Hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene over sulfided NiMo/γ-Al2O3, CoMo/γ-Al 2O3, and Mo/γ-Al2O3 catalysts

Article abstract of DOI:10.1016/j.jcat.2004.05.002

The hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) was studied over sulfided NiMo/γ-Al₂O₃, CoMo/γ-Al₂O ₃, and Mo/γ-Al₂O₃ catalysts. The Ni and Co promoters strongly enhanced the activity of the Mo catalyst in the direct desulfurization pathway of the HDS of DBT and 4,6-DMDBT and in the final sulfur-removal step in the hydrogenation pathway, while the hydrogenation was moderately promoted. H₂S had a negative effect on the HDS of DBT and 4,6-DMDBT, which was strongest for the NiMo catalyst and stronger for the direct desulfurization pathway than for the hydrogenation pathway. Because the direct desulfurization pathway is less important for the HDS of 4,6-DMDBT than the hydrogenation pathway, the conversion of 4,6-DMDBT was less affected by H ₂S than the conversion of DBT. The sulfur removal via the direct desulfurization pathway and the ultimate sulfur removal in the hydrogenation pathway were affected by H₂S to the same extent over all the catalysts. This suggests that the removal of sulfur from tetrahydrodibenzothiophenes takes place by hydrogenolysis, like the direct desulfurization of DBT to biphenyl. The CoMo catalyst performed better than the NiMo catalyst in the final desulfurization via the hydrogenation pathway in the HDS of 4,6-DMDBT at all partial pressures of H₂S.

Full text of DOI:10.1016/j.jcat.2004.05.002

Journal of Catalysis 225 (2004) 417–427
www.elsevier.com/locate/jcat
Hydrodesulfurization of dibenzothiophene and
4,6-dimethyldibenzothiophene over sulfided NiMo/γ -Al2O3,
CoMo/γ -Al2O3, and Mo/γ -Al2O3 catalysts
Marina Egorova and Roel Prins ∗
Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology (ETH), 8093 Zurich, Switzerland
Received 16 January 2004; revised 28 April 2004; accepted 3 May 2004
Available online 28 May 2004
Abstract
The hydrodesulfurization (HDS) of dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT) was studied over sulfided
NiMo/γ -Al O , CoMo/γ -Al O , and Mo/γ -Al O catalysts. The Ni and Co promoters strongly enhanced the activity of the Mo catalyst in
2
3
2
3
2 3
the direct desulfurization pathway of the HDS of DBT and 4,6-DMDBT and in the final sulfur-removal step in the hydrogenation pathway,
while the hydrogenation was moderately promoted. H S had a negative effect on the HDS of DBT and 4,6-DMDBT, which was strongest
2
for the NiMo catalyst and stronger for the direct desulfurization pathway than for the hydrogenation pathway. Because the direct desulfu-
rization pathway is less important for the HDS of 4,6-DMDBT than the hydrogenation pathway, the conversion of 4,6-DMDBT was less
affected by H S than the conversion of DBT. The sulfur removal via the direct desulfurization pathway and the ultimate sulfur removal in the
2
hydrogenation pathway were affected by H S to the same extent over all the catalysts. This suggests that the removal of sulfur from tetrahy-
2
drodibenzothiophenes takes place by hydrogenolysis, like the direct desulfurization of DBT to biphenyl. The CoMo catalyst performed better
than the NiMo catalyst in the final desulfurization via the hydrogenation pathway in the HDS of 4,6-DMDBT at all partial pressures of H S.
2
2004 Elsevier Inc. All rights reserved.
Keywords: Hydrotreating; Hydrodesulfurization; HDS; Dibenzothiophene; DBT; 4,6-Dimethyldibenzothiophene; 4,6-DMDBT; Transition metal sulfide
catalysts; Mo catalysts; CoMo catalyst; NiMo catalyst
1. Introduction
DMDBT reacts mainly via the HYD pathway [5,8,12–16].
The reactivity of 4,6-DMDBT over Co- or Ni-promoted cat-
alysts is 5 to 10 times smaller than that of DBT [10,12,17].
The difficulty in converting alkyl-substituted DBTs is due to
the steric hindrance of alkyl groups, which are close to the
sulfur atom and prevent the interaction of the sulfur atom
with the active site. Kabe et al. reported that the adsorption
constant of 4,6-DMDBT on the CoMo/Al2O3 catalyst was
twice as high as that of DBT and suggested that DBTs adsorb
by means of π bonding to the catalyst [6]. In this adsorption
mode, the methyl groups do not hinder the adsorption of
the molecule on the catalyst surface. Thus, they suggested
that C–S bond scission occurs when DBT is adsorbed via
the sulfur atom in the σ mode. In this case, the methyl
groups hinder C–S bond scission [6]. They also reported
that 4-methyltetrahydrodibenzothiophene, the hydrogenated
intermediate in the HYD pathway, is easier to desulfurize
than 4-methyldibenzothiophene. In our previous work we
observed that 4,6-DMDBT converts somewhat faster via the
To improve the hydrodesulfurization (HDS) process, it
is necessary to gain a better understanding of the conver-
sion mechanisms of various sulfur compounds. Dibenzoth-
iophene (DBT) and 4,6-dimethyldibenzothiophenes (4,6-
DMDBT) belong to the most refractory sulfur-containing
molecules present in gas oil [1–12]. They are, therefore,
often used as model sulfur compounds. DBT and alkyl-
substituted DBT undergo HDS via two reaction pathways:
(i) direct desulfurization (DDS), which leads to the for-
mation of biphenyls; (ii) hydrogenation (HYD) yielding
tetrahydro- and hexahydro-intermediates followed by desul-
furization to cyclohexylbenzenes and bicyclohexyls. DBT is
converted predominantly via the DDS pathway, whereas 4,6-
*
Corresponding author. Fax: +41 1 632 1162.
E-mail address: roel.prins@chem.ethz.ch (R. Prins).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.05.002
2004 Elsevier Inc. All rights reserved.

418
M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
HYD route than DBT over NiMo/Al2O3 but that the desulfu-
rization of 4,6-dimethyltetrahydrodibenzothiophene occurs
much slower than that of tetrahydrodibenzothiophene [16].
Thus, the two methyl substituents have a negative influence
on the transformation of the partially hydrogenated inter-
mediate. Further research is required to determine why the
reactivity of 4,6-DMDBT is low.
defined as τ = wcat/nfeed, where wcat denotes the cata-
lyst weight and nfeed the total molar flow to the reactor
(1 g min/mol = 1.8 × 10−2 g h/L). The weight time (τ)
was changed by varying the flow rates of the liquid and
the gaseous reactants while keeping their ratio constant. The
reaction was stable after 3 to 4 h; during the 2 weeks of op-
eration there was almost no deactivation of the catalyst.
A better knowledge of the catalytic sites involved in the
transformation of sulfur-containing molecules may help us
gain insight into the process of sulfur removal. Therefore,
we studied the HDS of DBT and 4,6-DMDBT over differ-
ent catalysts: sulfided NiMo/γ -Al2O3, CoMo/γ -Al2O3, and
Mo/γ -Al2O3. Some researchers described the HDS of hin-
dered sulfur-containing compounds over the CoMo or the
NiMo catalysts [8,10,18,19], while others compared the cat-
alysts [12,20–25]. Landau et al. found that a Co-promoted
catalyst is more active in the HDS of substituted DBTs [22].
Mochida’s group, on the other hand, reported that NiMo
performs better than CoMo in the HDS of a gas oil [20]
but is affected more strongly by aromatic compounds and
H2S [23]. Knudsen et al. studied catalysts for ultradeep HDS
of diesel fuel and found that a NiMo catalyst performed well
only when most of the heterocyclic compounds had been
removed [25]. Thereafter, the NiMo catalyst was more ac-
tive than the CoMo catalyst. At low H2 pressures and high
temperatures, a CoMo catalyst performs better than a NiMo
catalyst [25]. To clarify the effect of the promoter, we car-
ried out a detailed analysis of the HDS mechanisms of DBT
and 4,6-DMDBT over sulfided CoMo, NiMo, and Mo cata-
lysts supported on alumina. The effect of H2S was studied
to gain a better understanding of the nature of the DDS and
HYD sites that take part in the HDS.
3. Results
3.1. HDS of DBT over CoMo/γ -Al2O3
The HDS of DBT was studied at 340 ◦C and 35 kPa H2S
over the CoMo/γ -Al2O3 catalyst. Fig. 1 shows the relative
partial pressures of DBT and its products as well as the
product selectivities. Four reaction products were observed:
Biphenyl and cyclohexylbenzene were the final products
of the DDS and HYD pathways, respectively, and tetrahy-
drodibenzothiophene and hexahydrodibenzothiophene were
intermediates in the HYD pathway of the HDS. The HYD
curve in Fig. 1a represents the sum of all the hydrogenation
products (tetrahydrodibenzothiophene, hexahydrodibenzo-
thiophene, and cyclohexylbenzene).GC-MS showed that the
tetrahydro-intermediate was 1,2,3,4-tetrahydrodibenzothio-
phene, with the double bond at the bridge between the par-
2. Experimental
The Mo/γ -Al2O3, NiMo/γ -Al2O3, and CoMo/γ -Al2O3
catalysts used in this work contained 8 wt% Mo and 0 or
3 wt% promoter (Ni or Co). They were prepared by succes-
sive incipient wetness impregnation of γ -Al2O3 (Condea,
(a)
pore volume 0.5 cm3 g−1, specific surface area 230 m2 g−1
)
with an aqueous solution of (NH4)6Mo7O24 · 4H2O, fol-
lowed (for the promoted catalysts) by an aqueous solution
of Ni(NO3)2 · 6H2O or Co(NO3)2 · 6H2O (all Aldrich). Af-
ter each impregnation step the catalysts were dried in air at
ambient temperature for 4 h and then in an oven at 120 ◦C
for 15 h and finally calcined at 500 ◦C for 4 h.
Reactions were carried out in a continuous mode in a
fixed-bed inconel reactor, as described elsewhere [15]. All
the experiments were performed at 340 ◦C. The gas-phase
feed consisted of 130 kPa toluene (the solvent for the DBTs),
8 kPa dodecane (the reference for DBTs and their derivatives
in the GC analysis), 1 kPa DBT or 4,6-DMDBT, 0–100 kPa
H2S and 4.8 MPa H2.
(b)
Fig. 1. Relative partial pressures (a) and selectivities (b) of the prod-
ucts in the HDS of dibenzothiophene at 340 C and 35 kPa H S over
CoMo/γ -Al O as a function of weight time.
◦
The reaction products were analyzed by off-line gas chro-
matography, as described elsewhere [15]. Weight time was
2
2
3

M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
419
tially hydrogenated 6-ring and the thiophene ring. Only trace
amounts of hexahydrodibenzothiophene were detected; its
relative partial pressure did not exceed 0.5% and its selectiv-
ity was 1.5% at the lowest weight time, decreasing to 0.14%
at the highest weight time (values not plotted in Fig. 1).
The product distribution (Fig. 1b) shows that the selectiv-
ity of biphenyl formation, that is, the selectivity for the DDS
pathway, is 70% and remains constant during the reaction.
The selectivity toward the formation of tetrahydrodibenzoth-
iophene was rather low, 6% at the lowest and 0.45% at the
highest weight time. This shows that this intermediate is
quickly converted.
clohexyl is typical of Mo/γ -Al2O3 (Fig. 2b). This product
was not observed at all over the NiMo and CoMo catalysts.
The DDS selectivity approached 35% at low weight time
and decreased to 25% at higher weight time (Fig. 2b). Up
to a weight time of 4 g min/mol, the main reaction product
is tetrahydrodibenzothiophene, the intermediate in the HYD
pathway (Fig. 2b). The yield of hexahydrodibenzothiophene
was about 3 to 22 times higher in the presence of the Mo
catalyst than in the presence of the CoMo catalyst.
3.3. HDS of 4,6-DMDBT over CoMo/γ -Al2O3
Fig. 3 shows the results of the HDS of 4,6-DMDBT over
the CoMo/γ -Al2O3 catalyst at 340 ◦C and 35 kPa H2S.
3,3-Dimethylbiphenyl was the product of the DDS route
and 4,6-dimethyltetrahydrodibenzothiophene, 1-methyl-3-
(2-methylphenyl)cyclohexane (which we will refer to as
methylcyclohexyltoluene) and 3,3ꢀ-dimethylbicyclohexyl
were the products of the HYD pathway. The results of the
GC-MS analysis showed that the tetrahydro-intermediate
was 4,6-dimethyl-1,2,3,4-tetrahydrodibenzothiophene. The
conversion via the HYD route was much higher than that
via the DDS pathway (Fig. 3a), and the total conversion was
much smaller than that of DBT (cf. Figs. 1a and 3a).
3.2. HDS of DBT over Mo/γ -Al2O3
Fig. 2 shows the relative partial pressures and product se-
lectivities in the HDS of DBT over Mo/γ -Al2O3 at 340 ◦C
and 35 kPa H2S. The overall conversion was much lower
over Mo than over CoMo and, although the same products
were observed, the product distribution was completely dif-
ferent. In contrast to the CoMo (Fig. 1a) and NiMo cata-
lysts [15], Mo/γ -Al2O3 mainly catalyzed the HYD pathway
(Fig. 2a). Biphenyl, the product of the DDS pathway, was
detected in much smaller amounts over the Mo catalyst than
over the CoMo and NiMo catalysts. The presence of bicy-
(a)
(a)
(b)
(b)
Fig. 2. Relative partial pressures (a) and selectivities (b) of the products in
Fig. 3. Relative partial pressures (a) and selectivities (b) of the products in
◦
◦
the HDS of dibenzothiophene at 340 C and 35 kPa H S over Mo/γ -Al O
the HDS of 4,6-dimethyldibenzothiophene at 340 C and 35 kPa H S over
2
2
3
2
as a function of weight time.
CoMo/γ -Al O as a function of weight time.
2 3

420
M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
The selectivity toward the formation of 3,3ꢀ-dimethylbi-
product up to τ = 4 g min/mol (Fig. 4b). The main desul-
furized product of the HDS of 4,6-DMDBT over the Mo
catalyst was 3,3ꢀ-dimethylbicyclohexyl.
phenyl (DDS selectivity) was only 12% and remained
constant during the reaction (Fig. 3b). The main reac-
tion product is methylcyclohexyltoluene even at the lowest
weight time, and 3,3ꢀ-dimethylbicyclohexyl is the second
most abundant product at τ ꢀ 1.7 g min/mol (Fig. 3b).
The yield of the partially hydrogenated intermediate, 4,6-
dimethyltetrahydrodibenzothiophene, is higher than that of
tetrahydrodibenzothiophene in the HDS of DBT.
3.5. Effect of H2S on the HDS of DBT and 4,6-DMDBT
The effect of different initial partial pressures (0, 10,
35, and 100 kPa) of hydrogen sulfide on the HDS of
DBT and 4,6-DMDBT was studied over all three catalysts:
NiMo/γ -Al2O3, CoMo/γ -Al2O3, and Mo/γ -Al2O3. Over
NiMo/γ -Al2O3 the HDS of DBT was extremely fast in the
absence of H2S and reached 100% conversion already at
τ = 1.5 g min/mol. Over CoMo/γ -Al2O3 the same con-
version was reached at τ = 2.3 g min/mol and over the
Mo/γ -Al2O3 catalyst the conversion of DBT was only 55%
at τ = 3.7 g min/mol. Since the feed contained only 1 kPa
of DBT or 4,6-DMDBT, its influence on the H2S partial
3.4. HDS of 4,6-DMDBT over Mo/γ -Al2O3
The Mo/γ -Al2O3 catalyst was only slightly less active in
the HDS of 4,6-DMDBT than in the HDS of DBT under
the same reaction conditions (340 ◦C and 35 kPa H2S) (cf.
Figs. 2a and 4a). The same products were observed as over
the CoMo catalyst. On the Mo catalyst the HYD conversions
of DBT and 4,6-DMDBT were almost the same, while the
DDS conversion of 4,6-DMDBT was lower than that of DBT
(cf. Figs. 2a and 4a).
pressure during reaction was small at PH S(init) = 10 kPa
2
and negligible at PH S(init) = 35 and 100 kPa. For the
2
experiments at PH S(init) = 0 kPa, however, the amount
2
The DDS selectivity (3,3ꢀ-dimethylbiphenylformation) is
16% during the whole reaction (Fig. 4b), somewhat higher
than over the CoMo catalyst (12%). However, the yield
of 3,3ꢀ-dimethylbiphenyl was higher over the CoMo cata-
lyst because of the higher conversion of the reactant. 4,6-
Dimethyltetrahydrodibenzothiophene is the main reaction
of H2S present during reaction was equal to the amount
formed in the HDS reaction. The results obtained in the ab-
sence of added H2S therefore do not really correspond to
PH S = 0 kPa, but correspond to PH S due to the desulfur-
2
2
ization conversion.
H2S strongly inhibited the conversion of DBT over all
three catalysts. At 100 kPa H2S and τ = 3.7 min/mol the
conversion of DBT reached 60% over NiMo, 53% over
CoMo, and 24% over the Mo catalyst. The conversion of
DBT and 4,6-DMDBT was well described by pseudo-first-
order kinetics. The plots of ln(C/C0) versus weight time
were always linear with high R2 values (0.97–0.99).In Fig. 5
we present the plot for the HDS of DBT at 35 kPa H2S
over the NiMo catalyst, where the DBT conversion at the
highest weight time was 90%. To prove that we also deal
with pseudo-first-order reactions when the catalyst activ-
ity is lower (and thus lower conversions are obtained), we
performed experiments on the HDS of DBT at two initial
partial pressures of the reactant (0.5 and 2 kPa) over the
Mo catalyst. It was found that the conversion of DBT did
not depend on its initial concentration, showing that also
(a)
(b)
Fig. 4. Relative partial pressures (a) and selectivities (b) of the products in
◦
the HDS of 4,6-dimethyldibenzothiophene at 340 C and 35 kPa H S over
Fig. 5. Plot of ln(C/C ) versus weight time for the HDS of DBT at 35 kPa
2
0
◦
Mo/γ -Al O as a function of weight time.
H S and 340 C over NiMo/γ -Al O .
2
3
2
2 3

M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
421
over Mo/γ -Al2O3 the HDS reaction follows first-order be-
havior. The observed first-order behavior seems to contradict
the general belief that HDS is strongly inhibited by reactant
molecules. It should be realized, however, that most stud-
ies have been executed at a lower temperature than ours. In
accordance with the literature [3,26], we observed a decreas-
ing conversion when increasing the partial pressure of DBT
from 0.1 to 3 kPa at the lower temperature of 300 ◦C and cal-
culated an adsorption constant of 2 kPa−1. This means that
at 340 ◦C and with PH S ꢁ PDBT and KH S > KDBT [3,26]
presence of H2S the activity of all three catalysts decreased
strongly. At higher partial pressures of H2S, however, the
CoMo catalyst was less affected than the NiMo catalyst, and
the activity of the Mo catalyst hardly changed. Over all the
catalysts, the DDS pathway was suppressed to a greater ex-
tent than the HYD pathway. The rate constant of the HYD
pathway, kHYD, was almost the same at 0 and 100 kPa H2S
over the Mo catalyst. As a consequence, at 35 and 100 kPa
H2S the contribution of the HYD pathway was larger than
that of the DDS pathway (Table 1).
The yields of the partially hydrogenatedsulfur-containing
intermediates were much lower over the NiMo and CoMo
catalysts than over the Mo catalyst. Therefore, the desul-
furization (the total sulfur removal) of DBT via the HYD
pathway is also improved by the promoters. The selectiv-
ities of the partially hydrogenated intermediates increased
2
2
the equation
r = kKDBTPDBT/(1 + KDBTPDBT + KH SPH S
)
2
2
can be replaced by r = kꢀPDBT. The constant kꢀ is a pseudo
rate constant which contains the influence of PH and PH S
.
2
2
Table 1 gives the pseudo rate constants of the total conver-
sion of DBT (ktot) for the three catalysts at different partial
with increasing partial pressure of H2S. We estimated the
HYD
pressures of H2S. At PH S(init) = 0 kPa, the rate of total
rate constant k
of the final desulfurization step in the
2
DESULF
DBT conversion was almost 10 times higher over the NiMo
catalyst and 6 times higher over the CoMo catalyst than over
the Mo catalyst. This is in good agreement with published
results [10,17]. These ktot ratios hardly changed in the pres-
HYD pathway by analyzing this pathway as a series of first-
order reactions A → B → C with rate constants kHYD and
HYD
DESULF
k
. The concentration of the intermediate B is given
by the expression [27]:
ence of 10 kPa H2S. At higher PH S, the positive influence
of the promoters was less outspoken. This is clearly shown
in Fig. 6 in which the pseudo rate constants ktot are plotted
2
kHYD
CB = CA
0
HYD
k
− kHYD
DESULF
ꢀ
ꢁ
ꢂꢃ
× exp(−kHYDτ) − exp −kHYD
DESULFτ
.
as a logarithmic function of PH S. This figure indicates that
2
H2S inhibits the HDS of DBT more strongly over the NiMo
and CoMo catalysts than over Mo.
The rate constant of the HYD pathway (kHYD) is already
HYD
DESULF
known; thus, we can calculate k
tion of the intermediate B, tetrahydrodibenzothiophene, at
weight time τ. The rate constants k
from the concentra-
To analyze the product distributions at different partial
pressures of H2S, the pseudo rate constants of the DDS path-
way (kDDS) and of the HYD pathway (kHYD) in the HDS of
DBT were determined from ktot and the selectivities at low
weight time. The resulting kDDS and kHYD values are pre-
sented in Table 1. In the absence of H2S, all three catalysts
performed better via the DDS than via the HYD pathway.
The Ni and Co promoters improved the DDS activity of the
Mo catalyst greatly and the HYD activity moderately. In the
HYD
for the differ-
DESULF
ent catalysts and at different H2S partial pressures are given
in Table 1. The activity of the CoMo and NiMo catalysts
for the final sulfur removal via HYD is much higher than
that of the Mo catalyst. H2S suppresses kHYD
DESULF much more
strongly than kHYD, the rate constant of the hydrogenation
of DBT to tetrahydrodibenzothiophene. The ratio of the rate
constants of sulfur removal via the DDS and HYD pathways
(kDDS/kDHEYSDULF) remained constant at different partial pres-
Table 1
Rate constants of the total, DDS, and HYD (desulfurization) conversions
of DBT and of the desulfurization of tetrahydrodibenzothiophene over
NiMo/γ -Al O , CoMo/γ -Al O , and Mo/γ -Al O catalysts at different
2
3
2
3
2 3
initial partial pressures of H S
2
Catalyst
NiMo
Rate constant, mol/(g min)
P
(init), kPa
10
H S
2
0
35
100
k
k
k
k
2.33
2.05
0.28
1.06
0.90
0.16
15
0.39
0.32
0.07
5.3
0.25
0.17
0.08
2.7
tot
DDS
HYD
HYD
36
DESULF
CoMo
Mo
k
k
k
k
1.42
1.31
0.11
0.42
0.34
0.08
9.1
0.36
0.28
0.08
5.1
0.20
0.14
0.06
3.3
tot
DDS
HYD
HYD
29
DESULF
k
k
k
k
0.25
0.19
0.06
4.7
0.09
0.05
0.04
1.3
0.08
0.03
0.05
0.70
0.08
0.03
0.05
0.50
tot
DDS
Fig. 6. Rate constants of the total conversion of DBT over NiMo/γ -Al O ,
2
3
HYD
CoMo/γ -Al O , and Mo/γ -Al O catalysts at different partial pressures of
2
3
2 3
HYD
DESULF
H S.
2

422
M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
Table 2
Rate constants of the total, DDS, and HYD (desulfurization) conver-
sions of 4,6-DMDBT and of the desulfurization of 4,6-dimethyltetra-
hydrodibenzothiophene over NiMo/γ -Al O , CoMo/γ -Al O , and Mo/γ -
2
3
2
3
Al O catalysts at different initial partial pressures of H S
2
3
2
(init), kPa
10
Catalyst
Rate constant, mol/(g min)
P
H S
2
0
35
100
NiMo
k
k
k
k
0.20
0.08
0.12
3.6
0.13
0.04
0.09
1.9
0.11
0.03
0.08
1.2
0.08
0.02
0.06
0.91
tot
DDS
HYD
HYD
DESULF
CoMo
Mo
k
k
k
k
0.16
0.06
0.10
4.2
0.12
0.03
0.09
3.0
0.12
0.02
0.10
2.0
0.09
0.01
0.08
1.8
tot
DDS
HYD
HYD
DESULF
Fig. 7. Rate constants of the total conversion of 4,6-DMDBT over
NiMo/γ -Al O , CoMo/γ -Al O , and Mo/γ -Al O catalysts at different
k
k
k
k
0.09
0.02
0.07
0.88
0.08
0.01
0.07
0.52
0.07
0.01
0.06
0.50
0.07
0.01
0.06
0.36
tot
2
3
2
3
2 3
DDS
partial pressures of H S.
2
HYD
HYD
DESULF
NiMo and Mo, and 0.01 for CoMo). This indicates that
the DDS desulfurization of 4,6-DMDBT is inhibited to the
sures of H2S over all three catalysts. It was equal to 0.06
for NiMo, 0.045 for CoMo, and 0.04 for Mo. This means
that the desulfurizations of DBT to biphenyl and of tetrahy-
drodibenzothiophene to cyclohexylbenzene are inhibited to
the same extent by H2S. The kDDS/kDHEYSDULF values show
that the desulfurization via the DDS pathway is 16 to 25
times slower than the desulfurization of the hydrogenated
intermediate.
same extent by H2S as the HYD desulfurization of 4,6-
HYD
DESULF
dimethyltetrahydrodibenzothiophene. The k
values
for the HDS of DBT and 4,6-DMDBT show that the methyl
groups have a strong inhibitory effect not only on the DDS
pathway but also on the desulfurization via the HYD path-
way. The final desulfurization of 4,6-DMDBT via HYD is
always faster over the CoMo than over the NiMo catalyst
(Table 2).
The rate constants of the hydrogenation, kHYD, and the
rate constants of the desulfurizations via the DDS and HYD
pathways, kDDS and kHYD
DESULF, for the HDS of 4,6-DMDBT
4. Discussion
in the presence of different catalysts and at different partial
pressures of H2S were calculated in the same way as for
the HDS of DBT. The results are presented in Table 2. The
differences in the performance of the Co- or Ni-promoted
catalyst and the Mo catalyst were smaller in the HDS of
4,6-DMDBT than in the HDS of DBT (cf. Tables 1 and 2).
Thus, the pseudo rate constant ktot of the overall conversion
of 4,6-DMDBT was only 2.2 times higher over the NiMo
catalyst than over the Mo catalyst and 1.9 times higher over
the CoMo catalyst in the absence of added H2S. Increasing
the partial pressure of H2S led to a decrease of ktot for all
catalysts, but most strongly for NiMo (Fig. 7). As a con-
sequence, at 35 and 100 kPa H2S, the CoMo catalyst was
more active in the HDS of 4,6-DMDBT than the NiMo cata-
lyst. This indicates that CoMo is a better catalyst under HDS
conditions (at least when only S compounds are present).
The HYD pathway of the HDS of 4,6-DMDBT was less
strongly suppressed by H2S than the DDS pathway over
all three catalysts (Table 2). However, the fraction of the
partially hydrogenated intermediates in the HYD pathway
4.1. DBT
To enable a comparison of the CoMo and Mo catalysts
with the NiMo/γ -Al2O3 catalyst in the HDS of DBT and
4,6-DMDBT at 340 ◦C and 35 kPa H2S, we call to mind the
results obtained for the NiMo catalyst [15,16]. They are very
similar to those of the CoMo catalyst. The HDS of DBT was
slightly faster over the Ni-promoted catalyst than over the
CoMo catalyst, and both catalysts were much more active
than the Mo catalyst (Table 1).
At all partial pressures of H2S the DDS pathway of the
HDS of DBT is about an order of magnitude faster over the
CoMo and NiMo catalysts than over the Mo catalyst (Ta-
ble 1). The rate of the HYD pathway is just slightly higher
for the CoMo and NiMo catalysts than for the Mo catalyst
but the product distribution is different. In the presence of
NiMo and CoMo the main HYD product is cyclohexylben-
zene (Fig. 1b), whereas over Mo cyclohexylbenzene consti-
tutes only 45% of the HYD products at τ = 5 g min/mol
(Fig. 2b). Another 45% comes from the tetrahydro- and
hexahydrodibenzothiophenes. Therefore, the promotion of
the Mo catalyst by Ni or Co is due not only to the better
performance of the DDS pathway but also to faster desulfu-
increased at increasing H2S partial pressures. The pseudo
HYD
DESULF
rate constant k
of the Mo catalyst in the removal
of sulfur via HYD is enhanced by the Ni and Co promot-
ers but to a lesser extent than in the HDS of DBT (cf.
Tables 1 and 2). The ratio kDDS/kDHEYSDULF was constant at
all partial pressures of H2S and over all catalysts (0.02 for
rization via the HYD pathway. This was confirmed by the
HYD
DESULF
k
values (Table 1).

M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
423
(a)
Scheme 1. Reaction network of the HDS of dibenzothiophene in the pres-
ence of transition metal sulfide catalysts.
(b)
A detailed network of the HDS of dibenzothiophene was
proposed by Houalla et al. in 1978 [28]. Their data indi-
cate that the conversion of dibenzothiophene occurs prefer-
entially via the path with the least hydrogen consumption
(DDS pathway) and that hydrogenation of biphenyl and cy-
clohexylbenzene is comparatively slow. A similar network,
desulfurization to biphenyl and hydrogenation to tetrahydro-
and hexahydrodibenzothiophene followed by desulfuriza-
tion to cyclohexylbenzene, was proposed by other groups
[29–31]. Our data suggest that the overall network of the
HDS of DBT is as represented in Scheme 1. Over the CoMo
and NiMo catalysts the HDS of DBT most probably oc-
curs via reactions 1, 2, 3, and 4. The tetrahydro-intermediate
was observed over CoMo and NiMo. Trace amounts of the
hexahydro-intermediate were observed over CoMo but not
over NiMo. Over the Mo catalyst, hexahydrodibenzothio-
phene was detected in higher concentrations and bicyclo-
hexyl was observed as well. We assume that bicyclohexyl
is formed by hydrogenation of the partially hydrogenated
intermediates to perhydrodibenzothiophene and subsequent
desulfurization. The hydrogenation of cyclohexylbenzene to
bicyclohexyl is unlikely since the hydrogenation of the sec-
ond phenyl ring of biphenyl ought to be more difficult than
that of the first ring. Our results show that the latter does not
occur over CoMo and occurs only slowly over NiMo and
Mo. Therefore, the HDS of DBT over the Mo catalyst occurs
via reactions 1, 2, 3, 4, 5, and 6. Over the NiMo and CoMo
catalysts, the hydrogenation of the first phenyl ring is the
rate-limiting step in the HYD pathway, and the desulfuriza-
tion of the partially hydrogenated intermediates is rather fast,
Fig. 8. Relative partial pressures (a) and selectivities (b) of the products of
◦
the HDS of 4,6-dimethyldibenzothiophene at 340 C and 35 kPa H S over
2
NiMo/γ -Al O as a function of weight time.
2
3
since the yields of these intermediates are very low (Fig. 1b).
Therefore, further hydrogenation of hexahydrodibenzothio-
phene to perhydrodibenzothiophene does not occur. Over
the Mo catalyst, however, the lifetime of the partially hy-
drogenated molecules is longer. Thus, the hydrogenation to
perhydrodibenzothiophene, followed by desulfurization to
bicyclohexyl, becomes possible. Reaction 7, the hydrogena-
tion of biphenyl to cyclohexylbenzene, was observed over
the NiMo and Mo catalysts but not over the CoMo cata-
lyst. This suggests that the active sites facilitating HDS over
the NiMo and Mo catalysts are different from those over the
CoMo catalyst. One reason for this difference may be that,
according to DFT theory, the active sites of the NiMo and
Mo catalysts are at the molybdenum edge, whereas the Co
atoms are at the sulfur edge of the MoS2 slab [32,33].
4.2. 4,6-DMDBT
The HDS of 4,6-DMDBT over the NiMo and CoMo cat-
alysts was much slower than the HDS of DBT (cf. Tables 1
and 2). At 35 kPa H2S, the total conversion of 4,6-DMDBT
was slightly higher over the CoMo than over the NiMo cat-
alyst due to a somewhat smaller DDS conversion but higher
HYD conversion (cf. Figs. 3a and 8a). Thus, the first re-
action, the hydrogenation, was faster over CoMo. Table 2
shows that the subsequent desulfurization via the HYD path-

424
M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
way is also easier over the CoMo catalyst. Since HYD is the
main pathway in the transformation of 4,6-DMDBT, both its
overall conversion and its desulfurization are faster over the
CoMo catalyst. The pseudo rate constants of the total 4,6-
DMDBT conversion (ktot) over the three catalysts are plotted
The desulfurization of 4,6-DMDBT in the presence of
the NiMo, CoMo, and Mo catalysts proceeds mainly via
reactions 2, 3, 4, 5, and 6. Over the Mo catalyst reac-
tions 2, 3, 5, and 6 are the most important, since 3,3ꢀ-
dimethylbicyclohexyl was observed as the main product.
Over the Co- and Ni-promoted catalysts the most important
reactions are reactions 2, 3, and 4. The NiMo and CoMo cat-
alysts have a higher activity than the Mo catalyst in the DDS
pathway, but the difference is not as large as in the HDS
of DBT. This shows the strong steric hindrance of the two
methyl groups on the DDS reactivity of the DBT molecule.
as a logarithmic function of PH S in Fig. 7. The NiMo cat-
2
alyst was more active than the CoMo catalyst up to 29 kPa
H2S, but at higher PH S the CoMo catalyst performed bet-
2
ter. Therefore, it is assumed that NiMo performs better than
CoMo under deep HDS conditions (at low partial pressures
of H2S), without however considering the influence of im-
purities such as nitrogen-containing compounds.
4.3. Reaction mechanisms and catalytic sites
The activity of the Mo catalyst is somewhat lower
in the HDS of 4,6-DMDBT than the activity of the Ni-
and Co-promoted catalysts (Fig. 7), but for all three cat-
alysts HYD prevails over DDS. Furthermore, the hydro-
genation rate constants of the three catalysts are similar.
However, the amount of hydrogenated, but not desulfur-
ized, products was higher over the Mo catalyst than over
the promoted catalysts. The lowest selectivity for 4,6-
dimethyltetrahydrodibenzothiophene was observed for the
CoMo catalyst. We assume that the mechanism of the HDS
of 4,6-DMDBT is similar to that of DBT (Scheme 1).
Hexahydro- and perhydro-intermediates were not observed
for 4,6-DMDBT in our experiments, probably because
of their fast desulfurization. The hydrogenation of 3,3ꢀ-
dimethylbiphenyl to methylcyclohexyltoluene (reaction 7
in Scheme 1) clearly did not occur, as shown by the
fact that the DDS selectivity to 3,3ꢀ-dimethylbiphenyl re-
mained constant for all three catalysts. The hydrogenation
of methylcyclohexyltoluene to 3,3ꢀ-dimethylbicyclohexyl
will probably be even more difficult to achieve. There-
fore, we assume that it does not occur either. Thus, 3,3ꢀ-
dimethylbicyclohexyl can only be the product of further
hydrogenation of 4,6-dimethyltetrahydrodibenzothiophene
to 4,6-dimethylperhydrodibenzothiophene (reactions 3 and
5 in Scheme 1) followed by desulfurization. In the HDS of
4,6-DMDBT the lifetime of the partially hydrogenated inter-
mediate (4,6-dimethyltetrahydrodibenzothiophene) is quite
long, as is also the case for tetrahydrodibenzothiophene in
the HDS of DBT over the Mo catalyst. Therefore, further hy-
drogenation to 4,6-dimethylperhydrodibenzothiophene fol-
lowed by desulfurization can take place. The larger amount
of 3,3ꢀ-dimethylbicyclohexylthan that of methylcyclohexyl-
toluene in the reaction products over the Mo catalyst indi-
cates that hydrogenation of 4,6-dimethylhexahydrodibenzo-
thiophene occurs faster than its desulfurization to methylcy-
clohexyltoluene.
The cause of the inhibitory effect of the methyl groups
on the reactivity of the DBT molecule is still not completely
understood. Some researchers attributed the lower reactivity
of 4,6-DMDBT to its lower adsorption strength on the cat-
alyst surface [2,8,20], whereas others explained it by steric
hindrance of the methyl groups on the C–S bond scission re-
action [6,10,12]. Meille et al. reported that the adsorption
equilibrium constants of DBT, 4-methyldibenzothiophene,
4,6-DMDBT, and 2,8-DMDBT were almost the same on a
sulfided NiMo catalyst [10]. Kabe et al. reported that the ad-
sorption equilibrium constants of 4-methyldibenzothiophene
and 4,6-DMDBT were even higher than that of DBT on a
CoMo/Al2O3 catalyst [6]. The adsorption constants of the
different sulfur-containing molecules were estimated by tak-
ing into account the total conversion of the reactant and as-
suming that the DDS and HYD pathways take place at the
same active sites. This approach is the right one when the
DDS and HYD reaction routes go through a common inter-
mediate.
The presence of a common intermediate for the DDS
and HYD reaction pathways in the HDS of DBT was first
suggested by Singhal et al. [34]. As a common intermedi-
ate they suggested 1ꢀ,4ꢀ-dihydrodibenzothiophene, in which
the hydrogenated bond is between the 6-ring and 5-ring
(Scheme 2A). This proposal was supported by other re-
searchers, who suggested that elimination is the main re-
action through which sulfur is removed [10,12,35]. They
assumed that 4,4ꢀ-dihydrodibenzothiophene (Scheme 2B) is
the common intermediate and that biphenyl is the prod-
uct of four consecutive reactions, two hydrogenations, each
followed by elimination (Scheme 3). In essence, they rule
out the possibility that thiophenol can be hydrodesulfur-
ized by hydrogenolysis. The HYD pathway also consists
of hydrogenation and elimination reactions, but there are
Scheme 2. The structures of proposed intermediates in the HDS of DBT [12,35].

M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
425
hydrogenolysis of the two C–S bonds. Therefore, the as-
sumption of a common dihydro-intermediate for both DDS
and HYD reaction pathways seems to be unfounded.
For these reasons and in conformity with other authors
[37–39], we assume that the two HDS pathways do not have
a common intermediate and are determined by the confor-
mation of the adsorbed DBT molecule. DDS occurs through
σ adsorption of the DBT molecule via the sulfur atom, and
HYD proceeds through π adsorption of the reactant via the
aromatic system. This would require two adsorption con-
stants for the DDS and HYD pathways and these constants
may be different.
A vibrational study of organometallic complexes with
thiophene ligands was used to model the adsorption of thio-
phene on a hydrodesulfurization catalyst [40]. It was found
that in the η1(S) bonding mode of thiophene (σ adsorption)
the C–C bonds in the C4 hydrocarbon backbone are slightly
stronger, while the C–S bonds are substantially weaker rel-
ative to free thiophene. Therefore, it was suggested that the
cleavage of the C–S bond occurs when thiophene is adsorbed
on a coordinately unsaturated Mo site via the sulfur atom in
the σ mode. Vecchi et al. [41] described the effect of the
methyl groups on σ coordination in Ru complexes of DBT,
4-methyldibenzothiophene, 4,6-DMDBT, and 2,8-DMDBT.
They observed that the binding strength increased in the or-
der 4,6-DMDBT < 4-methyldibenzothiophene < DBT <
2,8-DMDBT, in good agreement with catalytic experiments
[6,10].
A study of the C–S bond cleavage in methyl-substituted
thiophenes and benzothiophenes in the presence of a binu-
clear Re complex [42] led to the conclusion that thiophenes
and benzothiophenes reacted with Re2(CO)10 to different
products. Therefore, it was suggested that the HDS of thio-
phenes and benzothiophenes may occur via different path-
ways. DFT calculations of different adsorption conforma-
tions of DBT derivatives indicated that the adsorption prop-
erties of the refractory DBT and 4,6-DMDBT molecules are
very different from those of smaller model compounds such
as thiophene and benzothiophene [43,44]. The main reason
for this difference is the aromaticity of the DBT structure,
making π adsorption more likely than σ adsorption. It was
shown that the methyl groups do indeed hinder the perpen-
dicular σ adsorption of DBT but hardly affect the flat ad-
sorption via the aromatic π system.
Scheme 3. Reaction network of the HDS of dibenzothiophene suggested by
Bataille et al. [12].
two additional hydrogenation steps at the beginning. It is
known, however, that the hydrogenation of a carbon–carbon
double bond with three substituents (RXC=CHRꢀ) is more
difficult than that of a double bond with two substituents
(RHC=CHRꢀ) [36]. Furthermore, the stability of the 4,4ꢀ-
dihydro-intermediate (Scheme 2B) is extremely low from an
energetical point of view. This intermediate B still seems to
have a conjugation between the π electrons of the phenyl
group and of the partially hydrogenated ring. In reality,
because of the presence of the two sp3-hybridized carbon
atoms, the phenyl and hexadiene rings are not coplanar. In-
termediate A (Scheme 2) has the cis conformation and, thus,
the plane of the diene fragment (1,2,3,4) is oriented at an an-
gle of 54.7◦ to the plane of the remaining benzothiophene
fragment. The conjugation in this dihydrodibenzothiophene
molecule is interrupted. Finally, in all the published studies
on the HDS of DBTs the only partially hydrogenated inter-
mediate observed was 1,2,3,4-tetrahydrodibenzothiophene,
with the double bond in the bridge position (Scheme 2C).
This suggests that this double bond is the most difficult to
hydrogenate and that such an intermediate is the most sta-
ble. The proposal for a common intermediate for the DDS
and HYD pathways is also based on the assumption that
elimination is the only reaction responsible for C–S bond
breaking [12]. This assumption is much too restrictive, how-
ever, because methanethiol and neopentanethiol can be eas-
ily hydrodesulfurized to H2S and methane and neopentane,
respectively. This shows that hydrogenolysis is a clear al-
ternative to elimination whenever no β-H atoms are avail-
able. This suggests that the HDS of DBT might occur by
The inhibitory effect of H2S on the HDS activity of
CoMo-, NiMo-, and Mo/Al2O3 catalysts was studied be-
fore by Kasahara et al. [45]. They found that H2S decreased
the conversions of benzothiophene and DBT on all the cat-
alysts. Moreover, the inhibitory effect of H2S on NiMo was
stronger than on the CoMo and Mo catalysts. In agreement
with their observations, our results show that the total activ-
ity in the HDS of DBT and 4,6-DMDBT is suppressed to a
greater extent by H2S over the NiMo catalyst. In our work,
the Mo catalyst was least affected by H2S in the HDS of
DBT and of 4,6-DMDBT (Tables 1 and 2). Moreover, the
inhibition by H2S was larger in the HDS of DBT than of

426
M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
4,6-DMDBT (Tables 1 and 2). This is due to the stronger in-
hibitory effect of H2S on the DDS pathway than on the HYD
pathway. Since the contribution of the DDS is larger in the
case of the HDS of DBT, the overall HDS is affected more
strongly in the case of DBT. These results are in agreement
with those of Bataille et al. [12]. They also observed that the
HDS of DBT was affected more strongly by H2S than that
of 4,6-DMDBT over a NiMo catalyst.
[32,50,51]. It was shown that a sulfur atom between a nickel
(or cobalt) and a molybdenum atom is less strongly bonded
than a sulfur atom between two molybdenum atoms and can
be more easily removed, thus creating a vacancy. Therefore,
the promoter may enhance the intrinsic rate of the C–S bond
cleavage by increasing the electron density on the active Mo
sites, as well as by creating a larger number of sulfur vacan-
cies on the Ni- and Co-promoted catalysts.
The equal inhibition by H2S of the DDS pathway and the
final sulfur removal via the HYD pathway, in the HDS of
DBT as well as 4,6-DMDBT over all three catalysts, suggest
that the desulfurization via the DDS and HYD pathways oc-
curs over the same active sites and that a similar mechanism
of sulfur removal is involved in the DDS pathway and in the
final step of the HYD pathway. Thus, we suggest that the
tetrahydro-intermediate can be desulfurized only via σ ad-
sorption on the active site after having been desorbed from
the catalyst surface. Whether the tetrahydro-intermediate is
desulfurized by elimination or hydrogenolysis is unknown.
Assuming that, as usual, elimination occurs only when the
β-H atom is in trans position to the S atom on the α-carbon
atom, then elimination is not possible for sulfur atoms in
ring structures and hydrogenolysis would be responsible for
breaking the first C–S bond in the thiophene ring. Hermann
et al. also suggested that common steps are involved in both
HDS pathways [46]. They found that the activities of the
transition metal sulfides in the HDS of 4,6-DMDBT fol-
lowed the same trends as in the HDS of DBT, although the
DDS pathway was of major importance for DBT, while the
HYD pathway was prevalent for 4,6-DMDBT.
The final desulfurization of the hindered 4,6-DMDBT via
HYD is substantially slower than that of DBT (Tables 1
and 2). Therefore, the methyl groups suppress not only the
σ adsorption of the reactant, but also of the partially hy-
drogenated intermediate. The DDS pathway is retarded to
a greater extent by the presence of the methyl groups than
is the sulfur removal via the HYD pathway, especially over
the CoMo catalyst, as shown by the lower the kDDS/kHYD
values (Section 3.5). This may be due to the lower fleDxiEbSiUliLtyF
in DBT and 4,6-DMDBT than in their partly hydrogenated
intermediates, making adsorption easier for the latter mole-
cules.
We suggest that the DDS active sites and the sites at
which sulfur is removed from the partially hydrogenated in-
termediates are the same and constitute a sulfur vacancy on
the catalyst surface. Therefore, H2S has a strong inhibitory
influence on the sulfur removal via both reaction pathways.
The hydrogenation of DBT and 4,6-DMDBT, on the other
hand, is hardly affected by H2S (Tables 1 and 2). This sug-
gests that vacancies are not required for the flat π adsorption
of the S-containing molecules and that hydrogenation may
even occur over the surface, which is completely covered
with sulfur. Another possibility is that the DDS active sites
are located at the edges of the MoS2 slabs and that the HYD
active sites are at the corners, which are less sensitive to the
presence of H2S and allow the π adsorption of the spacious
molecules.
5. Conclusion
The data obtained in this study indicate that the HDS of
DBT and 4,6-DMDBT occurs via the same reaction network
over NiMo, CoMo, and Mo catalysts. Ni and Co promoters
noticeably improve the DDS activity of the Mo catalyst in
the HDS of DBT and, to a lesser extent, in the HDS of 4,6-
DMDBT. Since the DDS is the main reaction pathway in the
HDS of DBT, the overall activity of the Mo catalyst in the
transformation of DBT is remarkably improved by promo-
tion. We suggest that the DDS and HYD pathways of HDS
are determined by the adsorption conformation of the reac-
tant molecule on the catalyst surface. The DDS occurs via
σ adsorption, and the HYD pathway requires π adsorption.
The resilience of 4,6-DMDBT is the result of the hindered
removal of sulfur in the DDS pathway as well as in the HYD
pathway.
It is generally assumed that the catalytically active sites
in Mo/γ -Al2O3 hydrotreating catalysts are the molybdenum
atoms at the edges and corners of the MoS2 crystallites,
which have at least one sulfur vacancy, which enables chem-
ical adsorption of the reacting molecule on the molybdenum
atom [9,47–49]. Upon the addition of nickel and cobalt, the
HDS and HDN activities of a MoS2/γ -Al2O3 catalyst in-
crease substantially. Density-functional theory (DFT) calcu-
lations show that the most stable position for the promoter
atom (Co or Ni) is at the edge; there it substitutes the molyb-
denum atom [50] and forms the so-called Co–Mo–S phase
(Ni–Mo–S phase for Ni–Mo catalysts) [9]. DFT calculations
suggest that the combined action of the promoter (Ni or Co)
and the molybdenum atom is responsible for the catalysis
H2S suppresses the transformation of DBT and 4,6-
DMDBT over all three catalysts. The NiMo catalyst is af-
fected most and the Mo catalyst is least affected by the pres-
ence of H2S. The DDS pathway was inhibited more strongly
by H2S than was the HYD pathway in the HDS of DBT and
4,6-DMDBT over all three catalysts. The final desulfuriza-
tion via the HYD pathway was affected to the same extent as
the DDS pathway, as shown by the constant kDDS/kDHEYSDULF
ratio at all H2S partial pressures over all catalysts. This sug-
gests that the same mechanism is active in the DDS and in
the final HYD step. In the HDS of 4,6-DMDBT the desulfur-
ization via the HYD pathway was inhibited less by H2S over
the CoMo than over the NiMo catalyst. Because the activity
of the CoMo catalyst was lower at low PH S, there is a cross-
2

M. Egorova, R. Prins / Journal of Catalysis 225 (2004) 417–427
427
over point, at which the CoMo catalyst becomes better than
the NiMo catalyst. As a consequence, at partial pressures of
H2S higher than 29 kPa, the CoMo catalyst performed better
in the HDS of 4,6-DMDBT than the NiMo catalyst.
[25] K.G. Knudsen, B.H. Cooper, H. Topsøe, Appl. Catal. A 189 (1999)
205.
[26] M.L. Vrinat, L. De Mourgues, Appl. Catal. 5 (1983) 43.
[27] O. Levenspiel, Chemical Reaction Engineering, Wiley, New York,
1999.
[28] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick, B.C. Gates,
AIChE J. 24 (1978) 1015.
References
[29] M. Nagai, T. Kabe, J. Jpn. Petrol. Inst. 23 (1980) 82.
[30] G.H. Singhal, R.L. Espino, J.E. Sobel, J. Catal. 67 (1981) 446.
[31] M.L. Vrinat, L. De Mourgues, J. Chim. Phys.-Chim. Biol. 79 (1982)
45.
[1] D.R. Kilanowski, H. Teeuwen, V.H.J. de Beer, B.C. Gates, G.C.A.
Schuit, H. Kwart, J. Catal. 55 (1978) 129.
[2] M. Houalla, D.H. Broderick, A.V. Sapre, N.K. Nag, V.H.J. de Beer,
B.C. Gates, H. Kwart, J. Catal. 61 (1980) 523.
[32] L.S. Byskov, J.K. Nørskov, B.S. Clausen, H. Topsøe, J. Catal. 187
(1999) 109.
[3] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[4] A. Ishihara, H. Tajima, T. Kabe, Chem. Lett. 4 (1992) 669.
[5] T. Kabe, A. Ishihara, H. Tajima, Ind. Eng. Chem. Res. 31 (1992) 1577.
[6] T. Kabe, A. Ishihara, Q. Zhang, Appl. Catal. A 97 (1993) L1.
[7] V. Lamure-Meille, E. Schulz, M. Lemaire, M. Vrinat, Appl. Catal.
A 131 (1995) 143.
[8] X.L. Ma, K. Sakanishi, I. Mochida, Ind. Eng. Chem. Res. 35 (1996)
2487.
[9] H. Topsøe, B.S. Clausen, F.E. Massoth, Catalysis: Science and Tech-
nology, Springer, Berlin, 1996.
[33] H. Schweiger, P. Raybaud, H. Toulhoat, J. Catal. 212 (2002) 33.
[34] G.H. Singhal, R.L. Espino, J.E. Sobel, G.A. Huff, J. Catal. 67 (1981)
457.
[35] J. Mijoin, G. Perot, F. Bataille, J.L. Lemberton, M. Breysse, S.
Kasztelan, Catal. Lett. 71 (2001) 139.
[36] M.B. Smith, J. March, Advanced Organic Chemistry: Reactions,
Mechanisms, and Structure, Wiley–Interscience, New York, 2001.
[37] P. Desikan, C.H. Amberg, Can. J. Chem.-Rev. Can. Chim. 42 (1964)
843.
[38] C.N. Satterfield, M. Modell, J.F. Mayer, AIChE J. 21 (1975) 1100.
[39] S. Harris, R.R. Chianelli, J. Catal. 86 (1984) 400.
[40] P. Mills, S. Korlann, M.E. Bussell, M.A. Reynolds, M.V. Ovchinnikov,
R.J. Angelici, C. Stinner, T. Weber, R. Prins, J. Phys. Chem. A 105
(2001) 4418.
[41] P.A. Vecchi, A. Ellern, R.J. Angelici, J. Am. Chem. Soc. 125 (2003)
2064.
[42] M.A. Reynolds, I.A. Guzei, R.J. Angelici, J. Am. Chem. Soc. 124
(2002) 1689.
[10] V. Meille, E. Schulz, M. Lemaire, M. Vrinat, J. Catal. 170 (1997) 29.
[11] T. Kabe, A. Ishihara, W. Qian, Hydrodesulfurization and Hydrodeni-
trogenation, Kodansha, Tokyo, Wiley–VCH, Weinheim, 1999.
[12] F. Bataille, J.L. Lemberton, P. Michaud, G. Perot, M. Vrinat, M.
Lemaire, E. Schulz, M. Breysse, S. Kasztelan, J. Catal. 191 (2000)
409.
[13] D.D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 42 (1998) 345.
[14] M. Macaud, A. Milenkovic, E. Schulz, M. Lemaire, M. Vrinat,
J. Catal. 193 (2000) 255.
[15] M. Egorova, R. Prins, J. Catal. 221 (2004) 11.
[43] H. Yang, C. Fairbridge, Z. Ring, Energy Fuels 17 (2003) 387.
[44] S. Cristol, J.F. Paul, E. Payen, D. Bougeard, F. Hutschka, S. Clémen-
dot, J. Catal. 224 (2004) 138.
[45] S. Kasahara, T. Shimizu, M. Yamada, Catal. Today 35 (1997) 59.
[46] N. Hermann, M. Brorson, H. Topsøe, Catal. Lett. 65 (2000) 169.
[47] R. Prins, V.H.J. de Beer, G.A. Somorjai, Catal. Rev.-Sci. Eng. 31
(1989) 1.
[48] R.R. Chianelli, M. Daage, M.J. Ledoux, Adv. Catal. 40 (1994) 177.
[49] P. Grange, X. Vanhaeren, Catal. Today 36 (1997) 375.
[50] P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan, H. Toulhoat, J.
Catal. 190 (2000) 128.
[51] P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan, H. Toulhoat, J.
Catal. 189 (2000) 129.
[16] M. Egorova, R. Prins, J. Catal. 224 (2004) 278.
[17] B.C. Gates, H. Topsøe, Polyhedron 16 (1997) 3213.
[18] T. Kabe, K. Akamatsu, A. Ishihara, S. Otsuki, M. Godo, Q. Zhang,
W.H. Qian, Ind. Eng. Chem. Res. 36 (1997) 5146.
[19] P. Steiner, E.A. Blekkan, Fuel Process. Technol. 79 (2002) 1.
[20] X.L. Ma, K. Sakanishi, T. Isoda, I. Mochida, Ind. Eng. Chem. Res. 34
(1995) 748.
[21] I. Mochida, K. Sakanishi, X.L. Ma, S. Nagao, T. Isoda, Catal. Today 29
(1996) 185.
[22] M.V. Landau, D. Berger, M. Herskowitz, J. Catal. 159 (1996) 236.
[23] E. Lecrenay, K. Sakanishi, I. Mochida, Catal. Today 39 (1997) 13.
[24] D.D. Whitehurst, H. Farag, T. Nagamatsu, K. Sakanishi, I. Mochida,
Catal. Today 45 (1998) 299.