Journal of the Japan Petroleum Institute, 48, (5), 281 289 (2005) 281

[Regular Paper]

Inhibition Effect of Hydrogen Sulfide and Ammonia on NiMo/Al2O3,

CoMo/Al2O3, NiCoMo/Al2O3 Catalysts in Hydrodesulfurization of
Dibenzothiophene and 4,6-Dimethyldibenzothiophene

Hiroyuki NAKAMURA*, Masaomi AMEMIYA, and Katsuaki ISHIDA

Petroleum Refining Research and Technology Center, Japan Energy Corp., 3-17-35 Niizo-Minami, Toda, Saitama 335-8502, JAPAN

(Received August 24, 2004)

Model hydrodesulfurization catalysts (Co2.75Mo11, Ni2.75Mo11 and Ni0.9Co1.85Mo11 wt%/γ-Al2O3) pre-

pared by impregnation of similar supports with similar starting materials were used in the following test reactions:
hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene, with and without
addition of H2S or NH3. The presence of H2S or NH3 has a profound effect on the catalyst ranking and the activ-
ities of various reaction pathways, e.g., the hydrogenation (HYD) route and the direct desulfurization (DDS)
route. In the case of the HDS of DBT and 4,6-DMDBT, hydrogen sulfide only weakly poisoned CoMo and
NiCoMo catalyst, whereas a NiMo catalyst is poisoned more strongly. On the other hand, ammonia strongly
poisoned HYD route of CoMo catalyst, whereas DDS route of NiMo catalyst is not at all poisoned in the case of
the HDS of DBT. The trimetallic NiCoMo catalyst exhibited the best performance in the conversion of DBT.
A comparison of the HYD and DDS rate constants among CoMo, NiMo and NiCoMo catalysts shows that in the
presence of rich H2S, the activity of HYD and DDS of the NiCoMo catalyst equals the sum of the activity of the
HYD of NiMo catalyst and of the DDS of CoMo catalyst. On the other hand, in the presence of rich NH3, the
activity of the NiCoMo catalyst equals the sum of the HYD and DDS of NiMo catalyst, which each in turn are
higher than those of CoMo catalyst. This dual promoting effect of NiCoMo catalyst causes higher activity than
either NiMo or CoMo catalyst in the presence of either rich H2S or rich NH3.

Keywords
Hydrodesulfurization, Inhibition effect, Nickel molybdenum catalyst, Cobalt molybdenum catalyst,
Hydrogen sulfide, Ammonia

1. Introduction free diesel fuel production. In deep hydrodesulfuriza-

tion (HDS), employed to accomplish this goal, the most
The development of more active catalysts for more refractory sulfur compounds in gas oil are alkyl-
efficient hydrotreating processes is important in view of dibenzothiophenes with alkyl groups near the sulfur
upcoming oil regulations. Further tightening of diesel atom (in positions 4 and 6)1),2). The other problem is
fuel sulfur restrictions has been proposed and will be inhibition by H2S3),4) and nitrogen compounds including
implemented worldwide in the near future. In order to NH35),6). These inhibitors are generated during HDS
reduce particulate matter and NOx emission from and HDN reactions and their concentration is higher
diesel-fueled vehicles, the maximum permissible sulfur under the conditions of deep HDS.
content of diesel fuel will be 50 wt ppm by the end of Two solutions for ultra-low sulfur diesel fuel produc-
the year 2004 in Japan. Requirement of substantially tion have been developed by the authors. A simple
“sulfur free” diesel fuel (10 wt ppm or less) has been solution known as the “CoMo/NiMo Catalyst Relay”
proposed for the future. The European Commission system was described in ref.7. This solution can
proposed a mandatory “zero sulfur” restriction (10 wt achieve 50 wt ppm sulfur production without major
ppm or less) from 2011. Japan is also considering revamp of conventional deep HDS units. In this sys-
requiring “sulfur free” diesel fuel by 2007. Therefore, tem, the first part of the catalyst bed is the CoMo type,
much attention is given to effective technological solu- which is the pretreatment catalyst for ultra deep HDS,
tions for ultra-low sulfur diesel fuel, particularly sulfur- arranged over the main NiMo catalyst which is the sec-
ond part. The main function of the CoMo catalyst is
* To whom correspondence should be addressed. to achieve HDS of reactive sulfur compounds such as
* E-mail: [email protected] dibenzothiophene (DBT). The NiMo catalyst in the

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 282

second bed is the main catalyst for ultra-low sulfur

diesel fuel production and achieves HDS of refractory
sulfur compounds such as 4,6-dimethyldibenzothio-
phene (4,6-DMDBT), even in the presence of high con-
centrations of catalyst poisoning materials such as
hydrogen sulfide and ammonia.
A more effective solution is a two-stage process with
gas/liquid separation (GLSP) in the middle of the unit8).
This process has a great potential for producing sulfur-
free diesel fuel production (S = 10 wt ppm or less). Scheme 1 Reaction Scheme for HDS Reaction of DBT (from
ref. 10)
Removal of produced H2S and NH3 in the middle of the
unit accelerates HDS in the second-stage reactor here
described. Typical straight gas oil derived from tains 1000 wt ppm sulfur, in a decaline solution. The
Middle East heavy oil contains 1.2 wt% of sulfur and amount of the sulfur compounds referred to here is that
100-300 wt ppm of nitrogen compounds. During measured in the inlet of the second-stage reactor in our
ultra-deep HDS, in the first-stage reactor most of the two stage process incorporating gas/liquid separation.
nitrogen compounds are converted to NH3, and some The effect of H2S on the catalyst performance was test-
1.0% of sulfur compounds is converted to H2S. The ed by addition of 0.05 wt% of a hydrogen sulfide-gen-
authors will examine the inhibition effect of these sub- erating compound (dimethyldisulfide, DMDS, Kanto
stances on ultra-deep HDS in the second-stage reactor. Chemical Co., Inc., purity > 99%, sulfur amount corre-
We have also developed a new HDS catalyst, using sponding to 10,000 wt ppm of H2S). To explore the
the promotion effect of both Ni and Co9). The activity effect of NH3 on the catalyst activity in the HDS of 4,6-
of the catalyst on HDS and HDN is higher than that of DMDBT and DBT, 0.13 wt% of a NH3-generating
conventional NiMo and CoMo catalysts in a limited compound (tri-n-butyl amine, TBA, Kanto Chemical
range of Ni/Co ratio, 0.5 < Co/(Co + Ni) < 0.75. Co., Inc., purity > 99%, sulfur amount corresponding to
In this work, this promoter effect was examined in 300 wt ppm of NH3) was added to the reactant mixture.
more detail, special attention being paid to the inhibi- The mixed catalysts (Co(2.75)Mo(11)/Ni(2.75)Mo(11)/
tion of NH3 and H2S. The Ni-, Co- and NiCo-promot- and Ni(0.9)Co(1.85)Mo(11)/γ-Al2O3, amount in paren-
ed Mo catalysts were prepared by pore-volume impreg- theses after a metal is wt%—hereafter the same) were
nation, each in a similar support. In this way, we were pre-sulfided in-situ with a decaline solution of 2 wt%
able to clearly understand the inhibition of NH3 and DMDS at 5.0 MPa and 573 K for 8 h.
H2S on the HDS catalysts in conversion of DBT and The catalysts were tested at 563-603 K at a total
4,6-DMDBT. pressure of 5.0 MPa, weight hourly space velocity of
30 h−1, under 6 l h−1 hydrogen flow for 4,6-DMDBT
2. Experimental reactions.
For the reactions of DBT, the catalysts were tested at
2. 1. Catalyst Preparation 583-623 K at a total pressure of 5.0 MPa, weight hourly
Catalysts were prepared by successive incipient wet- space velocity of 60 h−1 under 6 l h−1 hydrogen flow.
ness impregnations of γ-Al2O3 (pore volume: 0.78 cm3 g−1, 2. 3. Catalyst Activity and Reaction Pathways
specific surface area: 380 m2 g−1) with an aqueous The reaction products were quantitatively analyzed
solution of MoO3 and Ni carbonate or Co carbonate by gas chromatography with a Shimadzu GC-17A
and with an aqueous solution of phosphoric acid. The equipped with a 60 m WCOT fused silica capillary col-
catalysts contained 11 wt% Mo and 2.75 wt% Co or umn, a CP-Sil 8 CB column (VARIAN, 0.25 mm i.d.,
2.75 wt% Ni, or 1.85 wt% Co + 0.9 wt% Ni. 0.25 µm film thickness), and a flame ionization detector
After the impregnation, each catalyst was dried in an (FID) and also identified by GC-MS (Shimadzu
oven at 393 K for 15 h and finally calcined at 773 K for 17A/QP-5000) and GC-SCD (Sulfur chemiluminescent
30 min. The catalysts were crushed and sieved to the detector, ANTEK7090 with FID-GC, Shimadzu GC2010).
desired particle size before use. A sample of 0.5 g of On the other hand, the result of sulfur stripping of the
a catalyst was diluted with 9 cm3 α-Al2O3 to achieve product oil was checked by XRF (X-ray fluorescence
continuous up-flow in a fixed bed reactor, thereby pre- ZSX101e, RIKEN) to correct the FID results.
venting incomplete catalyst wetting and bypass flow. All the HDS activities are characterized by their first
2. 2. Model Feed Experiments order reaction rate (kHDS) for DBT or 4,6-DMDBT.
The model feed consisted of a solution of either 0.20 The HDS of DBT proceeds through two different
wt% of 4,6-DMDBT (Acros Organics, purity > 95%) parallel pathways10) (Scheme 1): the so-called direct
which contains 300 wt ppm sulfur, or 0.58 wt% DBT desulfurization pathway (DDS), where a direct C_S
(Kanto Chemical Co., Inc., purity > 98%) which con- bond cleavage of the DBT molecule leads to production

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 283

of biphenyl (BP), or the hydrogenation pathway cases, and therefore, they were omitted in the calcula-
(HYD), which yields cyclohexylbenzene (CHB) by an tion of the conversion or selectivity.
initial hydrogenation of one of the aromatic rings of the We estimated the first order reaction rate constant for
DBT followed by C_S bond cleavage. the conversion of DBT or 4,6-DMDBT (kHDS) is the
The HYD and the DDS selectivity are based on the summation of the first order reaction rate constant for
product concentration ratio ((CHB)/(BP + CHB)) and the DDS route (kD) and the first order reaction rate con-
(BP/(BP + CHB)), respectively. The HDS of 4,6- stant for the HYD route (kH), kD and kH being defined as
DMDBT proceeds through similar pathways (Scheme follows,
2)11). In the 4,6-DMDBT case, the DDS pathway kD = (HYD selectivity) × kHDS (1)
leads to production of 3,3’-biphenyl (BPs) and the HYD kH = (DDS selectivity) × kHDS (2)
pathway leads to 3,3’-dimethylcyclohexylbenzene kHDS = kD + kH (3)
(CHBs) and 3,3’-dimethylbicyclohexyl (BCHs).
In this case, the HYD and the DDS selectivity are 3. Results
based on the product concentration ratio (CHBs +
BCHs)/(BPs + CHBs + BCHs) and (BPs/(BPs + CHBs 3. 1. Model Feed Experiments with 4,6-DMDBT
+ BCHs)), respectively. The first order overall reaction rate constants of the
It should be noted that an intermediate primary different catalysts for the conversion of 4,6-DMDBT
hydrogenated product, like tetrahydrodibenzothiophenes with and without additives are shown in Table 1A.
(THDBTs), and an intermediate secondary hydrogenated NiMo is by far the most active catalyst in the blank
product, like hexahydrodibenzothiophenes (HHDBTs), test. The ratio of the first order reaction rate constant
are formed along the HYD pathway. These concen- of NiMo to that of the blank test at the same tempera-
trations are very low and are not detected in the most ture strongly decreases. The first order reaction rates
of other catalysts also decrease in the presence of H2S.
The apparent activation energy of CoMo catalysts is
lower than the other catalysts (Table 2).
The presence of NH3 significantly decreases the
overall reaction rate constant for the conversion of 4,6-
DMDBT for all tested catalysts to a greater extent than
that for the blank test. Interestingly, the reaction rate
constant of CoMo catalyst significantly decreases with
addition of NH3. The apparent activation energy of
NiMo is lower than the other catalysts (Table 2).
3. 2. Model Feed Experiments with DBT
The first order overall reaction rate constants of the
different catalysts for the conversion of DBT with and
Scheme 2 Reaction Scheme for HDS Reaction of 4,6-DMDBT without additives are shown in Table 1B.
(from ref. 11) The rate constant of NiCoMo catalyst in the blank

Table 1A The First Order Rate Constants of the Conversion for 4,6-DMDBT
[h−1 g−1]

Blank Addition of H2S (+H2S) Addition of NH3 (+NH3)

Temperature
563 K 573 K 583 K 593 K 583 K 593 K 603 K 613 K 583 K 593 K 603 K 613 K
NiCoMo 19.1 23.0 32.4 46.1 16.6 31.8 53.8 76.4 14.7 25.1 40.1 79.2
NiMo 39.1 43.3 61.5 86.6 21.0 31.2 57.6 90.9 29.0 44.8 61.5 72.0
CoMo 17.5 22.6 30.4 43.9 20.6 31.3 41.2 56.3 6.6 8.4 18.1 40.9

Table 1B The First Order Rate Constants of the Conversion for DBT
[h−1 g−1]

Blank Addition of H2S (+H2S) Addition of NH3 (+NH3)

Temperature
583 K 593 K 603 K 613 K 593 K 603 K 613 K 623 K 593 K 603 K 613 K 623 K
NiCoMo 82.9 104.6 121.2 141.4 47.0 74.8 108.8 152.5 65.5 90.2 132.2 178.0
NiMo 66.7 87.2 104.8 125.2 25.9 40.7 59.2 78.4 73.0 98.8 139.6 152.3
CoMo 58.1 86.6 109.8 129.5 44.8 66.6 85.4 93.6 50.1 82.4 106.1 120.5

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 284

Table 2 Apparent Activation Energy for HDS of 4,6-DMDBT and DBT

[kcal mol−1]

4,6-DMDBT DBT
Catalyst
NiMo NiCoMo CoMo NiMo NiCoMo CoMo
Blank test 17.8 19.7 20.0 14.5 12.1 18.5
+H2S 35.0 35.7 23.1 26.9 28.2 17.9
+NH3 21.4 38.3 43.7 18.5 24.4 21.0

kH (black bar), the first order reaction rate constant for HYD route; kD (white bar), the first order reaction rate constant for DDS route;
kHDS = kH + kD. A: NiMo, B: CoMo, C: NiCoMo.

Fig. 1 The First Order Reaction Rate Constant for the Conversion of 4,6-DMDBT

test is slightly higher than those of the other catalysts, catalyst are not decreased by NH3 as much.
the rate constants of the catalysts being the almost the 3. 3. Reaction Pathway
same. The first order reaction rate constants for the conver-
The presence of H2S significantly decreases the over- sion of 4,6-DMDBT and DBT were calculated sepa-
all reaction rate constant for the conversion of DBT for rately for the HYD route (kH) and for the DDS route
all tested catalysts to a greater extent than that for the (kD).
blank test (Table 1B). NiCoMo is the most active The rate constant divided into kH and kD components
catalyst for the conversion of DBT in the presence of for each catalyst for the conversion of 4,6-DMDBT are
H2S. The apparent activation energy of CoMo is shown in Figs. 1A-1C. The HYD route (kH) is domi-
lower than those of the other catalysts (Table 2). nant in all the test results of 4,6-DMDBT.
The presence of NH3 significantly decreases the The kH and kD of each catalyst for the conversion of
activity of CoMo; the rates of the NiMo and NiCoMo DBT are shown in Figs. 2A-2C. The DDS route is

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 285

kH (black bar), HYD activity; kD (white bar), DDS activity; kHDS = kH + kD. A: NiMo, B: CoMo, C: NiCoMo.

Fig. 2 The First Order Reaction Rate Constant for the Conversion of DBT

dominant in all the test results of DBT. The HYD

selectivity of NiMo and NiCoMo catalysts is higher
than that of CoMo.
We observed the differences resulting from different
routes among the catalysts, and found that the degree of
inhibition by H2S and NH3 are differ according to the
route. The reason for these differences will be dis-
cussed in the following section.
3. 4. Inhibition Effect of H2S and NH3
Each first order reaction rate constant is compared to
the rate constant of the blank test at the same tempera-
ture. The average of the activity relative to the blank
test is shown as relative activity in Fig. 3.
The ranking of the inhibition effect of H2S on the
: NiMo, : NiCoMo, : CoMo. conversion of 4,6-DMDBT is as follows,
Fig. 3 The Average Inhibition Degree of the Activity for the
CoMo < NiCoMo < NiMo
Conversion of 4,6-DMDBT and DBT against Each The ranking of the inhibition effect of NH3 on the
Blank Test at the Same Temperature in the Presence of conversion of 4,6-DMDBT is as follows,
H2S and NH3 NiMo, NiCoMo << CoMo
The ranking of the inhibition effect of H2S on the
conversion of DBT is as follows,
CoMo, NiCoMo < NiMo

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 286

: NiMo, : NiCoMo, : CoMo. : NiMo, : NiCoMo, : CoMo.

Fig. 4 The Average Inhibition Degree of kH and kD for the Fig. 5 The Average Inhibition Degree of kH and kD for the
Conversion of 4,6-DMDBT against Each Blank Test at Conversion of DBT against Each Blank Test at the
the Same Temperature in the Presence of H2S and NH3 Same Temperature in the Presence of H2S and NH3

The ranking of the inhibition effect of NH3 on the has been studied extensively in the literature2) 4). In
conversion of DBT is as follows, general, by adsorption of H2S, DDS active sites may be
NiMo < NiCoMo < CoMo converted into HYD active sites, which can change the
rate of elementary reaction steps13). Texier et al.14)
4. Discussion reported that the addition of H2S increases hydrodesul-
furization activity. Reinhoudt et al.5) also reported
4. 1. Inhibition Effect of H2S and NH3 on HYD and that the addition of DMDS, which means H2S addition,
DDS Reaction Pathways increases CoMo catalyst activity of HDS for 4,6-alkyl-
To identify the active site of each promoted catalyst, ated DBT and DBT. In our study, we did not observe
we estimated which route was poisoned by H2S and such an increase; however, the high tolerance to H2S
NH3 by comparing separately kH and kD activity of the poisoning of NiCoMo and CoMo has the same mech-
blank test with the tests using H2S- and NH3-containing anism as this increase in activity. The adsorption of
feeds. H2S on the CoMo catalyst surface may form a HYD
Figure 4 shows the average inhibition degree of kH active site.
and kD for the conversion of 4,6-DMDBT, against each Raybaud et al.15) reported the most stable Co-pro-
blank test at the same temperature in the presence of moted MoS2 (101̄0) surfaces, found using ab initio
H2S and in the presence of NH3. quantum chemical calculation. The study shows that
The inhibition of H2S in the HYD route on NiMo Co atoms are apt to be substituted for Mo atoms on the
catalyst is clearly shown. CoMo catalyst is weakly MoS2 (101̄0) edge and the most stable surface is not
poisoned by H2S. NiCoMo poisoning is moderate, covered at all by sulfur and presents only four-fold
between NiMo and CoMo. coordinated Co atoms. There are many coordinatively
On the other hand, direct desulfurization route of unsaturated (CUS) sites where a sulfur atom can adsorb
CoMo and NiCoMo are strongly poisoned by H2S. are located on the most stable Co-promoted MoS2 sur-
This implies H2S is adsorbed on the active site of DDS. face.
The inhibition effect of H2S and NH3 is more clearly Travert et al.16) also reported the most stable CoMo
shown in the conversion of 4,6-DMDBT than in the and NiMo catalyst surfaces under actual working con-
conversion of DBT. The sulfur atom in 4,6-DMDBT ditions, calculating the adsorption energy of hydrogen
is difficult to remove because of the steric hindrance of using the same calculation program as Raybaud et al.15)
the methyl groups12). The reaction route for HDS of and considering the thermodynamic equilibrium of these
4,6-DMDBT is usually HYD; therefore, this cause of surfaces with H2S and H2 in the gas phase. The results
the difference in inhibition has clearer effect than in the show that hydrogen is easily dissociated on CoMo cata-
reaction route for HDS of DBT. lyst surfaces of the Co_S_Co bridge site which is one of
The inhibition effect of H2S on each catalyst in the the most stable Co-promoted MoS2 surfaces under
HDS of DBT is shown in Fig. 5. actual working conditions, while hydrogen breaks Ni_S
The role of H2S on hydrotreating sulfide catalysts bonds on the NiMo catalyst surface edge of the S_Ni_S

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 287

Table 3 Comparison of kH and kD for the Conversion of 4,6-DMDBT or DBT between the Proportional
Distribution and the NiCoMo Catalyst (kH; h−1 g−1, kD; h−1 g−1)

Reaction Proportional
CoMo NiMo NiCoMo
Feed temperature distributiona)
[K] kH kD kH kD kH kD kH kD
Blank 583 8 51 26 41 17 66 14 47
test 593 13 74 36 52 22 83 20 66
603 12 97 35 70 30 91 20 88
+H2S 593 9 36 12 14 12 35 10 29
603 13 54 18 23 19 55 14 43
613 17 69 25 34 21 88 20 57
+NH3 593 7 44 16 57 9 56 10 48
603 12 71 27 71 15 75 17 71
613 16 90 47 93 28 110 26 91

a) The proportional distribution of 66%CoMo and 33%NiMo.

Mesh: Higher kH and kD values between CoMo and NiMo.

bridge structure. straight-run gas oil. The activity of CoMo catalyst is

The CUS site is a DDS active site, and after the strongly poisoned by NH3. Reinhoudt et al.5) also
adsorption of S atom on the CUS site, a Co_S_Co reported the poisoning effect of carbazole on CoMo
bridge structure is formed on the Co-promoted MoS2 catalyst in the case of 4,6-alkylated DBT. If the feed
(101̄0) surface. This Co_S_Co bridge structure acti- contains a lot of nitrogen compounds, CoMo catalyst is
vates hydrogen16) and likely influences the HYD active not suitable for HDS.
site. 4. 4. The Activity of NiCoMo Catalyst
The inhibition effect of NH3 is shown in Figs. 4 and NiCoMo catalyst is the best catalyst for DBT conver-
5. NH3 poisoned strongly both the HYD and DDS sion in all cases. After calculating HYD and DDS
active site on CoMo catalyst in the case both of DBT activity, we found an interesting result in the summa-
and 4,6-DMDBT. In the case of 4,6-DBT poisoned tion of kH and kD.
with NH3, NiCoMo catalyst behaves like NiMo cata- NiCoMo catalyst comprises Ni, Co and Mo impreg-
lyst. In the case of DBT poisoned with NH3, NiMo nated in the same support. On this surface, there are
catalyst shows a high tolerance for NH3. Especially, two active sites located together. Examining the inhi-
kD of NiMo catalyst exhibits almost the same activity as bition effect of H2S and NH3, we found that the nature
the blank test in the case of DBT poisoned with NH3. of NiMo catalyst and of CoMo catalyst are quite differ-
4. 2. The Activity of NiMo Catalyst ent from each other. CoMo catalyst has a high toler-
NiMo catalyst causes the highest level of HDS con- ance for H2S and NiMo catalyst has a high tolerance for
version of 4,6-DMDBT blank test. Among our exper- NH3. If there are two different active sites on
imental conditions, this is the case where H2S concen- NiCoMo catalyst, and kH and kD are different for NiMo
tration is the lowest. An ideal gas/liquid separation and CoMo, the NiCoMo HDS activity should be the
system with high HDS-promoting catalyst in the first- summation of higher values of kH and kD.
stage reactor realizes such a lean H2S condition. We list the higher of the values of kH and kD for
Hensen et al.17) also reported high activity of NiMo cat- NiMo and for CoMo, and compare the calculated val-
alyst under lean H2S conditions. ues of NiCoMo catalyst in Table 3.
The NiMo catalyst has a high tolerance for NH3. Although there are some variations at higher temper-
Figure 5 shows that the HYD active site on NiMo cata- atures, the activity of NiCoMo catalyst is the summa-
lyst is poisoned by NH3 as are the other catalysts; how- tion of the higher values of kH and kD for NiMo and for
ever, the DDS active site on NiMo catalyst is not poi- CoMo in the presence of rich H2S and NH3 conditions.
soned at all. This phenomenon suggests that the DDS The CoMo catalyst has a higher kD value than the
active site on NiMo catalyst has a high tolerance for NiMo catalyst while the NiMo catalyst has a higher
NH3. value of kH. NiCoMo catalyst activity is a summation
4. 3. The Activity of CoMo Catalyst of kD of CoMo and kH of NiMo; therefore the total HDS
The CoMo catalyst is not as active as the other two activity is higher than either or CoMo. It is known
catalysts except under H2S rich conditions. However, that catalytic reactions proceed through the lowest bar-
H2S rich conditions are the most typical actual condi- rier. NiCoMo catalyst has two kinds of active sites,
tions for HDS of straight-run gas oil, and therefore and the reaction pathway proceeds through the active
CoMo catalyst is commonly used for actual HDS of site with the lower barrier. Thus, NiCoMo catalyst

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 288

activity is the sum of the higher values of kH and kD for HYD values for NiMo and CoMo, and the higher of the
CoMo and for NiMo. DDS values for NiMo and CoMo. This implies the
The variation of kH among catalysts in the presence dual promoting effect of Ni and Co on NiCoMo cata-
of NH3 is larger than in the presence of H2S. As lyst.
shown in Table 3, the HYD route of all catalysts is The best combination of catalysts in the first-stage
poisoned by NH3. If we find the proportion of CoMo and second-stage reactor for the two-stage process with
and NiMo and incorporate this in calculations, the esti- gas/liquid separation is considered to be as follows.
mated HYD activity of NiCoMo is much more accu- In the first-stage reactor, NiCoMo catalyst is the best
rate. This implies that the HYD active sites are poi- catalyst because of the high tolerance for both H2S and
soned by NH3. NH3. In the second-stage reactor, NiMo catalyst is the
If the NiCoMo catalyst is just a mixture of NiMo and best catalyst because of the high HDS activity under
CoMo catalyst, the activity should be the proportional lean H2S conditions. Among industrial hydrotreating
distribution of NiMo and CoMo activity. The catalysts, a NiW catalyst would be an alternative for a
NiCoMo catalyst comprising 1.9 wt% Co and 0.85 wt% NiMo catalyst because NiW catalyst is well known its
Ni is loaded on an Al2O3 support. Co and Ni atoms high hydrogenation activity in lean H2S conditions.
are intended to substitute for atoms at the MoS2 edge
surface15), and for this reason the NiCoMo catalyst sur- Acknowledgment
face is 1/3 NiMo and 2/3 CoMo. The research of the two-stage process with gas/liquid
We compared the kH and kD calculated based on the separation was entrusted by the New Energy and
proportion of NiMo and CoMo catalyst with the actual Industrial Technology Development Organization
kH and kD of NiCoMo catalyst. Table 3 also shows a (NEDO) under a subsidy of the Ministry of Economy,
comparison between kH and kD for the conversion of Trade and Industry and this research was partially com-
4,6-DMDBT or DBT computed from the proportional missioned by NEDO.
distribution and those of the actual NiCoMo catalyst.
The kH and kD value of NiCoMo catalyst is higher than References
those calculated from the proportional distribution
between NiMo and CoMo except in the case of the 1) Houalla, M., Broderick, D., Sapre, A. V., Nag, N. K., de Beer,
V. J. H., Gates, B. C., Kwart, H., J. Catal., 61, 523 (1980).
blank test of HDS for 4,6-DMDBT. The blank test for
2) Kabe, T., Ishihara, A., Tajima, H., Ind. Eng. Chem. Res., 31,
the conversion of 4,6-DMDBT (300 wt ppm as S 1577 (1992).
atoms) generates very low H2S concentration. The 3) Isoda, T., Ma, X., Mochida, I., Sekiyu Gakkaishi (J. Jpn. Petrol.
other cases generate conditions rich in H2S or NH3, and Inst.), 37, (5), 506 (1994).
therefore NiCoMo catalyst shows higher activity than 4) Zhang, Q., Qian, W., Ishihara, A., Kabe, T., Sekiyu Gakkaishi
(J. Jpn. Petrol. Inst.), 40, (3), 185 (1997).
the proportional distribution would indicate. These
5) Reinhoudt, H. R., van Gorsel, M., van Langeveld, A. D., van
results imply the NiCoMo catalyst causes high activity Veen, J. A. R., Sie, S. T., Moulijn, J. A., Stad. Surf. Sci. and
of both HYD and DDS in the presence of rich H2S or Catal., 127, 211 (1999).
NH3 conditions. 6) LaVopa, V., Satterfield, C. N., J. Catal., 110, 375 (1988).
Finally, we determined the best combination for the 7) Koide, R., Goto, Y., Kawabata, M., Ishida, K., Prep. Petrol.
Chem. Div., Am. Chem. Soc., 46, 398 (2001).
two-stage process with gas/liquid separation as follows.
8) Amemiya, M., Minatoya, M., Koide, R., Goto, Y., Kawabata,
NiCoMo catalyst is the best catalyst for the first-stage M., Ishida, K., Segawa, H., Prepr. Div. Fuel Chem., Am. Chem.
reactor because NiCoMo has the high tolerance for Soc., 47, 461 (2002).
both H2S and NH3. NiMo catalyst is the best catalyst 9) Nakamura, H., JP3545943.
for the second-stage reactor because NiMo has high 10) Houalla, M., Broderick, D., de Beer, V. H. J., Gates, B. C.,
Kwart, H., Prepr. Div. Petrol. Chem., Am. Chem. Soc., 22, 941
HDS activity under lean H2S conditions.
(1977).
11) Isoda, T., Ma, X., Mochida, I., Sekiyu Gakkaishi (J. Jpn. Petrol.
5. Conclusion Inst.), 37, (4), 368 (1994).
12) Isoda, T., Takase, Y., Isumi, N., Kusakabe, K., Morooka, S.,
It was demonstrated that the presence of H2S and Sekiyu Gakkaishi (J. Jpn. Petrol. Inst.), 41, (5), 318 (1998).
13) Yang, S. H., Satterfield, C. N., J. Catal., 81, 168 (1993).
NH3 has a large effect on the catalyst ranking and rela-
14) Texier, S., Berhault, G., Pérot, G., Harlé, V., Diehl, F., J. Catal.,
tive importance of reaction pathways, i.e. the activity of 223, 404 (2004).
HYD route (kH) and the activity of DDS route (kD), of 15) Raybaud, P., Hafner, J., Kresse, G., Toulhoat, H., J. Catal., 190,
various catalysts for the conversion of both DBT and 128 (2000).
4,6-DMDBT. 16) Travert, A., Nakamura, H., van Santen, R. A., Cristol, S., Paul,
J.-F., Payen, E., J. Am. Chem. Soc., 124, 7084 (2002).
In the presence of H2S, NiCoMo catalyst is the best
17) Hensen, E. J. M., de Beer, V. H. J., van Veen, J. A. R., van
catalyst for the conversion of DBT. We examined the Santen, R. A., J. Catal., 215, 353 (2003).
HYD and DDS activity of NiCoMo, and found that the
NiCoMo catalyst activity is the sum of the higher of the

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005

 289

4,6- NiMo/Al2O3, CoMo/Al2O3,

NiCoMo/Al2O3

335-8502 3-17-35

NH3 NiMo
Co2.75Mo11 Ni2.75Mo11 Ni0.9Co1.85Mo11 wt /γ- NH3 NiCoMo
Al2O3 4,6- NiMo CoMo
H2S NH3
NiMo CoMo
NiMo CoMo NiCoMo
CoMo NiCoMo H2S NiMo CoMo
H2S NiMo
CoMo

J. Jpn. Petrol. Inst., Vol. 48, No. 5, 2005