Alkyldibenzothiophenes hydrodesulfurization-promoter effect, reactivity, and reaction mechanism

Article abstract of DOI:10.1006/jcat.1999.2790

The promoter effect of Co or Ni on the HDS activity of sulfided Mo/alumina was studied using dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (46DMDBT) as reactants at 340°C under a 4 MPa total pressure in a fixed-bed microreactor. The promoted and unpromoted catalysts had different characteristics concerning the two main pathways, i.e., direct desulfurization and hydrogenation (HYD), of the HDS of DBT and 46DMDBT. The origin of differences in reactivity of these compounds to steric effects upon adsorption on the catalytic surface lay in the kinetics of transformation of these two compounds. The DDS and HYD pathways could be decomposed into H₂ addition steps and C-S bond cleavage by elimination steps. The rate-determining step might be different depending on the reactant and on the catalyst. On the unpromoted catalyst, DBT and 46DMDBT had comparable reactivities because the C-S bond cleavages were rate-limiting for both reactants. On the promoted catalysts, DDS became the main pathway for the HDS of DBT. The HYD of DBT into dihydrodibenzothiophene was the rate-limiting step for the two pathways. With 46DMDBT, the influence of the promoter on C-S bond cleavage was limited due to steric constraints. Thus, C-S bond cleavage remained the rate-limiting step, particularly for the DDS pathway. The presence of the methyl groups in 46DMDBT changed significantly the reactivity concerning the two pathways. The low reactivity of 46DMDBT was due to the inhibition of the DDS pathway. Several explanations were proposed for this lower reactivity. The promoter improved the C-S bond cleavage activity of the MoS₂ on alumina catalyst by increasing the basicity of certain sulfur anions shared between the Mo and the promoter, e.g., Ni or Co.

Full text of DOI:10.1006/jcat.1999.2790

Journal of Catalysis 191, 409–422 (2000)
doi:10.1006/jcat.1999.2790, available online at http://www.idealibrary.com on
Alkyldibenzothiophenes Hydrodesulfurization-Promoter Effect,
Reactivity, and Reaction Mechanism
�
�
�
�
,1
Fr e´ d e´ ric Bataille, Jean-Louis Lemberton, Philippe Michaud, Guy P e´ rot, Michel Vrinat,†
Marc Lemaire,‡ Emmanuelle Schulz,‡ Mich e` le Breysse,§ and Slavik Kasztelank
�
Laboratoire de Catalyse en Chimie Organique, Universit e´ de Poitiers, 40 avenue du Recteur Pineau, 86022 Poitiers Cedex, France; †Institut de
Recherches sur la Catalyse, 2 avenue Albert Einstein, 69626 Villeurbanne Cedex, France; ‡Laboratoire de Catalyse et Synth e` se Organique,
IRC, UCBL, CPE, 43 bd du 11 novembre 1918, 69622 Villeurbanne Cedex, France; §Laboratoire de R e´ activit e´ de Surface, Universit e´
Pierre et Marie Curie,Tour 54-55, Casier 178, 2 e` e´ tage, 4 place Jussieu, 75252 Paris Cedex 05, France; and kInstitut Fran c¸ ais du
P e´ trole, 1&4 avenue de Bois-Pr e´ au, 92852 Rueil-Malmaison Cedex, France
Received September 20, 1999; revised December 13, 1999; accepted December 13, 1999
tation of more stringent specifications in many countries.
The promoter effect of Co or Ni on the hydrodesulfurization The gas chromatographic analysis of hydrotreated gas oils
HDS) activity of Mo/alumina was studied by using dibenzothio- compared to that of the corresponding straight-run gas
(
phene (DBT) and 4,6-dimethyldibenzothiophene (46DMDBT) as
oils shows clearly that alkyldibenzothiophenes and par-
ticularly 4,6-dimethyldibenzothiophene (46DMDBT) are
the sulfur impurities that are the most difficult to de-
compose (1–4). Actually, on typical CoMo/alumina and
NiMo/alumina hydrotreating catalysts, 46DMDBT was
found to be much less reactive than dibenzothiophene
�
reactants. The reaction was carried out at 340 C under a 4 MPa
total pressure in a fixed-bed microreactor. On the Mo/alumina cata-
lyst, both reactants had similar reactivities, 46DMDBT being even
slightly more reactive than DBT. However, as generally observed,
on the CoMo/alumina and NiMo/alumina catalysts, DBT was much
more reactive (5 to 6 times) than 46DMDBT. This was mainly be-
cause of a tremendous enhancement of the rate of the so-called “di-
rect desulfurization” (DDS) pathway of the HDS of DBT, whereas
(
DBT) (5–9). It has been suspected for a long time that
the low reactivity of 46DMDBT was due to steric effects
for 46DMDBT this effect was much more limited. It was therefore on adsorption (5, 6). However, recently, the measurement
concluded that the main effect of the promoter on the HDS of DBT- of the heat of adsorption of DBT and of alkyldibenzothio-
type molecules was to increase the rate of the C–S bond cleavage phenes on CoMo/alumina catalysts (7) as well as compet-
provided this reaction was not hindered by steric constraints. This
effect was attributed to an enhancement by the promoter of the
basicity of certain sulfur anions in its vicinity. It was also shown
that the lower reactivity of 46DMDBT compared to that of DBT
measured on the promoted catalysts could not be attributed to dif-
ferences in the adsorption strength of the reactants. Assuming that
C–S bond cleavage occurred through a �-elimination process, sev-
eral other explanations for the low reactivity of 46DMDBT were
proposed and discussed: (a) steric hindrance of the adsorption of
itive hydrodesulfurization (HDS) experiments with DBT
and 46DMDBT (8) indicated that the low reactivity of
4
6DMDBT was not due to an inhibition of the reactant
adsorption, but to steric hindrance of the C–S bond scis-
sion in the adsorbed sulfur compound, although the exact
origin of this inhibition remains unknown.
The product distribution obtained in DBT and in
4
6DMDBT HDS over typical CoMo and NiMo/alumina
the dihydrointermediates by the methyl groups; (b) steric hindrance catalysts shows that the reaction gives essentially two
by the methyl groups of the C–S bond cleavage; (c) the fact that only families of products: biphenyl-type compounds and tetra-
oneH atom isavailablefor theC–Sbond cleavage;(d)an effect ofthe hydrodibenzothiophene-type compounds (9–12). The lat-
methyl group on the acidity of the H atom involved in the elimina-
tion step. Proposals were also made concerning the catalytic centers
involved in the hydrogenation steps and in the C–S bond cleavage
ter lead in turn to cyclohexylbenzene-type products. More-
over, it has been shown that, under HDS conditions (i.e., in
the presence of an organic sulfur compound), biphenyl-type
steps.
�c 2000 Academic Press
compounds do not hydrogenate readily into cyclohexylben-
zene (9, 13–15). Despite the fact that this point is still ques-
tioned especially with NiMo catalysts (4, and references
therein), it was concluded that the HDS of DBT-type com-
poundsoccurred through two parallelreactionsasindicated
in Scheme 1: (i) direct desulfurization (DDS) which yields
biphenyl-type compounds, and (ii) desulfurization through
hydrogenation (HYD) which gives first tetrahydrodiben-
zothiophene and then cyclohexylbenzene-type compounds.
INTRODUCTION
The removal of sulfur from gasoline and diesel oil is be-
coming more and more necessary due to the implemen-
1
To whom correspondence should be addressed.
409
0021-9517/00 $35.00
Copyright �c 2000 by Academic Press
All rights of reproduction in any form reserved.

4
10
BATAILLE ET AL.
SCHEME 1. Hydrodesulfurization of DBT (R = H) or 46DMDBT
(
R = CH3) on sulfided NiMo/alumina catalyst. DDS, direct desulfurization
pathway; HYD, hydrogenation pathway.
However, depending on the reactant (DBT or 46DMDBT),
the contribution of the two pathways to the overall HDS
wasverydifferent. Under conventionalHDSconditions(9),
the DDS pathway contributed 80% to the overall HDS of
DBT, while only 20% to the HDS of 46DMDBT. It was also
found that the presence of methyl groups in the 4 and 6 po-
sitions in 46DMDBT inhibited the DDS pathway, whereas
the HYD pathway was hardly affected (7–9).
For the DDS pathway, one way to obtain a C–S bond
cleavage ending with two phenyl rings in the product is to
hydrogenate one of the double bonds in the vicinity of the
sulfur atom to obtain a dihydrogenated product and then
to open the C–S bond by an elimination process (8, 10,
Scheme 2). We note that this particular double bond is not
necessarily the easiest to be reduced. The second C–S bond
cleavage leading to the biphenyl compound possibly occurs
SCHEME 3. Reaction scheme for the hydrodesulfurization of DBT.
through the same mechanism. Accepting this mechanism, second C–Sbond cleavage in the HYD pathwaydoesnot re-
we will discuss in this paper the possible reasons why the quire the second aromatic ring to be fully hydrogenated and
existence of methyl groups in the 4 and 6 positions in DBT may occur through a DDS-type process leading to the pro-
alters its reactivity. However another possibility of obtain- duction of cyclohexylbenzene instead of dicyclohexyl-type
ing a C–S bond cleavage ending with aromatic rings is the compounds. This means that the second C–S bond cleavage
insertion of a metal atom in the C–S bond (16, 17). We will is not affected much by the presence of a methyl group in
not discuss this mechanism here, although it could also be the vicinity of the sulfur atom in the second aromatic ring.
sensitive to steric hindrance.
In the literature, many of the results reported on
For the HYD pathway which involves the hydrogena- alkyldibenzothiophene HDS have been obtained with pro-
tion of one aromatic ring, it is reasonable to consider a step moted hydrotreating catalysts, namely sulfided CoMo and
by step process beginning also with the hydrogenation of NiMo/alumina catalysts. Considering the important pro-
the substrate into a dihydrogenated intermediate. There- moter effect often measured for hydrotreating reactions, it
fore, as in earlier proposals (8, 10) we assume that the two appears of importance to know how the observations sum-
pathways have dihydrodibenzothiophene compounds as in- marized above can be influenced by the presence of the
termediates (Scheme 3), despite the fact that this type of promoter.
compound has never been observed under the usual HDS
In this paper we examine the effect of Co and Ni pro-
conditions. We also note that, whatever the reactant, the moters on the activity of a Mo/alumina catalyst in the HDS
of both DBT and 46DMDBT under typical hydrotreat-
ing conditions. We also report complementary experiments
on the competitive reaction in the gas phase of DBT and
4
6DMDBT as well as on the effect of hydrogen and hydro-
gen sulfide on the transformation of these compounds. Our
aim was to propose new explanations concerning the inhi-
bition of the reactivity of alkyldibenzothiophenes and to
discuss the nature of the catalytic centers involved in HDS
SCHEME 2. Hydrodesulfurization of DBT. The first steps of the
DDS pathway: C–S bond cleavage through an elimination reaction (E2)
+
�
leading to the rearomatization of the ring. X , vacancy or proton; B , basic
site.

HDS OF ALKYLDIBENZOTHIOPHENES
411
on molybdenum sulfide-based catalysts and more specifi- Catalysts
ally to address two questions: (i) Do the DDS and HYD
pathways occur on the same or on distinct catalytic centers?
The commercial NiMo/alumina hydrotreating catalyst
used for the competition and the pressure effect exper-
iments contained 3 wt% NiO and 14 wt% MoO3. The
Mo/alumina catalyst, containing 9.3 wt% Mo, was pre-
pared (18) by incipient wetness impregnation of � -alumina
(
ii) What are the possible structures of these centers?
2
� 1
3
� 1
(
240 m g , pore volume 0.56 cm g ) with an aqueous so-
EXPERIMENTAL
lution of ammonium heptamolybdate (Fluka). The catalyst
�
�
Standard Reaction Conditions
was dried at 120 C and calcined under air flow at 500 C. The
NiMo/alumina and CoMo/alumina catalysts used for stud-
ies of the promoting effect were obtained (18) by impreg-
nating the Mo/alumina catalyst with aqueous solutions of
nickel or cobalt nitrates (Aldrich). They contained 2.3 wt%
Ni and 2.5 wt% Co, respectively, which corresponded to
a molar ratio Ni (Co)/Ni (Co) + Mo equal to 0.3. All the
The HDS of DBT and 46DMDBT was carried out in a
�
flow reactor at 340 C under a 4 MPa total pressure (9). De-
calin was used as a solvent to which dimethyldisulfide was
added to generate H2S. Under standard reaction conditions,
the various partial pressures were DBT or 46DMDBT =
0
0
.01 MPa, decalin = 0.89 MPa, H2 = 3.0 MPa, H2S =
�
catalysts were dried at 120 C and calcined under flowing
air at 500 C. They were then sulfided in situ by a mixture
of 5 vol% dimethyldisulfide in n-heptane, under a 3.0 MPa
.05 MPa (plus 0.05 MPa CH4 resulting from the dimethyl-
�
disulfide decomposition).
hydrogen partial pressure and a 4.0 MPa total pressure. The
H S partial pressure was 0.125 MPa and that of n-heptane
2
DBT/46DMDBT Competition Experiments
was 0.75 MPa. The sulfiding feed was injected at a starting
Both reactants were mixed together in decalin and in-
jected into the reactor. To examine the effect of DBT on
the transformation of 46DMDBT, the partial pressure of
�
�
�
� 1
temperature of 150 C and raised to 350 C at a 5 C min
�
rate. After 14 h, the temperature was lowered to 340 C, and
the reaction mixture was substituted for the sulfiding feed.
4
6DMDBT was maintained at the standard 0.01 MPa value,
while that of DBT was 0.01, 0.02, or 0.04 MPa. All the other
partial pressures and the total pressure were kept constant
by changing the partial pressure of the solvent. To exam-
Products
Dibenzothiophene, decalin, and dimethyldisulfide were
ine the effect of 46DMDBT on the conversion of DBT, the purchased from Aldrich, whereas 46DMDBT was synthe-
partial pressure of DBT was kept constant at 0.01 MPa, sized as previously described (19). Owing to the high boil-
while that of 46DMDBT was modified (0.005 or 0.01 MPa). ing point of the reactants, on-line analysis of the reaction
Higher 46DMDBT partial pressures could not be obtained products was not convenient. Consequently, the reactor
because of the low solubility of this compound in decalin.
effluents were condensed, and liquid samples were peri-
odically collected to be analyzed by gas chromatography.
Gaseous products were not formed, except for methane
which was produced by dimethyldisulfide decomposition.
The analyses were carried out on a Varian 3400 chromato-
graph equipped with a 50-m DB17 capillary column (J&W
Scientific) with a temperature programming from 100 to
Effect of Hydrogen Partial Pressure
Hydrogen partial pressures of 2, 3, and 4 MPa were used,
which modified the total pressure (3, 4, and 5 MPa, respec-
tively), but not the partial pressures of the reactant, of the
solvent, and of H2S when compared to the standard par-
tial pressures. After each hydrogen pressure modification,
the reaction was carried out again under standard reaction
conditions to make sure that the catalyst had remained un-
altered.
� �
�
1
2
30 C (10 C min ). Unknown products were identified by
GC-MS (Finnigan INCOS 500).
RESULTS
Simultaneous Transformation of DBT and of 46DMDBT
on a Commercial NiMo/Alumina Catalyst
Effect of Hydrogen Sulfide Pressure
In this case, the total and the partial pressures of the re-
Table 1 shows the effect of 46DMDBT partial pressure
actant were kept constant. The changes in the H2S partial on the activity of the catalyst for DBT transformation (total
pressure, obtained by modifying the amount of dimethyld- activity and activity for each pathway: DDS or HYD). The
isulfide in the feed, were compensated by changes in the sol- results obtained indicate that 46DMDBT inhibits the trans-
vent partial pressure. The H2S partial pressures used were formation of DBT. The apparent kinetic order with respect
0
, 0.025, 0.05, and 0.1 MPa. As explained above, we also to 46DMDBT can be roughly estimated to be � 0.4. On the
checked after each H2S pressure change that the catalyst other hand, the ADDS/AHYD ratio for DBT is reduced by a
had not been modified. factor of about 2 when the 46DMDBT pressure is increased

4
12
BATAILLE ET AL.
TABLE 1
TABLE 3
Effect of 46DMDBT on the Transformation
of DBT and Vice Versa
Effect of Hydrogen Sulfide Pressure on the Transformation
of DBT and 46DMDBT
Reactant:
DBT
46DMDBT
Reactant:
DBT
46DMDBT
Competitor:
46DMDBT
DBT
P H2S (MPa)
AT
ADDS
AHYD
ADDS/AHYD
0
0.025 0.05 0.1
0
0.025 0.05 0.1
Partial pressure of
competitor (MPa)
AT
ADDS
AHYD
0
0.005 0.01
0
0.01 0.02 0.04
74.7 17.7
70.9 15.8
3.8
9.7
7.7
2.0
3.8
5.4 4.0
4.2 1.9
1.2 2.1
1.9
0.5
1.4
1.7
0.3
1.4
1.3
0.2
1.1
9.7
7.7
2.0
6.9
4.8
2.1
5.7
3.8
1.9
1.7
1.1
1.0
0.8
1.9
8.8
0.30 0.20 0.20 0.15
1.40 0.90 0.80 0.65
2.00 0.21 0.22 0.25 0.23
18.6
3.5 0.90 0.36
0.21 0.18
�
ADDS/AHYD
3.85 2.28
Note. Sulfided commercial NiMo/alumina catalyst at 340 C, 3.0 MPa
H2. Reactant pressure = 0.01 MPa. (See Table 1 for abbreviations and
units.) Conversions of DBT and of 46DMDBT � 20 and 30 mol% , respec-
tively.
�
Note. Sulfided commercial NiMo/alumina catalyst at 340 C; 3.0 MPa
H2. Reactant pressure = 0.01 MPa; H2S pressure = 0.05 MPa. AT = total
activity; ADDS = activity for the direct desulfurization pathway; AHYD =
�
1
� 1
activity for the hydrogenation pathway (mol h kg ). Conversions of
DBT and of 46DMDBT as reactants � 40 and 15 mol% , respectively.
there is quite a noticeable change in the case of 46DMDBT
(
Table 2).
from 0 to 0.01 MPa. This means that the DDS pathway for
DBT is apparently more sensitive to 46DMDBT inhibition Effect of Hydrogen Sulfide Pressure on the Transformation
than the HYD pathway. However, the number of experi-
ments is not sufficient here to come to a definite conclusion.
Table 1 shows also that DBT inhibits the transforma-
tion of 46DMDBT in the same manner as 46DMDBT in-
hibits the transformation of DBT. The apparent kinetic or-
der with respect to DBT is also � 0.4. The two pathways
of DBT and of 46DMDBT on a Commercial
NiMo/Alumina Catalyst
The influence of hydrogen sulfide partial pressure was
also studied by adding dimethyldisulfide to the feed. It must
be pointed out that the H2S pressures indicated in Table 3
correspond only to the H2S produced by the dimethyldisul-
fide decomposition; the H2S produced by the HDS of the
reactant was not taken into account. The results in Table 3
indicate that the addition of H2S brings about a significant
decrease in activity, especially in the case of DBT. In the
presence of 0.1 MPa H2S, the activity for DBT transfor-
mation is reduced by a factor of 14, while the activity for
(
DDS and HYD) are in this case equally affected since the
ADDS/AHYD ratio is not modified.
Effect of Hydrogen Pressure on the Transformation
of DBT and 46DMDBT on a Commercial
NiMo/Alumina Catalyst
Table 2 shows that an increase in hydrogen pressure en- 46DMDBT transformation is divided only by a factor of 3.
hances the conversion of DBT and of 46DMDBT. Both the
As shown in Table 4, the kinetic order with respect to H2S
reaction rates and the kinetic orders reported in Table 4 is largely more negative for DBT than for 46DMDBT. On
indicate that for DBT the DDS and the HYD pathways in- the other hand, for the two molecules, the ADDS/AHYD ratio
crease in a similar manner, while for 46DMDBT, the DDS is lower in the presence of H S, which means that the DDS
2
pathway is much less sensitive to hydrogen pressure than pathway is much more sensitive to H S than the HYD path-
2
the HYD pathway. Consequently, while there is almost way. The kinetic order with respect to H S is more negative
2
no change in ADDS/AHYD selectivity in the case of DBT, for the DDS pathway than for the HYD reaction. In other
words, H2S must be a better competitor for the intermediate
TABLE 2
TABLE 4
Effect of Hydrogen Pressure on the Transformation
of DBT and 46DMDBT
Apparent Kinetic Orders for the Transformation
of DBT and 46DMDBT
Reactant:
DBT
46DMDBT
Reactant:
Effect of:
DBT
46DMDBT
P H2 (MPa)
AT
ADDS
AHYD
ADDS/AHYD
2
3
4
2
3
4
6.7
5.2
1.3
4.0
9.7
7.7
2.0
3.8
13.0
9.6
2.7
3.5
1.2
0.3
0.9
0.33
1.7
0.3
1.4
0.21
3.1
0.4
2.6
0.15
H2
H2S
H2
H2S
Global
DDS
HYD
1.0
0.9
1.0
� 0.7
� 0.8
� 0.4
1.3
0.3
1.4
� 0.3
� 0.3
� 0.2
�
Note. Sulfided commercial NiMo/alumina catalyst at 340 C. Reactant
pressure = 0.01 MPa; H2S pressure = 0.05 MPa. (See Table 1 for abbrevi-
ations and units.) Conversions = 10–30 mol% .
�
Note. Sulfided commercial NiMo/alumina catalyst at 340 C, reactant
pressure = 0.01 MPa. (See Table 1 for abbreviations.)

HDS OF ALKYLDIBENZOTHIOPHENES
413
involved in the DDS pathway of HDS in the case of DBT
TABLE 6
than in the case of 46DMDBT. This explains why the
HDS of DBT is more sensitive to H2S than the HDS of
Reactivity Ratio between DBT and 46DMDBT on the Mo, CoMo,
�
and NiMo/Alumina Catalysts at 340 C, 3.0 MPa H2
4
6DMDBT, the contribution of the DDS pathway to the
overall HDS reaction being much more significant for DBT DBT/46DMDBT
Total
DDS
HYD
than for 46DMDBT.
Mo
CoMo
NiMo
0.61
6.0
4.7
2.0
31.5
22.6
0.50
0.9
0.8
Effect of Ni or Co Promoters on the Decomposition
of DBT and of 46DMDBT
Note. Reactant pressure = 0.01 MPa; H2S pressure = 0.05 MPa. Con-
The reactivity and product distribution of DBT and versions � 15 mol% .
of 46DMDBT were measured on a Mo/alumina catalyst
and compared to those obtained with CoMo/alumina and
NiMo/alumina catalysts.
As reported in the literature (5–9) the HDS reactivity of
Surprisingly, on the Mo/alumina catalyst, 46DMDBT is 46DMDBT on promoted catalysts is less than that of DBT
slightly more reactive than DBT. This is mainly due to a by a factor of 5 to 6 (Table 6). This is because the promoting
higher rate of the HYD pathway which is the prominent effect of Co and Ni for the HDS of 46DMDBT (1.8 and 2.6,
pathway for both compounds on this catalyst: 75% for DBT, respectively) is 10 times smaller than for the HDS of DBT.
9
2% for 46DMDBT (Table 5). However, if we consider the
HDS activity (products with no sulfur), DBT is about twice on all the catalysts was the HYD pathway (80% on the
as reactive as 46DMDBT.
promoted catalysts, 90% on Mo/alumina, Table 5). It can
The main pathway for the transformation of 46DMDBT
As generally reported, the CoMo/alumina and NiMo/ be noted that this selectivity is about the same as the one
alumina catalysts were much more active than the Mo/ obtained with DBT on Mo/alumina. The promoting effect
alumina catalyst for the HDS of DBT (Table 5). The pro- on the HYD pathway is small, about one-half that for DBT.
moting effect was around 20, which is in accordance with The effect is greater on the DDS pathway than on the HYD
previous studies (20, and references therein). However the pathway yet smaller than for DBT.
promoting effect was essentially due to the enhancement of
Table 6 reports the differences of reactivity between
the rate of the DDS pathway of DBT. In this case, the pro- DBT and 46DMDBT for the various catalysts. As we have
moting effect for the DDS pathway was over 60, whereas seen, on Mo/alumina DBT is less reactive than 46DMDBT,
it was equal to 3–4 for the HYD pathway. For DBT, the mainly because of the HYD pathway. As already reported
HYD pathway represented about 75% on the Mo/alumina (7–9) on the promoted catalysts both reactants have ap-
catalyst but only 13–15% with the promoted catalysts.
proximately the same reactivity regarding the HYD path-
way (reactivity ratio between DBT and 46DMDBT equal
to 0.8–0.9) but DBT is considerably more reactive than
TABLE 5
4
6DMDBT regarding the DDS pathway (reactivity ratio
Activities of Mo, CoMo, and NiMo/Alumina Catalysts for the
Transformation of DBT and of 46DMDBT at 340 C, 3.0 MPa H2
equal to 20–30).
�
Table 7 compares the product distributions obtained with
the Mo/alumina and with the CoMo and NiMo/alumina
catalysts, either at similar total conversions or at similar
conversions through the HYD pathway. In the case of DBT,
we can see that the promoter changes not only the contri-
butions of the DDS and of the HYD pathways with re-
spect to each other but also the product distribution of the
HYD pathway. The two promoted catalysts give approxi-
mately the same results. In particular, the observed amount
of tetrahydrodibenzothiophene (THDBT) is much greater
with the unpromoted catalyst than with the promoted cata-
lysts. At a total conversion of about 9% , we obtain 4.7% of
tetrahydrodibenzothiophene with the unpromoted catalyst
and about only 1% with the promoted catalysts. We can also
consider the results obtained at similar HYD conversions
Reactant:
Catalyst:
DBT
46DMDBT
CoMo
Mo
CoMo
NiMo
Mo
NiMo
Activities
AT
ADDS
AHYD
ADDS/AHYD
AHDS
Selectivities
SDDS
SHYD
Promoting effect
Total
On HYD
On DDS
0.4
0.1
0.3
0.3
0.15
7.2
6.3
0.9
7.0
6.5
8.0
6.8
1.2
5.7
7.1
0.65
0.05
0.60
0.1
1.2
0.2
1.0
0.2
0.6
1.7
0.3
1.4
0.2
0.7
0.08
25
75
87
13
85
15
8
92
17
83
18
82
18
3
63
20
4
68
1.8
1.7
4.0
2.6
2.3
6.0
(
compare line 1 to lines 3 and 5, Table 7) which are more
Note. Reactant pressure = 0.01 MPa; H2S pressure = 0.05 MPa. SDDS
and SHYD = selectivities (% ) for the DDS and HYD pathways, respec-
tively. (See Table 1 for other abbreviations and units.) Conversions
relevant. We note in particular that the molar ratio tetrahy-
drodibenzothiophene/cyclohexylbenzene (THDBT/CHB)
is much higher with the unpromoted catalyst than with
�
15 mol% .

4
14
BATAILLE ET AL.
TABLE 7
Product Distributions for the Transformation of DBT and 46DMDBT over Mo, CoMo, and NiMo/Alumina Catalysts
Product distribution (mol% )
Conversion (mol% )
HYD
DDS
BiP
Reactant
DBT
Catalyst
Total
DDS
HYD
THDBT
4.7
HHDBT
CHB
1.1
DCH
0.3
THDBT/CHB
4.3
Mo
8.7
2.0
6.7
2.0
0.6
NiMo
8.9
3.0
7.4
25.7
1.5
7.3
7.4
25.7
1.0
2.3
—
—
0.5
5.0
—
—
2.0
0.5
3
CoMo
8.8
2.2
7.6
45.2
1.2
7.0
7.6
45.2
0.8
1.9
—
—
0.4
5.1
—
—
2.0
0.4
5
Reactant
Catalyst
Mo
Total
DDS
HYD
DMBiP
THDMDBT
HHDMDBT
MCHT
DMDCH
THDMDBT/MCHT
4
6DMDBT
9.9
0.8
2.8
9.1
17.2
0.8
2.8
8.6
13.4
—
—
0.5
2.0
—
1.8
17.2
6.7
2
2
0.0
NiMo
CoMo
9.0
0.0
1.5
4.5
7.5
15.5
1.5
4.5
5.5
4.5
—
—
2.0
10.0
—
1.0
2.8
0.4
11.9
3.0
2.5
4.7
9.4
18.3
2.5
4.7
5.5
2.8
—
—
3.9
15.5
—
—
1.4
0.2
2
�
Note. Reaction conditions: 340 C, 3.0 MPa H2. Reactant pressure = 0.01 MPa; H2S pressure = 0.05 MPa. DDS, direct desulfurization pathway;
HYD, hydrogenation pathway; BiP, biphenyl; THDBT, tetrahydrodibenzothiophene; HHDBT, hexahydrodibenzothiophene; CHB, cyclohexylben-
zene; DCH, dicyclohexyl; DMBiP, dimethylbiphenyl; THDMDBT, tetrahydrodimethyldibenzothiophene; HHDMDBT, hexahydrodimethyldiben-
zothiophene; MCHT, methylcyclohexyltoluene; DMDCH, dimethyldicyclohexyl.
the promoted catalysts. We can also see that traces of hex- Then we attempt to deduce from these considerations a
ahydrodibenzothiophene (HHDBT) were detected in the reasonable description of the active edge sites. However,
transformation of DBT on the Mo/alumina catalyst.
before that it appeared useful to describe in detail the
Similar results were obtained with 46DMDBT except sequence of steps that are expected to occur during the
that in this case the DDS pathway was much inhibited so transformation of alkyldibenzothiophenes.
that there was not so much difference between the total
conversion and the conversion through the HYD pathway.
Nevertheless, we can see that at similar conversions through
the HYD pathway, the molar ratio tetrahydrodimethyldi-
Steps Involved in the HDS of Alkydibenzothiophenes
Scheme 3, a typical reaction scheme for the HDS of
DBT, shows that a common dihydrointermediate could ex-
ist for the DDS and HYD pathways. However, Scheme 3
is still oversimplified. Actually, nine dihydroisomers can be
formed through partial hydrogenation of either DBT or
benzothiophene/methylcyclohexyltoluene (THDMDBT/
MCHT) is much smaller on the promoted catalysts than
on Mo/alumina.
DISCUSSION
4
6DMDBT as shown in Scheme 4A. One can note that
As already reported in the literature (5–9), it is found in this step is expected to be the most difficult hydrogena-
this work that the two S compounds, DBT and 46DMDBT, tion step since it destroys the aromaticity of one benzenic
have different reactivities on promoted catalysts. However ring. The dihydroisomers 1–6 can be formed by 1,2 addi-
it is also shown that the effect of the promoter (Co or Ni) tion of two hydrogen atoms to either of the two mesomeric
on the activity of the Mo catalyst is notably different for structures M1 or M2 shown in Scheme 4A. Compounds 7–9
the DDS and HYD pathways of the two compounds. Un- result either from 1,4 addition of two hydrogen atoms to
expectedly, it is found that the activity of the Mo catalyst for DBT or 46DMDBT or from double-bond isomerization in
the HDS of 46DMDBT is almost equal to its activity for the compounds 1–6. It can be noted, however, that the double
HDS of DBT. This questions the explanation often found bonds in the dihydrogenated rings of compounds 7–9 are
in the literature (5, 6) of the low reactivity of 46DMDBT not conjugated. Consequently these compounds are proba-
based on steric hindrance of the 46DMDBT adsorption due bly less stable than compounds 1–6, especially compounds
to the presence of the methyl groups at the 4 and 6 positions. 8 and 9 in which the thiophenic ring is not preserved. In
In the following discussion we examine other possible Scheme 4A two types of reaction can be considered for
explanations for the differences in reactivity of DBT and the dihydroisomers, either C–S bond cleavage (a) or fur-
4
6DMDBT over the Mo, CoMo, and NiMo catalysts in or- ther hydrogenation into tetrahydrogenated products (b).
der to better understand the origin of the promoter effect. Scheme 4B decribes these two routes in detail.

HDS OF ALKYLDIBENZOTHIOPHENES
415
SCHEME 4. (A) Various isomers of the dihydrointermediates in the hydrodesulfurization of DBT (R = H) and of 46DMDBT (R = CH3). Reaction
rates: e, easy; m, medium; a, end of the sequence for the DDS pathway; b, end of the sequence for the HYD pathway. (B) Details for sequences a and
b of Scheme 4A given for compound 4 as an example.
We note immediately in Scheme 4A that only two of the hydrogenation to tetrahydroisomers (HYD pathway, route
dihydrointermediates can undergo directly an elimination “b” in Schemes 4A and 4B).
step leading to C–S bond cleavage (DDS pathway, route
a” in Scheme 4A and 4B) and regenerate an aromatic ring. fect on most of the steps leading to the HDS of DBT and
All of the other dihydroisomers will be converted by further 46DMDBT, namely on: (i) the adsorption of the reactants
The presence of the methyl groups could have an ef-
“

4
16
BATAILLE ET AL.
Regarding the HYD pathway, C–S bond cleavage is also
the rate-limiting step for both reactants on the unpromoted
catalyst (the hydrogenated intermediates were obtained
in significant quantities) but the formation of the dihy-
drointermediates is most likely the rate-limiting step for
the transformation of DBT on the promoted catalysts. For
4
6DMDBT it is not so clear: significant amounts of hydro-
genated intermediates were formed, which means that the
C–S bond cleavage is apparently not very rapid compared
to the hydrogenation steps. In the following sections we will
examine the possible effect of the methyl groups on the var-
ious steps of both pathways and try to understand why they
inhibit mainly the DDS pathway.
SCHEME 5. Mechanism of the C–S bond cleavage in the DDS path-
way on MoS2.
Steric and Electronic Effects on the Reactivity
of DBT-Type Compounds
or intermediates (especially the dihydrointermediates);
(
a) Effect on the adsorption of 46DMDBT. It was
(
ii) the addition of hydrogen atoms (to form the tetrahydro-
found that the methyl groups in 46DMDBT affected only
the DDS pathway. Therefore, if we assume that the DDS
and HYD pathways have a common intermediate, we can
conclude that it is neither the adsorption of the reactant
which is affected by the presence of alkyl groups in posi-
tions 4 and 6 nor the step leading to the common dihy-
drointermediate, but rather one of the subsequent steps of
the DDS pathway of 46DMDBT, most probably the first
C–S bond cleavage (7). Indeed, the HDS reaction of DBT
and 46DMDBT carried out in competition in a batch re-
actor (8) showed that the adsorption coefficients of the
two compounds on NiMo/alumina were similar. A value
of 1.3 for the ratio of the constants of the adsorption equi-
librium (KDBT/K46DMDBT) was obtained by comparing the
ratio of reactivity between the two molecules in competi-
tion (kDBT/k46DMDBT) with the ratio of the rate constants
and hexahydrointermediates) and (iii) the C–S bond cleav-
age steps through elimination (E2 mechanism shown in
Scheme 5). In particular, the presence of the methyl groups
could explain why the rate-limiting steps for DBT and for
4
6DMDBT are not the same.
The results of Tables 5–7 indicate that the rate-limiting
step of the two pathways (Table 8) depends both on the re-
actant and on the catalyst. This will be discussed in greater
detail later (see in particular the section dealing with the
effect of the promoter). It is clear, however, that on the
Mo/alumina catalyst C–S bond cleavage is the rate-limiting
step of the DDS pathways for both reactants. On the pro-
moted catalysts, C–S bond cleavage is still the rate-limiting
step for the DDS pathway of 46DMDBT while the forma-
tion of the dihydrointermediate is the rate-limiting step for
the DDS pathway of DBT.
(
kDBT/k46DMDBT) calculated for the transformation of each
molecule. In the experiments reported here, it was found
that the two reactants had the same reciprocal inhibit-
ing effect. In the present work the HDS of DBT and of
TABLE 8
Rate-Limiting Steps (X) for the HDS of DBT and 46DMDBT
on Mo, CoMo, and NiMo/Alumina Catalysts
4
6DMDBT (alone or in competition) was carried out in a
continuous flow reactor on NiMo/alumina (Table 1). Us-
ing the same method as Meille et al. (8), the activity values
for the two reactions led to a KDBT/K46DMDBT ratio of 1.1,
which confirms that the two molecules have similar adsorp-
tion coefficients.
We must consider then the hypothesis that with
46DMDBT the two pathways could have different rate-
limiting steps which would not be equally sensitive to steric
hindrance. Actually, in the case of DBT the kinetic orders
with respect to hydrogen of these pathways are close to 1
(Table 4). This result is consistent with the assumption (as
indicated in Table 8 for promoted catalysts) that the rate-
limiting step of DBT transformation is the step which is
common to the two pathways, i.e., the formation of dihy-
drodibenzothiophene as proposed recently by Olguin and
Reactant:
DBT
46DMDBT
CoMo or
CoMo or
NiMo
Catalyst:
DDS
Mo
/
Mo
NiMo
Formation of the
X
/
/
X
/
/
dihydroisomers 4 and 5
C–S cleavage in the
dihydroisomers 4 and 5
Formation of the
X
/
X
HYD
X
/
X ?
dihydroisomers (1–7)
Hydrogenation of the
dihydroisomers (1–7)
C–S cleavages in
/
/
/
/
X
/
X
tetrahydro or
hexahydroisomers

HDS OF ALKYLDIBENZOTHIOPHENES
417
Vrinat (21). In the case of 46DMDBT, the kinetic order with of the methyl groups would not interfere in the same man-
respect to hydrogen for the HYD pathway is also quite high ner, either because the sites are different or because of the
(
greater than 1), but the kinetic order for the DDS pathway presence or absence of hydrogen atoms on the sites (see
is only 0.3. This can be explained by supposing that the two below).
pathways do not have the same rate-limiting step; for exam-
(
c) Steric hindrance of the C–S bond cleavage and elec-
ple, the rate-limiting step of the HYD pathway is still the
formation of the dihydrointermediate (Table 8), whereas
for the DDS pathway the C–S bond cleavage is now the
rate-limiting step.
If we look at the mechanism of the C–S bond cleavage
shown in Scheme 5, it can be seen that there are several
other reasonswhyalkylgroupsin positions4and 6mayalter
the reactivity of DBT-type compounds and in particular
regarding the DDS pathway.
tronic effect on the lability of the proton involved in the reac-
tion. For C–S bond cleavage to take place according to the
DDS pathway (Scheme 5), compounds 4 and/or 5 must ei-
ther undergo the elimination reaction without desorption
or readsorb on their side bearing one hydrogen atom in
�
-position with respect to the sulfur atom. Moreover, it is
reasonable to consider that in the adsorbed state the sulfur
atom should be in interaction with the catalyst (which is not
necessarily so in the adsorbed state for hydrogenation) in
(
b) Steric effect of the methyl groups on the adsorption of order to favor the antielimination process. However, if the
the dihydrointermediates involved in the DDS pathway. If catalytic hydrogenation of a double bond consists (as gen-
the alkyl groups do not inhibit the adsorption of the sub- erally accepted) in a cis addition of the two hydrogen atoms,
strate, we must envisage that the alkyl groups can have an then the sulfur atom and the �-hydrogen atom in the dihy-
effect on the adsorption of the dihydrointermediates. This drointermediate (as shown for compound 4 in Scheme 6b)
must be considered especially if there is desorption of these should be in a trans configuration. This is favorable for an
intermediates (as shown in Scheme 4), which we will as- antielimination despite the fact that it must be very diffi-
sume despite the fact that they were not detected among cult to have, at the same time, both the sulfur atom and the
the products (probably because they undergo subsequent trans-hydrogen atom in �-position in interaction with the
hydrogenation readily).
surface (if the latter is assumed to be flat). In any case, the
As already pointed out, only two of the dihydrointerme- presence of the methyl group on carbon 4 of compound 4
diates (compounds 4 and 5 in Scheme 4A) can undergo C–S (Scheme 6b) can obviously hinder the elimination process
bond cleavage. The adsorption of one of these compounds by preventing either the sulfur atom or the �-hydrogen
(
compound 4) may be greatly affected by the presence of atom from approaching the catalytic centers. The methyl
the methylgroups. Actuallyonce the substrate hasbeen par- group in 46DMDBT can also make the �-hydrogen atom
tially hydrogenated into compound 4 the two sides of the involved in the elimination process (Scheme 6b) less acidic
molecule are no longer identical. Hence, while the analo- than in DBT (22). Moreover, only one hydrogen atom is
gous dihydrodibenzothiophene can adsorb equally on both available for the elimination reaction instead of two in the
sides, the adsorption of compound 4 on the side with the corresponding dihydrointermediate for DBT (if we do not
alkyl group out of the plane of the molecule can obviously exclude the possibility that a cis-�-hydrogen atom could
be hindered (Scheme 6b). We must assume then that de- be involved in the elimination process). All these effects
pending on whether the adsorption of these intermediates could lower the reactivity of 46DMDBT compared to that
leads to the DDS or HYD pathway, the steric hindrance of DBT regarding especially the DDS pathway.
SCHEME 6. Approaches of the dihydrointermediates of DBT (a) and of 46DMDBT (b) to the catalytic center (compound 4 of Scheme 4A chosen
as an example).

4
18
BATAILLE ET AL.
(
d) Effect of the methyl groups on the distribution of the on steric hindrance (sections b and c above). Indeed, the re-
dihydroisomers. The presence of alkyl groups in positions activity of 46DMDBT can be improved by releasing steric
and 6 can modify the kinetics of hydrogenation of the hindrance through acid-catalyzed methyl migration, while
4
substrate and the stability of the dihydroproducts. Con- no such effect can be expected with DBT (9).
sequently the distribution of the dihydroisomers obtained
(
e) Effect of the methyl groups on the hydrogenation
with DBT can be different from that of the analogous di-
hydroisomers obtained with 46DMDBT.
We can consider that the ratesofhydrogenation ofdouble
bonds follow the sequence,
steps. As we have seen, the methyl groups in the 4 and
positions hinder the DDS pathway significantly but have
practically no effect on the reactivity through the HYD
pathway of DBT-type compounds.
6
As suggested by Schemes 4A and 4B, all the dihydroin-
termediates can undergo a subsequent hydrogenation step
and lead to the HYD pathway. The dihydrointermediates
which are expected to be the most reactive regarding subse-
quent hydrogenation are those with a disubstituted double
bond (compounds 3–7). We note that compounds 4 and
(
(
the less substituted, the more reactive the double bond
22)) and that the stabilities of the dihydro compounds de-
5
which can lead to DDS can also undergo HYD quite
pend on the degree of substitution of their double bonds in
the order
readily. We also recall that tetrasubstituted double bonds
are difficult to hydrogenate; they are also the most stable.
This may explain why tetrahydrodibenzothiophene, prob-
ably the 1,2,3,4-tetrahydroisomer,
(
the less substituted, the less stable the double bond). With
this in mind and considering the fact that the thiophenic
ring is preserved or not in the dihydrointermediates, we
come to the conclusion that compounds 4 and 5 will most
likely be disfavored both kinetically and thermodynami- was found among the products. If we consider compounds
cally (Table 9), with one of them (compound 4) even more 3, 4, and 7 formed from 46DMDBT, we could expect that
disfavored in the case of 46DMDBT than in the case of the methyl group which is out of the plane of the molecule
DBT. Moreover these intermediates should be among the could hinder its adsorption and subsequent hydrogenation.
more reactive to undergo HYD since they possess a dis- This does not seem to be the case. The reason for this could
ubstituted double-bond (Scheme 4A). These observations be that in the case of a hydrogenation step the sulfur atom
could also explain why DBT is more reactive especially in of the adsorbed reactant does not necessarily interact with
the DDS pathway than 46DMDBT.
a catalytic center which, on the contrary, is probably the
However, if this were a critical point, the addition of any case for C–S bond cleavage (the rate-limiting step of the
cocatalyst which could help the migration of the double DDS pathway).
bonds in the dihydroisomers should in principle increase
Catalytic Centers
the reactivity of both DBT and 46DMDBT by producing
compounds 4 and 5 as soon as they disappear through the
DDS pathway. An increase in reactivity was indeed ob-
served in the case of 46DMDBT when an acid component
was added to a commercial hydrotreating catalyst, but not
for DBT (9), which makes this explanation of the effect of
the methyl groups less likely than the explanations based
The foregoing discussion has shown that the numerous
possible steps involved in the HDS of DBT-type molecules
belong to two categories: (i) the addition of H2 and (ii) C–S
bond cleavage by elimination. It is therefore of interest to
discuss the possible nature of the catalytic centers responsi-
ble for these two major steps. We will first consider a simpli-
fied representation of a MoS2 particle neglecting possible
surface reconstruction (23), a much debated matter, as well
as the presence of surface carbon species (24). In addition,
since it isnot well known yet which of the two exposed edges
TABLE 9
Kinetics of Dihydroisomers Formation from DBT (R = H) and
4
6DMDBT (R = CH3) and Expected Stabilities (See Scheme 4A for
¯ ¯
(
1010 or 1010) is the one containing the active centers, both
the Notation of the Dihydroisomers)
will be considered in the following discussion. We will also
consider that H2 dissociates heterolytically on sulfide cata-
lysts (25–28) and that hydrogen atoms (protons at least) are
mobile on the surface of sulfide catalysts at temperatures
R
Kinetics
Expected stabilities
H
CH3
1 � 2 � 3 > 4 = 6 > 5
1 � 2 > 3 � 6 > 4 � 5
1 � 3 � 7 > 2 > 4 � 6 > 5
1 > 2 > 3 > 6 > 7 > 4 � 5
�
as low as 80 C (28).

HDS OF ALKYLDIBENZOTHIOPHENES
419
The main features that must be taken into account are
that (i) the catalytic centers should be able to activate (ad-
sorb) the aromatic ring(s) of the substrate molecules in or-
der to make their hydrogenation possible, (ii) they must
obviously adsorb and dissociate dihydrogen, and (iii) the
catalyst should be able to retain S atoms resulting from the
decomposition of the organic molecule (29, 30).
Therefore the catalytic centers should display the quali-
ties (or functions) indicated below:
for H2 addition
—
at least one vacancy to absorb the substrate through
at least one of its aromatic rings;
one vacancy to adsorb a hydrogen atom with a hy-
dride character;
—
—
one neighboring S2� atom to adsorb a hydrogen
atom as a proton;
for C–S bond cleavage
—
one vacancy to adsorb the dihydro-, terahydro-, or
hexahydrointermediate;
—
one vacancy to “activate” (Lewis-type center) and
retain the sulfur atom;
—
one S2 atom acting as a basic site.
�
Hence it could be considered that before adsorption of
the reactants, the sites for the two steps should possess the
same features, namely at least two vacancies and one neigh-
boring sulfur atom. Such catalytic centers can be obtained
by removing the sulfur atoms from the edges and corners of
a MoS2 slab with all its edges lined by sulfur atoms. Indeed,
ab initio calculations indicate that the edges of a MoS2
slab are more stable when terminated by sulfur atoms than
by molybdenum atoms (31). We will consider the centers
which, resulting from the removal of a minimum number
of sulfur atoms, fit the requirements indicated above.
A catalytic center with at least two vacancies can be
¯
SCHEME 7. Catalytic centers with two sulfur vacancies on the 1010
¯
and 1010 edges. White or gray circles, sulfur atoms; black circles, molyb-
denum atoms; black circles with star, uncompletely coordinated molybde-
num atoms. , sulfur vacancy; ٗ, free coordination position.
obtained by removing two sulfur atoms from the edges basicity. Therefore, if the main requirement for the E step
2
¯
¯
2�
in the 1010 and 1010 planes of an hexagonal MoS2 slab (or C–S bond cleavage in general) is the basicity of the S
Scheme 7). For each edge plane, several configurations atoms involved in the elimination process (18, and refer-
(
¯
can be obtained. In the 1010 case, sites Va1, Va2 and Va3 ences therein), the greater sensitivity of the DDS pathway
are easy to obtain according to Byskov et al. (31). In the to H S may be explained by supposing that H S (due to
1
2
2
¯
010 case, it is possible to obtain Vb1 (or B site according to its acidic properties) adsorbs preferably on the most basic
Siegel’s proposal (32)), Vb2 and Vb3. Other sites may also sites.
be obtained at corners. Such sites (e.g., Vc, Scheme 7) with
The difference between C–S bond cleavage and hydro-
two vacancies are like site Vb1 except that on one side they genation centers could also arise once the catalyst is un-
have no S2 neighbors in the front row. If we consider just der running conditions. Under these conditions, we could
these types of sites, it is clear that although they possess consider that the E2 reaction (C–S bond cleavage) re-
the same number of vacancies, they have neither the same quires a “naked” site such as Va1 for instance (Scheme 8a),
�
number of S2 neighbors nor the same geometry and will while the hydrogenation reaction would require the same
not generate the same steric constraints to the approach- type of center occupied, however, by dissociated hydrogen
ing substrate. The adsorption of H2S on such centers will (Scheme 8b). A hydrogenating center must be capable of
in principle inhibit the HYD pathway, but should also in- delivering two H atoms. This would be definitely so if, under
hibit the DDS pathway. However, depending on whether these steady-state working conditions, dissociated hydro-
�
the removed S2 ions belong to the same or to different gen were available on the center. Interestingly, the adsorp-
�
Mo atoms, the remaining S2 atoms may not have the same tion of H on a “C–S bond cleavage site” is able to transform
�
2

4
20
BATAILLE ET AL.
it into a “hydrogenation site”, which is in accordance with
the idea of site interconversion (33).
Possible Role of the Promoter
According to the interpretation of Chianelli (34) the pro-
moter is supposed to decrease the strength of the bond be-
tween molybdenum and the sulfur atoms resulting from
the decomposition of the organic molecules. In the same
way it can be supposed that the promoter decreases the
metal–sulfur bond in the sulfide itself (20, and references
therein) and increases the electronic density on the sulfur
atoms (35), which can also be interpreted in terms of an
SCHEME 9. Possible promoted C–S bond cleavage centers.
enhancement of the basicity of particular S2 centers (18).
This obviously has a consequence on the reactivity of both
DBT and 46DMDBT as well as on their product distri-
bution. Both depend at the same time on the properties
of the catalysts (acid-base character in particular) and on
the structure of the reactants (steric effects of the methyl
groups for instance). A remarkable observation is that the
product distribution and in particular the ADDS/AHYD se-
lectivity obtained with DBT on the unpromoted catalyst
are very much the same as those obtained with 46DMDBT
on the promoted catalysts (Tables 5 and 7).
The main effect of the promoter is to enhance the rate
of the DDS pathway (or the C–S bond cleavage activity
in general). This is quite easy to understand if we assume
that the C–S bond cleavage occurs through an elimination
mechanism as described in Schemes 2 and 5. Such a mecha-
nism involves the attack of a hydrogen atom (in � position
relative to the sulfur atom in the organic molecule) by a sul-
fur anion acting as a basic site. Therefore if, as suggested,
the promoter enhances the basicity of the sulfur anions, it
should favor C–S bond cleavage. The effect is spectacular in
the case of DBT with which there are no significant steric
constraints. As explained in the foregoing discussion, on
the unpromoted catalyst the C–S bond cleavage reactions
involved in the HDS of DBT seem to be slow compared
to the hydrogenation reactions: the DDS pathway partic-
ipates to a minor degree to the overall process (Table 5),
which means that step 7 (Scheme 3) is slow. Regarding the
HYD pathway which is predominant, a large amount of
�
intermediate hydrogenated compounds is found (tetrahy-
drodibenzothiophene and even traces of hexahydrodiben-
zothiophene, Table 7), which means that steps 4 (Scheme 3)
is also relatively slow compared to step 2. This is even more
so with 46DMDBT, the difficulty to cleave the C–S bond
being amplified by steric constraints.
When the promoter is added to molybdenum the DDS
pathway becomes predominant for DBT and the amount
of intermediate hydrogenated compounds (tetrahydro-
dibenzothiophene or hexahydrodibenzothiophene) is
much smaller at similar conversion rates than with the un-
promoted catalyst. The tetrahydrodibenzothiophene/cyclo-
hexylbenzene molar ratio decreases significantly (Table 7).
Hence it can be concluded, at least as far as DBT is con-
cerned, that the C–S bond cleavage is no longer rate-
limiting. It is not so with 46DMDBT probably because
of steric constraints. The DDS pathway represents in this
case a minor contribution even with the promoted cata-
lysts. On the other hand, if the amount of hydrogenated
intermediates is also smaller on the promoted catalysts, it
can be seen that, at similar HYD conversion (about 7.5%
on NiMo/alumina for instance), the tetrahydrodimethyl-
dibenzothiophene/methylcyclohexyltoluene molar ratio is
much greater than the tetrahydrodibenzothiophene/cyclo-
hexylbenzene molar ratio. This explains why 46DMDBT
behaves with promoted catalysts more or less in the same
way as DBT on the unpromoted catalyst, and why we have
approximately the same product distribution in both cases.
The reason is that in both cases C–S bond cleavages are
slow compared to the hydrogenation steps whereas with
DBT on promoted catalysts the rate of the C–S bond cleav-
ages is increased to such an extent that the hydrogenation
steps become rate-limiting for the HDS.
It can therefore be concluded that a typical C–S bond
cleavage center on a promoted catalyst should contain at
least a promoter atom in the vicinity of a sulfide anion
(
Scheme 9).
CONCLUSION
The results obtained in the present work indicate
that promoted and unpromoted catalysts have different
SCHEME 8. Examples of catalytic centers. C–S bond cleavage cen-
ter, naked Va1 site (a); hydrogenation center, Va1 with adsorbed H2 (b).

HDS OF ALKYLDIBENZOTHIOPHENES
421
properties regarding the two main pathways (DDS and
The main effect of the promoter is to improve the C–S
HYD) of the HDS of DBT and 46DMDBT. This means bond cleavage activity of the MoS2 on alumina catalyst, pre-
that attributing the differences in reactivity of these com- sumably by increasing the basicity of certain sulfur anions
pounds to steric effects upon adsorption on the catalytic sur- shared between the Mo and the promoter (Co or Ni).
face is not correct. A detailed analysis of the activities and
product distributions for the transformation of DBT and
6DMDBT on sulfided Mo, CoMo, and NiMo supported
ACKNOWLEDGMENTS
4
This work was carried out within the framework of a contract entitled
Hydrod e´ sulfuration des gazoles, financially supported by CNRS-Ecodev,
Elf-Antar France, IFP, Total and Procatalyse. The authors are particularly
grateful to C. Traversa for the synthesis of 4,6-dimethyldibenzothiophene.
on alumina catalysts indicates that the origin of these dif-
ferences lies in the kinetics of transformation of these two
compounds. For both reactants the DDS and HYD path-
ways can be decomposed into H2 addition steps and C–S
bond cleavage by elimination (or E2) steps. Depending on
the reactant and on the catalyst the rate-determining step
may be different.
On the unpromoted catalyst, DBT and 46DMDBT have
similar reactivites because of the fact that C–S bond cleav-
ages are rate-limiting for both reactants. 46DMDBT is even
slightly more reactive than DBT although it leads mainly
to hydrogenation products.
On the promoted catalysts, DDS becomes the main path-
way for the HDS of DBT. The hydrogenation of DBT
into dihydrodibenzothiophene is the rate-limiting step for
the two pathways. With 46DMDBT, the effect of the pro-
moter on C–S bond cleavage is limited because of steric
constraints. Consequently, C–S bond cleavage remains the
rate-limiting step, for the DDS pathway in particular. The
presence of the methyl groups in 46DMDBT changes con-
siderably the reactivity regarding the two pathways. The
low reactivity of 46DMDBT is mainly due to the inhibition
REFERENCES
1
2
3
4
5
6
7
. Ma, X., Sakanishi, K., and Mochida, I., Ind. Eng. Chem. Res. 33, 218
(1994).
. Shih, S. S., Mizrahi, S., Green, L. A., and Sarli, M. S., Ind. Eng. Chem.
Res. 31, 1232 (1992).
. Kabe, T., Ishihara, A., and Tajima, H., Ind. Eng. Chem. Res. 31, 1577
(
1992).
. Whitehurst, D. D., Isoda, T., and Mochida, I., Adv. Catal. 42, 345
(1998).
. Kilanowski, D. R., Teeuwen, H., de Beer, V. H. J., Gates, B. C., Schuit,
G. C. A., and Kwart, H., J. Catal. 65, 129 (1978).
. Houalla, M., Broderick, D. H., Sapre, A. V., Nag, N. K., de Beer, V. H. J.,
Gates, B. C., and Kwart, H., J. Catal. 61, 523 (1980).
. Kabe, T., Ishihara, A., and Zhang, Q., Appl. Catal. A 97, L1 (1993).
8. Meille, V., Schulz, E., Lemaire, M., and Vrinat, M., J. Catal. 170, 29
1997).
. Michaud, P., Lemberton, J. L., and P e´ rot, G., Appl. Catal. A 169, 343
1998).
0. Singhal, G. H., Espino, R. L., Sobel, J. E., and Huff, G. A., J. Catal. 67,
57 (1981).
(
9
(
1
4
of the DDS pathway (which in this case means C–S bond 11. Houalla, M., Nag, N. K., Sapre, A. V., Broderick, D. H., and Gates,
B. C., AIChE J. 24, 1015 (1978).
2. Nag, N. K., Appl. Catal. 10, 53 (1984).
3. Rollmann, L. D., J. Catal. 46, 243 (1977).
4. Nagai, N., Sato, T., and Aiba, A., J. Catal. 97, 52 (1986).
5. Lamure-Meille, V., Schulz, E., Lemaire, M., and Vrinat, M., Appl.
Catal. A 131, 143 (1995).
6. Sanchez-Delgado, R. A., J. Mol. Catal. 86, 287 (1994).
7. Jones, W. D., Meyers, A. W., and Vicic, D. A., in “Proceedings of the
16th Meeting of the North American Catalysis Society,” p. 74. Boston,
MA, 1999.
cleavage). Several explanations can be proposed for this
lower reactivity:
1
1
1
1
—
a steric hindrance by the methyl groups to the adsorp-
tion of the dihydrointermediates;
—
the fact that only one H is available for the elimination
step in 46DMDBT;
a steric hindrance by the methyl groups during the E2
elimination step in the dihydrointermediates;
an effect of the methyl group on the acidity of the H
1
1
—
1
8. Mijoin, J., Th e´ venin, V., Garcia Aguirre, N., Yuze, H., Wang, J., Li,
W. Z., P e´ rot, G., and Lemberton, J. L., Appl. Catal. A 180, 95
—
atom in the position 4 (or 6).
(
1999).
1
2
9. Meille, V., Schulz, E., Lemaire, M., Faure, R., and Vrinat, M., Tetrahe-
dron 52, 3953 (1996).
0. Topsoe, H., Clausen, B. S., and Massoth, F. E., in “Catalysis, Science
and Technology” (J. R. Anderson and M. Boudard, Eds.), Vol. 11.
Springer-Verlag, Berlin, 1996.
Two types of catalytic centers can be considered to ac-
count for the results obtained, including the effect of hy-
drogen and of hydrogen sulfide:
—
the sites involved in the hydrogenation steps would be
2
2
1. Olguin Orozco, E., and Vrinat, M., Appl. Catal. A 170, 195 (1998).
2. March, J., in “Advanced Organic Chemistry—Reactions, Mechanism
and Structure,” 4th Ed., p. 998. Wiley, New York, 1992.
3. Eijsbouts, S., Appl. Catal. A 158, 53 (1997).
4. Berhault, G., and Chianelli, C., in “Proceedings of the 16th Meeting
of the North American Catalysis Society,” p. 20. Boston, MA, 1999.
5. Anderson, A. B., Al-Saigh, Z. Y., and Hall, W. K., J. Phys. Chem. 92,
composed of a vacancy associated with a SH group and with
a hydrogen atom adsorbed on a molybdenum atom;
2
2
—
the sites involved in the E2 direct desulfurization route
and more generally in C–S bond cleavage would be com-
posed of two vacancies associated with a sulfur anion.
2
2
2
8
03 (1988).
6. Kasztelan, S., and Guillaume, D., Ind. Eng. Chem. Res. 33, 203
1994).
The two centers could be considered as basically identical
(
in terms of Mo and S atoms involved) except that adsorbed
(
H2 should be considered as part of the hydrogenation
center.
7. Thomas, C., Vivier, L., Lemberton, J. L., Kasztelan, S., and P e´ rot, G.,
J. Catal. 167, 1 (1997).

4
22
BATAILLE ET AL.
2
8. Thomas, C., Vivier, L., Travert, A., Maug e´ , F., Kasztelan, S., and P e´ rot,
31. Byskov, L. S., Hammer, B., Norskof, J. K., Clausen, B. S., and Topsoe,
G., J. Catal. 179(2), 495 (1998).
H., Catal. Lett. 47, 177 (1997).
2
9. Isagulyants, G. V., Greish, A. A., and Kogan, V. M., in “Proceed- 32. Siegel, S., J. Catal. 30, 139 (1973).
ings, 9th International Congress on Catalysis” (M. J. Phillips and
33. Delmon, B., and Froment, G. F., Catal. Rev. Sci. Eng. 38, 69 (1996).
M. Ternan, Eds.), Vol. 1, p. 35. Chem. Inst. of Canada, Ottawa, 34. Chianelli, R. R., Catal. Rev. Sci. Eng. 26, 361 (1984).
1
988.
0. Kabe, T., Qian, W., Wang, W., and Ishihara, A., Catal. Today 29, 197
1996).
35. Topsoe, H., Clausen, B. S., Topsoe, N. Y., Pedersen, E., Niemann, W.,
M u¨ ller, A., B o¨ gge, H., and Lengeler, B., J. Chem. Soc. Faraday Trans.
I 83, 2157 (1987).
3
(