Chemical Engineering Journal 429 (2022) 132279

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Sulfated complex metal oxides solid acids with dual Brønsted-Lewis acidic
property for production of 5-ethoxymethylfurfural from
biomass-derived carbohydrates
Yunfei He a, b, Luxin Zhang a, *, Yuting Liu a, Simin Yi a, Han Yu c, Yujie Zhu d, Ruijun Sun a
a
College of Environmental and Municipal Engineering, Shaanxi Key Laboratory of Environmental Engineering, Key Lab of Northwest Water Resource, Environment and
Ecology, MOE, Xi’an University of Architecture and Technology, Xi’an 710055, PR China
b
State Key Joint Laboratory of Environment Simulation and Pollution Control, School of Environment, Tsinghua University, Beijing 100084, PR China
c
Division of Water Resources Engineering, Lund University, Sweden
d
College of Environmental Science and Engineering, Academy of Environment and Ecology, Tianjin University, Tianjin 300072, PR China

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

Keywords: The transformation of aldose-based carbohydrates into 5-ethoxymethylfurfural (EMF) is very challenging as
Sulfated complex metal oxides compared to ketose-based carbohydrates, but the formers are more abundant and cheaper. Here, a series of
5-ethoxymethylfurfural sulfated complex metal oxides were synthesized for the conversion of aldose-based mono-, di-, and poly-
Brønsted/Lewis dual acidic solid acid
saccharides, as well as starchy food waste into EMF. The catalysts were carefully characterized and the results
Biomass
showed that the type and strength of the acid sites were more important than their concentration. It was also
Biofuel
Kinetic and thermodynamic analysis shown that the efficiency of these catalysts was significantly affected by the metal species in the catalyst
composition and followed the order tetra- > tri- > bi- > mono-component metal oxides based catalyst. Among
the prepared catalysts, Zr-Sn-Fe-Al-O-S exhibited superior catalytic activity, with an EMF yield of 33.1% from
glucose, and yields ranging from 4.1− 26.3% for di-, poly-saccharides and starchy food waste in ethanol/
dimethyl sulfoxide solvent system under glucose/catalyst mass ratio of 4. The role of co-solvent in the reaction
pathway was also studied. It was found that the predominant reaction pathway for EMF production was closely
related to the co-solvent amount. A kinetic model of glucose conversion to EMF was developed and the ther­
modynamic analysis was performed, the main features of the experimental observations can be described by the
model. Zr-Sn-Fe-Al-O-S was reused for four runs without intermediate regeneration steps, showing a slight decay
in activity. After reactivation by calcination before the fifth cycle, the catalyst recovered its activity, indicating
good reusability and thermal stability.

1. Introduction etherification of HMF with ethanol. Due to the high energy density and
remarkable reduction of soot and SOx emissions, EMF has gained
The rapid depletion of energy resources and the continuous deteri­ considerable attentions [10–12].
oration of environment promoted the efforts of identifying renewable In the past few years, increasing attention has been paid to the
resources to supply fuels and chemicals [1,2]. The biomass-derived synthesis of EMF in various reaction systems by using the catalytic route
carbohydrates are abundant and relatively inexpensive, and thus, are [1,12–16]. The feedstocks used to produce EMF are mainly HMF [17],
considered as renewable feedstocks to produce fuels and valuable commercial saccharides (e.g., fructose, glucose, sucrose), and other
chemicals [2–4]. Among the various biomass-derived platform chem­ biomass-derived oligo- or poly-carbohydrates [18]. Generally, EMF can
icals, 5-hydroxymethylfurfural (HMF) has been widely studied and be synthesized from HMF or ketose-based carbohydrates with satisfac­
recognized as a characteristic platform compound for biomass and fuels tory yields. Yet, the large-scale production of EMF from HMF or ketose-
[5–7]. The reduction, oxidation, and esterification of HMF produce based sugars is limited due to the high price of the raw materials [19].
various high-value derivative chemicals [8,9]. Under acidic conditions, Therefore, good deal of work has been done to produce EMF using
HMF can be converted into 5-ethoxymethylfurfural (EMF) by heterogeneous acidic catalyst from glucose or other abundant aldose-

* Corresponding author.
E-mail address: [email protected] (L. Zhang).

https://doi.org/10.1016/j.cej.2021.132279
Received 8 May 2021; Received in revised form 27 August 2021; Accepted 1 September 2021
Available online 8 September 2021
1385-8947/© 2021 Elsevier B.V. All rights reserved.
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

based carbohydrates at a lower price [20,21]. addition of any co-solvents. Furthermore, the kinetic-thermo model was
In most cases, heterogeneous catalysts with Brønsted acid sites were developed, and the kinetic and thermodynamic parameters were
not effective for EMF production from carbohydrates consisting of presented.
glucose subunits. No EMF was detected when glucose was used even the
reaction system exhibited good performance in generating HMF 2. Materials and methods
[10,12,22]. The conversion of glucosyl-based carbohydrates into EMF
involves multiple parallel and cascade reactions, mainly ethanolysis of 2.1. Materials
glucose to ethyl glucoside (EG), isomerization of glucose to fructose,
isomerization of EG to ethyl fructoside (EF), dehydration of fructose (or D-glucose (≥99%), ethanol (99%), cellobiose (98%), sucrose
EF), and etherification of HMF. The isomerization of glucose to fructose, (≥99%), inulin (≥99%), starch (≥99%) and cellulose (≥99%) were
and the isomerization of EG to EF are the rate-limiting steps of the supplied by Alfa Aesar. Dimethyl sulfoxide (DMSO, 99%), tetrahydro­
process. Brønsted acids alone are inactive for these two isomerization furan (THF, 99%), γ-butyrolactone (GBL, 99%), γ-valerolactone (GVL,
reactions. In general, isomerization of glucose to fructose, and EG to EF, 99%), toluene (99%), ZrCl4, SnCl4⋅5H2O, FeCl3, AlCl3⋅6H2O, ZnCl2,
occurs on Lewis acid sites. Therefore, when projecting the production of CrCl3, (NH4)2SO4 were supplied by Tianjin Kermel Chemical Co., Ltd.
EMF from carbohydrates containing glucose subunits via an acid catal­ (Tianjin, China). Ethyl glucoside (EG, 98%) and ethyl fructoside (EF,
ysis route, both Lewis and Brønsted acid sites should be considered. 98%) were purchased from Bide Pharmatech Ltd. (Shanghai, China). All
Brønsted acid sites are more active than Lewis acid sites in reactions, commercial reagents were used as received without further purification.
such as hydrolysis, dehydration, and etherification [23], but excessive Kitchen waste, expired potato chips and instant noodles were dried,
strong Brønsted acid sites promote EMF conversion into ethyl levulinate milled and then passed through 80 mesh sieves before experiment.
(EL), which is a diesel additive [24]. Therefore, the rational control of
active Brønsted/Lewis sites in heterogeneous catalysts is crucial to 2.2. Synthesis of sulfated complex metal oxides solid acids catalysts
improve the yield of target product EMF.
The synergistic effect of Brønsted and Lewis acid sites was observed In a general procedure, M1 precursors (Table 1) were dissolved in
for the glucose conversion into EMF. The EMF yields of 10–20% were 500 mL deionized water and stirred for a few minutes at room temper­
obtained from glucose when commercial zeolites and -SO3H function­ ature to form the solution A. Then, a solution of NaOH (10 wt%) was
alized MIL were used [19]. In order to improve the reaction efficiency of added to the solution A to form a colloidal precipitate. After aging for 15
the ethanolysis of glucose into EMF, Lew et al. investigated the effect of h, the resulted precipitates were filtered and washed three times with
the Sn-Beta and Amberlyst-131 mixture on the conversion of glucose in deionized water and then dried in an oven at 105 ◦ C for 12 h. The ob­
anhydrous ethanol, for which a 31% EMF yield was obtained [25]. Since tained solids were milled and then intensively mixed with (NH4)2SO4
low to medium EMF yield can be obtained from glucose, the production (20 wt% of the solid mixture). The obtained powders were then calcined
of EMF from glucose-based polysaccharides or raw biomass is still far of at 600 ◦ C for 10 h. The resulted samples were labeled as Zr-O-S, Zr-Sn-O-
being satisfactory. The development of highly efficient and cost- S, Zr-Sn-Fe-O-S, Zr-Sn-Al-O-S, Zr-Sn-Zn-O-S, Zr-Sn-Cr-O-S, Zr-Sn-Fe-Al-
effective heterogeneous catalysts for direct conversion of glucose and O-S, and Zr-Sn-Zn-Cr-O-S.
glucosyl-based polysaccharides into EMF is therefore needed.
Due to the presence of acid centers on their surfaces, various metal
2.3. Catalysts characterization
oxides were commonly used as catalysts in chemical reaction [26,27]. It
is universally acknowledged that metal oxides delivered enhanced cat­
Scanning electron microscopy (SEM) images were recorded on a
alytic activity [28–30]. Thus, using complex metal oxides to catalyze
Quanta FEG 250 scanning electron microscope at 10 kV. The different
EMF formation is worth trying. In the present work, a series of sulfated
metal elements composition and distribution of catalysts were tested
complex metal oxides solid acids catalysts with both Lewis and Brønsted
using energy dispersive X-ray spectroscopy (EDS) and scanning trans­
acid sites were synthesized and used as heterogeneous catalysts for the
mission electron microscopy (STEM). The specific surfaces areas and the
transformation of biomass-derived carbohydrates and starchy food
pore volumes of the catalysts were calculated from nitrogen adsorption
waste into EMF. Due to the high acidity of ZrO2 compared with other
and desorption isotherms at − 196 ◦ C recorded on a Micromeritics ASAP
metal oxides, Zr was chosen as main metal element of the prepared
2020 instrument. The samples were degassed under N2 flow at 250 ◦ C
catalysts [31]. Moreover, transition element (Sn, Fe, Zn, Cr) and Al
overnight prior to analysis. The specific surface area was determined
metal were selected as subordinate elements because of their catalytic
using Brunauer–Emmett–Teller (BET) model. The Bru­
application in the synthesis of furan compounds [32,33]. It was reported
nauer–Joyner–Halenda model was employed to calculate the pore vol­
that the presence of SO42- promoted catalytic activity of metal oxides
ume and pore size. The crystalline structures of the synthesized samples
and increased the stability of physical texture to thermal treatment [34].
What’s more, thermal decomposition of ammonia from ammonium salt
at high temperature generated Brønsted sites [26], which were essential Table 1
M1 precursor composition of the catalysts.
to the catalytic cascade reaction. Considering the given reasons,
(NH4)2SO4 as precursor was introduced to provide NH4+ and SO42- Entry Catalysts M1 precursor Molar ratio of the substances
species in our catalysts. in M1 precursor

Besides acidic and characteristics of a catalyst, the physico-chemical 1 Zr-O-S ZrCl4 1

properties, the morphology and structure of the catalyst, as well as the 2 Zr-Sn-O-S ZrCl4: SnCl4⋅5H2O 3: 2
3 Zr-Sn-Fe-O- ZrCl4: SnCl4⋅5H2O: FeCl3 3: 1: 1
polarity and protonicity of co-solvent affect the activity of the catalyst,
S
the reaction rate and product distribution, and may lead to changes in 4 Zr-Sn-Al-O- ZrCl4: SnCl4⋅5H2O: 3: 1: 1
the reaction pathway. In this work, the physico-chemical properties of S AlCl3⋅6H2O
all prepared catalysts were studied by different techniques. The different 5 Zr-Sn-Zn-O- ZrCl4: SnCl4⋅5H2O: ZnCl2 3: 1: 1
reaction parameters of reaction system were investigated in relation to S
6 Zr-Sn-Cr-O- ZrCl4: SnCl4⋅5H2O: CrCl3 3: 1: 1
the EMF yield to optimize the reaction conditions for EMF production. S
The properties of solvent may affect the reaction rate and the products 7 Zr-Sn-Fe-Al- ZrCl4: SnCl4⋅5H2O: FeCl3: 7: 1: 1: 1
evolution. Therefore, the effect of co-solvent on the conversion of O-S AlCl3⋅6H2O
glucose to EMF was further investigated. The role of a co-solvent on the 8 Zr-Sn-Zn-Cr- ZrCl4: SnCl4⋅5H2O: ZnCl2: 7: 1: 1: 1
O-S CrCl3
reaction pathway was illustrated via control experiments without the

2
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

were analyzed on a Seifert 3003 T powder X-ray diffraction (XRD) with determined by a HPLC (Waters 2412) with an aminex HPX-87P column
Cu Kα radiation (λ = 0.154 nm) at a voltage of 40 kV and a current of 40 (Agilent) and a refractive index detector. A H2SO4 solution (0.005 M)
mA. Each sample was scanned in the range of 2θ = 5◦ − 80◦ at a rate of was used as mobile phase at a flow rate of 0.6 mL/min, and the tem­
2◦ /min. X-ray photoelectron spectra (XPS) was conducted on a Physical perature of column was maintained at 65 ◦ C.
Electronics PHI 5700 spectrometer with a mono-chromatic Al-Kα radi­ The starch content in kitchen waste, expired potato chips, and
ation (300 W, 15 kV, 1253.6 eV) to analyze the core level signals of the expired instant noodles were measured based on the method of
different elements, with a multi-channel detector. The elemental anal­ “Determination of starch in food from national standards for food safety”
ysis of all prepared catalysts was determined by inductively coupled (Chinese Standard GB 5009.9–2016). The contents of the starch in
plasma atomic emission spectroscopy (ICP-AES, ICPS-7500). The kitchen waste, expired potato chips, and expired instant noodles were
temperature-programmed desorption of ammonia (NH3-TPD) measure­ 65.1, 74.1, and 71.4%, respectively.
ments were performed to determine the acid amount and strength of the The conversion of the reactants and the yields of products were
catalysts, and the results were recorded on an AutoChem II 2920. Before calculated as follows (Eq. 1–2):
adsorption, the samples were degassed under He flow at 300 ◦ C for 0.5 h.
mol of reactant initial − mol of reactant final
Then, the catalysts were cooled to the desired temperature and kept for Conversion(%) = × 100%
mol of reactant initial
2 h. After that, the NH3 was adsorbed on the catalyst surface at 100 ◦ C,
(1)
followed by flushing with He for 0.5 h to remove the physically absorbed
NH3. Finally, NH3-TPD was carried out under a constant flow of He from mol of product produced
100 to 900 ◦ C at a heating rate of 10 ◦ C/min. An on-line gas chro­ Yield(%) = × 100% (2)
mol of reactant subunit initial
matograph (Shimadzu GC-14A) equipped with a thermal conductivity
detector (TCD) was used to quantify the evolved ammonia. The Brønsted
3. Results and discussion
and Lewis acid sites of the catalysts were evaluated based on the
pyridine-FTIR spectra, which were recorded on a Shimadzu Fourier
3.1. Catalyst characterization‘
Transformation Instrument (FTIR-8300). The samples were heated to
40, 200 and 350 ◦ C, respectively. And then the catalysts were evacuated
3.1.1. Morphology
at 10-2 Pa for 1 h, followed by cooling to room temperature. The cata­
SEM images of all the prepared catalysts are depicted in Figure S1.
lysts adsorbed saturated pyridine vapors for 0.5 h at room temperature
All catalyst samples displayed a brick-like morphology of irregular
and then outgassed at different temperatures before spectra recording.
lumps with a rough surface and small adherent particles. The particles
The thermal stability of the catalysts was measured by thermogravim­
had diameters ranging from several nanometers to tens of microns. The
etry (TG), which was carried out on a NETZSCHSTA 429 instrument in
metal contents analysis by ICP-AES (Table S1) revealed that Zr was the
the range of 30 – 800 ◦ C under N2 atmosphere with a heating rate of
main metal element for all the prepared catalysts. Furthermore, the
10 ◦ C/min.
element mapping images of all synthetic catalysts and their corre­
sponding STEM dark-field images (Figure S2) proved that all the metal
2.4. Catalytic conversion of glucose
elements were evenly distributed on the surface of these catalysts.
To analyze the textural properties, i.e., surface area, pore volume,
Based on the previous literature, the catalytic tests were performed
and pore size, of these catalysts, N2 physisorption was used. The iso­
in a 15 mL thick-walled glass reactor, which was placed in a pre-heated
therms of all samples were shown in Figure S3. In addition, the pore size
oil bath and maintained at 160 ◦ C (oil temperature detected by the
distribution curves (Figure S4) showed that these catalysts exhibited a
probe) under magnetic stirring. In general experiments, 5 mg catalyst
relatively large pore size distribution in the range of 10 to 40 nm. The
and 20 mg feedstock were added to pure ethanol or a solvent system
specific surface area and pore volume of all catalysts were summarized
consisting of ethanol and a co-solvent with a total volume of 2 mL. The
in Table 2. The differences in specific surface area of these catalysts were
feedstock-to-catalyst weight ratio was 4 unless otherwise specified. Time
mainly caused by the different metal oxides. Compared with a single
zero was taken when the set temperature was reached. After the desired
metal oxide-based catalyst, the specific surface area of the catalysts
reaction time, the reactor was quickly removed from the oil bath and
increased with an increasing number of metal elements due to the
cooled to room temperature by tap water. The sample was diluted with
generated tiny metal oxides adhering to the catalyst and the new created
deionized water and filtered with a 0.22 μm polytetrafluoroethylene
porosity. Therefore, the specific surface area of tri-component metal
filter membrane to remove insoluble solid prior to analysis. The catalyst
oxide-based catalysts was significantly higher than that of the Zr-O-S
recycling experiments were performed by mixing 1 mL DMSO, 1 mL
catalyst. The surface areas of the Zr-Sn-Fe-O-S and Zr-Sn-Al-O-S cata­
ethanol, 5 mg catalyst, and 20 mg glucose with magnetic stirring at
lysts were found to be higher than those of Zr-Sn-Zn-O-S and Zr-Sn-Cr-O-
160 ◦ C for 12 h. After each reaction run, the catalyst was separated by
S. This was possibly related to the surface area of the metal oxides, as
filtration from the reaction mixture, dried at room temperature over­
Al2O3 and Fe2O3 displayed larger surface areas than ZnO and Cr2O3
night, and directly reused in the next catalytic run without regeneration
[35–38]. However, the catalysts easily collapsed due to the further in­
treatment. However, the catalyst was calcined at 600 ◦ C prior to reuse in
crease in metal oxide content, causing collapsed pores and a reduced
the fifth run.
surface area [36]. Thus, the surface areas of the tetra-component metal

2.5. Quantification of products

Table 2
Quantitative analysis of HMF and EMF was carried out on a Thermo Physicochemical characterization of the catalysts.
Fisher UPLC (U3000). The analysis of HMF and EMF was performed
Entry Sample SBET (m2g− 1) Vpore (cm3g− 1) Daverage (nm)
using a XDB-C18 column and a DAD detector. The column temperature
was set at 30 ◦ C, and the mobile phase consisted of deionized water and 1 Zr-O-S 2.22 0.010 17.65
2 Zr-Sn-O-S 2.66 0.014 20.91
acetonitrile at a volume ratio of 85: 15 with a flow rate of 1.0 mL/min.
3 Zr-Sn-Fe-O-S 6.61 0.035 21.20
EL was analyzed by gas chromatography (GC) equipped with a DB-5 4 Zr-Sn-Al-O-S 7.38 0.020 10.69
column and flame ionization detector (FID). The temperature of the 5 Zr-Sn-Zn-O-S 3.78 0.013 14.05
injector was 270 ◦ C, the temperature of the column was maintained at 6 Zr-Sn-Cr-O-S 4.90 0.023 19.00
90 ◦ C for 3 min and then heated to 210 ◦ C with a rate of 10 ◦ C/min. The 7 Zr-Sn-Fe-Al-O-S 4.25 0.022 20.48
8 Zr-Sn-Zn-Cr-O-S 2.32 0.005 8.56
conversion of glucose and the yields of fructose, EG, and EF were

3
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

oxides based catalysts were lower than the surface areas of the corre­ spectrum of Zn 2p showed two peaks at 1022.9 eV (Zn 2p3/2) and
sponding tri-component metal oxide-based catalysts. The surface area of 1045.9 eV (Zn 2p1/2), which can be attributed to Zn2+ in the ZnO phase
Zr-Sn-Zn-Cr-O-S is similar with that of Zr-O-S, but the pore volume [44]. Cr 2p electron binding energies assigned to Cr2O3 were located at
decreased by two, mainly because some of the oxides blocked the pores. 587.36 eV for Cr 2p1/2 and 577.35 eV for Cr 2p3/2 [37]. Based on the
The Zr-Sn-Fe-Al-O-S catalyst had a higher specific surface area than Zr- above XPS results, the chemical formulas of all catalysts were shown in
Sn-Zn-Cr-O-S due to the presence of iron oxide and alumina oxide. Table S2.
The specific surface area and the average diameter were not posi­
tively correlated. The average diameter of the pores was jointly deter­ 3.1.3. Acid properties
mined by the specific surface area and the volume of the pores, and the NH3-TPD was carried out to quantify the acid sites of all the prepared
value of these diameters was usually related to the surface area-to- catalysts. The NH3-TPD profiles were illustrated in Fig. 2. Generally, an
volume ratio [38]. Thus, as the loading of metal oxides increased, the ammonia molecule was chemically absorbed on the surface of metal
pore structure severely collapsed, reducing the specific surface area and oxides in the following ways: (a) coordinated with electron-deficient
total volume of the pores. For the Zr-Sn-Fe-Al-O-S and Zr-Sn-Zn-Cr-O-S metal atom (Lewis acid site); (b) reacted with acidic OH group
catalysts containing four metal oxides, the difference in the average (Brønsted acid sites) to generate ammonium cation; and (c) dissociated
diameters of their pores was induced by the different characteristics of in the form of –NH2 or = NH ad-species and attached on the catalysts,
Al2O3, Fe2O3, ZnO, and Cr2O3. mainly in the center of Lewis acid [45]. In short, it was commonly
believed that mode (a) was dominant way for chemical adsorption, type
3.1.2. Catalyst structure and chemical state (b) and (c) contributed less than type (a). It can be seen that the profiles
The X-ray diffraction patterns of all prepared catalysts were shown in of the catalysts containing three or four metal elements displayed more
Fig. 1. The XRD patterns essentially displayed the same diffraction peaks obvious peaks, compared with those of the catalysts containing one or
at 2θ = 30.40◦ , 35.34◦ , 50.86◦ , 60.29◦ , 63.35◦ , and 74.84◦ , which were two metal elements. Zr-O-S and Zr-Sn-O-S showed similar profiles, with
assigned to the (0 1 1), (1 1 0), (0 2 0), (1 2 1), (2 0 2), (2 2 0) planes, relatively small peaks attributed to the medium-strong acids and of weak
respectively, of ZrO2 (JCPDS#50–1089). In addition, the XRD pattern of acids. Zr-Sn-Fe-O-S, Zr-Sn-Al-O-S, Zr-Sn-Zn-O-S, and Zr-Sn-Cr-O-S
Zr-O-S showed a series of peaks at 2θ = 28.20◦ , 30.20◦ , 31.42◦ , 34.54◦ , exhibited intense desorption peaks around 580, 707, 705, and 605 ◦ C,
50.22◦ , 60.10◦ , implying that ZrO2 in this catalyst was a mixture of which indicated the existence of strong acid sites on the surface of these
monoclinic (2θ = 28.20◦ , 31.42◦ ), tetragonal (2θ = 30.20◦ , 34.54◦ ) and catalysts. For catalysts containing four metal elements, the curves of Zr-
cubic (2θ = 50.22◦ , 60.10◦ ) phases [30,39]. The tetragonal and cubic Sn-Fe-Al-O-S and Zr-Sn-Zn-Cr-O-S displayed desorption peaks at around
phases of ZrO2 were also identified in other catalysts. However, no 770 and 630 ◦ C, respectively, implying the existence of strong and super
obvious diffraction peaks typical for other metal oxides were observed, strong acid sites. Furthermore, the peaks at low temperature (<230 ◦ C)
suggesting these phases were fully homogeneously dispersed with ZrO2 might correspond to the adsorption of NH3 on the surface hydroxyl
[4,40]. The species of Sn4+, Fe3+, Al3+, Zn2+, Cr3+ detected in XPS groups, as mentioned in previous report [20]. The S content of catalyst
measurement confirmed the co-existence of multi-metals (Figure S5). increased from 8.7% of Zr-O-S catalyst to 14.1% of Zr-Sn-Zn-Cr-O-S
XPS characterized the chemical composition and content of catalysts. catalyst (Table S2), which was in agreement with the increased acid
Specifically, Zr was the dominant element of the catalysts. The Zr 3d strength from mono-component metal oxide based catalyst to tetra-
spectrum displayed peaks at binding energies of 182.5 and 184.9 eV, component metal oxides based catalyst.
corresponding to Zr 3d5/2 and Zr 3d3/2, respectively. This is in agree­ Quantification of the acid sites of the catalyst samples were shown in
ment with reported data for Zr (IV) in ZrO2 [40]. The XPS spectrum of Sn Table 3. The total acidity of Zr-O-S and Zr-Sn-O-S catalysts was of 0.88
3d displayed peaks at binding energies of 495.6 and 487.1 eV, indicating and 0.84 mmol/g, respectively, which was obviously lower than those of
the existence of Sn 3d3/2 and 3d5/2 in accordance with SnO2 materials other catalysts. The total acidity of the catalysts containing three metal
[41]. The XPS spectrum of Fe 2p showed two major peaks centered at elements followed the order of Zr-Sn-Al-O-S < Zr-Sn-Fe-O-S < Zr-Sn-Cr-
711.6 and 725.2 eV, corresponding to Fe2O3 [42]. The Al 2p spectrum O-S < Zr-Sn-Zn-O-S, in line with the amount of strong acid sites. The
showed a peak located at 74.9 eV, attributed to Al2O3 [43]. The total acidity of Zr-Sn-Cr-O-S was 2.37 mmol/g, of which 2.00 mmol/g
corresponded to strong acidity. The total acidity of Zr-Sn-Fe-Al-O-S and

Fig. 1. XRD patterns of (a) Zr-O-S, (b) Zr-Sn-O-S, (c) Zr-Sn-Fe-O-S, (d) Zr-Sn-Al- Fig. 2. NH3-TPD profiles of (a) Zr-O-S, (b) Zr-Sn-O-S, (c) Zr-Sn-Fe-O-S, (d) Zr-
O-S, (e) Zr-Sn-Zn-O-S, (f) Zr-Sn-Cr-O-S, (g) Zr-Sn-Fe-Al-O-S, (h) Zr-Sn-Zn-Cr- Sn-Al-O-S, (e) Zr-Sn-Zn-O-S, (f) Zr-Sn-Cr-O-S, (g) Zr-Sn-Fe-Al-O-S, (h) Zr-Sn-Zn-
O-S. Cr-O-S.

4
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

Table 3
Acidity of the catalysts measured by NH3-TPD.
Entry Catalysts Weak Moderate Strong and Total
acidity acidity superstrong acidity
(mmol/g) (mmol/g) acidity(mmol/g) (mmol/g)

1 Zr-O-S 0.04 0.22 0.62 0.88

2 Zr-Sn-O-S 0.02 0.31 0.51 0.84
3 Zr-Sn-Fe- 0.08 0.35 1.33 1.76
O-S
4 Zr-Sn-Al- 0.07 0.47 1.08 1.62
O-S
5 Zr-Sn-Zn- 0.16 0.21 2.00 2.37
O-S
6 Zr-Sn-Cr- 0.07 0.22 1.76 2.05
O-S
7 Zr-Sn-Fe- 0.09 0.30 1.11 1.50
Al-O-S
8 Zr-Sn-Zn- 0 0.08 1.74 1.82
Cr-O-S

Zr-Sn-Zn-Cr-O-S was 1.50 and 1.82 mmol/g, respectively. Tri- and tetra-
metal component catalysts possess more than twice total acid amount as
Zr-O-S or Zr-Sn-O-S, which was consistent with the literature that the
acid amounts of the single oxides were generally low, the introduction of
other metal elements would generate more acid sites by an excess of
positive or negative charges in the complex oxides due to the different
valences and coordination numbers of the elements [46]. However, the
acid amount was not limitless increase as the number of metal elements
increased, even appeared slight decrease for Zr-Sn-Fe-Al-O-S catalyst,
similar results were observed in previous reports. [45].
FT-IR spectra of adsorbed pyridine was used to determine the nature
of acid sites, i.e., Brønsted (denoted as B) or Lewis (denoted as L) type,
and the results were displayed in Fig. 3. The bands around 1540 indi­
cated the presence of Brønsted acid sites, as well as the band at 1640
cm− 1. The bands at 1440, 1570, and 1610 cm− 1 were assigned to Lewis
acid sites. In addition, the band at 1490 cm− 1 corresponded to pyridine
adsorbed on both Brønsted and Lewis acid sites [23,47]. The amount of
Brønsted and Lewis acid sites and ratio between them were listed in
Table S3.
Apparently, both Lewis and Brønsted acid sites populated the sur­
faces of the catalysts. The Lewis acid site was attributed to the presence
of an excess of the positive charge, the sulfate group induced the
removal of electrons from metal elements and produced more Lewis
acidity [34]. Regarding the origin of Brønsted sites, it was assumed to
appear to associate with oxygen to keep electric neutrality [46], and it
was also possible that Brønsted sites were formed during the thermal
desorption of NH3 from (NH4)2SO4 during the calcination process
[26,48]. Specifically, the Lewis acidity increased with metal element
from Zr-O-S catalyst (single metal element) to Zr-Sn-Fe-Al-O-S and Zr-
Sn-Zn-Cr-O-S catalysts (four metal elements), which was assigned to
an excess of charges due to the addition of metal elements.

3.2. Catalytic results

3.2.1. Glucose conversion into EMF over the catalysts

DMSO is a widely used polar solvent for the transformation of car­
bohydrates into furan derivative chemicals due to its acceptable price,
Fig. 3. The adsorbed pyridine-IR spectra of (a) Zr-O-S, (b) Zr-Sn-O-S, (c) Zr-Sn-
competitive solubility for intermediates, and excellent thermal stability. Fe-O-S, (d) Zr-Sn-Al-O-S, (e) Zr-Sn-Zn-O-S, (f) Zr-Sn-Cr-O-S, (g) Zr-Sn-Fe-Al-O-S,
Based on our preliminary experiments, the synthetic catalysts were (h) Zr-Sn-Zn-Cr-O-S.
evaluated for their performance in the production of EMF from glucose
at 160 ◦ C in the DMSO/ethanol bi-solvent system. The yields obtained
Zr-Sn-Al-O-S, with the Zr-Sn-Al-O-S catalyst leading to a yield of 19.8%.
after 12 h of reaction were shown in Table 4. When Zr-O-S and Zr-Sn-O-S
For the quaternary metal oxides-based catalysts, Zr-Sn-Fe-Al-O-S was
were used as catalysts, the yield of EMF was only 4.1 and 4.0%,
more active than Zr-Sn-Zn-Cr-O-S, with EMF and EL yields of 33.1 and
respectively, which was due to the relatively low amount of acid sites on
8.6%, respectively, where EL is the downstream product of EMF.
the surfaces of these two catalysts. When the ternary metal oxides-based
Isomerization is considered as the most difficult step in the synthesis of
catalysts were used, the yield of EMF increased obviously. The EMF yield
EMF from glucose. Compared with other catalysts, Zr-Sn-Fe-Al-O-S had a
increased in the order of Zr-Sn-Zn-O-S < Zr-Sn-Cr-O-S < Zr-Sn-Fe-O-S <

5
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

Table 4 Table 5
Screen of catalysts for the formation of EMF from glucose. The effect of co-solvents added on the conversion of glucose in ethanol.
Entry Catalyst Product yield (%) Entry Solvent Product yield (%)
HMF EMF EL HMF yield EMF yield EL yield

1 Zr-O-S 0.1 4.1 ND 1a 2.0 mL ethanol 0.1 0.7 ND

2 Zr-Sn-O-S 0.2 4.0 0.7 2b 2.0 mL ethanol 0.1 3.9 0.8
3 Zr-Sn-Fe-O-S ND 11.7 ND 3a 1.0 mL GBL, 1 mL ethanol 6.5 ND 1.3
4 Zr-Sn-Al-O-S 0.1 19.8 ND 4a 1.0 mL GVL, 1 mL ethanol 4.9 ND 1.3
5 Zr-Sn-Zn-O-S 0.5 10.0 ND 5a 1.0 mL H2O, 1 mL ethanol 0.1 14.2 0.2
6 Zr-Sn-Cr-O-S 0.2 11.3 ND 6a 1.0 mL THF, 1 mL ethanol 0.1 5.6 0.9
7 Zr-Sn-Fe-Al-O-S 0.7 33.1 8.6 7a 1.0 mL DMSO, 1 mL ethanol 0.4 33.1 8.6
8 Zr-Sn-Zn-Cr-O-S ND 11.3 6.7 8a 1.0 mL Toluene, 1 mL ethanol 1.5 5.0 ND

Reaction condition: 20 mg glucose, 5 mg catalyst, solvent: ethanol + DMSO = 1 Reaction condition: 20 mg glucose, 5 mg Zr-Sn-Fe-Al-O-S, 2 mL solvent, 160 ◦ C,
+ 1 mL, 160 ◦ C, 12 h. ND: not detected. (a) 12 h, (b) 20 h. ND: not detected.

strong acidity of 1.11 mmol/g while the results of pyridine FTIR showed from glucose in our reaction system. Although it is reported that unlike
that it has abundant Lewis acid sites, which promoted the isomerization ethanol, water is an inefficient solvent for dehydration of fructose [53],
of glucose into fructose (and isomerization of EG to EF) [15]. 14.2% EMF was obtained when 1 mL of water was added to the ethanol.
Among the ternary/quaternary metal oxides-based catalysts, Zr-Sn- When THF and toluene were used as co-solvents, the yield of EMF was
Zn-O-S with 2.00 mmol/g strong acidic sites and a total acidity of only about 5%. The addition of DMSO improved the EMF yield to 33.1%,
2.37 mmol/g led to the lowest yield of EMF from glucose. It could be indicating DMSO as a better co-solvent for the formation of EMF, which
thus stated that the amount of acid sites was not the predominant factor was in accordance with previous studies [54–56]. The trend of DMSO ≫
affecting the reaction process, but a synergy of acidic properties was water > THF ≈ toluene on EMF yield appeared to be related to the
needed to change the profile of the product distribution. Therefore, it polarity and solubility of co-solvent. According to Guo et al., co-solvents
can be concluded that the quantity of acid sites was not the predominant having higher polarity than ethanol were readily protonated and acted
factor affecting the reaction process. Acid strength and the B/L site ratio as the key catalytic protonation species [57]. Although the polarity and
played a more vital role in changing the profile of the product distri­ solubility of water ranked first among the four co-solvents, DMSO
bution. On the other hand, the results of XRD and N2 physisorption exhibited better performance. Solvation modeling pointed out that the
showed similar crystalline phase and surface area, respectively, for Zr- glucose conversion to HMF (intermediate for EMF formation) proceeded
Sn-Fe-Al-O-S and Zr-Sn-Zn-Cr-O-S catalysts. Yet, these two samples do through the isomerization to fructose, which was thermodynamically
not exhibit similar acidity. Thus, Zr-Sn-Fe-Al-O-S had more appropriate favored in DMSO than in water. The rationale is the Gibbs free energies
strong/weak acidity and B/L ratios than Zr-Sn-Zn-Cr-O-S. A low ratio of of glucose conversion to HMF in DMSO and water, which are − 34.40
weak Brønsted acid sites to weak Lewis acid sites and a high ratio of and − 24.13 kcal/mol at 25 ◦ C, respectively. When the temperature
strong Brønsted acid sites to strong Lewis acid sites were beneficial for increased to 120 ◦ C, the free energy of glucose conversion to HMF in
the conversion of glucose into EMF. DMSO decreased to − 39.84 kcal/mol, but the free energy of the reaction
To note, considerable EL (the downstream product of EMF) was in water dropped slightly to − 26.13 kcal/mol, indicating the reaction in
obtained only in the presence of quaternary metal oxides/ammonium water requires a higher temperature [58]. Thus, it was reasonable that
salt catalyst, indicating the superiority of quaternary metal oxides based DMSO was the best co-solvent among those tested in this study, and the
solid acids in these whole cascade ethanolysis reactions. Generally, the effect of DMSO proportion on the production of EMF from glucose was
efficiency of these catalysts on ethanolysis of carbohydrates and total further investigated.
“EMF + EL” yield was closely related to the number of metal species in
the catalyst and followed the order of tetra- > tri- > bi- > mono- 3.2.3. Effect of DMSO content and reaction time on catalytic conversion of
component metal oxides based solid acids. Based on the above results, glucose-to-EMF catalyzed by Zr-Sn-Fe-Al-O-S
Zr-Sn-Fe-Al-O-S was selected as catalyst to evaluate the role of co- The effect of the content of DMSO in tandem with the reaction time
solvent in the conversion of glucose to EMF. was investigated for the Zr-Sn-Fe-Al-O-S catalyst (Fig. 4, a-d). The use of
DMSO had a small effect on glucose conversion, because the conversion
3.2.2. Effect of the co-solvent on the conversion of glucose to EMF reached 100% after 4 h for all the tested content of DMSO. However, it
The choice of solvent is a critical factor determining the reaction rate was obvious that the addition of DMSO changed the distribution of
and the distribution of products. Therefore, the effect of co-solvent on products. For instance, the addition of 0.2 mL DMSO maintained the
the conversion of glucose to EMF was further investigated by using Zr- yield of HMF below 1.0% even after 16 h of reaction, but the yields of
Sn-Fe-Al-O-S as catalyst. Considering toxicity, green solvents including EMF and EL increased significantly compared with those in pure
GBL, GVL, and water were selected as co-solvents. What’s more, THF, ethanol. Hence, the yields of EMF and EL reached 7.9 and 16.4%,
DMSO, and toluene were also tested as co-solvent because they were respectively, at 16 h, which was twice as high as those at 4 h. EMF was
commonly used solvents for carbohydrates conversion reactions not stable and could be easily converted into EL with the prolongation of
[11,12,19,49,50]. The results of the effect of these co-solvent on EMF the reaction time, as proved by the results obtained after 8 h of reaction.
yield were shown in Table 5. When pure ethanol was used as solvent at When 0.5 mL DMSO was added, the maximum yields of EMF and EL
160 ◦ C, the yield of EMF was 0.7% with a trace amount of HMF after 12 h increased to 14.3 and 20.0%, respectively.
of reaction. Even when the reaction time was prolonged to 20 h, only Compared with pure ethanol, the addition of 0.2 and 0.5 mL DMSO
3.9% EMF yield was obtained. Abundant EG was detected, indicating promoted the conversion of glucose into EMF and EL. Further increase in
that the catalyst could promote the reaction of glucose with ethanol to the concentration of DMSO to 1.0 mL resulted in an evident increase in
form EG, whereas EG could hardly be converted into EMF under these the yield of EMF before 10 h. Hence, the EMF yield reached a maximum
reaction conditions, as previously reported [25,50,51]. of 33.1% at 12 h and then began to decline. However, at this amount of
Green solvents (such as gamma-valerolactone) are often investigated DMSO, the yield of EL decreased significantly while HMF yield
in catalytic systems [49,52]. However, no EMF was detected when GBL increased. An increase in DMSO content to 1.5 mL caused a rapid
or GVL were used as the co-solvent in a binary solvent-ethanol system, decrease in EMF yield and an increase in HMF yield. The yields of EMF
indicating that GBL and GVL were inefficient solvents for EMF formation and EL decreased to 12.4 and 5.6%, respectively, while the HMF yield

6
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

Fig. 4. (a-d) Effect of DMSO content on the conversion of glucose over time. (a) 0.2 mL DMSO and 1.8 mL ethanol, (b) 0.5 mL DMSO and 1.5 mL ethanol, (c) 1.0 mL
DMSO and 1.0 mL ethanol, (d) 1.5 mL DMSO and 0.5 mL ethanol. Reaction condition: 20 mg glucose, 5 mg Zr-Sn-Fe-Al-O-S, 160 ◦ C. (e-f) Effect of reaction tem­
perature on the conversion of glucose over time. (e) 150 ◦ C, (f) 140 ◦ C. Reaction condition: 20 mg glucose, 5 mg Zr-Sn-Fe-Al-O-S, 1.0 mL DMSO and 1.0 mL ethanol.

reached 8.5% at 16 h. These results suggested that a certain amount of formation of EMF from glucose, with maximum EMF of 18.8% at 140 ◦ C
DMSO could promote EMF formation, but an excessive amount of DMSO increased to 33.1% at 160 ◦ C. Meanwhile, the similar results were
promoted the production of HMF instead of EMF and EL, which was in occurred that prolonged time accelerated the EMF conversion to EL and
line with previous reports [2,21]. by-products even in relative lower temperature (150 ◦ C and 140 ◦ C).
The effect of reaction temperature had investigated at 160 ◦ C, 150 ◦ C Thus, the optional reaction time was controlled at 12 h.
and 140 ◦ C, and the results were shown in Fig. 4 (c, e, f). It was
apparently observed that the higher temperature promoted the

7
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

3.2.4. The effect of the catalyst amount on the EMF yield Table 6
The influence of the amount of Zr-Sn-Fe-Al-O-S, ranging from 3 to 20 EMF production from glucose using various catalysts.
mg, on the production of EMF from glucose was investigated, and the Entry Catalyst Reactionconditions glucose EMF References
results were shown in Fig. 5. As a general observation, it could be noted /catalyst yield/
that this factor did not affect the conversion rate remarkably, a glucose mass ratio %
conversion of nearly 100% being obtained irrespective of the catalyst 1 Zr-Sn-Fe-Al- 160 ◦ C, 12 h 4 33.1 This work
loading. Yet, an effect on the product yield was noticed. Using 3 mg of O-S
catalyst produced 23.5% EMF and 12.6% EL. The maximum EMF yield 2 MIL-101- 130 ◦ C, 15 h 1.8 < 10 Ref. [19]
SO3H (1 0 0)
of 33.1% was obtained when 5 mg catalyst were used for the reaction. 3 Glu-Fe3O4- 140 ◦ C, 48 h 2 27 Ref. [59]
Additionally, the yield of EL decreased to 8.6% at this amount of cata­ SO3H
lyst. When the catalyst dose continued to increase, the yields of EMF and 4 H-USY (6) 131 C, 24 h
◦
1.5 26 Ref. [50]
EL decreased due to the increase in available active sites favorable for +
Amberlyst-
the formation of by-products [40]. At 20 mg of Zr-Sn-Fe-Al-O-S, the
15
yields of EMF and EL decreased to 20.4 and 5.4%, respectively. 5 Lys/PW (2) 120 ◦ C, 15 h 3.6 < 10 Ref. [1]
Regarding HMF, the yield was constantly kept below 2% due to the
facile transformation of HMF into EMF. Meanwhile, our catalytic system
was compared with other reaction system, and the results were listed in
dCEMF
Table 6. The EMF yield obtained with Zr-Sn-Fe-Al-O-S catalyst was = k1 CGLU − (k3 + k4 )CEMF = k1 CGLU − kEMF CEMF (4)
dt
higher than those of catalysts and the glucose/catalyst mass ratio
reached 4 in our catalytic system, indicating that per unit weight of Zr- dCEL
= k3 CEMF (5)
Sn-Fe-Al-O-S catalyst exhibited remarkable catalytic ability. Further­ dt
more, the raw materials used for synthesizing Zr-Sn-Fe-Al-O-S are easy
where
to be obtained and the preparative method is convenient.
k1 + k2 = kGLU (6)
3.2.5. Determination of kinetic and thermodynamic parameters for the
formation of EMF from glucose. k3 + k4 = kEMF (7)
In the process of producing EMF from glucose, humins and other Based on the Arrhenius plots (Eq. (8)), the reaction rate constants are
unidentified soluble products are inevitably formed. Meanwhile, EMF is applied to acquire the value of activation energies (Ea) and pre-
likely to be transformed to EL under long reaction time and high tem­ exponential factors (A) of each reaction. Besides, the Eyring equation
perature. To better understand the reaction, a kinetic modeling and (Eq. (9)) is used to attain thermodynamic parameter. The Gibbs free
thermodynamic assessment are put forward. energy of activation is determined from Eq. (10). From the equations,
As shown in Figure S6, the cascade reactions consist of four reaction ΔH, ΔS and ΔG represent activation enthalpies, activation entropies,
steps include (1) glucose conversion to EMF, (2) glucose transformation and Gibbs free energy of activation, respectively. Additionally, T is re­
to by-products, (3) EMF conversion to EL, (4) EMF degradation to by- action temperature (K), kB is the Boltzmann’s constant (1.381 × 10− 23
products. Based on this model, the concentration of glucose, EMF and J⋅K− 1), h is Plank’s constant (6.626 × 10− 34 J⋅s), and R is gas constant
EL as a function of time are represented as Eq. (3–7). CGLU , CEMF and CEL (8.314 J⋅mol− 1⋅K− 1).
refer to glucose concentration, EMF concentration and EL concentration,
respectively, g/L. k1 , k2 , k3 , k4 , kGLU and kEMF are the reaction rate co­ lnk = lnA −
Ea
(8)
efficients of reaction (1), 2, 3, 4, glucose conversion, EMF conversion, RT
respectively, h− 1. ( ) ( )
k kB ΔS ΔH
ln = ln + − (9)
dCGLU T h R RT
= − (k1 + k2 )CGLU = − kGLU CGLU (3)
dt
kh
ΔG = − RTln( ) (10)
TkB
The kinetic parameters of glucose transformation were evaluated by
MATLAB using non-linear least squares regression analyses and results
were presented in Table 7, the concentrations of the products were
calculated based on the data in Fig. 4 (c, e, f). The reaction rate constants
were closely related to temperature. With temperature increasing, the k
value increased which indicated higher temperature enhanced the
glucose decomposition. The rate constants of glucose conversion to EMF
(k1) and EMF to EL (k3) were higher compared with others (k2 and k4),
implying that reaction (1) and (3) were favorable to happen. The values
of kGLU and kEMF were the constants of glucose decomposition and EMF
transformation, respectively. However, k2 and k4 accounting for non-
negligible ration of kGLU and kEMF confirmed the fact that only a frac­
tion of glucose was converted to EMF and esterified to EL, while the
other parts were transformed to side-products.
As shown in Figure S7, the Arrhenius plots were used to obtain the
value of activation energies (Ea) and pre-exponential factors (A) for
glucose conversion catalyzed by Zr-Sn-Fe-Al-O-S catalyst in the ethanol/
DMSO solvent system, and the results were presented in Table 7. The
Fig. 5. Synthesis of EMF from glucose with different amount of Zr-Sn-Fe-Al-O-
higher Ea value represented that reaction was strongly affected by
S. Reaction conditions: 20 mg glucose, solvent: ethanol + DMSO = 1 + 1 mL, temperature, and vice versa. The lowest activation energy (52.01 kJ/
160 ◦ C, 12 h. mol) existed in EMF conversion into EL (Reaction (3)), implying that

8
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

Table 7
Kinetic and thermodynamic parameters of glucose decomposition over Zr-Sn-Fe-Al-O-S catalyst.
1 1 1 1 1 1
T/◦ C k1/h− k2/h− k3/h− k4/h− kGLU/h− kEMF/h−

140 0.01506 0.01023 0.0134 0.01316 0.02529 0.02656

150 0.02367 0.01311 0.0211 0.0157 0.03678 0.03680
160 0.03382 0.02678 0.02692 0.03407 0.06060 0.06099
Parameter Reaction
1 2 3 4 GLU EMF
Ea (kJ•mol− 1) 60.23 71.30 52.01 70.40 64.92 61.72
A (h− 1) 6.28 × 105 9.78 × 106 5.21 × 104 9.47 × 106 3.99 × 106 1.64 × 106
ΔH (kJ•mol− 1) 56.71 67.79 48.49 66.88 61.41 58.21
ΔS (J•mol− 1•K− 1) − 77.06 − 54.47 − 97.99 − 54.74 − 61.92 − 69.32
ΔG (kJ•mol− 1) 89.23 91.31 89.63 90.67 87.68 87.68

reaction was relatively easy to happen. The Ea of reaction (2) and re­ from cellobiose achieved only < 10% EMF yield. The results indicated
action (4) were 71.30 and 70.40 kJ/mol, respectively, which was higher the high performance of Zr-Sn-Al-Fe-O-S catalyst for cellobiose conver­
than reaction (1) and reaction (3). It also meant that more by-products sion into EMF in the ethanol/DMSO solvent system. Interestingly, the
would be generated as the temperature increased. Meanwhile, as illus­ conversion efficiency of starch into EMF was comparable to that of
trated in Figure S8, the Eyring plot was applied to determine the acti­ cellobiose or sucrose, with a 17.9% EMF yield. Furthermore, when cel­
vation parameters including activation enthalpies (ΔH), activation lulose was used as raw material, the yields of EMF and EL were of 4.1
entropies (ΔS), and Gibbs free energy of activation (ΔG), and the results and 4.6%, respectively. The depolymerization and decrystallization of
were calculated and listed in Table 7. The lower ΔH (48.49 kJ/mol, cellulose are much challenging [15], therefore harsher reaction condi­
Reaction (3)) signified the easy formation of EL, while higher ΔH (Re­ tions (high temperature and pressure) are often required to convert
action (2) and (4)) indicated complexity in by-products formation. The cellulose into EMF [60].
higher ΔH, the more energy provided to attain the transition state, Besides biomass-derived carbohydrates, starch-enriched food wastes
which was in accordance with Ea values from the Arrhenius plots. The were used as alternative feedstocks for EMF production. In this study,
extent of disorder between ground state and transition state in a reaction the conversion of kitchen waste, expired potato chips, and expired
was usually measured by ΔS. According to the second law of thermo­ instant noodles into EMF was also studied. The results showed that the
dynamics, lower ΔS was related to the ease of reaction and vice versa. In yields of EMF from these sources were 17.4, 17.7, and 12.1%, respec­
the present study, ΔS was observed to be lower for glucose to EMF and tively (Entry 8 – 10), clearly showing the potential of Zr-Sn-Al-Fe-O-S as
EMF into EL (Reaction (1) and (3)) when compared with glucose and an effective solid acid catalyst for the transformation of real food wastes
EMF to by-products (Reaction (2) and (4)), indicating less disorder and into EMF in ethanol/DMSO solvent.
more proper environment for EMF synthesis. In addition, the smaller
ΔG, the more feasibility and ease of a reaction. Thus, the glucose con­ 3.2.7. Reaction pathways for the conversion of glucose
version to EMF and EMF transformation to EL presented more feasibility The transformation of glucose into EMF over Zr-Sn-Fe-Al-O-S cata­
as compared to glucose and EMF to side-products. lyst in DMSO/ethanol solvent involves a multi-step cascade reaction.
Several possible reaction pathways for EMF production from glucose
3.2.6. Conversion of various substrates into EMF have been already proposed [18,40,50,61,62], but the reaction chem­
The use of Zr-Sn-Al-Fe-O-S catalyst was further extended for the istry of glucose conversion into EMF may be different depending on the
conversion of other carbohydrates, including disaccharides (sucrose and catalytic system. The results obtained in this work suggested the reac­
cellobiose), polysaccharides (inulin, starch and cellulose), and three tion pathways illustrated in Scheme 1.
types of starchy food waste, in ethanol/DMSO solvent. The results were The blank experiment demonstrated that catalyst is important for
summarized in Table 8. When inulin was used as starting material, the conversion of glucose into EMF. Only 2.5% EMF was detected without
total product yield of EMF and EL reached 41.7%, with 26.3% EMF and catalyst (Table S4, Entry 1). EG was generally formed by the conden­
15.4% EL. For sucrose, the most abundant and cheapest disaccharide sation reaction of C1-hydroxyl group in glucose and hydroxyl groups in
[11], the yield of EMF reached 20.1%. It was noteworthy that consid­ ethanol. Then, EG isomerized to EF, followed by dehydration to form
erable yields of EMF between 16.4 and 22.3% were obtained from EMF in the presence of Brønsted acid sites. It was reported that EG was
cellobiose (Entries 2–4). These results were encouraging because most stable in the reaction system and maintained unreactive for EMF for­
of the reported heterogeneous acid catalytic systems for EMF production mation [50]. However, 8.5% EMF was obtained in our catalytic system
when EG was used as starting material (Table S4, Entry 2), suggesting a
“glucose – EG – EF – EMF” route (pathway 1) in our reaction system.
Table 8 Another possible route was that the glucose isomerized into fructose in
The product yields from various substrates catalyzed by Zr-Sn-Al-Fe-O-S. the presence of Lewis acid sites. Subsequently, fructose underwent
Entry Substrate Product yields (%) downstream reactions to produce EMF. Fructose and HMF as substate
HMF EMF EL afforded 31.8 and 94.2% EMF yield in our reaction system, respectively
1a Inulin ND 26.3 15.4 (Table S4, Entry 3–4). Meanwhile, HMF and EF were detected in our
2a Cellobiose 2.1 22.3 1.3 reaction system, indicating that the “glucose – fructose – HMF - EMF”
3b Cellobiose 0.2 16.4 1 (pathway 2a) and “glucose-fructose- EF-EMF” (pathway 2b) routes co-
4c Cellobiose 0.2 18.3 1.5
5a Sucrose ND 20.1 ND
exist with pathway 1.
6a Starch 1.6 17.9 5.2 Notably, the addition of DMSO could not only improve the yield of
7a Cellulose ND 4.1 4.6 EMF and prevent the formation of EL from EMF, but also changed the
8a Kitchen waste 0.2 17.4 6.2 reaction pathway. When the proportion of DMSO in ethanol was low, the
9a Expired potato chips 0.2 17.7 ND
glucose preferred to convert to EG directly, and then dehydrated to form
10a Expired instant noodles 0.2 12.1 ND
EMF through the pathway 1. As the concentration of DMSO increased,
Reaction conditions: (a) 20 mg substrate, 5 mg catalyst; (b) 40 mg substrate, 5 the glucose was isomerized to fructose and then involved in the for­
mg catalyst; (c) 40 mg substrate, 10 mg catalyst. Solvent: DMSO + ethanol = 1 + mation of EMF (pathway 2). The adjustment of co-solvent amount
1 mL, 160 ◦ C, 12 h. ND: not detected.

9
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

Scheme 1. Possible reaction pathways of EMF formation from glucose in DMSO/ethanol solvent system catalyzed by Zr-Sn-Fe-Al-O-S.

affected the affinity of the substrates for the solvent system, which then
determined the trend of the reaction. The substrates investigated herein
had a high affinity for the solvents with high polarity rather than for
ethanol [57]. However, continuous increase in the amount of DMSO was
not feasible since after a limit, the yield of EMF was no longer improved.
As mentioned in the earlier literature, ethanol acted as both reactant and
solvent in this reaction. When ethanol volume was reduced to a certain
content (25 vol%), it acted as a reactant more than a solvent [12]. The
reduction of the ethanol volume negatively affected EMF yield, which
was in agreement with our results showing an HMF yield of 8.5% and an
EMF yield below 10% when 1.5 mL DMSO (75 vol%) was used in the
reaction.

3.2.8. Reusability of Zr-Sn-Fe-Al-O-S for one-pot conversion of glucose into

EMF
Zr-Sn-Fe-Al-O-S was used to study the stability and recyclability of
the catalysts. The experiments were performed with the optimal
amount, i.e., 5 mg catalyst and 50 vol% DMSO. After each reaction run, Fig. 6. Catalyst recycling experiments of Zr-Sn-Al-Fe-O-S catalyst. Reaction
the catalyst was separated by filtration from the reaction mixture, then conditions: solvent: DMSO + ethanol = 1 + 1 mL, 160 ◦ C, 12 h. 20 mg glucose,
the recovered catalyst was dried at room temperature overnight and 5 mg catalyst.
directly reused in the next catalytic run without regeneration treatment.
As shown in Fig. 6, the yield of EMF decreased after four consecutive structure and acid sites within this temperature range. The second
rounds without intermediate regeneration steps, and the carbon balance weight loss occurring in the range of 300–400 ◦ C was assigned to the
decreased from 54.7% during the first run to 41.6% after the catalyst dehydroxylation of hydroxyl groups on the surface of the catalyst [20].
was successively used for four times. This was mainly due to the humins The weight dropped at temperatures above 550 ◦ C was likely related to
produced during the reaction and deposited on the catalyst surface. The loss of SO42- [65]. Meanwhile, a hot filtration was performed to verify
humins covered the catalytic sites and inhibited the formation of EMF, the stability of Zr-Sn-Fe-Al-O-S catalyst, and the result was shown in
resulting in a decrease in the carbon balance. This has also been reported Table S4 (Entry 5). The EMF yield was similar with that of blank
by Wiesfeld et al. [63]. It was also found that the color of catalyst surface experiment, the results once again confirmed that Zr-Sn-Fe-Al-O-S
became dark grey after catalyst recycling. The catalyst was calcined at catalyst was stable under present reaction conditions.
600 ◦ C prior to reuse in the fifth run. The EMF yield reached 33% and the
carbon balance returned to 54.4% in the fifth cycle, which was com­ 4. Conclusions
parable to the EMF yield and carbon balance obtained in the first run.
This indicates the successful reactivation of the catalyst after calcina­ In summary, a series of sulfated complex metal oxide solid acid
tion. The results demonstrated that the performance of the catalyst was catalysts were prepared for the conversion of aldose-based carbohy­
generally preserved during recycling tests, which was in agreement with drates and starchy food wastes into EMF. The synthetic route for these
the results of thermal gravimetric analysis performed for the Zr-Sn-Fe- catalysts was conveniently based on readily available raw materials and
Al-O-S catalyst (Figure S9). In the first range of weight loss, the mass an easy preparation procedure. The characterization of catalysts by XRD
of the catalyst decreased by 1.5%, which was mainly attributed to the revealed a mixed tetragonal and cubic phase of ZrO2 in which various
removal of absorbed water [64], indicating the stability of the catalyst elements (Sn, Fe, Al, Zn, Cr) were doped to obtain bi-, tri-, and tetra-

10
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

component metal oxides based catalysts. Both Brønsted and Lewis acid Appl. Catal., B 200 (2017) 192–199, https://doi.org/10.1016/j.
apcatb.2016.07.004.
sites were identified on the surface of all catalysts. Yet, the tri- and tetra-
[4] C. García-Sancho, I. Fúnez-Núñez, R. Moreno-Tost, J. Santamaría-González,
component metal oxides based catalysts displayed a higher number of E. Pérez-Inestrosa, J.L.G. Fierro, P. Maireles-Torres, Beneficial effects of calcium
strong acid sites than Zr-O-S and Zr-Sn-O-S. The catalytic results ob­ chloride on glucose dehydration to 5-hydroxymethylfurfural in the presence of
tained for the conversion of glucose in DMSO/ethanol showed that the alumina as catalyst, Appl. Catal., B 206 (2017) 617–625, https://doi.org/10.1016/
j.apcatb.2017.01.065.
metal species doped in these catalysts affected EMF formation and that [5] J. Wang, Z. Zhang, S. Jin, X. Shen, Efficient conversion of carbohydrates into 5-
the efficiency increased with the incorporation of additional metal ox­ hydroxylmethylfurfan and 5-ethoxymethylfurfural over sufonic acid-functionalized
ides from unitary metal oxide-based catalysts to quaternary metal oxide- mesoporous carbon catalyst, Fuel 192 (2017) 102–107, https://doi.org/10.1016/j.
fuel.2016.12.027.
based catalysts. Acid strength and the B/L ratio on the catalyst surface [6] H. Wu, T. Huang, F. Cao, Q. Zou, P. Wei, P. Ouyang, Co-production of HMF and
played a more critical role than the quantity of acid sites in determining gluconic acid from sucrose by chemo-enzymatic method, Chem. Eng. J. 327 (2017)
the catalyst performance for EMF production. The cooperation of Lewis 228–234, https://doi.org/10.1016/j.cej.2017.06.107.
[7] A. Bohre, M.I. Alam, K. Avasthi, F. Ruiz-Zepeda, B. Likozar, Low temperature
acid sites and Brønsted acid sites was important for the reaction. A low transformation of lignocellulose derived bioinspired molecules to aviation fuel
weak B/weak L ratio and high strong B/strong L ratio were conducive precursor over magnesium–lanthanum mixed oxide catalyst, Appl. Catal., B 276
for generating EMF. Zr-Sn-Al-Fe-O-S converted glucose to EMF with a (2020) 119069, https://doi.org/10.1016/j.apcatb.2020.119069.
[8] R. Šivec, M. Grilc, M. Huš, B. Likozar, Multiscale Modeling of (Hemi)cellulose
yield of 33.1% in a bi-solvent (DMSO: ethanol = 1: 1) system. Other Hydrolysis and Cascade Hydrotreatment of 5-Hydroxymethylfurfural, Furfural, and
biomass-derived di- and poly-saccharides were successfully converted Levulinic Acid, Ind. Eng. Chem. Res. 58 (35) (2019) 16018–16032, https://doi.
into EMF over Zr-Sn-Al-Fe-O-S catalyst in ethanol/DMSO as solvent. It org/10.1021/acs.iecr.9b00898.
[9] M.M. Cajnko, U. Novak, M. Grilc, B. Likozar, Enzymatic conversion reactions of 5-
was also showed that the content of DMSO was vital in the orientation of
hydroxymethylfurfural (HMF) to bio-based 2,5-diformylfuran (DFF) and 2,5-fur­
the reaction routes. A low amount of DMSO favored the reaction andicarboxylic acid (FDCA) with air: mechanisms, pathways and synthesis
pathway via intermediate ethyl glucoside (EG), while a high proportion selectivity, Biotechnol. Biofuels 13 (2020) 66, https://doi.org/10.1186/s13068-
of DMSO was convincible to switch the reaction pathway via interme­ 020-01705-z.
[10] L. Bing, Z. Zhang, K. Deng, Efficient One-Pot Synthesis of 5-(Ethoxymethyl)furfural
diate fructose. Meanwhile, the kinetic parameters were calculated based from Fructose Catalyzed by a Novel Solid Catalyst, Ind. Eng. Chem. Res. 51 (47)
on Arrhenius plots, and the results were in agreement with first-order (2012) 15331–15336, https://doi.org/10.1021/ie3020445.
reaction. The kinetic model revealed the glucose conversion to EMF [11] B. Liu, Z. Zhang, K. Huang, Z. Fang, Efficient conversion of carbohydrates into 5-
ethoxymethylfurfural in ethanol catalyzed by AlCl3, Fuel 113 (2013) 625–631,
and EMF transformation to EL were dominant reactions. Furthermore, https://doi.org/10.1016/j.fuel.2013.06.015.
the Eyring equation was applied to determine the thermodynamic pa­ [12] H. Wang, T. Deng, Y. Wang, Y. Qi, X. Hou, Y. Zhu, Efficient catalytic system for the
rameters. The analysis revealed that the rate of reaction increased with conversion of fructose into 5-ethoxymethylfurfural, Bioresour Technol. 136 (2013)
394–400, https://doi.org/10.1016/j.biortech.2013.02.110.
elevating temperature, and EMF conversion into EL was likely to happen [13] M.I. Alam, S. De, B. Singh, B. Saha, M.M. Abu-Omar, Titanium hydrogenphosphate:
in all reactions, indicating it was necessary to suppress the EMF trans­ An efficient dual acidic catalyst for 5-hydroxymethylfurfural (HMF) production,
formation into EL in order to obtain more EMF as target product. Zr-Sn- Appl. Catal., A 486 486 (2014) 42–48, https://doi.org/10.1016/j.
apcata.2014.08.019.
Fe-Al-O-S catalyst was reactivated by calcination and maintained a [14] E.F. Dunn, D. Liu, E.Y.X. Chen, Role of N-heterocyclic carbenes in glucose
relatively high activity after five reaction cycles. conversion into HMF by Cr catalysts in ionic liquids, Appl. Catal., A 460-461
(2013) 1-7. https://doi.org/10.1016/j.apcata.2013.04.015.
[15] J. Liu, Y. Tang, K. Wu, C. Bi, Q. Cui, Conversion of fructose into 5-hydroxyme­
Declaration of Competing Interest thylfurfural (HMF) and its derivatives promoted by inorganic salt in alcohol,
Carbohydr Res. 350 (2012) 20–24, https://doi.org/10.1016/j.carres.2011.12.006.
[16] Y. Yang, M.M. Abu-Omar, C. Hu, Heteropolyacid catalyzed conversion of fructose,
The authors declare that they have no known competing financial
sucrose, and inulin to 5-ethoxymethylfurfural, a liquid biofuel candidate, Appl.
interests or personal relationships that could have appeared to influence Energy 99 (2012) 80–84, https://doi.org/10.1016/j.apenergy.2012.04.049.
the work reported in this paper. [17] F.u. Yang, J. Tang, R. Ou, Z. Guo, S. Gao, Y. Wang, X. Wang, L. Chen, A. Yuan, Fully
catalytic upgrading synthesis of 5-Ethoxymethylfurfural from biomass-derived 5-
Hydroxymethylfurfural over recyclable layered-niobium-molybdate solid acid,
Acknowledgement Appl. Catal., B 256 (2019) 117786, https://doi.org/10.1016/j.
apcatb.2019.117786.
This work was supported by the National Natural Science Foundation [18] S. Quereshi, E. Ahmad, K.K.K. Pant, S. Dutta, Insights into Microwave-Assisted
Synthesis of 5-Ethoxymethylfurfural and Ethyl Levulinate Using Tungsten Disulfide
of China (21607119), Key Research and Development Plan of Shaanxi as a Catalyst, ACS Sustainable Chem. Eng. 8 (4) (2020) 1721–1729, https://doi.
Province-Industrial Innovation Chain Project (2020ZDLNY06-08), org/10.1021/acssuschemeng.9b0323110.1021/acssuschemeng.9b03231.s001.
Young Talents Support Program of Colleges and Universities Association [19] X. Liu, H. Li, H. Pan, H. Zhang, S. Huang, K. Yang, W. Xue, S. Yang, Efficient
catalytic conversion of carbohydrates into 5-ethoxymethylfurfural over MIL-101-
for Science and Technology of Shaanxi Province (20190420), Special based sulfated porous coordination polymers, J. Energy Chem. 25 (3) (2016)
Funds of the Education Department of Shaanxi Province (19JK0475), 523–530, https://doi.org/10.1016/j.jechem.2016.01.015.
Innovative Talents Promotion Plan-Science and Technology Innovation [20] N.A.S. Ramli, N.A.S. Amin, Fe/HY zeolite as an effective catalyst for levulinic acid
production from glucose: Characterization and catalytic performance, Appl. Catal.,
Group of Shaanxi Province (2019TD-025), Special Program in Natural B 163 163 (2015) 487–498, https://doi.org/10.1016/j.apcatb.2014.08.031.
Science of Xi’an University of Architecture and Technology (ZR19020). [21] A.A. Marianou, C.M. Michailof, A. Pineda, E.F. Iliopoulou, K.S. Triantafyllidis, A.
A. Lappas, Effect of Lewis and Brønsted acidity on glucose conversion to 5-HMF
and lactic acid in aqueous and organic media, Appl. Catal., A 555 555 (2018)
Appendix A. Supplementary data 75–87, https://doi.org/10.1016/j.apcata.2018.01.029.
[22] B. Liu, Z. Zhang, One-pot conversion of carbohydrates into 5-ethoxymethylfurfural
Supplementary data to this article can be found online at https://doi. and ethyl d-glucopyranoside in ethanol catalyzed by a silica supported sulfonic
acid catalyst, RSC Adv. 3 (30) (2013) 12313–12319, https://doi.org/10.1039/
org/10.1016/j.cej.2021.132279. c3ra41043a.
[23] J. Guo, S. Zhu, Y. Cen, Z. Qin, J. Wang, W. Fan, Ordered mesoporous Nb–W oxides
References for the conversion of glucose to fructose, mannose and 5-hydroxymethylfurfural,
Appl. Catal., B 200 (2017) 611–619, https://doi.org/10.1016/j.
apcatb.2016.07.051.
[1] H. Li, K.S. Govind, R. Kotni, S. Shunmugavel, A. Riisager, S. Yang, Direct catalytic
[24] G. Xu, B. Chen, Z. Zheng, K. Li, H. Tao, One-pot ethanolysis of carbohydrates to
transformation of carbohydrates into 5-ethoxymethylfurfural with acid–base
promising biofuels: 5-ethoxymethylfurfural and ethyl levulinate, Asia-Pac. J.
bifunctional hybrid nanospheres, Energy Convers Manage. 88 (2014) 1245–1251,
Chem. Eng. 12 (4) (2017) 527–535, https://doi.org/10.1002/apj.v12.410.1002/
https://doi.org/10.1016/j.enconman.2014.03.037.
apj.2095.
[2] J. Zhang, K. Dong, W. Luo, H. Guan, Catalytic upgrading of carbohydrates into 5-
[25] C.M. Lew, N. Rajabbeigi, M. Tsapatsis, One-Pot Synthesis of 5-(Ethoxymethyl)
ethoxymethylfurfural using SO3H functionalized hyper-cross-linked polymer based
furfural from Glucose Using Sn-BEA and Amberlyst Catalysts, Ind. Eng. Chem. Res.
carbonaceous materials, Fuel 234 (2018) 664–673, https://doi.org/10.1016/j.
51 (14) (2012) 5364–5366, https://doi.org/10.1021/ie2025536.
fuel.2018.07.060.
[26] Y. Goto, K.-I. Shimizu, K. Kon, T. Toyao, T. Murayama, W. Ueda, NH3-efficient
[3] B. Chen, F. Li, Z. Huang, G. Yuan, Carbon-coated Cu-Co bimetallic nanoparticles as
ammoxidation of toluene by hydrothermally synthesized layered tungsten-
selective and recyclable catalysts for production of biofuel 2,5-dimethylfuran,

11
Y. He et al. Chemical Engineering Journal 429 (2022) 132279

vanadium complex metal oxides, J. Catal. 344 (2016) 346–353, https://doi.org/ [47] Qidong Hou, Meinan Zhen, Le Liu, Yu Chen, Fang Huang, Shiqiu Zhang, Weizun Li,
10.1016/j.jcat.2016.10.013. Meiting Ju, Tin phosphate as a heterogeneous catalyst for efficient dehydration of
[27] K.K.M. Belokopytov Y V, Gerei S V. , An infrared study of the surface properties of glucose into 5-hydroxymethylfurfural in ionic liquid, Appl. Catal., B 224 (2018)
metal oxides: 2. The interaction of ammonia with the surface of Fe2O3, ZnO, 183–193, https://doi.org/10.1016/j.apcatb.2017.09.049.
MoO3, and V2O5., J. Catal. 60(1) (1979) 1-7. [48] T. Murayama, N. Kuramata, W. Ueda, Hydrothermal synthesis of W-Ta–O complex
[28] Fu Yang, Xu Hu, Shuying Gao, Xianfeng Liu, Shijian Zhou, Yan Kong, Enabling metal oxides by assembling MO6 (M = W or Ta) octahedra and creation of solid
synchronous activation of inner-core and mesoporous outer-shell of monodispersed acid, J. Catal. 339 (2016) 143–152, https://doi.org/10.1016/j.jcat.2016.04.007.
Fe3O4@SiO2 by in-situ implanted MnO to synergistically deliver enhanced catalytic [49] Shanta Dutta, Iris K.M. Yu, Daniel C.W. Tsang, Zhishan Su, Changwei Hu, Kevin C.
activity, J. Alloys Compd. 842 (2020) 155817, https://doi.org/10.1016/j. W. Wu, Alex C.K. Yip, Yong Sik Ok, Chi Sun Poon, Influence of green solvent on
jallcom.2020.155817. levulinic acid production from lignocellulosic paper waste, Bioresour Technol. 298
[29] F. Yang, Y. Ding, J. Tang, S. Zhou, B. Wang, Y. Kong, Oriented surface decoration of (2020) 122544, https://doi.org/10.1016/j.biortech.2019.122544.
(Co-Mn) bimetal oxides on nanospherical porous silica and synergetic effect in [50] H. Li, S. Saravanamurugan, S. Yang, A. Riisager, Direct transformation of
biomass-derived 5-hydroxymethylfurfural oxidation, Mol. Catal. 435 (2017) carbohydrates to the biofuel 5-ethoxymethylfurfural by solid acid catalysts, Green
144–155, https://doi.org/10.1016/j.mcat.2017.03.034. Chem. 18 (3) (2016) 726–734, https://doi.org/10.1039/c5gc01043h.
[30] W. Zeng, D.-G. Cheng, F. Chen, X. Zhan, Catalytic Conversion of Glucose on Al–Zr [51] M.S. Holm, S. Saravanamurugan, E. Taarning, Conversion of sugars to lactic acid
Mixed Oxides in Hot Compressed Water, Catal. Lett. 133 (1–2) (2009) 221–226, derivatives using heterogeneous zeotype catalysts, Science 328 (5978) (2010)
https://doi.org/10.1007/s10562-009-0160-3. 602–605, https://doi.org/10.1126/science.1183990.
[31] Katsue Shibata, Tadamitsu Kiyoura, Jun Kitagawa, Takashi Sumiyoshi, [52] S. Dutta, I.K.M. Yu, D.C.W. Tsang, J. Fan, J.H. Clark, Z. Jiang, Z. Su, C. Hu, C.
Kozo Tanabe, Acidic properties of binary metal oxides, Bull. Chem. Soc. Jpn. 46 S. Poon, Efficient Depolymerization of Cellulosic Paper Towel Waste Using Organic
(10) (1973) 2985–2988. Carbonate Solvents, ACS Sustainable Chem. Eng. 8 (34) (2020) 13100–13110,
[32] I.K.M. Yu, D.C.W. Tsang, Conversion of biomass to hydroxymethylfurfural: A https://doi.org/10.1021/acssuschemeng.0c04102.
review of catalytic systems and underlying mechanisms, Bioresour Technol. 238 [53] Y. Yang, C. Hu, M.M. Abu-Omar, Conversion of glucose into furans in the presence
(2017) 716–732, https://doi.org/10.1016/j.biortech.2017.04.026. of AlCl3 in an ethanol-water solvent system, Bioresour Technol. 116 (2012)
[33] I.K.M. Yu, D.C.W. Tsang, A.C.K. Yip, S.S. Chen, Y.S. Ok, C.S. Poon, Valorization of 190–194, https://doi.org/10.1016/j.biortech.2012.03.126.
starchy, cellulosic, and sugary food waste into hydroxymethylfurfural by one-pot [54] Ryoichi Otomo, Toshiyuki Yokoi, Junko N. Kondo, Takashi Tatsumi, Dealuminated
catalysis, Chemosphere 184 (2017) 1099–1107, https://doi.org/10.1016/j. Beta zeolite as effective bifunctional catalyst for direct transformation of glucose to
chemosphere.2017.06.095. 5-hydroxymethylfurfural, Appl. Catal., A 470 470 (2014) 318–326, https://doi.
[34] P.J. M., Promotion of zirconia acidity by addition of sulfate ion., Catal. Today 15(3- org/10.1016/j.apcata.2013.11.012.
4) (1992) 481-490. [55] Hongliang Wang, Tiansheng Deng, Yingxiong Wang, Xiaojing Cui, Yongqin Qi,
[35] Jacques Védrine, Heterogeneous Catalysis on Metal Oxides, Catalysts 7 (11) (2017) Xindong Mu, Xianglin Hou, Yulei Zhu, Graphene oxide as a facile acid catalyst for
341, https://doi.org/10.3390/catal7110341. the one-pot conversion of carbohydrates into 5-ethoxymethylfurfural, Green Chem.
[36] S. Gholami, S.M. Alavi, M. Rezaei, Synthesis of Cr2O3–Al2O3 powders with various 15 (9) (2013) 2379, https://doi.org/10.1039/c3gc41109e.
Cr2O3/Al2O3 molar ratios and their applications as support for the preparation of [56] G. Morales, M. Paniagua, J.A. Melero, J. Iglesias, Efficient production of 5-ethox­
nickel catalysts in CO2 methanation reaction, Int. J. Hydrogen Energy 46 (7) ymethylfurfural from fructose by sulfonic mesostructured silica using DMSO as co-
(2021) 5311–5322, https://doi.org/10.1016/j.ijhydene.2020.11.066. solvent, Catal. Today 279 (2017) 305–316, https://doi.org/10.1016/j.
[37] A. Bumajdad, S. Al-Ghareeb, M. Madkour, F.A. Sagheer, Non-noble, efficient cattod.2016.02.016.
catalyst of unsupported alpha-Cr2O3 nanoparticles for low temperature CO [57] Haixin Guo, Alif Duereh, Yaqiong Su, Emiel J.M. Hensen, Xinhua Qi, Richard
Oxidation, Sci. Rep. 7 (1) (2017) 14788, https://doi.org/10.1038/s41598-017- Lee Smith, Mechanistic role of protonated polar additives in ethanol for selective
14779-x. transformation of biomass-related compounds, Appl. Catal., B 264 (2020) 118509,
[38] M.J. Nalbandian, M. Zhang, J. Sanchez, Y.H. Choa, J. Nam, D.M. Cwiertny, N. https://doi.org/10.1016/j.apcatb.2019.118509.
V. Myung, Synthesis and optimization of Fe2O3 nanofibers for chromate adsorption [58] V. Choudhary, R.I. Burnett, D.G. Vlachos, S.I. Sandler, Dehydration of Glucose to 5-
from contaminated water sources, Chemosphere 144 (2016) 975–981, https://doi. (Hydroxymethyl)furfural and Anhydroglucose: Thermodynamic Insights, J. Phys.
org/10.1016/j.chemosphere.2015.08.056. Chem. C 116 (8) (2012) 5116–5120, https://doi.org/10.1021/jp2113895.
[39] Jian He, Hu Li, Ye-Min Lu, Yan-Xiu Liu, Zhi-Bing Wu, De-Yu Hu, Song Yang, [59] R.S. Thombal, V.H. Jadhav, Application of glucose derived magnetic solid acid for
Cascade catalytic transfer hydrogenation–cyclization of ethyl levulinate to etherification of 5-HMF to 5-EMF, dehydration of sorbitol to isosorbide, and
γ-valerolactone with Al–Zr mixed oxides, Appl. Catal., A 510 510 (2016) 11–19, esterification of fatty acids, Tetrahedron Lett. 57 (39) (2016) 4398–4400, https://
https://doi.org/10.1016/j.apcata.2015.10.049. doi.org/10.1016/j.tetlet.2016.08.061.
[40] L. Peng, L. Lin, J. Zhang, J. Shi, S. Liu, Solid acid catalyzed glucose conversion to [60] M. Cheng, T. Shi, S. Wang, H. Guan, C. Fan, X. Wang, Fabrication of micellar
ethyl levulinate, Appl. Catal., A 397(1-2) (2011) 259-265. https://doi.org/ heteropolyacid catalysts for clean production of monosaccharides from
10.1016/j.apcata.2011.03.008. polysaccharides, Catal. Commun. 12 (15) (2011) 1483–1487, https://doi.org/
[41] C. Wang, J. Li, Y. Zhang, Y. Wei, J. Liu, The influence of phosphate on crystal grain 10.1016/j.catcom.2011.06.003.
growth of nanosized SnO2, J. Alloys Compd. 493 (1–2) (2010) 64–69, https://doi. [61] T. Flannelly, S. Dooley, J.J. Leahy, Reaction Pathway Analysis of Ethyl Levulinate
org/10.1016/j.jallcom.2009.12.131. and 5-Ethoxymethylfurfural fromd-Fructose Acid Hydrolysis in Ethanol, Energy &
[42] X.F. Lu, X.Y. Chen, W. Zhou, Y.X. Tong, G.R. Li, alpha-Fe2O3@PANI Core-Shell Fuels 29 (11) (2015) 7554–7565, https://doi.org/10.1021/acs.
Nanowire Arrays as Negative Electrodes for Asymmetric Supercapacitors, ACS energyfuels.5b01481.
Appl. Mater. Interfaces 7 (27) (2015) 14843–14850, https://doi.org/10.1021/ [62] Haixin Guo, Alif Duereh, Yuya Hiraga, Xinhua Qi, Richard Lee Smith, Mechanism
acsami.5b03126. of Glucose Conversion into 5-Ethoxymethylfurfural in Ethanol with Hydrogen
[43] Abdullah H. Alshehri, Kissan Mistry, Viet Huong Nguyen, Khaled H. Ibrahim, Sulfate Ionic Liquid Additives and a Lewis Acid Catalyst, Energy & Fuels 32 (8)
David Muñoz-Rojas, Mustafa Yavuz, Kevin P. Musselman, Quantum-Tunneling (2018) 8411–8419, https://doi.org/10.1021/acs.energyfuels.8b0071710.1021/
Metal-Insulator-Metal Diodes Made by Rapid Atmospheric Pressure Chemical acs.energyfuels.8b00717.s001.
Vapor Deposition, Adv. Funct. Mater. 29 (7) (2019) 1805533, https://doi.org/ [63] J.J. Wiesfeld, M. Kim, K. Nakajima, E.J.M. Hensen, Selective hydrogenation of 5-
10.1002/adfm.v29.710.1002/adfm.201805533. hydroxymethylfurfural and its acetal with 1,3-propanediol to 2,5-bis(hydroxy­
[44] R. Zha, R. Nadimicherla, X. Guo, Ultraviolet photocatalytic degradation of methyl methyl)furan using supported rhenium-promoted nickel catalysts in water, Green
orange by nanostructured TiO2/ZnO heterojunctions, J. Mater. Chem. A 3 (12) Chem. 22 (4) (2020) 1229–1238, https://doi.org/10.1039/c9gc03856f.
(2015) 6565–6574, https://doi.org/10.1039/c5ta00764j. [64] I. Elsayed, M. Mashaly, F. Eltaweel, M.A. Jackson, E.B. Hassan, Dehydration of
[45] E. Rombi, I. Ferino, R. Monaci, C. Picciau, V. Solinas, R. Buzzoni, Toluene glucose to 5-hydroxymethylfurfural by a core-shell Fe3O4@SiO2-SO3H magnetic
ammoxidation on α-Fe2O3-based catalysts, Appl. Catal., A 266(1) (2004) 73-79. nanoparticle catalyst, Fuel 221 (2018) 407–416, https://doi.org/10.1016/j.
https://doi.org/10.1016/j.apcata.2004.01.028. fuel.2018.02.135.
[46] Kozo Tanabe, Takashi Sumiyoshi, Katsue Shibata, Tadamitsu Kiyoura, [65] J. Pérez, E. Pérez, B. del Vas, L. García, J.L. Serrano, Analysis of (NH4)2SO4/(NH4)
Jun Kitagawa, A new hypothesis regarding the surface acidity of binary metal H2PO4 mixtures by thermogravimetry and X-ray diffraction, Thermochim Acta.
oxides, Bull. Chem. Soc. Jpn. 47 (5) (1974) 1064–1066. 443 (2) (2006) 231–234, https://doi.org/10.1016/j.tca.2006.02.003.

12