Bifunctional behavior of bulk MoOxNy and nitrided supported NiMo catalyst in hydrodenitrogenation of indole

Article abstract of DOI:10.1006/jcat.1998.2381

Oxynitrides of early transition metals are bifunctional catalysts (metallic and addic sites) active in hydrodenitrogenation (HDN). TheHDN of indole was used as a molecular probe reaction to study the metallic and acidic properties of the dual sites on oxy

Full text of DOI:10.1006/jcat.1998.2381

Journal of Catalysis 183, 63–68 (1999)
Article ID jcat.1998.2381, available online at http://www.idealibrary.com on
Bifunctional Behavior of Bulk MoOxNy and Nitrided Supported NiMo
Catalyst in Hydrodenitrogenation of Indole
,1
K. Miga, K. Stanczyk, C. Sayag,† D. Brodzki,† and G. Dje´ga-Mariadassou†
Institute of Coal Chemistry, Polish Academy of Sciences, ul. Sowinskiego 5, 44-102 Gliwice, Poland; and †Universite´ P. et M. Curie,
Laboratoire Re´activite´ de Surface, CNRS-UMR 7609, 4 Place Jussieu, Casier 178, 75252 Paris Cedex 05, France
Received July 16, 1998; revised October 22, 1998; accepted December 10, 1998
gen surface monolayer. Subsequently, oxygen atoms can
react with hydrogen atoms in the bulk during the nitrida-
tion or carburization processes. Hydrogen atomscan diffuse
from the bulk to the surface and lead to OH surface groups.
Several fundamental studies on tungsten carbides (Ribeiro
et al. (2) and Muller et al. (3)) and on oxynitrides (Sellem
et al. (4)) recently demonstrated the existence of surface
Brønsted acidity in the isomerization of alkanes.
As a result, two types of sites exist on the surface of ni-
trides or carbides prepared at high temperature after expo-
sure to oxygen: metallic sites (early transition metal atoms
modified by C and/or nitrogen) and acidic sites (Brønsted
groups). Lewis sites, due to low oxidation states of the
transition metal, cannot be excluded. Therefore, nitrides
and carbides of transition metals prepared at high tem-
perature are bifunctional. Oxygen-containing nitrides and
carbides are called “oxynitrides” (MeOxNy) and “oxycar-
bides” (MeOxCy) respectively.
Oxynitrides of early transition metals are bifunctional catalysts
(metallic and acidic sites) active in hydrodenitrogenation (HDN).
The HDN of indole was used as a molecular probe reaction to study
the metallic and acidic properties of the dual sites on oxynitrides.
The behavior of those sites was found to be different from that of
supported and sulfided NiMo. The reaction was conducted at low
partial pressure of H2S over both a bulk MoOxNy and a supported
NiMo catalyst, nitrided before the reaction. Under high hydrogen
pressure, HDN of indole was shown to occur with or without prior
hydrogenation of the aromatic ring of orthoethylaniline, leading
to either ethylbenzene or ethylcyclohexane. Secondary reactions,
such as the joining of heterocyclic rings (leading to 1-4 tetrahy-
droquinoline) and dimerization, were found to occur as a result of
the presence of acidic sites. Hydrogenolysis of the lateral chain of
c
orthoethylaniline was also observed.
1999 Academic Press
Key Words: molybdenum oxynitride; bulk; supported; high pres-
sure HDN.
In the case of molybdenum carbides, there is some confu-
sion between the high temperature carbides, prepared ac-
cording to Volpe and Boudart (5), and the low temperature
molybdenum carbides prepared by Ledoux and co-workers
(6). The latter materials are prepared by inserting carbon
into the lattice of molybdenum oxide and were found to be
monofunctional catalysts in alkane isomerization (6).
A characteristic of high temperature nitrides and car-
bides is that Brønsted groups and transition metal atoms
can exist side by side on the surface of the solid. The concept
of a “dual site” led us to propose a model for MoOxNy (7)
(Fig. 1). MoOxNy, prepared at high temperature, presents
a NaCl-like cubic structure. The fcc sublattice of molyb-
denum atoms defines octahedral sites located either in the
center or on the edges of the cube, where N, C, or O atoms
can be inserted. A dual site can be defined by a metallic
center such as a Mo atom, electronically modified by N or
C atoms, and an acidic site (Brønsted group). Vacancies can
exist in the sublattice of molybdenum, as shown by Gouin
et al. (8). Therefore, these materials are nonstoichiomet-
ric, but counting metallic sites can be done by specific CO
chemisorption under well-defined conditions (9).
1. INTRODUCTION
The main bulk or surface characteristics of nitrides and
carbides of early transition metals will be described again in
order to clarify the main concepts used in this work. Inser-
tion of nitrogen or carbon into the lattice of early transition
metals (groups 4 to 6) is now known to transfer some of
the properties of precious metals to these transition metals.
Carbon, in particular, causes them to hydrogenate stronger
than does the insertion of nitrogen (1). Another important
feature is the ability of these materials to adapt their own
chemical composition to that of the feed during reactions.
Carbon, nitrogen, and oxygen atoms can then exchange,
leading to the real catalytically active phase. A fundamen-
tal question remains: can sulfur or phosphorus atoms also
be inserted into the nitride or carbide lattice?
After synthesis, nitrides or carbides are pyrophoric, and
a passivation step is required to prevent the bulk oxida-
tion of these materials on exposure to air. A treatment with
1 vol% O2/helium at room temperature leads to an oxy-
1 To whom correspondence should be addressed.
63
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

64
MIGA ET AL.
Supported MoOxNy. The precursor of 76aN was a Pol-
ish NiMo catalyst referred to as 76a from the Institute of
Petroleum and Coal Chemistry, Wroclaw. In recent years,
this catalyst has been generally used for hydrotreating coal-
derived fuels after sulfidation. With such a feedstock, no
limitation of external diffusion was observed. This precur-
sor was used in the present study for two reasons: (i) it was
of interest to nitride, instead of sulfide, the oxidic form of
this supported catalyst to determine whether the nitridation
process can be extrapolated to already existing hydrotreat-
ing catalyst precursors, and (ii) it permitted stronger acidity
of the molybdenum nitride-supported catalyst. The precur-
sor consisted of 4 wt% NiO + 16 wt% MoO3 supported on
modified alumina (Al2O3 + 20 wt% of NiH–ZSM-5). The
nickel-exchanged form of H–ZSM-5 was obtained accord-
ing to the following steps:
FIG. 1. Acidicsite (OH group), metallicsite (Mo), and dualsite model
for MoOxNy.
Na-ZSM-5 H→Cl H-ZSM-5 Ni(NO→3)2 NiH-ZSM-5
The XRD pattern of 76aN showed that the crystallinity
of the zeolite was preserved after nitridation, as observed
previously with the EMT zeolite (10). The total specific
One aim of the present work was to study the effect of sur-
face acidity on the product distribution of indole reactions
during HDN of indole. To study the effect of the strength
and number of Brønsted groups, strong acidic sites are cre-
ated by supporting MoOxNy on a mixed [Al2O3/H–ZSM-5]
support.
A second aim deals with the study of the reaction path-
ways of HDN of indole and secondary reactions occurring
via acidic centers.
surface area of 76aN was 205 m2. g 1, and its pore volume
1
was 0.36 cm3
g
.
According to published data on bulk (2, 5, 11) and sup-
ported (10, 11) oxynitrides, both bulk molybdenum oxyni-
trides and nitrided supported commercial NiMo catalysts
were black, a characteristic of nitrided materials.
Temperature-programmed desorption (TPD) in the vac-
uum of preadsorbed ammonia at 370 K showed that 76a
and 76aN materials mainly exposed medium strength acid
sites (desorption temperature of NH3 between 570 and
720 K). No drastic difference was observed between the
NH3 TPD profiles of 76a and 76aN, indicating that acidity
was preserved after nitridation. The MoOxNy acidic sites
were mainly weak, as shown by the lower temperature
TPD peaks of ammonia. This was also observed by Nagai
et al. (12).
2. EXPERIMENTAL
2.1. Catalysts
Two kinds of nitrides were prepared: a bulk MoOxNy and
a nitrided NiMo supported catalyst referred to as 76aN.
Bulk MoOxNy. Molybdenum trioxide (Janssen-Chim-
ica) was used as the precursor of MoOxNy. Nitridation was
carried out according to the procedure of Volpe and
Boudart (5). Precursors (about 2 g) were placed in a quartz
reactor and nitrided in pure flowing NH3 (VHSV =
37,500 h 1). Temperature was increased in three steps, first
quickly from 293 to 633 K (17 K min 1), then more slowly
from 633 to 733 K (33 K h 1), and from 733 to 973 K
(100 K h 1). The temperature was held at 973 K for 3 h.
The system was then cooled to room temperature in flow-
ing NH3 and flushed in helium for 0.5 h. The sample was
passivated by flowing 1 vol% O2 in He for 1 h, before ex-
posure to air. The X-ray diffraction pattern of the bulk
molybdenum oxynitride showed the characteristic lines of
2.2. Catalytic Tests
Runs were performed in a high-pressure flow reactor,
equipped with a Deoxo device for hydrogen purification
(13). The main part of the apparatus is a fixed bed steel
reactor containing 2 cm3 of catalyst. The initial liquid feed
(cyclohexane or indole in cyclohexane) was provided by a
high-pressure pump.
2.3. Catalytic Runs
Allrunswere conducted in flowinghydrogen at high pres-
-Mo2N. Elemental chemical analysis indicated that the sure (9.0 MPa) in the presence of 70 ppm H2S without any
material contained 7 wt% oxygen after passivation. Its total prereduction of the catalysts. The liquid hourly space ve-
specific surface area was 130 m2
g
1, and its pore volume locity of the feed was 2 h 1, with a volume ratio between
hydrogen and the feed equal to 1000 : 1. The feed was a
1
was 0.17 cm3
g
.

BULK MoOxNy AND NITRIDED SUPPORTED NiMo CATALYST
65
TABLE 1
solution of 2 wt% indole (Fluka, purity > 99% ) in cyclo-
hexane. The catalyst volume was always 2 cm3 (76aN, 2.1 g
and MoOxNy, 2.7 g).
Product Distribution (mol%) Formed at 630 K,
9.1 MPa under Standard Conditions (1 wt% indole
in cyclohexane, 70 ppm H2S) on Nonnitrided 76a
The temperature profile of each run as well as the changes
in the feed during the runs were as follows. A run cycle
consisted of five 12-h periods (60 h) and of two 1-h periods
while the temperature increased. Each cycle started and
ended with 12-h periods of blank feed when only cyclohex-
ane (CH) was supplied. The liquid product was collected
in a high-pressure separator. Sampling was performed ev-
ery 2 h. After the initial temperature increase, steady state
was achieved after 2 h. Analytical data of the two last sam-
plings were taken into account for each period. Both cata-
lysts (bulk and supported) were subjected to the above
procedure.
Catalyst
76a
Indole and N-containing compounds
Indole
Indoline
0.8
0
0
0
0
0
0.8
Orthoethylaniline
Orthomethylaniline
Aniline
C4-aniline
Sum
Indole HDN products
Ethylcyclohexane
Ethylcyclohexene
Ethylbenzene
C4-cyclohexane
C4-benzene
0
0
0
0
0
0
3. RESULTS AND DISCUSSION
3.1. Reaction Network of HDN of Indole
Sum
Solvent
Cyclohexane
Figure 2 presents the reaction network of HDN of in-
dole according to Abe and Bell (14). The acronyms are also
listed. Two main routescan be defined for the removalofthe
nitrogen atom from indole. The first, route 1, leads to ethyl-
benzene (EB) through orthoethylaniline (OEA), without
saturation of the aromatic ring of OEA. This route is most
desirable, because it consumes less hydrogen.
The second route (Fig. 2, route 2), observed in the pre-
sence of sulfided commercial catalysts, goes through he-
xahydroindole (HHI) and orthoethylcyclohexylamine
(OECHA) to ethylcyclohexane (ECH) or, with saturation
of the aromatic ring of OEA, to OECHA, prior to the re-
moval of the nitrogen atom as NH3. Two of the reactions
involved in these HDN routes are hydrogenation and hy-
drogenolysis of C–N bonds. Both can be considered char-
acteristic of the metallic function of the catalyst. Whether
hydrogenolysis involves an acidic function remains an open
90.6
Solvent products
Methylcyclohexane
Ethylcyclohexane
Sum
0.5
0.9
1.4
Solvent isomers
Methylcyclopentane
Cyclopentane
Sum
6.9
1.1
8.0
Total
100.8
question. As will be shown, route 1 of HDN of indole is far
from negligible on bulk MoOxNy and 76aN, even at high
hydrogen pressure [9.1 MPa].
3.2. Blank Runs
Nickel in 76a is present as Ni species associated with
molybdenum as well as Ni exchanged in H–ZSM-5. In or-
der to rule out possible HDN activity of Ni species in 76a in
the presence of low H2S pressure, without prior nitridation
(not nitrided molybdenum being considered as inactive in
HDN), blank runs were made at 630 K under our standard
conditions (especially in the presence of 9.0 MPa H2 and
70 ppm H2S). Table 1 reports the product distribution of
the corresponding runs. It clearly shows that there is no
HDN of indole and only some alkylation or isomerization
of the solvent.
Ni in the presence of sulfur, at high temperature and high
pressure, is easily sulfided and becomes inactive (15–17).
3.3. HDN of Indole over MoOxNy at 600 K
FIG. 2. General scheme for HDN of indole at 630 K and 9.1 MPa,
going beyond Abe and Bell’s proposal (14). (1) Nitrogen-containing com-
pounds. HHI, hexahydroindoline; OEA, orthoethylaniline; OECHA, or-
thoethylcyclohexylamine. (2) EB, ethylbenzene; ECH, ethylcyclohexane.
Table 2 shows that 93.3 mol% indole is transformed
at 600 K, but 65.7 mol% of the products still N-contain

66
MIGA ET AL.
TABLE 2
3.3.2. Distribution of indole HDN products. Ethylcy-
clohexane, cyclohexene, and ethylbenzene are the main
products. The two routes in Fig. 2 can be considered to be in-
dependent, taking into account that EB cannot transform
to ECH in the presence of OEA because of the inhibit-
ing effect of OEA, which strongly adsorbs on the surface.
This behavior has already been described in previous stud-
ies of HDN on commercial catalysts (19, 20). In a parallel
work, the same behavior was observed for bulk molybde-
num oxynitrides: hydrogenation of propylbenzene (instead
of EB) on MoOxNy decreased from 98 to 0% when ortho-
propylaniline (instead of OEA) was added to the feed at
543 K and 5 MPa. Therefore, the ratio (ECH + ECHE)/EB
(18.7/6.2) = 3.0 permits the quantification of these two
routes over MoNxOy.
Ethylcyclohexene (ECHE) (2.8 mol% ) is the intermedi-
ate olefin between OECHA and ECH (Fig. 2), following
a Hoffmann-type -hydrogen elimination (21). At 600 K,
hydrogenation is not sufficiently strong to completely trans-
form ethylcyclohexene to ECH. Weak alkylations still oc-
cur and lead to 0.4 mol% butylcyclohexane and 0.1 mol%
butylbenzene.
Distribution of Reaction Products and Intermediates (in mol%)
Formed at 600 K at 9.1 MPa over MoOxNy and 76aN Catalysts
Catalysts
MoOxNy
93.3
76aN
63.5
Total indole conversion
Indole and N-containing compounds
Indole
Indoline
Orthoethylaniline
Orthomethylaniline
Aniline
6.7
3.2
45.2
9.3
36.5
15.2
15.25
7
1.2
0.2
C4-aniline
Sum
0.1
65.7
0.0
74.1
Indole HDN products
Ethylcyclohexane
Ethylcyclohexene
Ethylbenzene
C4-cyclohexane
C4-benzene
15.9
2.8
6.2
0.4
0.1
14.6
1.1
1.6
0
0
Sum
25.4
17.3
THQ cyclic compounds
1-4 THQ
2.8
0.9
0.1
0.5
1
3.4
0.1
0.0
0
0.6
4.1
Orthopropylaniline
Propylbenzene
C3-cyclohexene
C3-cyclohexane
Sum
3.3.3. Distribution of secondary products. The main
product observed is 1-4 THQ (2.8 mol% ), due to the en-
largement reaction of indoline. This enlargement requires
the presence of surface acidic sites. All intermediates (OPA,
propylcyclohexene (C3–CHene)), and products (ethylben-
zene (EB), propylcyclohexane (PCH)) ofthe 1-4THQ cycle
were detected. Another active acidic site is also shown by
the formation of dimers (3.4 mol% ) (Table 2).
5.3
Dimers
Sum
3.4
4.1
Total
99.8
99.6
compounds, whereas 25.4 mol% lead to HDN products
following the routes 1 and 2 in Fig. 2. Two secondary re-
actions, involving acidic sites (enlargement of indoline to
1-4 THQ and dimerization), lead to 5.3 mol% of 1-4 THQ
with all its HDN intermediates and products and 3.4 mol%
of dimers. In the next section, a detailed analysis of each
group of compounds will be given.
3.4. Comparison with Nitrided Supported Material (76 aN)
The detailed analysis of the intermediates and final prod-
ucts in the liquid phase resulting from the HDN of indole
over 76aN was conducted in a similar manner as the analysis
done over bulk MoOxNy (Table 2).
Table 2 (column 3) shows the lower global HDN activ-
3.3.1. Distribution of indole and N-containing com- ity of 76aN compared to MoOxNy (column 2) leading to a
pounds. Table 2 shows all the intermediates of route 1 higher mol% of N-containing compounds (74.1 mol% in-
(Fig. 2). Indole hydrogenates to indoline (3.2 mol% ), the stead of 65.7 mol% ) and a lower mol% of products of HDN
latter of which undergoes a first C–N hydrogenolysis lead- (17.3 mol% instead of 25.4 mol% ).
ing to OEA, the major intermediate product in the gas
In contrast, secondary reactions linked to the acidic func-
phase (45.2 mol% ). Secondary C–C hydrogenolysis of the tion are quite similar. The lower HDN activity over 76aN
lateral alkyl chain of OEA can occur, leading to 9.3 mol% appears in the intermediate distribution of indole and N-
OMA and 1.2mol% aniline. A veryweak alkylation process containing compounds. A higher amount of indole remains
produces butylaniline (C4–aniline) (0.1 mol% ).
(36.5 mol% instead of 6.7 mol% for bulk MoOxNy), and
According to a “rake” mechanism (18), quasi-equilibria a higher concentration of indoline (15.2 mol% instead of
between compounds in the gas phase and their correspond- 3.2 mol% ) is found; furthermore, the mol% of OEA is low-
ing adsorbed species may be assumed. Only low concentra- ered by a factor of 3. Hydrogenolysis of the alkyl chain
tions of indole (6.7 mol% ) and indoline (3.2 mol% ) remain leading to orthomethylaniline (OMA) and aniline is also
in the gas phase, showing that both molecules are trans- weaker.
formed and, therefore, that the catalyst surface is mainly
covered by adsorbed OEA, the major product in the gas route 2 is favored on the supported material, as shown by
phase.
the ratio (ECH + ECHE)/EB which is equal to (15.7/1.6) =
For a lower HDN conversion of indole at 600 K,

BULK MoOxNy AND NITRIDED SUPPORTED NiMo CATALYST
67
TABLE 3
C–C hydrogenolysis in accord with the metallic behavior of
MoOxNy at this temperature.
Distribution of Reaction Products and Intermediates (in mol%)
Formed at 630 K and 9.1 MPa over MoOxNy and 76aN Catalysts
Table 3 (column 2, second chemical group) shows the two
major products of HDN: ethylbenzene (23.9 mol% ) for
route 1 and ethylcyclohexane (63.2 mol% ) for route 2, with
a ratio (ECH + ECHE)/EB (64.4/23.9) = 2.7 corresponding
to a high hydrogenating function.
Catalyst
MoOxNy
100
76aN
93.3
Total indole conversion
Indole and N-containing compounds
Indole
Indoline
0
0
0.8
0.4
0
6.7
2.3
16.2
7.7
0.6
0
3.6. Effect of Temperature: HDN of Indole
at 630 K over 76aN
Orthoethylaniline
Orthomethylaniline
Aniline
C4-aniline
Sum
Table 3 (column 3) reports the detailed analysis of the in-
termediatesand productsover 76aN at 630K. The moderate
HDN activity (only 47.5 mol% of indole HDN products) is
due to a moderate hydrogenating function as shown by the
presence of unsaturated N-containing compounds.
In contrast, the activity of the acidic function of the sup-
port is enhanced when the reaction temperature is raised,
leading to 7.3 mol% of 1-4 THQ cycle compounds and
11.3 mol% of dimers.
0
1.2
33.5
Indole HDN products
Ethylcyclohexane
Ethylcyclohexene
Ethylbenzene
C4-cyclohexane
C4-benzene
63.2
1.2
23.9
0.9
0
89.2
36.8
1.1
8.7
0.8
0.1
Sum
47.5
THQ cyclic compounds
1-4 THQ
3.6.1. Distribution of indole and N-containing com-
pounds. The low HDN activity and the low hydrogenating
function are shown by the 6.7 mol% of indole that did not
transform and the 2.3 mol% of indoline which still did not
undergo C–N hydrogenolysis. OEA remains the major N-
containing intermediate (16.2 mol% ), leading to 7.7 mol%
OMA by C–C bond scission. The same HDN elementary
steps apply for both bulk and supported MoOxNy.
As a consequence of the moderate hydrogenating func-
tion of the 76aN material, 1.1 mol% of ethylcyclohexene
still remains in the final liquid phase. Route 2 tends to be
favored, as in the case of commercial sulfided NiMo/Al2O3
catalysts, with (ECH + ECHE)/EB (37.9/8.7) = 4.4 at 630 K
(Table 3), instead of (ECH + ECHE)/EB (15.7/1.6) = 9.8 at
600 K (Table 2).
0
0
0.4
0
5.8
6.2
0.3
0.6
0.6
0
5.8
7.3
Orthopropylaniline
Propylbenzene
C3-cyclohexene
C3-cyclohexane
Sum
Dimers
Sum
3.2
11.3
99.6
Total
99.9
9.8 over 76aN instead of (18.7/6.2) = 3.0 over MoOxNy
(Table 2).
3.5. Effect of Temperature: HDN of Indole at 630 K
over MoOxNy
Table 3 shows that products of the HDN of indole now
3.6.2. Comparison with bulk sample (MoOxNy). The
represent 89.2 mol% of the total final liquid instead of differences between bulk MoOxNy and 76aN materials in
25.4 mol% at 600 K (Table 2, column 2), whereas only 1.2% HDN activity, global hydrogenating, and metallic behavior,
of indole and N-containing compounds remain. In contrast, as well as the acidic function are now quite obvious from
secondary reactions—enlargement of indoline to 1-4 THQ the data in Table 3.
and its products of HDN (6.2 mol% ) and dimers (3.2 mol% )
In the presence of 70 ppm H2S, 33.5 mol% of indole and
—still remain in the same range as at 600 K: 6.2 mol% in- N-containing compounds still remain with the 76aN catalyst
stead of 5.3 mol% at 600 K and 3.2 compared to 3.4 mol% compared to only 1.2 mol% with the bulk material (Table 3,
at 600 K for dimers. This must be linked to a lower number first chemical group).
of acidic sites or to fewer active acidic sites when the tem-
The 76aN material presents only about 53% (47.5/89.2)
perature is raised. This may also be due to a lower number of the total HDN activity of the bulk MoOxNy (Table 3,
of surface OH groups due to a reduction of the surface by second chemical group).
H2 at higher temperature.
Secondary reactions, occurring on acidic sites, lead to a
Table 3 (column 2, first chemical group) shows that the higher hydroconversion of indoline to 1-4 THQ cyclic com-
higher HDN conversion observed corresponds to a total pounds and to a higher concentration of dimers (Table 3,
hydrogenation of indole to indoline and to a total C–N third chemical group).
bond scission of the heterocycle of indoline to give mainly
Data in Table 3 clearly show the large difference be-
OEA + OMA. At 630 K, OEA is clearly the most abun- tween the two catalysts: (i) MoOxNy gives pratically no
dant reactive intermediate, OMA corresponding again to a N-containing compounds (only low amounts of OEA and

68
MIGA ET AL.
OMA), whereas 76aN gives large amounts of all initial and the Action Concerte´e, France–Poland, the French Ambassy, and the Polish
Science Committee (KBN), Warsaw.
intermediate compounds and (ii) different distributions of
HDN products are observed where ECH + EB are the two
major products for both materials.
REFERENCES
CONCLUSIONS
1. Dje´ga-Mariadassou, G., Boudart, M., Bugli, G., and Sayag, C., Catal.
Lett. 31, 411 (1995).
2. Ribeiro, F. H., Boudart, M., Dalla-Betta, R. A., and Iglesia, E., J. Catal.
130, 498 (1991).
3. Muller, A., Keller, V., Ducros, R., and Maire, G., Catal. Lett. 35, 65
(1995).
There are two ways to design new catalysts or new cata-
lytic functions. The first is to prepare new materials (here
the bulk MoOxNy) or to modify existing catalysts (here a
commercial NiMo-supported catalyst). In hydrotreating,
NiMo-supported catalysts are generally sulfided to proceed
to either the HDN or HDS reaction. The correspond-
ing molybdenum sulfide can activate both H2 and the
N-containing molecules. One of the main properties of sul-
fides, nitrides, and carbides is that they are able to exchange
S, N, or C (7, 22). During HDN over a sulfided molybdenum
catalyst, nitrogen insertion probably occurs. Oxynitrides
and oxycarbides are efficient for activating dihydrogen and
N-containing molecules. Thus, nitridation of a commercial
NiMo-supported catalyst (76aN) can be achieved. Fur-
thermore, bulk MoOxNy and 76a nitrided (76aN) present
metallic and acidic functions. Since the 76aN material
4. Sellem, S., Potvin, C., Manoli, J. M., Contant, R., and Dje´ga-
Mariadassou, G., J. Chem. Soc. Chem. Commun., 359 (1995).
5. Volpe, L., and Boudart, M., J. Solid State Chem. 59, 332 (1985).
6. Ledoux, M. J., Guille, J., and Cuong, P. H., J. Catal. 194, 176
(1988).
7. Kim, H. S., Sayag, C., Bugli, G., Dje´ga-Mariadassou, G., and Boudart,
M., Mat. Res. Sci. Symp. Proc. Series 368, 3 (1995).
8. Gouin, X., Marchand, R., L’Haridon, P., and Laurent, Y., J. Solid
State Chem. 109, 175 (1994).
9. Sayag, C., Bugli, G., Havil, P., and Dje´ga-Mariadassou, G., J. Catal.
167, 372 (1997).
10. Sellem, S., Potvin, C., Manoli, J. M., Maquet, J., and Dje´ga-
Mariadassou, G., Catal. Lett. 41, 89 (1996).
11. Be´cue, T., Manoli, J. M., Potvin, C., and Dje´ga-Mariadassou, G.,
J. Catal. 170, 123 (1997).
contains HZSM-5, its acidity is stronger. Sellem et al. (10) 12. Nagai, M., Kusagaya, T., Miyata, A., and Omi, S., Bull. Soc. Chim.
Belg. 104, Nos. 4–5 (1995).
13. Kaernbach, W., Kisielow, W., Warzecha, L., Miga, K., and Klecan, R.,
Fuel 69, 221 (1990).
14. Abe, H., and Bell, A. T., Catal. Lett. 18, 1 (1993).
showed that, for W2N/EMT, the nitridation step preserved
the activity of the acid sites for isomerization and cracking.
One of the main results of this paper is that route 1 (Fig. 2)
of the HDN of indole is possible over molybdenum oxyni-
tride. This is in good agreement with the work of Abe and
Bell (14), and Li and Lee (23) who performed HDN at
low hydrogen pressure. A low hydrogen pressure does not
thermodynamically favor hydrogenation, and these authors
found a large amount of ethylbenzene, indicating the pos-
sibility of HDN via route 1, as described in our work. This
route does not require the hydrogenating step of indole be-
fore C–N scission, and it is observed even under 9.1 MPa
of hydrogen, as shown here.
15. Bartholomew, C. H., Weatherbee, G. D., and Jarvi, G. A., J. Catal. 60,
257 (1979).
16. Breysse, M., Cattenot, M., Decamp, T., Frety, R., Gachet, C., Lacroix,
M., Leclercq, C., de Mourgues, L., Portefaix, J. L., Vrinat, M., Houari,
M., Grimblot, J., Kasztelan, S., Bonnelle, J. P., Housni, S., Bachelier,
J., and Duchet, J. C., Catal. Today 4, 39 (1988).
17. Suzuki, M., Tsutsumi, K., Takahashi, H., and Saito, Y., Zeolites 9, 98
(1989).
18. Germain, J. E., “Catalytic Conversion of Hydrocarbons,” p. 259.
Academic Press, New York, 1969.
19. Olive´, J. L., Biyoko, S., Moulinas, C., and Geneste, P., Appl. Catal. 19,
165 (1985).
20. Jian, M., Kapteijn, F., and Prins, R., J. Catal. 168, 491 (1997).
21. Perot, G., Catal. Today 10, 447 (1991).
ACKNOWLEDGMENTS
22. Marchand, R. J., Gouin, X., Tessier, F., and Laurent, Y., Proc. Mat.
Res. Soc. Symp. 368, 15 (1994).
23. Li, S., and Lee, J. S., J. Catal. 173, 134 (1998).
This work was supported by the International Program for Scientific
Cooperation, France–Poland, PICS 508, CNRS/ADEME/MAE, and by