Kinetics of the HDS of 4,6-dimethyldibenzothiophene and its hydrogenated intermediates over sulfided Mo and NiMo on γ-Al2O3

Article abstract of DOI:10.1016/j.jcat.2007.06.005

To study the hydrodesulfurization (HDS) reaction network of 4,6-dimethyldibenzothiophene (DMDBT), three hydrogenated intermediates-tetrahydro-, hexahydro-, and dodecahydro-DMDBT-were synthesized, and their HDS was investigated over sulfided Mo and NiMo on γ-Al₂O₃ catalysts at 300 °C and 5 MPa. Tetrahydro-DMDBT reacted by hydrogenation to hexahydro-DMDBT, which in turn reacted to dodecahydro-DMDBT by hydrogenation and to 3,3′-dimethylcyclohexylbenzene by desulfurization. All four diastereoisomers of hexahydro-DMDBT were observed, all of which interconverted rapidly during HDS. Dodecahydro-DMDBT reacted by desulfurization to 3,3′-dimethylbicyclohexyl. The rate constants of all steps in the kinetic network of the HDS of DMDBT could be measured over Mo/γ-Al₂O₃, and those of some steps could be measured over NiMo/γ-Al₂O₃. The first step-hydrogenation of DMDBT to tetrahydro-DMDBT-is the slowest, rate-determining step, but the hydrogenation of and the sulfur removal from hexahydro-DMDBT are also slow. Opening of the sulfur-containing ring in DMDBT and its hydrogenated intermediates occurs by C{single bond}S hydrogenolysis rather than by elimination.

Full text of DOI:10.1016/j.jcat.2007.06.005

Journal of Catalysis 250 (2007) 283–293
www.elsevier.com/locate/jcat
Kinetics of the HDS of 4,6-dimethyldibenzothiophene and its hydrogenated
intermediates over sulfided Mo and NiMo on γ -Al2O3
Xiang Li a,b, Anjie Wang a,b, Marina Egorova a, Roel Prins a,∗
a
Institute of Chemical and Bioengineering, ETH Zurich, 8093 Zurich, Switzerland
b
State Key Laboratory of Fine Chemicals, Dalian University of Technology, Dalian, China
Received 20 April 2007; revised 4 June 2007; accepted 5 June 2007
Available online 30 July 2007
Abstract
To study the hydrodesulfurization (HDS) reaction network of 4,6-dimethyldibenzothiophene (DMDBT), three hydrogenated intermediates—
tetrahydro-, hexahydro-, and dodecahydro-DMDBT—were synthesized, and their HDS was investigated over sulfided Mo and NiMo on γ -Al O
2
3
◦
catalysts at 300 C and 5 MPa. Tetrahydro-DMDBT reacted by hydrogenation to hexahydro-DMDBT, which in turn reacted to dodecahydro-
ꢀ
DMDBT by hydrogenation and to 3,3 -dimethylcyclohexylbenzene by desulfurization. All four diastereoisomers of hexahydro-DMDBT were
ꢀ
observed, all of which interconverted rapidly during HDS. Dodecahydro-DMDBT reacted by desulfurization to 3,3 -dimethylbicyclohexyl. The
rate constants of all steps in the kinetic network of the HDS of DMDBT could be measured over Mo/γ -Al O , and those of some steps could
2
3
be measured over NiMo/γ -Al O . The first step—hydrogenation of DMDBT to tetrahydro-DMDBT—is the slowest, rate-determining step, but
2
3
the hydrogenation of and the sulfur removal from hexahydro-DMDBT are also slow. Opening of the sulfur-containing ring in DMDBT and its
hydrogenated intermediates occurs by C–S hydrogenolysis rather than by elimination.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Hydrodesulfurization; HDS; Kinetics; 4,6-Dimethyldibenzothiophene; 4,6-DMDBT; Hydrogenated intermediates; Mo/γ -Al O ; NiMo/γ -Al O
2 3
2
3
1. Introduction
dihydro-intermediate, which loses H2S by elimination [5],
whereas others have proposed that the C–S bonds of DMDBT
are broken by hydrogenolysis [6–9]. In the hydrogenation
(HYD) pathway, the reactant molecule is first hydrogenated
to tetrahydro-, hexahydro-, and decahydro-intermediates of
DMDBT (designated TH-DMDBT, HH-DMDBT, and DH-
DMDBT, respectively), the C–S bonds of which are then broken
to form 1-methyl-3-(3-methylcyclohexyl)benzene (designated
3,3ꢀ-dimethylcyclohexylbenzene [DM-CHB]) and 1-methyl-
3-(3-methylcyclohexyl)cyclohexane (or 3,3ꢀ-dimethylbicyclo-
hexyl [DM-BCH]). These are presented in Scheme 1, with all
sulfur-containing compounds connected by (de)hydrogenation
on the top horizontal line and the final hydrocarbon products
on the bottom line. The HDS of DBT occurs mainly though
the DDS pathway, but the HYD pathway dominates the HDS
of DMDBT [1–4]. It is commonly assumed that this is because
of the fact that in the DDS pathway, the reactant needs to ad-
sorb in a σ mode, perpendicular to the catalyst surface. In this
mode, the neighboring methyl groups hinder bonding with the
catalytic sites because they extend further into space than the
By the year 2010, the maximum allowed amount of sulfur
in gasoline and diesel fuel will be only about 10 ppm in many
parts of the world. Deep hydrodesulfurization (HDS) technol-
ogy must be implemented to attain this low sulfur level. This
also means that 4,6-dialkyldibenzothiophene molecules, which
are very difficult to desulfurize because of the alkyl groups ad-
jacent to the sulfur atom, will have to undergo HDS [1–4]. Con-
sequently, 4,6-dimethyldibenzothiophene (DMDBT) is often
used as a model molecule in HDS studies. Extensive research
over metal sulfide catalysts has shown that the HDS of diben-
zothiophene (DBT) and DMDBT occurs through two reaction
pathways. In the direct desulfurization (DDS) pathway, 3,3ꢀ-
dimethylbiphenyl (DM-BP) is formed. Some authors believe
that the DDS of DMDBT occurs via a partially hydrogenated
*
Corresponding author. Fax: +41 44 6321162.
E-mail address: prins@chem.ethz.ch (R. Prins).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.06.005

284
X. Li et al. / Journal of Catalysis 250 (2007) 283–293
Scheme 1. Reaction network of the HDS of DMDBT.
lone pairs of the sulfur atom, which are responsible for the σ
bonding [1–4,10,11].
of DMDBT over the Mo catalyst became rather slow; therefore,
some experiments were carried out at 320 and 340 ◦C.
The gas-phase feed consisted of 1 kPa of reactant (DMDBT,
TH-DMDBT, HH-DMDBT, or DH-DMDBT), 130 kPa of
toluene (solvent for the reactant), 8 kPa of dodecane (GC refer-
ence for the DMDBT derivatives), 35 kPa of H2S, and 4.8 MPa
of H2. Previous tests verified that toluene does not influence
the HDS of DMDBT [13]. For the experiments at different
H2S partial pressures (0–100 kPa), the partial pressure of H2
was varied to keep the total pressure constant at 5 MPa. The
reaction products were analyzed by off-line gas chromatog-
raphy, as described elsewhere [13]. Weight time was defined
as τ = wcat/nfeed, where wcat denotes the catalyst weight and
nfeed denotes the total molar flow to the reactor (1 g min/mol
= 0.15 g h/l). The weight time was changed by varying the
flow rates of the liquid and the gaseous reactants while keeping
their ratio constant. The reaction was stable after 3–4 h; during
the 2 weeks of operation, almost no deactivation of the catalyst
occurred.
The HDS of DMDBT occurs predominantly through the
HYD pathway and thus is dependent on the hydrogenating abil-
ity of the catalyst. Nevertheless, in the HYD pathway C–S
bonds also must be broken; this raises the question of whether
this occurs through hydrogenolysis or through H2S elimina-
tion. Up to now, information on the network of the HDS of
DMDBT has been obtained only by studying DMDBT itself
[1–5]. Sometimes the reaction network itself was discussed
even though none of the TH-DMDBT, HH-DMDBT, and DH-
DMDBT intermediates were observed [12]. But because this
network is complex, studying only the reaction of DMDBT it-
self is not sufficient; the reactions of the intermediates should
be studied as well. Consequently, we have synthesized the
three key intermediates, TH-DMDBT, HH-DMDBT, and DH-
DMDBT, and studied their HDS. The questions that we address
are how fast or slow the different reaction steps in the HDS net-
work occur, how DDS occurs, and how the last part of the HYD
route (the final removal of sulfur from the hydrogenated DBT
intermediates) occurs.
2.2. Preparation of DMDBT intermediates
2. Experimental
TH-DMDBT was synthesized by hydrogenation of DMDBT
over 40 mg of sulfided Mo/γ -Al2O3 catalyst at 320 ◦C, 5 MPa
total pressure, and weight time τ = 4.9 gmin/mol. The reaction
conditions were chosen to maximize the yield of TH-DMDBT
and minimize the loss of sulfur-containing products to hydro-
carbons. The gas-phase feed consisted of 130 kPa of toluene,
2.5 kPa of DMDBT (almost the limit of the solubility in toluene,
45 mg/ml), 20 kPa of H2S, and 4.85 MPa of H2. The selectivity
to TH-DMDBT was 58% at 31% conversion of DMDBT. The
product was separated and purified by crystallization, column
chromatography, and vacuum distillation, as described previ-
ously [14,15]. MS and 1H and 13C NMR spectroscopy charac-
terized the final TH-DMDBT product as 1,2,3,4-tetrahydro-4,6-
DMDBT (Scheme 2A).
HH-DMDBT and DH-DMDBT were prepared by hydro-
genation of DMDBT in a 300-ml stainless steel autoclave at the
lower temperature of 200 ◦C and higher pressure of 15 MPa H2
for 5 h over 10 g of a 10 wt% Pd/C catalyst. The autoclave was
loaded with 10 g of DMDBT, 10 g of catalyst, and 180 ml of
glacial acetic acid. (This solvent is crucial for high conversion.)
At 53% conversion, the selectivity to HH-DMDBT was 79%.
Separation and purification were performed as described previ-
2.1. Catalyst preparation and reaction procedure
Mo/γ -Al2O3 and NiMo/γ -Al2O3 catalysts with 8 wt% Mo
and 0 or 3 wt% Ni were prepared by incipient wetness im-
pregnation of γ -Al2O3 (Condea, pore volume 0.5 cm3 g−1
,
specific surface area 230 m2/g) with an aqueous solution of
(NH4)6Mo7O24·4H2O, followed by an aqueous solution of
Ni(NO3)2·6H2O (all Aldrich) for the promoted catalyst. After
each impregnation step, the catalysts were dried in air at am-
bient temperature for 4 h and then in an oven at 120 ◦C for
15 h, and finally calcined at 500 ◦C for 4 h. Before the reaction,
the catalysts were activated by in situ sulfidation with a mix-
ture of 10% H2S in H2 (50 ml/min) at 400 ◦C and 1 MPa for
4 h. HDS reactions were carried out in a continuous mode in
a fixed-bed Inconel reactor, as described elsewhere [13]. Most
experiments were performed at 300 ◦C, because at higher tem-
peratures, the reactions of TH-DMDBT and HH-DMDBT were
too fast. This also allowed comparison with published results
for the HDS of DMDBT over Pd/Al2O3 [14]. A consequence
of the choice of 300 ◦C was that the rate of the HDS reaction

X. Li et al. / Journal of Catalysis 250 (2007) 283–293
285
Scheme 2. (A) 1,2,3,4-Tetrahydro-4,6-dimethyldibenzothiophene (TH-DMDBT), (B) hexahydro-4,6-dimethyldibenzothiophene (HH-DMDBT), and (C) dodecahy-
dro-4,6-dimethyldibenzothiophene (DH-DMDBT) with chiral carbon atoms indicated by arrows.
ously [15]. HH-DMDBT has three chiral centers (Scheme 2B).
Thus, the hydrogen atoms on the carbon atoms in positions 4,
4a, and 9b can be in either cis or trans configuration relative
to each other, leading to 4 different diastereomers. Three HH-
DMDBT diastereomers could be isolated: 28% isomer A, 9%
isomer B, and 63% isomer C. 1H and 13C NMR spectroscopy,
MS/MS experiments, and X-ray crystal structure determination
[15] revealed that isomer A had a 4,4a-trans;4a,9b-cis config-
uration, isomer B had a 4,4a-trans;4a,9b-trans configuration,
and isomer C had a 4,4a-cis;4a,9b-cis configuration. A small
amount of the fourth diastereomer, D, which must have a 4,4a-
cis;4a,9b-trans configuration, was observed, but it was insuffi-
cient to allow isolation for characterization.
rather low; it decreased from 4.8 to 3.6% with weight time in
the absence of H2S.
H2S had only a slight influence on the conversion of
DMDBT. The yields of TH-DMDBT, HH-DMDBT, and DM-
BCH increased, whereas the yields of DM-BP and DM-CBH
decreased slightly in the presence of 10 kPa H2S. This led to a
small increase in the selectivity to HH-DMDBT and decreases
in the selectivity to DM-CHB and DM-BP (Fig. 1). Whereas
DM-CHB was the main hydrocarbon product in the absence of
H2S, DM-BCH was the major HDS product in the presence of
H2S. Increasing the H2S partial pressure to 100 kPa had little
effect on the HDS of DMDBT over sulfided Mo/γ -Al2O3.
The sulfided Ni–Mo/γ -Al2O3 catalyst exhibited a much
higher HDS activity than the Mo/γ -Al2O3 catalyst in the ab-
sence of H2S. At 300 ◦C, 0 kPa H2S, and τ = 5.2 gmin/mol,
the conversion of DMDBT reached 51% (Fig. 2a). DM-CHB
became the most abundant product at τ > 0.8 gmin/mol, and
its yield reached 32% at τ = 5.2 gmin/mol (Fig. 2a), almost
12 times higher than over Mo/γ -Al2O3. Although conversion
The hydrogenation of DMDBT in the autoclave at 200 ◦C
and 15 MPa over 15 g of a 10 wt% Pd/C catalyst (conver-
sion, 71%; selectivity to DH-DMDBT, 32%) led almost exclu-
sively to the production of one DH-DMDBT diastereomer with
the 4,4a-cis;4a,9b-cis;5a,6-trans;5a,9a-trans;9a,9b-cis configu-
ration of the six chiral centers of the molecule (Scheme 2C).
3. Results
3.1. HDS of DMDBT
The sulfided Mo/γ -Al2O3 catalyst exhibited low HDS ac-
tivity; only 18% of DMDBT was converted at 300 ◦C and
τ = 5.2 gmin/mol in the absence of H2S. The same intermedi-
ates and products were observed as in the HDS of DMDBT over
Pd/γ -Al2O3 [14,16]: DM-BP, the product of the DDS route,
and TH-DMDBT, HH-DMDBT, DH-DMDBT, DM-CHB, and
DM-BCH, the intermediate and final products of the HYD path-
way (Scheme 1). Only trace amounts of DH-DMDBT were
detected. The yields of the other five compounds increased
with weight time. The selectivity to TH-DMDBT decreased
with weight time (Fig. 1), but in all cases TH-DMDBT was
the most abundant compound, with a selectivity >47% at τ =
5.2 gmin/mol and 0 kPa H2S. The selectivity to HH-DMDBT,
the second most abundant product, first increased with weight
time and then leveled off at τ > 2.2 gmin/mol. As shown in
Fig. 1a, these two partially hydrogenated sulfur compounds
represented 89% of the reaction products at short weight time
(τ = 0.8 gmin/mol) and 73% at high weight time. This indi-
cates that sulfur removal is difficult over Mo/γ -Al2O3 and that
under our conditions, the HDS reaction was still in its initial
stage. DM-CHB and DM-BCH, the final products of the HYD
pathway, exhibited continuously increasing selectivity. The se-
lectivity toward the formation of DM-BP, the DDS product, was
(a)
(b)
Fig. 1. Product selectivities in the HDS of DMDBT over Mo/γ -Al O at
2
3
◦
300 C and 0 (a) and 10 (b) kPa H S as a function of weight time ((ꢀ)
2
TH-DMDBT; (a) HH-DMDBT; (Q) DM-BP; (") DM-CHB; (2) DM-BCH).

286
X. Li et al. / Journal of Catalysis 250 (2007) 283–293
(a)
(b)
◦
Fig. 2. Partial pressures of reactant and products in the HDS of DMDBT over Ni–Mo/γ -Al O at 300 C and 0 (a) and 10 (b) kPa H S as a function of weight time
2
3
2
((ꢁ) DMDBT; (ꢀ) TH-DMDBT; (a) HH-DMDBT; (Q) DM-BP; (") DM-CHB; (2) DM-BCH).
(a)
(b)
(c)
(d)
◦
Fig. 3. Product selectivities in the HDS of DMDBT over Ni–Mo/γ -Al O at 300 C and 0 (a), 10 (b), 35 (c) and 100 (d) kPa H S as a function of weight time
2
3
2
((ꢀ)TH-DMDBT; (a) HH-DMDBT; (Q) DM-BP; (") DM-CHB; (2) DM-BCH).
was higher over the Ni–Mo/γ -Al2O3 catalyst than over the
Mo/γ -Al2O3 catalyst, the selectivity to DM-BP was higher over
Ni–Mo/γ -Al2O3 than over Mo/γ -Al2O3, 16.6% at the lowest
weight time and 9.4% at the highest weight time (Fig. 3a).
H2S strongly inhibited the conversion of DMDBT and the
formation of DM-CHB over Ni–Mo/γ -Al2O3 (Fig. 2b). At
τ = 5.2 gmin/mol and 10 kPa H2S, the conversion of DMDBT
and the yield of DM-CHB decreased drastically, to 24 and 10%,
respectively. Similar to Mo/γ -Al2O3, a further increase in the
partial pressure of H2S from 10 to 100 kPa had little influence
on the total conversion of DMDBT. However, the selectivities
to the three HDS products (DM-CHB, DM-BCH, and DM-BP)
decreased with H2S partial pressure, and TH-DMDBT became
the most abundant product at all weight times at 100 kPa H2S
(Fig. 3d). In contrast to the HDS of DMDBT over Mo/γ -
Al2O3, DM-CHB was always the major desulfurized hydro-
carbon product over Ni–Mo/γ -Al2O3 in both the presence and
absence of H2S.
Assuming pseudo-first-order kinetics for the HDS of
DMDBT, we calculated the rate constant of the overall con-
version of DMDBT (kDMDBT) at short weight time. By taking
the selectivities of the pathways into account, we could also cal-
culate the rate constants of the two routes (kHYD and kDDS). The
results over Mo/γ -Al2O3 and Ni–Mo/γ -Al2O3 at different H2S

X. Li et al. / Journal of Catalysis 250 (2007) 283–293
287
Table 1
in the presence of 35 kPa H2S (Fig. 4a). HH-DMDBT was a
primary product, and its selectivity decreased with increasing
weight time; it was 82% at the lowest weight time (Fig. 4b).
Based on its constant (low) selectivity, DM-CHB seemed to be
a primary product as well. The HH-DMDBT intermediate was
hydrogenated to DH-DMDBT, the yield of which was very low
(only 0.7% at the highest weight time), and desulfurized to DM-
BCH. For τ > 1.6 gmin/mol, the major final reaction product
was DM-BCH (Fig. 4).
The conversion of TH-DMDBT over NiMo/γ -Al2O3 at
300 ◦C and in the presence of 35 kPa H2S was faster and
the yield of HH-DMDBT was lower (Fig. 5a) than over the
Mo/γ -Al2O3 catalyst (Fig. 4a). Only a trace of DH-DMDBT
was detected at the highest weight time. However, the total
amount of desulfurized products (DM-CHB and DM-BCH)
at τ = 5.3 gmin/mol was much higher over Ni–Mo/γ -Al2O3
(87%) than over Mo/γ -Al2O3 (36%). The final product distrib-
ution was different over the NiMo/γ -Al2O3 catalyst. The major
reaction product was DM-CHB instead of DM-BCH (Figs. 5b
and 4b, respectively), with a selectivity of 87% at the highest
weight time. These results are in good agreement with our pre-
vious results on the HDS of DMDBT [17].
The pseudo-first-order rate constants for the HDS of DMDBT over Mo/γ -
◦
Al O and Ni–Mo/γ -Al O at 300 C and different H S partial pressures
2
3
2
3
2
Catalyst
Rate constant
(mol/(g min))
H S partial pressure (kPa)
2
0
10
35
100
Mo/γ -Al O
k
k
k
0.036
0.002
0.034
0.035
0.001
0.034
0.038
0.001
0.037
0.049
0.001
0.048
DMDBT
DDS
2
3
HYD
NiMo/γ -Al O
k
k
k
0.19
0.03
0.16
0.080
0.009
0.071
0.076
0.006
0.070
0.050
0.004
0.046
DMDBT
DDS
2
3
HYD
partial pressures are given in Table 1. In the absence of H2S,
the pseudo-rate constant kDMDBT of the overall conversion of
DMDBT over the Ni–Mo/γ -Al2O3 was 5.3 times higher than
that over Mo/γ -Al2O3. However, when the H2S partial pres-
sure was increased to 100 kPa, the values of kDMDBT for the
two catalysts became almost the same, indicating a strong in-
hibitory effect of H2S on the HDS activity of Ni–Mo/γ -Al2O3,
as mentioned above.
3.2. HDS of TH-DMDBT
From the conversion of TH-DMDBT at low weight time and
assuming pseudo-first-order kinetics, we calculated kTH-DMDBT
The partially hydrogenated TH-DMDBT converted rapidly
to HH-DMDBT over the Mo/γ -Al2O3 catalyst at 300 ◦C and
(a)
(b)
◦
Fig. 4. Partial pressures of reactant and products (a) and product selectivities (b) in the HDS of TH-DMDBT over Mo/γ -Al O at 300 C in the presence of 35 kPa
2
3
H S as a function of weight time ((ꢀ) TH-DMDBT; (a) HH-DMDBT; (ꢂ) DH-DMDBT; (") DM-CHB; (2) DM-BCH).
2
(a)
(b)
◦
Fig. 5. Partial pressures of reactant and products (a) and product selectivities (b) in the HDS of TH-DMDBT over Ni–Mo/γ -Al O at 300 C in the presence of 35
2
3
kPa H S as a function of weight time ((ꢀ) TH-DMDBT; (a) HH-DMDBT; (") DM-CHB; (2) DM-BCH).
2

288
X. Li et al. / Journal of Catalysis 250 (2007) 283–293
= 0.8 0.1 and 1.0 0.1 mol/(gmin) for Mo/γ -Al2O3 and
Ni–Mo/γ -Al2O3, respectively.
H2S partial pressures (10 and 35 kPa). Because of their similar
HDS behavior, only the HDS results of isomer C are shown here
(Fig. 6).
3.3. HDS of HH-DMDBT
HH-DMDBT was very reactive in the presence of 10 kPa
H2S (Fig. 6a). A conversion of 65% was already reached at
τ = 0.6 gmin/mol, due mainly to the interconversion of HH-
DMDBT to TH-DMDBT (with a selectivity of almost 90%).
The TH-DMDBT and HH-DMDBT partial pressures leveled
off thereafter. TH-DMDBT, formed by dehydrogenation, was
always the most abundant product, with 56% selectivity at high
weight time (τ = 3.9 gmin/mol and 10 kPa H2S). A trace
of DMDBT was observed as the product of subsequent dehy-
drogenation. The selectivities of DH-DMDBT, DM-BCH, and
DM-CHB all increased with weight time (Fig. 6b), with DM-
BCH being much more abundant than DM-CHB. The conver-
sion of the HH-DMDBT isomer C and the yield of TH-DMDBT
were hardly affected by an increase in the H2S partial pres-
sure from 10 to 35 kPa. However, the yield of DM-BCH at
τ = 2.0 gmin/mol decreased from 15% at 10 kPa H2S to 9%
at 35 kPa H2S. At the same time, the yield of DH-DMDBT in-
creased slightly.
Only two isomers of HH-DMDBT, the 4,4a-trans;4a,9b-cis
isomer A and the 4,4a-cis;4a,9b-cis isomer C, were prepared in
sufficient amounts to carry out HDS experiments. Their HDS
was studied over the Mo/γ -Al2O3 catalyst at 300 ◦C and two
During the HDS reaction, isomerization of the trans–cis and
cis–cis isomers to the other isomers of HH-DMDBT occurred.
GC–MS identified five isomers with mass 218. The MS patterns
of four of these could be identified [15] as those of isomers A
(trans–cis), B (trans–trans), C (cis–cis), and D (cis–trans). An-
other GC peak belonged to mass 218 as well, but the identity
of this isomer X could not be determined. The selectivities of
the isomers, normalized to the total amount of HH-DMDBT,
demonstrate that the trans–cis isomer quickly converted to the
cis–cis isomer, so that equilibrium was approached already at
τ = 1.0 gmin/mol and 10 kPa H2S (Fig. 7a). The cis–cis iso-
mer also quickly converted to the trans–cis and trans–trans
isomers at 10 kPa H2S (Fig. 7b). The distribution of the iso-
mers in the HDS of the trans–cis and cis–cis isomers at high
weight time did not change on an increase of the H2S partial
pressure to 35 kPa; only the selectivity to X decreased.
HH-DMDBT can convert in three different ways over Mo/γ -
Al2O3: (1) by hydrogenation to DH-DMDBT and further
desulfurization to DM-BCH, (2) by dehydrogenation to TH-
DMDBT, and (3) by direct sulfur removal to DM-CHB. Thus,
(a)
(b)
Fig. 6. Partial pressures of reactant and products (a) and product selectivities (b)
in the HDS of HH-DMDBT isomer C over Mo/γ -Al O at 300 C and 10 kPa
H S as a function of weight time ((e) HH-DMDBT (C); (ꢀ) TH-DMDBT;
(ꢂ) DH-DMDBT; (") DM-CHB; (2) DM-BCH).
◦
2
3
2
(a)
(b)
◦
Fig. 7. Distribution of hexahydro-isomers in the HDS of HH-DMDBT isomer A (a) and isomer C (b) over Mo/γ -Al O at 300 C and 10 kPa H S as a function of
2
3
2
weight time.

X. Li et al. / Journal of Catalysis 250 (2007) 283–293
289
Table 2
with mass 192 is 1-methyl-3-(3-methylcyclohexyl)cyclohexene
(3,3ꢀ-dimethylcyclohexylcyclohexene), a desulfurized product
with one double bond. Actually, several isomers with mass 192
were observed; these could be isomers with different double-
bond locations, or different stereoisomers (due to the presence
of chiral centers). The low yields of these isomers did not allow
further analysis. The product with mass 220 must be octahydro-
4,6-dimethyldibenzothiophene (OH-DMDBT), the dehydro-
genated reactant with two double bonds. These double bonds
are most likely located in the thiophene ring. The products with
mass 192 and 220 were not detected above τ = 2.5 gmin/mol.
Four different isomers of DH-DMDBT were observed in the
HDS experiments for TH-DMDBT and HH-DMDBT. The re-
actant in the HDS of DH-DMDBT itself had the 4,4a-cis;4a,9b-
cis;5a,6-trans;5a,9a-trans;9a,9b-cis configuration of the hydro-
gen atoms on the six chiral centers of the molecule [15]. The se-
lectivities of the main five isomers observed in the HDS of this
DH-DMDBT reactant, normalized to the total amount of DH-
DMDBT, showed that the 4,4a-cis;4a,9b-cis;5a,6-trans;5a,9a-
trans;9a,9b-cis isomer isomerized quickly to four other DH-
DMDBT isomers (Fig. 9). The structures of these isomers
were not determined. Equilibrium between the isomers was not
reached at the highest weight time used.
The influence of H2S on the HDS of DH-DMDBT was stud-
ied at 300 ◦C. At this temperature, no dehydrogenation products
were observed except for OH-DMDBT, the product with mass
220. The conversion of DH-DMDBT at τ = 0.3 gmin/mol was
45% in the absence of H2S and 25% in the presence of 100 Pa
of H2S. The main HDS product was 3,3ꢀ-dimethylcyclohexyl-
cyclohexene, the desulfurized alkene with mass 192. Its selec-
tivity decreased from 75% in the absence of H2S to 70% in
the presence of 100 kPa H2S. Assuming the reaction of DH-
DMDBT to be first order with respect to the reactant, and taking
into account the sum of the selectivities of the desulfurized
products (DM-BCH and 3,3ꢀ-dimethylcyclohexylcyclohexene),
we obtained pseudo-HDS rate constants of 1.65, 1.35, 1.09,
and 0.74 mol/(g min) at 0, 10, 35, and 100 kPa H2S, respec-
The pseudo-first-order rate constants for the HDS of HH-DMDBT isomer C
over Mo/γ -Al O at 300 C and different H S partial pressures
◦
2
3
2
Rate constant
(mol/(g min))
H S partial pressure (kPa)
2
10
35
k
k
k
k
2.40 0.10
0.04 0.03
0.07 0.03
2.29 0.10
2.56 0.10
0.07 0.03
0.06 0.03
2.43 0.10
HH-DMDBT
hydrogenation
desulfurization
dehydrogenation
the sums of the selectivities of the products in each pathway and
their extrapolation to τ = 0 quantify the initial proportions of
the three independent routes. Based on these three proportions
and a pseudo-first-order kinetic assumption, the rate constants
kHH-DMDBT, khydrogenation, kdehydrogenation, and kdesulfurization of
HH-DMDBT (C) over Mo/γ -Al2O3 at different H2S partial
pressures were calculated (Table 2).
3.4. HDS of DH-DMDBT
The reactivity of the fully hydrogenated DH-DMDBT mole-
cule over the Mo/γ -Al2O3 catalyst at 300 ◦C and 35 kPa
H2S was much higher than that of HH-DMDBT. The ma-
jor reaction product was DM-BCH (Fig. 8). Two other prod-
ucts were identified by a combination of mass spectrome-
try and chromatography; these had masses of 192 and 220
and behaved as primary products (Fig. 8b). The product
(a)
Fig. 9. Distribution of dodecahydro-isomers in the HDS of 4,4a-cis;4a,9b-
◦
(b)
cis;5a,6-trans;5a,9a-trans;9a,9b-cis-DH-DMDBT over Mo/γ -Al O at 300 C
2
3
as a function of weight time ((F) 4,4a-cis;4a,9b-cis;5a,6-trans;5a,9a-trans;
9a,9b-cis-DH-DMDBT; (Q), (ꢂ) and (2) other unknown DH-DMDBT iso-
mers).
Fig. 8. Partial pressures of reactant and products (a) and product selectivities
◦
(b) in the HDS of DH-DMDBT over Mo/γ -Al O at 300 C and 35 kPa H S
2
3
2
as a function of weight time.

290
X. Li et al. / Journal of Catalysis 250 (2007) 283–293
Table 3
◦
Product distribution in the HDS of DH-DMDBT at 340 C, τ = 0.3 gmin/mol, and different H S partial pressures
2
H S (kPa)
2
Conversion (%)
DM-BCH
M-192
DM-CHB
HH-DMDBT
TH-DMDBT
M-220
DM-DBT
0
92.0
83.1
79.1
63.7
33.0
30.0
27.4
40.1
33.9
25.7
23.7
1.4
7.6
6.5
6.1
3.3
4.0
1.9
3.3
4.2
5.5
3.6
4.7
2.9
0.9
4.4
3.0
3.9
3.6
10
35
100
11.3
11.8
10.2
tively. These findings indicate that the removal of sulfur from
DH-DMDBT is suppressed with increasing partial pressure of
H2S.
time by assuming first-order kinetics and by lumping the in-
fluences of the H2 and H2S pressure into the kinetic constants.
Because DMDBT and its hydrogenated intermediates have dif-
ferent aromaticity, their adsorption constants and thus their ki-
netic order could be different. Once we have synthesized more
TH-DMDBT and HH-DMDBT, we will study this further. For
the moment, note that the rate constants are pseudo-first-order
rate constants that are valid only under the specified conditions.
Nevertheless, these rate constants are valuable for a semiquan-
titative comparison.
To allow a comparison with results obtained in the HDS of
DMDBT at 340 ◦C [17], the HDS of DH-DMDBT was also
studied at 340 ◦C. The DMDBT results had indicated that H2S
has the same inhibitory influence on the sulfur removal step in
the hydrogenation pathway as in the direct desulfurization path-
way, but because the hydrogenation pathway of DMDBT con-
sists of several consecutive and parallel reactions, results for the
sulfur removal step were only semiquantitative. More accurate
results can be obtained by taking the fully hydrogenated inter-
mediate as a reactant, because this intermediate can react only
by desulfurization. The HDS of DH-DMDBT demonstrates that
the total conversion of DH-DMDBT and the sum of the desul-
furized products (DM-BCH, product with mass 192, and DM-
CHB) decreased with increasing partial pressure of H2S (Ta-
ble 3). At the higher temperature of 340 ◦C, DH-DMDBT also
underwent dehydrogenation to HH-DMDBT, TH-DMDBT, and
DMDBT. Desulfurization of HH-DMDBT and/or TH-DMDBT
occurred with the formation of DM-CHB, the yield of which
decreased with increasing H2S partial pressure from 7.6 to
3.3% (Table 3). DMDBT was hardly desulfurized via the di-
rect desulfurization pathway, because the amount of DM-BP
never exceeded 0.1%. The product with mass 192 behaved as a
primary product, and its production was suppressed by H2S al-
most as much as the production of DM-BCH up to 35 kPa H2S.
At 100 kPa H2S, we observed only 1.4% of the product with
mass 192, whereas the amount of DM-BCH was higher than at
lower partial pressure of H2S.
The kinetic results of the HDS of DMDBT and its interme-
diates at low conversion over sulfided Mo/γ -Al2O3 (Scheme 3)
show that the smallest measurable rate constant was obtained
for the reaction of DMDBT to DM-BP. This rate constant
(0.001 mol/(g min)) was 34 times smaller than the next smallest
rate constant, that for the reaction of DMDBT to TH-DMDBT
(0.034 mol/(g min)). Thus, the rate-determining steps in the
HDS of DMDBT are the reactions of DMDBT itself. The
next steps in the hydrogenation part of the HDS network—
the hydrogenation of TH-DMDBT and dehydrogenation of
HH-DMDBT—are more than an order of magnitude faster.
This explains why it appears as if TH-DMDBT and HH-
DMDBT quickly come to equilibrium (Figs. 4 and 6). The
hydrogenation reaction rate constant of the subsequent reac-
tion of HH-DMDBT to DH-DMDBT is, within the uncer-
tainty of the measurement ( 0.01 mol/(gmin)), equal to that
of DMDBT to TH-DMDBT, whereas the rate constant of the
desulfurization reaction from HH-DMDBT to DM-CHB is
about twice as large. The rate constant of the desulfurization
reaction of DH-DMDBT to DM-BCH is the largest of all rate
constants. The rate constant for desulfurization thus strongly
increases from DMDBT (0.001 mol/(g min)) to HH-DMDBT
(0.07 mol/(g min)) to DH-DMDBT (1.35 mol/(g min)). This
must be due to weakening of the C–S bond (going from an
arylthiol bond to an alkanethiol bond), as well as to decreased
4. Discussion
The rate constants presented in Tables 1 and 2 and Scheme 3
were calculated from the experimental data at short weight
◦
Scheme 3. Pseudo-first-order rate constants (in mol/(g min)) in the reaction network of the HDS of DMDBT over Mo/γ -Al O at 300 C and 10 kPa H S.
2
3
2

X. Li et al. / Journal of Catalysis 250 (2007) 283–293
291
steric hindrance. Hydrogenation of a phenyl ring makes the
cyclohexyl ring in the hydrogenated DMDBT intermediate flex-
ible and allows the methyl group to rotate away from the C(4a)–
S–C(5a) plane (cf. Scheme 2A). Consequently, less open sites
on the catalyst surface can bind the hydrogenated intermediates
via the lone pair of the sulfur atom, and breaking of the C–S
bond can occur faster.
Although the amounts of TH-DMDBT and HH-DMDBT
that we had synthesized were sufficient to measure the rate
constants of all reaction steps over Mo/γ -Al2O3, insufficient
material was available for measurements over NiMo/γ -Al2O3.
The rate constants for the reactions of DMDBT to DM-BP
and TH-DMDBT and for the desulfurization of TH-DMDBT
to DM-CHB were larger over NiMo/γ -Al2O3 than over Mo/γ -
Al2O3. This may be due to the more open structure of the Ni
edges of NiMo/γ -Al2O3 compared with the Mo and S edges
of Mo/γ -Al2O3. According to DFT calculations [18–22], a Mo
atom at the Mo edge is coordinated by six sulfur atoms and
one at the S edge is coordinated by four sulfur atoms, but the
Ni atoms on the metal edge of a NiMo catalyst (where the Ni
atoms are preferentially located) are coordinated by four or five
sulfur atoms. The rate constant for the hydrogenation of TH-
DMDBT to HH-DMDBT (1.0 mol/(g min)) is similar to that
over Mo/γ -Al2O3 (0.8 mol/(g min)).
The hydrogenation rate constants for DMDBT and HH-
DMDBT over Mo/γ -Al2O3 and NiMo/γ -Al2O3 are about an
order of magnitude smaller than over Pd/γ -Al2O3 [14]. The de-
hydrogenation rate constants are much smaller as well. Over
Mo/γ -Al2O3, only a trace of DMDBT was observed in the
HDS of TH-DMDBT, and only a trace of HH-DMDBT was
found in the HDS of DH-DMDBT. This demonstrates that the
rate constants for the dehydrogenation of TH-DMDBT and
DH-DMDBT are very small (<0.001 mol/(g min)) over Mo/γ -
Al2O3 under our conditions. Over Pd/γ -Al2O3, on the other
hand, substantial amounts of DMDBT were measured in the
HDS of TH-DMDBT, and in the HDS of DH-DMDBT, not
only HH-DMDBT and TH-DMDBT, but also DMDBT, were
found [14]. Over Pd, the hydrogenation of DMDBT and the
hydrogenation and dehydrogenation of its hydrogenated inter-
mediates were fast, and DMDBT as well as all hydrogenated
intermediates were always observed regardless of the starting
compound. The fast hydrogenation and dehydrogenation over
Pd are probably due to stronger π bonding. The flat terraces of
Pd crystallites offer a better opportunity for extended π bond-
ing and hydrogenation than the curved edges of the MoS2 and
NiMoS2 particles. This favors the demanding hydrogenation of
the first and second phenyl rings of DMDBT. Hydrogenation of
the 4a,9b double bond in 1,2,3,4-TH-DMDBT to HH-DMDBT
and the reverse dehydrogenation require only a small part of
the catalyst surface, however. This explains why the rate con-
stants of these reactions are almost the same over Mo/γ -Al2O3,
NiMo/γ -Al2O3, and Pd/γ -Al2O3.
favors elimination, and high H2 pressure favors hydrogenol-
ysis [24,25]. The mechanism of the C–S bond-breaking in
arylthiols (thiophenol) and cyclic sulfides (thiophene, tetrahy-
drothiophene, DMDBT, TH-DMDBT, HH-DMDBT, and DH-
DMDBT) remains a matter of debate. Elimination and hy-
drogenolysis have not only been proposed for the C–S bond
breaking in alkanethiols, but also for the breaking of the aryl-
thiol C–S bond in thiophenol and the C–S sulfide bond in
DBT and DMDBT. In the elimination mechanism, a phenyl
ring must be hydrogenated before desulfurization by elimina-
tion of the SH group and a β-H atom can occur. A hydro-
genated intermediate as a common precursor for the DDS and
HYD pathways was first proposed by Singhal et al. [26]. How-
ever, this hydrogenation–elimination DDS mechanism needs a
highly energetic 4,4a- or 4a,9b-dihydro-DMDBT compound as
an intermediate and requires that H2S elimination occur by the
less-favorable syn elimination [5,27]. Another difficulty is that
hydrogenation of a double bond requires π adsorption, which
makes it difficult to understand why the methyl groups consti-
tute a steric problem in the DDS of DMDBT.
An alternative DDS mechanism assumes that the C–S bond
is broken by hydrogenolysis [6–9]. Hydrogenolysis is known to
occur because it is the only possible route for breaking the C–S
bond of methanethiol, which quickly desulfurizes to methane
over MoS2 and WS2 [28]. DFT calculations have confirmed that
hydrogenolysis of CH3SH to CH4 [8], C2H5SH to C2H6 [9],
2,5-dihydrothiophene to 2-butene [29], and C6H5SH to C6H6
[30] have relatively low activation energies. Studies in homo-
geneous solution have shown that organometallic complexes
can insert into the C–S bonds of thiophene, benzothiophene,
and DBT [31–34] (Scheme 4). HDS can then occur via M–C
cleavage of the resulting metallathiacycle to a thiolate com-
plex, which is known to desulfurize more easily than a cyclic
sulfide. All of these results from heterogeneous and homoge-
neous catalysis studies indicate that hydrogenolysis is a likely
mechanism for HDS.
Another mechanism for the breaking of the first C–S bond
in cyclic sulfides could involve acid-catalyzed ring opening by
H2S to a dithiol or by a metal-sulfide fragment on the catalyst
surface to a dithiolate as the first step. Such a reaction is anal-
ogous to the acid-catalyzed reactions of ethers to alcohols and
is used in the synthesis of tetrahydrofurans to tetrahydrothio-
phenes and pyrrolidines [35]. In this way, tetrahydrothiophene
could react with H2S on a metal cation site (Lewis acid) to
1,4-butanedithiol, which quickly reacts to a mixture of bu-
tane (by double hydrogenolysis), butene (by elimination and
hydrogenolysis or by double elimination and hydrogenation),
and butadiene (by double elimination). Similarly, HH-DMDBT
could open the sulfur-containing ring to 4a,5a-dithio-DM-CHB
Metal-sulfide catalysts can break the C–S bond in aliphatic
alkanethiols in two ways, by elimination and by hydrogenoly-
sis, as demonstrated by experimental [23] and theoretical [9]
studies of the HDS of ethanethiol by elimination to ethene
and by hydrogenolysis to ethane. Low (atmospheric) pressure
Scheme 4. Reaction of a cyclic sulfide to a metal thiolate via insertion of a metal
hydride complex in an C–S bond.

292
X. Li et al. / Journal of Catalysis 250 (2007) 283–293
Scheme 5. Formation of a dithiolate complex at the catalyst surface as interme-
diate for opening of the thiophene ring in DMDBT.
Scheme 7. Possible intermediates in the HDS of 1,2,3,4-tetrahydro-dimethyldi-
benzothiophene (TH-DMDBT).
Scheme 6. Direct desulfurization of 4,6-dimethyldibenzothiophene (DMDBT)
ꢀ
to 3,3 -dimethylbiphenyl (DM-BP).
to the question of whether C–S bond-breaking occurs in TH-
DMDBT or whether this molecule is first hydrogenated to HH-
DMDBT, which then undergoes C–S bond-breaking. 1,2,3,4-
TH-DMDBT cannot undergo direct elimination, but a facile
double-bond-shift isomerization from 1,2,3,4-TH-DMDBT to
2,3,4,4a-TH-DMDBT would allow elimination (Scheme 7).
Hydrogenolysis of the C–S bond in TH-DMDBT is expected
to be slightly easier than in DMDBT, because the C(4a)–S
bond might be slightly easier to break in TH-DMDBT than in
DMDBT, because of a weaker resonance effect between a sul-
fur atom and a cyclohexene ring compared with that between a
sulfur atom and a phenyl ring. In principle, HH-DMDBT and
DH-DMDBT can undergo elimination directly, and the first
C–S bond-breaking in these molecules should be faster than
in TH-DMDBT and DMDBT. Moreover, C–S bond-breaking
of HH-DMDBT and DH-DMDBT by hydrogenolysis should
be faster, because C(sp3)–S bonds are weaker than C(sp2)–S
bonds.
It is not clear how the first C–S bond-breaking in HH-
DMDBT and DH-DMDBT occurs. If it occurs by elimination,
then molecules containing a double bond should be formed.
HH-DMDBT should react to 1-methyl-3-(3-methylcyclohexe-
nyl)benzene [also designated 1-methyl-3-(3-methylphenyl)-
cyclohexene or 3,3ꢀ-dimethylphenylcyclohexene], and DH-
DMDBT should react to 3,3ꢀ-dimethylbicyclohexenyl, with
a double bond in each of the two methylcyclohexyl rings.
Neither of these two molecules was observed in our ex-
periments. It could be that 3,3ꢀ-dimethylphenylcyclohexene
is rapidly hydrogenated to DM-CHB. Unfortunately, 3,3ꢀ-di-
methylphenylcyclohexene and even phenylmethylcyclohexene
are not commercially available. Instead, we added phenylcy-
clohexene to the HDS of DBT and checked whether it re-
acts too rapidly to allow observation under HDS conditions.
Although phenylcyclohexene reacted quite rapidly, we were
able to observe it. Thus, it seems that fast hydrogenation of
3,3ꢀ-dimethylphenylcyclohexene to DM-CHB is not the rea-
son for the absence of 3,3ꢀ-dimethylphenylcyclohexene in the
HDS of DMDBT. Rather, it appears as if ring opening of a
cyclic sulfide is slower by elimination than by hydrogenoly-
sis and that an alkanethiol (or alkylthiolate), rather than an
alkenylthiol, is predominantly formed. Acyclic alkanethiols,
on the other hand, can undergo elimination, which would ex-
plain why the major product in the HDS of DH-DMDBT is
3,3ꢀ-dimethylcyclohexylcyclohexene. The first C–S breaking
occurs inside a ring due to hydrogenolysis, whereas the second
and DH-DMDBT to 4a,5a-dithio-DM-BCH. These dithiols will
undergo fast desulfurization by hydrogenolysis or by a com-
bination of hydrogenolysis and elimination. The ring opening
also could occur by reaction of the substrate with a metal atom
and a sulfur atom at the catalyst surface to a dithiolate structure
(cf. Scheme 5 for HH-DMDBT). The metal–dithiolate com-
plex can react directly by hydrogenolysis of both C–S bonds.
This acid-catalyzed ring opening is less likely for aromatic sul-
fides such as DBT and DMDBT, because aromatic nucleophilic
substitution is much more difficult than aliphatic nucleophilic
substitution [36].
DDS is the minor route for the removal of sulfur from
DMDBT (Scheme 3). The DDS reaction must be a two-step
process, with first a reaction to 2-methyl-6-(3-methylphenyl)-
thiophenol (3,3ꢀ-dimethyl-2-thiobiphenyl) and then a reaction
to DM-BP (Scheme 6). The second step must be fast, be-
cause an intermediate thiophenol has never been observed
in the HDS of either DBT or DMDBT. Moreover, in the
HYD route, one C–S bond in a cyclic sulfide must be broken.
Whichever C–S bond-breaking mechanism prevails, it most
likely will be easier to break the first C–S bond of the hy-
drogenated intermediates TH-DMDBT, HH-DMDBT, and DH-
DMDBT than that of DMDBT. The reason for this is that in
the hydrogenation–elimination mechanism, hydrogenation to a
dihydro-intermediate is not needed, whereas in the hydrogenol-
ysis mechanism, the C–S bond should be more easily cleaved
in an alkanethiol than in an arylthiol.
Our experimental results seem to indicate that the rate of
C–S bond-breaking of TH-DMDBT over Mo/Al2O3 is greater
than that of HH-DMDBT, because the DM-CHB selectivity ex-
trapolates to 12% at τ = 0 (Fig. 4b) in the former compared
with only 3% in latter (Fig. 6b). However, TH-DMDBT and
HH-DMDBT react mainly by interconversion, and thus the un-
certainties in the much smaller rate constants for C–S bond
breaking are substantial. Actually, the results of the HDS of
TH-DMDBT and HH-DMDBT are quite similar if we omit
the yields of HH-DMDBT from TH-DMDBT and TH-DMDBT
from HH-DMDBT. The yields of DM-BCH and DM-CHB are
about equal (Figs. 4a and 6a); only the yield of DH-DMDBT is
different. It is lower in the HDS of TH-DMDBT, because DH-
DMDBT is then a secondary product, not a primary product.
Our experimental results do not allow an unequivocal answer

X. Li et al. / Journal of Catalysis 250 (2007) 283–293
293
C–S breaking of the acyclic cyclohexanethiol occurs mainly by
elimination.
[5] F. Bataille, J.L. Lemberton, P. Michaud, G. Pérot, M.L. Vrinat, M.
Lemaire, E. Schulz, M. Breysse, S. Kasztelan, J. Catal. 191 (2000) 409.
[6] H. Schulz, M. Schon, N.M. Rahman, Stud. Surf. Sci. Catal. 27 (1986) 201.
[7] S. Cristol, J.F. Paul, E. Payen, D. Bougeard, J. Hafner, F. Hutschka, Stud.
Surf. Sci. Catal. (1999) 327.
[8] T. Todorova, R. Prins, Th. Weber, J. Catal. 236 (2005) 190.
[9] T. Todorova, R. Prins, Th. Weber, J. Catal. 246 (2007) 109.
[10] 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.
[11] T. Kabe, A. Ishihara, H. Tajima, Ind. Eng. Chem. Res. 31 (1992) 1577.
[12] V. Vanrysselberghe, R. Le Gall, G.F. Froment, Ind. Eng. Chem. Res. 37
(1998) 1235.
[13] M. Egorova, R. Prins, J. Catal. 224 (2004) 278.
[14] A. Röthlisberger, R. Prins, J. Catal. 235 (2005) 229.
[15] P. Kukula, V. Gramlich, R. Prins, Helv. Chim. Acta 89 (2006) 1623.
[16] A. Niquille-Röthlisberger, R. Prins, J. Catal. 242 (2006) 207.
[17] M. Egorova, R. Prins, J. Catal. 225 (2004) 417.
[18] L.S. Byskov, J.K. Nørskov, B.S. Clausen, H. Topsøe, J. Catal. 187 (1999)
109.
The HYD pathways in the HDS of DBT and DMDBT over
NiMo/γ -Al2O3 are strongly inhibited by 2-methylpiperidine
and 2-methylpyridine. In their presence, cyclohexylbenzene
and DM-CHB, the final products of the HYD pathway, have
been observed in small amounts only [37]. In the present
study, we confirmed the strong inhibition by adding 0.1 kPa
2-methylpiperidine during the HDS of 1 kPa HH-DMDBT (iso-
mer C) over Mo/γ -Al2O3 at 300 ◦C. At high weight time, only
1.5% conversion of HH-DMDBT occurred. This demonstrates
a correlation between the ability to break the C–S bond of DBT
and DMDBT and the ability to break the C–S bond of their hy-
drogenated intermediates.
5. Conclusion
[19] P. Raybaud, J. Hafner, G. Kresse, S. Kasztelan, H. Toulhoat, J. Catal. 190
(2000) 128.
[20] S. Cristol, J.F. Paul, E. Payen, D. Bougeard, S. Clémendot, F. Hutschka, J.
Phys. Chem. B 104 (2000) 11220.
[21] H. Schweiger, P. Raybaud, H. Toulhoat, J. Catal. 212 (2002) 33.
[22] M. Sun, A.E. Nelson, J. Adjaye, J. Catal. 226 (2004) 32.
[23] P. Kieran, C. Kemball, J. Catal. 4 (1965) 380.
[24] Y. Zhao, P. Kukula, R. Prins, J. Catal. 221 (2004) 441.
[25] Y. Zhao, R. Prins, J. Catal. 229 (2005) 213.
[26] G.H. Singhal, R.L. Espino, J.E. Sobel, G.A. Huff, J. Catal. 67 (1981) 457.
[27] V. Meille, E. Schulz, M. Lemaire, M. Vrinat, J. Catal. 170 (1997) 29.
[28] R.L. Wilson, C. Kemball, J. Catal. 3 (1964) 426.
[29] P.G. Moses, B. Hinnemann, H. Topsøe, J.K. Nørskov, J. Catal. 248 (2007)
188.
[30] Th. Weber, J.A.R. van Veen, Catal. Today, in press.
[31] M.D. Curtis, S.H. Drucker, J. Am. Chem. Soc. 119 (1997) 1027.
[32] D.A. Vicic, W.D. Jones, Organometallics 17 (1998) 3411.
[33] R.L. Angelici, in: Th. Weber, R. Prins, R.A. van Santen (Eds.), Transition
Metal Sulfides, Chemistry and Catalysis, Kluwer Academic, Dordrecht,
1998, p. 89.
The tetrahydro-, hexahydro-, and decahydro-intermediates
in the HDS of DMDBT were synthesized and their rates of
reaction under HDS conditions were measured. The slowest,
rate-determining steps in the HDS of DMDBT turned out to be
the hydrogenation and hydrogenolysis of DMDBT itself. The
hydrogenation of TH-DMDBT to HH-DMDBT, its reverse de-
hydrogenation, and the desulfurization of DH-DMDBT were
the fastest reactions in the network of the HDS of DMDBT. The
correlation between the ease of breaking the C–S bond of DBT
and DMDBT and that of breaking the C–S bond of their hydro-
genated intermediates suggests that all C–S bonds are broken
in the same way. Ring opening of the dihydrothiophene ring in
HH-DMDBT and the tetrahydrothiophene ring in DH-DMDBT
by hydrogenolysis rather than by elimination explains why the
resulting products do not contain a double bond.
References
[34] C. Bianchini, A. Meli, in: Th. Weber, R. Prins, R.A. van Santen (Eds.),
Transition Metal Sulfides, Chemistry and Catalysis, Kluwer Academic,
Dordrecht, 1998, p. 129.
[1] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick, B.C. Gates, AIChE
J. 24 (1978) 1015.
[35] Y. Ono, Heterocycles 16 (1981) 1755.
[36] M.B. Smith, J. March, Advanced Organic Chemistry, fifth ed., Wiley, New
York, 2001, p. 850.
[2] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[3] D.D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 42 (1998) 345.
[4] T. Kabe, A. Ishihara, W. Qian, Hydrodesulfurization and Hydrodenitro-
genation, Kodansha, Wiley–VCH, Weinheim, Tokyo, 1999.
[37] M. Egorova, R. Prins, J. Catal. 241 (2006) 162.