Hydrodesulfurization of alkyldibenzothiophenes: Evidence of highly unreactive aromatic sulfur compounds

Article abstract of DOI:10.1006/jcat.2000.2897

Crude oil usually contains ~ 1 wt % of sulfur and SO_x (a major air pollutant) is emitted, during combustion of fuels. Deep HDS is hindered mainly by the alkyldibenzothiophenes. 4,6-Dimethyldibenzothiophene (4,6-DMDBT) is always present in hydrotreated gas oils, even after hard HDS conditions. The synthesis of 4,6-DMDBT is used as a model to improve the reactivity of HDS catalysts. Various dialkylbenzothiophenes containing methyl, ethyl, diisobutyl, and diisopropyl substituents in the fourth and sixth positions were synthesized and their reactivities compared over an NiMo/Al₂O₃ industrial catalyst. In a batch reactor at 573 K and at a total pressure of 5 MPa, studies were conducted. Kinetic studies and competitive reactions showed that adsorption on the surface of the catalyst was not the rate-determining step for their transformation. HDS was proposed to proceed via a network of parallel reactions after partial hydrogenation of one aromatic ring of the sulfur compound. Variations in the reactivities of the different dibenzothiophene derivatives agreed well to the steric hindrance generated by the alkyl groups near the sulfur atom. 4,6-Diisopropyldibenzothiophene showed very low reactivity, and the steric effect led to the disappearance of desulfurization under the reaction conditions.

Full text of DOI:10.1006/jcat.2000.2897

Journal of Catalysis 193, 255–263 (2000)
doi:10.1006/jcat.2000.2897, available online at http://www.idealibrary.com on
Hydrodesulfurization of Alkyldibenzothiophenes: Evidence of Highly
Unreactive Aromatic Sulfur Compounds
,
,
,1
M. Macaud, † A. Milenkovic, † E. Schulz,† M. Lemaire,† and M. Vrinat
Institut de Recherches sur la Catalyse, 2 Av. A. Einstein, 69626 Villeurbanne Cedex, France; and †Laboratoire de Catalyse et Synthe`se Organique, IRC,
UCBL, CPE, 43 Bd.du 11 Novembre 1918, 69622 Villeurbanne Cedex, France
Received December 15, 1999; revised March 10, 2000; accepted April 10, 2000
lysts based on new supports or new active phases (3–5) and
to improve catalytic processes by using two- or three-stage
HDS processes (6, 7). However, until now none of them
proved to be economically feasible for removing “hard
sulfur” species from distillates. Recent findings on the HDS
of polyaromatic sulfur compounds has been reviewed by
Whitehurst et al. (8).
Various dialkyldibenzothiophenes containing methyl, ethyl, di-
isobutyl, and diisopropyl substituents in the fourth and sixth po-
sitions were synthesized and their reactivities compared over an
NiMo/Al2O3 industrial catalyst. Studies were made in a batch reac-
tor at 573 K and at a total pressure of 5 MPa. Kinetic studies and
competitive reactions demonstrated that adsorption on the surface
of the catalyst was not the rate-determining step for their trans-
formation. Hydrodesulfurization was proposed to proceed via a
network of parallel reactions after partial hydrogenation of one
aromatic ring of the sulfur compound. Variations in the reactivities
of the different dibenzothiophene derivatives are well correlated to
the steric hindrance generated by the alkyl groups near the sulfur
atom. Moreover, 4,6-diisopropyldibenzothiophene showed very low
reactivity, and the steric effect led to the disappearance of desulfu-
c
It is now well established that deep HDS is hindered
mainly by one class of sulfur compounds, the alkyldibenzo-
thiophenes (DBT), and more precisely by the -substituted
alkyldibenzothiophenes (1, 9–13). 4,6-Dimethyldibenzo-
thiophene (4,6-DMDBT) has been recognized as a highly
refractory sulfur-containing compound. This molecule has
alwaysbeen present in hydrotreated gasoils, even after hard
HDS conditions (14). The synthesis of 4,6-DMDBT has
been reported and its use as a model to improve the reactiv-
ity of HDS catalysts has been discussed. Kinetic investiga-
tions of the behavior of such alkyldibenzothiophene deriva-
tives under HDS conditions have led to two contradictory
explanations. The first hypothesis suggests that the trans-
formation of 4-alkyl- and 4,6-dialkyldibenzothiophenes is
limited by the adsorption step; it implies adsorption by the
sulfur atom. In this case the alkyl groups sterically hinder
adsorption (15, 16). The second proposal assumes that the
adsorption of the dibenzothiophene compounds occurs via
the -electrons of the aromatic system. The alkyl groups do
not hinder adsorption on the catalyst surface; instead the
elimination step is assumed to be hindered by steric effects.
Based on our previous studies, which included competitive
experiments (17, 18), we assume that, after adsorption
via an aromatic ring, the molecule is first hydrogenated to
dihydrodibenzothiophene as postulated by Singhal et al.
(19). This unstable intermediate can then be transformed
via two pathways on similar catalytic sites by hydrogena-
tion into tetrahydro- and hexahydrodibenzothiophene or
by desulfurization according to an elimination mechanism
resulting in biphenyl derivatives. Both propositions are
being investigated by several groups.
rization under our reaction conditions.
2000 Academic Press
Key Words: hydrodesulfurization; hydrogenation; alkyldiben-
zothiophenes; kinetics.
INTRODUCTION
Crude oil typically contains about 1 wt% of sulfur; dur-
ing combustion of fuels, SOx, a major air pollutant, is emit-
ted. Nevertheless, legislation should be in force by the year
2005, stipulating that diesel fuels cannot contain more than
0.05 wt% of sulfur. Due to increasingly restrictive legisla-
tion on the number of heteroatoms, especially sulfur, the
production of “cleaner” distillate products is a major chal-
lenge for refiners.
Conventional catalytic hydrotreating processes have
been used for many years to eliminate sulfur from dis-
tillates. These processes proved capable of removing the
so-called “easy sulfur” compounds which include nonaro-
matic sulfur, thiophenes, and benzothiophenes. However,
considerable amounts of “hard sulfur” compounds (0.2 to
0.3 wt% ) are found in the diesel fraction, after conventional
hydrodesulfurization (HDS) (1, 2). Considerable efforts
have been made in past years to develop more active cata-
1 To whom correspondence should be addressed. Fax: 33.4.72.44.53.99.
E-mail: vrinat@catalyse.univ-lyon1.fr.
Increasingly sophisticated analytical devices, especia-
lly gas chromatographs equipped with sulfur-specific
255
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

256
MACAUD ET AL.
detectors, have shown that other sulfur-containing mole-
cules with a higher molecular mass are particularly resis-
tant to classical HDS (20). Many research groups are de-
termining the precise structure of these molecules and are
trying to understand the reasons for their low reactivity.
Following our previous investigations, we now report the
synthesis of various alkyldibenzothiophenes (4,6-diethyl-
dibenzothiophene, 4,6-diisobutyldibenzothiophene, and
4,6-diisopropyldibenzothiophene) that bear bulky sub-
stituents in both positions, which hinder the sulfur atom.
We attemped to identify these structures in hydrotreated
gas oils and we compared their reactivities in a batch reactor
and over an industrial NiMo/Al2O3 hydrotreating catalyst.
FIG. 1. Synthesis of 4,6-diethyl- and 4,6-diisobutyldibenzothiophene.
iodide. 4,6-Diisobutyldibenzothiophene (4,6-DiBuDBT)
was analogously prepared in one step (60% yield) by
metalation of 4,6-DMDBT and subsequent reaction with
isopropyl iodide. The schemes of these syntheses are
shown in Fig. 1.
We were also interested in preparing a dibenzoth-
iophene derivative that was sterically more hindered
near the sulfur atom. We thus aimed to synthesize 4,6-
diisopropyldibenzothiophene (4,6-DiPrDBT) by metala-
tion and methylation of 4,6-DEDBT. This resulted in a
mixture of numerous dialkylated dibenzothiophenes from
which it was not possible to isolate the desired com-
pound. Attempts to prepare 4,6-DiPrDBT by direct met-
alation and subsequent alkylation of DBT with isopropyl
iodide failed. Numerous other electrophiles (e.g., acetone,
N,N-dimethylformamide, dimethyl carbonate, trimethyl or-
thoacetate, or 1-acetylimidazole), precursors for the forma-
tion of the isopropyl substituent, were further tested. None
of them led to the formation of the desired product, and
only very poor conversion of DBT was observed.
It proved, however, possible to quench 4,6-dilithiodi-
benzothiophene with dry ice with a high yield (88% ).
The corresponding dicarboxylic compound was esterified
and then reacted with methylmagnesium bromide to the
tertiary diol. This product was dehydrated and hydro-
genated with Pd/C under H2 atmospheric pressure to 4,6-
diisopropyldibenzothiophene with high purity (14% over-
all yield). A schematic representation of the synthesis is
given in Fig. 2.
SYNTHESIS OF VARIOUS
4,6-DIALKYLDIBENZOTHIOPHENES
General
Dialkyldibenzothiophenes are generally not commer-
cially available, because they are not easily accessible. Very
pure 4,6-DMDBT was prepared in a four-step procedure
(25% global yield), starting with a coupling reaction of
o-thiocresol and 2-bromo-3-nitrotoluene (21). However, to
obtain a sufficient quantity of this compound, a more rapid
and inexpensive method was developed by modifying the
procedure published by Caube`re et al. (22). Indeed, bis-
lithiation of dibenzothiophene, which forms an insoluble
species, was performed by adding tetramethylethylenedi-
amine (TMEDA) to activate the metallating reagent, i.e., n-
BuLi. Methylation wasthen performed usingmethyliodide.
Caube`re et al. (22) also tested this procedure for prepar-
ing higher homologues of 4,6-DMDBT. When ethyl iodide,
instead of methyl iodide, was used with a larger amount of
electrophile, 4,6-DEDBT was obtained but with a yield of
only 40% . The successful preparation of 4,6-DEDBT was
also achieved by performing the lithiation at position 6 of
4-ethyldibenzothiophene followed byalkylation usingethyl
iodide (23, 24); the yield was 47% .
In the course of our studies on the synthesis of 4,6-
DMDBT, we observed that the introduction of both
methyl groups by two-step synthesis (i.e., lithiation
and methylation of isolated 4-methyldibenzothiophene)
results in 4-ethyldibenzothiophene as the major product.
The lithiation of the methyl group is preferred to the trans-
formation of the 6 position of 4-methyldibenzothiophene.
This result is explained by better solubility of the dilithiated
4,6-DMDBT than of the 4,6-dilithiodibenzothiophene in
cyclohexane. We thus decided to use this procedure to
prepare various 4,6-dialkyldibenzothiophene derivatives,
starting from 4,6-DMDBT and performing its dilithiation
on both methyl groups.
We prepared 4,6-DEDBT (71% yield) starting from 4,6-
DMDBT in the presence of n-BuLi, TMEDA, and methyl
FIG. 2. Synthesis of 4,6-diisopropyldibenzothiophene.

HYDRODESULFURIZATION OF ALKYLDIBENZOTHIOPHENES
257
FIG. 3. Gas chromatography (Sievers detection) of a desulfurized gas oil: identification of some refractory compounds.
Sulfur derivatives obtained by these syntheses have been dide (Fig. 1). First, 10.2 ml of BuLi was slowly added to a so-
used, with the help of gas chromatography, as authentic lution (2.5 ml) of TMEDA in 15 ml of anhydrous cyclohex-
samples to identify the structures of some refractory com- ane under stirring at 273 K. The mixture was stirred further
pounds contained in desulfurized gasoils (see Fig. 3). We for 45 min at that temperature and then diluted with 30 ml of
then compared their reactivity in a batch reactor, on an anhydrous cyclohexane. At room temperature, 1.14 g of 4,6-
industrial NiMo/Al2O3 hydrotreating catalyst.
DMDBT (5.4 mmol) was added, and the resulting reddish
solution was refluxed for 3 h. Isopropyl iodide (2.15 ml) was
then slowly added (1 h) at 258 K, and the mixture was stirred
at room temperature overnight. The mixture was then hy-
drolyzed, and the organic phase was washed with saturated.
NH4Cl, and dried over magnesium sulfate. After purifica-
tion in a silica column (CH2Cl2/cyclohexane 8/2) and re-
crystallization in isopropanol, 1.18 g of 4,6-DiBuDBT was
obtained.
Experimental
4,6-Dimethyldibenzothiophene (4,6-DMDBT) was pre-
pared according to the method of Caube`re et al. (22) after
optimization by metalation and subsequent methylation of
DBT.
4,6-Diethyldibenzothiophene (4,6-DEDBT) was pre-
pared in one step (71% yield) by metalation and subsequent
methylation of 4,6-DMDBT, as described in Fig. 1. First,
40.8 ml of BuLi was slowly added to a solution (10 ml) of
TMEDA in 20 ml of anhydrous cyclohexane under stirring
at 273 K. The mixture was stirred further for 45 min at 273 K
and then diluted with 40 ml of anhydrous cyclohexane.
At room temperature, 5.76 g of 4,6-DMDBT (27.17 mmol)
was added, and the resulting reddish solution was re-
fluxed for 3 h. Methyl iodide (5.36 ml) was then slowly
added (1 h) at 258 K, and the mixture was stirred at
room temperature overnight. The mixture was then hy-
drolyzed, and the organic phase was washed with saturated.
NH4Cl, and dried over magnesium sulfate. After purifica-
tion in a silica column (CH2Cl2/cyclohexane 8/2) and re-
crystallization in isopropanol, 4.97 g of 4,6-DEDBT was
obtained.
1H NMR (CDCl3): 0.9 (d, 12H), 2.15 (m, 2H), 2.7
(d, 4H), 7.15 (d, 2H), 7.35 (t, 2H), 7.9 (d, 2H). 13C NMR
(CDCl3): 21.9, 30.2, 41.6, 124.8, 126.6, 128.2, 134.9, 138.2,
144.1.
4,6-Diisopropyldibenzothiophene (4,6-DiPrDBT) was
prepared in five steps (14% overall yield, not optimized)
starting from DBT (Fig. 2). First 108.8 ml of BuLi was slowly
added to a solution (26.2 ml) of TMEDA in 50 ml of anhy-
drous cyclohexane under stirring at 273 K. The mixture was
stirred further for 45 min at 273 K and then diluted with
100 ml of anhydrous cyclohexane. At room temperature,
8 g of DBT (43.5 mmol) was added, and the resulting
reddish solution was refluxed for 3 h. Dry ice was then
slowly added (1 h bubbling) at 258 K, and the mixture
was stirred at room temperature overnight. The mixture
was then hydrolyzed, and the organic phase was washed
several times with water. The combined aqueous phases
were washed with ether and then acidified with 1 N
HCl, at which point the desired product precipitated. Af-
ter filtration, the white crystals were washed with water
and ether, and dried over P2O5. A total 10.45 g of 4,6-
dicarboxydibenzothiophene was obtained with 88% yield.
1H NMR (CDCl3): 1.4 (t, 6H), 3.0 (q, 4H), 7.2 (d, 2H),
7.4 (t, 2H), 8.0 (d, 2H). 13C NMR (CDCl3): 13.6, 28.1,
119.3, 124.8, 125.1, 136.3, 138.2, 138.6.
4,6-Diisobutyldibenzothiophene (4,6-DiBuDBT) was
analogously prepared in one step (60% yield) by metalation
of 4,6-DMDBT and subsequent reaction with isopropyl io-

258
MACAUD ET AL.
Next, 5 g of 4,6-dicarboxydibenzothiophene (18.38 mmol) column (HP 5, 5% phenylmethylsilicone, 30 m 0.53 mm)
was stirred with 50 ml of thionyl chloride at 313 K for equipped with an FID detector. The different intermedi-
10 h and then at room temperature for 12 h. The ex- ates and products of the reaction were identified by GC/MS
cess thionyl chloride was evaporated at lower pressure, analyses. The reactant conversion and product yields
and 15 ml of methanol and 10.2 ml of triethylamine were were determined relative to an internal standard using
added to the residue. The mixture was refluxed for 4 h hexadecane as the reference.
and stirred overnight at room temperature. The usual pro-
Control of the kinetic regime. Previous experiments
cedure (water/ethyl acetate) gave 2.87 g (52% ) of 4,6-
were carried out to ensure the absence of intraparticle and
dicarboxymethyldibenzothiophene. This compound (2.5 g,
interphase mass-transfer limitations. First the effects of the
8.3 mmol) was dissolved in 80 ml of anhydrous THF, and
stirring rate and the weight and granulometry of the cata-
the resulting suspension was cooled to 273 K. Next, 27.8 ml
lyst were controlled with a nonhindered molecule (DBT);
of MeMgBr (3 M in ether, 83 mmol) was added dropwise
the results indicated that the kinetic regime was established
and the mixture was finally refluxed for 4 h. The usual
in all cases (25). Then, the internal diffusion of a more hin-
procedure (water/ethyl acetate) gave 2.33 g (93% ) of the
dered compound (4,6-DiBuDBT) was studied by varying
desired diol. This compound was then quantitatively dehy-
the granulometry (50–80 m and 80–125 m) of the cat-
drated in refluxing acetic acid with HCl. The corresponding
alyst. It appeared that there were no diffusional problems
unsaturated substituted dibenzothiophene was dissolved
under our operating conditions when the mass of the cata-
in 25 ml of a 1/1 mixture of MeOH/CH2Cl2, 650 mg of
lyst was limited to 1 g and the stirring rate was fixed to at
Pd/C (10 mol% ) was added, and the mixture was stirred
least 25 Hz.
overnight under atmospheric H2 pressure. After filtra-
tion and purification using silica gel chromatography
(CH2Cl2/heptane 1/9), 533.8 mg of 4,6-DiPrDBT was ob-
Kinetic Calculations
Some experiments were performed with single products
and others with equimolar mixtures of two compounds re-
acting in competition.
tained (33% yield).
1H NMR (CDCl3): 1.5 (d, 12H), 3.3 (m, 2H), 7.3 (d, 2H),
7.4(t, 2H), 8.0(d, 2H). 13C NMR (CDCl3): 22.7, 33.6, 119.3,
122.6, 125.0, 136.4, 138.1, 142.9.
Hydrodesulfurization ofa singlecompound. The results
of a typical reaction course over time are presented for the
transformation of 4,6-DEDBT in Fig. 4 . The initial rate of
the transformation can be obtained from the slope of the
4,6-DEDBT disappearance curve at the origin. The yields
of the different products can also be plotted against the
conversion in order to determine the reaction scheme and
the product selectivities (Fig. 5).
KINETIC STUDY
Experimental Procedure
Materials. The catalyst used for the hydrodesulfu-
rization of alkyldibenzothiophenes was an industrial
NiMo/Al2O3 catalyst (Mo, 9 wt% ; Ni, 2.4 wt% ), provided as
extrudates (bulk density, 0.71 g cm 3; pore volume, 0.48 cm3
Hydrodesulfurization in competitive experiments. In
competitive experiments, two compounds A and B compete
g
1; surface area, 185 m2 g 1). The extrudates were crushed,
screened (80–125 m), and then sulfided ex situ in an H2–
H2S (15% ) flow (4 l/h) at 673 K for 4 h before use.
Reactor and analytical. Studies were performed in a
stirred slurry tank reactor (250 ml) operating in batch
mode. The autoclave was equipped with a hollow-shaft, six-
bladed magnetically driven turbine with four baffles on the
wall to prevent vortex formation. The samples were col-
lected through a tube ( 116 -in. diameter); hydrogen was in-
troduced through a pressure controller, which maintains a
constant pressure during the course of the experiment. In
each run, the autoclave was charged with 250 mg of reac-
tant dissolved in 80 ml of dodecane. Freshly sulfided catalyst
(400 mg) was added to the solution. The reactor was flushed
with nitrogen and heated under stirring to reach the reac-
tion temperature of 573 K; hydrogen was then introduced
(Ptot = 5 MPa); this step was considered as the beginning
of the reaction. Samples were periodically removed dur-
ing the course of the reaction and analyzed by gas chro-
FIG. 4. Transformation of a single molecule (4,6-DEDBT): conver-
sion versus time. Experimental conditions: 573 K, 5 MPa, NiMo catalyst;
4, 4,6-DEDBT; +, alkylbiphenyl; ᭹, alkylcyclohexylbenzene; ᭜, hydro-
matography. GC analyses were performed using an apolar genated compounds.

HYDRODESULFURIZATION OF ALKYLDIBENZOTHIOPHENES
259
FIG. 5. Diagrams of yield versus conversion for the tranformation of DBT, 4,6-DMDBT, 4,6-DEDBT, 4,6-DiBuDBT, and 4,6-DiPrDBT
(᭹, alkylcyclohexylbenzenes; ᭜, hydrogenated compounds; + , alkylbiphenyls).
for adsorption on the same sites. R is the common reactant where the underlined components (i.e., A) correspond to
in the transformation of A and B.
the adsorbed species, k1i and k represent the rate con-
1i
The reactions can be schematized as follows:
stant of adsorption and desorption, respectively, of compo-
nent i and ki the rate constant of the surface reaction of
component i.
At the stationary state of A the balance between the for-
mation and the disappearance of this species could be ex-
pressed by
kA
R
A
R
A
A + R → C
k 1 A
k B
B + R → D ,
k
1 A
k 1 B
B
B
k1A[A][S] = k 1A[A] + kA[A][R].
[1]
k
1 B

260
MACAUD ET AL.
The formation rate of C is
That was also observed for 4,6-DMDBT and was assumed
to be true for all other compounds.
= kA[A][R],
[2]
[3]
c
Therefore, a comparison of the R values with the rate
ratios kA/kB, calculated from the transformation of each
product in separate experiments, enables the determination
of the ratio KA/KB of the adsorption equilibrium constants.
where [A] can be expressed by
k
1A[A][S]
[A] =
.
k
1A + kA[R]
Kinetic Results and Discussion
In this expression, [i] represents the concentration of com-
ponent i after a known reaction time, [S] is the concen-
tration of free active sites (unoccupied sites available for
adsorption) at the surface, and [R] is the concentration
of hydrogen adsorbed on the catalyst surface. In this ex-
pression no assumption was made about the mechanism
of hydrogenation via molecular or dissociated hydrogen,
though recent results (26, 27) strongly suggest that over su-
fide catalysts hydrogenation and hydrodesulfurization are
performed with heterolytic dissociated hydrogen.
Transformation of the single molecules. The reaction
products of the HDS of DBT were H2S, biphenyl (BP),
cyclohexylbenzene (CHB), and traces of partially hydro-
genated sulfur-containingintermediatessuch astetrahydro-
and hexahydrodibenzothiophene. The same substituted
analogues are observed for the substituted DBT. More-
over, the curves representing the yields of products versus
the conversion of the studied molecule (Fig. 5) suggest that
these alkyldibenzothiophenes are transformed according
to two parallel routes in agreement with previous propos-
als in the literature reported for DBT, 4-methylDBT, and
4,6- DMDBT (8, 17, 19, 20). The first route, referred to as
the direct desulfurization pathway (DDS), gives biphenyl
derivatives. These desulfurized molecules are then slowly
hydrogenated into cyclohexylbenzenes. The second path-
way, referred to asthe hydrogenation route (HYD), consists
of a preliminary hydrogenation of one aromatic ring, giving
tetrahydro- and hexahydrodibenzothiophene or analogues
(HN). These intermediates can then be desulfurized.
From the curves in Fig. 5 we can determine the selec-
tivities to all products; these appear to be related to the
presence and size of the alkyl groups. In the case of DBT,
the major product observed is BP, with CHB appearing
significantly as a secondary product only at conversions
higher than 30% . Only trace amounts of hydrogenated and
nondesulfurized DBT are present. For 4,6-DMDBT, 4,6-
DEDBT, and 4,6-DiBuDBT, the products result mainly
from the hydrogenation pathway, leading to substi-
tuted cyclohexylbenzene and hydrogenated undesulfurized
molecules; the concentration of biphenyl derivatives re-
mains very low whatever the conversion. Selectivities cal-
culated at 10% conversion are reported in Table 1, which
also contains the relative rates of transformation.
Similarly, [B] can be expressed by
k
1B[B][S]
+ kB[R]
[B] =
.
[4]
k
1B
If [R] is the same in both expressions, which means that the
same hydrogen species are involved in the transformation
of A and B, then the ratio is expressed by
d[A]
d[B]
kAk1A[B]
kBk1B[A]
k
k
+ kB[R]
.
C
D
1B
=
=
[5]
1A + kA[R]
By integration of this equation, the reactivity ratio RA/B of
the molecules A and B can be obtained (28–30):
log [A]/[A0]
log [B]/[B0]
kAk1A
kBk1B
k
k
+ kB[R]
1B
=
= RA/B
.
[6]
1A + kA[R]
The ratio RA/B at constant temperature does not depend
on the initial composition of the mixture or on the extent
of conversion. At relatively low hydrogen pressure, the sur-
face reaction is the rate-determining step and RA/B can be
expressed as a function of adsorption equilibrium constants
and rate constants.
kA KA
RA/B
=
,
[7]
kB KB
TABLE 1
Catalytic Activity Results (Selectivities Are Calculated
at 10% Conversion)
with
k1i
Ki =
[8]
k
1i
Transformation
rate (relative to
DBT = 100)
S(BP),
%
S(CHB),
S(HN),
%
representing the adsorption equilibrium constant of com-
pound i at the surface. We verified that, under our exper-
imental conditions, a pressure of 5 MPa still corresponds
to the range of low hydrogen pressure and, thus, that the
above expression can be used (25). Moreover, a separate
study of DBT indicated that, under our experimental con-
ditions, the order of reaction relative to DBT was zero (25).
%
DBT
100
15.9
11.9
6.3
90
9
13
10
—
10
30
29
40
—
0
61
58
50
100
4,6-DMDBT
4,6-DEDBT
4,6-DiBuDBT
4,6-DiPrDBT
2.3

HYDRODESULFURIZATION OF ALKYLDIBENZOTHIOPHENES
261
For 4,6-DiPrDBT, the tendency of the reactant to be
transformed according to the hydrogenation pathway is
even more pronounced. Only hydrogenated products are
observed, and no desulfurization is detected under our ex-
perimental conditions. These results confirm that substitu-
tions of alkyldibenzothiophenes in the 4 and 6 positions
significantly affect the direct desulfurization rate. The hy-
drogenation pathway appears to be less affected; therefore,
the hydrogenation ability of the catalyst is of great impor-
tance in the conversion of such substituted compounds.
The initial tranformation rates of the different com-
pounds were calculated from the slopes of the reactant
transformation curves similar to the curve shown in Fig. 4.
These values were compared to the DBT transformation
rate, and ratios of these initial rates are given in Table 1.
The rates decrease in the order
DBT > 4,6-DMDBT > 4,6-DEDBT
> 4,6-DiBuDBT > 4,6-DiPrDBT,
FIG. 6. ln(A/A0) versus ln(B/B0) obtained in competitive experi-
ments. ᭜, A, 4,6-DiBuDBT, B, DBT; ᭹, A, 4,6-DiBuDBT, B, 4,6-DEDBT;
+, A, 4,6-DiPrDBT, B, 4,6-DMDBT; ᭿, A, 4,6-DiBuDBT, B, 4,6-DMDBT;
᭡, A, 4,6-DEDBT, B, 4,6-DMDBT; —, A, 4,6-DMDBT, B, DBT.
which confirms that HDS of the substituted DBTs is in-
hibited by the presence of alkyl groups at the carbon to
the sulfur atom. However, though 4,6-DMDBT (which is
about 6 times less reactive than DBT) is generally consid-
ered to be a model compound for resistant sulfur molecules,
it should be noted that the transformation of 4,6-DiPrDBT
is more than 6 times slower than that of 4,6-DMDBT, and
only hydrogenated products are observed. With regard to
classical hydrodesulfurization, such a molecule would be
considered to be a highly unreactive polyaromatic sulfur
compound. A competitive study was carried out to deter-
mine the phenomena that might explain such inhibition.
validity of the competitive experimental approach and en-
able the calculation of the corresponding reactivity ratios
RA/B
.
The results in Table 2 show that, for most of the
alkylDBT couples, except for experiments including the
4,6-DiPrDBT, the rate ratio kA/kB calculated from separate
experiments is close to the reactivity ratio RA/B obtained
from competitive experiments. Therefore, the adsorption
equilibrium constants are almost the same for all studied
molecules, suggesting that these compounds are adsorbed
in the same way on the catalyst surface. Taking these re-
sults into account, adsorption by the sulfur atom is unreal-
istic. The constants of adsorption would otherwise be very
different for molecules sterically hindered by alkyl groups
around the sulfur atom. We can thus assume that adsorp-
tion of the DBT derivatives takes place via the -electrons
of the aromatic ring.
Competitive experiments. Competitive studies were
performed in order to determine whether the very low re-
activities of some alkylated DBTs were due to difficulties
with adsorption on a catalytic site or to differences in the
surface reaction rates.
According to the theory of competitive hydrogenation,
competitive experiments enable the determination of the
reactivity ratio RA/B = kAKA/kBKB. The rate constant ratio
kA/kB can be easily derived from the ri values (initial trans-
formation rate of compound i) obtained under conditions at
which a zero order relative to the hydrocarbon compounds
is verified. This rate ratio, compared to the reactivity ratio
obtained in competition, enables the calculation of the ra-
tio KA/KB of the adsorption equilibrium constants of the
two molecules. This calculation was made for various cou-
ples of alkyldibenzothiophenes using an equimolar mix-
ture of the two compounds under standard experimental
conditions (NiMo/Al2O3 catalyst, total pressure of 5 MPa,
573 K). Samples were periodically removed at increasing
conversion to calculate the concentration of each reactant
(A and B). The curves ln(A/A0) versus ln(B/B0) are given
in Fig. 6, and the straight lines obtained demonstrate the
If the adsorption strength is not responsible for the dif-
ference in the reactivity of the various alkylDBT molecules,
TABLE 2
Results of Competitive Experiments
Products A and B
kA/kB
RA/B
KA/KB
4,6-DMDBT/DBT
0.16
0.77
0.53
0.06
0.40
0.024
0.15
0.13
1
0.81
1.30
1
1.66
2.07
3.5
4,6-DEDBT/4,6-DMDBT
4,6-DiBuDBT/4,6-DEDBT
4,6-DiBuDBT/DBT
4,6-DiBuDBT/4,6-DMDBT
4,6-DiPrDBT/DBT
0.53
0.10
0.83
0.08
0.56
4,6-DiPrDBT/4,6-DMDBT
3.7

262
MACAUD ET AL.
groups of this compound is lower than for 4,6-DiBuDBT,
the steric hindrance generated near the sulfur atom is more
important. The tertiary carbon atom is closer to the sulfur
in the isopropyl compound compared to the isobutyl com-
pound. Such evolution of the transformation rates with the
size as well as with the nature of the alkyl substituents is in
good agreement with the proposed mechanism.
CONCLUSION
Kinetic studies of individual compounds and competitive
experiments with mixtures have proved that the drastic de-
crease in the global reaction rate of alkyl- substituted diben-
zothiophenes, compared to DBT, is due not to adsorption
but to steric hindrance in the elimination reaction involved
in the C–S bond scission. Such inhibition is increasingly pro-
nounced on going from 4,6-DMDBT to 4,6-DEDBT and to
4,6-DiBuDBT, in agreement with the increase in the size of
the substituent.
In the case of 4,6-DiPrDBT, the steric effect leads to the
elimination of the desulfurization under the chosen condi-
tions. This compound might be considered as a better model
of highly unreactive polyaromatic sulfur, since its transfor-
mation rate is about 2 orders of magnitude lower than that
of dibenzothiophene.
ACKNOWLEDGMENTS
This work was conducted within the scope of the contract “Post traite-
ment de coupes hydrotraite´es” and was supported by CNRS-Ecodev, Elf,
IFP, Total, and Procatalyse.
FIG. 7. Reaction scheme for the transformation of alkyl-DBT.
REFERENCES
an explanation must be sought in the mechanism of evolu-
tion of the adsorbed species. For 4,6-DMDBT, 4,6-DEDBT,
and 4,6-DiBuDBT, the decrease in the transformation rate
is more pronounced with increasing size of the alkyl group,
and these results seem to agree with the reaction scheme
involving a partially hydrogenated intermediate, as previ-
ously proposed for the transformation of DBT, 4-MDBT,
and 4,6-DMDBT (8, 17, 19, 20). This mechanism is shown in
Fig. 7. Indeed, the lower reactivity would be related to the
lower reaction rate for the C–S bond cleavage (elimination
reaction) due to steric hindrance in the basic attack; this
hindrance is generated by the alkyl group near the sulfur
atom.
However, a detailed examination of the results obtained
for the 4,6-DiPrDBT was interesting. This compound not
only has the lowest reactivity of the various sulfur com-
pounds studied, but competitive experiments indicate that,
compared to DBT or 4,6-DMDBT, it also shows slightly
stronger adsorption on the catalyst surface (KA/KB = 3.5).
Therefore, this low reactivity and the absence of desulfu-
rization (only hydrogenated products are observed) was
explained by more pronounced steric hindrance in the ba-
sic attack. Though the number of carbon atoms of the alkyl
1. Kabe, T., Ishihara, A., and Tajima, H., Ind. Eng. Chem. Res. 31, 1577
(1992).
2. Landau, M. V., Catal. Today 36, 393 (1997).
3. Breysse, M., Portefaix, J. L., and Vrinat, M., Catal. Today 10, 489
(1991).
4. Luck, F., Bull. Soc. Chim. Belg. 100, 781 (1991).
5. Grange, P., and Vanhaeren, X., Catal. Today 36, 375 (1997).
6. Sakanishi, K., Ando, M., Abe, S., and Mochida, I., Sekiyu Gakkaishi
34, 553 (1991).
7. Ma, X., Sakanishi, K., and Mochida, I., Fuel 73, 1667 (1994).
8. Whitehurst, D. D., Isoda, T., and Mochida, I., Adv. Catal. 42, 345(1998).
9. Ma, X., Sakanishi, K., and Mochida, I., Ind. Eng. Chem. Res. 34, 748
(1994).
10. Houalla, M., Broderick, D., de Beer, V. J. H., Gates, B. C., and Kwart,
H., Pap. Am. Chem. Prepr. Div. Petrol. Chem. 22, 941 (1977).
11. Amorelli, A., Amos, Y. D., Halsig, C. O., Kosman, J. J., Jonker, R. J.,
de Wind, M., and Vrieling, J., Hydrocarbon Processing 93 (1992).
12. Ishihara, A., Tajima, H., and Kabe, T., Chem Lett. 669 (1992).
13. Sakanishi, K., Ando, M., and Mochida, I., Sekiyu Gakkaishi 35, 403
(1992).
14. Kilanowski, D. R., Teeuwen, H., de Beer, V. J. H., Gates, B. C., Schuit,
G. C. A., and Kwart, H., J. Catal. 55, 129 (1978).
15. Houalla, M., Broderick, D., Sapre, A. V., Nag, N. K., de Beer, V. J. H.,
Gates, B. C., and Kwart, H., J. Catal. 61, 523 (1980).
16. Ma, X., Sakanishi, K., Isoda, T., and Mochida, I., Pap. Am. Chem.
Prepr. Div. Petrol. Chem. 39, 622 (1994).

HYDRODESULFURIZATION OF ALKYLDIBENZOTHIOPHENES
263
17. Meille, V., Schulz, E., Lemaire, M., and Vrinat, M., J. Catal. 170, 29
23. Boberg, F., Bruns, W., and Musshoff, D., Phosphorus, Sulfur Silicon
Relat. Elem. 79, 113 (1993).
(1997).
18. Meille, V., Schulz, E., Lemaire, M., and Vrinat, M., Appl. Catal. 187,
179 (1999).
19. Singhal, G. H., Espino, R. L., Sobel, J. E., and Huff, G. A., J. Catal. 67,
457 (1981).
24. Katritzky, A. R., and Perumal, S., J. Heterocycl. Chem. 27, 1737 (1990).
25. Lamure-Meille, V., Schulz, E., Lemaire, M., and Vrinat, M., Appl.
Catal. 131, 143 (1995).
26. Kasztelan, S., and Guillaume, D., Ind. Eng. Chem. Res. 2(33), 203
(1994).
20. Depauw, G. A., and Froment, G. F., J. Chromatogr. A 761, 231 (1997).
21. Meille, V., Schulz, E., Lemaire, M., Faure, R., and Vrinat, M., Tetrahe- 27. Olguin Orozco, E., and Vrinat, M., Appl. Catal. 170, 195 (1998).
dron 52, 3953 (1996). 28. Maurel, R., and Tellier, J., Bull. Soc. Chim. Fr. 10, 4191 (1968).
22. Kuehm-Caube`re, C., Adach-Becker, S., Fort, Y., and Caube`re, P., Tetra- 29. Radel, C. P., and Smith, H. A., J. Am. Chem. Soc. 84, 1443 (1962).
hedron 52, 9087 (1996). 30. Vrinat, M., and Germain, J. E., J. Chim. Phys. 74, 524 (1977).