Mutual influence of the HDS of dibenzothiophene and HDN of 2-methylpyridine

Article abstract of DOI:10.1016/S0021-9517(03)00264-1

The influence of 2-methylpyridine and 2-methylpiperidine on the hydrodesulfurization of dibenzothiophene (DBT) and the effect of DBT on the hydrodenitrogenation of 2-methylpyridine and 2-methylpiperidine were studied over a sulfided NiMo/Al₂O₃ catalyst at 5 MPa, 35 kPa H₂S, and 300 and 340°C. Both N-containing molecules strongly suppressed the hydrogenation pathway of the hydrodesulfurization of DBT and inhibited the direct desulfurization route at both reaction temperatures. The inhibitory effect on the direct desulfurization was stronger for 2-methylpyridine than for 2-methylpiperidine. H₂S promoted the hydrogenation of 2-methylpyridine up to 10 kPa and inhibited it at higher partial pressures. H₂S had a positive influence on the hydrodenitrogenation conversions of 2-methylpiperidine and 2-methylpyridine. DBT had a negative effect on the hydrogenation of 2-methylpyridine, but did not influence the C-N bond cleavage of 2-methylpiperidine. Therefore, C-N and C-S bond breaking takes place at different active sites, whereas the hydrogenation sites for N- and S-containing molecules may be the same.

Full text of DOI:10.1016/S0021-9517(03)00264-1

Journal of Catalysis 221 (2004) 11–19
www.elsevier.com/locate/jcat
Mutual influence of the HDS of dibenzothiophene and HDN
of 2-methylpyridine
Marina Egorova and Roel Prins ∗
Institute for Chemical and Bioengineering, Swiss Federal Institute of Technology (ETH), 8093 Zurich, Switzerland
Received 16 April 2003; revised 6 June 2003; accepted 6 June 2003
Abstract
The influence of 2-methylpyridine and 2-methylpiperidine on the hydrodesulfurization of dibenzothiophene (DBT) and the effect of DBT
on the hydrodenitrogenation of 2-methylpyridine and 2-methylpiperidine were studied over a sulfided NiMo/Al O catalyst at 5 MPa, 35 kPa
2
3
◦
H S, and 300 and 340 C. Both N-containing molecules strongly suppressed the hydrogenation pathway of the hydrodesulfurization of DBT
2
and inhibited the direct desulfurization route at both reaction temperatures. The inhibitory effect on the direct desulfurization was stronger for
2-methylpyridine than for 2-methylpiperidine. H S promoted the hydrogenation of 2-methylpyridine up to 10 kPa and inhibited it at higher
2
partial pressures. H S had a positive influence on the hydrodenitrogenation conversions of 2-methylpiperidine and 2-methylpyridine. DBT
2
had a negative effect on the hydrogenation of 2-methylpyridine, but did not influence the C–N bond cleavage of 2-methylpiperidine. There-
fore, C–N and C–S bond breaking takes place at different active sites, whereas the hydrogenation sites for N- and S-containing molecules
may be the same.
2003 Elsevier Inc. All rights reserved.
Keywords: Hydrotreating; Hydrodesulfurization; HDS; Dibenzothiophene; DBT; Hydrodenitrogenation; HDN; 2-Methylpyridine; 2-Methylpiperidine
1. Introduction
fluence of S- and N-containing molecules during hydrotreat-
ing showed that N-compounds have an inhibitory effect on
HDS reactions, whereas the presence of S-containing mole-
cules in some cases promotes HDN [10,11]. In the simu-
lation of the kinetics, it is usually assumed that one active
site is present on which HDS and HDN reactions take place
[7,12], although the presence of two different sites (desul-
furization and hydrogenation) was already reported in 1964
for the HDS of thiophene [13].
A systematic study of the interaction between catalytic
hydrodesulfurization and hydrodenitrogenation was made
by Satterfield et al. [2]. They studied the mutual influences of
the HDS of thiophene and HDN of pyridine in the presence
of sulfided CoMo/Al2O3, NiMo/Al2O3, NiW/Al2O3, and
NiW/SiO2–Al2O3 catalysts. Over all four catalysts pyridine
inhibited the HDS reaction, whereas thiophene had a dual
effect on HDN. At low temperatures thiophene inhibited the
HDN reaction by competing with pyridine for hydrogena-
tion sites and at high temperatures the dominant effect was
an improvement of C–N bond cleavage by H2S, the prod-
uct of the HDS reaction. The enhancement of the HDN by
H2S was ascribed to the conservation of the catalyst in a
completely sulfided state, in which it has a better HDN ac-
tivity. The HDS sites were classified into two types: (i) active
New environmentallegislation has put the oil industry un-
der increased pressure to limit the level of sulfur in gasoline
and diesel fuel with a view to reducing exhaust emissions.
A sulfur specification of 350 ppm is currently practiced in
the EEC and a sulfur content as low as 50 or even 10 ppm
is foreseen for the year 2005. Basic nitrogen compounds in-
hibit the hydrodesulfurization (HDS) of sulfur compounds
through competitive adsorption, but in normal HDS this in-
fluence is negligible because the amount of N-containing
molecules in gasoil is much smaller than that of S-containing
molecules. Nitrogen compounds will be harmful in deep
HDS, however, because then the amounts of N- and S-
containing molecules are comparable.
The significance of competitive adsorption in hydrotreat-
ing reactions has been recognized for a long time [1]
and several research groups investigated the effect of N-
compounds on HDS [2–8] and of S-compounds on hydro-
denitrogenation (HDN) [2,3,9,10]. Studies of the mutual in-
*
Corresponding author.
E-mail address: prins@tech.chem.ethz.ch (R. Prins).
0021-9517/$ – see front matter
2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0021-9517(03)00264-1

12
M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
sites that are responsible for the majority of the HDS activity,
but that are extremely sensitive to basic nitrogen compounds,
and (ii) active sites that are less active in HDS, but also less
susceptible to poisoning. Furthermore it was shown [3] that
the first type of sites was responsible for the hydrogenation
of S-containing molecules, because tetrahydrothiophenewas
not observed when pyridine was present in the feed.
fided in situ with a mixture of 10% H2S in H2 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 re-
action temperature, and the liquid reactant was fed to the
reactor by means of a Gilson 307 piston pump. The experi-
ments were carried out at 300 and 340 ◦C. The composition
of the gas-phase feed consisted of 130 kPa toluene (as sol-
vent for the DBT and amine), 8 kPa dodecane (as reference
for DBT and its derivatives in the GC analysis), 11 kPa hep-
tane (as reference for the N-compounds in the GC analysis),
1 kPa dibenzothiophene, 2–10 kPa amine reactant, 35 kPa
H2S (unless noted otherwise), and 4.8 MPa H2. Toluene was
chosen as the solvent due to its good ability to dissolve aro-
matic heterocycles like dibenzothiophene. We checked that
in the presence of S- or N-containing molecules toluene did
not undergo hydrogenation to methylcyclohexane. The par-
tial pressure of H2S was kept sufficiently higher than that of
DBT to avoid the influence of H2S formed during the HDS
reaction.
The reaction products were analyzed by on- and off-
line gas chromatography with a Varian 3800 GC instrument
equipped with a CP-Sil 8 CB fused silica capillary column
(50 m × 0.25 mm × 0.25 µm). Off-line analysis was used
for the HDS of DBT because of the high boiling points
of the reactant and reaction products. Detection was per-
formed with a flame ionization detector as well as with a
pulsed flame photometric detector, which is very useful for
detecting small amounts of S- and N-containing compounds.
Weight time was defined as τ = wcat/nfeed, where wcat de-
notes the catalyst 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 keep-
ing their ratio constant. The reaction was stable after 3 to 4 h,
and during 2 weeks of operation almost no catalyst deactiva-
tion was observed.
Two reaction pathways were found for the HDS of diben-
zothiophene (DBT), which take place on different sites. The
first type of sites is responsible for the hydrogenation path-
way, and the second type of sites is facilitating the direct
sulfur removal or hydrogenolysis [4–6,14–18]. N-containing
molecules were found to be strong inhibitors for the mi-
nor, hydrogenation pathway of the HDS of DBT. The overall
conversion of DBT decreased only slightly, suggesting that
the amount of product formed via the hydrogenolysis path-
way increased. Some authors have even claimed an enhance-
ment of this route in the presence of N-compounds [4–6].
We have studied the mutual influences of the HDS of
DBT and the HDN of 2-methylpyridine and 2-methylpipe-
ridine over a sulfided NiMo/Al2O3 catalyst. DBT was cho-
sen because it allows the study of the removal of sulfur by the
direct desulfurization pathway (hydrogenolysis) as well as
by the hydrogenation pathway (hydrogenation followed by
desulfurization). Pyridine is the smallest model molecule for
studying HDN, but, although the reactions taking place in its
HDN are now well understood, the study of its HDN kinetics
proved to be extremely difficult. The reason for this difficulty
is that two molecules of piperidine, the first intermediate in
the HDN of pyridine, disproportionate to N-pentylpiperidine
and ammonia [19]. However, the introduction of a methyl
group on the α carbon atom of piperidine strongly sup-
presses the disproportionation, so that it hardly interferes
with the other reactions taking place during the HDN of 2-
methylpyridine and 2-methylpiperidine [20]. Therefore, we
decided to use 2-methylpyridine and 2-methylpiperidine in
our study of the influence of N-containing molecules on the
HDS of DBT.
3. Results
3.1. HDS of dibenzothiophene
2. Experimental
The results of the experiments of the HDS of DBT carried
out at 300 and 340 ◦C (Figs. 1 and 2) show that the reaction
goes via two parallel pathways (Scheme 1): direct desul-
furization (DDS), which yields biphenyl (BP), and hydro-
genation (HYD) followed by desulfurization, which gives
first tetrahydrodibenzothiophene (TH-DBT) and then cyclo-
hexylbenzene (CHB). The two final products are biphenyl
and cyclohexylbenzene. At 300 ◦C the selectivity of the
biphenyl formation is about 85% at low and 80% at high
weight time (Fig. 1b), whereas at the higher temperature of
340 ◦C those values are 90 and 70%, respectively (Fig. 2b).
These results indicate that the DDS route is much easier than
the HYD route, since the amount of biphenyl is six to nine
times higher than the sum of tetrahydrodibenzothiophene
and cyclohexylbenzene. Slow hydrogenation of biphenyl to
The NiMo/γ -Al2O3 catalyst used in this work contained
8 wt% Mo and 3 wt% Ni and was prepared by successive
incipient wetness impregnation of γ -Al2O3 (Condea, pore
volume 0.5 cm3 g−1, specific surface area 230 m2 g−1) with
an aqueous solution of (NH4)6Mo7O24·4H2O (Aldrich), fol-
lowed by an aqueous solution of Ni(NO3)2·6H2O (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.
Reactions were carried out in a continuous mode in a
fixed-bed inconel reactor (15 mm i.d., 320 mm long) that
was heated by an oven. A sample of catalyst (0.05 g) was
diluted with 8 g SiC to achieve plug-flow conditions in
the continuous-flow fixed-bed reactor. The catalyst was sul-

M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
13
◦
Fig. 1. Relative partial pressures (a) and selectivities (b) of the products of the HDS of dibenzothiophene 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 dibenzothiophene at 340 C as a function of weight time.
consequence, the hydrogenation of biphenyl is not only in-
trinsically faster but also less inhibited by DBT at 340 ◦C.
The HDS of DBT is generally assumed to be a first-order
reaction with respect to DBT [11]. Our DBT conversions at
300 ◦C confirm this, since we obtained a linear dependency
when plotting ln[(CDBT)/(CDBT)0] versus weight time and
a high R2 value. Also in the presence of the N-containing
molecules the HDS of DBT could be approximated with
first-order kinetics. Therefore, all the rate constants of the
HDS of DBT with and without N-compounds at 300 ◦C were
evaluated assuming the HDS of DBT to be a first-order re-
action.
Scheme 1. Reaction network of the HDS of dibenzothiophene.
cyclohexylbenzene took place at both reaction temperatures,
since the biphenyl selectivity decreased with weight time
and the increase of the cyclohexylbenzene selectivity with
weight time was higher than the decrease of the tetrahy-
drodibenzothiopheneselectivity. The resulting overall mech-
anism of the HDS of DBT is represented in Scheme 1.
The DDS pathway is slightly enhanced at higher temper-
atures, since the selectivity toward biphenyl formation at low
weight time is higher at 340 than at 300 ◦C. Thus, an increase
of the temperature has a positive effect on the desulfurization
reaction. At high weight time, the biphenyl selectivity was
lower at 340 than at 300 ◦C, due to increased hydrogenation
of biphenyl. At 340 ◦C and τ = 5.5 g min/mol the DBT con-
version was about 90%, but at 300 ◦C it was only 45%. As a
3.2. Inhibition of the HDS by N-containing molecules
The HDS of DBT was studied in the presence of 2-
methylpyridine (2-MPy) and 2-methylpiperidine (2-MPiper)
at 300 and 340 ◦C at different conversion levels of the N-
compounds. The HDN network of 2-MPy and 2-MPiper
was determined earlier [20] and is presented in Scheme 2.
As expected, 2-MPiper is the only primary product in the
HDN of 2-MPy, since the HDN of heterocyclic N-containing
aromatic molecules can only occur after ring hydrogena-
tion [21]. The HDN of 2-MPiper showed that four com-
pounds, 1-aminohexane, 2-aminohexane, 2-methylpyridine,
and 2-methyl-3,4,5,6-tetrahydropyridine, have nonzero se-
lectivity at zero conversion of 2-MPiper and thus might be

14
M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
Scheme 2. Reaction network of the HDN of 2-methylpyridine and 2-meth-
ylpiperidine.
considered to be primary products. Actually, aminoalkenes
and aminoalkanethiols are expected to be primary products.
Aminoalkenes, which are the primary products of the ring
opening via elimination of alkylpiperidine, were not ob-
served, probably because of their fast hydrogenation to the
corresponding amines. Aminoalkanethiols can be formed
by nucleophilic substitution of the alkylpiperidine by H2S.
They are probably not observed because of a fast H2S elim-
ination or C–S hydrogenolysis. Comparison of the con-
version of 2-aminohexane and 1-aminohexane showed that
the reactivity of 2-aminohexane is higher than that of 1-
aminohexane. Despite this higher reactivity, much more 2-
aminohexane than 1-aminohexane was detected in the HDN
of 2-MPiper [20]. Therefore, it can be concluded that the
first C–N bond breaking in 2-MPiper occurs predominantly
between the nitrogen atom and the carbon atom of the
methylene group and that the methyl group actually has a
negative rather than a positive influence on the C–N bond
breaking. This suggests that the HDN of 2-MPiper occurs by
nucleophilic substitution to 5-aminohexanethiol rather than
via elimination to 5-aminohexene-1.
Fig. 3. Inhibition of the HDS of 1 kPa DBT in the presence of 2-methylpyri-
dine (") and 2-methylpiperidine (F) at 300 C.
◦
and 2-MPiper not only suppress the hydrogenationroute, but
inhibit the hydrogenolysis (DDS) pathway as well. Since the
transformation of DBT is still well described by first-order
kinetics, we calculated kDBT (the rate constant of the HDS of
DBT in the presence of the N-compound) and plotted the ra-
tio kDBT/kDBT, where kDBT is the rate constant of the HDS
of DBT alone, versus the partial pressure of the N-containing
molecule in the gas phase (Fig. 3). One can see that the
inhibition effect of 2-MPy is much stronger than that of 2-
MPiper. While no conversion of 2-MPiper was observed in
the competitive experiments, the conversion of 2-MPy var-
ied from 50 to 20% when increasing its partial pressure in
the feed.
At 340 ◦C, DBT could be fully converted and 2-MPy and
2-MPiper underwent hydrodenitrogenation in the simultane-
ous HDS and HDN reactions. Different partial pressures of
2-MPy and 2-MPiper (2, 6, and 10 kPa) were used. Again,
just trace amounts of products formed via the HYD pathway
of the HDS of DBT were observed; this route is strongly
suppressed at 340 ◦C as well. The inhibition of the direct
desulfurization pathway is clearly seen at high partial pres-
sures of 2-MPy and 2-MPiper (Fig. 4). The decrease of the
DBT conversion is almost the same in the presence of 10
kPa of both N-compounds. However, at 2 and 6 kPa the
inhibitory influence of 2-MPy is stronger, as was observed
at 300 ◦C. For the lowest concentration of 2-MPy and 2-
MPiper studied (2 kPa) one can hardly see any influence on
We first studied the HDS of 1 kPa DBT at 300 ◦C in the
presence of 2, 6, and 10 kPa of 2-MPy and 2-MPiper. At this
temperature 2-MPy only converts to 2-MPiper and, because
of thermodynamics [22,23], this reaction is irreversible. In
the presence of the N-containing molecules the HYD path-
way of the HDS of DBT is strongly suppressed, since the
selectivity toward biphenyl formation is increased from 85 to
96–98%. The conversion of DBT decreased with increasing
partial pressure of the N-compounds, indicating that 2-MPy
◦
Fig. 4. Inhibition of the HDS of DBT in the presence of 2-methylpyridine (a) and 2-methylpiperidine (b) at 340 C.

M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
15
◦
Fig. 5. Relative partial pressures of biphenyl in the HDS of DBT in the presence of 2-methylpyridine (a) and 2-methylpiperidine (b) at 340 C.
the DBT conversion and in these experiments we observed
an enhancement of the biphenyl formation (Fig. 5). This can
be explained by the blocking of the HYD pathway, so that
a higher amount of DBT is available for DDS. We verified
this by kinetic calculations, assuming the DDS and HYD
reactions of the DBT HDS to be first-order irreversible re-
actions running in parallel. The rate constants were obtained
using the selectivity data (90% for DDS and 10% for HYD).
Then excluding the HYD pathway from the HDS network
of DBT we obtained 72% conversion toward biphenyl at
τ = 3.5 g min/mol. In Fig. 5 the theoretical conversion of
DBT to biphenyl in the absence of the HYD pathway is
presented by a dashed line. Therefore, we conclude that the
◦
Fig. 6. HDN of 2-methylpiperidine at 340 C and different H S partial pres-
2
sures.
enhancement of the biphenyl formation is due to the absence
of the HYD route. At low concentrations of 2-MPy and 2-
MPiper, the inhibition of the DDS pathway is minor and the
DDS pathway profits fully from the blocking of the HYD
pathway. At higher concentrations of 2-MPy and 2-MPiper,
the DDS pathway is slowed down as well and the biphenyl
curves lie under the curve obtained in the absence of the N-
containing molecules (Fig. 5).
3.3. Effect of H2S on the HDN of 2-methylpyridine
Before studying the influence of DBT on the HDN of 2-
MPy, we investigated the effect of H2S on the two reactions
that take place in the HDN of 2-MPy: the hydrogenation of
2-MPy and the removal of nitrogen from 2-MPiper. Exper-
iments on the hydrogenation of 2-MPy were performed at
300 ◦C, because at this temperature only hydrogenation to
2-MPiper occurred; 2-MPiper, in turn, did not react further.
The removal of nitrogen from 2-MPy was studied at 340 ◦C.
The cleavage of the C(sp3)–N bond is known to be pro-
moted by H2S [24–26]. In agreement with these results, our
experiments of the HDN of 2-MPiper showed an enhance-
ment of the HDN conversion with increasing H2S partial
pressure from 30 to 160 kPa (Fig. 6).
◦
Fig. 7. Conversion of 2-methylpyridine at 300 C and different H S partial
2
pressures and a weight time of 3 g min/mol.
hydrogenation of N-containing aromatic molecules by H2S
has been reported [27,28]. The occurrence of an optimum
H2S concentration may be caused by the fact that at high
concentrations H2S adsorbs on the active sites and poisons
the hydrogenation reaction, whereas at very low H2S con-
centrations the catalyst surface reconstructs and the number
of hydrogenation sites decreases.
The hydrogenation of 2-MPy at 300 ◦C starts quite fast
and levels off with weight time (Fig. 8). To determine if
poisoning by the product was responsible for this leveling
off, the reaction was repeated in the presence of 1 kPa 2-
MPiper at a ratio of 2-MPy/2-MPiper = 6. The results show
The conversion of 2-MPy at a fixed weight time of
3 g min/mol and a temperature of 300 ◦C is presented in
Fig. 7 as a function of the H2S pressure. The 2-MPy con-
version increased up to 10 kPa H2S and decreased at higher
H2S partial pressures. In the literature only inhibition of the

16
M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
◦
◦
Fig. 10. HDN of 6 kPa 2-methylpyridine at 340 C in the presence (- - -) and
absence (—) of 1 kPa DBT.
Fig. 8. Hydrogenation of 6 kPa 2-methylpyridine at 300 C in the presence
(- - -) and absence (—) of 1 kPa 2-methylpiperidine.
◦
Fig. 9. Total conversion and product distribution in the HDN of
2-methylpyridine at 340 C and different H S partial pressures.
2
Fig. 11. HDN of 6 kPa 2-methylpiperidine at 340 C in the presence (- - -)
and absence (—) of 1 kPa DBT.
◦
that such a small amount of 2-MPiper indeed strongly in-
hibits the hydrogenation reaction (Fig. 8). It indicates that
2-MPiper strongly adsorbs on the hydrogenation sites.
Studying the effect of H2S on the HDN of 2-MPy at
340 ◦C, we observed a decrease of 2-MPy conversion with
an increase of H2S partial pressure from 0 to 10 kPa. The
loss in the conversion was mainly due to the decrease of 2-
MPiper formation, whereas the HDN conversion of 2-MPy
to nitrogen-free C6 products was enhanced almost by a fac-
tor of four (Fig. 9). Successive increases of the H2S partial
pressure up to 100 kPa (30, 70, and 100) led to a further de-
crease of the 2-MPy conversion and 2-MPiper formation and
to the increase of the HDN conversion of 2-MPy. However,
these changes were not that significant as for the increase in
H2S partial pressure from 0 to 10 kPa (Fig. 9). Therefore,
we can conclude that at 340 ◦C H2S has a negative effect
on the hydrogenation of 2-MPy and a positive influence on
the HDN conversion of 2-MPy as well as 2-MPiper. These
results are in good agreement with those reported in [24,25].
DBT has a negative effect on the hydrogenation of 2-MPy
at 300 ◦C. The conversion of 6 kPa 2-MPy to 2-MPiper de-
creased almost by a factor of two in the presence of 1 kPa
DBT (not shown). This cannot be caused by the H2S released
in the HDS reaction since the partial pressure of H2S in the
feed was 35 kPa, 35 times higher than that of DBT. The ob-
served inhibition effect must therefore be due to the presence
of DBT.
The influence of DBT on the HDN of the N-containing
molecules was also studied at 340 ◦C, when C–N bond
cleavage took place as well. The results of the HDN of 2-
MPy, carried out in parallel with the HDS of DBT, and of the
HDN of 2-MPy alone are presented in Fig. 10. In the pres-
ence of DBT, the conversion of 2-MPy and the yield of 2-
MPiper decreased, while the amount of C6 products (hexane,
hexenes, and hexylamines) hardly changed. The conversion
of 2-MPy was lower at 340 than at 300 ◦C, since the hy-
drogenation of 2-MPy to 2-MPiper becomes reversible. The
inhibitory influence of DBT on the hydrogenation of 2-MPy
was less strong at 340 than at 300 ◦C. The influence of DBT
on the C–N bond cleavage of 2-MPiper to hexane, hexenes,
and hexylamines [20] was studied at 340 ◦C. Fig. 11 shows
that DBT does not have any influence on the HDN of 2-
MPiper.
3.4. Influence of DBT on the HDN of N-compounds
The N-containing molecules had a negative influence on
both pathways of the HDS of DBT. They almost blocked
the HYD pathway and inhibited the DDS pathway. In turn,

M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
17
4. Discussion
H2S inhibits the DDS pathway of the HDS of DBT
strongly, but the HYD pathway only slightly [34]. This sug-
gests that the hydrogenolysis and hydrogenation of DBT
proceed on different catalytic sites. The sites active for the
hydrogenation of DBT are very sensitive to poisoning by ni-
trogen bases, whereas the sites responsible for direct C–S
bond cleavage are less susceptible to poisoning [4,5]. It was
even reported that acridine has a promotion effect on the
direct desulfurization of DBT over sulfided NiMo- and NiW-
supported catalysts [35,36]. However, it has never been con-
firmed by systematic studies.
As originally proposed by Houalla et al. [29], the reaction
pathways for the HDS of DBT are (i) direct desulfurization
to form biphenyl and (ii) hydrogenation to give 1,2,3,4-
tetrahydrodibenzothiophene or 1,2,3,4,10,11-hexahydrodi-
benzothiophene, which are further desulfurized to cyclo-
hexylbenzene. Our results show that over NiMo the desul-
furization pathway is six times faster than the hydrogenation
pathway at 300 ◦C and nine times faster at 340 ◦C (Figs. 1b
and 2b). Desulfurization of the hydrogenated intermediate
tetrahydrodibenzothiophene is also much faster than hydro-
genation, since only amounts of tetrahydrodibenzothiophene
as small as 1–2% were observed during the HDS reaction.
The biphenyl formed in the DDS pathway was slowly hy-
drogenated to cyclohexylbenzene at 300 ◦C and somewhat
faster at 340 ◦C (Figs. 1b and 2b). The hydrogenation of
biphenyl to cyclohexylbenzene during the HDS of DBT has
also been observed in other studies [15,16,18]. It is more evi-
dent over NiMo catalysts than over CoMo catalysts [30], be-
cause NiMo is a better hydrogenation catalyst in comparison
with CoMo. The higher activity of a nickel-containing cata-
lyst has been ascribed to the structure of the catalyst. NiMo
catalysts are more sensitive to the inhibiting effect of H2S
than CoMo catalysts in HDS reactions [31–33]. This indi-
cates that the state of sulfidation of a NiMo catalyst changes
more strongly with the H2S partial pressure than that of a
CoMo catalyst. At lower H2S partial pressure, the NiMo ca-
talyst may be more easily depleted of sulfur and thus have
a better hydrogenation activity. No further hydrogenation of
cyclohexylbenzene to bicyclohexyl was observed under our
reaction conditions. Therefore, Scheme 1 describes our re-
sults of the HDS of DBT best.
The hydrogenation of biphenyl, in the absence of S- or
N-containing molecules, was about 10 times slower over our
NiMo/Al2O3 catalyst than the hydrogenation of the biphenyl
formed during the HDS reaction, however. Similarly low hy-
drogenation rates of biphenyl were obtained by Nagai et
al. [5]. In their study of the HDS of DBT, they changed
a feed, that contained 5 wt% DBT, to a feed containing
5 wt% DBT and 1 wt% biphenyl, and later switched back
to the initial feed. No noticeable changes in the concentra-
tion of the cyclohexylbenzene formed were observed after
the changes in the feed composition. We ascribe the differ-
ence between the hydrogenation of added biphenyl and of
biphenyl formed in situ to hidden kinetics. When biphenyl
is formed in situ, by the DDS of DBT, it is still adsorbed
on the catalyst surface. In this flat, adsorbed state it can di-
rectly be hydrogenated, before desorbing from the catalytic
site. When added to the gas phase, biphenyl has to diffuse
to the catalytic sites and adsorb. In the gas phase, the two
phenyl rings of biphenyl are not coplanar and some activa-
tion energy has must brought up to adsorb biphenyl in a flat
conformation. This may explain the difference in the rate of
hydrogenation between added and in situ formed biphenyl.
In the competitive HDS and HDN experiments at 300 ◦C,
with feeds containing 1 kPa DBT and different partial pres-
sures (2, 6, and 10 kPa) of 2-MPy or 2-MPiper, a decrease
of the total conversion of DBT was observed in all reac-
tions. The inhibition was stronger in the presence of 2-MPy
than in 2-MPiper (Fig. 3). The HDS rate constant (kDBT
)
changed slightly in the presence of 2 kPa 2-MPiper, whereas
in the presence of 2 kPa 2-MPy it decreased by 50%. The
products of the HYD pathway of the HDS of DBT, such as
tetrahydrodibenzothiophene and cyclohexylbenzene, were
observed in very small amounts (2–4% of the total conver-
sion). This shows that even small amounts of N-compounds
significantly inhibit the hydrogenation of DBT and means
that the slight inhibition of the HDS of DBT at 2 kPa of 2-
MPiper is mainly due to the small contribution of the HYD
route. At higher concentrations of 2-MPiper the inhibition of
the DDS route becomes obvious. In the presence of 2-MPy,
the DDS pathway is inhibited already at low 2-MPy partial
pressures.
At the higher temperature of 340 ◦C, the DDS pathway
was enhanced (Fig. 2b) and less sensitive to the low con-
centrations of N-compounds (Fig. 4). An amount of 2 kPa
2-MPy decreased the DBT conversion slightly, while 2 kPa
2-MPiper hardly changed the DBT conversion. The forma-
tion of biphenyl was enhanced (Fig. 5). Similar increases
in one product, induced by inhibition of another product,
were reported by others [4,5]. However, this increase in the
concentration of biphenyl is not the result of a promoting ef-
fect of the N-containing molecules on the HDS of DBT, but
the result of an increase of the DBT concentration, due to
the absence of the HYD pathway. As a result, the ultimate
yield of biphenyl can be 100 instead of 90% (dashed line in
Fig. 5). Moreover, in the presence of N-containing molecules
biphenyl is not further hydrogenated to cyclohexylbenzene.
At higher 2-MPy and 2-MPiper concentrations, the inhi-
bition of DDS became noticeable. Like at 300 ◦C, also at
340 ◦C the inhibitory effect of 2-MPy is larger than that of 2-
MPiper. Hence, the N-containing molecules have a negative
influence on both HYD and DDS pathways at both reaction
temperatures, but, whereas the HYD route is strongly inhib-
ited, the DDS route is less affected.
The less basic 2-MPy was found to be a stronger in-
hibitor than the more basic 2-MPiper for the DDS of DBT.
This is in agreement with report that pyridine is a stronger
poison than piperidine in the HDS of DBT [5]. Although

18
M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
the aqueous basicity (pKa) of aliphatic heterocyclic com-
pounds is almost two times higher than that of heterocyclic
aromatic compounds [5,37], their adsorption constants are
lower [5,37]. Apparently, the aqueous basicity is a poor pre-
dictor of adsorption strength. A more appropriate parameter
is the gas-phase proton affinity, which has a good correlation
with the adsorption constant [7]. Indeed, the proton affinity
and the adsorption constant of piperidine on sulfided NiMo
catalyst are lower than those of pyridine [5]. This suggests
that the stronger inhibitory effect of 2-MPy than of 2-MPiper
is due to a higher adsorption constant on the catalyst surface.
All observations are basically related to the major DDS path-
way of the HDS of DBT. The picture for the HYD site can
be different, because of the different nature of the site. In-
deed, our results show that the more basic 2-MPiper inhibits
the HYD pathway somewhat stronger than 2-MPy. This is
in accordance with results reported by Zeuthen et al. [38],
who showed that basic N-compounds influenced the HDS of
diesel strongly. Since most sulfur compounds in their diesel
were of the 4,6-dialkylDBT-type, these results are in good
agreement with ours.
(Fig. 6). Therefore, the effect of H2S on the overall con-
version of 2-MPy could be promoting at lower and retarding
at higher H2S partial pressures. However, our experiments
showed that H2S has a small negative influence on the over-
all conversion of 2-MPy at 340 ◦C but a positive effect on the
HDN conversion (Fig. 9). At first sight, it looks as if these re-
sults contradict the promotion effect of low amounts of H2S
on both hydrogenation and C–N bond cleavage. However,
at the higher temperature of 340 ◦C one more reaction takes
part in the HDN of 2-MPy and that is the reverse reaction
of dehydrogenation of 2-MPiper to 2-MPy. This reaction
can be also promoted by a small amount of H2S, just as the
hydrogenation reaction, therefore, the overall conversion of
2-MPy would decrease with increase of H2S partial pressure.
These results are in good agreement with those evidencing
the dual effect of H2S on HDN [25,39,40].
A detailed study of the interaction between catalytic hy-
drodesulfurization and hydrodenitrogenation was first per-
formed by Satterfield et al. [2,3] with thiophene and pyridine
as model compounds. Thiophene had a dual effect on the
HDN of pyridine. At low temperatures, the competitive ad-
sorption of thiophene on hydrogenation sites retarded the
hydrogenation of pyridine to piperidine, and thus reduced
the overall reaction rate. At higher temperatures, H2S pro-
duced in the HDS of thiophene increased the rate of piperi-
dine hydrogenolysis and enhanced the overall HDN reac-
tion rate. Since these results were obtained in the absence
of hydrogen sulfide, it is not clear if the inhibitory effect
of the S-containing compound was due to H2S released in
the HDS reaction or due to the S-containing molecule itself.
Therefore, we performed our experiments at an H2S partial
pressure, which was 35 times higher than that of DBT. The
results show that DBT adsorbs much stronger on the hydro-
genation sites than H2S, since the conversion of 2-MPy to
2-MPiper in the presence of DBT decreased by a factor of
two at 300 ◦C. At the higher temperature of 340 ◦C, the in-
hibitory effect of DBT on the hydrogenation of 2-MPy was
less drastic as at 300 ◦C, but still obvious (Fig. 10).
The adsorption constants of 2-MPy and 2-MPiper on the
HDS sites were determined by analyzing our experimental
results with the Langmuir-Hinshelwood model, by assuming
that the HDS of DBT proceeds on a site that can be poisoned
by 2-MPy and 2-MPiper:
∂PDBT
∂τ
kKDBTPDBT
=
.
1 + KDBTPDBT + KNPN
In the analysis we used the data of the HDS of DBT itself
and of the HDS of DBT in the presence of 2, 6, and 10 kPa
2-MPy and 2-MPiper obtained at 300 ◦C. The adsorption
constants of DBT, 2-MPy, and 2-MPiper on the HDS sites of
the NiMo/γ -Al2O3 catalyst were found to be equal to 0.029,
0.53, and 0.076 kPa−1, respectively. This shows that the ad-
sorption constant of DBT is about 20 times smaller than that
of 2-MPy and 2 times smaller than that of 2-MPiper and that
the adsorption constant of 2-MPy is 7 times higher than that
of 2-MPiper.
Satterfield et al. showed that sulfur compounds as well
as H2S promote the rupture of C–N bonds [2]. They also
demonstrated that H2S released in the HDS of thiophene,
and not thiophene itself, was responsible for the enhance-
ment effect on HDN. This was confirmed by our results,
which showed that, in the HDN of 2-MPy and of 2-MPiper,
there was no change in the formation of C6 products when
DBT was added to the feed (Figs. 10 and 11). This proves
that the active sites that facilitate C–N bond cleavage are
different from those responsible for the HYD and DDS of
S-containing molecules. At the same time, as noted above,
N-containing compounds retard both DDS and HYD path-
ways, but to a different extent. The HYD pathway of the
HDS of DBT was strongly suppressed in the presence of 2-
MPy and 2-MPiper at 300 ◦C as well as at 340 ◦C, whereas
the DDS pathway was inhibited to a smaller extent. At high
temperatures and low partial pressures of 2-MPy and 2-
MPiper it was almost not affected. Thus, while C–N bond
Two reactions take part in the HDN of 2-MPy (Scheme 2),
hydrogenation–dehydrogenation and C–N bond cleavage.
This suggests the presence of two active sites: sites re-
sponsible for the hydrogenation–dehydrogenation and sites
promoting the C–N bond breaking. H2S has a dual effect on
the HDN reactions of pyridine and quinoline [25,39,40]. It
slightly inhibits the hydrogenation and dehydrogenation, but
markedly accelerates the overall HDN rate because it pro-
motes C(sp3)–N bond cleavage. Our results show that there
is an optimum H2S concentration for the hydrogenation of
2-MPy at 300 ◦C (Fig. 7). On the one hand, H2S creates
acidic sites on the catalyst surface, thereby increasing the
adsorption of a basic aromatic molecule. On the other hand,
at higher concentrations H2S covers the vacancies on the
catalyst surface and hinders the adsorption of N-containing
molecules. 2-MPiper at 340 ◦C undergoes mainly HDN and
H2S has a positive influence on the C–N bond cleavage

M. Egorova, R. Prins / Journal of Catalysis 221 (2004) 11–19
19
cleavage needs another site than the HYD and DDS of S-
containing molecules (see above), the hydrogenation of S-
and N-containing molecules may take place at the same ac-
tive sites.
[10] U.S. Ozkan, S.G. Ni, L.P. Zhang, E. Moctezuma, Energy Fuels 8
(1994) 249.
[11] M.J. Girgis, B.C. Gates, Ind. Eng. Chem. Res. 30 (1991) 2021.
[12] P. Zeuthen, P. Stolze, U.B. Pedersen, Bull. Soc. Chim. Belg. 96 (1987)
985.
The enhancement of the C–N bond rupture and bet-
ter HDN activity of the catalyst in the presence of higher
amounts of H2S were previously associated with the main-
tenance of the catalyst in a completely sulfided state [3],
since the active S-containing species will be unstable on
the catalyst surface and will rapidly decompose in the ab-
sence of sufficient sulfur in the feed. However, we observed
an increase in the HDN conversion of 2-methylpiperidine at
much higher H2S partial pressures than needed to keep the
catalyst in a completely sulfided state. 2-MPiper undergoes
HDN via elimination or via nucleophilic substitution of the
amine group by an SH group followed by hydrogenolysis
of the C–S bond. In the latter case, H2S should be of im-
portance in the HDN of 2-MPiper. It was recently shown
that aliphatic amines undergo HDN mainly via nucleophilic
substitution [41]. Therefore, the enhancement of the HDN
of 2-MPiper with increasing H2S partial pressure may be as-
cribed to the reaction mechanism and not only to the sulfided
state of the catalyst.
[13] P. Desikan, C.H. Amberg, Can. J. Chem. 42 (1964) 843.
[14] G.H. Singhal, R.L. Espino, J.E. Sobel, G.A. Huff, J. Catal. 67 (1981)
457.
[15] T.C. Ho, J.E. Sobel, J. Catal. 128 (1991) 581.
[16] B.C. Gates, H. Topsøe, Polyhedron 16 (1997) 3213.
[17] T. Kabe, K. Akamatsu, A. Ishihara, S. Otsuki, M. Godo, Q. Zhang,
W.H. Qian, Ind. Eng. Chem. Res. 36 (1997) 5146.
[18] V. Meille, E. Schulz, M. Lemaire, M. Vrinat, J. Catal. 170 (1997) 29.
[19] R. Prins, Adv. Catal. 46 (2002) 399.
[20] M. Egorova, Y. Zhao, P. Kukula, R. Prins, J. Catal. 206 (2002) 263.
[21] N. Nelson, R.B. Levy, J. Catal. 58 (1979) 485.
[22] C.N. Satterfield, J.F. Cocchetto, AICHE J. 21 (1975) 1107.
[23] J.F. Cocchetto, C.N. Satterfield, Ind. Eng. Chem. Proc. Des. Dev. 15
(1976) 272.
[24] R.T. Hanlon, Energy Fuels 1 (1987) 424.
[25] M. Jian, J.L.R. Cerda, R. Prins, Bull. Soc. Chim. Belg. 104 (1995) 225.
[26] F. Rota, R. Prins, J. Mol. Catal. A 162 (2000) 359.
[27] S.H. Yang, C.N. Satterfield, J. Catal. 81 (1983) 168.
[28] A. Olalde, G. Perot, Appl. Catal. 13 (1985) 373.
[29] M. Houalla, N.K. Nag, A.V. Sapre, D.H. Broderick, B.C. Gates,
AICHE J. 24 (1978) 1015.
[30] T. Kabe, A. Ishihara, Q. Zhang, Appl. Catal. A 97 (1993) L1.
[31] S. Kasahara, T. Shimizu, M. Yamada, Catal. Today 35 (1997) 59.
[32] W. Robinson, J.A.R. van Veen, V.H.J. de Beer, R.A. van Santen, Fuel
Proc. Technol. 61 (1999) 89.
References
[33] E.J.M. Hensen, V.H.J. de Beer, J.A.R. van Veen, R.A. van Santen,
J. Catal. 215 (2003) 353.
[1] F.W. Kirsch, H. Shalit, H. Heinemann, Ind. Eng. Chem. 51 (1959)
1379.
[2] C.N. Satterfield, M. Modell, J.F. Mayer, AICHE J. 21 (1975) 1100.
[3] C.N. Satterfield, M. Modell, J.A. Wilkens, Ind. Eng. Chem. Proc. Des.
Dev. 19 (1980) 154.
[34] 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.
[35] M. Nagai, Ind. Eng. Chem Prod. Res. Dev. 24 (1985) 489.
[36] M. Nagai, Chem. Lett. 6 (1987) 1023.
[37] A. Albert, Heterocyclic Chemistry, The Athlone Press, London, 1959.
[38] P. Zeuthen, K.G. Knudsen, D.D. Whitehurst, Catal. Today 65 (2001)
307.
[39] C.N. Satterfield, S. Gultekin, Ind. Eng. Chem. Proc. Des. Dev. 20
(1981) 62.
[40] S.H. Yang, C.N. Satterfield, Ind. Eng. Chem. Proc. Des. Dev. 23 (1984)
20.
[4] M. Nagai, T. Kabe, J. Catal. 81 (1983) 440.
[5] M. Nagai, T. Sato, A. Aiba, J. Catal. 97 (1986) 52.
[6] V. La Vopa, C.N. Satterfield, Chem. Eng. Commun. 70 (1988) 171.
[7] V. La Vopa, C.N. Satterfield, J. Catal. 110 (1988) 375.
[8] G.C. Laredo, J.A. De los Reyes, J.L. Cano, J.J. Castillo, Appl. Catal.
A 207 (2001) 103.
[9] M. Cerny, Collect. Czech. Chem. Commun. 47 (1982) 1465.
[41] Y. Zhao, P. Kukula, R. Prins, submitted for publication.