Competitive hydrodesulfurization of 4,6-dimethyldibenzothiophene, hydrodenitrogenation of 2-methylpyridine, and hydrogenation of naphthalene over sulfided NiMo/γ-Al2O3

Article abstract of DOI:10.1016/j.jcat.2004.03.005

The hydrodesulfurization (HDS) of 4,6-dimethyldibenzothiophene (4,6-DMDBT) was studied at 340°C and 5 MPa in the presence of 2-methylpyridine and 2-methylpiperidine. Both N-containing molecules inhibited the HDS. The inhibitory effect of 2-methylpiperidine on the direct desulfurization and hydrogenation pathways of the HDS was slightly stronger than that of 2-methylpyridine. The desulfurization of 4,6-dimethyltetrahydrodibenzothiophene, an intermediate in the hydrogenation pathway, was extremely difficult in the presence of N-containing molecules. 4,6-DMDBT, in turn, inhibited the hydrogenation of 2-methylpyridine to 2-methylpiperidine but did not affect the C-N bond cleavage in the hydrodenitrogenation of 2-methylpiperidine. The use of toluene as an aromatic solvent had no effect on the HDS of dibenzothiophene and 4,6-DMDBT at 340°C. Naphthalene inhibited the HDS of dibenzothiophene and 4,6-DMDBT without changing the product distributions. Both S-containing molecules suppressed the hydrogenation of naphthalene to the same extent.

Full text of DOI:10.1016/j.jcat.2004.03.005

Journal of Catalysis 224 (2004) 278–287
www.elsevier.com/locate/jcat
Competitive hydrodesulfurization of 4,6-dimethyldibenzothiophene,
hydrodenitrogenation of 2-methylpyridine, and hydrogenation
of naphthalene over sulfided NiMo/γ -Al2O3
Marina Egorova and Roel Prins ∗
Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology (ETH), 8093 Zurich, Switzerland
Received 16 December 2003; revised 20 February 2004; accepted 5 March 2004
Available online 16 April 2004
Abstract
◦
The hydrodesulfurization (HDS) of 4,6-dimethyldibenzothiophene (4,6-DMDBT) was studied at 340 C and 5 MPa in the presence of
2-methylpyridine and 2-methylpiperidine. Both N-containing molecules inhibited the HDS. The inhibitory effect of 2-methylpiperidine on
the direct desulfurization and hydrogenation pathways of the HDS was slightly stronger than that of 2-methylpyridine. The desulfuriza-
tion of 4,6-dimethyltetrahydrodibenzothiophene, an intermediate in the hydrogenation pathway, was extremely difficult in the presence of
N-containing molecules. 4,6-DMDBT, in turn, inhibited the hydrogenation of 2-methylpyridine to 2-methylpiperidine but did not affect the
C–N bond cleavage in the hydrodenitrogenation of 2-methylpiperidine. The use of toluene as an aromatic solvent had no effect on the HDS
◦
of dibenzothiophene and 4,6-DMDBT at 340 C. Naphthalene inhibited the HDS of dibenzothiophene and 4,6-DMDBT without changing
the product distributions. Both S-containing molecules suppressed the hydrogenation of naphthalene to the same extent.
2004 Elsevier Inc. All rights reserved.
Keywords: Hydrotreating; Hydrodesulfurization; HDS; Dibenzothiophene; DBT; 4,6-Dimethyldibenzothiophene; 4,6-DMDBT; Hydrodenitrogenation;
HDN; 2-Methylpyridine; 2-Methylpiperidine; Hydrogenation; Aromatics; Naphthalene
1. Introduction
teristics [2,3]. Organic nitrogen compounds are constituents
of liquid fuels that contribute to harmful NOx emissions
and poison acidic catalysts. They are among the strongest
inhibitors of hydrodesulfurization (HDS) [4–10]. This in-
hibitory effect becomes more pronounced under conditions
of deep HDS when the amounts of S and N compounds are
comparable.
Dibenzothiophene (DBT) and its substituted derivatives
undergo HDS via two parallel pathways: (i) direct desul-
furization (DDS) or hydrogenolysis leading to the forma-
tion of biphenyls and (ii) hydrogenation (HYD) followed
by desulfurization giving first tetrahydro- and hexahydro-
dibenzothiophenes, which are further desulfurized to cyclo-
hexylbenzenes and bicyclohexyls [11–14]. DBT and 4,6-di-
methyldibenzothiophene(4,6-DMDBT) are used as model S
compounds in our studies, because they represent the family
of the most refractory S-containing molecules in oil distil-
lates. It is well known that, over a NiMo/Al2O3 catalyst,
DBT reacts preferentially via the DDS pathway, whereas
4,6-DMDBT converts mainly via the HYD pathway [15,16].
This is due to the strong steric hindrance of the two methyl
Research on the purification of fuels, including desul-
furization, denitrogenation, and dearomatization, has be-
come an important subject of environmental catalysis studies
worldwide. The three major types of transportation fuels,
gasoline, jet, and diesel fuel, differ in composition and prop-
erties [1]. Diesel fuel contains the most refractory sulfur
compounds, alkylated benzothiophenes and alkylated and
nonalkylated dibenzothiophenes.The sulfur content in diesel
fuel is of environmental concern because, upon combustion,
sulfur-containing molecules are converted to hydrocarbons
and SOx; the latter compounds not only contribute to acid
rain but also poison the catalytic converter for the treatment
of exhaust emission. Reducing the contents of aromatics as
well as sulfur is generally desirable with respect to the qual-
ity of diesel fuel; the reduction of aromatics increases the
cetane number and generally improves combustion charac-
*
Corresponding author. Fax: +41 1 632 1162.
E-mail address: roel.prins@chem.ethz.ch (R. Prins).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.03.005
2004 Elsevier Inc. All rights reserved.

M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
279
groups in the 4 and 6 positions adjacent to the sulfur atom of
DBT. The methyl substituents hinder the perpendicular one-
point adsorption of the molecule on the catalyst surface that
is needed for the DDS pathway of the HDS. The overall reac-
tivity of DBT over Ni- or Co-promoted Mo/Al2O3 catalysts
is one order of magnitude higher than that of 4,6-DMDBT
[13,17,18].
erence for DBT, 4,6-DMDBT, and their derivatives in the GC
analysis), 11 kPa heptane (as reference for the N compounds
in the GC analysis), 1 kPa DBT, or 4,6-DMDBT, 2–10 kPa
2-MPy or 2-MPiper, 1–10 kPa naphthalene, 35 kPa H2S, and
4.8 MPa H2. The partial pressure of H2S in all the experi-
ments was 35 times higher than that of DBT or 4,6-DMDBT
to avoid an effect of the H2S formed during the HDS reac-
tion.
The first systematic studies on simultaneous catalytic
HDS and hydrodenitrogenation (HDN) were performed by
Satterfield et al. [19,20]. They investigated the interaction
in the HDS of thiophene and HDN of pyridine. However,
when using thiophene as S-containing molecules, one can-
not distinguish the DDS and HYD pathways in the HDS.
In our previous work on the mutual influence of HDS
and HDN we used DBT as a model S compound and 2-
methylpyridine (2-MPy) and 2-methylpiperidine(2-MPiper)
as N-containing molecules [16]. Both N compounds had a
strong inhibitory effect on the HYD pathway of the HDS
of DBT, with 2-MPiper having a somewhat stronger effect
than 2-MPy. Both N-containing molecules also inhibited the
DDS pathway of the HDS of DBT, 2-MPy being a stronger
inhibitor than 2-MPiper. In the work reported here we stud-
ied the influence of 2-MPy and 2-MPiper on the HDS of
4,6-DMDBT, which undergoes HDS predominantly via the
HYD route, and compared the results with those obtained for
the HDS of DBT. We investigated the effect of the solvent
and of naphthalene on the HDS of DBT and 4,6-DMDBT.
The reverse influence of S-containing molecules on the HDN
of 2-MPy and 2-MPiper and on the HYD of naphthalene was
investigated as well.
The reaction products were analyzed by on- and off-line
gas chromatography as described earlier [16]. Weight time
was defined as τ = wcat/nfeed, where wcat denotes the cata-
lyst weight and nfeed the total molar flow to the reactor. The
weight time (τ) was changed by varying the flow rates of
the liquid and the gaseous reactants while keeping their ra-
tio constant. The reaction was stable after 3 to 4 h; during
2 weeks of operation almost no deactivation of the catalyst
was observed. A single experiment lasted 36–48 h. The com-
petitive experiments with different amounts of N-containing
molecules were always carried out on the same sample of a
catalyst.
3. Results
3.1. HDS of 4,6-dimethyldibenzothiophene
The rate of the HDS of 4,6-DMDBT and the resulting
product distribution were investigated at 300 and 340 ◦C.
4,6-DMDBT undergoes HDS via the same reaction path-
ways as DBT: direct desulfurization and hydrogenation fol-
lowed by desulfurization. The HYD pathway predominates
at both reaction temperatures, since the amount of products
obtained via this route is four to six times higher than that of
3,3ꢀ-dimethylbiphenyl formed via the DDS route (Figs. 1a
and 2a). The selectivity toward 3,3ꢀ-dimethylbiphenyl for-
mation or the DDS selectivity is 15% at 300 ◦C and 25% at
340 ◦C; the selectivity toward the HYD pathway is 85 and
75%, respectively (Figs. 1b and 2b). The selectivity toward
the formation of 3,3ꢀ-dimethylbiphenyl remains constant
during the course of the reaction, showing that this prod-
uct is not hydrogenated further. 4,6-Dimethyltetrahydrodi-
benzothiophene formed via hydrogenation of 4,6-DMDBT
is desulfurized to methylcyclohexyltoluene. We consider
that 3,3ꢀ-dimethylbicyclohexyl is the product of further hy-
drogenation of the tetrahydro-intermediate to the perhydro-
intermediate and subsequent desulfurization and not the
product of the hydrogenation of methylcyclohexyltoluene.
This is because the hydrogenation of the second ring in the
biphenyl-like structures should be much more difficult than
that of the first ring, since the partially hydrogenated mole-
cule is not flat any more and it is more difficult to adsorb such
a molecule on the catalyst surface as compared to biphenyl.
Moreover, we know that the hydrogenation of the first ring
of 3,3ꢀ-dimethylbiphenyl does not occur during the HDS of
4,6-DMDBT. The perhydro-intermediate was not observed
in the product, probably because of its fast conversion and
2. Experimental
The NiMo/γ -Al2O3 catalyst contained 8 wt% Mo and
3 wt% Ni and was prepared by successive incipient wet-
ness impregnation of γ -Al2O3 (Condea, pore volume
0.5 cm3 g−1, specific surface area 230 m2 g−1 with an aque-
ous solution of (NH4)6Mo7O24 · 4H2O followed by an aque-
ous solution of Ni(NO3)2 · 6H2O (both Aldrich). After each
impregnation step the catalyst was dried in air at ambient
temperature for 4 h, then dried in an oven at 120 ◦C for 15 h,
and finally calcined at 500 ◦C for 4 h.
A sample of 0.05 g of catalyst was diluted with 8 g SiC
to achieve plug-flow conditions in the continuous flow fixed-
bed reactor. The catalyst was sulfided in situ with a mixture
of 10% H2S in H2 (25 ml min−1) at 400 ◦C and 1.0 MPa for
4 h. After sulfidation, the pressure was increased to 5.0 MPa,
the temperature was decreased to reaction temperature, and
the liquid reactant was fed to the reactor by means of a
Gilson 307 piston pump.
Reactions were carried out in continuous mode in a fixed-
bed inconel reactor as described previously [16]. The exper-
iments were carried out at 300 and 340 ◦C. The composition
of the gas-phase feed consisted of 130 kPa toluene or decane
(as solvent for the DBTs and amine), 8 kPa dodecane (as ref-

280
M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
◦
Fig. 1. Relative partial pressures (a) and selectivities (b) of the products of the HDS of 4,6-dimethyldibenzothiophene at 300 C as a function of weight time.
◦
Fig. 2. Relative partial pressures (a) and selectivities (b) of the products of the HDS of 4,6-dimethyldibenzothiophene at 340 C as a function of weight time.
suppresses the unwanted side reaction of disproportionation,
which usually takes place in the HDN of pyridine and piperi-
dine [21].
The results of competitive experiments performed at 2
and 6 kPa 2-MPy and 2-MPiper are shown in Figs. 3a and 3b.
Both N-containing molecules strongly inhibit the overall
HDS of 4,6-DMDBT. Already at 2 kPa 2-MPy or 2-MPiper,
the conversion of 4,6-DMDBT decreased by a factor of 5
to 6, and at 6 kPa 2-MPy or 2-MPiper, the conversion of
4,6-DMDBT hardly reached 4% at τ = 5 g min mol−1. The
inhibitory effects of 2-MPy and 2-MPiper are almost the
same (Figs. 3a and 3b). Also the product distribution was al-
most identical; therefore, only the results of 2-MPiper are
presented in Figs. 4a and 4b. The selectivity of the DDS
Scheme 1. Reaction network of the HDS of 4,6-dimethyldibenzothiophene.
pathway in the HDS of 4,6-DMDBT increased in the pres-
ence of the N-containing molecules from 25 to 35% at 2 kPa
2-MPy or 2-MPiper and to 45% at 6 kPa 2-MPy or 2-MPiper,
but the HYD route still remained the major pathway. How-
ever, at 6 kPa 2-MPy and 2-MPiper, the only product of the
HYD route is the partially hydrogenated 4,6-dimethyltetra-
hydrodibenzothiophene. At 2 kPa of 2-MPy or 2-MPiper,
this compound is the main HYD product. The character of
the selectivity curve of 4,6-dimethyltetrahydrodibenzothio-
phene (Fig. 4a) shows that when the N compounds are con-
verted this intermediate can react further. Sulfur removal is,
thus, strongly suppressed in the presence of N-containing
molecules.
thus low concentration. Therefore, the overall reaction net-
work of the HDS of 4,6-DMDBT is as shown in Scheme 1.
3.2. Influence of N-containing molecules on HDS
The effect of the N compounds on the HDS of 4,6-
DMDBT was studied at 340 ◦C, where the conversion of
4,6-DMDBT was high enough to observe inhibitory effects.
2-MPy and 2-MPiper were used as model N-containing
molecules since it was shown previously that the presence of
the methyl group on the α-carbon atom of pyridine strongly

M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
281
◦
Fig. 3. Inhibition of the HDS of 4,6-DMDBT in the presence of 2-methylpyridine (a) and 2-methylpiperidine (b) at 340 C. HDS of 4,6-DMDBT alone (2),
in the presence of 2 kPa (Q) and 6 kPa (!) N-compounds.
◦
Fig. 4. Product distribution in the HDS of 4,6-DMDBT at 340 C in the presence of 2 kPa (a) and 6 kPa (b) of 2-MPiper.
Scheme 2. Reaction network of the HDN of 2-methylpyridine and 2-methylpiperidine.
3.3. Influence of S-containing molecules on HDN
The influence of DBT and 4,6-DMDBT on the HDN
of 2-MPy and 2-MPiper was studied at 340 ◦C. The HDN
network of 2-MPy and 2-MPiper was determined earlier
[21] (Scheme 2). 2-MPiper is the primary product in the
HDN of 2-MPy, since the cleavage of the C–N bond in
heterocyclic N-containing aromatic molecules can only oc-
cur after ring hydrogenation [22]. We found previously that
the first C–N bond breaking in 2-MPiper occurs predomi-
nantly between the nitrogen atom and the carbon atom of
the methylene group and that the methyl group has a nega-
◦
tive rather than a positive influence on the C–N bond break-
ing [21].
Fig. 5. HDN of 6 kPa 2-MPy at 340 C in the presence (– – –) and absence
(—) of 1 kPa 4,6-DMDBT.
In the presence of 1 kPa 4,6-DMDBT, the conversion of
6 kPa 2-MPy and the yield of 2-MPiper decreased, while the
amount of C6 products (hexane, hexenes, and hexylamines)
increased slightly (Fig. 5). No changes were observed in
the HDN of 2-MPiper in the presence of 4,6-DMDBT, nei-
ther in the conversion of 2-MPiper, nor in the formation of
C6 products (not shown). The effect of 4,6-DMDBT on the
HDN of 2-MPy and 2-MPiper was, thus, identical to that of
DBT [16]. Hence, both S-containing molecules have an in-
hibitory effect on the hydrogenation of 2-MPy and no effect
on the C–N bond cleavage in 2-MPiper.

282
M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
◦
Fig. 6. Effect of the solvent on the HDS of DBT (a) and 4,6-DMDBT (b) at 340 C in decane (") and toluene (P).
◦
Fig. 7. Inhibition of the HDS of DBT (a) and 4,6-DMDBT (b) in the presence of naphthalene at 340 C.
3.4. Effect of solvent on HDS
Toluene was the solvent in most of our HDS reactions,
since the solubility of DBT and 4,6-DMDBT is much bet-
ter in an aromatic solvent. However, aromatic molecules
may compete with the S-containing reactant for the active
sites on the catalyst surface. In order to determine whether
the solvent had any effect on the HDS conversion we also
performed HDS experiments of DBT and 4,6-DMDBT in
decane. The results show that the conversions of both DBT
(Fig. 6a) and 4,6-DMDBT (Fig. 6b) are the same in toluene
and decane at 340 ◦C. The product distribution was also the
same. Therefore, we conclude that the use of toluene as a sol-
vent does not influence the HDS reactions of DBT and 4,6-
DMDBT at 340 ◦C. Toluene itself did not undergo hydro-
genation when S- or N-containing molecules were present
in the feed.
Fig. 8. Hydrogenation of 10 kPa naphthalene to tetralin in the presence
(– – –) and absence (—) of 1 kPa DBT (Q) or 4,6-DMDBT (P).
HDS of 4,6-DMDBT decreased by a factor of 1.8, whereas
that of DBT was affected by a factor of 1.34 at 10 kPa naph-
thalene (cf. Figs. 7a and 7b). The selectivities in the HDS of
4,6-DMDBT were not influenced by the presence of naph-
thalene. Thus, naphthalene affected the DDS and the HYD
pathways of the HDS of both S-containing molecules to the
same extent. The inhibitory effect of naphthalene was much
weaker than that of the N-containing molecules.
3.5. Effect of naphthalene on HDS
The effect of naphthalene on the HDS of DBT and 4,6-
DMDBT was studied at 340 ◦C. Naphthalene was used to
study the effect of condensed aromatics on HDS. Already at
1 kPa naphthalene the conversion of DBT decreased slightly
and at 10 kPa naphthalene by 10% (Fig. 7a); the product
distribution remained the same however. The inhibitory ef-
fect of 10 kPa naphthalene was stronger in the HDS of 4,6-
DMDBT than in the HDS of DBT. The initial rate of the
3.6. Influence of S-containing molecules on HYD
The effects of the S-containing molecules on the hydro-
genation of naphthalene are presented in Fig. 8. At a partial

M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
283
◦
Fig. 9. Relative partial pressures (a) and selectivities (b) of the products of the HDS of dibenzothiophene at 340 C as a function of weight time.
pressure of 1 kPa, DBT and 4,6-DMDBT inhibited the hy-
drogenation of 10 kPa naphthalene equally strongly. The
partial pressure of hydrogen sulfide in these experiments
was kept high to avoid an effect of H2S formed during the
HDS reaction. Under the conditions of our study (5 MPa to-
tal pressure and 340 ◦C) naphthalene is mainly converted to
tetralin over NiMo/Al2O3. Further hydrogenation took place
extremely slowly, as demonstrated by the negligible amounts
of decalin: at 75% conversion of naphthalene the sum of cis-
decalin and trans-decalin was only 1%.
Scheme 3. Reaction network of the HDS of dibenzothiophene.
Table 1
4. Discussion
Rate constants of the DDS and HYD pathways in the HDS of DBT and
◦
4,6-DMDBT at 300 and 340 C
The aim of the present study was to compare the influ-
ence of N-containing molecules and aromatics on the HDS
of DBT and 4,6-DMDBT in order to gain insight into the na-
ture of the DDS and HYD active sites. The main difference
between the HDS of DBT and 4,6-DMDBT is that DBT re-
acts predominantly via the DDS pathway and 4,6-DMDBT
mainly via the HYD route. Thus, the selectivities for the
DDS and HYD pathways in the HDS of DBT were 85 and
15% at 300 ◦C and 90 and 10% at 340 ◦C, respectively [16].
The product distribution showed that slow hydrogenation
of biphenyl to cyclohexylbenzene took place, since the se-
lectivity toward biphenyl formation decreased with weight
time and the increase in the cyclohexylbenzene selectiv-
ity with weight time was higher than the decrease in the
tetrahydrodibenzothiophene selectivity (Fig. 9). Therefore,
the overall HDS network of DBT is as shown in Scheme 3.
In the HDS of 4,6-DMDBT, the selectivities toward the
DDS and HYD are 15 and 85% at 300 ◦C and 25 and 75% at
340 ◦C, respectively (Figs. 1b and 2b). Thus, the product dis-
tribution is the reverse as in the HDS of DBT, where the DDS
prevails. The selectivity toward the formation of 3,3ꢀ-di-
methylbiphenyl remained constant during the reaction. This
means that the ratio of the products obtained via two par-
allel pathways, DDS and HYD, did not change. Therefore,
we conclude that the further hydrogenation of 3,3ꢀ-dimethyl-
biphenyl to methylcyclohexyltoluene in the HDS of 4,6-
DMDBT does not take place. Temperature has a stronger
promotional effect on the direct sulfur removal from both
◦
◦
300 C
340 C
k
k
k
k
HYD
HYD
DDS
DDS
DBT
4,6-DMDBT
0.102
0.006
0.018
0.033
0.35
0.03
0.04
0.09
DBT and 4,6-DMDBT, since the DDS pathway is enhanced
at 340 ◦C compared to 300 ◦C. Furthermore, the removal
of sulfur from the partially hydrogenated intermediate im-
proved, because the amount of 4,6-dimethyltetrahydrodi-
benzothiophene decreased and the amount of methylcyclo-
hexyltoluene increased more strongly in the course of the
reaction at higher temperature.
The HDS of DBT and 4,6-DMDBT can be well described
as pseudo-first-order reactions with respect to the reactant,
in good agreement with the literature [23–26]. The rate con-
stants of the DDS and HYD pathways, kDDS and kHYD, can
be obtained from the selectivity data. The kinetic results at
300 and 340 ◦C show that the overall HDS (kDDS + kHYD
)
of DBT is about three times faster than that of 4,6-DMDBT
(Table 1). This value is somewhat smaller than the values
of 5 to 6 [17,27,28] and 10 [13] reported in the literature.
This can be explained by the high partial pressure of H2S
in our experiments to avoid the influence of H2S released
during HDS. Because hydrogen sulfide inhibits the HYD
pathway to a lesser extent than the DDS pathway [18] and
since HYD is the main route for the transformation of 4,6-
DMDBT, the HDS of this molecule is less affected in the

284
M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
ered and one methylene group is under and one above the
plane of the molecule. As a consequence, one of the hydro-
gen atom, at the carbon atom below the plane, hinders the ad-
sorption. When a methyl group is present in the partially hy-
drogenated ring the structure is more rigid and the hydrogen
of the methylene group extends further toward the catalyst
suface. Therefore, the molecule is tilted on the surface and
the adsorption of 4,6-dimethyltetrahydrodibenzothiophene
is weaker than that of tetrahydrodibenzothiophene. Because
of the resulting longer lifetime of 4,6-dimethyltetrahydro-
dibenzothiophene than of tetrahydrodibenzothiophene, the
hydrogenationof the second benzene ring attains importance
and 3,3ꢀ-dimethylbicyclohexyl is formed.
Fig. 10 shows that kHYD, the rate constant for the HYD
pathway in the HDS of DBT, strongly decreases in the pres-
ence of 2-MPy and 2-MPiper [16]; 2-MPiper retards kHYD
to a somewhat greater extent than 2-MPy. The major prod-
uct of the HYD pathway is tetrahydrodibenzothiophene, but
even at the highest partial pressure of 6 kPa 2-MPy and
2-MPiper the reaction of tetrahydrodibenzothiophene to cy-
clohexylbenzene was not totally inhibited. The DDS path-
way was much less affected by 2-MPy and 2-MPiper; kDDS
was hardly affected at 2 kPa 2-MPiper and only slightly at
2 kPa 2-MPy. Since the HYD pathway was strongly inhib-
ited already at low partial pressures of the N compounds
and the total conversion changed slightly, the formation of
biphenyl was enhanced. This enhancement resulted from the
larger amount of DBT available for DDS, because HYD
hardly occurred. This was proven by calculations in which
it was assumed that the rate constant of the DDS path-
way did not change in the presence of small amounts of
2-MPy and 2-MPiper and that the HYD pathway was com-
pletely blocked. At higher partial pressure of the N-contain-
ing molecules (6 kPa), the DDS pathway was more inhibited
and 2-MPy had a stronger inhibitory effect than 2-MPiper
(Fig. 10).
Scheme 4. Conformation structure of 4,6-dimethyltetrahydrodibenzothio-
phene.
presence of H2S. We also conclude from the data in Table 1
that the DDS of 4,6-DMDBT is 12 to 17 times slower than
that of DBT. However, the HYD pathway of 4,6-DMDBT is
two times faster than that of DBT. This must be due to the
positive influence of the two methyl groups on the hydro-
genation of an aromatic ring. Meille et al. observed a higher
reactivity in the HDS of 2,8-DMDBT than in the HDS of
DBT [17] and proposed an analogy with the hydrogenation
of aromatics: toluene is more easily hydrogenated than ben-
zene [29] and 3-methylbiphenyl more easily than biphenyl
over metal sulfide catalysts [30]. In agreement with this ex-
planation, we found that the hydrogenationof 3,3ꢀ-dimethyl-
biphenyl at 340 ◦C was two times faster than that of biphenyl
over the NiMo/Al2O3 catalyst.
Although the HYD pathway is faster in the HDS of
4,6-DMDBT, the removal of sulfur from 4,6-dimethyltetra-
hydrodibenzothiophene takes place much slower than from
tetrahydrodibenzothiophene, since the selectivity toward the
formation of this intermediate is much higher in the HDS of
4,6-DMDBT than in that of DBT (cf. Figs. 2b and 9b). This
means that the two methyl groups in the 4 and 6 positions
not only sterically hinder the sulfur removal in the DDS but
also in the HYD pathway. 4,6-DMDBT has a flat structure
like DBT. The methyl groups adjacent to the sulfur atom are
more spacious than the σ orbitals on the sulfur atom and
hinder the molecule to approach the catalyst surface via the
sulfur atom. Therefore, the adsorption of 4,6-DMDBT in the
σ mode is weaker than that of DBT. When two double bonds
of one benzene ring are hydrogenated, the hydrogenated part
of the molecule is not flat any more (Scheme 4). The methyl
group of the hydrogenated ring does not have to affect the
adsorption because it can be turned away from the catalyst
surface. However, the partially hydrogenated ring is puck-
In the HDS of 4,6-DMDBT the difference between the ef-
fect of 2-MPy and 2-MPiper (Fig. 3) is less pronounced than
in the case of DBT, as shown too by the rate constants of the
DDS and HYD pathways (Fig. 11). For the HYD pathway
the HDS of DBT and 4,6-DMDBT are similar: N-contain-
Fig. 10. Rate constants of the DDS and HYD pathways in the HDS of DBT in the presence of 2-MPy and 2-MPiper.

M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
285
Fig. 11. Rate constants of the DDS and HYD pathways in the HDS of 4,6-DMDBT in the presence of 2-MPy and 2-MPiper.
ing molecules have a strong inhibitory influence and the
effect of 2-MPiper is slightly stronger than that of 2-MPy
(cf. Figs. 10 and 11). The inhibition can be indicated by the
factor kHYD/kHꢀ YD, where kHYD is the rate constant of the
HYD pathway in a single HDS reaction and kHꢀ YD is the
HYD rate constant in a competitive experiment. The inhi-
bition factors for the HYD route in the presence of 2-MPy
and 2-MPiper were very similar in the HDS of DBT and
4,6-DMDBT. However, in the HDS of DBT the final prod-
uct of the HYD pathway, cyclohexylbenzene, was observed
at all partial pressures of N-containing molecules, whereas
in the HDS of 4,6-DMDBT, at 6 kPa 2-MPy and 2-MPiper,
the only product of the HYD route was 4,6-dimethyltetra-
hydrodibenzothiophene (Fig. 4b). Therefore, the real desul-
furization of 4,6-DMDBT in the presence of N-compounds
occurs predominantly via the DDS pathway. This indicates
again the greater difficulty in the removal of sulfur from 4,6-
dimethyltetrahydrodibenzothiophene than from tetrahydro-
dibenzothiophene. As indicated above, this difficulty may be
due to the weaker adsorption of the tetrahydro-intermediate
of 4,6-DMDBT than of DBT. Thus, this intermediate can
hardly compete with the N-containing molecules for the ac-
tive site. The DDS of 4,6-DMDBT was suppressed to a
greater extent by the N compounds than that of DBT (cf.
Figs. 10 and 11), which results from the weaker σ adsorp-
tion of 4,6-DMDBT than of DBT. Another difference in the
HDS of DBT and 4,6-DMDBT is that the product of the
DDS pathway is hydrogenated further in the case of DBT
and does not react in the case of 4,6-DMDBT, despite the
faster hydrogenation of 3,3ꢀ-dimethylbiphenyl as compared
to biphenyl.
strongly affected by 2-MPy, and DBT had an inhibitory ef-
fect on the hydrogenation of 2-MPy but not on the HDN of
2-MPiper. In the case of 4,6-DMDBT, however, 2-MPiper
has a stronger retarding effect than 2-MPy on both DDS and
HYD routes (Fig. 11). 4,6-DMDBT, in turn, had no influ-
ence on the C–N bond cleavage. DBT and 4,6-DMDBT can
adsorb in the π mode and thus hinder the hydrogenation of
2-MPy. In the σ mode, however, the S-containing molecules
adsorb apparently much weaker than 2-MPiper and therefore
do not hinder the transformation of 2-MPiper.
A special case is the influence of 2-MPiper on the DDS
of DBT [31]. Up to 1 kPa 2-MPiper promotes the DDS of
DBT, but does not promote the DDS of 4,6-DMDBT. This
enhancement was suggested to be the result of a structural
or electronic modification of the catalyst surface through the
adsorption of 2-MPiper in the σ mode. Because 2-MPiper
promotes the DDS of DBT at low concentrations, the over-
all HDS is affected only slightly [31]. The DDS of 4,6-
DMDBT is inhibited more strongly in the presence of 2 kPa
2-MPiper than that of DBT. This may be the result of the
small adsorption constant of 4,6-DMDBT in the σ mode
K2-MPiper ꢁ KDσBT ꢁ K4σ,6-DMDBT because of the hindrance
of the methyl groups. This is why the DDS of 4,6-DMDBT
is not improved at low partial pressures of 2-MPiper, in
contrast to the DDS of DBT. Because the adsorption of
2-MPiper is much stronger than that of the S-containing
molecules, the HDN of 2-MPiper is not affected by DBT
and 4,6-DMDBT.
The composition of the hydrocarbons in the feed may af-
fect the catalytic activity and selectivity under deep HDS
conditions. Aromatic hydrocarbons in fuels have a detri-
mental effect on the catalyst activity [32]. In a series of pa-
pers Kabe and Ishihara and co-workers discussed the influ-
ence of solvents on the activity and selectivity of a sulfided
CoMo/Al2O3 catalyst in the HDS of benzothiophene (BT)
and DBT [33–35]. They found that the catalytic activity de-
creased in the order toluene > decalin > n-pentadecane >
1-methylnaphthalene in the HDS of BT and in the order
n-heptane > xylene > decalin > tetralin in the HDS of DBT.
The activation energies were unaffected by the solvents, but
the heats of adsorption of BT and DBT on the CoMo/Al2O3
To better understand the nature of the inhibitory influence
we studied the effect of 4,6-DMDBT on the HDN reac-
tions. 4,6-DMDBT decreases slightly the hydrogenation of
2-MPy to 2-MPiper but has no influence on the C–N bond
scission, since the conversion of 2-MPiper did not change
in the presence of 4,6-DMDBT. The same results were ob-
tained in the simultaneous HDN of 2-MPy and the HDS of
DBT [16]. Previously, we had concluded that C–N and C–S
bond cleavage take place at different sites, since the DDS
of DBT and, correspondingly, the overall HDS was more

286
M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
Fig. 12. Rate constants of the DDS and HYD pathways in the HDS of DBT and 4,6-DMDBT in the presence of naphthalene.
catalyst depended on the solvent and correlated well with
the catalytic activity. Solvents with the largest retarding ef-
fects decreased the heats of adsorption of BT and DBT to a
greater extent. Furthermore, in the HDS of DBT the solvent
mainly affected the conversion to biphenyl, whereas it hardly
affected the formation of cyclohexylbenzene [35]. Accord-
ing to a Langmuir–Hinshelwood model, with reactant and
solvent molecules in the gas phase, the heat of adsorption
of a reactant should not be affected by the presence of an-
other molecule. However, if we assume that a solvent layer
is present on the catalyst surface, then there will be an in-
teraction between reactant and solvent molecules and the
interaction energy will influence the apparent heat of adsorp-
tion. Ishihara et al. worked at rather low temperatures of 150
to 250 ◦C, which favors the formation of a liquid layer on
the catalyst surface.
We found no difference between decane and toluene in
the HDS rates of DBT and 4,6-DMDBT (Fig. 6); the prod-
uct distribution was also unaffected. Toluene did not un-
dergo hydrogenation to methylcyclohexane when DBT or
4,6-DMDBT were present in the feed. This indicates that
the adsorption of toluene is much weaker than that of the
S-containing molecules. The results reported by Kabe et al.
[33] indicate that the effects of the solvent are more pro-
nounced below 300 ◦C. At 300 ◦C, there was hardly any dif-
ference between n-heptane and xylene, which are very simi-
lar to the solvents used in our study. Moreover, the effects of
the solvent were more pronounced at 0.1 than at 1 wt% DBT
in the feed [34]. Our experiments were performed at 340 ◦C
with 1.4 wt% DBT and 1.6 wt% 4,6-DMDBT. Therefore,
our results do not contradict the results of Ishihara et al. At
340 ◦C, toluene, as an aromatic molecule, does not compete
with the S compounds for the active sites, although there is
a large excess amount of it in the feed.
naphthalene (Figs. 7a and 7b). Because naphthalene did not
change the selectivities, the rate constants of the DDS and
HYD pathways are affected to the same extent (Fig. 12).
Farag et al. reported that a 100-fold excess of naphthalene
relative to 4,6-DMDBT resulted in only slight decrease in
the HYD selectivity [40]. They also noted, however, that the
HYD/DDS ratio varied with the conversion of 4,6-DMDBT.
These changes in the HYD/DDS ratios might be due to the
fact that their study was carried out in an autoclave. In that
case, the quantitative analysis can easily have an error of
10 to 15%. We believe that the HYD/DDS ratio stays con-
stant within a single experiment on the HDS of 4,6-DMDBT,
since the selectivity toward 3,3ꢀ-dimethylbiphenyl formation
is constant in the course of the reaction (Figs. 1b and 2b).
The HYD/DDS ratio is also unaffected by the presence of
naphthalene. In the case of the HDS of DBT, biphenyl is hy-
drogenated further to cyclohexylbenzene. The presence of
naphthalene does not change the product distribution in the
HDS of DBT (Fig. 12).
Under our reaction conditions naphthalene converts pre-
dominantly to tetralin; only small amounts of decalin were
observed, in good agreement with results obtained by Isoda
et al. [38,39]. The effect of S-containing molecules on the
hydrogenation reaction has been studied mainly in terms of
the influence of H2S [41–43]. We investigated the effect of
DBT and 4,6-DMDBT on the hydrogenation of naphthalene
at a reasonably high partial pressure of H2S in order to avoid
the influence of the H2S released during the HDS reaction.
Our data indicate that hydrogenation is inhibited in the pres-
ence of S compounds and that DBT and 4,6-DMDBT have
the same effect (Fig. 8). Both DDS and HYD pathways of the
HDS of DBT and 4,6-DMDBT are affected to the same ex-
tent in the presence of naphthalene, and the hydrogenationof
naphthalene is equally suppressed by DBT and 4,6-DMDBT.
Thus, we conclude that the hydrogenation of naphthalene
takes place at both the DDS and the HYD sites. Moreover,
we assume that the DDS and the HYD sites are the same and
that the only factor, which determines the HDS pathway, is
the adsorption conformation of the S-containing molecule.
Kogan et al. also suggested that the HYD pathway of the
HDS of thiophene takes place at the same catalytic sites as
the DDS pathway [44].
Only limited information is available about the inhibition
of the HDS of thiophene [5], DBT [36] and 4,6-DMDBT
[9,37–40] by polycyclic aromatic compounds. Most studies
indicate that the influence of aromatics on HDS is not neg-
ligible but that it is much weaker than that of basic organo-
nitrogen compounds. Our results show that the rate constant
of the HDS of DBT decreased by a factor of 1.34 and that
of 4,6-DMDBT by a factor of 1.8 in the presence of 10 kPa

M. Egorova, R. Prins / Journal of Catalysis 224 (2004) 278–287
287
5. Conclusion
[13] B.C. Gates, H. Topsøe, Polyhedron 16 (1997) 3213.
[14] R. Shafi, G.J. Hutchings, Catal. Today 59 (2000) 423.
[15] D.D. Whitehurst, T. Isoda, I. Mochida, Adv. Catal. 42 (1998) 345.
[16] M. Egorova, R. Prins, J. Catal. 221 (2004) 11.
[17] V. Meille, E. Schulz, M. Lemaire, M. Vrinat, J. Catal. 170 (1997) 29.
[18] F. Bataille, J.L. Lemberton, P. Michaud, G. Perot, M. Vrinat, M. Le-
maire, E. Schulz, M. Breysse, S. Kasztelan, J. Catal. 191 (2000) 409.
[19] C.N. Satterfield, M. Modell, J.F. Mayer, AIChE J. 21 (1975) 1100.
[20] C.N. Satterfield, M. Modell, J.A. Wilkens, Ind. Eng. Chem. Proc. Des.
Dev. 19 (1980) 154.
[21] M. Egorova, Y. Zhao, P. Kukula, R. Prins, J. Catal. 206 (2002) 263.
[22] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[23] V. Vanrysselberghe, G.F. Froment, Ind. Eng. Chem. Res. 35 (1996)
3311.
[24] X.L. Ma, K. Sakanishi, I. Mochida, Ind. Eng. Chem. Res. 35 (1996)
2487.
[25] T. Kabe, K. Akamatsu, A. Ishihara, S. Otsuki, M. Godo, Q. Zhang,
W.H. Qian, Ind. Eng. Chem. Res. 36 (1997) 5146.
[26] T.C. Ho, J. Catal. 219 (2003) 442.
[27] D.R. Kilanowski, H. Teeuwen, V.H.J. de Beer, B.C. Gates, G.C.A.
Schuit, H. Kwart, J. Catal. 55 (1978) 129.
[28] M. Houalla, D.H. Broderick, A.V. Sapre, N.K. Nag, V.H.J. de Beer,
B.C. Gates, H. Kwart, J. Catal. 61 (1980) 523.
[29] O. Weisser, S. Landa, Sulphide Catalysts, Pergamon, Oxford, 1973.
[30] V. Lamure-Meille, E. Schulz, M. Lemaire, M. Vrinat, Appl. Catal.
A 131 (1995) 143.
In studying the mutual influence of the HDS of 4,6-
DMDBT and HDN of 2-MPy and 2-MPiper, we found that
both N-containing molecules are strong inhibitors of HDS.
Moreover, the inhibitory effect of 2-MPiper was somewhat
stronger than that of 2-MPy for the DDS and HYD path-
ways of the HDS of 4,6-DMDBT. 4,6-DMDBT and DBT
suppressed the hydrogenationof 2-MPy but did not affect the
C–N bond cleavage in the HDN of 2-MPiper. Therefore, we
assume that adsorption of 2-MPiper on both DDS and HYD
sites is much stronger than that of 4,6-DMDBT or DBT.
When toluene is used as a solvent, neither the rates of the
HDS of DBT and 4,6-DMDBT nor the product distributions
are affected. Toluene itself does not undergo hydrogenation
in the presence of S or N compounds. Thus, the molecules of
the solvent do not compete with reactant for the active sites.
Naphthalene inhibited the DDS and HYD pathways in the
HDS of DBT and 4,6-DMDBT to the same extent. Thus, the
hydrogenation of naphthalene takes place at both the DDS
and the HYD sites. DBT and 4,6-DMDBT suppressed the
hydrogenationof naphthalene to the same extent. We assume
that the adsorption of naphthalene is much weaker than that
of S-containing molecules.
[31] M. Egorova, R. Prins, Catal. Lett. 92 (2004) 87.
[32] P.T. Vasudevan, J.L.G. Fierro, Catal. Rev.-Sci. Eng. 38 (1996) 161.
[33] T. Kabe, A. Ishihara, M. Nomura, T. Itoh, P.Y. Qi, Chem. Lett. (1991)
2233.
[34] A. Ishihara, T. Itoh, T. Hino, M. Nomura, P.Y. Qi, T. Kabe, J. Catal. 140
(1993) 184.
References
[35] A. Ishihara, T. Kabe, Ind. Eng. Chem. Res. 32 (1993) 753.
[36] M. Nagai, T. Kabe, J. Catal. 81 (1983) 440.
[37] X.L. Ma, K. Sakanishi, T. Isoda, I. Mochida, Ind. Eng. Chem. Res. 34
(1995) 748.
[38] T. Isoda, S. Nagao, X.L. Ma, Y. Korai, I. Mochida, Energy Fuels 10
(1996) 482.
[39] T. Isoda, S. Nagao, X.L. Ma, Y. Korai, I. Mochida, Energy Fuels 10
(1996) 487.
[40] H. Farag, K. Sakanishi, I. Mochida, D.D. Whitehurst, Energy Fuels 13
(1999) 449.
[41] S. Gultekin, S.A. Ali, C.N. Satterfield, Ind. Eng. Chem. Proc. Des.
Dev. 23 (1984) 179.
[42] D.H. Broderick, A.V. Sapre, B.C. Gates, H. Kwart, G.C.A. Schuit,
J. Catal. 73 (1982) 45.
[43] A.V. Sapre, B.C. Gates, Ind. Eng. Chem. Proc. Des. Dev. 21 (1982) 86.
[44] V.M. Kogan, R.G. Gaziev, S.W. Lee, N.N. Rozhdestvenskaya, Appl.
Catal. A 251 (2003) 187.
[1] S. Mayo, E. Brevoord, L. Gerritsen, F. Plantenga, Hydrocarb.
Process. 80 (2001) 84.
[2] B.H. Cooper, B.B.L. Donnis, Appl. Catal. A 137 (1996) 203.
[3] C. Song, X.L. Ma, Appl. Catal. B 41 (2003) 207.
[4] M. Nagai, T. Sato, A. Aiba, J. Catal. 97 (1986) 52.
[5] V. La Vopa, C.N. Satterfield, J. Catal. 110 (1988) 375.
[6] S. Chand, Res. Ind. 39 (1994) 194.
[7] G.C. Laredo, J.A. De los Reyes, J.L. Cano, J.J. Castillo, Appl. Catal.
A 207 (2001) 103.
[8] G.C. Laredo, E. Altamirano, J.A. De los Reyes, Appl. Catal. A 243
(2003) 207.
[9] T. Koltai, M. Macaud, A. Guevara, E. Schulz, M. Lemaire, R. Bacaud,
M. Vrinat, Appl. Catal. A 231 (2002) 253.
[10] C. Kwak, J.J. Lee, J.S. Bae, S.H. Moon, Appl. Catal. B 35 (2001) 59.
[11] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick, B.C. Gates,
AIChE J. 24 (1978) 1015.
[12] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.