Role of hydrogenolysis and nucleophilic substitution in hydrodenitrogenation over sulfided NiMo/γ-Al2O3

Article abstract of DOI:10.1006/jcat.2001.3277

The HDS of cyclohexanethiol and 2-methylcyclohexanethiol was studied over a sulfided NiMO/γ-Al₂O₃ catalyst. About 66% of the thiols reacted by elimination to (methyl)cyclohexene and 33% by hydrogenolysis of the C-S bond to (methyl)cyclohexane. These values were slightly lower than those for the selectivity to methylcyclohexene and slightly higher than those for the selectivity to methylcyclohexane in the HDN of 2-methylcyclohexylamine. HDN occurred predominantly in aliphatic molecules that contained H atoms in the β position relative to the nitrogen atom by elimination of ammonia. Part of the remaining HDN occurred by nucleophilic substitution of the amine by H₂S, followed by elimination of H₂S from the resulting thiol and, to a lesser extent, by C-S bond hydrogenolysis, the rest of the remaining HDN occurred by direct hydrogenolysis of the C-N bond.

Full text of DOI:10.1006/jcat.2001.3277

Journal of Catalysis 202, 195–199 (2001)
doi:10.1006/jcat.2001.3277, available online at http://www.idealibrary.com on
RESEARCH NOTE
Role of Hydrogenolysis and Nucleophilic Substitution
in Hydrodenitrogenation over Sulfided NiMo/ -Al2O3
F. Rota and R. Prins1
Laboratory for Technical Chemistry, Federal Institute of Technology (ETH), 8092 Zurich, Switzerland
Received March 16, 2001; revised May 11, 2001; accepted May 11, 2001
higher hydrogen consumption in HDN than in hydrodesul-
furization (HDS) and the increase in rate as a result of the
addition of H2S (2–5). Portefaix et al. (6) and Cattenot et al.
(7) showed that the aliphatic C–N bond breaking occurs
by Hofmann elimination because an increase in the num-
ber of hydrogen atoms in the position with respect to the
nitrogen atom in pentylamines and piperidines increased
the rate of ammonia removal. Even when there are no hy-
drogen atoms on the carbon atom and elimination is not
possible, C–N bond breaking is still possible, as shown by
Vivier et al. for benzylamine (8). They ascribed the C–N
bond breaking to nucleophilic substitution and subsequent
C–S hydrogenolysis.
In agreement with these findings, we observed that o-
toluidine reacts mainly by hydrogenation to 2-methyl-
cyclohexylamine and subsequent elimination of ammonia
from 2-methylcyclohexylamine to methylcyclohexene (9–
11). Part of the 2-methylcyclohexylamine reacted, how-
ever, to methylcyclohexane (9). This methylcyclohexane
may have been formed by nucleophilic substitution
of 2-methylcyclohexylamine to 2-methylcyclohexanethiol
and subsequent hydrogenolysis to methylcyclohexane
(10, 11). Methylcyclohexane may also form, however,
through direct hydrogenolysis of the C–N bond of 2-
methylcyclohexylamine. The possibility of C(sp3)–N bond
breaking by hydrogenolysis is suggested by the observa-
tion that, in the HDN of aniline (12, 13), o-toluidine (9–11),
2-propylaniline (14), and naphthylamine (15), part ofthe re-
actant reacted directly to benzene, toluene, propylbenzene,
and naphthalene, respectively. This indicates that C(sp2)–N
bond breaking is possible on metal sulfide catalysts. If this is
the case, then C(sp3)–N bond breaking should also be pos-
sible because a C(sp3)–N bond is weaker than a C(sp2)–N
bond.
The hydrodesulfurization of cyclohexanethiol and 2-
methylcyclohexanethiol was studied over a sulfided NiMo/ -Al2O3
catalyst. About two-thirds of the thiols reacted by elimination
to (methyl)cyclohexene and one-third by hydrogenolysis of the
C–S bond to (methyl)cyclohexane. These values are slightly lower
than those for the selectivity to methylcyclohexene and slightly
higher than those for the selectivity to methylcyclohexane in the
hydrodenitrogenation of 2-methylcyclohexylamine. In aliphatic
molecules that contain H atoms in the position relative to the
nitrogen atom, hydrodenitrogenation occurs predominantly (70–
80%) by elimination of ammonia. Part of the remaining (20–30%)
hydrodenitrogenation takes place by nucleophilic substitution
of the amine by H2S, followed by elimination of H2S from the
resulting thiol and, to a lesser extent, by C–S bond hydrogenolysis;
the rest of the remaining hydrodenitrogenation takes place by
c
direct hydrogenolysis of the C–N bond.
2001 Academic Press
Key Words: hydrodenitrogenation; hydrodesulfurization; hy-
drogenolysis; nucleophilic substitution; cyclohexanethiol; aniline;
cyclohexylamine; sulfide catalyst; NiMo/Al2O3.
INTRODUCTION
Nelson and Levy were the first to publish a general
mechanism of hydrodenitrogenation (HDN) of aromatic
nitrogen-containing molecules (1). They proposed that the
aromaticity of the nitrogen-containing aromatic ring must
be broken by hydrogen before C–N bond breaking and re-
moval of the nitrogen atom in the form of ammonia can
take place. It wasassumed that elimination and nucleophilic
substitution are responsible for the C–N bond breaking. In
nucleophilic substitution, the NH2 group is replaced by an
SH group through the reaction of the aliphatic amine with
H2S. The resulting alkanethiol then undergoes desulfuriza-
tion by means of a hydrogenolysis reaction with hydrogen
to an alkane and H2S. These proposals explained the much
To determine whether nucleophilic substitution or hy-
drogenolysis is responsible for the C(sp3)–N bond break-
ing, we studied the HDS of cyclohexanethiol over sulfided
NiMo/ -Al2O3. Cyclohexanethiol was chosen as a model
compound because it is commercially available and it is
1 To whom correspondence should be addressed. Fax: 41-1-6321162.
E-mail: prins@tech.chem.ethz.ch.
195
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

196
ROTA AND PRINS
the product of the nucleophilic substitution of the NH2 In the event of a change in the partial pressure of the reac-
group by the SH group in cyclohexylamine. Thus, cyclo- tant, the octane flow was adapted to keep the partial pres-
hexanethiol can be used to perform reactions and obtain sures of hydrogen, heptane, and hydrogen sulfide constant.
information about the importance of the nucleophilic sub- The HDS reactions of CHT were performed between 250
stitution in the HDN mechanism. To ensure that methyl and 375 C at CHT partial pressures of 10, 15, and 50 kPa as
substitution does not lead to a drastic change in the reac- well as with 15 kPa CHT in the presence of 25 kPa methyl-
tivity of cyclohexanethiol, we synthesized a small amount cyclohexylamine (MCHA). The HDS reaction of 15 kPa
of 2-methylcyclohexanethiol and compared its HDS with MCHT in the presence of 25 kPa cyclohexylamine (CHA)
that of cyclohexanethiol. NiMo/ -Al2O3 was chosen as the was performed only at 250 C. The mass balances for the
catalyst because the same catalyst was used in our study HDS reactions were generally better than 90% ; only in the
of the stereochemistry of elimination of ammonia from cy- case of the reaction of MCHT together with CHA did we
clohexylamine (16). Furthermore, the elimination and nu- experience plugging of the reactor and the mass balance
cleophilic substitution reactions are about equally fast over for CHA was only 80% . The mass balance for MCHT was,
sulfided Mo/ -Al2O3 as over sulfided NiMo/ -Al2O3 (9).
however, almost 100% . The HDN of 25 kPa CHA was per-
formed between 250 and 375 C in the same way as the
HDS of CHT. Samples were collected at the reactor outlet
using an automated valve maintained at 300 C and reac-
tor pressure. The reaction products were injected from the
valve into a gas chromatograph equipped with a 30-m DB-5
fused silica capillary column (0.32-mm i.d. and 0.25- m film
thickness) and a flame ionization as well as a pulsed flame
photometric detector.
EXPERIMENTAL
Cyclohexanethiol was obtained from Fluka; a mixture of
cis- and trans-2-methylcyclohexanethiol (MCHT) was pre-
pared by the free radical addition of thiolacetic acid to 1-
methylcyclohexene followed by hydrolysis of the formed
thiol acetates, as reported by Bordwell and Hewett (17).
This procedure provides an excellent synthetic route to
pure thiols since yields are high and the reaction pro-
ceeds exclusively by anti-Markovnikov addition. In short,
thiolacetic acid (Aldrich, 96% ) was slowly added to 1-
RESULTS
Reactions of CHT with 20 kPa H2S were performed at 10,
methylcyclohexene (Aldrich, 97% ) while irradiating with a 15, and 50 kPa of CHT and at 250, 300, 325, 350, and 375 C;
150-W lamp. The 2-methylcyclohexylthiol acetate was dis- the only products detected were cyclohexene (CHE) and
tilled under vacuum and then hydrolyzed by refluxing for cyclohexane (CH). Because of the high reactivity of CHT,
1 h in a solution of KOH in aqueous alcohol. After neu- its conversion was 100% at and above 325 C, even at the
tralization with glacial acetic acid and separation of the lowest weight time. Consequently, the CHE and CH selec-
aqueous and nonaqueous phases, the aqueous phase was tivities were influenced by the subsequent conversion of
extracted three times with n-hexane. These extracts and CHE to CH. Even at 250 C, the conversion of CHE to CH
the nonaqueous phase were combined and dried over mag- was not sufficiently slow to enable the measurement of the
nesium sulfate. After the hexane evaporated, the product real selectivity of CHT to CH. To inhibit the conversion
was distilled under vacuum. The final purity of the mixture of CHE to CH (11), 25 kPa MCHA was added during the
of cis and trans isomers was 97% , and the ratio cis-to trans- reaction of 15 kPa CHT. MCHA was used because it in-
MCHT was 81 : 16.
fluences elimination and direct C(sp3)–N bond breaking in
The preparation of the NiMo/ -Al2O3 catalyst, which the same way (11). In the presence of MCHA, the CHE
contained 8 wt% Mo and 3 wt% Ni, has been described and CH selectivities were only slightly dependent on CHT
elsewhere (10). In short, it was prepared by incipient wet- conversion; the real selectivities of CHE and CH during
ness impregnation, dried at 120 C for 15 h, and calcined at the CHT reaction were determined by extrapolation of the
500 C for 4 h. The resulting catalyst was crushed and sieved CH selectivities to zero weight time. Thus, the initial se-
to a particle size of 230 mesh to avoid diffusion effects on lectivities for the direct transformation of CHT to CH in
product distribution and conversion (9). The catalyst was the presence of 20 kPa H2S were 15, 25, and 30% at 250,
then diluted with SiC and loaded into a continuous-flow 325, and 350 C, respectively (Fig. 1). At 350 C and at the
fixed-bed reactor. In situ sulfidation was carried out with higher H2S partial pressure of 200 kPa, 35% CHT reacted
10% H2S in H2 at 370 C and 1 MPa for 4 h. Thereafter, the to CH.
pressure was increased to 5 MPa and the liquid feed of cy-
With the limited amount of MCHT that was synthesized
clohexanethiol (CHT, Fluka) or MCHT, octane (solvent), (10 g), we could perform only a few experiments at 250 C
and heptane (internal standard) was fed to the reactor by and 15 kPa MCHT. The addition of 25 kPa CHA slowed
a high-pressure syringe pump. A mixture of 10% H2S in down the conversion ofmethylcyclohexene (MCHE, result-
H2 was added to keep the H2S partial pressure constant at ing from the ammonia elimination from MCHT) to methyl-
20 kPa in all the experiments (unless indicated otherwise). cyclohexane (MCH); thus, it was possible to measure the

HYDRODENITROGENATION OVER SULFIDED NiMo/ -Al2O3
197
FIG. 1. Cyclohexane (CH) selectivity in the HDS of cyclohexanethiol
at 250, 325, and 350 C in the presence of 25 kPa 2-methylcyclohexylamine
and 20 kPa (solid symbols) and 200 kPa (open symbols) H2S.
FIG. 3. Arrhenius plots for the rates of formation of cyclohexene
(CHE) and cyclohexane (CH) in the HDN of 20 kPa cyclohexylamine
in the presence of 20 kPa H2S.
true MCHE and MCH selectivities. The results(Fig. 2) show
that cis-MCHT reacts fast and partly to trans-MCHT. The
selectivities of the products MCHE-1, MCHE-3, and MCH
were 50, 30, and 20% , respectively.
The HDN of CHA was studied to compare the rates of
formation of the CHE and CH products with those ob-
tained in the HDS of CHT. The rates were obtained from
data measured at 250, 295, 325, 350, and 395 C and were
extrapolated to zero space time; they are presented in the
form of Arrhenius plots in Fig. 3.
the role of nucleophilic substitution and hydrogenolysis,
we investigated the HDS of CHT. The main product was
CHE, in agreement with a former study, which demon-
strated that ethene is the main elimination product in the
HDS of ethanethiol (18). The selectivity to the minor prod-
uct CH was 15 to 35% , as was the case for the minor product
ethane in the HDS of ethanethiol (18). The rate of hy-
drogenolysis of ethanethiol to ethane (18) was equal to that
of methanethiol to methane (19).
The selectivity to CH increased with increasing temper-
ature (Fig. 1), showing that the C–S bond hydrogenolysis
has a higher activation energy than the elimination of H2S.
Moreover, the CH selectivity increased with increasing H2S
partial pressure. This observation may explain why an in-
crease in the partial pressure of H2S increased the forma-
tion of MCHT in the HDN of MCHA (10, 11). Assuming
that MCHT can easily react via hydrogenolysis to MCH,
it was suggested that the formation of MCH from MCHA
occurs via nucleophilic substitution of the NH2 group by an
SH group, followed by hydrogenolysis of the intermediate
MCHT to MCH.
DISCUSSION
As explained in the Introduction, the aim of this study
was to determine the roles of nucleophilic substitution
and hydrogenolysis in the HDN of aniline-like molecules
and their intermediates. The HDN ofCHA can be described
by three reactions: (A) the elimination of NH3, leading to
cyclohexene (CHE), (B) the nucleophilic substitution of
the NH2 group by H2S, leading to CHT, and the subse-
quent reaction of the formed CHT via hydrogenolysis to
cyclohexane (CH) or via elimination to CHE, and (C) the
direct hydrogenolysis of CHA to CH (compare Fig. 4 for
the equivalent scheme of reactions for MCHA). To study
We want to compare the CH selectivities of 30 and 35% ,
obtained in the HDS of CHT in the presence of MCHA
FIG. 4. Selectivities for elimination (A), nucleophilic substitution
(B), and hydrogenolysis (C) in the HDN of 2-methylcyclohexylamine
(MCHA) and the observed selectivities of methylcyclohexene (MCHE)
and methylcyclohexane (MCH) in the HDN of 2-methylcyclohexylamine
(normal figures) and in the HDS of 2-methylcyclohexanethiol (MCHT)
(bold figures) in the presence of 20 kPa (left-hand side) and 200 kPa (right-
hand side) H2S.
FIG. 2. Conversion of cis- and trans-2-methylcyclohexanethiol
(MCHT) in the presence of 25 kPa cyclohexylamine and 20 kPa H2S at
250 C.

198
ROTA AND PRINS
TABLE 1
of MCHE-1 at low temperature because it must react via
syn elimination, which is known to be more difficult than
anti elimination. Figure 2 indeed shows that trans-2-MCHT
reacts very slowly at 250 C. However, cis-2-MCHA reacts
fast to MCHE-1 (16). Therefore, substitution of the NH2
group by an SH group cannot be dominant in MCHA, and
the cases (Table 1) in which the substitution is relatively
low ( 34% ) seem most likely. As Table 1 shows, this also
means that a direct contribution of 6 to 18% C(sp3)–N bond
breaking (via hydrogenolysis) must occur.
Case 5, in which nucleophilic substitution does not occur,
is not possible either, as proven by the detection of MCHT
and the increasing MCH selectivity with increasing H2S
partial pressure (10). H2S also increases the CH selectivity
Five Cases of Selectivities of Elimination, Substitution, and Hy-
drogenolysis of Methylcyclohexylamine at 350 C and 20 kPa H2S
A
B
C
Hydrogenolysis
(% )
Elimination
(% )
Substitution
(% )
Case
1
2
3
4
5
49
60
71
76
82
51
34
17
9
0
6
12
15
18
0
and 20 or 200 kPa H2S, respectively (Fig. 1), with those in the HDS of CHT (Fig. 1). Consequently, C(sp3)–N bond
obtained in the HDN of MCHA (10). To do so, we assume breaking of methylcyclohexylamine must (at least partially)
that the MCH selectivity obtained in the MCHT reaction take place via nucleophilic substitution of MCHT followed
at 350 C is 5% larger than that of CH in the reaction of by hydrogenolysis to MCH.
CHT, as observed at 250 C. These adapted selectivities are
The above discussion shows that neither nucleophilic
used in the network of MCHA (Fig. 4) to explain the MCH substitution nor hydrogenolysis dominates. A possibility
selectivities of 18 and 22% and the MCHE selectivities of for the HDN of MCHA at 350 C in the presence of
82 and 78% for the reaction of MCHA at 350 C and 20 and 20 kPa H2S may thus be 71% elimination, 17% nucle-
200 kPa H2S, respectively (10).
ophilic substitution, and 12% hydrogenolysis. In the pres-
Since the selectivity for hydrogenolysis of MCHT to ence of 200 kPa H2S these values may be 60, 30, and 10% ,
MCH does not differ dramatically from the selectivity for respectively.
the formation of MCH in the HDN of MCHA, it is not pos-
The HDN of CHA showed that also in the reaction of
sible to determine the contributions of nucleophilic substi- this molecule elimination to CHE dominated over the
tution and hydrogenolysis unequivocally. Nevertheless, the reaction to CH; 80% of CHA reacted to CHE and 20% to
following discussion enables us to draw some qualitative CH at 325 C. The Arrhenius plots of the rates of formation
conclusions. Table 1 lists five possibilities for the rates of of CHE and CH in the HDN of CHA show two temper-
the three different pathways (elimination (A), nucleophilic ature regions for each product (Fig. 3). Between 250 and
substitution (B), and hydrogenolysis (C)) in the presence of 325 C, the activation energy for CH formation (60 10 kJ)
20 kPa H2S. The MCH selectivities are equal to 0.35B + C is higher than that of CHE (30 10 kJ), leading to an
(in % ) and were calculated on the basis of the observed increase in CH selectivity with increasing temperature.
MCH selectivity of 18% in the HDN of MCHA and of Above 325 C, the activation energies for the rates of
35% for the reaction of MCHT to MCH (Fig. 4). Table 1 formation of CHE and CH are about the same (140 10
also lists the extreme cases in which neither hydrogenolysis and 130 10 kJ, respectively). The changes in the activation
nor nucleophilic substitution occurs.
energy of CHE and CH clearly indicates that two mecha-
Case 1, in which hydrogenolysis does not take place, nisms are operating in both molecules during the HDN of
would lead to the conclusion that the substitution path- CHA. This supports the network presented in Fig. 4, which
way contributes 51% to the HDN of MCHA. This result shows that MCHE can be formed by the direct elimination
differs from the results of a stereochemical investigation of MCHA and by substitution to MCHT followed by
of the HDN of cis- and trans-2-MCHA (16). In that study, elimination. MCH can also be formed by two reactions,
elimination was shown to be much faster for the cis di- through direct hydrogenolysis of MCHA and through
astereomers than for the trans diastereomers of several cy- substitution to MCHT followed by C–S hydrogenolysis.
clohexylamines because they allow for an anti-geometric
The way in which hydrogenolysis of the aliphatic C–N
relationship in the chair conformation between the amino bond takes place is not yet clear. Hydrogenolysis of a C–S
group and a hydrogen atom in the position. If substitution bond is an accepted phenomenon, although in this case, too,
were to account for 51% of the total rate of the cyclohexy- the mechanistic aspects have not been studied in depth. It
lamines, then a substantial contribution of Walden inver- is unlikely that the C–S or C–N bond is broken as in the
sion would follow at the -carbon atom when MCHA re- hydrogenolysis of hydrocarbons on metals. During the lat-
acts to MCHT by nucleophilic substitution (20). That means ter hydrogenolysis, the C–C bond is assumed to be parallel
that cis-2-MCHA, which was found to react much faster to two metal surface atoms and the breaking of the C–C
than trans-2-MCHA, would be transformed into trans-2- bond and the formation of two M–C bonds are assumed
MCHT. Trans-MCHT should not result in the formation to take place in a concerted reaction. However, neither

HYDRODENITROGENATION OVER SULFIDED NiMo/ -Al2O3
199
the greater metal–metal distance in metal sulfides than in
REFERENCES
metals nor the intermediate sulfur atoms favor a similar
hydrogenolysis reaction on the surface of a metal sulfide.
The mechanism proposed for C–S bond breaking in thiols
on a Mo2Co2S3 cluster (21) is a more likely hydrogenolysis
mechanism on metal sulfides. During this process, the S–R
group is 3-bonded to three metal atoms (perpendicular in-
stead of parallel to the surface), and the H atom binds to
a neighboring sulfur atom. The strong bonding of the sul-
fur atom in the SR group to the metal atoms weakens the
S–C bond, causing it to break and, thus, enabling the RH
molecule to form. A similar mechanism, in which an N–R
group is 3-bonded to three metal atoms, may explain the
hydrogenolysis of the C–N bond.
1. Nelson, N., and Levy, R. B., J. Catal. 58, 485 (1979).
2. Cochetto, J. F., and Satterfield, C. N., Ind. Eng. Chem. Proc. Des. Dev.
15, 272 (1976).
3. Shih, S. S., Katzer, J. R., Kwart, H., and Stiles, A. B., Prep. Am. Chem.
Soc. Div. Pet. Chem. 22, 919 (1977).
4. Satterfield, C. N., and Cocchetto, J. F., Ind. Eng. Chem. Proc. Des. Dev.
20, 53 (1981).
5. Perot, G., Catal. Today 10, 447 (1991).
6. Portefaix, J. L., Cattenot, M., Guerriche, M., Thivolle-Cazat, J., and
Breysse, M., Catal. Today 10, 473 (1991).
7. Cattenot, M., Portefaix, J. L., Afonso, J., Breysse, M., Lacroix, M., and
Perot, G., J. Catal. 173, 366 (1998).
8. Vivier, L., Dominguez, V., Perot, G., and Kasztelan, S., J. Mol. Catal.
67, 267 (1991).
9. Rota, F., and Prins, R., Stud. Surf. Sci. Catal. 127, 319 (1999).
10. Rota, F., and Prins, R., J. Mol. Catal. A: Chem. 162, 359 (2000).
11. Rota, F., and Prins, R., Top. Catal. 11/12, 327 (2000).
12. Geneste, P., Moulinas, C., and Olive´, J. L., J. Catal. 105, 254 (1987).
13. Moreau, C., Joffre, J., Saenz, C., and Geneste, P., J. Catal. 122, 448
(1990).
14. Jian, M., Kapteijn, F., and Prins, R., J. Catal. 168, 491 (1997).
15. Moreau, C., Bekakra, L., Olive´, J. L., and Geneste, P., in “Proceedings,
9th International Congress on Catalysis, Calgary, 1988” (M. J. Philips
and M. Ternan, Eds.), Vol. 1, p. 58. Chem. Institute of Canada, Ottawa,
1988.
16. Rota, F., Ranade, V., and Prins, R., J. Catal. 200, 000 (2001).
17. Bordwell, F. G., and Hewett, W. A., J. Am. Chem. Soc. 79, 3493 (1957).
18. Kieran, P., and Kemball, C., J. Catal. 4, 380 (1965).
19. Wilson, R. L., and Kemball, C., J. Catal. 3, 426 (1964).
20. March, J., “Advanced Organic Chemistry,” 3rd ed., Chap. 10. Wiley,
New York, 1985.
APPENDIX: ABBREVIATIONS
CH
Cyclohexane
CHA
CHE
CHT
MCH
Cyclohexylamine
Cyclohexene
Cyclohexanethiol
Methylcyclohexane
MCHA 2-Methylcyclohexylamine
MCHE Methylcyclohexene
MCHT 2-Methylcyclohexanethiol
ACKNOWLEDGMENT
21. Curtis, M. D., and Druker, S. H., J. Am. Chem. Soc. 119, 1027 (1997).
We thank Dr. P. Kukula for synthesizing 2-methylcyclohexanethiol.