Study of the preparation of bulk powder tungsten carbides by temperature programmed reaction with CH4 + H2 mixtures

Article abstract of DOI:10.1006/jcat.1996.0015

The synthesis of bulk tungsten carbides by carburization of W metal or of WO₃ with mixtures of CH₄ in hydrogen at various pressures has been studied in temperature programmed experiments. The resulting solids have been characterized by elemental analysis, X-ray diffraction, XPS analysis, and specific surface area measurements. The carburization occurs in two distinct steps: W₂C is formed in the first step taking place at about 650°C at atmospheric pressure with a 20% CH₄-H₂ mixture, while the formation of WC occurs only at higher temperatures. During carburization some free carbon is deposited, the importance of which is very much dependent on CH₄ partial pressure and on the temperature of carburization. It has also been shown that direct carburization of WO₃ by CH₄-H₂ does not take place, but that the carburization occurs via the reduction of WO₃ to W metal. The rate of reduction of WO₃ and that of carburization of W metal are very much dependent on, respectively, hydrogen partial pressure and CH₄ partial pressure. The extent of reduction of WO₃ into W metal required for carburization which takes place also depends on CH₄ partial pressure, indicating a competition between carburization of W metal at the surface and diffusion of W metal into the bulk of the Solid.

Full text of DOI:10.1006/jcat.1996.0015

JOURNAL OF CATALYSIS 158, 142–169 (1996)
ARTICLE NO. 0015
Study of the Preparation of Bulk Powder Tungsten Carbides by
Temperature Programmed Reaction with CH ϩ H Mixtures
4
2
G. Leclercq,* M. Kamal,* J. M. Giraudon,* P. Devassine,* L. Feigenbaum,* L. Leclercq,* A. Frennet,†
J. M. Bastin,† A. L o¨ fberg,† S. Decker,† and M. Dufour†
*
Laboratoire de Catalyse H e´ t e´ rog e` ne et Homog e` ne, URA CNRS 402, Universit e´ des Sciences et Technologies de Lille, 59655 Villeneuve d’Ascq
Cedex, France; and †Catalyse H e´ t e´ rog e` ne CP 243, Universit e´ Libre de Bruxelles, 1050 Bruxelles, Belgium
Received February 24, 1994; revised July 26, 1995; accepted August 25, 1995
needed to identify the factors which are of importance in
The synthesis of bulk tungsten carbides by carburization of controlling the surface composition and in particular the
W metal or of WO with mixtures of CH in hydrogen at surface carbon and amount of oxygen impurity. Two exper-
3
4
various pressures has been studied in temperature programmed
experiments. The resulting solids have been characterized by
elemental analysis, X-ray diffraction, XPS analysis, and specific
surface area measurements. The carburization occurs in two
imental approaches can be defined to solve this problem:
i) Synthesis of high purity bulk powder tungsten car-
bide catalysts. The aim here is to obtain well-defined struc-
tures with precisely-known surface compositions and with
reasonable surface areas of a few square meters per gram.
(ii) Synthesis of bulk carbides with high specific surface
(
distinct steps: W
at about 650؇C at atmospheric pressure with a 20% CH
mixture, while the formation of WC occurs only at higher
2
C is formed in the first step taking place
4
–H
2
temperatures. During carburization some free carbon is depos- areas and porous textures suitable for catalysis. These char-
ited, the importance of which is very much dependent on CH
acteristics are often accompanied by lower purity of sam-
partial pressure and on the temperature of carburization. It ples, suited for studying the effects of surface oxygen and
4
has also been shown that direct carburization of WO
CH –H does not take place, but that the carburization occurs
via the reduction of WO to W metal. The rate of reduction of
WO and that of carburization of W metal are very much
dependent on, respectively, hydrogen partial pressure and CH
partial pressure. The extent of reduction of WO into W metal
3
by
excess carbon on the catalytic properties.
4
2
We have mainly focused our research work on the first
approach (11). The second approach has been extensively
used by Boudart and his group, who have achieved mean-
ingful results on catalysts with specific surface areas be-
3
3
4
3
2
Ϫ
1
required for carburization which takes place also depends on tween 50 and 200 m g (12–14).
CH partial pressure, indicating a competition between carburi-
Various methods of preparing transition metal carbides
zation of W metal at the surface and diffusion of W metal into are listed in Table 1. In this paper, we have chosen to
4
the bulk of the solid.

1996 Academic Press, Inc.
investigate in detail the preparation steps of tungsten car-
bide from tungsten metal and tungsten trioxide, using CH4
as carbiding agent in a mixture with hydrogen. CH has
4
INTRODUCTION
been chosen as carbiding agent in order to reduce as much
as possible the formation of carbon deposits at the surface
of the carbide. It is well known that methane is the least
reactive of the alkanes, and accordingly requires the high-
est temperature for reaction, conditions where diffusion
of carbon in the solid is no longer the rate limiting step.
The objective of this paper is to optimize the conditions
of the preparation of tungsten carbides by varying the total
pressure, the ratio of the reactants, and the temperature
of carburization.
Transition metal carbides are expected to gain wide ap-
plication due to their unusual properties (1), in particular
in catalysis. Their effectiveness as catalysts was demon-
strated 20 years ago. For example, tungsten carbide was
shown to exhibit an interesting catalytic behavior in reac-
tions catalyzed by noble metals such as Pt and Pd (2).
Other investigations (3–5) have since tried to explore the
suitability of using highly dispersed tungsten carbides as
inexpensive alternative catalysts to Pt or other noble metals
in reactions such as hydrocarbon reforming (6–8) and hy-
drogenation of alkenes (4), and in the electro-oxidation of
hydrogen or methanol in fuel cells (9, 10).
EXPERIMENTAL
Tungsten Precursors
For a binary alloy such as WC, which is a rather complex
The precursors containing tungsten used in this study
heterogeneous catalyst, crucial synthesis experiments are were either WO (Fluka, high purity grade 99.9%, BET
3
1
42
0
021-9517/96 $12.00
Copyright  1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

PREPARATION OF TUNGSTEN CARBIDES
143
TABLE 1
General Methods of Synthesis of Transition Metal Carbides
Preparation method
Reaction
References
A. Solid–solid reaction
1) Direct reaction of metal or metal hydride with carbon black
xMe ϩ yC UUǞ Me
(
(
x
C
y
(1, 3, 11, 15)
(1, 3, 11, 15)
1
800 C
Њ
2) Direct reaction of metal oxide and carbon black
MeO ϩ C Ǟ MeC ϩ CO
B. Solid–gas reaction
—disproportionation of CO
Me ϩ CO Ǟ MeC ϩ CO
MeO ϩ CO Ǟ MeC ϩ CO
(
1) Reaction of metal or oxide with carbiding reagents
2
(11)
(16, 17)
2
—
—
with CO–CO
with hydrocarbons
2 2
and CO–H mixtures
(11)
Me ϩ C
MeO ϩ C
CH
Ex. ͱW
x
H
y
Ǟ MeC ϩ H
2
(18, 19)
(12–14)
x
H
y
2
Ǟ MeC ϩ CO ϩ H O
—
4
–H
N Ǟ ͱW
2
mixtures
(
(
2) Reaction of metal nitride with CH
3) Pyrolysis of an organometallic complex under hydrogen
4
–H
2
mixtures
2
2
C
(6)
(20)
(21)
Ex. 1 Cp
Ex. 2 Cp
2
W
2
(CO)
4
(dmad)a
(dmad)
2
Mo (CO)
2
4
C. Gas–gas reaction
WCl ϩ C
Mo(CO) ϩ CO (or H
WO ϩ 3C Ǟ WC ϩ 2CO
2MoO ϩ 7C Ǟ Mo C ϩ 6CO
(
1) Reaction of metal chloride or metal carbonyl with hydrocarbons,
hydrogen, or CO by chemical vapor deposition
2) Reaction of metal oxide vapor with gasified carbon (under
vacuum)
6
x
H
y
ϩ H
2
Ǟ WC ϩ HCl ϩ C
m n
H
(
15)
6
2
) Ǟ Mo C ϩ CO ϩ CO ϩ H O
2
2 2
(
2
(
22, 23)
3
2
a
Cp is cyclopentadienyl and dmad is dimethylacetylenedicarboxylate.
2
Ϫ
1
surface area about 3 m g , mixture of monoclinic and to Ϫ196ЊC in liquid nitrogen. The sample was then submit-
orthorombic crystallographic phases) or WO prepared ted to a mixture N (30%)–He, and the cell was warmed
3
2
in our laboratory by ammonium metatungstate to room temperature. The amount of nitrogen desorbed
(NH ) H W O , xH O, Fluka, high purity grade 97%) was measured with a thermoconductivity detector.
(
4
6
2
12 40
2
decomposition (24).
X-ray photoelectron spectroscopy. XPS measurements
were carried out with an AEI ES 200B (Bruker) spectrom-
eter equipped with an Al anode (h␯ ϭ 1486.6 eV, 300 W).
Surface and Bulk Characterization
X-ray diffraction. The bulk structure of the materials The binding energies (BE) were determined using the C 1s
was determined by X-ray diffraction using either a Philips peak of contamination (BE ϭ 285 eV) as a reference.
(
Norelco PW 1051) apparatus (CuKͰ radiation ␭ ϭ 1.54178
˚
A, Ni filter) or a diffractometer (Siemens D 5000) equipped
with a specially designed in situ X-ray diffraction cell. A
chosen gas mixture was passed through the platinum cell
while the XRD patterns were collected using the diffracto-
meter. These XRD patterns were assigned using the
JCPDS data base (Joint Committee on Powder Diffrac-
tion Standards).
Reduction and Carburization of the Precursors
Two different setups were used for reduction experi-
ments.
Temperature programmed reduction. The first setup
(
called apparatus TPR hereafter) employed a TCD as de-
tector for hydrogen in a flow of H –Ar mixture which
2
Chemical analysis. Elemental analyses were per- contained initially 30% H . First the sample (about 60 mg
2
formed at the ‘‘Service Central de Microanalyse’’ of the of WO ) was dried in a flow of nitrogen at 500ЊC for a few
3
CNRS (Vernaison, France) by coulometry for C and O, hours and cooled to room temperature. Then nitrogen was
Ϫ
1
and by plasma emission spectroscopy for W after fusion replaced by the reducing mixture (flow rate 2.6 liter h ).
of the samples in NaOH ϩ KNO3 .
Water formed during reduction was trapped by a 13X
molecular sieve at Ϫ196ЊC. The data from the thermocou-
ple and from the TCD were collected and processed by a
Hewlett Packard HP 85 microcomputer.
Surface areas. Surface areas were measured with a
sorptometer (Quantasorb Jr.), using the single point BET
surface area determination. Samples (200 to 300 mg) were
placed in a cell and heated with nitrogen to 150ЊC. After
Thermogravimetry (TPG). The second setup em-
outgassing, the cell was cooled to room temperature, then ployed an electronic microbalance (Sartorius GmbH) to

144
LECLERCQ ET AL.
follow the weight loss when samples were heated (60ЊC working at atmospheric pressure with partial pressures of
Ϫ
1
Ϫ
6
h ) in vacuum (10 Torr with an oil diffusion pump). (1 reactive gases ranging from 0.02 to 0.3 atm and argon as
Ϫ
2
Torr ϭ 133.3 N m .)
balance gas. The analysis of the gas composition at the
outlet of the reactor was performed by a quadrupole mass
spectrometer (QMG-420-5 from Balzers) equipped with
a GEV-010 (Balzers) inlet device. Gas flow rates were
measured and regulated by Rosemount Sho-Rate 150
flowmeters equipped with needle valves. Typical flow rates
Preparation of Carbides by Temperature Programmed
(
TP) Experiments
Three different systems were used according to the total
pressure requirements, and will be referred to as APR1,
APR2, and LPR. Two of them were operated at atmo-
spheric pressure (APR1 and APR2) and the third at lower
pressures (LPR).
Ϫ
1
ranged from 1.2 to 5 liter h . A Eurotherm 818 controller
using chromel-alumel thermocouples was used for measur-
ing and controlling the reactor temperature.
Quadstar 3.0 software (Balzers) running on an IBM-
type computer was used for controlling the mass spectrom-
eter and for collecting temperature data. Samples of 40 to
Atmospheric pressure reactor No. 1 (APR1). This reac-
tor, located at Villeneuve d’Ascq, was a flow reactor, which
was used previously in catalytic studies (25). For the pres-
ent carburization experiments, the flow rates were regu-
lated by membrane flow regulators (Brookes 8802) and
measured by soap bubble flowmeters. The reactor was
about 10-cm long and 2-cm wide. A thermocouple well
made it possible for the temperature of the sample to be
measured. The size of the sample was usually around 3 g
of WO , but, in some cases, was up to 12 g. The gas flow
rate of the carburizing mixture was about 8–10 liter h .
The heating rate in the TP experiments was 50ЊC h . When
reduction preceded carburization it was carried out at
00ЊC for 10 h after heating from room temperature at a
rate of 50ЊC h . Before each experiment the sample was
pretreated in a nitrogen flow (about 10 liter h ) at 550ЊC
and then cooled to room temperature. After carburization
the sample was cooled to room temperature in the CH –H
mixture (unless mentioned otherwise), swept by a nitrogen
flow (10 liter h ) until no hydrogen was detected by chro-
matography, and finally passivated in an O –N mixture
1.5% O ) at a flow rate of 10 liter h for at least 10 h
in order to avoid bulk oxidation when the carbide was
contacted with air.
1
1
50 mg of WO were placed in a quartz reactor (volume
–3 ml) connected to the apparatus through Cajon adap-
3
tors. They were then purged in pure argon at room temper-
ature until the mass signals typical of air had returned to
the baseline. The sample was then dried at 500ЊC in flowing
argon. Meanwhile the mixture of the reactants (CH and
4
hydrogen) was prepared in the bypass. After cooling the
sample to room temperature, this mixture was switched
to the reactor and, after a few minutes, the temperature
programming was started (typically a linear ramp of 180
3
Ϫ
1
Ϫ
1
Ϫ
1
to 360ЊC h ).
6
The gases used in this apparatus were argon N 50, 10.0%
H in argon ‘‘certified standard,’’ 10.0% CH in argon ‘‘cer-
Ϫ
1
2
4
Ϫ
1
tified standard’’ from UCAR, and 10.0% O in argon
2
‘
‘alpha standard’’ from BOC for experiments in argon or
with reactant pressure lower than 0.1 atm. Pure hydrogen
N 50 from BOC) diluted in argon was used for hydrogen
4
2
(
Ϫ
1
partial pressures between 0.1 and 0.23 atm.
2
2
Ϫ
1
Low pressure reactor (LPR). The following experi-
ments were performed on this apparatus which was located
in Brussels.
(
2
The composition of the gas mixture at the outlet of the
(i) Determination of total surface areas, obtained
reactor was periodically measured by gas chromatography through krypton adsorption measurements and evaluation
using a Shimadzu chromatograph GC9A equipped with using the BET and Kaganer (26) equations. It has been
both TCD and flame ionization detectors (FID). Products shown (27) that, when applied in their respective appro-
were separated at 90ЊC in a column (3-m long and 1/8- priate coverage ranges, these two methods provide values
inch wide) filled with ‘‘Carbosphere’’ molecular sieve. Peak in excellent agreement.
areas were measured with an integrator (Spectra Physics
270).
(ii) Reduction or carburization reactions in temperature
programmed conditions with total pressures lower than
4
Hydrogen (Air Liquide—U quality) was purified by 10 Torr.
flowing through a Pd catalyst (Deoxo) and through a 13X
molecular sieve. Air (Air Liquide) was purified by a 13X
molecular sieve and nitrogen (Air Liquide—U quality)
was passed through copper heated at 300ЊC in order to
remove oxygen, and then through a 13X molecular sieve.
Methane (Air Liquide—N 30 quality, 99.9%) was purified
by passing through a 4A molecular sieve.
The reaction vessel was built completely of commercial
stainless steel elements, connected through 2.75 inch Con-
flant flanges, with the exception of the reactor itself (made
of quartz) and the copper and Viton seals and gaskets.
The high vacuum in this reaction vessel was obtained
through an oil diffusion pump, while an ion pump enabled
Ϫ
8
the sample to be kept at pressure below 10 Torr between
Atmospheric pressure reactor No. 2 (APR2). This ap- different experiments. The pressure was measured during
paratus, located in Brussels, was a classical flow reactor experiments by a membrane capacitance micromanometer

PREPARATION OF TUNGSTEN CARBIDES
145
3 2
FIG. 1. TPR of WO in 30% H
–Ar (300ЊC hϪ1, 64 mg). TCD signal (in arbitrary units) versus temperature.
Ϫ
4
(
Barocel from Datametric) within the range 10 to 10 it was submitted to a flow of a 20% CH –D mixture at 5
4 2
Ϫ
1
Torr, the UHV pressure when working with the ion pump or 0.5 Torr at a heating rate of 360ЊC h from 25 to
was measured with a VG Bayard-Alpert gauge. 1200ЊC, where it was left at 1200ЊC for 35 min. During this
The introduction of gases into the apparatus was con- experiment the peaks at mass 2, 3, 12, 15 to 20, 28, and 44
trolled by two Phillips-Granville ‘‘type C ultra-high vac- were recorded at various times of reaction.
uum’’ valves while the outlet flow rate was determined by
fritted disks. The apparatus was equipped with two such
RESULTS AND DISCUSSION
fritted disks, each of them having different values of vol-
Ϫ
1
ume flow rate (14 and 55 ml s for H ). The advantage
2
1
. Reduction of Tungsten Trioxide
of working with these fritted disks is that in the whole
Ϫ
4
working pressure range (10 to 10 Torr) their volume
flow rate is only dependent on the nature of the flowing
molecules and is independent of the pressure (molecular
1
.1. Reduction by Hydrogen at a Pressure of 0.3 atm
In a first experiment, TPR (300ЊC) of WO (Fluka, 64
3
flow conditions). A LMD 100 manipulator (VG) allowed mg) was studied in apparatus TPR in a H (30%)–Ar flow
2
Ϫ
1
the insertion of the most appropriate fritted disk for the ex- (2.6 liter h ) at atmospheric pressure. The variation of H
2
periment.
consumption, as measured by the TCD signal, as a function
During experiments conducted in flow conditions the of the temperature (Fig. 1) shows that the reduction of
nature of the gaseous molecules could be determined using WO into W takes place in at least three steps. The hydro-
3
a GD 150 mass spectrometer from Atlas Werk.
gen consumption starts at about 480ЊC, then there is a
The gases used were the following: H N 55 and Kr N shoulder at about 640ЊC which may indicate the formation
2
4
5 from BOC, and D N 28, O N 50, and CH N 55 of a metastable oxidic form of stoichiometry between WO3
2
2
4
from UCAR.
and WO . The hydrogen consumption for this shoulder is
2
A sample of 74 mg WO (Fluka) was calcined in a flow about 9.3% of the total consumption which would corre-
3
of oxygen at 8 Torr from 25 to 1000ЊC. Then, after cooling, spond to the reduction of WO into W O or W O ,
3
18 49
20 58

146
LECLERCQ ET AL.
Ϫ
1
FIG. 2. Reduction of WO
3 2 2
in pure H at atmospheric pressure. Area of H chromatographic peak versus time. Heated at 50ЊC h up to 600ЊC,
then at constant temperature. m ϭ 3.05 g.
the formation of which has already been reported (28). A after drying at 550ЊC in a flow of nitrogen, was heated in
Ϫ
1
first maximum of hydrogen consumption is observed at a flow of hydrogen (9.6 liter h ) up to 600ЊC where it
about 670ЊC. The uptake made up of these two features stayed for 10 h. The composition of the gas flowing out of
(
640 and 670ЊC) is about one half of that of the second the reactor was determined by GC in order to follow the
peak at 800ЊC. Hence the first peak probably corresponds variations of the hydrogen partial pressure (which is pro-
to the formation of WO . In the final step which starts at portional to the area of the hydrogen peaks in GC). Figure
2
about 710ЊC and leads to the second peak, the maximum 2 shows the result. It is not possible to determine precisely
reduction rate being at 800ЊC, WO is reduced into W when hydrogen starts to be consumed, but certainly it is
2
metal. These results are in good qualitative agreement with being consumed at 400ЊC. Hence, under 1 atm of hydrogen,
those of Vermaire and van Berge (29) and of Gr u¨ nert et the temperature where WO reduction starts is noticeably
3
al. (30) who, when studying the reduction of WO by XPS, lowered compared to the previous experiment (480ЊC)
3
with PH2 ϭ 0.3 atm.
have shown that it proceeds through intermediate oxida-
tion states W , W , W , and finally W .
5
ϩ
4
ϩ
2
ϩ
0
Knowing the pressure of hydrogen in the exit gas and its
inlet flow rate we have calculated the amount of hydrogen
consumed at various times (Fig. 3). At the end of the
reduction the total amount of hydrogen consumed corres-
1.2. Reduction by Hydrogen at a Pressure of 1 atm
(
Apparatus APR1)
ponds to about 3.2 mol H per mol W which shows, consid-
2
The above study has shown that, under our conditions ering the margin of error of the measurements, that WO3
of TPR, WO is completely reduced by hydrogen at 850ЊC. is completely reduced into W at 600ЊC by hydrogen at 1
3
However, at such a temperature W metal is likely to be atm. The steps of reduction of WO cannot be clearly
3
severely sintered. Consequently, we have checked whether distinguished in this experiment, but probably the maxi-
WO could be completely reduced in pure hydrogen at mum of hydrogen consumption at 600ЊC (minimum in Fig.
3
atmospheric pressure in the flow reactor used for carburiza- 2) corresponds to the maximum rate of reduction of WO3
Ϫ
2
tion (APR1). A sample of WO (Fluka, 1.32 ϫ 10 mol), into WO . The first step corresponding to the formation
3
2

PREPARATION OF TUNGSTEN CARBIDES
147
3
FIG. 3. Hydrogen consumption as a function of time during reduction of WO .
3 2
FIG. 4. Reduction of WO in a mixture of 30% H
in Ar at atmospheric pressure heated at rate of 180ЊC hϪ1. 74 mg.

148
LECLERCQ ET AL.
TABLE 2
increase of PH2 from 0.3 to 1 atm has evidently increased
the reduction rate.
2
TPR Results under Various H Pressures
Temperatures (ЊC)
1.3. Reduction by Hydrogen at a Pressure of 0.1 atm
(
Apparatus APR2)
Maximum rate of
formation
WO (Fluka, 76 mg) was reduced in a flowing mixture
3
Heating rate
Beginning
of reduction WO
Ϫ
1
of 10% H in argon (P ϭ 0.1 atm) with a heating rate of
P
H2 (Torr)
(ЊC h
)
3
Ϫ
x
WO
2
W
2
1
H2
Ϫ
360ЊC h from 600 to 900ЊC in apparatus APR2. The
7
2
60
30
50
150
360
180
300–400
480
600
ء
600
670
ء
extent of reduction has been monitored using the height
640
730
ء
800
of the H O peak at mass 18 by mass spectroscopy. The
2
7
6
800 Ͼ900
Ͼ1000
results are given in Fig. 4 where we have plotted the hydro-
gen pressure and the temperature versus reduction time.
Reduction starts only at around 600ЊC and clearly occurs
in three steps. The position of the maxima are reported in
Table 2. The third step starts at about 850ЊC but the posi-
1
700–750
ء
*
Not determined.
of a suboxide cannot be distinguished from the second tion of the third peak maximum has not been determined
one. The reduction of WO into W metal probably mainly since the experiment was stopped at 900ЊC.
2
occurs during the almost constant H consumption in the
It is evident that lowering the hydrogen pressure results
2
isothermal stage at 600ЊC. The heating rate was three times in a considerable decrease in the rate of reduction of WO3
lower in this experiment than in the previous one. The into WO3Ϫx (W O or W O ), WO , or W metal, and
18
49
20 58
2
separation of the steps WO Ǟ WO and WO Ǟ W is consequently better separates the three steps of reduction.
3
2
2
accordingly worse than that obtained with a lower H pres- This effect of the hydrogen pressure indicates clearly that
2
sure. In Fig. 3 the constant H consumption starts at a the rate limiting step of reduction is the surface reaction
2
ratio n /n of about 1.7, showing that, at 1 atm, the reduc- with hydrogen and not the diffusion rate of various species
H
W
2
tion of WO into W metal starts much before the total (probably W metal atoms) in the solid which should be
2
reduction of WO into WO . Taken overall, however, an independent of hydrogen pressure. Consequently, the com-
3
2
2
FIG. 5. Domains of existence for W, W C, WC pure, and with free carbon.

PREPARATION OF TUNGSTEN CARBIDES
149
2 4
FIG. 6. Changes of the areas of H and CH chromatographic peaks versus time during temperature programmed carburization of W metal
Ϫ
1
(
4 2
50ЊC h , 2.60 g W, 20% CH –H ).
position of the solid is likely to be approximately uniform with carbon. Clearly, for the preparation of carbides free
within the bulk of the solid whatever the extent of re- of carbon, it is better to use a carburizing mixture with a
duction.
low molar fraction of methane. On the other hand, for very
low CH contents, the domains of existence of tungsten
4
2
. Carburization of Tungsten Metal
carbides are shifted toward higher temperatures and carbu-
rizing at too high a temperature could lead to solids with
very low surface areas. Consequently we have decided to
start this study of the carburization of W using a mixture
2
.1. Thermodynamics of the System CH /W/C
4
Three reactions are likely to occur during carburization
of W:
containing 20% CH in hydrogen with the same composi-
4
tion as the one often used by Boudart and co-workers
(12–14, 32).
2
W ϩ CH ϭ W C ϩ 2H
[1]
[2]
[3]
4
2
2
W C ϩ CH ϭ 2WC ϩ 2H
2
4
2
2
.2. Carburization of Reduced WO at 1 atm (Apparatus
3
CH ϭ C ϩ 2H
4
2
APR1)
If a is the molar fraction of methane, then for each equilib-
2.2.1 Carburization with the mixture 20% CH –H :
4
2
2
rium the equilibrium constant is K ϭ (1 Ϫ a) /a if the WCR1. The carburization of WO (Fluka) reduced in
i
3
total pressure is 1 atm. Using the thermodynamic data of hydrogen into W metal as described in the Experimental
Barin and Knacke (31) we have calculated the values of a section was performed by temperature programmed reac-
Ϫ
1
corresponding to each equilibrium at various temperatures tion (13) at atmospheric pressure in a flow (10.2 liter h )
and at 1 atm total atmospheric pressure. The results are of the mixture 20% CH –H increasing the temperature
4
2
Ϫ
1
reported in Fig. 5 where we can see the domains of exis- up to 800ЊC (50ЊC h ) where the sample was kept for 10 h.
tence for W, W C, WC free of carbon or contaminated This sample was passivated according to the procedure
2

150
LECLERCQ ET AL.
FIG. 7. Fraction of CH
4
consumed (X/N
0
4 2
) as a function of time during temperature programmed carburization of W metal in 20% CH –H .
reported in the Experimental section. This solid will be consumed at a given state of reaction to the initial one
called WCR1. (here corresponding to 20% CH ), as a function of the
4
In Fig. 6, the areas of the chromatographic peaks (in areas of CH peaks and of a, the initial ratio of the molar
4
arbitrary units) of CH and of H , which are proportional fractions of hydrogen and of methane (here a ϭ 4).
4
2
to their partial pressures, are plotted versus time of reac-
tion and temperature. At temperatures lower than 600ЊC
a slight decrease in the methane partial pressure associated
with a corresponding increase in the hydrogen partial pres-
sure is observed. Then the composition of the mixture
stabilizes at approximately 250ЊC. Methane consumption
starts at about 600ЊC and exhibits two peaks respectively
centered at 660 and 720ЊC. At 800ЊC the percentage of
methane in the mixture reaches a value close to that ob-
served between 250 and 600ЊC, but seemingly slightly
lower.
X
(a ϩ 1)[S(CH ) Ϫ S(CH )]
4 0 4
ϭ
N0 (a ϩ 1)[S(CH ) ] ϩ S(CH )
4 0
4
S(CH ) is the area of CH peak at a given time and S(CH )
4 0
is the initial area of the CH peak. By integrating the curve
using the trapezium method), X/N versus time (Fig. 7),
we obtain the amount of CH consumed as a function of
4
4
4
(
0
4
time of carburization (Fig. 8).
The curve in Fig. 7 shows three regimes, the first starting
at about 76ЊC, the second at 630ЊC (corresponding to the
first peak in Fig. 6), and the third at 690ЊC. The two waves
When methane is consumed the hydrogen partial pres-
sure increases, first, because of the lowering of methane
partial pressure and second, because of hydrogen produc-
at 630 and 690ЊC correspond to real CH consumption.
4
tion during carbon consumption. To quantify the methane However, the first feature probably does not correspond
consumption, it is necessary to establish the relationship to CH consumption; it is very likely to be a simple artefact
4
between the number of CH molecules and the CH pres- due to different changes of hydrogen and of methane flow
4
4
sure (proportional to the CH signal in chromatography). rates as a result of changes in the pressure drop due to
4
In a very simple reasoning (see Appendix I) we have calcu- the presence of the catalyst when the temperature in-
lated the ratio X/N of the molar flow rate of methane creases. This results in changes of the mixture composition
0

PREPARATION OF TUNGSTEN CARBIDES
151
FIG. 8. Number of CH
4
4 2
moles consumed versus time during temperature programmed carburization of W metal in 20% CH –H .
and in a drift of the baseline. To support this view, let us of the exhausting of the W. For the first step, N ϭ 7.1 ϫ
1
Ϫ
3
mention that if this first regime is considered as resulting 10 mol C and N /W ϭ 0.5, and for the second step, N ϭ
from CH consumption the total amount of CH used in 1.01 ϫ 10 mol C and N /W ϭ 0.72. Hence we can conclude
this carburization would be 3.41 ϫ 10 mol CH . Compari- that in the first step of carburization one obtains essentially
1
2
Ϫ
2
4
4
2
Ϫ
2
4
son of this value to the amount of W in the sample leads W C, whereas WC is noticeably formed only at tempera-
2
to a ratio CH /W of the number of molecules of CH
tures higher than 690ЊC. The C/W ratio of 1.2 indicates
4
4
leading to C deposit to that of atoms of W of 2.2. Chemical that some carbon deposit at the carbide surface occurs
analysis gives a ratio C/W in the final sample of 1.2. Hence during the carburization of W at 800ЊC. The surface area
the integrated curve of Fig. 8 has been obtained by consid- of the final solid after passivation, measured by N phy-
2
2
Ϫ
1
ering that CH consumption starts only at 630ЊC. In such sisorption, is 8.6 m g . The X-ray diffraction pattern of
4
a case the final value of the ratio CH /W is 1.22, which is this carbide WCR1 exhibits the main lines characteristic
4
in good agreement with chemical analysis. Therefore, our of the simple hexagonal structure of WC.
methodology enables us to calculate the number of CH4
moles used for the carburization of W and for carbon product are reported in Fig. 9. The W 4f signal exhibits
deposit resulting from CH decomposition. the doublet W 4f7/2,5/2 at binding energies of 32.2 eV (Ϯ0.2
The W 4f, C 1s, and O 1s XPS spectra of the passivated
4
Carburization has been shown to occur in two very dis- eV) for W 4f7/2 and at 34.2 (Ϯ0.2 eV) for W 4f5/2 character-
tinct steps, one starting at about 630ЊC and the other at istic of tungsten carbide (11, 32). The left side of the peak
6
ϩ
6
90–700ЊC. We have tried to estimate the number of moles indicates the presence of some oxide (W ) at about 38
of CH (N and N ) used in each of these two steps, taking eV for W 4f5/2
.
4
1
2
into account the fact that at the minimum (at 690ЊC) be-
The decomposition of the photopeak C 1s using a
tween the two peaks of the curve X/N vs time (Fig. 7) FWHM of 1.7 eV typical of photopeaks of C 1s level (34)
0
the second step of carburization has already started while shows the participation of three components. The first at
the first is not yet finished due to its rate decreasing because lower binding energy (283.3 eV) is specific for carbon in

152
LECLERCQ ET AL.
FIG. 9. XPS analysis of WC (WCR1) prepared by temperature programmed carburization of W metal up to 800ЊC in 20% CH
photopeak; (b) C 1s photopeak; (c) O 1s photopeak.
4 2
–H . (a) W 4f
a transition metal carbide (35). The highest peak at 285 eV). Binding energies near this value (285.9 (35), 286.4
eV is assigned to polymeric or free carbon and is a result (34), and 286.7 (36)) have been reported in the literature
partly of oil contamination in the spectrophotometer but for carbon atoms linked to oxygen atoms by single bonding
mainly of carbon deposit arising from CH decomposition. in poly-ether-ether-ketones. Hence this third type of car-
4
The third component is at higher binding energy (286.6 bon can be attributed to surface carbon atoms bound to

PREPARATION OF TUNGSTEN CARBIDES
153
FIG. 9—Continued
oxygen of passivation. This last assignment is supported of W metal in a CH –H mixture containing 10% of meth-
4
2
by the shape of the O 1s signal which is very wide (half ane. The characteristics of this preparation are reported
width about 4 eV) and is composed of at least two compo- in Table 4. This sample was cooled from 800ЊC to room
nents: there is one at 531.6 eV corresponding to transition temperature in pure hydrogen. It is named WCR2. The
metal oxides (37), while the other, centered at about 533 variations of CH partial pressure are reported in Fig. 10.
4
eV, could be attributed to oxygen in water (533 eV (38)) CH consumption starts at 685ЊC (a temperature that is
4
or hydroxyl groups (532.6 eV (39)) or to oxygen singly higher than when 20% CH –H was used), steeply increases
4
2
linked to carbon 533.8 eV (33) or 534.0 eV (34).
XPS results clearly show the formation of tungsten car-
bide. However, after carburization at 800ЊC, the surface
of WC is covered with polymeric carbon and part of the
carbide surface has been oxidized during passivation. This
excess carbon and oxygen will have to be removed before
catalytic reactions in order to characterize clean carbide
surfaces.
and approximately stabilizes from 740 to 780ЊC, and
slightly increases again up to 800ЊC where temperature
was kept constant. The integration of the curve X/N versus
0
time of reaction leads to the curves in Fig. 11 where the
ratio of CH /W of the number of moles of CH consumed
4
4
to that of W in the sample is plotted versus time and
temperature. The value of CH /W corresponding to the
4
end of the plateau of PCH4 (780–785ЊC) is 0.5, which indi-
The composition of the surface layers analyzed by XPS
cates that, at this stage, W C is likely to be formed; at the
2
(
about 5 nm in depth) is reported in Table 3. The C/W
end of carburization at 800ЊC the ratio CH /W is 0.94.
4
and O/W ratios determined by XPS analysis are definitely
higher than those measured by chemical analysis. This
shows that most of the ‘‘free’’ carbon and oxide phases
are localized at the surface of the carbide.
Chemical analysis gives the formula WC0.91O0.15 for this
carbide WCR2, which is in good agreement with the previ-
ous result.
The X-ray diffraction pattern is characteristic of WC.
2
Ϫ
1
2
.2.2. Carburization of reduced WO with the mixture The specific surface area (8 m g ) is on the same order
3
1
0% CH –H : WCR2. The thermodynamics of the WC/ of magnitude as that of WCR1. The comparison of these
4
2
CH /C system predict an increase in the equilibrium tem- results with those of the previous carburization in the mix-
4
peratures for reactions [1–3] when the CH percentage ture 20% CH –H shows that, when the percentage of CH
4
4
2
4
decreases, at 1 atm total pressure (Fig. 5). This could result is only 10%, CH consumption starts at higher temperature
4
in a lowering of the amount of free carbon at the surface. (685ЊC instead of 630ЊC). Therefore a decrease in the CH4
In order to check this, we have carried out carburization pressure results in a noticeable decrease in the carburiza-

154
LECLERCQ ET AL.
TABLE 3
Bulk and Surface (XPS) Compositions and Specific Surface Areas of Bulk Tungsten Carbides
Composition
XPS
Temperature of
carburization
(ЊC)
Characterization
From
elemental analysis
Specific surface
2
Ϫ
1
Catalysts
From XPS
W
c
/W
t
C
c
/C
t
areas (m g
)
WCR1
WCR2
WC1
WC2
WC3
WC4
WCR3
WC5
WCR4
WC6
800
800
900
600
630
660
700
730
900
900
WC1.22
WC0.90
WC0.90
WC0.43
WC0.44
WC0.52
WC0.86
WC0.90
WC2.23
WC2.58
O
O
O
O
O
O
O
O
O
O
0.02
0.15
0.2
WC3.4
WC1.44
O
O
0.23
0.86
0.64
0.69
0.52
0.50
0.68
0.57
0.62
0.87
0.88
8.6
8.0
0.83
0.55
0.62
0.56
0.56
0.65
0.50
0.36
—
WC1.2
WC1.1
WC1.3
WC0.74
WC1.30
O
0.80
10.2
11.6
21.0
not det.
5.2
6.9
11.4
15.0
0.66
0.74
0.50
0.34
0.41
0.01
0.01
O
O
1.3
1.5
O
O
1.0
1.13
WC2.7
O
0.6
0.3
0.4
WC11
O
WC12.5
O
—
tion rate. Here again, carburization seems to occur in two
As a partial conclusion, the carburization of W metal
steps which would correspond to the formation of W C with a mixture CH –H is faster when the CH pressure
2
4
2
4
then of WC, but these two steps are not as well separated is increased and allows one to distinguish better between
as with 20% CH –H . Both chemical (Table 3) and chro- the two steps of carburization, but it increases the carbon
4
2
matographic analyses seem to indicate that a duration of deposit at the carbide surface.
h in the isothermal stage at 800ЊC is not quite enough
5
to carburize completely 4.3 g of WO with the flow rates
used here.
3
2
.3. Carburization of Reduced WO at 5 Torr (Apparatus
LPR)
3
XPS analysis (Fig. 12) shows that tungsten (Fig. 12a) is
in its carbidic form (W 4f7/2 at a BE ϭ 32.8 eV) associated
with a minor oxidic form (W 4f5/2 at a BE ϭ 38.1 eV) that
arises probably from passivation. This oxidic form is in
a slightly higher proportion than in the previous sample
carburized in 20% CH /H which contained more free car-
bon at its surface. The C 1s photopeak (Fig. 12b) indicates
that carbidic carbon (BE ϭ 283.3 eV) is more abundant
than the free carbon (BE ϭ 285 eV) which is in contrast
to what was obtained for WCR1. The maximum of O 1s
is at 531.3 eV characteristic of oxygen associated with tung-
sten in oxide form. The broadening at higher binding ener-
gies probably originates from oxygen adsorbed on free
carbon. The surface composition of WCR2 is reported in
Table 3.
WO (Fluka, 76 mg) was reduced under 100 Torr of
3
hydrogen at 1000ЊC for 1 h. The surface area of W metal
2
measured by the Kaganer method (26) is very low: 0.1 m
Ϫ
1
g . Then it was submitted to a flow of 6% CH –D mixture,
4
2
at a total pressure of 5 Torr with a heating rate of 360ЊC
4
2
Ϫ
1
h
from 25 to 1000ЊC. Changes in the methane partial
pressure and the temperature versus time are reported in
Fig. 13. The amount of CH does not change until 750ЊC,
4
and then one can observe a sudden decrease in CH at
4
partial pressure between 750 and 815ЊC. This decrease is
not accompanied with increases in the heights of peaks at
mass 17 to 20, hence there is no CH –D exchange. In
4
2
contrast, a corresponding increase in the heights of peaks
at mass 2 and 3 indicates that methane is decomposed into
carbon and hydrogen adsorbed at the surface allowing
dihydrogen desorption and exchange between deuterium
and H adsorbed, and, probably, the beginning of W carbu-
rization.
It is clearly seen that using a lower CH pressure leads
4
to less pollution of the carbide surface with free carbon.
However, despite a possible global carbon deficit of the
carbide, excess free carbon is deposited at the surface but
The complete absence of deuterated methane clearly
at a much lower extent than when 20% CH –H is used.
4
2
indicates that each molecule of CH which becomes ad-
4
Moreover, this carbon deficit does not seem to manifest
itself at the surface (C /W ϭ 1.2). As a consequence of
sorbed on W at 1000ЊC does not desorb but remains on
the solid and is decomposed. It should be noted that a
c
c
lower carbon deposit, oxygen enters to a higher extent tremendous decrease in CH pressure from 76 Torr (0.1
4
during the passivation step and the tungsten carbide sur- atm) to 0.25 Torr only results in a modest increase in the
face is more oxidized (W /W ϭ 0.64 and O/W ϭ 0.83 temperature of the beginning of CH consumption from
c
t
4
against 0.86 and 0.2 for WCR1).
685 to 750–815ЊC.

PREPARATION OF TUNGSTEN CARBIDES
155
TABLE 4
Experimental Conditions of Preparation of Bulk Tungsten Carbides at Atmospheric Pressure
Preparation
Posttreatment in H
2
Passivation
Precursor
weight (g)
Fa
T
Gb
Duration
(h)
F
T
Duration
(h)
Cooling
to RT
F
Duration
(h)
Catalyst
WCR1
(liter hϪ1) (ЊC) (ЊC hϪ1)
(liter hϪ1) (ЊC)
(liter hϪ1) % O
2
c
WO
3.27
3
TN ϭ 9.95 550
TR ϭ 8.0 600
TC ϭ 10.2 800
50
50
50
10
10
10
d
No
CH
4
ϩ H
2
11.0
3.5
10
1.4
2.0
1.0
15
15
15
e
WO
4.35
3
TN ϭ 3.6
TR ϭ 3.6
TC ϭ 3.1
500
600
800
60
60
60
14
12
5
WCR2
WC1
No
H
2
2
Ammonium
metatungstate
TN ϭ 10.0 550
TC ϭ 10.4 900
50
50
10
10
8.3
800
5
H
3
.51
WO
3.14
3
TN ϭ 9.7
TC ϭ 10.0 600
550
50
50
10
33
WC2
WC3
WC4
WCR3
No
No
No
No
CH
CH
4
4
ϩ H
2
2
8.5
10.3
3.0
0.5
0.6
2.0
2.0
15
15
15
15
WO
3.11
3
TN ϭ 10.1 550
TC ϭ 10.3 630
50
50
10
28
ϩ H
WO
1.02
3
TN ϭ 6.0
TC ϭ 6.1
550
660
60
60
5
4
H
2
WO
5.20
3
TN ϭ 2.5
TC ϭ 2.7
TC ϭ 3.4
500
600
700
30
30
30
10
10
10
CH
CH
CH
CH
4
4
4
4
ϩ H
ϩ H
ϩ H
ϩ H
2
2
2
2
3.0
WO
3.37
3
TN ϭ 10.2 550
TC ϭ 10.9 730
50
50
10
10
WC5
No
No
No
8.9
10.1
10.0
1.3
1.6
0.6
15
15
6
WO
3.53
3
TN ϭ 10.1 550
TC ϭ 8.2
TC ϭ 11.0 900
50
10
11
15
WCR4
600
50
50f
WO
3.72
3
TN ϭ 10.4 550
TC ϭ 10.0 900
50
10
6
WC6
50f
a
Flow rate (liter h 1).
Ϫ
b
c
d
e
f
Gradient of temperature (ЊC hϪ1).
Treatment in N
Treatment of reduction.
Treatment of carburization.
0ЊC h from RT to 450ЊC, then by steps of about 10 h every 30ЊC.
2
.
Ϫ
1
5
The integration of the amount of CH decomposed cor- prepared by metatungstate decomposition. The maximum
4
responds to a ratio C/W of 0.99.
temperature of carburization was 900ЊC, a temperature at
which the sample was left for 10 h.
During this experiment the chromatographic analyses
indicated, besides CH and H consumption, the formation
3. Carburization of Unreduced WO3
4
2
of H O, CO, and CO . The composition of the gas flowing
at the outlet of the reactor is reported in Fig. 14.
3
.1. Carburization at 1 atm (20% CH –H ): WC1
2
2
4
2
(
Apparatus APR1)
The curve of H O formation is very much like that ob-
2
In their papers on tungsten carbide preparation, Boudart tained during TPR of WO with the mixture 30% H –Ar,
3
2
and co-workers (12–14) state that they have directly carbu- but it is shifted toward lower temperatures as expected
rized WO without previous reduction. Hence we have with a hydrogen pressure of 0.8 atm instead of 0.3 atm.
3
studied the carburization of samples of a few grams of Starting at 550ЊC, CO production occurs, it is maximum
Ϫ
2
WO . Here the precursor WO (1.42 ϫ 10 mol) was at 740ЊC, and it continues until 900ЊC. It is accompanied
3
3

156
LECLERCQ ET AL.
chromatographic peak versus time during temperature programmed carburization of W metal (60ЊC.h 1
Ϫ
,
FIG. 10. Changes of the area of the CH
.35 g, 10% CH –H ).
4
4
4
2
by CO from 700ЊC. CO and CO are produced during reported in Appendix II we have calculated the ratios
2
2
tungsten oxide reduction either directly in reactions such as X/N (already defined), and Y/N , Z/N , N /N using
0
0
0
H O
0
2
the areas of the various product peaks. X, Y, and Z are
WO ϩ CH ϭ WO ϩ CO ϩ 2H
2
[4] the number of CH moles consumed for carburization
4
3
4
2
and carbon deposit (X), and for CO and CO formation
2
WO ϩ 2CH ϭ W ϩ 2CO ϩ 4H
2
[5]
2
4
2
(
Y and Z) per unit of time, and N
is the number of
H O
2
molecules of water formed per unit of time.
WO ϩ CH ϭ W ϩ CO ϩ 2H ,
[6]
2
4
2
2
Then one can plot the curves of X/N and (Y ϩ 2Z ϩ
0
N
)/N as a function of time which represent, respec-
or in methane reforming by H O produced in the reduction
H O
2
0
2
tively, the relative amounts of carbon consumed (for carbu-
rization and carbon deposit) and of oxygen eliminated in
of WO with H :
3
2
WO reduction. Their integrations lead to the curves in
3
CH ϩ H O ϭ CO ϩ 3H
2
[7]
[8]
4
2
Figs. 15 and 16. The total amount of oxygen in H O, CO,
2
CH ϩ 2H O ϭ CO ϩ 4H .
4
2
2
2
and CO formed (Fig. 16) corresponds to an O/W ratio of
2
0
3
.14 which shows that all WO has been reduced into W
3
Reactions [7] and [8] are catalyzed by transition metals in the reaction. The amount of oxygen consumed during
40, 41) and they usually occur at temperatures between the first peak of reduction (max: 630ЊC) corresponds to a
00 and 800ЊC, as in this experiment. At this stage, we ratio O/W close to 1.4. Hence, at that temperature WO3
(
6
cannot identify the principal steps leading to CO and CO2 is reduced into WO but also some W metal is already
2
production. However, in each of the reactions [4] to [8], formed.
the disappearance of one molecule of CH is accompanied
Coming back to Fig. 14, a decrease in CH pressure is
4
4
by the formation of three molecules (one of CO or CO2 noticeable from 600ЊC, and the pressure then goes through
and two of H ) in the gas phase. By the simple calculation a mild minimum around 650ЊC, increases slightly before
2

PREPARATION OF TUNGSTEN CARBIDES
157
FIG. 11. Ratio of the number of CH
W metal in a 10% CH –H mixture.
4
moles consumed to that of W metal versus time during the temperature programmed carburization of
4
2
decreasing again at 720ЊC, and goes through two minima
at 750 and 785ЊC. From Fig. 15, one can estimate the ratios and passivation, chemical analysis indicates a composition
C/W corresponding to these three peaks (Table 5). corresponding to WC O which shows that the hydrogen
After a treatment in flowing hydrogen for 5 h at 800ЊC
0
.9 0.2
In this case the carbon consumption in the first step of treatment removes most of the free carbon. X-ray diffrac-
carburization (600–720ЊC) is much lower than that corre- tion shows that the solid obtained in this preparation is
sponding to the first step of carburization of W metal into the hexagonal WC.
W C (C/W ϭ 0.5). Here, only around 50% of tungsten
2
W 4f, C 1s, and O 1s XP spectra are very similar to
would have led to W C. From Fig. 16, the amount of oxygen those obtained for WCR2 (Fig. 12). Table 3 shows that
2
consumed up to 720ЊC, which corresponds to the end of the resulting solid is tungsten carbide and that the amount
Ϫ
2
the first peak of carburization, is 2.8 ϫ 10 mol O, an of free carbon is low (C/W ϭ 1.2), but also that part of
amount which allows the complete reduction of WO into tungsten carbide was more oxidized during passivation
3
Ϫ
2
WO (1.42 ϫ 10 mol O) and of 48.5% of WO into W (W /W ϭ 0.69) than WCR1 which had not been treated
2
2
c
t
metal. These figures seem to indicate that only W metal in a flow of pure hydrogen after carburization.
can be carburized by CH . The explanation for the second The temperature programmed carburization at 1 atm of
and the third peaks of CH consumption is not easy. The WO (Fluka) in a flow of 20% CH –H was also studied
4
4
3
4
2
second peak could correspond to the completion of W C in situ by X-ray diffraction. In the X-ray diffractometer,
2
formation together with the beginning of WC formation the sample was submitted to drying (RT to 550ЊC) and
and of free carbon deposit. The third peak could be due carburizing treatments similar to those of the previous
to the end of WC formation accompanied by free car- sample WC1 up to 800ЊC where it was left for 5 h. An X-
bon deposit.
ray diffraction pattern was registered every 100ЊC up to
The total amount of C consumed in this experiment 500ЊC, then every 50ЊC during the temperature program-
C/W ϭ 1.51) is higher than that in the preparation of ming, and after every hour during the isothermal
(
WCR1. This shows that at 900ЊC carbon deposit is more treatment.
important than at 800ЊC. The various X-ray diffraction patterns are reported in

158
LECLERCQ ET AL.
FIG. 12. XPS analysis of WC (WCR2) prepared by temperature programmed carburization of W metal up to 800ЊC in 10% CH
f photopeak; (b) C 1s photopeak; (c) O 1s photopeak.
4 2
–H . (a) W
4
Fig. 17. During the N drying treatment no change occurs; could be hydrogen bronzes or/and tungsten suboxides. At
2
only WO is present. During the carburizing treatment, 550ЊC, W O and bulk W O (or W C) can be identified.
3
20 58
3
3
noticeable modifications begin to be observed at 500ЊC However, no WO has been detected by XRD. At 600ЊC
2
where a small amount of some new phases were formed the pattern is completely modified and indicates the pres-
which were difficult to identify with certainty but which ence of a single phase which is either W O or W C (Fig.
3
3

PREPARATION OF TUNGSTEN CARBIDES
159
FIG. 12—Continued
18). At 650ЊC a small proportion of carbide which could 600 and 810ЊC at 0.3 atm and between 830 and 1000ЊC at 0.1
correspond to W C or to W C is formed. Then, at 700 atm. The total amount of oxygen consumed corresponds to
6
2.54
2
and 750ЊC, only W C is observed and finally at 800ЊC WC a ratio O/W close to 3, indicating that WO has been
2
3
is progressively formed at the expense of W C which disap- completely reduced into W metal.
2
pears almost completely after 2 h of isothermal treatment
In the experiment performed at 0.3 atm, CH consump-
4
at 800ЊC.
tion is important at temperatures between 760 and 840ЊC.
This XRD study is in good agreement with the previous It starts very slightly at lower temperatures (about 600ЊC)
experiments, since it indicates a progressive reduction of but carbon mass balance shows that CH is mainly trans-
4
WO starting at about 500ЊC. No W metal has been de-
3
formed into CO. The real CH consumption for carburiza-
4
tected by X-ray diffraction, but the formation of W C when
2
tion starts only at 760ЊC; that is at a much higher tempera-
ture (⌬T ϭ 160ЊC) than during carburization at 1 atm. The
difference is even much more important for the experiment
W O (or W C) disappears probably indicates that with a
3
3
CH pressure of 0.2 atm, W metal is almost immediately
4
transformed into W carbide. The progressive carburization
of W, first into W C and then into WC, has been evi-
denced here.
at 0.1 atm since, there, CH consumption begins at about
4
2
870ЊC, is roughly constant up to 970–980ЊC, sharply in-
creases at 980ЊC, and finally decreases and becomes negligi-
ble. The first consumption of CH up to 980ЊC corresponds
only, or mainly, to CO production and the second sharp
4
3
.2. Carburization at Low Pressures (Mixtures of 20%
CH –H from 0.3 atm to 0.5 Torr)
4
2
CH peak is due to carburization which occurs when a
4
3
.2.1. Carburization at mixture pressures of 0.3 atm and large part of WO has been reduced into W metal. Hence,
3
0
.1 atm (apparatus APR2). The carburizing mixture was at lower CH pressure the relative position of the peak of
4
composed of 20% CH /H (0.1 or 0.3 atm) diluted with carburization compared to that of the W formation is
4
2
argon.
shifted toward higher temperatures. Such a behavior can
The variations of the composition of the products flowing easily be understood on the basis of two competitive phe-
out of the reactor were rather similar to those observed nomena occurring to W metal formed at the surface of the
with sample WC1 (Fig. 14) during the carburization at solid. The first is its carburization and the second is the
atmospheric pressure, but they are shifted toward higher diffusion of W atoms from the surface into the bulk of the
temperatures (Table 6).
sample. The diffusion of W metal is unaffected by CH4
A low production of CO has been observed between pressure and it was shown previously to be fast compared

160
LECLERCQ ET AL.
4 3 4 2
FIG. 13. Temperature programmed carburization of W metal at 5 Torr. Change of CH , 20% CH –D ).
pressure versus time (360ЊC hϪ1, 76 g WO
to the rate of reduction of W oxides by H . In contrast, the kept at 1200ЊC for 15 min, and then it decreases while
2
carburization of W metal is probably very much lowered by carburization starts to take place.
a decrease of CH pressure. Hence, if the diffusion of W
CH exchange is better followed by the height of the
4
4
is much faster than its carburization one can understand peak at m/z ϭ 15 (h ) than by h , since water decomposi-
1
5
16
that carburization of W occurs only when a large part of tion participates at mass 16. One can observe (Fig. 20)
W oxide has been reduced into W.
that, up to about 1000ЊC, no methane exchange occurs,
The stoichiometry of the final product carburized at 0.1 and then its importance continuously increases with tem-
atm has been determined by temperature programmed perature. It is interesting to notice that the temperature
oxidation from the amounts of CO and CO produced
during TPO. It corresponds to the formula WC1.1 which is
very close to WC.
2
where CH exchange starts to take place is the same as
4
that where H and HD appear in the gas phase (Fig. 21).
2
Therefore, probably, CH exchanges hydrogen atoms with
4
The total carbon consumption in the carburization at
D and not with D O. Peaks at m/z ϭ 17, 18, 19, and 20
2
2
0.3 atm corresponds to a ratio C/W close to 1.44. This
start to increase at the same temperature, about 650ЊC,
which corresponds to the beginning of the reduction of
WO by D ; then they go through maxima (at the same
indicates that the CH pressure has a large importance for
4
the amount of free carbon at the surface of the catalyst
since the ratios C/W are 1.51, 1.44, and 1.1 for temperatures
of carburization, respectively, of 900, 840, and 1000ЊC in
CH –H mixtures at pressures of 1, 0.3, and 0.1 atm.
3
2
temperature for m/z ϭ 17, 18, and 20) and become negligi-
ble after about 30 min at 1000ЊC indicating that the reduc-
4
2
tion of WO is complete. The simultaneous increase of
3
3
.2.2. Carburization at a total pressure of 5 Torr (appara- peaks at mass 20 and 17 to 19 indicates that exchange
tus LPR). The reactions that can take place during this between D O and some hydrogen-containing species oc-
2
experiment are numerous, thus there are the reduction of curs. However, since no CH is exchanged before 1000ЊC,
4
WO by D and by CH (which will form CO and CO ), HDO is probably formed at lower temperatures by ex-
3
2
4
2
the exchange of CH with D or with D O formed by the change of D O with some hydroxyl groups at the surface
4
2
2
2
reduction of WO , and the consumption of methane for of WO or with some water molecules of crystallization of
3
3
the carburization or for free carbon formation (decomposi- WO . Let us notice that, at 1000ЊC, the height of peak
3
tion of CH ). Information on this last reaction of methane 19 increases much more rapidly when the temperature
4
consumption is given by the height of the peak at m/z ϭ increases than that of peak 20 (Fig. 22), probably in relation
1
2 (h ) independently of the exchange of CH . It has been to CH exchange which gives HDO. Moreover, the height
1
2
4
4
observed (Fig. 19) that, within the margin of error, the of peak 17, which was very low, suddenly increases at that
total methane pressure is constant until the solid has been temperature, very probably and in the main because of

PREPARATION OF TUNGSTEN CARBIDES
161
FIG. 14. Changes of the area of CH
of the TCD) during the temperature programmed carburization of WO
S(CO)/42; छ S(H O)/33; ᭝ (S(CO )/48) ϫ 30.
4 2 2
, H O, CO, and CO chromatographic peaks (normalized, taking into account the relative molar response
Ϫ
1
3
4 2 4
(50ЊC h , 3.51 g ammonium metatungstate, 20% CH –H ). Ճ S(CH )/36;
᭜
2
2
the formation of CH D. An argument for the assignment
3.2.3. Carburization at a total pressure of 0.5 Torr (appa-
3
of peak 17 mainly to CH D after 1000ЊC can be found ratus LPR). At such a low pressure, peaks 17 to 20, which
3
when considering the ratio of the heights of peaks 17 and indicate the formation of water, hence the occurrence of
19 (Table 7) (these heights have to be corrected for the tungsten oxide reduction, start to increase at only about
residuals). This ratio is roughly constant until 1000ЊC and 800ЊC, the temperature where CO starts to appear. Hence
is approximately equal to 0.12 Ϯ 0.02 which is characteristic at that pressure, water and carbon monoxide are simultane-
ϩ
of HDO fragmentation into HO . Then it noticeably in- ously formed. No decrease in the height of peak 16 has
creases with temperature because of CH D formation. In been observed. Moreover, no formation of H or of HD
3
2
contrast, the ratio h /(h ϩ h ) is roughly constant what- has been detected. Hence, no CH exchange with D takes
18
20
19
4
2
ever the temperature, which indicates that h18 mainly, if place, indicating that no W metal is present at the surface
not only, originates from D O and HDO fragmentation. of the solid. Consequently no carburization can occur. A
CO and CO were also observed at high temperature further analysis has confirmed these conclusions, since only
2
2
(
approximately from 830 to 860ЊC).
an amount of carbon corresponding to 0.5% of the W atoms
The comparison of the curves in Fig. 19 (CH consump- present has been detected, which corresponds to surface
4
tion) and 22 (water production) indicates that carburiza- carbon deposit.
tion occurs only when the reduction of WO is complete
3
(
end of water formation). This is very well explained by a
4
. Influence of the Maximum Temperature of
Carburization
very fast migration of W metal atoms into the bulk of the
sample and, since at such a low CH pressure of 1 Torr
4
the rate of carburization is certainly very low, carburization
The maximum temperature of carburization was varied
cannot occur before complete WO reduction into W metal between 600 and 900ЊC. In each case, this temperature was
3
which occurs after about 15 min at 1200ЊC (Fig. 22).
maintained constant for several hours. The experimental

162
LECLERCQ ET AL.
FIG. 15. Ratio of the number of CH
0% CH –H
4
3
moles consumed to that of W versus time during the temperature programmed carburization of WO in
2
4
2
.
conditions of preparation of the various samples are sum-
Carburization at 700 or 730ЊC (samples WCR3 and
marized in Table 4 and their compositions (elemental anal- WC5) leads to mixtures of W C and WC (XRD and ele-
2
ysis and XPS) together with their physical and chemical mental analysis, Table 3) with a large predominance of the
characterizations are in Table 3.
WC phase. After carburization at higher temperatures (800
At 600ЊC (sample WC2), the solid obtained is a mixture to 900ЊC) the only phase detected by X-ray diffraction
of W C, W metal, and probably W oxycarbide as seen by is WC.
2
X-ray diffraction. Its overall formula is WC0.43O0.66. Hence
00ЊC is too low a temperature to obtain pure tungsten temperature of carburization, it is possible to synthesize
carbides. In contrast, a temperature of carburization of pure W C or WC.
It is interesting to note that, according to the selected
6
2
6
30ЊC (WC3) or 660ЊC (WC4) leads to pure W C (checked
Table 3 clearly shows that the higher the temperature
2
by XRD). The amount of CH consumed during the prepa- of carburization the higher the proportion of free carbon
4
ration of the sample WC3 corresponds to a ratio C/W of and the lower the proportion of W oxide after passivation.
0.52. Its overall composition after 1 year storage in air was
WC0.44O , which shows that it has been heavily oxidized.
CONCLUSION
1.1
For these three compounds, XPS analysis (Table 3) indi-
cates the presence of W carbide and oxide together with
This study of the carburization of W metal and of WO3
carbidic carbon and carbon of pollution. The proportions by CH –H mixtures has clearly shown that it occurs in
4
2
of carbidic C (C /C ϭ 0.56 and 0.65) and of W oxides two distinct steps. W C is formed in the first step at about
c
t
2
(
W /W ϭ 0.32 and 0.5) are high, indicating that, at these 650–700ЊC in a CH pressure of 0.2 atm, while WC is
4
o t
low temperatures, the amount of free carbon is low, and formed in the second step. Hence it is possible to synthesize
consequently it cannot play its protecting role during the either W C or WC by adjusting the conditions of carburiza-
2
passivation step, leading to significant surface oxidation tion (mainly the temperature). It has also been shown
of W C.
that the rate of reduction of WO together with that of
2
3

PREPARATION OF TUNGSTEN CARBIDES
163
FIG. 16. Number of H
2
O, CO, and CO
) ϫ 30.
2
moles formed during temperature programmed carburization of WO
3
in 20% CH
4
–H
2
. Ճ N(CO);
ᮀ
N(H O); ᭜ N(O total); ᭡ N(CO
2
2
carburization of W are very much dependent on the total to the conclusion that tungsten oxides cannot be directly
pressure (hence on the pressure of H for the first reaction carburized under the conditions studied here, but only W
2
and on that of CH for the second reaction), since the metal can be carburized. When W metal is formed at the
4
temperature of the beginning of reduction by hydrogen surface of the incompletely reduced oxide, the phenomena
shifts from about 300–400 to 700ЊC when hydrogen pres- observed are very different when the methane pressure is
sure is decreased from 1 atm to 1 Torr and the carburization decreased, because of the competition between the diffu-
of W metal starts at about 630 and 685ЊC when the percent- sion of W metal from the surface into the bulk of the
age of CH is equal to, respectively, 20 and 10% in CH –H
mixtures at atmospheric pressure.
sample and its carburization at the surface. At high meth-
ane pressure (0.2 atm), W metal is carburized almost as
4
4
2
The direct carburization of WO by CH –H has led soon as it is formed, as evidenced by in situ X-ray diffrac-
3
4
2
tion experiments during the temperature programmed
carburization and by the amounts of oxygen (from the
oxide) and of methane consumed as a function of the
TABLE 5
Ratio C/W in Sample WC1 and Ratio O/W (O consumed temperature. However, when the methane pressure is
per W during reduction) as a Function of the Temperature of lower (0.02 atm or 1 Torr), the rate of W carburization
Reaction
is strongly decreased while that of W diffusion is not
changed. Hence most of the W migrates into the solid
Peak
Temperature
before being carburized, and carburization occurs only
at higher extents of reduction of W oxides into W metal.
This results in a large shift of the temperature of the
beginning of carburization. However, at such low methane
(CH
4
consumption)
(ЊC)
C/W
O/W
1
2
3
600–720
720–770
770–900
0.26
0.85
1.51
2.0
3.14
3.14
pressures the consumption of CH is very steep, indicating
4

164
LECLERCQ ET AL.
(60ЊC h 1, 20%
Ϫ
FIG. 17. X-ray diffraction patterns at different temperatures during in situ temperature programmed carburization of WO
CH –H ).
3
4
2
3 4 2 3 3
FIG. 18. X-ray diffraction pattern of WO treated at 600ЊC in a mixture 20% CH –H . ᭡ W O; ᭹ W C.

PREPARATION OF TUNGSTEN CARBIDES
165
TABLE 6
Results of the Temperature Programmed Carburization of Unreduced WO
3
4 2
at Various Pressures of 20% CH –H
Carburization
Reduction temperature (ЊC)
temperature (ЊC)
P
CH4 ϩ PH2
(
atm)
Diluent
Beginning
First max.
Second max.
Beginning
Max.
1
None
Ar
Ar
300
500
540
650
630
640
800
750
780
ȁ1000
600
760
870
650–720
810
980
0
0
.3
.1
Ϫ
3
6
.6 ϫ 10
None
1000
1200
—
that carburization is very fast as soon as, due to the When 1 mol CH disappears, 2 mol H are formed.
4
2
extended reduction of WO , some W metal becomes
available at the surface of the solid.
3
N denotes the number of moles CH in the initial mix-
ture per second,
0
4
For all the carbides obtained by carburization with a
CH –H mixture it has been shown that some free carbon is
X the number of moles CH transformed per second, and
4
4
2
a the initial ratio of the molar fractions of H and of CH .
deposited at their surface and that some surface oxidation
occurs during the passivation and the storage of the solids.
However, their composition strongly depends on the tem-
2
4
The number of moles of CH and H flowing out of the
4
2
reactor are
perature of carburization and on the CH pressure, since
4
the proportion of free carbon at the surface increases with
methane pressure at a given temperature and with the
temperature of carburization at a given CH pressure. Only
NCH4 ϭ N Ϫ X and NH2 ϭ aN ϩ 2X.
0 0
4
a small amount of carbon of pollution is observed after
carburization at 650ЊC, but only free carbon can be ob-
served by XPS after carburization at 900ЊC. However, the
presence of free carbon can be advantageous, since the
more free carbon at the surface the less the extent of
oxidation of the W carbide surface, which shows that car-
bon of pollution protects W carbides from oxidation. Nev-
ertheless, the presence of the oxide phases and of free
carbon at the surface of carbides is obviously detrimental
for their use in catalysis. Consequently some pretreatments
for cleaning their surfaces will be necessary before catalytic
experiments. These pretreatments will be described in a
further paper.
One can calculate the partial pressure of CH4
(a ϩ 1)(PCH4) (N Ϫ X)
P (N Ϫ X)
0
0
T
0
P
ϭ
ϭ
,
CH4
(
a ϩ 1)N ϩ X
(a ϩ 1)N ϩ X
0
0
and the ratio X/N is
5
[(PCH4)0 Ϫ PCH4]
X
N0
APPENDIX I
ϭ
.
5(PCH4)0 Ϫ PCH4
Calculation of the Ratio X/N Fraction of CH4
0
Consumed Per Unit of Time during the Carburization
of W Metal
Or, since PCH4 is proportional to the area of the chromato-
graphic peak corresponding to CH4,
The reactions that take place are
2
W ϩ CH ϭ W C ϩ 2H
4 2 2
W ϩ CH ϭ WC ϩ 2H
4
2
(
a ϩ 1)[(SCH4)0 Ϫ SCH4]
X
N0
W C ϩ CH ϭ 2WC ϩ 2H
2
4
2
ϭ
.
(a ϩ 1)(SCH4)0 ϩ SCH4
CH ϭ C ϩ 2H .
4
2

166
LECLERCQ ET AL.
3 4 2
FIG. 19. Temperature programmed carburization of WO in a mixture of 20% CH –D
at 5 Torr (360ЊC hϪ1, 74 mg). Height of peak at mass
1
2 versus time.
3 4 2
FIG. 20. Temperature programmed carburization of WO in a mixture of 20% CH –D
at 5 Torr (360ЊC hϪ1, 74 mg). Heights of peaks at mass
1
5 and 16 versus time.

PREPARATION OF TUNGSTEN CARBIDES
167
3 4 2
FIG. 21. Temperature programmed carburization of WO in a mixture of 20% CH –D
at 5 Torr (360ЊC hϪ1, 74 mg). Heights of peaks at mass
2
and 3 versus time.
3 4 2
FIG. 22. Temperature programmed carburization of WO in a mixture of 20% CH –D
at 5 Torr (360ЊC hϪ1, 74 mg). Heights of peaks at mass
1
9 and 20 versus time.

168
LECLERCQ ET AL.
TABLE 7
Carburization of WO
3
with 20% CH
4
–D
2
at a Pressure of 5 Torr Mass Spectrometry Analysis
Temperature
h
20 corr.
(V)
h
19 corr.
(V)
h
18 corr.
(V)
h
17 corr.
(V)
h
17 corr.
19 corr.
h17 corr.
(
ЊC)
h
h
19 corr. ϩ h20 corr.
6
7
7
8
8
8
9
9
9
002
049
106
129
152
187
200
200
200
200
52
18
98
29
60
84
07
38
75
0.0013
0.0046
0.0175
0.0405
0.0740
0.1090
0.1540
0.2920
0.4350
0.6500
0.7650
0.9500
0.9940
1.0450
1.145
0.0036
0.0081
0.0195
0.0317
0.0442
0.0547
0.0652
0.0867
0.1032
0.1212
0.1242
0.1432
0.1582
0.1832
0.2632
0.3032
0.2982
0.0432
0.0102
0.0024
0.0057
0.0129
0.0216
0.0316
0.0421
0.0549
0.0916
0.1416
0.1816
0.2106
0.2776
0.3076
0.3406
0.4376
0.4576
0.4276
0.0626
0.0156
0.0009
0.0012
0.0027
0.0035
0.0055
0.0060
0.0075
0.0110
0.0150
0.0225
0.0375
0.1225
0.1465
0.1715
0.1965
0.1965
0.1815
0.0315
0.0255
0.25
0.15
0.14
0.11
0.12
0.11
0.12
0.13
0.14
0.19
0.30
0.85
0.93
0.94
0.75
0.65
0.61
0.73
2.5
0.49
0.45
0.35
0.30
0.27
0.26
0.25
0.24
0.26
0.24
0.24
0.25
0.27
0.28
0.31
0.30
0.32
0.33
0.51
1
1
1
1
1
1
1
1
1
1
1.195
1,0150
0.1450
0.0205
APPENDIX II
Calculation of the Ratios X/N , Y/N , Z/N , and
Hence,
͚
N ϭ (a ϩ 1)N ϩ X ϩ 2Y ϩ 2Z using the
0
same meaning for a and N as in Appendix I.
The various partial pressures are
0
0
0
0
N
/N (X, Y, Z, N ؍ Number of Moles CH4
H O
0
H O
2
2
Consumed Per Second for Carburization (X), for CO
P (N Ϫ X Ϫ Y Ϫ Z)
T
0
Formation (Y), for CO Formation (Z), and Number of
2
PCH4 ϭ
͚
N
Moles H O Formed Per Second) during the
2
Carburization of WO3
Y
Z
PCO ϭ PT
PCO2 ϭ PT
The reactions taking place are
͚
N
͚
N
H ϩ ‘‘O’’ ϭ H O
and
2
2
WO ϩ CH ϭ WO ϩ CO ϩ 2H
2
3
4
2
N
H O
2
P
H O
2
ϭ PT
WO ϩ 2CH ϭ W ϩ 2CO ϩ 2H
2
2
4
͚
N
WO ϩ CH ϭ W ϩ CO ϩ 2H
2
2
4
2
since
CH ϩ H O ϭ CO ϩ 3H
2
4
2
CH ϩ 2H O ϭ CO ϩ 4H
P ϭ (a ϩ 1)(P ) .
T CH4 0
4
2
2
2
and the reactions of carburization (see Appendix I) and/
or any combination of these equations.
One can easily calculate
One can see that when 1 mol of CO or CO is formed
there is an increase of 2 mol in the gas phase; when 1 mol
2
(
a ϩ 1)[(PCH4)0 Ϫ PCH4] Ϫ (a ϩ 3)[PCO ϩ PCO2]
(a ϩ 1)(PCH4)0 ϩ PCH4 Ϫ PCO Ϫ PCO2
X
N0
ϭ
ϭ
of CH is consumed for carburization or carbon deposition,
4
there is an increase of 1 mol in the gas phase, and the
Y
N0
(a ϩ 2)PCO
disappearance of 1 mol of H does not change the number
2
(a ϩ 1)(PCH4)0 ϩ PCH4 Ϫ PCO Ϫ PCO2
of mol in the gas phase.

PREPARATION OF TUNGSTEN CARBIDES
169
‘‘Preparation of Catalysts II’’, (B. Delmon, P. Grange, and G. Pon-
(
a ϩ 2)PCO2
celet, Eds.), p. 627 Elsevier, Amsterdam, 1978.
Z
N0
ˆ
ϭ
ϭ
16. Lemaitre, J., Vidick, B., and Delmon, B., J. Catal. 99, 415 (1986).
(a ϩ 1)(PCH4)0 ϩ P
Ϫ PCO Ϫ PCO2
ˆ
CH4
17. Vidick, B., Lemaitre, J., and Delmon, B., J. Catal. 99, 428 (1986).
8. Boudart, M., and Leclercq, L., U.S. Patent 4, 271, 041, 1981.
1
N
(a ϩ 2)P
19. Boudart, M., Oyama, S. T., and Leclercq, L., ‘‘Proceedings 7th Inter-
national Congress on Catalysis, Tokyo, 1980’’ (T. Seiyama and K.
Tanabe, Eds.), p. 578. Elsevier, Amsterdam, 1981.
H O
2
H O
2
.
N0
(a ϩ 1)(PCH4)0 ϩ P
CH4 CO2
Ϫ PCO
P
2
0. Laine, R. M., and Hirschon, A., in ‘‘Better Ceramics Through Chemis-
try II’’ (C. J. Brinker, D. E. Clark, and D. R. Ulrich, Eds.). Materials
Research, New York, 1984.
PCH4, PCO, PCO2, and P
are proportional to the areas of
H O
2
2
2
2
2
1. Giraudon, J. M., Leclercq, L., Leclercq, G., L o¨ fberg, A., Frennet, A.,
J. Mater. Sci. 28, 2449 (1993).
2. Ledoux, M. J., Pham Huu, C., Marin, S., Guille, J., and Weibel, M.,
C. R. Acad. Sci. Ser. II 707 (1990).
3. Ledoux, M. J., Pham Huu, C., Guille, J., and Dunlop, H., J. Catal.
the chromatographic peaks corresponding to, respectively,
CH , CO, CO , and H O, taking into account the molar
4
2
2
response factors of the TCD.
1
34, 383 (1992).
ACKNOWLEDGMENTS
4. Mooney, W., Chiola, V., Hoffman, C. W. W., and Vanderpool, C. D.,
J. Electro. Chem. Soc. 109, 1179 (1962).
This work was supported by a Stimulation Contract ST2J-0467-C(TT)
from the EEC.
2
2
2
5. Leclercq, G., Leclercq, L., and Maurel, R., J. Catal. 5, 87 (1977).
6. Kaganer, M. G., Zhur. Fiz. Khim. 33, 2202 (1959).
7. Delaunois, Y., Frennet, A., and Lienard, G., J. Chim. Phys. 63,
9
06 (1966).
REFERENCES
2
8. (a) Tilley, R. D. J., ‘‘Defect Crystal Chemistry and its Applications,’’
Blackie Glasgow/London, 1987; (b) Hansen, M., ‘‘Constitution of
Binary Alloys’’ p. 1076. McGraw–Hill, New York, 1958.
9. Vermaire, D. C., and Van Berge, P. C., J. Catal. 116, 309 (1989).
0. Gr u¨ nert, W., Shpiro, E. S., Feldhaus, R., Anders, K., Antoshin,
G. V., and Minachev, K. M., J. Catal. 107, 522 (1987).
1. Barin, I., and Knacke, O., ‘‘Thermochemical Properties of Inorganic
Substances.’’ Springer Verlag, Berlin, 1973.
2. Lee, J. S., Locatelli, S., Oyama, S. T., and Boudart, M., J. Catal. 125,
1
. Toth, L. E., ‘‘Transition Metal Carbides and Nitrides.’’ Academic
Press, New York, 1971.
. Levy, R. B., and Boudart, M., Science 181, 547 (1973).
2
3
2
3
4
5
. Levy, R. B., Adv. Mater. Catal. 101 (1977).
. Oyama, S. T., and Haller, G. L., Catal. Spec. Per. Rep. 5, 333 (1982).
. Leclercq, L., in ‘‘Proceedings of the Nato Advanced Study Institute,
Lille, France, 1982’’ (J. P. Bonnelle, B. Delmon, and E. Derouane,
Eds.), p. 433. Reidel, Dordrecht, 1983.
. Ribeiro, F. H., Boudart, M., Dalla Betta, R. A., and Iglesia, E., J.
Catal. 130, 498 (1991).
. Iglesia, E., Baumgartner, J. E., Ribeiro, R. H., and Boudart, M., J.
Catal. 131, 137 (1991).
. Keller, V., Cheval, M., Vayer, M., Ducros, R., and Maire, G., Catal.
Lett. 10, 137 (1991).
3
3
1
57 (1990).
6
7
8
9
3
3
3. Rodero, A., Ph.D. Thesis, University of South Florida, 1976.
4. Jama, C., Dessaux, O., Goudmand, P., Gengembre, L., and Grimblot,
J., Surf. Interface Anal. 18, 751 (1992).
5. Bonzel, H. P., and Krebs, H. J., Surf. Sci. 91, 199 (1980).
6. Bou, M., Martin, J. M., Lemogne, T., and Vovelle, L., Appl. Surf.
Sci. 47, 149 (1991).
7. Dianis, W. P., and Lester, J. E., Anal. Chem. 45, 1416 (1973).
8. Blanchard, L., Grimblot, J., and Bonnelle, J. P., J. Catal. 98, 229 (1986).
9. Grimblot, J., Payen, E., and Bonnelle, J. P., ‘‘Proceedings of the
Fourth Climax Molybdenum International Conference Sponsored
(Ann Arbor, Michigan)’’ (H. F. Barry and P. C. H. Mitchell, Eds.),
p. 261. University of Reading, England, 1982.
3
3
. Miles, R., J. Chem. Tech. Biotechnol. 30, 35 (1980).
3
3
3
1
1
0. Nakazawa, N., and Okamoto, H., Appl. Surf. Sci. 24, 75 (1985).
1. Leclercq, L., Provost, M., Pastor, H., Grimblot, J., Hardy, A. M.,
Gengembre, L., and Leclercq, G., J. Catal. 117, 371 (1989).
2. Boudart, M., Oyama, S. T., and Volpe, L., U.S. Patent 4, 515, 763, 1985.
3. Volpe, L., and Boudart, M., J. Solid State Chem. 59, 332 and 348
1
1
(
1985).
40. Rostrup-Neilsen, J. R., J. Catal. 31, 173 (1973).
41. Vitidsant, T., Laguerie, C., Gilot, B., and Damronglerd, S., Bull. Soc.
Chim. France 194 (1989).
1
1
4. Lee, J. S., Oyama, S. J., and Boudart, M., J. Catal. 106, 125 (1987).
5. Leclercq, L., Imura, K., Yoshida, S., Barbee, T., and Boudart, M., in