Characterization and HDS activity of cobalt molybdenum nitrides

Article abstract of DOI:10.1021/jp002133c

The characterization of cobalt molybdenum nitrides with various cobalt contents has been studied using XRD, TEM, XPS, and temperature-programmed reactions with NH₃. The hydrodesulfurization of thiophene on the cobalt molybdenum nitride was carried out using a differential microreactor at atmospheric pressure. The XRD analysis of Co(50)973 indicated that Co₃Mo₃N was formed the nitriding of a CoMo₄ precursor prepared from Co(NO₃)₂?·6H₂O and (NH₄)₆ Mo₇O₂₄?·4H₂O but not from the mixture of CoO and MoO₃. Co(10-25)973 had a high surface area of 120-129 m² g^-1, but Co(50)973 had that of 12 m² g^-1. The BET surface area was related to the molybdenum ions (Mo²⁺ and Mo³⁺) and cobalt ion (Co³⁺) in the bulk sample. The XPS and elemental analyses of Co(50)973 showed that the bulk composition was Co₃Mo₃N?·N_0.2 and the surface composition was Co₃Mo₃N?·N₂. The cobalt molybdenum nitride Co(0-35)973 was more active than cobalt molybdenum sulfide on basis of catalyst weight in the hydrodesulfurization of thiophene at 623 K.

Full text of DOI:10.1021/jp002133c

4084
J. Phys. Chem. B 2001, 105, 4084-4093
ARTICLES
Characterization and HDS Activity of Cobalt Molybdenum Nitrides
Kenichiro Hada, Masatoshi Nagai,* and Shinzo Omi
Graduate School of Bio-Applications and Systems Engineering, Tokyo UniVersity of Agriculture and
Technology, 2-24 Nakamachi, Koganei, Tokyo 184-8588, Japan
ReceiVed: June 13, 2000; In Final Form: March 1, 2001
The characterization of cobalt molybdenum nitrides with various cobalt contents has been studied using XRD,
TEM, XPS, and temperature-programmed reactions with NH
cobalt molybdenum nitride was carried out using a differential microreactor at atmospheric pressure. The
XRD analysis of Co(50)973 indicated that Co Mo N was formed from the nitriding of a CoMoO precursor
prepared from Co(NO ‚6H O and (NH Mo O but not from the mixture of CoO and MoO . Co-
3
. The hydrodesulfurization of thiophene on the
3
3
4
3
)
2
2
4
)
6
7
O
24‚4H
2
3
2
-1
2
-1
(
10-25)973 had a high surface area of 120-129 m g , but Co(50)973 had that of 12 m g . The BET
2
+
3+
3+
surface area was related to the molybdenum ions (Mo and Mo ) and cobalt ion (Co ) in the bulk sample.
The XPS and elemental analyses of Co(50)973 showed that the bulk composition was Co Mo N‚N0.2 and the
surface composition was Co Mo
3
3
3
3
N‚N . The cobalt molybdenum nitride Co(0-35)973 was more active than
2
cobalt molybdenum sulfide on basis of catalyst weight in the hydrodesulfurization of thiophene at 623 K.
Introduction
Transition-metal nitrides are active during hydrodesulfuriza-
attention has been paid to whether bimetallic nitrides had high
surface areas and the composition on the surface and in the bulk
of the cobalt molybdenum nitrides. Although the preparation
and characterization of passivated bimetallic nitrides were
1
-12
2,8,13-21
tion (HDS)
and hydrodenitrogenation
and are re-
ported to have a turnover frequency higher than commercial
22-24,31-33
extensively studied ex situ on the basis of XRD,
there
8
,15
20
Co-Mo
and Ni-Mo sulfide catalysts on the basis of the
have been few studies on the in situ characterization of the
nitrides using XPS with a complement to XRD. The XPS
analysis can provide the oxidation states and compositions of
metal atoms in the bulk and on the surface of the samples. More
recently, the catalytic activities of supported and unsupported
Co-Mo nitrides for hydroprocessing over bimetallic nitrides
CO and O2 adsorption. Recently, novel bimetallic nitrides and
22-30
25,31
31,32
oxynitrides such as Co-Mo,
Fe-Mo,
Ni-Mo,
and
V-Mo2
3,24,33
were studied with respect to the catalytic activity
and properties of the samples when synthesizing by temperature-
programmed nitriding of the bimetallic oxides and a mixture
31
of the monometallic oxides. Bem et al. reported the synthesis
of Ni3Mo3N by nitriding NiMoO4 at 973 K with ammonia and
Fe3Mo3N by nitriding FeMoO4 at 1073 K. Also, Co3Mo3N was
6
,22,27,29
22
27
were studied.
Kim et al. and Bussell and co-workers
reported that cobalt molybdenum nitrides had greater HDS
activity than γ-Mo2N. In this study, the cobalt molybdenum
nitrides are synthesized at a temperature of 973 K via nitriding
of the precursor prepared from Co(NO3)2‚6H2O and (NH4)6-
Mo7O24‚4H2O and from a mixture of CoO and MoO3. Also,
the formation and composition of cobalt molybdenum nitride
in the bulk sample and on the surface are confirmed by XRD
and XPS along with changing the ratio of Co/(Co + Mo) in
the cobalt molybdenum oxides. The relationship between the
surface area and the cobalt and molybdenum species of the
samples are analyzed on the basis of the XPS spectra of Co 2p,
Mo 3d, and N 1s of the samples. The catalytic activities of the
cobalt molybdenum nitride samples were tested in thiophene
HDS reactions at 623 K and atmospheric pressure.
prepared from temperature-programmed nitriding of the Co-
MoO4 precursor at a final temperature of 973 K2
2,25-29
and of
NH5(CoOHMoO4) via Co metal and γ-Mo2N intermediates at
3
0
23,24
about 880 K. On the other hand, Oyama et al.
reported
Co-Mo oxynitrides during nitriding of a physical mixture of
cobalt and molybdenum oxides via a solid-state reaction.
Although the formation of cobalt molybdenum nitride from an
oxidic precursor has been the subject of study for a long time,
there is little agreement as to the precursor of cobalt molybde-
num nitride. Furthermore, only a few attempts have so far been
made to determine the oxidation states of the cobalt and
molybdenum species in the nitrides. For the surface area of the
nitrides, the Co-Mo oxynitride had a high surface area (103
2
-1 23,24
m g )
(
on one side and Co3Mo3N had a low surface area
Experimental Section
2
-1 22
19 m g ) on the other. Although a large number of studies
Preparation of Cobalt Molybdenum Nitride. Oxidic pre-
cursors with several Co/(Co + Mo) ratios were prepared using
a mixture of an aqueous solution of cobalt nitrate (Co(NO3)2‚
H2O) (Kishida Chemical Co., 99%) and ammonium hepta-
molybdate (NH4)6Mo7O24‚4H2O) (Kishida Chemical Co., 99%).
The solid products were dried at 373 K overnight and calcined
have been carried out on the high surface area, little is know
about the relation between the high surface area and the
composition of the cobalt molybdenum nitrides. Moreover, no
*
To whom correspondence should be addressed. Phone and fax: +81
(
0)42 388 7060. E-mail: mnagai@cc.tuat.ac.jp.
1
0.1021/jp002133c CCC: $20.00 © 2001 American Chemical Society
Published on Web 04/19/2001

Cobalt Molybdenum Nitrides
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4085
at 773 K for 7 h in dry air. The precursors were placed on a
fritted quartz plate in a 10 mm i.d. quartz microreactor and then
nitrided by a temperature-programmed reaction (TPR) with
ammonia (99.99%). A 0.2 g sample was oxidized in dry air
and mortar, dispersed in ethanol with an ultrasonic apparatus,
placed on a copper microgrid, and transferred to the analysis
chamber in the TEM.
XPS and Elemental Analysis. X-ray photoelectron spec-
troscopy was carried out using a Shimadzu ESCA 3200
photoelectron spectrometer with Mg KR radiation (1253.6 eV,
-
1
(
1.11 mL s ) at 723 K for 1 h for an oxidized sample. For
nitrided samples, the 0.2 g sample was oxidized in dry air at
23 K for 1 h, cooled from 723 to 573 K in dry air, nitrided
7
8
kV, 30 mA). The samples after nitriding were removed from
-
1
-1
from 573 to 973 K at a rate of 0.0167 K s with 1.11 mL s
the microreactor and then transferred to a glovebag in which
the atmosphere was exchanged with argon (99.9999%) five
times and filled with argon gas. The samples were mounted on
a holder with carbon tape in the glovebag without exposure to
air. Argon etching was done for the analysis of the bulk samples
of ammonia, held at this temperature for 3 h, and then cooled
to room temperature in flowing ammonia. The other oxidic
precursor was prepared from a physical mixture of CoO and
MoO3 (Co/(Co + Mo)) ) 0.5, after nitriding the sample for
comparison. To compare the catalytic properties of the nitrided
catalysts with the sulfided catalyst, Co(25) was sulfided at 623
-4
at a pressure of 8 × 10 Pa for 1 min. The analysis was
-6
typically done at a pressure of 5 × 10 Pa and at a scan speed
-
1
K under a stream of 10% H2S/H2 (4 L h ) after the catalyst
was heated at 723 K in air. The sample names of the cobalt
-1
-1
of 0.33 eV s with 0.05 eV (step) . The binding energy of C
1
s (284.6 eV) was taken as a reference to correct the binding
-
2
molybdenum samples are abbreviated Co(x)y, where x × 10
energy of the samples. The envelopes of the XPS binding
energies of Mo 3d, Mo 3p, Co 2p, and N 1s were deconvoluted
using Origin package software (Microcal Co.) after the analytical
data were transferred from the Shimadzu ESCA 3200 to a PC
data station. Since the binding energies of XPS Mo 3d3/2 and
is the atomic ratio of Co/(Co + Mo) in the oxides and nitrides
and y is 723 K for oxidizing or is 773-1073 K for the nitriding.
For example, Co(20)723 denotes the cobalt molybdenum
samples (Co(NO3)2‚H2O and (NH4)6Mo7O24‚4H2O with Co/(Co
Mo) ) 0.2) oxidized at 723 K. Co(50)973 and Co(5-25)-
73 denote the cobalt molybdenum samples with Co/(Co + Mo)
0.5 and 0.05-0.25, respectively, nitrided with ammonia at
BET Surface Area. The specific surface areas of the samples
+
9
)
0
2+
3+
3
d5/2 and Co 2p3/2 and 2p1/2 had many peaks (Mo , Mo , Mo ,
4+
5+
6+
0
2+
3+
Mo , Mo , and Mo and Co , Co , and Co ), the binding
energies and fwhms of Mo 3d and Co 2p were fixed and
9
73 K.
0
deconvoluted, except for those of Co (-0.5 to 0.5 eV deviation
from the binding energy of Co0 at 777.6 eV). The spectra of
XPS Mo 3d of the cobalt molybdenum nitrided samples were
were measured using an Omnisorp 100CX (Beckman Coulter
Co.). The surface area of the samples (0.1 g) was measured by
nitrogen adsorption at liquid-nitrogen temperature after the
deconvoluted using an intensity ratio of 2/3 and a splitting of
34
3
.2 eV and referenced to the data reported by Hercules et al.
-
4
sample was evacuated at 473 K and 1.3 × 10 Pa for 2 h.
35
0
and Hada et al.: Mo (Mo 3d5/2 binding energy, 227.7 ( 0.2;
2+
3+
XRD. The sample before and after nitriding was measured
fwhm, 1.2 ( 0.2), Mo (228.4 ( 0.1; 1.4 ( 0.2), Mo (229.2
4
+
5+
by XRD. The nitrided sample was passivated in 1% O2/He
( 0.1; 1.5 ( 0.3), Mo (230.1; 1.6 ( 0.2), Mo (231.6; 1.65
-
1
6+
(
0.167 mL s ) for longer than 12 h at room temperature after
nitriding. The diffraction pattern was obtained using a RAD-II
Rigaku Co.) equipped with Cu KR radiation (λ ) 1.542 Å).
( 0.2), and Mo (233.0; 1.7 ( 0.2). Since the Co 2p spectra
were very broad and difficult to deconvolute for both Co 2p3/2
0
2+
(
and 2p1/2 in the region of 775-805 eV, the peaks of Co , Co ,
3+
2+
The peaks were identified on the basis of the JCPDS card
references: MoO3 (JCPDS 5-0508, 2θ ) 12.8, 25.7, and 27.4°;
this study, 2θ ) 12.9, 25.8, and 27.4°), MoO2 (JCPDS 32-671,
Co , and Co satellites were fitted for only the Co 2p3/2 region
for all the samples. The peak fittings of the XPS Co 2p3/2 binding
energies were referenced to the data of cobalt molybdenum
3
6
2
5
θ ) 26.1, 36.9, and 53.2°; this study, 2θ ) 26.0, 36.9, and
3.6°), γ-Mo2N (JCPDS 25-1366, 2θ ) 37.4, 43.5, and 63.2°;
2 3
sulfides obtained by Pawelec et al., of CuCo/Al O by
Figueiredo et al., and of Co/Al O by Zsoldos et al. The Co
3
7
38
2
3
this study, 2θ ) 37.3, 43.5, and 63.0°), CoO (JCPDS 43-1004,
2p3/2 binding energies for the cobalt molybdenum nitrided
0
2
6
θ ) 36.5, 42.4, and 61.6°; this study, 2θ ) 36.7, 42.3, and
1.4°), Co metal (JCPDS 15-806, 2θ ) 44.8, 47.6, and 76.0°;
samples were deconvoluted to Co (BE, 777.6 ( 0.8 eV; fwhm,
2
+
3+
1.3 ( 0.1 eV), Co (779.9 ( 0.4 eV; 4.2 eV), Co (781.6 (
2
+
this study, 2θ ) 44.8, 47.5, and 75.9°), and CoMoO4 (JCPDS
0.3 eV; 4.7 ( 0.1 eV), and Co satellite (786.1 ( 0.3; 5.0
eV). The base line corrections made of the peak fitting of Co
2p3/2 and Mo 3d were carried out with the Shirley method
provided by the Shimadzu spectrometer manufacturer. Further-
more, the XPS N 1s binding energy of the samples at 397 (
0.3 eV (fwhm 2.0 eV) was used to calculate the nitrogen content
in the samples on the basis of a new determination method39
for the deconvolution of Mo 3p3/2 and N 1s in the region of
390 and 410 eV. Elemental analysis of the samples was carried
out using a Shimadzu CHN 123 elemental analyzer (oxygen
burning method) after the samples were evacuated at 673 K
2
)
1-868 and 25-1434, 2θ ) 14.2, 26.7, and 28.5°; this study, 2θ
14.2, 26.5, and 28.5°). The peaks of cobalt molybdenum
25
nitride were consistent with the data reported by Jackson et al.
and Kim et al.22 such that Co3Mo3N had peaks at 2θ ) 40.0,
4
2.6, 46.6 and 72.8°. The peaks of Co-Mo oxynitride were
identified by comparison with the data reported by Oyama and
2
3
co-workers at the broad peaks of 2θ ) 37.0, 43.0, and 63.0°.
Temperature-Programmed Reaction with NH3 (TPR). The
oxidic precursor after oxidizing at 723 K was heated from 573
-
1
-1
to 973 K at a rate of 0.167 K s with 1.11 mL s of ammonia
99.99%). The desorbed gases were monitored using a Balzers
-
4
(
and 1.3 × 10 Pa for 2 h.
Quadstar 422 quadrupole mass spectrometer. Ammonia and
water in the desorbing gases during TPR were qualitatively
analyzed at m/z 15 and 18, respectively, because they overlapped
at m/z 17. Hydrogen and nitrogen were monitored at m/z 2 and
HDS of Thiophene. The flow system for the HDS of
40
thiophene is described elsewhere. The system consisted of a
single-pass, differential microreactor. The HDS of thiophene
on the nitride and sulfide catalysts (ca. 0.2 g) were carried out
at 623 K and atmospheric pressure. The reaction feed, consisting
of 3.8 vol % thiophene in pure hydrogen, was introduced at a
2
8, respectively.
TEM. The morphology of the samples was determined using
-
1
a JEM-2000F transmission electron microscope (JEOL Co.)
operating at 200 kV equipped with an energy-dispersive X-ray
spectrometer (EDS). The sample was crushed in an agate pestle
rate of 50 mL min into the reactor. Quantitative analysis was
performed by injecting a sample from a sampling loop into the
gas chromatograph to analyze the amount of thiophene (column

4086 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Hada et al.
3
Figure 2. TPR profile of Co(50)723 under a stream of NH .
oxynitride in Co(50)773-Co(50)923. Moreover, a subsequent
nitridation to Co-Mo oxynitride with NH3 started above 820
K to desorb H2 and N2. The plateau of N2 formation at 820-
Figure 1. XRD patterns of cobalt molybdenum samples nitrided at
various temperatures: (a) Co(50)773; (b) Co(50)873; (c) Co(50)923;
(d) Co(50)973; (e) Co(50)1023; (f) Co(50)1073.
8
80 K and the H2O formation peak at 900 K were ascribed to
1
0% Silicon DC-550) at 348 K and reaction products (column
the complete formation of Co3Mo3N from the nitriding of Co-
Mo oxynitride. From the XRD results in Figure 1c, very broad
peaks of Co-Mo oxynitride were observed for the Co(50)923
sample (the broad peaks at approximately 37, 43, and 63°), while
cobalt molybdenum nitrides were observed for the Co(50)973
sample (Figure 1d). On the basis of the XRD and TPR data,
we propose that the reaction of CoMoO4 with NH3 proceeds
through the reaction pathway
VZ-8) at 313 K.
Results and Discussion
Formation of Co3Mo3N. The XRD patterns of Co(50)
nitrided at various temperatures are shown in Figure 1. The XRD
measurement showed that CoMoO4 was observed in Co(50)-
7
23 before nitriding. In Co(50)773-923 (Figure 1a-c), broad
NH3
peaks were observed at around 2θ ) 37, 43, and 63° and were
probably characteristic of the formation of Co-Mo oxynitride.
CoMoO4
8 CoMoO4-YN + YH O (<820 K)
Y 2
23
The Co(50)923 sample (Figure 1c) was observed at new peaks
of 2θ ) 40.0, 42.6, 46.6, and 72.8° which were in agreement
with 2θ ) 40.0, 42.5, and 46.5° due to Co3Mo3N reported by
Jackson et al.25 and Kim et al. The peak intensities increased
with increasing nitriding temperature and sharpen with progres-
sive increases in calcination temperature, indicating an increase
in crystallinity of the cobalt molybdenum nitride. The patterns
after nitriding at 1073 K are dominated by the characteristic
patterns of Co3Mo3N. On the other hand, a very small peak
was observed at about 37.4° in the XRD patterns of Co(50)973
NH3
8
Co Mo N + (4 - Y)H O + N +
3 3 2 2
H (>820 K)
2
22
Cobalt molybdenum samples nitrided at 973 K with various
cobalt contents were analyzed using XRD before and after
nitriding the mixture of cobalt nitrate and ammonium hepta-
molybdate at 973 K. The representative spectra are shown in
Figure 3. γ-Mo2N (Figure 3a) was formed, together with MoO2
for Co(5-25)973 (Figure 3b). γ-Mo2N exhibited weak and
broad peaks at 2θ ) 37.4, 43.5, and 63.2°, which were close to
(Figure 1d). This peak was different from that of Co-Mo
oxynitride in the Co(50)773-923 samples but consistent with
that of the γ-Mo2N (111) phase. The amorphous γ-Mo2N was
slightly formed, together with Co3Mo3N in the Co(50)973
sample.
23
the peaks of Co-Mo oxynitrides (37.0, 43.0, and 63.0°).
Therefore, Co(25)973 contained γ-Mo2N and probably amor-
phous and small particles of Co-Mo oxynitrides. Co3Mo3N had
characteristic peaks of 2θ ) 40.0, 42.7, and 46.4° for Co(50)-
973 (Figure 3d). Since the Co(0-25)973 samples contained
γ-Mo2N and MoO2 which were contained in the Co(50)973
sample, the presence of a small amount of cobalt not only
prevented the formation of Co3Mo3N but also prevented
nitriding of the molybdenum oxides. For Co(35)973 and Co-
(65)973, small amounts of γ-Mo2N and Co metal were observed
together with Co3Mo3N, respectively, as shown in Figure 3c,e.
For the Co(90,100)973 samples, Co metal was only formed.
On the other hand, CoMoO4 was only formed for Co(50)723
in Figure 3g and formed together with MoO3 for Co(10-35)-
723 and with CoO for Co(65)723. From the results, Co3Mo3N
The profile of a temperature-programmed reaction (TPR) of
the Co(50)723 sample, upon heating from 573 to 973 K at a
-
1
rate of 1 K min under a stream of excess NH3, is shown in
Figure 2. The formation peaks of H2 and N2 were observed
above 780 K, while the formation peaks of H2O were observed
at about 750 and 900 K and the former peak of H2O was larger
than the latter peak. Below 820 K, ammonia decomposed and
the hydrogen from ammonia was consumed for reduction of
CoMoO4 to desorb H2O and the nitrogen from ammonia-nitrided
CoMoO4 to form Co-Mo oxynitride. Thus, the broad XRD
patterns (in Figure 1a-c) showed the formation of Co-Mo

Cobalt Molybdenum Nitrides
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4087
did not consist of a single phase but contained the two phases
γ-Mo2N and Co3Mo3N.
High Surface Area. The BET surface area of the cobalt
molybdenum compounds nitrided at 973 K as a function of the
Co/(Co + Mo) ratio is shown in Figure 5. The surface area of
2
-1
Co(0)973 was 112 m g . The surface area of Co(10,25)973
2
-1
increased to 123 and 129 m g . The surface area of Co(50)-
2
-1
2
-1
9
73 was 12 m g , but that of Co(100)973 was only 4 m g .
Thus, the addition of cobalt at 10-25% of the samples increased
the high surface area but Co(50)973 had one-tenth the surface
area of Co(10-25)973. The distributions of the mesopores and
micropores in the samples are shown in Figure 6. The mesopores
at 4-5 nm and micropores at 0.4-0.6 nm for Co(25)973 were
larger than those for Co(0)973. Since the Co(0-25)973 samples
contained γ-Mo2N from the XRD analysis and had micropores
and mesopores with a high surface area, the addition of a small
amount of cobalt to the molybdenum compounds increased the
micropores of γ-Mo2N. For Co(35-65)973, the micropores at
0
.4-0.6 nm disappeared, and therefore, the Co(35-65)973
samples had mesopores without micropores. Moreover, we
found that the Co3Mo3N particles with a low crystallinity
4
1
increased the surface area of γ-Mo2N. The low-surface-area
sample without micropores was probably due to the growth of
the Co3Mo3N particle.
Oxidation States and Compositions of Bulk by XPS. The
XPS binding energies of Mo 3d and Co 2p3/2 for Co(0-100)-
9
73 were analyzed using XPS to determine the oxidation state
Figure 3. XRD patterns of cobalt-molybdenum nitrides and oxide
with various Co loadings prepared from the mixture of cobalt nitrate
and molybdenum heptamolybdate: (a) Co(0)973; (b) Co(25)973; (c)
Co(35)973; (d) Co(50)973; (e) Co(65)973; (f) Co(100)973; (g) Co-
and composition of cobalt molybdenum nitride. The representa-
tive spectra are shown in Figure 7. Co(0,50)973 had a large
peak of Mo centered at the binding energy of Mo 3d5/2 of 227.7
eV. Co(5-25,90)973 had a large peak of Mo located at the
binding energy of 228.4 eV, which was reported to be at 228.4
and 228.8 eV for γ-Mo2N. The Co 2p3/2 level of a zerovalent
Co was observed at 777.9 ( 0.5 eV (777.6 ( 0.5 eV37,38,42,43)
for all the samples nitrided at 973 K. The distribution of
molybdenum oxidation states in the Mo 3d spectra was
calculated as the individual Mo peak area divided by the total
Mo peak area, as shown in Figure 8A. Mo was distributed
from 25 to 52% for all the catalysts, and Mo was distributed
at more than 30% for Co(20-90)973. The Mo and Mo ions
were at 10-20% and Mo and Mo ions were present at less
than 10%. Thus, the Co(35-65)973 sample with Co3Mo3N
mainly consisted of Mo and Mo . Furthermore, from the XRD
and TEM analyses, γ-Mo2N was partially contained together
with Co3Mo3N in Co(50)973, as mentioned before. Since the
Co(50)973 sample included six types of Mo ions, these ions
were probably due to the mixture of Co3Mo3N and γ-Mo2N in
the samples. In Figure 8B, on the other hand, Co increased for
Co(5-25)973 and was relatively constant at 40% for Co(25-
0
2
+
(
50)723; (1) MoO
2
2 3 3
; (b) γ-Mo N; (2) Co Mo N; (9) Co metal; (4)
25
CoMoO
4
.
2
was formed from CoMoO4 prepared from a mixture of cobalt
nitrate and ammonium heptamolybdate. The cobalt molybdenum
sample with Co/(Co + Mo) ) 0.5 was analyzed using XRD
before and after nitriding the physical mixture of CoO and MoO3
at 973 K. Co3Mo3N did not have peaks of 2θ ) 40.0, 42.7, and
2
+
0
4
6.4°, but γ-Mo2N and Co metal were observed in the XRD
3+
4+
patterns (Figure 3a). Also, CoMoO4 was not formed but CoO
and MoO3 were formed in the oxidic precursor (Figure 3b).
CoMoO4 was not formed by oxidizing the physical mixture of
CoO and MoO3 at 723 K. Co3Mo3N was not produced until
CoMoO4 was formed in the sample. Therefore, Co3Mo3N was
not formed by nitriding the physical mixture of CoO and MoO3
but was formed by nitriding CoMoO4 in Co(35-65)973, which
was prepared using the aqueous solution of the mixture of cobalt
nitrate and ammonium heptamolybdate.
Morphology of C3Mo3N. The TEM images of Co(50)973
and Co(0)973 are shown in Figure 4. Molybdenum and cobalt
atoms were observed in the samples from the EDS analysis for
Figure 4A, while only molybdenum atoms are observed for
Figure 4B. From the image, the aggregation of large particles
5
+
6+
0
2+
0
2
+
9
5
0)973. The distribution of Co ion was almost constant at
0%. The Co and Co ions were distributed at 90% in the all
0
2+
3
+
samples. The amount of Co ion was much less than that of
Co and Co , except for Co(5-25)973. Thus, Mo , Mo , Co ,
and Co species were the most abundant species in the Co-
0
3+
0
2+
0
2
+
(a) was observed in all samples, while small particles (b) (ca.
(35-65)973 samples.
5
nm) were partially agglomerated. The particles (a) (20-30
nm) were as large as Co3Mo3N particles at Dc ) 36.3 nm from
the XRD results using Scherrer’s equation for the peak of Co-
The XPS binding energies of N 1s and Mo 3p for Co(0,25,-
50)973 are shown in Figure 9 and Table 1. One peak of N 1s
and four peaks of Mo 3p3/2 (Mo (δ ) 0-3), Mo , Mo ,
and Mo ) were deconvoluted in the range of 390-410 eV.
The binding energy of N 1s was observed at 397.1 ( 0.2 eV
(397.4 eV; dissolved N) in Co(0-25)973 but shifted to 396.2
(396.3) eV in Co(35-65)973, at which the binding energy was
close to 396.7 eV (Mo-N band ) or lower than those of the
Mo-N bond (397.1, 397.4 eV
δ+
4+
5+
(50)973 at 42.5°. Since the Co/Mo atomic ratio of the particles
6
+
39
(a) was about 1.0 from the EDS analysis, the particles (a)
consisted of Co3Mo3N. On the other hand, despite the Co/Mo
ratio of 0.3, the image of the particles (b) was similar to that of
the particles (c) in Co(0)973 (a small peak of γ-Mo2N in Figure
3
9
3
9
1d). These results indicated that the small particles (b) consisted
4
4,45
of γ-Mo2N with highly dispersed Co species. Thus, Co(50)973
) and Co-N bond (398.1

4088 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Hada et al.
Figure 4. TEM images of the (A) Co(50)973 and (B) Co(0)973 samples.
for the Mo-N bond, compared to that of the Co-N bond due
to transferring electrons to the nitrogen atoms from cobalt. In
Table 1, the ratios of N to Mo, calculated from the XPS data,
were constant at 0.33-0.55 in Co(5-65)973, although the ratio
was 1.2 for Co(90)973. The N/Co ratio of Co(5)973 was 5 but
decreased with the cobalt content and reached the ratio of 0.3
for Co(65)973. The nitrogen atoms of cobalt molybdenum
nitride in Co(35-65)973 were most likely bonded to molyb-
denum atoms rather than to the cobalt atoms.
On the other hand, the formation of MoO2 was observed for
Co(25)973 by XRD analysis as mentioned before (Figure 3).
Since Co(50)973 contained a large amount of Co3Mo3N, the
composition of cobalt molybdenum nitride was Co:Mo:N ) 3:3:
1
.2 (Z2.3:2.3:1 from the XPS data; 2.5:2.5:1.0 from the
elemental analysis). Since the Co/Mo ratio of cobalt molybde-
num nitride in Co(50)973 was also 1.0 from the EDS analysis,
the composition of cobalt molybdenum nitride was Co3Mo3N1.2.
Cobalt molybdenum nitride in the bulk sample had a composi-
2
2
tion similar to Co3Mo3N obtained by Kim et al. and Jackson
Figure 5. BET surface area as a function of Co/(Co + Mo) ratio for
the 973 K nitrided cobalt-molybdenum.
25
31
et al. Also, Bem et al. reported similar compositions of Fe3-
Mo3N and Ni3Mo3N simulated by the Rietveld analysis method
of the XRD data. Fe3Mo3N was reported to consist of the
46
eV ). The binding energy of the N 1s peak in the Co-Mo-N
2
5,31
compounds was relatively close to the binding energy of N 1s
octahedral structure of NMo6.
The variety of the oxidation

Cobalt Molybdenum Nitrides
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4089
Figure 6. Pore size distribution of (A) mesopores and (B) micropores
for the 973 K nitrided cobalt molybdenum: (a) Co(0)973; (b) Co(25)-
9
73; (c) Co(50)973.
states of Co and Mo can be explained by the coexistence of
Co3Mo3N, γ-Mo2N, and Co metal (or Co nitride) from the XRD
results in the Co(35-90)973 samples. Co(25)973 contained
MoO2 and γ-Mo2N (probably overlapped with Co-Mo oxyni-
tride) in the XRD patterns (Figure 3). The presence of oxygen
compounds such as MoO2 and Co-Mo oxynitride probably
resulted in various Co and Mo oxidation states in the Co(5-
2
5) samples. Another possibility is that the deconvolution of
the Co and Mo spectra gave rise to some artifacts for obtaining
a good curve fitting of the spectra of Mo 3d5/2 and Mo 3d3/2.
Surface Composition Measured by XPS. The spectra of
the XPS Mo 3d, Co 2p, and N 1s were deconvoluted in order
to analyze the surface composition of the samples without argon
etching in the XPS analysis. The distribution of Mo and Co
ions on the surface of Co(0-100)973 is shown in Figure 10.
0
2+
The distribution of Mo and Mo was about 30% for Co(20-
3
+
4+
5+
6+
6
5)973. The Mo , Mo , Mo , and Mo ions were less than
1
5% for the Co(5-65)973 sample, except for the Co(90)973
sample. The cobalt-rich samples contained a high abundance
of Mo and Mo . The tendency of the Mo oxidation states
4
+
5+
on the surface was similar to those in the bulk of Co(35-65)
2+
samples. Furthermore, the distribution of Co was more than
0
3+
5
0% for all the samples, while Co and Co were present in
0
less than 40% amount. The distribution of Co was relatively
low on the surface, compared to the bulk. The Co oxidation
state on the surface became much higher than that in the bulk.
The surface of Co(35-65)973 probably consisted of Mo , Mo ,
and Co ions, which were the same species as in the bulk as
already discussed. In Table 2, the surface composition of Co-
Figure 7. XPS spectra of (A) Mo 3d and (B) Co 2p3/2 of bulk 973 K
nitrided cobalt molybdenum. For (A): (a) Co(0)973; (b) Co(25)973;
(c) Co(50)973; (d) Co(90)973. For (B): (a) Co(5)973; (b) Co(25)973;
0
2+
(c) Co(50)973; (d) Co(100)973.
2
+
Co(50)973 was Co3Mo3N3 (Co3Mo3N‚N2). From the results, the
surface of the samples contained a number of nitrogens with
molybdenum compared to the nitrogen content in the bulk.
High Surface Area and Bulk Composition. The relationship
between the surface area and the bulk Mo 3d/Co 2p ratio is
shown in Figure 11a, to determine the effect of the different
(
9
6
0-35)973 had the N/Co ratio of greater than 7 while Co(50)-
73 was 1.1. The N/Mo ratio of the surface was 1.0 for Co(0-
5)973 and 1.4 for Co(90)973. From the result, the surface
nitrogen content of Co(50)973 was 2.5 times greater than that
of Co3Mo3N in the bulk sample. The surface composition of

4090 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Hada et al.
Figure 9. XPS spectra of N 1s and Mo 3p3/2 in the bulk for the 973
K nitrided cobalt molybdenum: (a) Co(0)973; (b) Co(25)973; (c) Co-
(50)973; (d) Co(100)973.
TABLE 1: Atomic Composition of Cobalt Molybdenum
Nitride in Bulk with Various Co/(Co + Mo) Ratios
XPS/eV
Co/ N/
N/
elemental anal.
N 1sa Mo Mo Co compositionb composition
c
sample
Co(0)973
Co(5)973
397.1 0 0.69
396.8 0.10 0.50 5.12 Co0.2Mo1.4
Mo1.4
N
Mo1.2
Co0.07Mo1.3
Co0.2Mo1.4
Co0.33Mo1.3
Co0.42Mo1.25
CoMo1.9
N
N
N
N
N
Co(10)973 396.9 0.10 0.40 4.01 Co0.19Mo
2
N
Co(20)973 396.8 0.10 0.37 3.41 Co0.28Mo2.7
Co(25)973 396.9 0.10 0.33 3.09 Co0.3Mo
Co(35)973 396.2 0.31 0.54 1.77 Co0.59Mo1.9
N
N
3
N
N
N
Co(50)973 396.3 1.0 0.44 0.42 Co2.3Mo2.2
Co(65)973 396.3 1.4 0.39 0.26 Co3.6Mo2.5
Co(90)973 397.3 4.8 1.15 0.24 Co4.1Mo0.82
N
N
N
Co2.5Mo2.5
N
Co4.6Mo2.5
nm
N
d
Co(100)973 397.7
0.20 Co
5
N
Co20N
a
Binding energies of N 1s. b Atomic composition obtained by XPS.
c
Atomic composition derived from the N/(Co,Mo) ratio obtained by
elemental analysis. d Not measured.
Figure 8. (A) Distribution of the Mo 3d spectra in the bulk as a
function of Co/(Co + Mo) ratio for the 973 K nitrided cobalt
3+
was proportional to the content of Co ion but not on those of
0
2+
3+
4+
5+
molybdenum: (9) Mo ; (b) Mo ; (2) Mo ; (0) Mo ; (O) Mo
;
2+
0
Co and Co in the bulk. No cobalt species on the surface were
related to the surface area. From these results, the insertion of
Co ion into the bulk increases the distribution of Mo ion
in the bulk which led to the increase in the surface area of the
samples.
6+
(
4) Mo . (B) Distribution of the Co 2p spectra in the bulk as a function
of Co/(Co + Mo) ratio for the 973 K nitrided cobalt molybdenum:
3
+
2+
0
2+
3+
(
9) Co ; (b) Co ; and (0) Co .
molybdenum ions in the bulk samples on the surface area. Since
the surface area increased with the increasing Mo 3d/Co 2p ratio,
the high surface area of the Co(10-25)973 samples was
attributed to the Mo species in the bulk sample. However, the
surface area was not related to the Mo 3d/Co 2p atomic ratio
on the surface of the samples. In Figure 6, Co(0-25)973 had
Activity of Cobalt Molybdenum Nitrides for Thiophene
HDS. The rates of thiophene HDS on the various samples at
623 K with time on stream are shown in Figure 12. The Co-
(25)973 sample was the most active, but the Co(100)973 was
the least. The HDS rate of the low Co-containing nitride samples
(Co/(Co + Mo) ) 0-0.35) reached a steady state after 1 h on
stream, but the rates were kept constant for the high Co-
containing samples (Co/(Co + Mo) ) 0.5-1.0). The low Co-
containing samples had higher activity than the high Co-
containing samples on the basis of catalyst weight. The Co(25)
sulfide catalyst was less active than the Co(0-25)973 catalysts.
Concerning the HDS rate on surface area basis, the Co(35,50)-
973 catalysts had a considerably higher HDS rate than the other
catalysts at the initial stage and were more stable during the
reaction. Since the sulfided catalyst had very low surface area,
the sulfided Co-Mo catalyst exhibited higher HDS activity than
the nitrided Co-Mo catalysts on surface area basis. Table 3
summarized the amount of irreversible CO adsorption and the
catalytic data of the cobalt molybdenum nitride samples in
2
-1
surface areas (120-129 m g ) that contained γ-Mo2N. The
2
-1
surface area of Co(50)973 was 12 m g during the formation
of Co3Mo3N, which was one-tenth the surface area of γ-Mo2N.
Furthermore, Co3Mo3N in the Co(35-65)973 samples had
mesopores without micropores at 0.4-0.6. In the relationship
2
+
between the distribution of the surface area and the Mo ion
in the bulk (Figure 11b), the surface area was proportional to
the distribution of the Mo2 ion. However, the other Mo ions
in the bulk and on the surface were not related to the BET
surface area. The Co(5-25)973 samples mainly contained the
+
2+
Mo ion (about 50% in the distribution). Therefore, the surface
2
+
area of the samples depended on the Mo ion in the bulk.
Furthermore, the influence of the cobalt species in the bulk on
these surface areas (Figure 11c) indicated that the surface area

Cobalt Molybdenum Nitrides
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4091
Figure 11. Relationship between the BET surface area and (a) bulk
2+
3+
Mo 3d/Co 2p ratio, (b) the distribution of Mo and Mo in bulk, and
(
c) the distribution of Co3 in bulk.
+
Figure 10. (A) Distribution of the surface Mo 3d spectra as a function
of Co/(Co + Mo) ratio for the 973 K nitrided cobalt molybdenum:
0
2+
3+
4+
5+
6+
(
9) Mo ; (b) Mo ; (2) Mo ; (0) Mo ; (O) Mo ; (4) Mo . (B)
Distribution of the surface Co 2p spectra as a function of Co/(Co +
0
Mo) ratio for the 973 K nitrided cobalt molybdenum: (9) Co ; (b)
2
+
3+
Co ; (0) Co
.
TABLE 2: Atomic Composition of Cobalt Molybdenum
Nitride on the Surface with Various Co/(Co + Mo) Ratios
by XPS
sample
Co/Mo
N/Co
N/Mo
composition
Co(0)973
Co(5)973
0
0.92
1.00
0.95
0.93
1.04
1.00
1.02
1.01
1.4
Mo1.1N
0.05
0.08
0.10
0.10
0.11
0.97
1.37
3.58
18.7
11.9
6.75
11.3
8.7
1.05
0.68
0.39
0.35
Co0.04MoN
Co0.08Mo1.1
Co0.16MoN
Co0.08Mo0.96
Co0.13MoN
Co0.95Mo0.98
Co1.4MoN
Co(10)973
Co(20)973
Co(25)973
Co(35)973
Co(50)973
Co(65)973
Co(90)973
Co(100)973
N
Figure 12. HDS of thiophene on various cobalt molybdenum nitrides
with time on stream at 623 K: (9) Co(0)973; (0) Co(5)973; (b) Co-
(10)973; (O) Co(20)973; (2) Co(25)973; (4) Co(35)973; (1) Co(50)-
973; (3) Co(65)973; ([) Co(90)973; (]) Co(100)973; (+) Co(25).
N
N
Co2.7Mo0.7
Co2.9
N
N
γ-Mo2N (Co(0)973). Co3Mo3N was mainly formed in the Co-
(35,50)973 samples, although Co(65-100)973 exhibited am-
biguous TOFs due to very small CO adsorption. Therefore, the
formation of Co3Mo3N enhanced the activity of the catalysts
thiophene HDS. The low Co-containing samples exhibited high
surface area and large amounts of CO adsorption. The amount
of CO adsorption decreased with increasing Co content from
Co(0) to Co(65)973 and then was present constantly in only a
small amount. The samples of higher site densities for CO
adsorption were Co(0,5)973 with high Mo content and Co-
2
2
for thiophene HDS. Kim et al. reported that the HDS activity
of Co3Mo3N was 4 times higher than that of γ-Mo2N on a
2
7
surface area basis. Bussell and co-workers also reported that
the HDS activity of nitrided 6.7 wt % Co-12.8 wt % Mo/Al2O3
(Co/Co + (Mo) ) 0.48) catalysts was 4 times higher than that
(35,50)973 with Co3Mo3N. The TOFs of the low Co-containing
samples (Co(5-50)973) exhibited two times or higher TOF than

4092 J. Phys. Chem. B, Vol. 105, No. 19, 2001
Hada et al.
TABLE 3: Thiophene HDS Activity for Cobalt Molybdenum Nitrides
HDS rate/µmol min-
1
2
m
-2
C4 selectivityc
n-butane 1-butene/
n-butane 1-butene cis-2-butene trans-2-butene 2-butenes trans-2-butene
CO
uptake /
µmol g
c
a
(BET surface area /m g-1)
d
C4 product distribn /%
TOF /
min
-
1
initialb
steadyc
-1
sample
Co(0)973
Co(5)973
372
279
207
200
166
136
21
4.0
5.8
7.6
0.60 (112)
0.89 (99)
0.54 (123)
0.66 (129)
0.70 (124)
0.96 (60)
1.19 (12)
0.75 (20)
0.66 (10)
0.69(4)
0.68 (74)
0.75 (61)
0.69 (54)
1.94 (25)
1.19 (45)
0.99 (35)
1.12 (11)
0.43 (11)
0.17 (4)
0.19
0.32
0.32
0.43
0.54
0.42
0.68
3.76
1.14
0.35
1.88
60.4
57.3
54.7
53.3
54.1
46.7
37.2
25.8
18.4
9.8
10.5
10.1
11.6
11.3
13.7
17.2
20.1
26.3
13.0
13.6
14.8
15.3
14.6
16.8
18.7
22.7
23.3
16.8
18.5
20.4
19.8
20.0
22.8
26.9
31.4
32.0
2.03
1.79
1.55
1.52
1.56
1.18
0.82
0.48
0.33
0.58
0.57
Co(10)973
Co(20)973
Co(25)973
Co(35)973
Co(50)973
Co(65)973
Co(90)973
Co(100)973
0.495
0.586
0.565
0.601
0.639
0.640
0.822
e
e
e
e
0.11 (1)
21.6 (2)
nd
39.3
nd
15.0
nd
26.2
nd
19.5
f
Co(25)
25
15.8 (10)
0.65
0.769
a
Irreversible CO adsorption. After 10 min on stream. After 8 h on stream. d TOF (min-1) ) initial HDS rate (µmol min m-2) × BET
b
c
-1
2
-1
-1
e
f
2 2
surface area (m g )/CO uptake (µmol g ). Not detected. Sulfided at 623 K for 3 h under a stream of 10% H S/H .
Co3Mo3N had small micropores, some remaining cobalts were
crystallized with γ-Mo2N in the presence of hydrogen to form
Co3Mo3N in the bulk and led to the decrease in the surface
area.
The reaction products were n-butane, 1-butene, cis-2-butene,
and trans-2-butene with a very small amount of tetrahydro-
thiophene in thiophene HDS, as shown in Table 3. The
distribution of n-butane formation was 60% for Co(0)973 with
γ-Mo2N, 54% for Co(10-25)973, and not more than 50% for
Co(35-65)973 with Co3Mo3N. The hydrogenation selectivity
(butane/2-butene ratio) of Co3Mo3N was 2 times lower than
that of γ-Mo2‚N and the sulfided catalyst. For the isomerization
selectivity, 1-butene to trans-2-butene for all the samples slightly
differed from a thermal equilibrium value (0.46). From the
results, Co3Mo3N greatly desulfurized thiophene with lower
hydrogen consumption than γ-Mo2N. Therefore, cobalt molyb-
denum nitrides promise to be highly active catalysts for
hydroprocessing of fuel oils and are more active than cobalt
molybdenum sulfide on a catalyst weight basis.
Conclusions
Cobalt molybdenum nitride was formed by the nitriding of
CoMoO4 as a precursor via formation of Co-Mo oxynitride,
while was not formed by the nitriding of CoO and MoO3. The
structure of the nitride samples depended on the oxidic
precursors; that is, γ-Mo2N was formed from MoO3 in Co(0-
Figure 13. (A) XRD patterns after the reaction. (B) Pore size
distribution of mesopores and micropores for Co(25)973: (a) after the
reaction; (b) before the reaction.
2
5)723, Co3Mo3N from CoMoO4 in Co(35-65)723, and Co
metal and nitrides from CoO in Co(90-100)723. The TEM
images showed that Co(50)973 consisted of several phases (γ-
Mo2N and Co3Mo3N). Co3Mo3N contained predominantly cobalt
and molybdenum species. The BET surface area was related to
the bulk Mo 3d/Co 2p ratio, and furthermore, cobalt species
penetrated into the bulk and increased the surface area of
γ-Mo2N for the low cobalt contents (Co/(Co + Mo) ) 0.1-
of γ-Mo2N/Al2O3 on a O2 adsorption basis and found that the
incorporation of Co species in the Mo2N catalyst enhanced HDS
activity of the nitrided Co-Mo/Al2O3 catalyst.
The decrease in the BET surface area before and after the
reaction was observed in Table 3. The surface area of the Co-
(
0-35)973 samples extremely decreased after the reaction. The
0
.25). The cobalt molybdenum nitrides were more active and
surface area of the Co(25)973 sample decreased from 124 to
4
selective than cobalt molybdenum sulfide on a catalyst weight
basis for the HDS of thiophene with low hydrogenation
selectivity.
2
-1
5 m g (36% decrease) after the reaction. Especially, the
low-Co-containing samples exhibited great decrease in the
surface area after the reaction. However, the surface area of
Co(50)973 was unchanged before and after the reaction,
indicating that Co3Mo3N was very stable. Although Co3Mo3N
was not observed in Co(25)973 before the reaction (Figure 3b),
Co3Mo3N was formed after the HDS reaction (Figure 13A). The
increase in the activity of Co(25)973 on surface area basis (from
Acknowledgment. We are grateful to Mr. T. Watabe and
Ms. Y. Kondo (JEOL HIGH TECH, JEOL, Co., Ltd., Tokyo,
Japan) for TEM analysis.
References and Notes
-
1
-2
0
.70 to 1.19 µmol min m ) is probably due to the formation
(1) Markel, E. J.; Van Zee, J. W. J. Catal. 1990, 126, 643.
(2) Ozkan, U. S.; Zhang, L.; Clark, P. A. J. Catal. 1997, 172, 294.
(3) Nagai, M.; Miyao, T.; Tuboi, T. Catal. Lett. 1993, 18, 9.
of Co3Mo3N during the reaction. Moreover, the (left) micropores
and (right) mesopores for Co(25)973 in Figure 13B were formed
largely before the reaction but decreased after the reaction. Since
(4) Nagai, M.; Takada, J.; Omi, S. J. Phys. Chem. B 1999, 103, 10180.

Cobalt Molybdenum Nitrides
J. Phys. Chem. B, Vol. 105, No. 19, 2001 4093
5) Nagai, M.; Irisawa, A.; Omi, S. J. Phys. Chem. B 1998, 102, 7619.
(26) Li, Y.; Zhang, Y.; Raval, R.; Li, C.; Zhai, R.; Xin, Q. Catal. Lett.
1997, 48, 239.
(27) Logan, J. W.; Heiser, J. L.; McCrea, K. R.; Gates, B. D.; Bussell,
M. E. Catal. Lett. 1998, 56, 165.
(
(28) Kojima, R.; Aika, K. Chem. Lett. 2000, 514.
(
29) Ihm, S.-K.; Kim, D.-W.; Lee, D.-K. In Catalyst DeactiVation;
Bartholomew, C. H, Fuentes, G. A., Eds.; Elsevier Science: New York,
1997; p 343.
(30) Jacobsen, C. J. H. Chem. Commun. 2000, 1057.
(
31) Bem, D. S.; Gibson, C. P.; Loye, H.-C. Chem. Mater. 1993, 5,
3
97.
32) Chu, Y.; Wei, Z.; Yang, S.; Li, C.; Xin, Q.; Min, E. Appl. Catal.
A 1999, 176, 17.
33) Yu, C. C.; Ramanathan, S.; Sherif, F.; Oyama, S. T. J. Phys. Chem.
994, 98, 13038.
34) Quincy, B. R.; Houalla, M.; Proctor, A.; Hercules, D. M. J. Phys.
(
(
1
(
Chem. 1990, 94, 1520.
(
14) Nagai, M.; Miyao, T. Catal. Lett. 1992, 15, 105.
(35) Hada, K.; Nagai, M.; Omi, S. J. Phys Chem. B 2000, 104, 2090.
(36) Pawelec, B.; Navarro, R.; Fierro, J. L. G.; Vasudevan, P. T. Appl.
Catal. A 1998, 168, 205.
(37) Figueiredo, R. T.; Granados, M. L.; Fierro, J. L. G.; Vigas, L.;
Piscina, P. R.; Homs, N. Appl. Catal. A 1998, 170, 145.
(15) Choi, J.-G.; Brenner, J. R.; Colling, C. W.; Demczyk, B. G.;
Dunning, J. L.; Thompson, L. T. Catal. Today 1992, 15, 201.
16) Nagai, M.; Miyata. A.; Kusagaya, T.; Omi. S. In The Chemistry of
(
Transition metal Carbides and Nitrides; Oyama, S. T., Ed.; Blackie:
London, 1996; p 327.
(
9.
38) Zsoldos, Z.; Garin, F.; Hiaire, L.; Guczi, L. Catal. Lett. 1995, 33,
(17) Nagai, M.; Goto, Y.; Miyata, A.; Kiyoshi, M.; Hada, K.; Oshikawa,
3
K.; Omi, S. J. Catal. 1999, 182, 292.
(39) Nagai, M.; Omata, T.; Omi, S. Manuscript in preparation.
(40) Nagai, M.; Koizumi, K.; Omi, S. Catal. Today 1997, 35, 393.
(41) Hada, K.; Nagai, M.; Omi, S. Manuscript in preparation.
(42) Guczi, L.; Sundararajan, R.; Kopp a´ ny, Zs.; Zsoldos, Z.; Schay, Z.;
(18) Lee, K. S.; Abe, H.; Reimer, J. A.; Bell, A. T. J. Catal. 1993, 139,
3
4.
(
(
19) Colling, C. W.; Thompson, L. T. J. Catal. 1994, 146, 193.
20) Sajokowski, D. J.; Oyama, S. T. Appl. Catal. A.; Gen. 1996, 134,
Mizukami, F.; Niwa, S. J. Catal. 1997, 167, 482.
43) Okamoto, Y.; Nakano, H.; Imanaka, T.; Teranishi, S. Bull. Chem.
Soc. Jpn. 1975, 48, 1163.
44) Fix, R.; Gordon, R. G.; Hoffman, D. M. Thin Solid Films 1996,
3
39.
(
(21) Li, S.; Lee, J. S.; Hyeon, T.; Suslick, K. S. Appl. Catal. A 1999,
1
84, 1.
(
(
(
(
(
22) Kim, D.-W.; Lee, D.-K.; Ihm, S.-K. Catal. Lett. 1997, 43, 91.
2
88, 116.
23) Yu, C. C.; Ramanathan, S.; Oyama, S. T. J. Catal. 1998, 173, 1.
24) Ramanathan, S.; Yu, C. C.; Oyama, S. T. J. Catal. 1998, 173, 10.
25) Jackson, S. K.; Layland, R. C.; zur Loye, H.-C. J. Alloys Compd.
(45) Nakajima, T.; Shirasaki, T. J. Electrochem. Soc. 1997, 144, 2096.
(46) Milad, I. K.; Smith, K. J.; Philip C. W.; Mitchell, A. R., Catal.
1
999, 291, 94.
Lett. 1998, 52, 113.