the overall rate. During the reaction,
30 °C. Moreover, glycerol acetalization is an exothermic reaction. We solketal is formed and accumulated. When the solketal yield exceeded
observed the equilibrium conversion decreased slightly with increasing 25%, it behaved as a solubilizing agent to make a homogenous mixture
temperature using HY catalysts at 720 min (12 h) (Fig. 5C). This slight between glycerol and acetone, followed by a rapid conversion of gly-
decrease in equilibrium conversion (< 2%) at elevated temperature cerol because the interfacial mass transfer was eliminated.
was consistent with previous studies in both flow reactor with ethanol We further explored how this phase transition affected the inter-
as a solvent and solvent-free batch reactor [67–69]. An increase in facial mass transfer limitations in glycerol acetalization over HY cata-
lysts at 30 °C. When we performed the reaction under two-phase system

Fig. 6. Phase transition of the glycerol-acetone mixture during the course of reaction.

6
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

poor miscibility of acetone in glycerol phase caused a slow supply of

acetone for the formation of solketal. When the solketal content ex-
ceeded 25%, it acted as a solubilizing agent and the reaction mixture
became homogeneous, followed by a rapid production of solketal. We
proposed to overcome this initial interfacial mass transfer limitation by
grafting the n-octadecyltrichlorosilane (OTS) onto an HY catalyst. The
modified HY catalyst was hydrophobic and assisted the emulsion for-
mation during agitation. The emulsion formation improved the contact
between the ordinarily immiscible reactants and active sites of the OTS-
HY catalyst, yielding 89% glycerol conversion, compared with 28% for
unmodified HY, at 30 °C. In addition, we revealed that the formation of
reaction product, solketal, beyond 25% acted as a solvent to solubilize
glycerol and acetone into a single phase. Hydrophobic zeolites of this
type are promising for other acid-catalyzed reactions to upgrade by-
product glycerol from biodiesel industry, such as esterification, dehy-
dration to acrolein, and etherification with alcohols to produce fuel
additives.

CRediT authorship contribution statement

Fig. 7. Effect of catalyst loading on the glycerol conversion between 20–140
min.
Mohammad Shahinur Rahaman: Investigation, Formal analysis.
Thanh Khoa Phung: Investigation, Formal analysis, Writing - original
(i.e., conversion < 25%, see Fig. 7), the glycerol conversion increased draft. Md. Anwar Hossain: Investigation, Formal analysis. Emtias
with increasing the catalyst loading only up to 5 wt.%. The higher Chowdhury: Investigation. Sarttrawut Tulaphol: Conceptualization,
amount of catalyst loading did not increase the glycerol conversion, Methodology, Validation, Visualization. Shashi B. Lalvani:
suggesting that the reaction was controlled by the mass transfer of the Conceptualization, Supervision. Martin O’Toole: Supervision. Gerald
reactant from one phase to the other (i.e., the mass transfer of acetone A. Willing: Supervision. Jacek B. Jasinski: Methodology,
to the glycerol phase). However, when the reaction proceeded and Investigation, Visualization. Mark Crocker: Conceptualization, Writing
conversion is > 50%, the glycerol conversion continuously increased - review & editing, Supervision. Noppadon Sathitsuksanoh:
with increasing catalyst loading because the interfacial mass transfer Conceptualization, Methodology, Writing - review & editing,
limitation was minimized (i.e., the initial two-phase system became Supervision, Project administration.
homogenous (a single-phase) and the reaction became kinetically-con-
trolled. These findings supported the existence of interfacial mass Acknowledgment
transfer limitations. Moreover, these results are consistent with the
phase change behavior over the course of the reaction time (Fig. 6), A part of this material is based upon work supported by the National
affecting the apparent reaction rates. Science Foundation under Cooperative Agreement No. 1355438. This
Previous investigators have proposed the use of solvent (methanol, work was performed in part at the Conn Center for Renewable Energy
ethanol, and/or dichloromethane) to improve miscibility of glycerol Research at the University of Louisville, which belongs to the National
and acetone, overcoming the mass transfer limitation of the glycerol Science Foundation NNCI KY Manufacturing and Nano Integration
acetalization [71,72]. Glycerol conversion reached 73% at 40 °C be- Node, supported by ECCS-1542174. The authors would like to thank Dr.
cause the reaction rate was increased by both reactants being in the Howard Fried for his valuable comments and suggestions on the
same phase [72]. However, the use of solvents to enhance the mis- manuscript.
cibility of glycerol and acetone complicates downstream processing
because of additional unit operations for product separation/purifica- Appendix A. Supplementary data
tion. Our hydrophobically modified zeolite catalysts had higher cata-
lytic activity for glycerol acetalization than that of HY catalysts at low Supplementary material related to this article can be found, in the
temperature because OTS-HY catalysts possessed two key advantages. online version, at doi:https://doi.org/10.1016/j.apcata.2019.117369.
First, the grafted hydrophobic layers of OTS-HY catalysts form emulsion
between two immiscible reactants, minimizing the mass transfer lim- References
itation by improving contacts between the two immiscible reactants
without added solvents. Second, the OTS-HY catalysts still had the [1] M.M. Rojas-Downing, A.P. Nejadhashemi, T. Harrigan, S.A. Woznicki, Climate
comparable acid site density to that of HY, enabling the glycerol acet- change and livestock: Impacts, adaptation, and mitigation, Clim. Risk Manage. 16
alization at the interface of two immiscible reactants. The high catalytic (2017) 145–163.
[2] U.S. Environmental Protection Agency, Inventory of U.S. Greenhouse Gas Emissions
activity of OTS-HY at low temperatures makes it a promising candidate and Sinks: 1990-2017, Executive Summary, (2019) . (access Nov 19, 2019) https://
for glycerol acetalization and other acid-catalyzed two-phase reactions. www.epa.gov/ghgemissions/inventory-us-greenhouse-gas-emissions-and-sinks-
The alkyl chain length of the organosilane surfactant affects the degree 1990-2017.
[3] A. Paquette, J. Vayreda, L. Coll, C. Messier, J. Retana, Climate change could negate
of hydrophobicity of the catalysts. We are currently correlating the positive tree diversity effects on forest productivity: a study across five climate
degree of hydrophobicity of the modified catalyst with its catalytic types in Spain and Canada, Ecosystems 21 (2018) 960–970.
activity in glycerol acetalization. [4] C. Nolan, J.T. Overpeck, J.R. Allen, P. Anderson, J. Betancourt, H. Binney,
S. Brewer, M. Bush, B. Chase, R. Cheddadi, Past and future global transformation of
terrestrial ecosystems under climate change, Science 361 (2018) 920–923.
4. Conclusion [5] O. Edenhofer, Climate Change 2014: Mitigation of Climate Change, Cambridge
University Press, 2015.
[6] T. Boden, R. Andres, G. Marland, Global, Regional, and National Fossil-fuel CO2
Acetalization of glycerol with acetone is an acid-catalyzed reaction. Emissions (1751-2014)(v. 2017), Oak Ridge, TN: Carbon Dioxide Information
The poor miscibility between glycerol and acetone results in an initial Analysis Center (CDIAC), Oak Ridge National Laboratory (ORNL), 2017.
two-phase mixture and presents a considerable interfacial mass transfer [7] K. Anderson, G. Peters, The trouble with negative emissions, Science 354 (2016)
182–183.
limitation between liquid phases and catalyst’s active sites. Initially, the

7
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

[8] R.W. Howarth, Ideas and perspectives: is shale gas a major driver of recent increase reaction, React. Chem. Eng. 4 (2019) 235–243.
in global atmospheric methane? Biogeosci. 16 (2019) 3033–3046. [39] M. Moreira, R. Faria, A. Ribeiro, A. Rodrigues, Solketal Production from Glycerol
[9] U.S.E.I. Administration, Monthly Biodiesel Production Survey, (2018). Ketalization with Acetone: Catalyst Selection and Thermodynamic and Kinetic
[10] F. Yang, M.A. Hanna, R.C. Sun, Value-added uses for crude glycerol–a byproduct of Reaction Study, Ind. Eng. Chem. Res. 58 (2019) 17746–17759.
biodiesel production, Biotechnol, Biotechnol. Biofuels 5 (2012) 13. [40] H.C. Genuino, S. Thiyagarajan, J.C. van der Waal, E. de Jong, J. van Haveren,
[11] M. Ayoobi, I. Schoegl, Non-catalytic conversion of glycerol to syngas at inter- D.S. van Es, B.M. Weckhuysen, P. Bruijnincx, Selectivity Control in the Tandem
mediate temperatures: Numerical methods with detailed chemistry, Fuel 195 Aromatization of Bio‐Based Furanics Catalyzed by Solid Acids and Palladium,
(2017) 190–200. ChemSusChem 10 (2017) 277–286.
[12] M.R.K. Estahbanati, M. Feilizadeh, M.C. Iliuta, Photocatalytic valorization of gly- [41] P.A. Zapata, Y. Huang, M.A. Gonzalez-Borja, D.E. Resasco, Silylated hydrophobic
cerol to hydrogen: Optimization of operating parameters by artificial neural net- zeolites with enhanced tolerance to hot liquid water, J. Catal. 308 (2013) 82–97.
work, Appl. Catal. B 209 (2017) 483–492. [42] H. Wang, H. Ruan, M. Feng, Y. Qin, H. Job, L. Luo, C. Wang, M. Engelhard, E. Kuhn,
[13] S. García-Fernández, I. Gandarias, J. Requies, F. Soulimani, P. Arias, X. Chen, M. Tucker, B. Yang, One-Pot Process for Hydrodeoxygenation of Lignin to
B. Weckhuysen, The role of tungsten oxide in the selective hydrogenolysis of gly- Alkanes Using Ru-Based Bimetallic and Bifunctional Catalysts Supported on Zeolite
cerol to 1, 3-propanediol over Pt/WOx/Al2O3, Appl. Catal. B 204 (2017) 260–272. Y, ChemSusChem 10 (2017) 1846–1856.
[14] M. Massa, A. Andersson, E. Finocchio, G. Busca, Gas-phase dehydration of glycerol [43] P. Prinsen, R. Luque, C. González-Arellano, Zeolite catalyzed palmitic acid ester-
to acrolein over Al2O3-, SiO2-, and TiO2-supported Nb-and W-oxide catalysts, J. ification, Microporous Mesoporous Mater. 262 (2018) 133–139.
Catal. 307 (2013) 170–184. [44] E. Pires, J. Magalhães, U. Schuchardt, Effects of oxidant and solvent on the liquid-
[15] C.H. Zhou, J.N. Beltramini, Y.X. Fan, G.Q. Lu, Chemoselective catalytic conversion phase cyclohexane oxidation catalyzed by Ce-exchanged zeolite Y, Appl. Catal. A
of glycerol as a biorenewable source to valuable commodity chemicals, Chem. Soc. Gen. 203 (2000) 231–237.
Rev. 37 (2008) 527–549. [45] E. Astorino, J.B. Peri, R.J. Willey, G. Busca, Spectroscopic Characterization of
[16] D.C. de Carvalho, A.C. Oliveira, O.P. Ferreira, J.M. Filho, S. Tehuacanero-Cuapa, Silicalite-1 and Titanium Silicalite-1, J. Catal. 157 (1995) 482–500.
A.C. Oliveira, Tehuacanero-Cuapa, A.C. Oliveira, Titanate nanotubes as acid cata- [46] T.K. Phung, M.M. Carnasciali, E. Finocchio, G. Busca, Catalytic conversion of ethyl
lysts for acetalization of glycerol with acetone: Influence of the synthesis time and acetate over faujasite zeolites, Appl. Catal. A Gen. 470 (2014) 72–80.
the role of structure on the catalytic performance, Chem. Eng. J. 313 (2017) [47] M. Junaidi, C. Khoo, C. Leo, A. Ahmad, The effects of solvents on the modification
1454–1467. of SAPO-34 zeolite using 3-aminopropyl trimethoxy silane for the preparation of
[17] G.P. da Silva, M. Mack, J. Contiero, A promising and abundant carbon source for asymmetric polysulfone mixed matrix membrane in the application of CO2 se-
industrial microbiology, Biotechnol. Adv. 27 (2009) 30–39. paration, Microporous Mesoporous Mater. 192 (2014) 52–59.
[18] V.K. Garlapati, U. Shankar, A. Budhiraja, Bioconversion technologies of crude [48] D. Flinn, D. Guzonas, R.-H. Yoon, Characterization of silica surfaces hydrophobized
glycerol to value added industrial products, Biotechnol. Rep. Amst. (Amst) 9 (2016) by octadecyltrichlorosilane, Colloids Surf. A Physicochem. Eng. Asp. 87 (1994)
9–14. 163–176.
[19] M. da Silva, F. de Ávila Rodrigues, A. Júlio, SnF2-catalyzed glycerol ketalization: A [49] A. Kumar, J. Richter, J. Tywoniak, P. Hajek, S. Adamopoulos, U. Šegedin, M. Petrič,
friendly environmentally process to synthesize solketal at room temperature over Surface modification of Norway spruce wood by octadecyltrichlorosilane (OTS)
on solid and reusable Lewis acid, Chem. Eng. J. 307 (2017) 828–835. nanosol by dipping and water vapour diffusion properties of the OTS-modified
[20] T.N. Pham, T. Sooknoi, S.P. Crossley, D.E. Resasco, Ketonization of carboxylic acids: wood, Holzforschung 72 (2017) 45–56.
mechanisms, catalysts, and implications for biomass conversion, ACS Catal. 3 [50] K. Yamagishi, S. Namba, T. Yashima, Defect sites in highly siliceous HZSM-5 zeo-
(2013) 2456–2473. lites: a study performed by alumination and IR spectroscopy, J. Phys. Chem. 95
[21] M.R. Nanda, Y. Zhang, Z. Yuan, W. Qin, H.S. Ghaziaskar, C. Xu, Catalytic conver- (1991) 872–877.
sion of glycerol for sustainable production of solketal as a fuel additive: A review, [51] T. Montanari, E. Finocchio, G. Busca, Infrared spectroscopy of heterogeneous cat-
Renew. Sustain. Energy Rev. 56 (2016) 1022–1031. alysts: acidity and accessibility of acid sites of faujasite-type solid acids, J. Phys.
[22] S. Gadamsetti, N.P. Rajan, G.S. Rao, K.V.R. Chary, Acetalization of glycerol with Chem. C 115 (2010) 937–943.
acetone to bio fuel additives over supported molybdenum phosphate catalysts, J. [52] T. Korányi, F. Moreau, V. Rozanov, E. Rozanova, Identification of SH groups in
Mol. Cat. A Chem. 410 (2015) 49–57. zeolite-supported HDS catalysts by FTIR spectroscopy, J. Mol. Struct. 410-411
[23] M. Gonçalves, R. Rodrigues, T.S. Galhardo, W.A. Carvalho, Highly selective acet- (1997) 103–110.
alization of glycerol with acetone to solketal over acidic carbon-based catalysts [53] M. Makarova, A. Ojo, K. Karim, M. Hunger, J. Dwyer, FTIR study of weak hydrogen
from biodiesel waste, Fuel 181 (2016) 46–54. bonding of Broensted hydroxyls in zeolites and aluminophosphates, J. Phys. Chem.
[24] C.J.A. Mota, C.X.A. da Silva, N. Rosenbach, J. Costa, F. da Silva, Glycerin 98 (1994) 3619–3623.
Derivatives as Fuel Additives: The Addition of Glycerol/Acetone Ketal (Solketal) in [54] I. Tsuchiya, Infrared spectroscopic study of hydroxyl groups on silica surfaces, J.
Gasolines, Energy Fuel. 24 (2010) 2733–2736. Phys. Chem. 86 (1982) 4107–4112.
[25] P.H.R. Silva, V.L.C. Gonçalves, C.J.A. Mota, Glycerol acetals as anti-freezing ad- [55] I.-S. Chuang, G.E. Maciel, Probing hydrogen bonding and the local environment of
ditives for biodiesel, Bioresour. Technol. 101 (2010) 6225–6229. silanols on silica surfaces via nuclear spin cross polarization dynamics, JACS 118
[26] L. Fertier, M. Ibert, C. Buffe, R. Saint-Loup, C. Joly-Duhamel, J.J. Robin, O. Giani, (1996) 401–406.
New biosourced UV curable coatings based on isosorbide, Prog. Org. Coat. 99 [56] S. Mirji, S. Halligudi, D.P. Sawant, N.E. Jacob, K. Patil, A. Gaikwad, S. Pradhan,
(2016) 393–399. Adsorption of octadecyltrichlorosilane on mesoporous SBA-15, Appl. Surf. Sci. 252
[27] B.A.J. Noordover, A. Heise, P. Malanowksi, D. Senatore, M. Mak, L. Molhoek, (2006) 4097–4103.
R. Duchateau, C.E. Koning, R.A.T.M. van Benthem, Biobased step-growth polymers [57] J. Parise, D. Corbin, L. Abrams, D. Cox, Structure of dealuminated Linde Y-zeolite;
in powder coating applications, Prog. Org. Coat. 65 (2009) 187–196. Si139. 7Al52. 3O384 and Si173. 1Al18. 9O384: presence of non-framework Al
[28] M. Durand, V. Molinier, T. Féron, J.-M. Aubry, Isosorbide mono- and di-alkyl ethers, species, Acta Crystallogr, Sect. C: Cryst. Struct. Commun 40 (1984) 1493–1497.
a new class of sustainable coalescents for water-borne paints, Prog. Org. Coat. 69 [58] V. Felice, A.C. Tavares, Faujasite zeolites as solid electrolyte for low temperature
(2010) 344–351. fuel cells, Solid State Ion. 194 (2011) 53–61.
[29] J.I. Garcia, H. Garcia-Marin, E. Pires, Glycerol based solvents: synthesis, properties [59] P.A. Zapata, J. Faria, M.P. Ruiz, R.E. Jentoft, D.E. Resasco, Hydrophobic zeolites for
and applications, Green Chem. 16 (2014) 1007–1033. biofuel upgrading reactions at the liquid–liquid interface in water/oil emulsions,
[30] C. Chen, J.-T. Hu, Y.-J. Tu, J.-C. Wu, J. Liang, L.-L. Gao, Z.-G. Wang, B.-F. Yang, D.- JACS 134 (2012) 8570–8578.
L. Dong, Effects of isosorbide mononitrate on the restoration of injured artery in [60] N.-Y. Topsøe, K. Pedersen, E.G. Derouane, Infrared and temperature-programmed
mice in vivo, Eur. J. Pharm. 640 (2010) 150–156. desorption study of the acidic properties of ZSM-5-type zeolites, J. Catal. 70 (1981)
[31] Z.Q. Li, X. He, X. Gao, Y.Y. Xu, Y.-F. Wang, H. Gu, R.-F. Ji, S.-J. Sun, Study on 41–52.
dissolution and absorption of four dosage forms of isosorbide mononitrate: Level A [61] B. Binks, S. Lumsdon, Catastrophic phase inversion of water-in-oil emulsions sta-
in vitro–in vivo correlation, Eur. J. Pharm. Biopharm. 79 (2011) 364–371. bilized by hydrophobic silica, Langmuir 16 (2000) 2539–2547.
[32] S.-A. Park, J. Choi, S. Ju, J. Jegal, K.M. Lee, S.Y. Hwang, D.X. Oh, J. Park, [62] B. Binks, J. Rodrigues, W. Frith, Synergistic interaction in emulsions stabilized by a
Copolycarbonates of bio-based rigid isosorbide and flexible 1,4-cyclohex- mixture of silica nanoparticles and cationic surfactant, Langmuir 23 (2007)
anedimethanol: Merits over bisphenol-A based polycarbonates, Polymer 116 (2017) 3626–3636.
153–159. [63] S. Arditty, C. Whitby, B. Binks, V. Schmitt, F. Leal-Calderon, Some general features
[33] C. da Silva, V. Gonçalves, C. Mota, Water-tolerant zeolite catalyst for the acet- of limited coalescence in solid-stabilized emulsions, Eur. Phys. J. E 11 (2003)
alisation of glycerol, Green Chem. 11 (2009) 38–41. 273–281.
[34] C. Ferreira, A. Araujo, V. Calvino-Casilda, M. Cutrufello, E. Rombi, A. Fonseca, [64] B. Binks, J. Clint, from surface energy components: relevance to Pickering emul-
M. Bañares, I.C. Neves, Y zeolite-supported niobium pentoxide catalysts for the sions, Langmuir 18 (2002) 1270–1273.
glycerol acetalization reaction, Microporous Mesoporous Mater. 271 (2018) [65] S. Pickering, Chem, J. Chem. Soc., Trans. 91 (1907) 2001–2021.
243–251. [66] M. Pera‐Titus, L. Leclercq, J.M. Clacens, F. De Campo, V. Nardello‐Rataj, Pickering
[35] C.-N. Fan, C.-H. Xu, C.-Q. Liu, Z.-Y. Huang, J.-Y. Liu, Z.-X. Ye, Catalytic acetaliza- interfacial catalysis for biphasic systems: from emulsion design to green reactions,
tion of biomass glycerol with acetone over TiO2–SiO2 mixed oxides, Reac. Kinet. Angew. Chem. Int. Ed. 54 (2015) 2006–2021.
Mech. Cat. 107 (2012) 189–202. [67] A. Cornejo, M. Campoy, I. Barrio, B. Navarrete, J. Lázaro, Solketal production in a
[36] A. Behr, D. Obst, B. Turkowski, Isomerizing hydroformylation of trans-4-octene to solvent-free continuous flow process: scaling from laboratory to bench size, React.
n-nonanal in multiphase systems: acceleration effect of propylene carbonate, J. Mol. Chem. Eng. (2019).
Catal. A Chem. 226 (2005) 215–219. [68] I. Agirre, I. Garcia, J. Requies, V. Barrio, M. Güemez, J. Cambra, P. Arias, Glycerol
[37] C. Sievers, Y. Noda, L. Qi, E. Albuquerque, R. Rioux, S. Scott, Phenomena affecting acetals, kinetic study of the reaction between glycerol and formaldehyde, Biomass
catalytic reactions at solid–liquid interfaces, ACS Catal. 6 (2016) 8286–8307. Bioenergy 35 (2011) 3636–3642.
[38] N. Weeranoppanant, Enabling tools for continuous-flow biphasic liquid–liquid [69] I. Agirre, M. Güemez, A. Ugarte, J. Requies, V. Barrio, J. Cambra, P. Arias, Glycerol

8
M.S. Rahaman, et al. Applied Catalysis A, General 592 (2020) 117369

acetals as diesel additives: Kinetic study of the reaction between glycerol and and kinetic studies of a catalytic process to convert glycerol into solketal as an
acetaldehyde, Fuel Process. Technol 116 (2013) 182–188. oxygenated fuel additive, Fuel 117 (2014) 470–477.
[70] J. Yu, N. Grossiord, C.E. Koning, J. Loos, Controlling the dispersion of multi-wall [72] J. Deutsch, A. Martin, H. Lieske, Investigations on heterogeneously catalysed con-
carbon nanotubes in aqueous surfactant solution, Carbon 45 (2007) 618–623. densations of glycerol to cyclic acetals, J. Catal. 245 (2007) 428–435.
[71] M. Nanda, Z. Yuan, W. Qin, H. Ghaziaskar, M.-A. Poirier, C. Xu, Thermodynamic

9
 Fuel 291 (2021) 120207

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Synthesis of solketal from glycerol over modified SiO2 supported

p-phenolsulfonic acid catalyst
Ruru Zhou a, 1, Yuanyuan Jiang a, 1, Huaiyuan Zhao a, Boyong Ye a, Lina Wang b, Zhaoyin Hou a, c, *
a
Key Laboratory of Biomass Chemical Engineering of Ministry of Education, Department of Chemistry, Zhejiang University, Hangzhou 310028, China
b
Key Laboratory of Advanced Textile Materials and Manufacturing Technology (Zhejiang Sci-tech University), Hangzhou 310012, China
c
Center of Chemistry for Frontier Technologies, Department of Chemistry, Zhejiang University, Hangzhou 310028, China

A R T I C L E I N F O A B S T R A C T

Keywords: Acetalization of glycerol with acetone could convert surplus glycerol into an important fuel additive (solketal). In
Solid acid catalyst this work, a stable solid acid catalyst was synthesized in a facile polycondensation of p-phenolsulfonic acid on
Resin KH560 modified SiO2 (PSF/K-SiO2) and used in the acetalization of glycerol at room temperature. Character­
p-Phenolsulfonic acid
ization results confirmed that PSF/K-SiO2 possessed larger surface area and surface acidity than that of PSF resin
Glycerol
(copolymer of p-phenolsulfonic acid and formaldehyde) and non-modified SiO2 supported PSF resin (PSF/SiO2).
Acetalization
As a result, PSF/K-SiO2 was extremely active and stable in the acetalization of glycerol. At 25 ◦ C, the detected
conversion of glycerol reached 86.3% with a 97.7% selectivity of solketal at 1.5 h. The best yield of solketal over
PSF/K-SiO2 could reach 93.1 g-solketal/g-cat/h, and PSF/K-SiO2 could be recycled at least 5 times without
obvious decrease in reactivity.

1. Introduction metal phosphates [24,25] and sulfated carbon [8,26,27]. All these works
confirmed that acidity is the most important factor for acetalization of
Glycerol, one of the top 12 platform chemicals, is an inevitable by- glycerol with acetone, and the yield of solketal depended mainly on the
product during the production of biodiesel and the yield of glycerol acid strength and/or surface acid density of catalysts [7,8,16,17].
increased rapidly with the expanding application of biodiesel [1–5]. It Meanwhile, the acetalization reaction between glycerol and acetone is
was estimated that the global production of glycerol could reach 7.66 an exothermic reversible process, and low reaction temperature is more
million tons in 2020 [6]. Thus, catalytic transformation of surplus favorable for increasing the yield of solketal [24,28].
glycerol to value-added products was a hot topic in the past decade. Solid acid catalyst is one kind of important materials in traditional
Acetalization [7,8], dehydration [9,10], esterification [11,12], and petroleum refinery industry, and it played a crucial role in the cracking,
etherification of glycerol [13,14] were the popularly reported routines alkylation and isomerization processes [29]. And more recently, it was
for the high valued application of glycerol, and all these reactions were found that solid acid catalyst is also indispensable for the effective bio-
catalyzed by acids. Among of which, acetalization of glycerol with refinery technology, such as hydrolysis of cellulose [30–32], dehydra­
acetone is one promising way to convert glycerol into a value-added tion of sugar [33–35] and catalytic upgrading of platform chemicals
product, solketal (4-hydroxymethyl-2,2-dimethyl-1,3-dioxolane) [15]. [36–40]. The preparation and application of solid acids have attracted
Solketal is widely recommended as a fuel additive, which can reduce the attentions of many scientists for they are environmentally friendly, non-
formation of gum and soot as well as particle emissions, enhance the corrosive to instruments, easy in separation and recycle. But the main
octane number and improve the cold flow properties of liquid fuels drawback of solid acid catalyst lies in its uneven acid strength, low
[16–18]. Additionally, solketal is also a versatile solvent in polymer surface area, and tendency in deactivation [41–43].
industry and pharmaceutical production [19]. Recently, metal-free SO3H-functionalized polymers have been tried
In published works, several solid acids were tested in the synthesis of popularly in many acid-catalyzed reactions due to its strong acidity, high
solketal from glycerol, including zeolite [20,21], metal oxides [22,23], moisture tolerance and easiness in preparation and separation [44,45].

* Corresponding author at: Key Laboratory of Biomass Chemical Engineering of Ministry of Education, Department of Chemistry, Zhejiang University, Hangzhou
310028, China.
E-mail address: [email protected] (Z. Hou).
1
R. Zhou and Y. Jiang contributed equally to this work.

https://doi.org/10.1016/j.fuel.2021.120207
Received 24 October 2020; Received in revised form 14 December 2020; Accepted 11 January 2021
Available online 5 February 2021
0016-2361/© 2021 Elsevier Ltd. All rights reserved.
R. Zhou et al. Fuel 291 (2021) 120207

Scheme 1. Synthesis routine of PSF/K-SiO2.

Previous work in our group found that a copolymer of p-phenolsulfonic butanone, cyclohexanone, butyraldehyde and hexaldehyde with
acid and phenol exhibited excellent activity and stability in the esteri­ analytical reagent grade were purchased from Sinopharm Chemical
fication of glycerol with acetic acid under mild reaction conditions [11], Reagent Co. Ltd. and used as received without further purification.
but the surface area of these copolymers was quite low (<5.5 m2/g). Nano-silica powder (Aladdin) was calcined at 500 ◦ C for 4 h before use.
Therefore, it is of great importance to find a way for the preparation of
solid acid catalyst with strong acidity, high surface area and enhanced 2.2. Catalyst preparation
surface acid sites.
Herein, a stable solid acid catalyst (PSF/K-SiO2) with high surface At first, 1.2 g SiO2 powder was dispersed in cyclohexane, and then
area was synthesized in a facile polycondensation of p-phenolsulfonic 0.1 g KH560 was added dropwise under vigorous stirring. The mixture
acid on γ-(2,3-epoxypropoxy) propytrimethoxysilane (KH560) modified was stirred at 25 ◦ C for 4 h to ensure the formation of K-SiO2. After that,
SiO2. The structure and property of PSF/K-SiO2 were characterized and a mixture of 8 mmol p-phenolsulfonic acid and 16 mmol para­
compared with that of pure PSF resin (copolymer of p-phenolsulfonic formaldehyde in ethanol was added into the above suspension, raised to
acid and formaldehyde) and non-modified SiO2 supported PSF resin 120 ◦ C and kept at this temperature for 2 h under vigorous stirring.
(PSF/SiO2). The performance and stability of PSF/K-SiO2 catalyst were Subsequently, the reaction mixture was cooled to room temperature,
used in the acetalization of glycerol as a model reaction. and the resulted product was centrifuged and washed with ethanol for
several times. Finally, the solid product was dried at 60 ◦ C in vacuum for
2. Materials and methods 12 h, and denoted as PSF/K-SiO2. And this synthesis routine of PSF/K-
SiO2 was presented in Scheme 1. For comparison, PSF/SiO2 catalyst was
2.1. Materials also synthesized in the same process as above without KH560 in feed.

p-Phenolsulfonic acid, γ-(2,3-epoxypropoxy) propytrimethoxysilane 2.3. Characterization

(KH560), paraformaldehyde, cyclohexane, ethanol, glycerol, acetone,
Fourier transform infrared (FTIR) spectrometry was recorded in a
Nicolet Is10 FTIR spectrometer. 13C solid-state nuclear magnetic reso­
nance (13C SSNMR) experiments were carried out on a Bruker Avance III
HD 600 MHz spectrometer. Thermogravimetric analysis (TG) was con­
ducted on a TA Instruments SDT Q600 thermoanalyzer from 30 to
750 ◦ C with a ramp of 10 ◦ C/min in air flow (30 mL/min). Nitrogen
adsorption-desorption isotherms were measured on an ASAP 2010
analyzer after pretreated at 50 ◦ C for 4 h. Surface area of these samples
was calculated using Brunauer-Emmett-Teller (BET) method according
to the adsorption data. The morphology of these catalysts was obtained
on HITACHI Series SEM (SU8010, Japan) at an accelerating voltage of 5
kV, and samples were covered with platinum in order to avoid charging.
X-ray photoelectron spectroscopy (XPS) measurement was performed on
a Perkin-Elmer PHI ESCA System, and the X-ray source was Mg standard
anode 146 (1253.6 eV) operated at 12 kV and 300 W.

2.4. Acetalization of glycerol with acetone

Acetalization of glycerol with acetone was performed in a 50 mL

three-necked flask. A typical reaction mixture consisted of 10.86 mmol
glycerol, 108.6 mmol acetone and 0.05 g catalyst. Then the flask was
Fig. 1. FTIR spectra of SiO2, K-SiO2, PSF/K-SiO2 and PSF. placed in a preheated oil bath with magnetic stirring and refluxing.

2
R. Zhou et al. Fuel 291 (2021) 120207

13
Fig. 2. C SSNMR spectrum of PSF/K-SiO2.

During the reaction, the reaction mixture was vigorously stirred with a
strong magnetic stirrer (MAG-NEO, RV-06 M, Japan) at 1000 rpm in
order to eliminate possible heat and mass transfer limitations. After a
stipulated period of time of reaction, the mixture was cooled to room
temperature and centrifuged to separate the solid catalyst. Liquid phase
was analyzed by gas chromatograph (Shimadzu, 14B), equipped with a
flame ionization detector and a 30 m capillary column (DB-WAX 52CB,
USA). All products were identified by a gas chromatography-mass
spectrometry system (GC–MS, Agilent 6890) and quantified via an
external calibration method.
Recycle experiments were conducted in the same process and
aforementioned reaction conditions. And the recovered catalyst was Fig. 3. Thermal analysis curves of PSF (a) and PSF/K-SiO2 (b).
washed with ethanol and dried at 60 ◦ C for 12 h in a vacuum oven before
next recycle. structure and the binding of PSF resin with K-SiO2 (Fig. 2). The shift of
33.7 ppm was the methylene bridge (a) between p-phenolsulfonic acid.
3. Results and discussion The aromatic carbon (e) bonded with O was seen at 149.1 ppm [49],
while the small peak at 141.7 ppm could be ascribed to the carbon (b)
3.1. Characterization of catalyst bonded with sulfonic group [50]. The peaks at 129.2 and 134.5 ppm
indicated the existence of substituted and unsubstituted aromatic carbon
Fig. 1 compared the FTIR spectra of bare SiO2, K-SiO2, PSF/K-SiO2 (d and c) [49,51]. Several carbons (k, j, f, h, i and g) in KH560 could be
and PSF. The typical peaks at 1104, 799 and 471 cm− 1 were assigned to identified at the shift of 9.3, 23.3, 63.4, 67.4, 72.9 and 85.0 ppm,
the asymmetric stretching, symmetric stretching and bending vibration respectively. The methoxyl carbon (l) in KH560 appeared at the shift of
of Si–O–Si, respectively [46]. The band at 3437, 1638 and 958 cm− 1 59.0 ppm. Peaks at 63.4 and 85.0 ppm suggested that the epoxy ring in
disclosed that there are hydroxyl groups and adsorbed water molecules KH560 was condensed with the phenolic hydroxyl group of PSF [52].
on the surface of bare SiO2. And the broad band at 3437 cm− 1 (from the And this newly formed structure could improve the uniform dispersion
hydrogen bonds between water molecules) might cover the stretching of PSF resin and strengthen the interaction between PSF resin and K-
vibration of Si–OH. It was found that the intensity of band at 3437, 1638 SiO2. Besides, the peak at 16.0 ppm might be the methyl carbon from
and 958 cm− 1 decreased obviously when KH560 was introduced on residual solvent ethanol [53].
SiO2, indicating the less hydrophilic character of the resulting K-SiO2. In Fig. 3 compared the thermogravimetric analysis results of PSF and
addition, the characteristic peaks of –CH3 and –CH2– appeared at 2925 PSF/K-SiO2. It was found that there were two major mass loss stages in
and 2875 cm− 1 in K-SiO2. These results indicated that KH560 was suc­ PSF (Fig. 3a). The first stage below 250 ◦ C was the volatilization of
cessfully grafted on the surface of SiO2 via the silylation reaction be­ adsorbed solvent. Then the sulfonic groups began to decompose, fol­
tween the surface Si–OH groups and the terminal –Si(OCH3)3 group in lowed by the destruction of the polymeric framework above 250 ◦ C
KH560 (see Scheme 1) [47,48]. Moreover, it could be found that the [11,54]. The strong exothermic DTA peak at 463.9 ◦ C confirmed the
intensity of peaks at 2925 and 2875 cm− 1 was strengthened in the combustion of bulk resin. Compared with above phenomena of pure PSF
spectrum of PSF/K-SiO2 due to the stretching of aromatic and aliphatic resin, a new exothermic DTA peak at 300–400 ◦ C appeared in PSF/K-
C–H in PSF [49]. And the stretching of aromatic C– – C bond was also
SiO2 (see Fig. 3b), which might be ascribed to the combustion of linking
− 1
observed at 1478 cm , which indicated that the PSF resin was suc­ groups (HOCH2CH(O)CH2O(CH2)3Si(OCH3)2, derived from KH560, see
cessfully immobilized on K-SiO2 support. However, it is regretted to note Fig. 2) between PSF resin and SiO2. And this suggestion was also
that several bands, such as the asymmetric stretching of O– – S–– O at
confirmed by the TG-DTA curves of K-SiO2 (Fig. S1a) and reference [55].
1215 and 1165 cm− 1, the symmetric stretching of O– – S–– O at 1120 and
According to above mass loss in different stages, the calculated mass
1010 cm− 1 and the stretching of C–S at 1032 cm− 1, were overlapped ratio of linking groups in K-SiO2 was 6.5 wt% of SiO2 support (see
with the broad intense peak of Si–O–Si at 1104 cm− 1 [11]. Fig. S1a). And then, the calculated content of PSF resin in PSF/K-SiO2
The 13C SSNMR analysis of PSF/K-SiO2 further disclosed the fine catalyst was 48.5 wt% (see Fig. 3b). At the same time, it was interesting

3
R. Zhou et al. Fuel 291 (2021) 120207

Table 1 agglomeration and separated SiO2 grains was covered (Fig. S3b), indi­
Textural properties of SiO2, K-SiO2, PSF/K-SiO2, PSF/SiO2 and PSF. cating that KH560 was successfully grafted on SiO2. Many irregular
Sample Surface Pore PSF S content Acidity plates formed in PSF/K-SiO2 and it made the shaggy K-SiO2 particles
area (m2/ volume content (wt (wt%)b (mmol/g)c further coated with the PSF resin (Fig. 5a1 and a2), and this coverage
g) (cm3/g) %)a could make the lower surface area of PSF/K-SiO2 (than that of SiO2 and
SiO2 234.6 1.378 – – – K-SiO2) well understood. Furthermore, EDS analysis confirmed that C, O
K-SiO2 117.2 0.579 – – – and S elements distributed uniformly on K-SiO2 (Fig. 5a3 and a4).
PSF/K- 77.8 0.373 48.5 8.4 2.6 XPS analysis disclosed that the binding energy of S 2p in PSF/SiO2
SiO2
PSF/ 47.5 0.249 41.6 7.2 2.3
and PSF/K-SiO2 was 169.8 and 169.4 eV, respectively (Fig. 6b), and
SiO2 these data were similar as that reported in literature [56]. The lower
PSF 1.5 0.002 100.0 17.2 5.4 binding energy of S 2p in PSF/K-SiO2 would be ascribed to the electron
a
Concluded from TG results, and detailed calculation method was listed in
transfer from coupling agent to PSF resin, which also confirmed that the
supplementary materials. coupling agent strengthened the interaction between PSF resin and K-
b
S content, defined as (PSF content) × (atomic mass of S)/(calculated mo­ SiO2. The calculated relative bulk and surface elemental composition
lecular weight of PSF). (except H) of PSF/SiO2 and PSF/K-SiO2 that measured via TG and XPS
c
Calculated based on the S content. analysis were summarized in Table 2. It was found that the relative
surface content of S in PSF/K-SiO2 increased obviously from 0.9% (in
PSF/SiO2) to 1.4%. These results also confirmed that the coupling agent
strengthened the interaction between PSF resin and K-SiO2, which
facilitated the uniform dispersion of PSF resin. These data were in
agreement with above 13C SSNMR, TG and SEM analysises.

3.2. Catalytic performance

The performance of bulk PSF resin, PSF/SiO2 and PSF/K-SiO2 for the
acetalization of glycerol with acetone at 25 ◦ C was summarized in
Table 3. It was found that bare SiO2 and K-SiO2 were inactive, which
meant the active sites for this reaction came from PSF resin. The
detected conversion of glycerol over bulk PSF resin, PSF/SiO2 and PSF/
K-SiO2 reached 75.0, 86.6 and 86.3%, and the calculated space time
yield of solketal over these catalysts (STY, defined as (gram of solketal
formed)/(gram of catalyst)/(reaction time)) was 14.1, 16.3 and 16.1 g-
solketal/g-cat/ h, respectively. The lower activity of bulk PSF resin
might be ascribed to its lowest surface area (1.5 m2/g), which retarded
the accessibility of acid sites (see Table 1). On the other hand, the
enlarged surface area of PSF/SiO2 and PSF/K-SiO2 could promote the
Fig. 4. N2 adsorption-desorption isotherms of SiO2, K-SiO2, PSF/K-SiO2, PSF/ dispersion and accessibility of acid sites, and enhance their activity for
SiO2 and PSF.
the acetalization of glycerol.
Table 4 summarized the catalytic performance of HZSM-5, HBEA,
to find that the thermal stability of PSF resin was enhanced when it was Amberlyst-45, H3PW12O40, PSF/K-SiO2 in the acetalization of glycerol
immobilized on K-SiO2, as the exothermic peak (of combustion of bulk with acetone at 25 ◦ C. It was found that HZSM-5 was less active for this
resin) shifted obviously from 463.9 to 505.9 ◦ C. Furthermore, it was reaction at low temperature as the detected conversion of glycerol and
observed that the decomposition of PSF resin in PSF/K-SiO2 was slower the selectivity of solketal were the lowest. HBEA, Amberlyst-45 and
than that of bulk PSF resin and PSF/SiO2 (see Fig. S1b). And the H3PW12O40 were active and the detected conversion of glycerol
calculated PSF and S content as well as acidity of these catalysts were increased to 70.9, 80.6 and 84.5%, respectively. It is quite interesting to
summarized in Table 1. find that the conversion of glycerol and the selectivity of solketal over
N2 adsorption-desorption isotherms of bare SiO2, K-SiO2, PSF/K- PSF/K-SiO2 reached 86.3% and 97.7%. Compared with those published
SiO2, PSF/SiO2 and PSF were compared in Fig. 4. It was found that all sulfated carbon catalysts (see Table 4), these results also demonstrated
samples showed type III isotherms without obvious hysteresis loop at that PSF/K-SiO2 is an excellent solid acid catalyst for its higher activity
low p/po. The BET surface area of bare SiO2 that calculated on the basis at lower temperature (25 ◦ C).
of adsorption data was 234.6 m2/g, and it decreased to 117.2 m2/g when Fig. 7 displayed the performance of recycled PSF/K-SiO2, PSF and
SiO2 was modified with KH560. The reduced surface of K-SiO2 might be PSF/SiO2. It was found that bulk PSF resin and PSF/SiO2 deactivated
attributed to that KH560 was successfully grafted on the surface of SiO2 quickly as the conversion of glycerol in the 5th cycle declined sharply to
and it made some pore channels blocked (as the detected pore volume 29.5 and 30.5% (see Fig. 7b and 7c), respectively. On the other hand, the
also decreased obviously from 1.378 to 0.579 cm3/g). When PSF resin conversion of glycerol over PSF/K-SiO2 remained stable in the consec­
was further immobilized on K-SiO2, the calculated surface area and pore utive five recycles. FTIR, N2 adsorption-desorption and XPS analysises of
volume of PSF/K-SiO2 decreased slightly to 77.8 m2/g and 0.373 cm3/g, spent PSF/K-SiO2 were compared with the fresh one in Fig. S4. It can be
respectively, while the surface area of bare SiO2 supported PSF resin found that the detected functional groups (Fig. S4a), N2 adsorption-
(PSF/SiO2) and separated PSF resin were quite low (see Table 1). These desorption and XPS curves (Fig. S4b and c) of fresh and spent catalyst
results further confirmed that the modification of SiO2 with KH560 were similar, confirming that the structure of PSF/K-SiO2 retained well
strengthened the interaction between PSF resin and K-SiO2 (see Fig. 2), after five recycles. But the surface content of S decreased slightly from
improved the dispersion of PSF resin and the surface area of K-SiO2 was 1.4 (in fresh sample) to 1.2% (in spent sample) (see Table S1), which
preserved (compared with that of PSF/SiO2). might be attributed to the vigorous stirring during the reaction. The
Fig. 5 showed the SEM images of PSF/K-SiO2 catalyst. It was found excellent stability of PSF/K-SiO2 could be attributed to the strong
that bare SiO2 particles were stacked up of numerous grains (Fig. S3a). interaction between PSF resin and K-SiO2 (see Figs. 1 & 2) and enhanced
The outline of KH560 modified SiO2 particles looks like a shaggy thermal stability (see Fig. 3).

4
R. Zhou et al. Fuel 291 (2021) 120207

Fig. 5. SEM images of PSF/K-SiO2 (a1, a2); line-EDS of PSF/K-SiO2 (a3, a4).

Fig. 6. XPS spectra of PSF/SiO2 and PSF/K-SiO2 (a) and binding energy of S 2p in PSF/SiO2 and PSF/K-SiO2 (b).

5
R. Zhou et al. Fuel 291 (2021) 120207

Table 2
Bulk and surface element composition of PSF/SiO2 and PSF/K-SiO2.
Sample Bulk composition (%)a Surface composition (%)b

C Si S O C Si S O

PSF/SiO2 28.0 17.3 4.0 50.7 30.6 17.5 0.9 51.0

PSF/K-SiO2 33.7 14.3 4.5 47.5 48.6 11.2 1.4 38.8
a
Concluded from TG results, see Table 1.
b
Detected by XPS.

Table 3
Acetalization of glycerol with acetone over different PSF catalysts.a
Catalyst Conversion (%) Selectivity (%) STY (g-solketal/g-cat/h)b

Solketal Acetal

PSF 75.0 98.0 2.0 14.1

PSF/SiO2 86.6 98.1 1.9 16.3
PSF/K-SiO2 86.3 97.7 2.3 16.1
K-SiO2 <1 – – 0
SiO2 <1 – – 0
a
Reaction conditions: 10.86 mmol glycerol, 108.6 mmol acetone, 0.05 g
catalyst, 25 ◦ C, 1.5 h.
b
STY, defined as (gram of solketal formed)/(gram of catalyst)/(reaction
time).

Table 4
Acetalization of glycerol with acetone over acid catalysts.a
Catalyst Conversion Selectivity (%) STY (g- Ref.
(%) solketal/g-
Solketal Acetal
cat/h)b

HZSM-5c 7.2 65.4 34.6 0.9 This

work
HBEA 70.9 97.5 2.5 13.2 This
work
Amberlyst-45 80.6 97.4 2.6 15.0 This
work
H3PW12O40 84.5 98.6 1.4 15.9 This
work
PSF/K-SiO2 86.3 97.7 2.3 16.1 This
work
d
C-SO3H 94.0 99.5 0.5 26.8 [8]
Acid carbon 80.0 95.0 5.0 9.1 [26]
catalyst e
Hollow sphere 79.1 99.2 0.8 3.5 [27]
carbon-SO3H f
a
Reaction conditions: 10.86 mmol glycerol, 108.6 mmol acetone, 0.05 g
catalyst, 25 ◦ C, 1.5 h.
b
STY, defined as (gram of solketal formed)/(gram of catalyst)/(reaction
time).
c
Si/Al = 25, and pretreated at 500 ◦ C for 4 h.
d
Reaction conditions:10.86 mmol glycerol, 65.2 mmol acetone, 0.05 g cata­
lyst, 40 ◦ C, 1 h.
e
Reaction conditions:10.1 mmol glycerol, 39.6 mmol acetone, 0.028 g cata­
lyst, 25 ◦ C, 4 h.
f
Reaction conditions:10 mmol glycerol, 10 mmol acetone, 0.05 g catalyst,
80 ◦ C, 6 h.

To further demonstrate the stability of PSF/K-SiO2, a leaching test

Fig. 7. The reusability of PSF/K-SiO2 (a), PSF (b) and PSF/SiO2 (c).
was carried out under the optimal reaction condition and the results
were shown in Fig. 8. In brief, PSF/K-SiO2 was removed from the re­
action mixture by centrifugation after 10 min (blue line), then the su­ under varied reaction temperature for 20 min was illustrated in Fig. 9. It
pernatant was kept reacting for another 80 min. It was found that the was found that the conversion of glycerol increased rapidly from 46.4 to
conversion of glycerol and the selectivity of solketal did not increase 60.2% and the selectivity of solketal also increased from 85.0 to 88.6%
anymore after removing PSF/K-SiO2 (black line). These results with the increasing temperature from 15 to 25 ◦ C. Then the conversion
confirmed that PSF/K-SiO2 was stable in the reaction mixture and the of glycerol and the selectivity of solketal increased slightly to 74.7% and
leaching of acid sites was negligible. 92.1% when further raising the temperature to 55 ◦ C, respectively.
The performance of PSF/K-SiO2 for the acetalization of glycerol These results indicated that this reaction over PSF/K-SiO2 catalyst could

6
R. Zhou et al. Fuel 291 (2021) 120207

Fig. 8. The leaching test of PSF/K-SiO2. Fig. 11. Acetalization of glycerol with acetone over different amounts
of catalyst.

Fig. 9. Acetalization of glycerol with acetone under varied temperature.

Fig. 12. Acetalization of glycerol with acetone at varied reaction time on PSF/
K-SiO2.

and glycerol in feed on the performance of PSF/K-SiO2. It was found that

the conversion of glycerol increased quickly from 58.7 to 86.3% with the
increasing molar ratio from 2:1 to 10:1, and then the conversion
increased slightly. The selectivity of solketal remained stable in these
experiments. These results indicated that excess amount of acetone was
favorable for the conversion of glycerol [17,23–25].
At the same time, it was found that the acetalization of glycerol with
acetone could be activated even with trace amount of PSF/K-SiO2 as the
detected conversion of glycerol reached 25.0% with a 77.8% selectivity
of solketal (the catalyst dosage was 0.2 wt% of glycerol) (see Fig. 11),
and the calculated space time yield reached 93.1 g-solketal/g-cat/h. The
conversion of glycerol increased quickly to 80.3% with the increasing
amount of catalyst at first (below 3.0 wt% of glycerol), and then
increased slightly to around 86%. The selectivity of solketal also
increased from 77.8 to 97.7%, and remained stable even in the reaction
mixture with 10.0 wt% dosage of catalyst.
Fig. 12 demonstrated the time course of acetalization of glycerol with
acetone over PSF/K-SiO2 at 25 ◦ C. It was confirmed that the reaction
Fig. 10. Acetalization of glycerol with varied molar ratio of acetone/glycerol.
initiated quickly as the conversion of glycerol reached 45.1% within
0.17 h, and then increased continuously to 86.3% at 1.5 h. Meanwhile,
be activated in a short time even at low temperature (15 ◦ C) and reaction the selectivity of solketal also rose from 83.4% to 97.7%. After 1.5 h, the
temperature plays a vital role [28,57]. detected conversion of glycerol and the selectivity of solketal changed
Fig. 10 presented the influence of the molar ratio between acetone

7
R. Zhou et al. Fuel 291 (2021) 120207

Table 5
Acetalization of glycerol with different ketones/aldehydes over PSF/K-SiO2.a
Entry Ketones or Aldehydes Temperature (◦ C) Conversion (%) Selectivity (%)

5-Member ring 6-Member ring

1 25 84.4 99.4 0.6

2 25 93.9 89.2 10.8

3 40 85.1 84.1 15.3

4 40 62.1 94.3 5.7

a
Reaction conditions: 10.86 mmol glycerol, 108.6 mmol ketones/aldehydes, 0.05 g catalyst, 1.5 h.

slightly with the prolonged time because of the thermodynamic Appendix A. Supplementary data
equilibrium.
What’s more, it was found that PSF/K-SiO2 was also active for the (1) Thermal analysis curves of K-SiO2 and PSF/SiO2, (2) Detailed
acetalization of glycerol with butanone and cyclohexanone at 25 ◦ C (see calculation of TG results, (3) SEM images of SiO2, K-SiO2 and PSF/SiO2,
Table 5), and the conversion of glycerol reached to 84.4 and 93.9%, (4) FTIR spectra, N2 adsorption-desorption isotherms and XPS spectra of
respectively. Besides, the catalytic performance of PSF/K-SiO2 was also fresh and recycled PSF/K-SiO2, (5) Surface element composition of fresh
tested in the acetalization of glycerol with butyraldehyde and hex­ and recycled PSF/K-SiO2. Supplementary data to this article can be
aldehyde at higher temperature (40 ◦ C), and the detected conversion of found online at https://doi.org/10.1016/j.fuel.2021.120207.
glycerol was 85.1% and 62.1%, respectively. In all experiments, the
selectivity of 5-member ring product was higher than 84%. These results References
suggested the PSF/K-SiO2 is versatile for the acetalization of glycerol.
[1] Serrano-Ruiz JC, Luque R, Sepúlveda-Escribano A. Transformations of biomass-
derived platform molecules: from high added-value chemicals to fuels via aqueous-
4. Conclusions phase processing. Chem Soc Rev 2011;40:5266–81.
[2] Ruppert AM, Weinberg K, Palkovits R. Hydrogenolysis goes bio: from
carbohydrates and sugar alcohols to platform chemicals. Angew Chem Int Ed 2012;
In summary, it was confirmed that a stable solid acid catalyst with
51:2564–601.
high surface area (PSF/K-SiO2) could be synthesized in a facile method. [3] Brock D, Koder A, Rabl H-P, Touraud D, Kunz W. Optimising the biodiesel
Characterizations results disclosed that the modification of SiO2 with production process: implementation of glycerol derivatives into biofuel
formulations and their potential to form hydrofuels. Fuel 2020;264:116695.
KH560 could strengthen the interaction between PSF resin and K-SiO2
[4] Zhang X, Cui G, Feng H, Chen L, Wang H, Wang B, et al. Platinum-copper single
support, enhance its thermal stability. As a result, PSF/K-SiO2 exhibited atom alloy catalysts with high performance towards glycerol hydrogenolysis. Nat
prominent activity for the acetalization of glycerol with acetone at room Commun 2019;10.
temperature, and the best yield of solketal reached 93.1 g-solketal/g- [5] Okoye PU, Hameed BH. Review on recent progress in catalytic carboxylation and
acetylation of glycerol as a byproduct of biodiesel production. Renew Sustain
cat/h. It was also found that trace amount of PSF/K-SiO2 can activate the Energy Rev 2016;53:558–74.
reaction without any induction period, and the catalyst was versatile for [6] Monteiro MR, Kugelmeier CL, Pinheiro RS, Batalha MO, da Silva César A. Glycerol
the reaction between glycerol and different ketones or aldehydes. These from biodiesel production: technological paths for sustainability. Renew Sustain
Energy Rev 2018;88:109–22.
results could provide a new strategy for the design, synthesis and [7] Miao Z, Li Z, Liang M, Meng J, Zhao Y, Xu L, et al. Ordered mesoporous titanium
application of solid acid catalysts, and also accelerate the research in phosphate material: a highly efficient, robust and reusable solid acid catalyst for
catalytic conversion of glycerol. acetalization of glycerol. Chem Eng J 2020;381:122594.
[8] Konwar LJ, Samikannu A, Mäki-Arvela P, Boström D, Mikkola J-P. Lignosulfonate-
based macro/mesoporous solid protonic acids for acetalization of glycerol to bio-
CRediT authorship contribution statement additives. Appl Catal B Environ 2018;220:314–23.
[9] Yun D, Yun YS, Kim TY, Park H, Lee JM, Han JW, et al. Mechanistic study of
glycerol dehydration on Brønsted acidic amorphous aluminosilicate. J Catal 2016;
Ruru Zhou: Data curation, Formal analysis, Investigation, Writing - 341:33–43.
original draft, Writing - review & editing. Yuanyuan Jiang: Investiga­ [10] Ali B, Lan X, Arslan MT, Gilani SZA, Wang H, Wang T. Controlling the selectivity
and deactivation of H-ZSM-5 by tuning b-axis channel length for glycerol
tion, Formal analysis, Writing - original draft, Writing - review & edit­
dehydration to acrolein. J Ind Eng Chem 2020;88:127–36.
ing. Huaiyuan Zhao: Investigation, Formal analysis. Boyong Ye: [11] Jiang Y, Li X, Zhao H, Hou Z. Esterification of glycerol with acetic acid over SO3H-
Investigation, Formal analysis. Lina Wang: Methodology, Formal functionalized phenolic resin. Fuel 2019;255:115842.
analysis. Zhaoyin Hou: Conceptualization, Methodology, Resources, [12] An S, Sun Y, Song D, Zhang Q, Guo Y, Shang Q. Arenesulfonic acid-functionalized
alkyl-bridged organosilica hollow nanospheres for selective esterification of
Supervision, Writing - review & editing. glycerol with lauric acid to glycerol mono- and dilaurate. J Catal 2016;342:40–54.
[13] Saxena SK, Al-Muhtaseb AH, Viswanadham N. Enhanced production of high octane
oxygenates from glycerol etherification using the desilicated BEA zeolite. Fuel
2015;159:837–44.
Declaration of Competing Interest [14] Cannilla C, Bonura G, Maisano S, Frusteri L, Migliori M, Giordano G, et al. Zeolite-
assisted etherification of glycerol with butanol for biodiesel oxygenated additives
The authors declare that they have no known competing financial production. J Energy Chem 2020;48:136–44.
[15] Nda-Umar U, Ramli I, Taufiq-Yap Y, Muhamad E. An overview of recent research in
interests or personal relationships that could have appeared to influence the conversion of glycerol into biofuels, fuel additives and other bio-based
the work reported in this paper. chemicals. Catalysts 2018;9:15.
[16] Trifoi AR, Agachi PŞ, Pap T. Glycerol acetals and ketals as possible diesel additives.
A review of their synthesis protocols. Renew Sustain Energy Rev 2016;62:804–14.
Acknowledgements [17] da Silva MJ, de Ávila Rodrigues F, Júlio AA. SnF2-catalyzed glycerol ketalization: a
friendly environmentally process to synthesize solketal at room temperature over
This work was financially supported by the National Natural Science on solid and reusable Lewis acid. Chem Eng J 2017;307:828–35.
[18] Nanda MR, Zhang Y, Yuan Z, Qin W, Ghaziaskar HS, Xu C. Catalytic conversion of
Foundation of China (Contract Nos. 21773206, 21473155) and Natural glycerol for sustainable production of solketal as a fuel additive: a review. Renew
Science Foundation of Zhejiang Province (Contract No. LZ12B03001). Sustain Energy Rev 2016;56:1022–31.

8
R. Zhou et al. Fuel 291 (2021) 120207

[19] Bivona LA, Vivian A, Fusaro L, Fiorilli S, Aprile C. Design and catalytic applications [39] Kothe V, Melfi DT, dos Santos KC, Corazza ML, Ramos LP. Thermodynamic
of 1D tubular nanostructures: improving efficiency in glycerol conversion. Appl analysis, experimental and kinetic modeling of levulinic acid esterification with
Catal B Environ 2019;247:182–90. ethanol at supercritical conditions. Fuel 2020;260:116376.
[20] Manjunathan P, Maradur SP, Halgeri AB, Shanbhag GV. Room temperature [40] Gérardy R, Debecker DP, Estager J, Luis P, Monbaliu JM. Continuous flow
synthesis of solketal from acetalization of glycerol with acetone: effect of crystallite upgrading of selected C2–C6 platform chemicals derived from biomass. Chem Rev
size and the role of acidity of beta zeolite. J Mol Catal A Chem 2015;396:47–54. 2020.
[21] Rahaman MS, Phung TK, Hossain MA, Chowdhury E, Tulaphol S, Lalvani SB, et al. [41] Tanabe K, Hölderich WF. Industrial application of solid acid–base catalysts. Appl
Hydrophobic functionalization of HY zeolites for efficient conversion of glycerol to Catal A Gen 1999;181:399–434.
solketal. Appl Catal A Gen 2020;592:117369. [42] Lee AF, Bennett JA, Manayil JC, Wilson K. Heterogeneous catalysis for sustainable
[22] Li X, Zheng L, Hou Z. Acetalization of glycerol with acetone over Co[II](Co biodiesel production via esterification and transesterification. Chem Soc Rev 2014;
[III]xAl2− x)O4 derived from layered double hydroxide. Fuel 2018;233:565–71. 43:7887–916.
[23] Nair GS, Adrijanto E, Alsalme A, Kozhevnikov IV, Cooke DJ, Brown DR, et al. [43] Chen SS, Maneerung T, Tsang DCW, Ok YS, Wang C-H. Valorization of biomass to
Glycerol utilization: solvent-free acetalisation over niobia catalysts. Catal Sci hydroxymethylfurfural, levulinic acid, and fatty acid methyl ester by
Technol 2012;2:1173. heterogeneous catalysts. Chem Eng J 2017;328:246–73.
[24] Li X, Jiang Y, Zhou R, Hou Z. Layered α-zirconium phosphate: an efficient catalyst [44] Zhang Y, Riduan SN. Functional porous organic polymers for heterogeneous
for the synthesis of solketal from glycerol. Appl Clay Sci 2019;174:120–6. catalysis. Chem Soc Rev 2012;41:2083–94.
[25] Li X, Jiang Y, Zhou R, Hou Z. Acetalization of glycerol with acetone over [45] Liu F, Huang K, Zheng A, Xiao F-S, Dai S. Hydrophobic solid acids and their
appropriately-hydrophobic zirconium organophosphonates. Appl Clay Sci 2020; catalytic applications in green and sustainable chemistry. ACS Catal 2018;8:
189:105555. 372–91.
[26] Gonçalves M, Rodrigues R, Galhardo TS, Carvalho WA. Highly selective [46] Han L, Ren W, Wang B, He X, Ma L, Huo Q, et al. Extraction of SiO2 and Al2O3 from
acetalization of glycerol with acetone to solketal over acidic carbon-based catalysts coal gangue activated by supercritical water. Fuel 2019;253:1184–92.
from biodiesel waste. Fuel 2016;181:46–54. [47] Hosseini M-S, Masteri-Farahani M. Surface functionalization of magnetite
[27] Wang L, Zhang J, Yang S, Sun Q, Zhu L, Wu Q, et al. Sulfonated hollow sphere nanoparticles with sulfonic acid and heteropoly acid: efficient magnetically
carbon as an efficient catalyst for acetalisation of glycerol. J Mater Chem A 2013;1: recoverable solid acid catalysts. Chem Asian J 2019;14:1076–83.
9422–6. [48] Xia Y, He Y, Chen C, Wu Y, Zhong F, Chen J. Co-modification of polydopamine and
[28] Esteban J, Ladero M, García-Ochoa F. Kinetic modelling of the solventless synthesis KH560 on g-C3N4 nanosheets for enhancing the corrosion protection property of
of solketal with a sulphonic ion exchange resin. Chem Eng J 2015;269:194–202. waterborne epoxy coating. React Funct Polym 2020;146:104405.
[29] Okuhara T. Water-tolerant solid acid catalysts. Chem Rev 2002;102:3641–65. [49] Laskar IB, Rajkumari K, Gupta R, Rokhum L. Acid-functionalized mesoporous
[30] Tondro H, Zilouei H, Zargoosh K, Bazarganipour M. Nettle leaves-based sulfonated polymer-catalyzed acetalization of glycerol to solketal, a potential fuel additive
graphene oxide for efficient hydrolysis of microcrystalline cellulose. Fuel 2021; under solvent-free conditions. Energy Fuel 2018;32:12567–76.
284:118975. [50] Liu F, Meng X, Zhang Y, Ren L, Nawaz F, Xiao F-S. Efficient and stable solid acid
[31] Questell-Santiago YM, Galkin MV, Barta K, Luterbacher JS. Stabilization strategies catalysts synthesized from sulfonation of swelling mesoporous
in biomass depolymerization using chemical functionalization. Nat Rev Chem polydivinylbenzenes. J Catal 2010;271:52–8.
2020;4:311–30. [51] Kundu SK, Bhaumik A. Pyrene-based porous organic polymers as efficient catalytic
[32] Questell-Santiago YM, Zambrano-Varela R, Talebi Amiri M, Luterbacher JS. support for the synthesis of biodiesels at room temperature. ACS Sustain Chem Eng
Carbohydrate stabilization extends the kinetic limits of chemical polysaccharide 2015;3:1715–23.
depolymerization. Nat Chem 2018;10:1222–8. [52] Hsissou R, Berradi M, El Bouchti M, El Bachiri A, El Harfi A. Synthesis
[33] Morales-Leal FJ, Rivera de la Rosa J, Lucio-Ortiz CJ, De Haro-Del Rio DA, Solis characterization rheological and morphological study of a new epoxy resin
Maldonado C, Wi S, et al. Dehydration of fructose over thiol- and sulfonic-modified pentaglycidyl ether pentaphenoxy of phosphorus and their composite (PGEPPP/
alumina in a continuous reactor for 5-HMF production: study of catalyst stability MDA/PN). Polym Bull 2019;76:4859–78.
by NMR. Appl Catal B Environ 2019;244:250–61. [53] Wu SQ, Wang JW, Shao J, Wei L, Yang K, Ren H. Building a novel chemically
[34] Johnson RL, Perras FA, Hanrahan MP, Mellmer M, Garrison TF, Kobayashi T, et al. modified polyaniline/thermally reduced graphene oxide hybrid through pi-pi
Condensed phase deactivation of solid Brønsted acids in the dehydration of interaction for fabricating acrylic resin elastomer-based composites with enhanced
fructose to hydroxymethylfurfural. ACS Catal 2019;9:11568–78. dielectric property. ACS Appl Mater Interfaces 2017;9:28887–901.
[35] Mellmer MA, Sanpitakseree C, Demir B, Ma K, Elliott WA, Bai P, et al. Effects of [54] Zhang J, Dong K, Luo W, Guan H. Catalytic upgrading of carbohydrates into 5-
chloride ions in acid-catalyzed biomass dehydration reactions in polar aprotic ethoxymethylfurfural using SO3H functionalized hyper-cross-linked polymer based
solvents. Nat Commun 2019;10:1132. carbonaceous materials. Fuel 2018;234:664–73.
[36] Sharninghausen LS, Campos J, Manas MG, Crabtree RH. Efficient selective and [55] Li S, Chen F, Han Y, Zhou H, Li H, Zhao T. Enhanced compatibility and morphology
atom economic catalytic conversion of glycerol to lactic acid. Nat Commun 2014;5: evolution of the hybrids involving phenolic resin and silicone intermediate. Mater
5084. Chem Phys 2015;165:25–33.
[37] Torres-Pacheco LJ, Osornio-Villa A, García-Gómez NA, Olivas A, Valdez R, Guerra- [56] Zhao M, Shi J, Hou Z. Selective hydrogenation of phenol to cyclohexanone in water
Balcázar M, et al. Effect of AuM (M: Ag, Pt & Pd) bimetallic nanoparticles on the over Pd catalysts supported on Amberlyst-45. Chin J Catal 2016;37:234–9.
sorbitol electro-oxidation in alkaline medium. Fuel 2020;274:117864. [57] Vannucci JA, Nichio NN, Pompeo F. Solketal synthesis from ketalization of glycerol
[38] Lee SY, Kim HU, Chae TU, Cho JS, Kim JW, Shin JH, et al. A comprehensive with acetone: a kinetic study over a sulfated zirconia catalyst. Catal Today 2020.
metabolic map for production of bio-based chemicals. Nat Catal 2019;2:18–33.

9
 Fuel 299 (2021) 120923

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Solvent-free solketal production from glycerol promoted by yeast

activated carbons
Raphael Rodrigues a, *, Mikaela S. Santos b, Renan S. Nunes c, Wagner A. Carvalho c,
Georgia Labuto b, d
a
Faculty of Exact Sciences and Technology, Federal University of Grande Dourados, Dourados, MS 79804-970, Brazil
b
LabInSciences – Laboratory of Integrated Sciences, Universidade Federal de São Paulo, Faculty of Exact Sciences and Technology, 09913-030 SP, Brazil
c
Center for Natural Sciences and Humanities, Federal University of ABC (UFABC), Santo André, SP 09210-580, Brazil
d
Departamento de Química, Universidade Federal de São Paulo, 09913-030 Diadema, SP, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: Chemically activated carbons from yeast biomass waste were produced and later modified by treatment with
Activated-carbon HNO3 aiming at tuning their surface acid features. The obtained materials were successfully applied as catalyst in
Yeast a solvent-free solketal production process and the modified series of activated carbons displayed above average
Solketal
catalytic performance reaching glycerol conversion of up to 91% (TOF of 215 h− 1) with 97% of selectivity toward
Glycerol
Acetalization
solketal, as well as remarkable efficiency in consecutive catalytic runs. The physicochemical and surface
chemistry properties of the activated carbons were characterized by means of N2 adsorption/desorption iso­
therms, thermal analysis, elemental analysis, surface functional groups titration (Boehm titration), X-rays
diffraction and Raman spectroscopy.

1. Introduction as a by-product, which can represent an economic drawback to the

process viability. Crude glycerol has several impurities, such as meth­
Economic growth and industrial development have been continu­ anol, soap, triglycerides, fatty acids, salts and metals, which may affect
ously contributing to a sharp increase of natural resources consumption its application. However, several techniques for purifying crude glycerol
and anthropogenic greenhouses gases (GHG) concentration in atmo­ have been used with cost feasibility and reduced environmental effects
sphere. Despite global attempts to mitigate GHGs emissions, the world [5,6]. An increasing trend in the production of this fuel led to an over­
has been witnessing a significant rise in annual CO2 emissions, which is supply and consequently to a drop in quotations of glycerol [7].
projected to reach 72 billion tons by 2050 if appropriate climate change The chemical transformation of glycerol into value-added products
mitigation measures are not put in place [1,2]. In this sense, the search has attracted attention of the scientific community and several viable
for alternative sources of energy and sustainable processes aiming at synthetic routes have been considered, giving rise to a wide range of
reducing environmental pollution and global warming has stimulated versatile organic products [8–10]. The acetalization of glycerol with
the global market for clean fuels, among which biodiesel has recently aldehydes or ketones can produce two branched oxygen-containing
gained prominence being considered a promising substitute for compounds, [1,3]dioxan-5-ols (six-membered ring) and mainly [1,3]
petroleum-based diesel due to its intrinsic advantages including dioxolan-4-yl-methanols (five-membered ring), well known as solketal,
renewability, lower toxicity, biodegradability, biocompatibility and which possess a broad diversity of direct commercial application such as
lower emission profile [3]. solvent, low-temperature heat-transfer fluid, surfactant and fuel addic­
First-generation of biodiesel, which is chemically described as a tive [11,12].
mixture of fatty acid esters typically obtained by transesterification re­ Conventionally, condensation of glycerol with carbonyl compounds
action of vegetable oil with a short chain alcohol (ethanol or methanol), is carried out in a homogeneous Brønsted acid catalytic system via a
has been massively predominant worldwide [4]. It is estimated that the typical acetal formation mechanism [13]. However, economic issues
production of 100 m3 of biodiesel generates ca. 10 m3 of crude glycerol and environmental disadvantages of this process have stimulated studies

* Corresponding author.
E-mail address: [email protected] (R. Rodrigues).

https://doi.org/10.1016/j.fuel.2021.120923
Received 14 October 2020; Received in revised form 26 November 2020; Accepted 20 April 2021
Available online 30 April 2021
0016-2361/© 2021 Elsevier Ltd. All rights reserved.
R. Rodrigues et al. Fuel 299 (2021) 120923

toward the development of a novel solid material with suitable acidic impregnation as AC1 (1:0.5), AC2(1:0.8), AC3(1:1.1) and AC4(1:1.4).
characteristics for playing a major role as heterogeneous catalyst in
these reactions. In this sense, several approaches based on heteroge­
neous catalysis technologies have been successfully proposed, showing 2.2. Surface treatment of ACs
viability of a wide spectrum of catalysts, such as metal–organic frame­
works, heteropolyacids, mixed oxides, zeolites, carbon-based materials The surface treatment of the activated carbons was carried out by
among others [14–18]. adding 10 g of each AC to 100 mL of an HNO3 wt. 65% solution under
Activated carbons (ACs) are a broadly versatile class of materials magnetic stirring, at temperature between 343 and 353 K for 3 h. Af­
with remarkable surface chemistry nature and noticeable porous terwards, the material was washed with distilled water in a Soxhlet
framework which have been proving efficiency in a wide variety of extractor in order to remove the remaining physisorbed oxidant agent,
processes either as adsorbent material or as heterogeneous catalyst. and then dried at 373 K for 15 h. The modified ACs were denoted as
Traditionally, carbonaceous precursors are used in the ACs production ACN1, ACN2, ACN3 and ACN4, respectively.
protocols due to their carbon-rich structures and such characteristic can
be found in organic-based solids such as coal, glycerol, PET, lignocel­ 2.3. Characterizations
lulosic materials among others [19–21]. Agricultural activities generate
large amounts of waste with a high economic potential for application, Textural properties of the parent and modified ACs were determined
typically related to their chemical and structural composition. In this by N2 adsorption/desorption measurements in a NOVA 1200e device
sense, the scientific community has been driven toward the develop­ (Quantachrome Instruments). The samples were pre-treated at 523 K
ment/improvement of ACs production protocols using the above- under vacuum for 3 h and the analysis were carried out at 78 K. The
mentioned sources aiming at an extensive range of commercial possi­ surface area, total pore volume, micropore volume and pore size dis­
bilities [22–24]. tribution were calculated by means of the Brunauer–Emmett–Teller
Brazil is the largest sugarcane producer worldwide with a cultivated (BET) equation [32], Dubinin–Radushkevich (DR) equation [33] and
area of roughly 9 million hectares and the increasing production trend density functional theory (DFT) [34]. Powder X-ray diffraction patterns
has been stimulated by local policies focused on boost the role of (XRD) were measured on a Bruker D8 Advance diffractometer using a Cu
renewable fuels in its energy matrix, which predict an expansion in Kα radiation (1.542 Å) with a high resolution Lynxeye detector. Each
ethanol production form 28 billion liters in 2015 to around 50 billion sample was scanned in a 2θ range of 10–60◦ with step size of 0.02◦ /s.
liters by 2030 [25]. Nevertheless, ethanol production cycle leads to the Raman spectra were recorded in a Raman Renishaw microscope (Mod­
side generation of different biomass wastes with potential for alternative elInVia) with exciting radiation at 830 nm (laser diode) in a range of
destinations. The yeast biomass (YB) has been later revalued either to 900–2000 cm− 1. The number of acid sites per gram of the material was
the animal feed or as a source of β-glucan (molecule of commercial in­ evaluated using Boehm titration [35]. For this test, 0.2 g of each carbon
terest), however, after extracting this substance, yeast biomass waste sample was added to 25 mL of an aqueous basic solution (0.05 mol L-1
(YBW) has been found useless. Hence, considering both sugar and NaHCO3 to titrate carboxylic acid sites and 0.1 mol L-1 NaOH to deter­
alcohol industry scenario, in which the production of each cubic meter mine the total amount of acid sites). The suspensions were kept for 24 h
of ethanol side generates around 30 Kg of yeast biomass, and also under magnetic stirring at room temperature and filtered prior to
considering low-cost nature, logistical feasibility as well as circular titration. Aliquots containing 10 mL of the filtered solutions were
economy concepts [25–28], the fabrication of activated carbons from titrated with 0.1 mol L-1 HCl using an automatic titrator (Metrohm 905
carbon-rich YBW appears to be an appropriate destination for this Titrando). The carbon, hydrogen, nitrogen content of all materials was
agricultural waste, as first demonstrated by Modesto et. al. [29]. quantified by means of elemental analysis in a series II CHNS/O − 2400
The surface of the ACs may contain several organic groups and Analyzer (PerkinElmer) and the oxygen content was obtained by dif­
functionalities that can be tailored in order to achieve suitable proper­ ference. Thermal stability of the ACs and ACNs was investigated by
ties according to the final application. Aiming at tuning the AC acidic thermogravimetric analysis using a DTG – 60 Shimadzu under N2 at­
surface, treatment with oxidizing agent, such as nitric acid, has been mosphere (gas flow of 50 cm3min− 1) at a heating rate of 10 K min− 1 in
typically applied to upgrade the number of surface carbon oxygen the range of 303–1173 K.
contain groups with Brønsted acid properties (i.e. carboxyl groups,
carboxylic anhydrides, lactones, and phenolic hydroxyl groups)
[21,30,31].Therefore, in this work, a series of novel ACs from YBW was 2.4. Catalytic tests
prepared and modified in order to investigate the relationships of both
textural and acidic properties with their catalytic performance in the The acetalization of glycerol with acetone was carried out in 30 mL
solvent-free conversion of glycerol to solketal via acetalization reaction capped glass vials under vigorous magnetic stirring (500 rpm) at
different temperatures (room temperature, 318 K or 328 K). In a typical
2. Materials and methodologies catalytic test, 0.921 g (0.01 mol) of highly purified glycerol (99%), a
selected amount of acetone [i.e. 0.581 g (0.01 mol), 1.162 g (0.02 mol)
2.1. Preparation of ACs or 2.324 g (0.04 mol)] and 0.132 g (0.0015 mol) of 1,4-dioxane, used as
the GC internal standard, were weighed in a 30 mL glass vial containing
Activated carbons were prepared by chemical activation of 30 mg of the catalyst. The mixture was stirred for 5 h at the selected
powdered YBW, obtained after extracting of β-glucan from yeast temperature. The reaction products were analyzed by gas chromatog­
biomass, provided by Biorigin Company (São Paulo – Brazil) and pre­ raphy with flame ionization detector (GC-FID) using a PerkinElmer
viously dried by a spray dryer, which gives to it a presentation form of Clarus 600 device, equipped with a DB-WAX capillary column.
fine powder. Typically, the raw material was impregnated by a specific
amount of 85 wt% H3PO4 at 357 K for 4 h, with degree of impregnation 3. Results and discussions
(weight of YBW : weight of activation agent) ranging from 1:0.5 to 1:1.4.
The activation process was carried out in a horizontal glass reactor In the present study, a series of activated carbons from yeast biomass
under a nitrogen gas flow of 100 mL min− 1 at 723 K for 2 h. The residual waste were prepared for the first time through a chemical activation
activating agent was removed by washing the solids several time with process with H3PO4, accordingly modified and successfully applied as
hot distilled water and the obtained materials were dried at 373 K for 12 heterogeneous catalyst in the herein investigated glycerol upgrading
h. The series of ACs were labeled according to their degree of process.

2
R. Rodrigues et al. Fuel 299 (2021) 120923

Table 1
Textural properties and acidity of the parents and acid-treated activated
carbons.
SBET Vmicro Vmeso Vtotal COOH Total
(m2g− 1) (cm3g− 1) (cm3g− 1) (cm3g− 1) (mmol groups
g− 1) (mmol
g− 1)

AC1 983 0.44 0.06 0.50 1.2 1.6

AC2 1088 0.50 0.07 0.57 1.0 1.9
AC3 1122 0.50 0.18 0.68 0.7 1.2
AC4 1208 0.53 0.26 0.79 0.6 1.3
ACN1 559 0.28 0.01 0.29 2.1 4.8
ACN2 327 0.12 0.01 0.13 1.9 4.6
ACN3 501 0.25 0.01 0.26 2.4 5.3
ACN4 520 0.23 0.05 0.28 3.6 5.1

Fig. 1. N2 adsorption/desorption isotherms (a) and pore size distribution (b) of

parent activat ed carbons (AC).

3.1. Characterization of ACs and ACNs

Textural properties of the obtained activated carbons were evaluated

by means of N2 adsorption/desorption isotherms at 78 K and the results
are presented in Fig. 1(a). For instance, according to IUPAC classifica­
tion [36], AC1 and AC2 present a typical microporous adsorbent
isotherm (Type I), with a slightly slit-shaped mesoporous contribution,
as indicated by a subtle H4 hysteresis opening. Yet adsorption isotherms
of AC3 and AC4 can be classified as a combination of Type I and Type IV,
presenting a sharp increase of N2 adsorption at low relative pressure,
which is related to the microporous nature of both materials, followed
by a linear rise in the multilayer relative pressure range (P/P0 = 0,3 –
0,7) up to its capillary condensation, visualized at P/P0 > 0,7, empha­
sizing a remarkable presence of mesoporous. It also must be highlighted
the existence of a conspicuous H2 hysteresis loop, which indicates
Fig. 2. N2 adsorption/desorption isotherms (a) and pore size distribution (b) of
variability of the mesopore widths in AC3 and AC4 porous structure the modifie d activated carbons (ACN).
[36]. Fig. 1(b) illustrates pore size distributions for the unmodified
materials and their extensions comprehend the micropore and early
accompanied by further side chemical reactions such as degradation,
mesopore range, corroborating the information presented by the
dehydration and condensation [37,38]. It is also noticeable physical and
isotherms.
chemical transformations throughout impregnation and evaporation
Chemical activation process of a lignocellulosic precursor with
steps due the formation of tar on the surface of the particles. This is
phosphoric acid has already been widely detailed and begins with the
result of a depolymerization of cellulose followed by many complex
acid hydrolysis of glycosidic linkages in hemicellulose as well as the
dehydration and condensation reactions [39,40]. Lastly, as heat treat­
cleavage of aryl ether bonds into the chemical structure of lignin,
ment takes place, volatile matters and gaseous products evolve from the
reducing its mechanical resistance and leading to a swelling of the
structure of the carbonized carbon and the extraction of the activation
particles. Then, constituents are broken down into small units

3
R. Rodrigues et al. Fuel 299 (2021) 120923

agent controls the pore size structure of the final activated carbon
[37,41].
Therefore, as YBW is a cellular material with a cell wall rich in
polysaccharides such as mannan, chitin and glucans, containing roughly
43% of carbon in its composition and typically found in a fine powdered
form [26,42], it is expected that phosphoric acid plays a similar role
during the whole activating process. Thus, the activation agent likely
penetrates the YBW structure promoting acid hydrolysis of the poly­
saccharides as well as some chemical bond cleavages decreasing both
mechanical properties and molecular weight of the biopolymers. Then,
it probably induces the same effects and reaction protocols as previously
described for lignocellulosic materials. Indeed, reasonable amount of tar
surrounding the particles can be observed during the impregnation/
evaporation processes, which indicates the beginning of the precursor
conversion to carbon. Furthermore, the relationship between impreg­
nation ratio and the conforming porosity is also clear since the porous
distribution size and the isotherm profiles of the obtained materials are
highly dependent on the concentration of the activation agent corrob­
orating the results presented by Molina-Sabio et.al [41]. Surface area
and total pore volume of the ACs obviously follow the same trend and
AC4 stands out displaying specific surface area as high as 1208 m2 g− 1
and total pore volume of 0.76 cm3 g− 1, as can be seen in Table 1.
Microporous structure of the ACN series can be highlighted by the
Type 1 adsorption isotherm profiles visualized in all materials, as
depicted in Fig. 2(a). However, textural properties of the carbons are
severally modified under the action of the oxidation process with nitric
acid. Surface area and micropore volume of the ACNs meaningly
decreased whereas the mesopores are almost vanished. In fact, it is likely
that, as the oxidizing agent diffuses through the pore structure of the
ACs, oxygen containing functional groups stick at the entrance and on
the wall of micropores, preferentially where there is wide micropososity,
leading to a decrease in both surface area and pore volume (Table 1)
[43,44]. Such effect is more pronounced in ACN2, which might be
related to differences in surface chemical structure as well as textural
properties.
Furthermore, it is expected that such structural modification has Fig. 3. Thermogravimetric analysis un der nitrogen atmosphere and derivative
given rise to a complex ill-connected irregular topology consisting of thermogravimetric analysis obtained for a) the parent activated carbons (AC)
opened and restricted-access micropores linked through narrow throat- and b) the modified activated carbons (ACN).
type channels. Typically, the open micropores are rapidly filled at low
relative pressure, whereas the filling of the restricted-access pores takes and porous structure unmodified ACs, specially the presence of meso­
place as the P/P0 increases. Then, based on the thermodynamic pores, appear to be crucial to the effectiveness of the herein applied
approach, the fluid tends to remain trapped within those pores up to a oxidation process. In fact, the production of both total and stronger
very low relative pressure during the desorption process, which justifies oxygen containing groups on the surface of the AC3 and AC4 were
the unclosed hysteresis loop in all ACNs isotherms [45,46]. It is also notably higher than on the essentially microporous AC1 and AC2. As
noteworthy to mention that, aiming at reliability of the data, only the depicted in Table 1, the concentration of carboxylic acid sites on ACN1
adsorption branches of the ACNs isotherms were used to determine their and ACN2 roughly doubled compared to each patent one, whereas the
porous characteristics (Fig. 2(b)) [47]. number of such group on the surface of the micro/mesoporous ACN3
Concerning surface chemistry of ACs and ACNs, Boehm titration was and ACN4 increased by up to 6 times. Moreover, the same effect was
accordingly performed in order to characterize the population of car­ observed in all materials regarding total acid sites.
boxylic groups and total acidity, i.e. carboxyl groups, carboxylic anhy­ Thermal stability of the functional groups of the activated carbons
drides, lactones, quinone carbonyl groups, and phenolic hydroxyl were investigated by thermogravimetric analysis (TGA). Untreated ACs
groups [35]. As can be seen in Table 1, untreated activation carbons thermogram profiles showed an initial weight loss at temperatures
display a low concentration of total acid sites with a significant fraction below 383 K, assigned to vaporization of moistures, and a second ther­
of stronger acid sites. Furthermore, it can be observed that microporous mal event at 923 K, which has been associated to the evolution of CO2,
carbons (AC1 and AC2) presented slightly higher amount of both total ascribed to the decomposition of carboxyl, anhydride and lactone
and stronger acid sites compered to AC3 and AC4. It has been reported groups (Fig. 3 (a)). Likewise, these events are observed and highly
that the functional groups formed during the impregnation/activation pronounced in the ACN series, as consequence of their greater amount of
processes using phosphoric acid can be ascribed to the phosphorous- oxygen containing species, as well as a singular thermal event at
containing species along with functional groups arising from the pre­ approximately 573 K, as confirmed by derivative thermogravimetric
cursor [48]. In this sense, it suggests that the extraction of the inner curve (DTG) (Fig. 3 (b)), which has been related to the decomposition of
impregnated phosphorous species during the activation protocol of AC3 quinone and mono-oxygen species (carbonyls and phenols) into CO
and AC4, which gives rise to the mesoporous, also leads to a decrease in [49,50]. Elemental analysis corroborates the trends indicating a
the associated acidity. On the other hand, upon treatment with nitric remarkable increment of the O content upon treatment with HNO3 in all
acid, the amount of oxygen containing groups expressively increased treated AC structures, likely due to the introduction of surface oxygen-
compared to each parent material and the population of carboxylic containing groups. Regarding the nature of the surface groups formed
groups followed the same trend. It is also noticed that textural properties

4
R. Rodrigues et al. Fuel 299 (2021) 120923

Table 2 carbon atom. Moreover, introduction of a small amount of N, as

Weight loss (%), C, H, N and O mass content (%) and intensity ratios of D (1355 observed in all treated materials, takes place through a similar reaction
cm− 1) and G (1600 cm− 1) bands of the parent and acid treated activated car­ as in nitration of benzene [21,30]. Table 2 comparatively displays the
bons, as determined by TGA, elemental analysis and Raman spectroscopy. weight loss percentage of each thermal event observed in the TGA
Weight loss (%) Elemental Analysis(%) ID/IG analysis and the elemental composition of ACs and ACNs.
398 K 723 K 1173 K C H O N Structural features of both parent and modified series of ACs were
AC1 6,6 0,5 20,8 66.2 3.3 29.3 1.2 0.827 described by means of Raman spectra an XRD. Spectra profile of all
AC2 11,7 0,6 27,3 64.4 3.6 31.0 1.0 0.701 catalysts (Fig. 4 (a,b)) are dominated by D and G bands at 1355 and
AC3 7,4 0,3 17,6 70.9 3.8 24.1 1.2 0.708 1600 cm− 1, respectively. The former band can be ascribed to the
AC4 11,1 0,1 15.5 68.7 3.3 27.0 1.0 0.724
ACN1 11,2 14.6 20.7 56.1 2.4 37.6 3.9 0.856
breathing vibration mode of A2g symmetry that is related to structural
ACN2 8.7 16.4 38.6 56.9 2.8 36.6 3.7 0.875 disorder and its intensity is associated to sp2 carbon in six-fold aromatic
ACN3 11.5 17.3 29.7 56.8 2.6 36.9 3.7 0.843 ring in the carbon-based framework. The G mode owns E2g symmetry
ACN4 7.4 15.7 54.6 56.3 2.7 37.6 3.4 0.813 and is related to in-plane bond-stretching motion of C = C bonds, which
occurs in all hybridized sp2 sites present in graphitic-like materials
[20,52]. Moreover, intensity ratios of D and G peaks ID/IG (Table 2)
thought the oxidation process (i.e. carboxyl groups, carboxylic anhy­
characterize the level of disorder and their increment upon acid treat­
drides, lactones, and phenolic hydroxyl groups), such increment is
ment, compared to each unmodified AC, is associated to the increase of
directed related to the improvement of the acidity features of the ACNs
defect density in ACNs, likely due to the drastic nature of the oxidation
series. The most accepted mechanism for the herein applied oxidation
process [53,54]. Furthermore, it is observed a nonlinear behavior within
protocol is analogous to the oxidation reaction of 9,10-dihydrophenan­
the AC series, which appears to be associated to the variations in con­
threne and diphenylmethane with nitric acid [21,51]. Ketones are ob­
centration of activating agent. Indeed, the presence of remaining phos­
tained by the oxidation of the methylene groups and dicarboxylic groups
phorus in the activated carbons after the activation process promotes
are likely formed on the aliphatic side of the chain with more than one

Fig. 4. Raman spectra (a, b) and XRD patterns (c, d) obtained for activated carbons (AC) and modified activated carbons (ACN).

5
R. Rodrigues et al. Fuel 299 (2021) 120923

Table 3 with glycerol to acetone molar ratio of 1:1. The results were compared to
Acetalization of acetone with glycerol over ACs and ACNs at room temperature the modified series ACNs at the same reaction set up. As can be seen in
(glycerol:acetone molar ratio of 1:1, 3 wt% of catalyst, 5 h). Table 3, ACs show limited catalytic performance reaching maximum
Glycerol Selectivity to Selectivity to TOFa TOFb glycerol conversion of 18% for AC1, on the other hand, a remarkable
Conversion (%) Solketal (%) 6MR (%) h− 1 h− 1 increase of both subtracted consumption (around 50%) and solketal
AC1 18 86 14 5.5 8.4 selectivity (up to 97%) is noticed for all modified catalysts after 5 h of
AC2 16 82 18 4.6 8.9 reaction. This behavior corroborates to the acidic profile of the activated
AC3 11 85 15 4.2 7.9 carbons, despite of textual proprieties or porous structure.
AC4 8 80 20 3.7 7.4
Turnover frequency was also calculated considering both total and
ACN1 49 96 4 5.6 12.8
ACN2 47 97 3 6.4 15.6 carboxylic acid sites in order to correlate the role of the surface groups to
ACN3 52 97 3 14.4 30.2 the catalytic performance of the materials. Regarding total acidity, it is
ACN4 52 97 3 21.2 31.8 observed a slightly/negligible enhancement in catalytic efficiency
a – moles of glycerol converted per mole of total acid sites of the catalyst per comparing AC1 and AC2 with ACN1 and ACN2, whereas such trend is
hour; differently pronounced for AC3 (4.2 h− 1) and AC4 (3.7 h− 1) in which the
b – moles of glycerol converted per mole of stronger acid sites of the catalyst per surface acid treatment leads to a remarkable increase in TOF of their
hour. modified peers ACN3(14.4 h− 1) and ACN4 (21.2 h− 1). On the other
hand, when considering only carboxylic acid sites as the reactive center,
graphitic microcrystalline domains to form graphite [55], which must be it is noticed an indisputable increment in TOF for all ACN series
related to the lower concentration of sp2 graphitic-like carbon in AC1 (compared to each parent AC) and in higher magnitude for the micro/
compared to AC2, AC3 and AC4. On the hand, the increment in hy­ mesoporous ACN3 and ACN4 (Table 3). It suggests that the herein
drolysis capacity, as the concentration of activating agent increases, studied catalytic process is mainly driven by the presence of stronger
amplifies degradation process on biopolymers, which might lead to a acid sites (carboxylic groups), since fraction of such group on the surface
decrease in the graphitic nature of the activated carbons corroborating of the materials apparently determine its catalytic efficiency. Further­
the trend observed for AC2, AC3 and AC4 [56]. more, it is expected that the carboxyl and hydroxyl adjacent groups
The XRD patterns of ACs and ACNs (Fig. 4 (c, d)) show two broad (weaker acid sites) contribute to a further stabilization of the adsorbed
reflecting peaks at around 24◦ (200 plane) and 43◦ (100 and 101 planes) intermediate through hydrogen bonds. Thus, the acidity of the reactive
revealing a predominantly amorphous arrangement of the materials and center become even stronger increasing its catalytic activity [60]. This
also indicating the presence of some micro-crystallographic structure cooperative effect appears to play an import role in all modified cata­
characteristics of the activated carbons [24,38,53]. It also noteworthy to lysts, especially highlighted in ACN3 (Tables 1 and 3).
mention that both impregnation ratio and acid treatment did not pro­ Aiming at higher degree of glycerol conversion, excess of acetone
vide a major crystallographic structural modification, indicating that the relative to glycerol was employed using only the ACNs as catalysts.
materials were accordingly tailored. Indeed, a synergistic enhancement in catalytic performance was noticed
by varying molar ratio of glycerol to acetone from 1:1 to 1:4 (Fig. 5 (a))
3.2. Catalytic tests for all catalytic evaluation carried out at room temperature, reaching
glycerol conversion of up to 72% with 97% of selectivity toward solketal
Surface features and textural/structural properties of the herein for ACN4. In fact, based on a typical acetalization reaction mechanism of
presented activated carbons from YBW drew attention to the possibility ketal formation over acid sites [61], it is expected that the catalytic cycle
of their direct application as catalyst in the acid catalyzed and solvent- initiates with the chemisorption of the carbonyl group of the acetone
free acetalization reaction of glycerol with acetone under mild condi­ over the stronger Brønsted acid center of the carbon-based catalysts,
tions. Based on our previous studies [57–59], catalyst weight will be which is likely facilitated by the weaker acid groups, followed by the
maintained at 3% for all catalytic evaluations. The catalytic potential of nucleophilic attack of a primary hydroxyl group of the glycerol pro­
the unmodified activated carbons was evaluated at room temperature ducing five-membered ring acetal through dehydration. In this sense,

Fig. 5. a) E ffect of the variation of the glycerol to acetone ratio on the acetalization reaction (3 w t % of catalyst, 5 h at room temperature) and b) Effect of
temperature on the acetalizati on reaction (glycerol:acetone molar ratio of 1:4, 3 wt% of catalyst and 5 h of reaction) em ploying the modified activated car­
bons (ACN).

6
R. Rodrigues et al. Fuel 299 (2021) 120923

to its hydrophilic profile. Therefore, mass transfer appears to be the rate

controlling variable at the initial stage of the process up to the moment
enough amount of solketal is produced, leading it to a typical
kinetically-controlled one phase system [62], as can be demonstrated in
Fig. 7.
Furthermore, experiments at different temperatures were carried out
using glycerol:acetone molar ratio of 1:4 (Fig. 5 (b)). It is clearly
observed an increase in glycerol conversion after 5 h of reaction
comparing experiments at room temperature, 318 K or 328 K, despite of
the previously demonstrated exothermic nature of this process [63].
However, such thermodynamic characteristic seems to play a notable
role in the reactions carried out at higher temperatures, due to the fact
that equilibrium must have already been reached after 5 h of reaction
since similar glycerol conversion in the acetalization reaction with
acetone was noticed for all tested modified activated carbons at 318 K
and 328 K. Moreover, selectivity toward five-membered ring was not
substantially affected by temperature due to the higher thermodynamic
stability of solketal in comparison to six-membered ring acetal. As pre­
viously demonstrated [64], energy of the five-membered ring acetal was
found to be 1.7 Kcal mol− 1 more stable than its isomer, confirming
solketal as the thermodynamic product of the condensation of glycerol
with acetone.
It was further explored the catalytic activity of all ACNs throughout
time, by performing a kinetic investigation of the herein studied solketal
production process at 318 K and 328 K maintaining glycerol to acetone
molar ratio of 1:4. Reaction catalyzed by ACN3 and ACN4 at 318 K
apparently approximate equilibrium slightly faster than ACN1 and
ACN2, which might be related to the concentration of strong acid sites
along with the aforementioned cooperative effects (Fig. 7 (a)). On the
other hand, similar catalytic behavior is observed in the process carried
out at 328 K for all tested catalysts, reaching glycerol conversion of up to
91% after 60 min of reaction for ACN4 and TOF of 215 h− 1 (calculated
considering only the stronger acid sites), as depicted in Fig. 7 (b).
Furthermore, the condensation of glycerol with acetone has been
accurately described by Langmuir–Hinshelwood–Hougen–Watson
(LHHW) reaction rate law [63,65]. In this sense, considering the
Arrhenius equation for the temperature dependence of the reaction rate
constant, it is noticed that the rise in reaction temperature elevates the
kinetic constant to a plateau in which the concentration of solketal in the
early stages of the process became relevant (glycerol conversion roughly
75% upon 20 min of reaction) to a point that interface mass transfer
effects are negligible and thermodynamic equilibrium is rapidly reached
regardless differences in acid sites concentration of the ACN series of
catalysts. Moreover, it has been previously demonstrate [66,67] that in
acetalization reactions described by LHHW rate law, the kinetic constant
is expressively sensitive to the increase in temperature whereas water
adsorption on the surface of the catalyst is minimized thereby reducing
the probability of the reverse reaction to occur. Indeed, it is observed in
Fig. 7b a slightly decrease in conversion of glycerol after 60 min of re­
action (more pronounced for ACN4), which is probably related to the
contribution of the reverse reaction, i.e. hydrolysis of solketal, due to its
enthalpic nature.
Finally, catalytic stability was studied by performing consecutive
catalytic runs applying ACN4 as catalyst at 318 K for 4 h. After each
catalytic run, ACN4 was separated from the reaction medium, washed
three times with ethanol aiming at removing physisorbed compounds
and dried at 363 K for 24 h. It was noticed that, catalytic stability re­
Fig. 6. Kinetic study of the acetalization of glycerol with acetone catalyzed by mains approximately constant after each catalytic process with glycerol
the modified activated carbons (CAN) at room temperature (glycerol:acetone conversion ranging from 89% to 81% holding selectivity to solketal al­
molar ratio of a) 1:4b) 1:2 and c) 1:1, 3 wt% of catalyst). ways higher than 96%. These data indicate that, such less energetic and
environmental clean recovery protocol has been successfully proven
excess of acetone trends to shift equilibrium towards the products, efficient to almost fully reestablish initial catalytic activity of ACN4.
justifying the rise in glycerol conversion levels, as well as to increase However, the subtle and steady decrease in catalytic activity of ACN4
initial rate of reaction (Fig. 6). Nevertheless, glycerol and acetone are can be correlate to the undesired formation of carboxylate esters though
unmissable at room temperature and the catalyst majorly stay in the reaction of the glycerol with the stronger acid sites (–COOH) [57]. It is
higher polar phase (glycerol) at the beginning of the process, likely due also noteworthy to mention that, no or negligible leaching was observed,

7
R. Rodrigues et al. Fuel 299 (2021) 120923

Fig. 7. Kinetic study of the acetalization of glycerol with acetone catalyzed by the modified activated carbons (ACN), glycerol:acetone molar ratio of 1:4, 3 wt% of
catalyst at different temperatures a) 318 K and b) 328 K.

215 h− 1 as well as surface chemistry stability in consecutive catalytic

runs for ACN4.

CRediT authorship contribution statement

Raphael Rodrigues: Conceptualization, Methodology, Investiga­

tion, Writing - original draft, Writing - review & editing, Visualization,
Project administration. Mikaela S. Santos: Investigation. Renan S.
Nunes: Investigation. Wagner A. Carvalho: Conceptualization, Re­
sources, Writing - review & editing, Visualization, Funding acquisition.
Georgia Labuto: Conceptualization, Resources, Writing - review &
editing, Visualization, Funding acquisition.

Declaration of Competing Interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Acknowledgment

Fig. 8. Recycling test employing ACN4 in the acetalization of glycerol with

The authors are grateful to FAPESP (2016/06271-4, 2017/24931-4
acetone at 318 K (glycerol:acetone molar ratio of 1:4, 3 wt% of catalyst).
and 2016/05006-5), CAPES (Finance Code 001), CNPq (404843/
2018-2) for financial support and Biorigin Company (São Paulo, Brazil)
which can be inferred by the reusability of the catalyst, confirming the by supply the raw material.
heterogeneous nature of the process, as presented in Fig. 8.
References
4. Conclusions
[1] CO2 Emissions from Fuel Combustion 2019 https://www.iea.org/reports/co2-
Production of a new series of activated carbons from yeast biomass emissions-from-fuel-combustion-2019 (accessed January 11, 2020).
[2] Nejat P, Jomehzadeh F, Taheri MM, Gohari M, Abd. Majid MZ. A global review of
waste (undesired waste from ethanol production cycle) by a chemical energy consumption, CO2 emissions and policy in the residential sector (with an
activation process with H3PO4 have been successfully proven efficient. overview of the top ten CO2 emitting countries). Renew Sustain Energy Rev 2015;
The obtained materials hold remarkable textural proprieties with spe­ 43:843–62. doi: 10.1016/j.rser.2014.11.066.
[3] Goh BHH, Ong HC, Cheah MY, Chen W-H, Yu KL, Mahlia TMI. Sustainability of
cific surface area ranging from 983 to 1208 m2g− 1 and meso/micropore direct biodiesel synthesis from microalgae biomass: A critical review. Renew
structural design dependent on the activation agent concentration. Sustain Energy Rev 2019;107:59–74. doi: 10.1016/j.rser.2019.02.012.
Tuning their surface chemistry by a treatment with nitric acid gave rise [4] Singh D, Sharma D, Soni SL, Sharma S, Kumar Sharma P, Jhalani A. A review on
feedstocks, production processes, and yield for different generations of biodiesel.
to a new series of modified activated carbons detaining certain oxygen
Fuel 2020;262:116553. https://doi.org/10.1016/j.fuel.2019.116553.
containing groups, such as carboxylic groups, in relevant concentration [5] Chol CG, Dhabhai R, Dalai AK, Reaney M. Purification of crude glycerol derived
that affectively acted as an active and selective center for catalyzing the from biodiesel production process: Experimental studies and techno-economic
solketal production process. Indeed, it demonstrated above average analyses. Fuel Process Technol 2018;178:78–87. https://doi.org/10.1016/j.
fuproc.2018.05.023.
catalytic performance, achieving glycerol conversion of up to 91% after [6] Kumar LR, Yellapu SK, Tyagi RD, Zhang X. A review on variation in crude glycerol
60 min of reaction, with 97% of selectivity toward solketal, and TOF of composition, bio-valorization of crude and purified glycerol as carbon source for

8
R. Rodrigues et al. Fuel 299 (2021) 120923

lipid production. Bioresour Technol 2019;293:122155. https://doi.org/10.1016/j. [31] Gomes HT, Miranda SM, Sampaio MJ, Silva AMT, Faria JL. Activated carbons
biortech.2019.122155. treated with sulphuric acid: Catalysts for catalytic wet peroxide oxidation. Catal
[7] Fasahati P, Liu JJ, Ohlrogge JB, Saffron CM. Process design and economics for Today 2010;151(1-2):153–8. https://doi.org/10.1016/j.cattod.2010.01.017.
production of advanced biofuels from genetically modified lipid-producing [32] Lippens BC, de Boer JH. Studies on pore systems in catalysts: V. The t method. J
sorghum. Appl Energy 2019;239:1459–70. doi: 10.1016/j.apenergy.2019.01.143. Catal 1965;4:319–23. doi: 10.1016/0021-9517(65)90307-6.
[8] Jiang Y, Li X, Zhao H, Hou Z. Esterification of glycerol with acetic acid over SO3H- [33] Hutson ND, Yang RT. Theoretical basis for the Dubinin-Radushkevitch (D-R)
functionalized phenolic resin. Fuel 2019;255:115842. doi: 10.1016/j. adsorption isotherm equation. Adsorption 1997;3:189–95. doi: 10.1007/
fuel.2019.115842. BF01650130.
[9] Tang Z, Cao H, Tao Y, Heeres HJ, Pescarmona PP. Transfer hydrogenation from [34] Olivier JP, Conklin WB, Szombathely M v. Determination of Pore Size Distribution
glycerol over a Ni-Co/CeO2 catalyst: a highly efficient and sustainable route to from Density Functional Theory: A Comparison of Nitrogen and Argon Results. In:
produce lactic acid. Appl Catal B 2020;263:118273. https://doi.org/10.1016/j. Rouquerol J, Rodríguez-Reinoso F, Sing KSW, Unger KKBT-S in SS and C, editors.
apcatb.2019.118273. Charact. Porous Solids III, vol. 87, Elsevier; 1994, p. 81–9. doi: 10.1016/S0167-
[10] de Paula FGF, Rosmaninho MG, Teixeira AP de C, de Souza PP, Lago RM. 2991(08)63067-0.
Alcoxycle: A novel route for glycerol reform into H2 and COx in separate stages. [35] Boehm HP. Some aspects of the surface chemistry of carbon blacks and other
Catal Today 2017;289:127–32. doi: 10.1016/j.cattod.2016.06.039. carbons. Carbon N Y 1994;32:759–69. doi: 10.1016/0008-6223(94)90031-0.
[11] Aghbashlo M, Hosseinpour S, Tabatabaei M, Rastegari H, Ghaziaskar HS. Multi- [36] Physisorption of gases, with special reference to the evaluation of surface area and
objective exergoeconomic and exergoenvironmental optimization of continuous pore size distribution (IUPAC Technical Report) . Pure Appl Chem 2015;87:1051.
synthesis of solketal through glycerol ketalization with acetone in the presence of https://doi.org/10.1515/pac-2014-1117.
ethanol as co-solvent. Renew Energy 2019;130:735–48. doi: 10.1016/j. [37] Nakagawa Y, Molina-Sabio M, Rodríguez-Reinoso F. Modification of the porous
renene.2018.06.103. structure along the preparation of activated carbon monoliths with H3PO4 and
[12] Fraile JM, Mallada R, Mayoral JA, Menéndez M, Roldán L. Shift of multiple ZnCl2. Microporous Mesoporous Mater 2007;103(1-3):29–34. https://doi.org/
incompatible equilibriums by a combination of heterogeneous catalysis and 10.1016/j.micromeso.2007.01.029.
membranes. Chem. Eur. J. 2010;16(11):3296–9. [38] Prahas D, Kartika Y, Indraswati N, Ismadji S. Activated carbon from jackfruit peel
[13] Deutsch J, Martin A, Lieske H. Investigations on heterogeneously catalysed waste by H3PO4 chemical activation: Pore structure and surface chemistry
condensations of glycerol to cyclic acetals. J Catal 2007;245(2):428–35. https:// characterization. Chem Eng J 2008;140(1-3):32–42. https://doi.org/10.1016/j.
doi.org/10.1016/j.jcat.2006.11.006. cej.2007.08.032.
[14] Ghosh A, Singha A, Auroux A, Das A, Sen D, Chowdhury B. A green approach for [39] Molina-Sabio M, Rodrı ́guez-Reinoso F. Role of chemical activation in the
the preparation of a surfactant embedded sulfonated carbon catalyst towards development of carbon porosity. Colloids Surf, A 2004;241(1-3):15–25. https://
glycerol acetalization reactions. Catal. Sci. Technol. 2020;10(14):4827–44. doi.org/10.1016/j.colsurfa.2004.04.007.
https://doi.org/10.1039/D0CY00336K. [40] Jagtoyen M, Derbyshire F. Activated carbons from yellow poplar and white oak by
[15] Kowalska-Kuś J, Held A, Nowińska K. A continuous-flow process for the H3PO4 activation. Carbon 1998;36(7-8):1085–97. https://doi.org/10.1016/S0008-
acetalization of crude glycerol with acetone on zeolite catalysts. Chem Eng J 2020; 6223(98)00082-7.
401:126143. https://doi.org/10.1016/j.cej.2020.126143. [41] Molina-Sabio M, RodRíguez-Reinoso F, Caturla F, Sellés MJ. Porosity in granular
[16] da Silva MJ, Rodrigues AA, Pinheiro PF. Solketal synthesis from glycerol and carbons activated with phosphoric acid. Carbon 1995;33(8):1105–13. https://doi.
acetone in the presence of metal salts: a Lewis or Brønsted acid catalyzed reaction? org/10.1016/0008-6223(95)00059-M.
Fuel 2020;276:118164. https://doi.org/10.1016/j.fuel.2020.118164. [42] Aguilar-Uscanga B, Francois JM. A study of the yeast cell wall composition and
[17] Bakuru VR, Churipard SR, Maradur SP, Kalidindi SB. Exploring the Brønsted structure in response to growth conditions and mode of cultivation. Lett Appl
acidity of UiO-66 (Zr, Ce, Hf) metal–organic frameworks for efficient solketal Microbiol 2003;37(3):268–74. https://doi.org/10.1046/j.1472-765X.2003.01394.
synthesis from glycerol acetalization. Dalton Trans. 2019;48(3):843–7. https://doi. x.
org/10.1039/C8DT03512A. [43] El-Hendawy A-NA. Influence of HNO3 oxidation on the structure and adsorptive
[18] Chen L, Nohair B, Zhao D, Kaliaguine S. Highly Efficient Glycerol Acetalization properties of corncob-based activated carbon. Carbon N Y 2003;41:713–22. doi:
over Supported Heteropoly Acid Catalysts. ChemCatChem 2018;10:1918–25. doi: 10.1016/S0008-6223(03)00029-0.
10.1002/cctc.201701656. [44] Jaramillo J, Álvarez PM, Gómez-Serrano V. Oxidation of activated carbon by dry
[19] Castro CS de, Viau LN, Andrade JT, Mendonça TAP, Gonçalves M. Mesoporous and wet methods: Surface chemistry and textural modifications. Fuel Process
activated carbon from polyethyleneterephthalate (PET) waste: pollutant Technol 2010;91:1768–75. doi: 10.1016/j.fuproc.2010.07.018.
adsorption in aqueous solution. New J Chem 2018;42:14612–9. doi: 10.1039/ [45] Jeromenok J, Weber J. Restricted Access: On the Nature of Adsorption/Desorption
C8NJ02715C. Hysteresis in Amorphous, Microporous Polymeric Materials. Langmuir 2013;29:
[20] Cui Y, Atkinson JD. Tailored activated carbon from glycerol: role of acid 12982–9. doi: 10.1021/la402630s.
dehydrator on physiochemical characteristics and adsorption performance. [46] Braida WJ, Pignatello JJ, Lu Y, Ravikovitch PI, Neimark AV, Xing B. Sorption
J. Mater. Chem. A 2017;5(32):16812–21. https://doi.org/10.1039/C7TA02898A. hysteresis of benzene in charcoal particles. Environ. Sci. Technol. 2003;37(2):
[21] Ternero-Hidalgo JJ, Rosas JM, Palomo J, Valero-Romero MJ, Rodríguez-Mirasol J, 409–17. https://doi.org/10.1021/es020660z.
Cordero T. Functionalization of activated carbons by HNO3 treatment: Influence of [47] Lai W, Yang S, Jiang Y, Zhao F, Li Z, Zaman B, et al. Artefact peaks of pore size
phosphorus surface groups. Carbon N Y 2016;101:409–19. doi: 10.1016/j. distributions caused by unclosed sorption isotherm and tensile strength effect.
carbon.2016.02.015. Adsorption 2020;26(4):633–44. https://doi.org/10.1007/s10450-020-00228-1.
[22] Kharrazi SM, Mirghaffari N, Dastgerdi MM, Soleimani M. A novel post- [48] Danish M, Hashim R, Ibrahim MNM, Sulaiman O. Effect of acidic activating agents
modification of powdered activated carbon prepared from lignocellulosic waste on surface area and surface functional groups of activated carbons produced from
through thermal tension treatment to enhance the porosity and heavy metals Acacia mangium wood. J Anal Appl Pyrolysis 2013;104:418–25. doi: 10.1016/j.
adsorption. Powder Technol 2020;366:358–68. doi: 10.1016/j. jaap.2013.06.003.
powtec.2020.01.065. [49] Iruretagoyena D, Bikane K, Sunny N, Lu H, Kazarian SG, Chadwick D, et al.
[23] Jiang C, Yakaboylu GA, Yumak T, Zondlo JW, Sabolsky EM, Wang J. Activated Enhanced selective adsorption desulfurization on CO2 and steam treated activated
carbons prepared by indirect and direct CO2 activation of lignocellulosic biomass carbons: Equilibria and kinetics. Chem Eng J 2020;379:122356. doi: 10.1016/j.
for supercapacitor electrodes. Renew Energy 2020;155:38–52. doi: 10.1016/j. cej.2019.122356.
renene.2020.03.111. [50] Jaramillo J, Álvarez PM, Gómez-Serrano V. Preparation and ozone-surface
[24] Danish M, Ahmad T, Hashim R, Said N, Akhtar MN, Mohamad-Saleh J, et al. modification of activated carbon. Thermal stability of oxygen surface groups. Appl
Comparison of surface properties of wood biomass activated carbons and their Surf Sci 2010;256(17):5232–6. https://doi.org/10.1016/j.apsusc.2009.12.109.
application against rhodamine B and methylene blue dye. Surf Interfaces 2018;11: [51] Vinke P, van der Eijk M, Verbree M, Voskamp AF, van Bekkum H. Modification of
1–13. https://doi.org/10.1016/j.surfin.2018.02.001. the surfaces of a gasactivated carbon and a chemically activated carbon with nitric
[25] Bordonal R de O, Carvalho JLN, Lal R, de Figueiredo EB, de Oliveira BG, La Scala acid, hypochlorite, and ammonia. Carbon 1994;32(4):675–86. https://doi.org/
N. Sustainability of sugarcane production in Brazil. A review. Agron Sustain Dev 10.1016/0008-6223(94)90089-2.
2018;38:13. doi: 10.1007/s13593-018-0490-x. [52] Ferrari AC, Robertson J. Interpretation of Raman spectra of disordered and
[26] Labuto G, Trama B, Gueller GC da S, Guarnieri B de S, Silva FV da, Collazo R. amorphous carbon. Phys. Rev. B 2000;61(20):14095–107. https://doi.org/
Metals uptake by live yeast and heat-modified yeast residue . Rev Ambient Água 10.1103/PhysRevB.61.14095.
2015;10:510–9. [53] Yang T, Lua AC. Textural and chemical properties of zinc chloride activated
[27] Korhonen J, Honkasalo A, Seppälä J. Circular Economy: The Concept and its carbons prepared from pistachio-nut shells. Mater Chem Phys 2006;100:438–44.
Limitations. Ecol Econ 2018;143:37–46. doi: 10.1016/j.ecolecon.2017.06.041. doi: 10.1016/j.matchemphys.2006.01.039.
[28] Labuto G, Carrilho ENVM. Chapter 22 - Bioremediation in Brazil: Scope and [54] Childres I, Jauregui LA, Park W, Cao H, Chen YP. Raman spectroscopy of graphene
Challenges to Boost Up the Bioeconomy. In: Prasad MNVBT-B and B, editor., and related materials. New Dev Phot Mater Res 2013:1.
Elsevier; 2016, p. 569–88. doi: 10.1016/B978-0-12-802830-8.00022-8. [55] Yu L, Tatsumi D, Zuo S, Morita M. Promotion of Crystal Growth on Biomass-based
[29] Modesto HR, Lemos SG, dos Santos MS, Komatsu JS, Gonçalves M, Carvalho WA, Carbon using Phosphoric Acid Treatments. Bioresour Vol 10, No 2 2015.
et al. Activated carbon production from industrial yeast residue to boost up circular [56] Khadiran T, Hussein MZ, Zainal Z, Rusli R. Textural and Chemical Properties of
bioeconomy. Environ Sci Pollut Res 2020. doi: 10.1007/s11356-020-10458-z. Activated Carbon Prepared from Tropical Peat Soil by Chemical Activation
[30] Chingombe P, Saha B, Wakeman RJ. Surface modification and characterisation of a Method. Bioresour Vol 10, No 1 2014.
coal-based activated carbon. Carbon 2005;43(15):3132–43. https://doi.org/ [57] Rodrigues R, Gonçalves M, Mandelli D, Pescarmona PP, Carvalho WA. Solvent-free
10.1016/j.carbon.2005.06.021. conversion of glycerol to solketal catalysed by activated carbons functionalised
with acid groups. Catal Sci Technol 2014;4:2293–301. doi: 10.1039/C4CY00181H.

9
R. Rodrigues et al. Fuel 299 (2021) 120923

[58] Gonçalves M, Rodrigues R, Galhardo TS, Carvalho WA. Highly selective [63] Nanda MR, Yuan Z, Qin W, Ghaziaskar HS, Poirier M-A, Xu CC. Thermodynamic
acetalization of glycerol with acetone to solketal over acidic carbon-based catalysts and kinetic studies of a catalytic process to convert glycerol into solketal as an
from biodiesel waste. Fuel 2016;181:46–54. doi: 10.1016/j.fuel.2016.04.083. oxygenated fuel additive. Fuel 2014;117:470–7. doi: 10.1016/j.fuel.2013.09.066.
[59] Rodrigues R, Mandelli D, Gonçalves NS, Pescarmona PP, Carvalho WA. [64] Ozorio LP, Pianzolli R, Mota MBS, Mota CJA. Reactivity of glycerol/acetone ketal
Acetalization of acetone with glycerol catalyzed by niobium-aluminum mixed (solketal) and glycerol/formaldehyde acetals toward acid-catalyzed hydrolysis . J
oxides synthesized by a sol–gel process. J Mol Catal A Chem 2016;422:122–30. doi: Brazilian Chem Soc 2012;23:931–7.
10.1016/j.molcata.2016.02.002. [65] Moreira MN, Faria RP V, Ribeiro AM, Rodrigues AE. Solketal Production from
[60] Yu X, Chu Y, Zhang L, Shi H, Xie M, Peng L, et al. Adjacent acid sites cooperatively Glycerol Ketalization with Acetone: Catalyst Selection and Thermodynamic and
catalyze fructose to 5-hydroxymethylfurfural in a new, facile pathway. J Energy Kinetic Reaction Study. Ind Eng Chem Res 2019;58:17746–59. doi: 10.1021/acs.
Chem 2020;47:112–7. doi: 10.1016/j.jechem.2019.11.020. iecr.9b03725.
[61] Kao LC, Kan WC, Martin-Aranda RM, Guerrero-Perez MO, Bañares MÁ, Liou SYH. [66] Faria RP V, Pereira CSM, Silva VMTM, Loureiro JM, Rodrigues AE. Glycerol
SiO2 supported niobium oxides with active acid sites for the catalytic acetalization valorisation as biofuels: Selection of a suitable solvent for an innovative process for
of glycerol. Catal Today 2019. doi: 10.1016/j.cattod.2019.08.007. the synthesis of GEA. Chem Eng J 2013;233:159–67. doi: 10.1016/j.
[62] Rahaman MS, Phung TK, Hossain MA, Chowdhury E, Tulaphol S, Lalvani SB, et al. cej.2013.08.035.
Hydrophobic functionalization of HY zeolites for efficient conversion of glycerol to [67] Graca NS, Pais LS, Silva VMTM, Rodrigues AE. Oxygenated Biofuels from Butanol
solketal. Appl Catal A Gen 2020;592:117369. doi: 10.1016/j.apcata.2019.117369. for Diesel Blends: Synthesis of the Acetal 1,1-Dibutoxyethane Catalyzed by
Amberlyst-15 Ion-Exchange Resin. Ind Eng Chem Res 2010;49:6763–71. https://
doi.org/10.1021/ie901635j.

10