Aromatization of methane over supported and unsupported Mo-based catalysts

Article abstract of DOI:10.1006/jcat.1997.1478

The interaction of methane with unsupported and supported molybdenum compounds (Mo, MoO₂, MoO₃, Mo₂C, and MoC_(1-x)) has been investigated at 973 K. ZSM-5 was used as a support. Reaction products were analyzed using gas chromatography. Changes in the composition of catalyst samples were followed by X-ray photoelectron spectroscopy. Molybdenum metal and oxides interacted strongly with methane at 973 K to give H₂ (Mo) and H₂O and CO₂ (oxides), but only a trace amount of ethane. When these compounds were contacted with ZSM-5, the reaction pathway of methane initially was the same. Afterward, however, a dramatic change occurred in the product distribution: the formation of ethane, ethylene, and benzene came into prominence. This was particularly the case when these compounds were highly dispersed on ZSM-5. The selectivity to benzene was 80-85%. XPS analysis of Mo-containing catalysts demonstrated the formation of Mo carbides during the catalytic reaction. Unsupported Mo carbides behaved like metallic Mo; the dominant process was the decomposition of methane to hydrogen and carbon. The deposition of Mo₂C on ZSM-5 in a well-dispersed state, however, produced a very active and selective catalyst for the conversion of methane into benzene. The results suggest that Mo₂C is the active surface species in the Mo-containing catalysts, which converts methane into ethylene, the primary compound for the production of benzene on the zeolite surface.

Full text of DOI:10.1006/jcat.1997.1478

JOURNAL OF CATALYSIS 165, 150–161 (1997)
ARTICLE NO. CA971478
Aromatization of Methane over Supported and Unsupported
Mo-Based Catalysts
F. Solymosi, J. Csere´nyi, A. Szo¨ke, T. Ba´nsa´gi, and A. Oszko´
Institute of Solid State and Radiochemistry, A. Jo´zsef University, and Reaction Kinetics Research Group of the Hungarian
1
Academy of Sciences, P.O. Box 168, H-6701 Szeged, Hungary
Received March 4, 1996; revised June 19, 1996; accepted September 25, 1996
attributed to the migration of CHx species from the Rh onto
the Cu, and to the ease of the coupling of CHx fragments
on copper surface compared to Rh (9).
The interaction of methane with unsupported and supported
molybdenum compounds (Mo, MoO2, MoO3, Mo2C, and MoC(1ꢀx)
)
has been investigated at 973 K. ZSM-5 was used as a support. Re-
action products were analyzed using gas chromatography. Changes
in the composition of catalyst samples were followed by X-ray pho-
toelectron spectroscopy. Molybdenum metal and oxides interacted
strongly with methane at 973 K to give H2 (Mo) and H2O and CO2
(oxides), but only a trace amount of ethane. When these compounds
were contacted with ZSM-5, the reaction pathway of methane ini-
tially was the same. Afterward, however, a dramatic change oc-
curred in the product distribution: the formation of ethane, ethy-
lene, and benzene came into prominence. This was particularly the
case when these compounds were highly dispersed on ZSM-5. The
selectivity to benzene was 80–85%. XPS analysis of Mo-containing
catalysts demonstrated the formation of Mo carbides during the
catalytic reaction. Unsupported Mo carbides behaved like metal-
lic Mo; the dominant process was the decomposition of methane
to hydrogen and carbon. The deposition of Mo2C on ZSM-5 in a
well-dispersed state, however, produced a very active and selective
catalyst for the conversion of methane into benzene. The results sug-
gest that Mo2C is the active surface species in the Mo-containing
catalysts, which converts methane into ethylene, the primary com-
c
Similar studieson MoO3/ZSM-5catalystsproduced more
attractive results: it appeared that methane can be directly
transformed into aromatic products (10–15). Wang et al.
(10) reported 100% selectivity of benzene formation at
7.2% conversion, but details on the calculation of this
value were not given. They assumed that the activation of
methane proceeds via the polarization of the CH4 molecule
on the MoO3, and the reaction of the polarized molecule,
+
Hꢀ CH3 , on the Brønsted acid sites of H ZSM-5 zeolite
(10). In our study the maximum selectivity for benzene for-
mation (taking into account only the C-containing gaseous
products) was 61–75% at 973 K and at 2.0–5.7% CH4 con-
version (11). New findings, not observed in the previous
study (10), were as follows: (i) the reaction is preceded by
an induction period, (ii) the reduction of Mo6+ to a lower
oxidation state occurs and carbon is deposited during the
early stage of the contact of methane with the catalyst,
(iii) the high catalytic activity for methane–benzene con-
version is not restricted to MoO3/ZSM-5, as MoO3/SiO2
and even MoO3/Al2O3 exhibited reasonable activity for
this reaction. Similar features have been observed by Wang
et al. (14). For the selectivity of benzene formation they
obtained values between 60 and 65% . Recently, the con-
version of methane into higher hydrocarbons and benzene
was observed on K2MoO4/ZSM-5 catalyst with somewhat
lower CH4 conversion (1.2–6.3% at 973 K), but still with
high selectivity (60–67%) of benzene formation (12). In
this case the use of silica as a support led to an almost in-
active catalyst (CH4 conversion is <0.4–0.1% at 973 K),
which indicates the complex nature of this catalytic sys-
tem. Recently, Chen et al. reported on a more detailed
investigation (15). In this study the reduction of MoO3
and the deposition of carbon were also observed. It was
assumed that a synergistic effect between well-dispersed
molybdenum oxide species and the acidic sites of HZSM-5
is required for the catalysis of dehydrooligomerization of
methane.
–
–
pound for the production of benzene on the zeolite surface.
ꢁ 1997
Academic Press
1. INTRODUCTION
Recently many attempts have been made to activate
methane under nonoxidative conditions and to convert
it into higher hydrocarbons and aromatic compounds. In
the case of supported Pt metals the dehydrogenation of
methane occurs readily, yielding a small amount of ethane
in addition to the main products, H2, and surface carbon
(1–8). The subsequent hydrogenation of the most reactive
carbon gives several higher hydrocarbons and even ben-
zene (1–9). The formation of the latter compounds has been
enhanced byaddingCu metal to Rh/SiO2 catalyst which was
1 This laboratory is a part of the Center for Catalysis, Surface, and
Material Science at the Attila Jo´zsef University of Szeged.
150
0021-9517/97 $25.00
c
Copyright ꢁ 1997 by Academic Press
All rights of reproduction in any form reserved.

AROMATIZATION OF METHANE OVER Mo-BASED CATALYSTS
151
The above features of the reaction of methane and the
The gases used were of commercial purity (Linde). Ar
initial product formation clearly indicated that methane (99.996% ) and H2 (99.95% ) were deoxygenated with an
strongly interacted with MoO3 at the reaction temperature. oxytrap. The other impurities were adsorbed by a 5A
It appeared likely that the active phase for the methane molecular sieve at the temperature of liquid nitrogen. The
–
conversion is produced in the high temperature surface re- H ZSM-5 support was obtained by five times repeated ion
–
action. XPS analysis of the used catalyst did reveal the re- exchange of Na ZSM-5 (Si/Al = 55.0) with an aqueous so-
duction of Mo6+; the peaks for Mo5+, Mo4+, Mo2+, and Mo0 lution of ammonium nitrate (1 M), and calcined in air at
were identified in the XP spectra (11). Taking into account 863 K for 5 h. MoO3/ZSM-5 catalysts were prepared by im-
–
the large amount of carbon deposited on the catalyst in pregnating H ZSM-5 with a basic solution of ammonium
the course of the high temperature methane conversion, paramolybdate to yield a nominal 2 wt.% of MoO3. The
we may also reckon with the presence of Mo2C in the cata- suspension was dried at 373 K and calcined at 873 K for 5 h.
lysts. Recently, Wang et al. (14) identified the formation Before catalytic measurements, each sample was oxidized
of Mo2C from MoO3/ZSM-5 following the methane treat- in an O2 stream at 973 K in situ and then flushed with Ar
ment at 973 K by means of XPS, and we did the same from for 15 min.
K2MoO4/ZSM-5and from MoO3/SiO2 (12, 13). It appeared,
however, that the commercial Mo2C is rather inert toward the reduction of MoO3 (Aldrich) at 813 K in a thermobal-
the methane decomposition (13). ance until a constant weight-loss corresponding to the for-
MoO2 was a product of Aldrich. It was also prepared by
The present investigation was undertaken to elucidate mation of MoO2 (ꢂ2 h) was reached. Supported MoO2 was
the nature of the active surface species involved in the obtained from 2 wt.% MoO3/ZSM-5 by TPR: the highest
conversion of methane to benzene over Mo-containing reduction temperature reached was 808 K.
catalysts. For this purpose the effects of different Mo com-
While bulk MoO3 is readily reduced to metallic Mo by
pounds (metallic Mo, Mo oxides, and carbides) on the ac- hydrogen at 973 K, similar treatment of supported MoO3
tivation of methane and on the formation of benzene are does not give reduction below Mo4+ (16, 17). Supported
examined.
Mo in zerovalent state was prepared following the method
of Burwell et al. (16, 17) applied for the preparation of
–
Mo/Al2O3 ꢃ H ZSM-5 was dehydroxylated by the gradual
2. EXPERIMENTAL
heating of the sample in a dry Ar flow up to about 1173 K
and then Mo(CO)6 was evaporated onto the sample kept at
423 K. This was followed by the decomposition of Mo(CO)6
to Mo(CO)x in an argon stream at 423–573 K, and then by
hydrogenation at 973 K. In order to prevent the oxidation
of metallic Mo formed, the preparation was performed in
the catalytic reactor.
Hexagonal Mo2C was prepared by the method of Lee
et al. (18). Briefly, about 0.5 g of MoO3 was heated in
1 : 4 methane–H2 mixture flowing at 300 ml (STP)/min in
a quartz cell with two stopcocks. Preparation temperature
was increased rapidly to 773 K and at 30 K/h between 773
and 1023 K, and maintained at 1023 K for 3 h. Following
the suggestion of Lee et al. (18), the sample was deacti-
vated at 300 K with air, or used in situ for catalytic studies.
The BET surface area of this sample was 9.6 m2/g. Cubic
ꢀ-MoC1ꢀx(x = 0.45–0.5) was made by the carburization of
Mo2N powders (19). Mo2N was prepared by TPR between
MoO3 and ammonia (20).
Supported Mo2C catalysts were prepared in the follow-
ing ways. The first was a slurry method. ZSM-5 was sus-
pended in triply distilled water and the fine powder of
Mo2C (prepared as above) was added slowly to a stirred
slurry of ZSM-5. H2O was evaporated while the slurry was
stirred. The solid was homogenized and further dried in
an oven at 403 K and ground to a powder. The powder
was slightly pressed into wafers and calcined in situ. In
another case, Mo2C/ZSM-5 was produced by carburation
Catalytic reactions were carried out at 1 atm of CH4 pres-
sure in a fixed-bed, continuous flow reactor consisting of a
quartz tube (20-mm i.d.) connected to a capillary tube (11).
The flow rate was 12 ml/min. Generally 0.5 g of loosely com-
pressed catalyst sample was used. Reaction products were
analyzed gas chromatographically using a Hewlett-Packard
5890gaschromatograph and a Porapak QScolumn. In some
cases the initial interaction between the catalysts and the
reacting gases was followed by mass spectrometry using a
Balzers quadrupole mass spectrometer (type OMS 200).
The methane conversion was calculated from the H2 and
H2O balance. The selectivity values of product formation
represent the fraction of methane that has been converted
to specific hydrocarbon products taking into account the
number of carbon atoms in the molecules.
The XPS measurements were performed in a Kratos
XSAM 800 instrument at a base pressure of 10ꢀ8 Torr us-
ing MgKꢀ primary radiation (14 kV, 15 mA). To compen-
sate for possible charging effects, binding energies (BE)
were normalized with respect to the position of the Si(2p)
peak in SiO2 (BE = 103.4 eV) for supported samples and
to the Fermi-level for the Mo2C. The pass energy was
set at 40 eV, and an energy step width of 50 meV and
dwell time of 300 ms were used. Typically 10 scans were
accumulated for each spectrum. Fitting and deconvolu-
tion of the spectra were made using the VISION software
(Kratos).

152
SOLYMOSI ET AL.
of calcined MoO3/ZSM-5 in the catalytic reactor, in a way MoO3). The difference was that the evolution of H2 was
similar to that described above for the preparation of bulk observed at the very beginning of the reaction.
Mo2C. Mo2C/ZSM-5 catalysts were also prepared by the
Powders of Mo metal exhibited less activity toward
decarbonylation of the Mo(CO)6/ZSM-5 sample (obtained methane: the initial products were hydrogen and a small
as described above) and carburized in a CH4/H2 mixture amount of ethane. The rate of the formation of both com-
(20% CH4) at 973 K for 2 h (20). A detailed description of pounds decayed by a factor of 20 and 2, respectively, after
the properties of Mo2C samples can be found in the reviews 20 min (Fig. 1B).
of Omaya and Haller (21, 22).
3.2. Reaction of Methane on ZSM-5-Supported MoO3,
MoO2, and Mo0
3. RESULTS
Mixing of fine powders of MoO3 with ZSM-5 (weight
3.1. Reactions of Methane on Unsupported MoO3,
ratio 1 : 2) caused a dramatic change in the product distri-
MoO2, and Mo0 Metal
bution. In the first stage of the contact of methane with
the catalyst, which lasted up to 35–45 min, we observed the
same features as those for unsupported MoO3. Then the
evolution ofH2 (main product) started, together with that of
ethane, ethylene, benzene, and toluene. In the steady state
the conversion of methane was 6.0–7.5% and the selectivity
to benzene attained 83–85% . Similar catalytic performance
was observed when 2 wt.% MoO3 was dispersed on ZSM-5
by the impregnation method. The steady state conversion
(6.0–7.0% ) and the selectivity to benzene (83–85% ) were
higher on this catalyst than on the 2 wt.% MoO3/ZSM-5
samples prepared in our laboratory in previous studies
(11–13).
In order to detect the primary products of the high tem-
perature interaction of CH4 with Mo compounds, the re-
action was first followed by mass spectrometry. Results are
plotted in Fig. 1A. In the case of MoO3 (0.5 g), the product
formation started after 2 min yielding H2O (amu 18), CO2
(amu 44), and a minor amount of C2H6 (amu 30). Max-
imum rate of their formation was reached at 20–25 min.
Afterward, the amount of H2O decreased suddenly while
that of CO2 and ethane only moderately. At the same time,
the evolution of H2 was observed.
Similar products were identified in the reaction of
methane with MoO2 (prepared by the reduction of 0.5 g
FIG. 1. Interaction of CH4 with unsupported (A) MoO3 (0.5 g) and (B) Mo metal (produced by the reduction of 0.5 g MoO3) at 973 K followed
by mass spectrometric analysis.

AROMATIZATION OF METHANE OVER Mo-BASED CATALYSTS
153
FIG. 2. Conversion of methane, rate, and selectivity of the formation of various products on (A) MoO2/ZSM-5 by mixing MoO2 with ZSM-5 in
1 : 2 ratio by the slurry method, and (B) MoO2/ZSM-5 prepared by the partial reduction of 2 wt.% MoO3/ZSM-5 in situ.
When MoO2 was mixed with ZSM-5 by the slurry method with Tp = 723–753 K, then the formation of methane. The
(1 : 2 mol ratio), we obtained a high initial methane conver- hydrogenation was continued at 973 K until methane pro-
sion (15–20% ), which decreased to 0.5% in 40 min. During duction was detected, which took several hours. This pro-
this active period only H2O, CO2, and CO were formed: cedure was followed by a flushing out of the hydrogen with
ethylene, ethane, propane, and benzene started to evolve argon for 15–30 min. The Mo/ZSM-5 catalyst so formed
only after 40 min. The selectivity to benzene in this stage exhibited a high initial activity toward methane decompo-
attains a value of 35% (Fig. 2A).
sition at 973 K (Fig. 3B). The conversion, determined from
Preparation of MoO2 in situ on ZSM-5 support (by the the amount of hydrogen produced at a reaction time of
reduction of highly dispersed 2 wt.% MoO3) resulted in a 3 min, was around 35% . Trace amounts of ethane and ethy-
better catalyst (Fig. 2B). The decomposition of methane lene were also detected, but only about 0.057% of the de-
started with a maximum conversion (ꢂ14% ), which de- composed methane wastransformed into these compounds.
cayed to about 7.5–6.0% . At the beginning of the reaction, When the next sample was taken for analysis at 24 min, the
the evolution of H2O, CO2, and CO was observed in this conversion ofmethane dropped to 0.25% . At the same time,
case as well. Afterward, the rate of hydrocarbon formation the rate of ethane production also decreased by a factor of
significantly increased. The main hydrocarbon product was 5, whereas that of ethylene formation increased by a factor
benzene with a constant selectivity of 80–85% .
of 4. Afterward, both rates remained practically constant in
As supported metallic Mo cannot be prepared from the the time of methane stream (Fig. 3B). A small amount of
reduction of MoO3 on supports (16, 17), Mo/ZSM-5 was propane was also detected, but benzene was identified only
produced from the decarbonylation of Mo(CO)6. As this is in trace amounts.
a rather complicated procedure, in Fig. 3A we present the
When this sample was oxidized at 973 K after the reac-
data obtained for the various steps of the preparation of tion, we measured an initial methane conversion of 11.5%
this catalyst. It shows that the dehydroxylation of ZSM-5 is and a steady state conversion of 7% . The product dis-
fast at 473–573 K, but is completed only above 873 K. The tribution was nearly the same as observed for the 2%
rate of decarbonylation of Mo(CO)6 to Mo(CO)x exhibits MoO3/ZSM-5 catalyst; actually the selectivity to benzene
a maximum at 523 K; CO evolution ceases around 573 K. was somewhat higher, 85–86% , and that to toluene was
In the subsequent decomposition of Mo(CO)x in hydro- around 6% . Another feature is that the steady state values
gen to metallic Mo, first we observed the evolution of H2O remained constant for several hours.

154
SOLYMOSI ET AL.
FIG. 3. (A) Dehydroxylation of ZSM-5 (a); decomposition of Mo(CO)6 deposited on ZSM-5 (b) and decomposition of Mo(CO)x in H2 stream (c).
(B) Conversion of methane and rate of the formation of various products on Mo0/ZSM-5 at 973 K.
3.3. Reactions of Methane on Unsupported Mo2C Samples ration of Mo2C a significant amount of excess carbon is
deposited on Mo2C (18), which may influence its catalytic
As reported earlier, Mo2C (Aldrich) exhibited very lit- performance. This excess carbon can be removed by hy-
tle activity toward the decomposition of methane even at drogenation at 873 K. This treatment led to an appreciable
973 K (12). The initial conversion was very low (0.1–0.2% ) increase in the catalytic activity of Mo2C (the initial conver-
and decreased further with time.
sion of methane attains 6–8% ), but it also decreased with
A higher initial activity was measured for the Mo2C pre- time. A small amount of ethane was also detected: inter-
pared by the method of Lee et al. (18). The reaction started estingly, its formation rate exhibited a maximum with time
with a maximum rate (3–4% conversion at 973 K) which on methane stream. Variation of the temperature gave the
gradually decayed to a very low value, 0.1–0.3% . The only same picture. The C2H6/H2 ratio at the maximum of ethane
gaseous product was H2. Ethane, the other product ob- production was about 2.0 ꢄ 10ꢀ3 at 943 K, and increased
served in the decomposition of CH4 on supported Pt metals with the reaction temperature to the value of 6.0 ꢄ 10ꢀ3 at
(1–8), was not formed in detectable amounts. No change 973 K. Figure 4 shows the formation of H2 and C2H6 in
in the product formation occurred after several hours of the decomposition of methane at different temperatures.
reaction. It was observed before that during the prepa- A new sample was used for each temperature.

AROMATIZATION OF METHANE OVER Mo-BASED CATALYSTS
155
FIG. 4. Decomposition of CH4 on Mo2C at different temperatures. 0.5 g Mo2C was used.
The decomposition of methane on Mo2C at 973 K pro-
duced additional surface carbon, which reacted with hydro-
gen to give methane above 800 K. No other products of the
hydrogenation of this carbon were identified.
Some experiments have been performed with the cubic
MoC1ꢀx sample. We obtained a very high initial conversion,
ꢂ28–30% , which subsequently decreased to about 2.0–
2.5% . The main product was hydrogen. CO was also pro-
duced initially: its formation practically ceased after 20 min.
Small amounts of ethane and ethylene were also produced
(Fig. 5). The C2Hx/H2 ratio varied between 2.8 ꢄ 10ꢀ4 and
6.5 ꢄ 10ꢀ4
.
3.4. Reactions of Methane on Supported Mo2C
First the catalytic behavior of 2 wt.% Mo2C/ZSM-5 pre-
pared by the slurry method was tested. The conversion of
CH4 at the first point of the product analysis (ꢂ2 min)
was around 1.0% . In this case only a small amount of
ethane was detected. Afterward, the conversion decreased
to about 0.6% (60 min) and, besides hydrogen, the forma-
tion of all the hydrocarbons identified on the most active
MoO3/ZSM-5 catalysts occurred (Fig. 6A). The selectiv-
ity to benzene was around 53–55% . When this sample was
treated with hydrogen at 873 K before the catalytic test, to
remove the excess carbon, the conversion of methane deter-
mined after 2 min contact time was about 5% , and ethane,
FIG. 5. Conversion of methane, rate, and selectivity of the formation
ethylene, and benzene were also formed at the very begin- of various products on ꢀ-MoC1ꢀx at 973 K.

156
SOLYMOSI ET AL.
FIG. 6. Conversion of methane, rate, and selectivity of the formation of various products on Mo2C/ZSM-5 at 973 K. The catalyst was prepared by
mixing Mo2C with ZSM-5 by the slurry method. The Mo2C content was 1.5 wt.% . (A) No H2 treatment, (B) H2 treatment before the catalytic test at
873 K for 2 h.
ning of the reaction (Fig. 6B). The conversion of methane, that to ethylene was below 2% . In Fig. 7 we show the rates
however, suddenly decreased below 1% at the next analysis and slectivities of the formation of various products on this
(17 min). At the same time, the rate of ethylene formation catalyst.
increased and that ofethane production decreased. The rate
When Mo2C was prepared by decarbonylation and car-
of benzene production slowly decreased after a transitory burization of Mo(CO)6 deposited on ZSM-5 we also ob-
maximum. The selectivity to benzene diminished from 47 tained an active and selective catalyst. The initial conver-
to 40% in 120 min.
sion of methane was 7.5% , which decayed to 2% in 120 min
When the amount of Mo2C was significantly increased (Fig. 8). In addition to hydrogen, the hydrocarbons ethane,
(the Mo2C/ZSM-5 weight ratio was 1 : 2), we observed sim- ethylene, propane, benzene, and toluene were produced.
ilar picture. The initial conversion of methane on the pre- H2O, CO, and CO2 were not detected even in the initial
reduced catalyst attained 18% , but it decayed soon to 1% . phase ofthe reaction. (Note that we found these compounds
The product distribution and selectivities were practically when the catalyst sample was prepared not in situ and was
the same as before.
exposed to air.) The selectivity to benzene was high, >85% ,
A completely different picture was obtained when Mo2C and constant during the experiment. The reoxidized cata-
was prepared from 2 wt.% MoO3 deposited on ZSM-5 by lyst exhibited the same behavior as MoO3/ZSM-5 prepared
the impregnation of molybdenum compound. The induc- by impregnation. After an induction period, the conversion
tion period observed for supported MoO3 was missing: the of methane attained a value of 10–12% .
reaction started with the highest rate. The initial conver-
Similarly as in the previous papers (10–15), in the cal-
sion of methane was about 10% which diminished slowly culation of the selectivities to various products only the
to 5–6% . At the very beginning of the reaction CO was C-containing gaseous compounds were taken into account.
also produced but its amount soon decreased to a low level. During the CH4 conversion, however, a significant fraction
The other products of the reaction of methane were H2, of CH4 decomposed to carbon. In order to be clear as re-
C2H6, C2H4, and C6H6. The rates of their formation in- gards the reaction pathways of CH4, we evaluated quan-
creased with time on stream. The selectivity to benzene titatively the amount of CH4 transformed into carbon de-
became practically constant after 15 min and reached a posit and also into hydrocarbons and aromatic compounds.
value of 85% . The selectivity to ethane was 17–20% and Data are given in Table 1. It appears that the complete

AROMATIZATION OF METHANE OVER Mo-BASED CATALYSTS
157
FIG. 7. Conversion of methane, rate, and selectivity of the formation of various products on Mo2C/ZSM-5 at 973 K. Mo2C was prepared by the
carburization of 2 wt.% MoO3 on ZSM-5.
FIG. 8. Conversion of methane, rate, and selectivity of the formation of various products on Mo2C/ZSM-5 at 973 K. Mo2C was prepared by
decarbonylation and carburization of Mo(CO)6 deposited on ZSM-5.

158
SOLYMOSI ET AL.
TABLE 1
Characteristic Data for the Conversion of CH4 into Various Products on Different Mo-Based Catalysts at 973 K
Amount of CH4 transformed [% ]
Conversion [% ]
Ethylene
Ethane
Benzene
Toluene
Coke
Catalysts
2 min
60 min
2 min
60 min
2 min
60 min
2 min
60 min
2 min
60 min
2 min
60 min
MoO3/ZSM-5
MoO2/ZSM-5
Mo/ZSM-5
Mo2C/ZSM-5 from
MoO3/ZSM-5
Mo2C/ZSM-5 from
Mo(CO)6
1.6
14.0
35.5
9.72
4.15
6.42
0.12
5.71
0
0
0.01
0.37
5.52
3.49
7.83
1.74
0.59
0.15
0.05
0.97
2.92
2.29
3.07
2.09
0
0
0
36.8
39.5
0
0
0
0
0.2
2.76
2.23
0
0
50.5
52.4
89.2
42.86
39.5
99.9
88.39
10.1
50.0
3.15
7.23
2.62
0.19
1.26
10.15
5.06
12.69
57.32
0.33
3.18
76.9
40.8
decomposition of CH4 is significant even for the Mo2C de- the MoO3 was deposited on ZSM-5 and calcined in air at
posited on ZSM-5.
873 K, we observed a slight shift of these binding energies
(BE) to 235.7 and 232.7 eV. On treatment of this sample
with CH4 at 973 K for 2 h, the BE values for Mo(3d) were
registered at 231.2 and 227.9 eV. A strong C(1s) peak ap-
peared at 283.8 eV. Note that the carbon contamination of
MoO3 and MoO3/ZSM-5 gave a peak at higher energies,
285.1–284.9 eV. The BE values for Mo(3d) in MoO2 were
found at 232.9 and 229.8 eV, and that for 0(1s) at 529.8 eV,
corresponding well to literature data for MoO2 (23, 24).
When MoO2/ZSM-5 was applied as a catalyst after 1 h of
reaction with methane, we obtained practicallythe same BE
3.5. Characterization of Mo Catalysts
The surface area of the Mo2C sample prepared by us
is about 4.5–5.0 times larger than that of the commercial
Mo2C.
All the catalysts used in this study have been character-
ized by XPS measurements before and after the reaction.
Some spectra are presented in Fig. 9. For oxidized, unsup-
ported MoO3 we found the doublet for Mo(3d) at 236.2 and
233.0 eV. The BE for 0(1s) was located at 530.8 eV. When
FIG. 9. XPS spectra of Mo-containing catalysts. (1) MoO3 (oxidized at 973 K for 1 h); (2) 2 wt.% MoO3/ZSM-5 (oxidized at 973 K for 1 h);
(3) MoO2 (Aldrich, bombarded with argon ions for 15 min); (4) Mo2C (Aldrich); (5) Mo2C (prepared by the method of Lee et al. (18)); (6) Mo2C/ZSM-
5 (Mo2C was prepared from Mo(CO)6); (7) 2 wt.% MoO3/ZSM-5 treated with CH4 at 973 K for 2 h. Samples (4–7) have been treated with H2 at 873 K
for 2 h to remove the excess carbon.

AROMATIZATION OF METHANE OVER Mo-BASED CATALYSTS
159
final conclusion as regards the active sites and the possible
pathways of the transformation of methane into ethylene,
let us summarize the main findings of previous and present
studies.
TABLE 2
XPS Data for Mo Catalysts
Binding energy (eV)
Sample
MoO3
MoO3/ZSM-5
MoO2
Mo2C (Aldrich)
Mo2Ca
Mo2C/ZSM-5b
MoO3/ZSM-5c
Mo(3d5/2
)
Mo(3d3/2
)
C(1s)
4.1. Characteristic Features of the Reaction of Methane
on Mo Oxides and Mo Metal
233.0
232.7
229.8
228.0
227.8
227.6
227.9
236.2
235.7
232.9
231.1
231.0
230.7
231.2
The results of the present study clearly show that
methane reacts readily with unsupported MoO3 and MoO2
at high temperature to give H2O and CO2. The product for-
mation in the initial stage of the surface reaction can be
described by the overall equation
283.85
283.85
284.5
283.8
a From MoO3 [18].
b From Mo(CO)6 [19].
CH4 + 4Os = CO2 + 2H2O.
[1]
c Treated with CH4 at 973 K for 2 h.
The obvious first step in the activation of CH4 is an oxidative
dehydrogenation process
as for MoO3/ZSM-5 treated in the same way. The situation
was similar when the Mo0/ZSM-5 catalyst was examined
after its use as a catalyst.
CH4 + O(s) = CH3 + OH
followed by the recombination of OH groups
2OH = H2O + O(s)
[2]
In Fig. 9 we also show the XPS spectra for Mo2C sam-
ples received from Aldrich and prepared by us. The binding
energies of Mo(3d5/2) were 228.0 eV (Aldrich) and 227.8–
227.6 eV, respectively. A value of 227.9 eV was measured
for the MoO3/ZSM-5 catalyst used for CH4 conversion.
Note that these values are slightly lower than obtained
previously for Mo2C prepared by other methods (229.0 eV
(24) and 228.1–228.3 (25)), but agree well with the value
(227.8 eV) reported by Wang et al. (14) for Mo2C produced
from MoO3/ZSM-5 during the CH4 conversion into ben-
zene. The C(1s) peaks were relatively broad indicating the
coexistence of several forms of C with different binding en-
ergies. Following the XPS work of Bonzel and Krebs (26)
three C species can be distinguished (24, 25): carbidic car-
bon (283.3 eV), polymeric carbon (284.5 eV) and amor-
phous or graphitic carbon (285.0 eV). For our Mo2C sam-
ples we obtained a value of 283.8 eV, which is close to
the value attributed to Mo2C (14, 24, 25). XPS data for
Mo-containing catalysts used in this study are collected in
Table 2.
[3]
and by the oxidation of CH3 species
CH3 + Os = CH + H2O
[4]
[5]
1
CH + 2Os = CO2 + H2
2
by surface oxygen. Competing processes are the dehydro-
genation of CH3 species
CH3 = CH2 + H
CH2 = CH + H
CH = C + H
2H = H2
[6]
[7]
[8]
[9]
as indicated by the evolution of H2.
The respective couplings of CH3 and CH2 species
4. DISCUSSION
2CH3 = C2H6
2CH2 = C2H4
[10]
[11]
The primary aim of this work was to elaborate the na-
ture of the active surface site involved in the conversion of
methane to benzene over MoO3/ZSM-5 catalysts. The ba-
sic requirements for that center are to activate the methane on molybdenum oxides are negligible as compared to the
molecule to produce CH3 and CH2 fragments, to exhibit less oxidation and decomposition steps of CHx fragments, but
activity toward the subsequent decomposition of these CHx become detectable on partially reduced surfaces. In the case
species to carbon, and to promote their couplings to ethane of the Mo metal, the dominant process is the decomposition
and particularly to ethylene. Taking into account the high of methane to carbon and hydrogen, similarly as on sup-
–
efficiency of the acidic sites of H ZSM-5 in the oligomer- ported Pt metals (1–9); only about 7–8% of decomposed
ization and cyclization reactions (27, 28), it was envisaged methane is transfomed into ethane and ethylene.
that the transformation of ethylene into benzene is accom-
In all three cases a significant decay in the rate of CH4
plished on the ZSM-5 surface (12–15). Before making a reaction and changes in the product distribution were

160
SOLYMOSI ET AL.
experienced after the initial interaction of methane with the ment at 873 K (Fig. 4). In this case, however, we cannot
catalysts at high temperature. As a result, the formation of exclude the possibility that the high temperature treatment
C2H6 and C2H4 in very small amounts was observed.
with H2 removes not only the excess carbon, but also a frac-
The above results showed that although Mo oxides and tion of surface carbon of Mo2C, creating very active centers
metallic Mo can activate the CH4 molecule and catalyze its for the subsequent decomposition of CH4. This latter reac-
decomposition, neither of these catalysts alone is selective tion restores the surface stoichiometry of Mo2C leading to
enough to produce C2H6 and/or C2H4, the assumed start- a much less active catalyst. The fact that MoC1ꢀx was found
ing componds of benzene and aromatics, in larger quantity to be considerably more active in the decomposition of CH4
(>1% of decomposing methane).
(Fig. 5) than the Mo2C may support this speculation. The
When Mo oxides were mixed with ZSM-5, or particu- catalytic activity of this sample, however, also soon dimin-
larly when they were dispersed on the ZSM-5 support, ished. Coupling reactions of CHx species were very limited
at the very beginning of the contact of methane with the on this sample, too.
catalyst we observed similar features. In the latter stages,
The situation was completely different when Mo2C was
however, the production of hydrocarbons, particularly ben- added to ZSM-5 or deposited on its surface. In these cases,
zene, with a selectivity of 60–80% , came into prominence. in addition to H2, we detected all the hydrocarbons (ethy-
This dramatic change in the catalytic behavior can be at- lene, ethane, benzene, and small amounts of propane and
tributed to the formation of a new compound from the toluene), the formation of which was observed on the highly
catalyst during the high temperature reaction. The situa- active 2 wt.% MoO3/ZSM-5 catalyst. As in these cases the
tion was different for Mo0/ZSM-5. In this case the initial Mo2C samples were prepared by the carburization of sup-
interaction of Mo0 with methane at 973 K was followed ported or unsupported MoO3, one can assume that the high
by a drastic deactivation as indicated by the decrease catalytic performance of Mo2C/ZSM-5 samples is due to the
of CH4 conversion from 35 to 0.25% in 24 min. It ap- presence of unreacted MoO3 and/or Mo2C/MoOx contact.
pears that the deposition of coke and/or polymeric car- This idea was apparently supported by the observation that
bon (18) in a large quantity poisoned the active sites of the a partial oxidation of a less active Mo2C mixed with ZSM-5
catalyst.
improved the yield of hydrocarbons and benzene produced
(13). However, in the present case, we could not identify
the formation of H2O and CO2 even at the beginning of the
reaction. Therefore, we can exclude the dominant role of
MoO3 in the catalytic performance of Mo2C/ZSM-5.
This conclusion may be supported by the results obtained
for Mo2C/ZSM-5 prepared in situ by decarbonylation and
carburization of Mo(CO)6 (Fig. 8). This sample also proved
to be an active and selective catalyst in the conversion of
methane into benzene. In this case we cannot count on the
presence of Mo oxides, although if the reactive OH groups
4.2. Catalytic Behavior of Mo2C and Mo2C/ZSM-5
XPS examination of MoOx-containing catalysts sup-
ported by ZSM-5 clearly showed that the oxides are not
only reduced by methane at 973 K, but at least the sur-
face layers of oxides have been transformed into some kind
of Mo carbide. The same holds when the starting catalyst
was Mo0/ZSM-5. The binding energies of Mo(3d) and C(1s)
agree with those obtained for commercial Mo2C and for
Mo2C samples prepared by us. This is not surprising as the
conditions used for the preparation of Mo2C are nearly the
same as applied for the catalytic conversion of methane.
Accordingly, we may assume that the dramatic change in
the product distribution following the strong interaction of
methane with supported Mo oxides is due to the formation
of Mo2C.
2+
of the support are not removed completely, 2ꢁꢀO– Mo
–
species can be transitorily formed during the preparation
in the reaction
2ꢁ–OH + Mo(CO)3 → 2ꢁ–Oꢀ–Mo2+ + 3CO + H2, (12)
as it was postulated for alumina (17). However, we did not
In a preliminary study, however, Mo2C exhibited very find H2O, CO, and CO2 in the reaction products which sug-
–
little activity toward methane decomposition, and mixing gests that the Mo O species was not present.
it with ZSM-5 brought no positive result (13). This Mo2C, Accordingly, we may conclude that Mo2C on the ZSM-5,
however, was a commercial product of very low surface produced by the transformation of highly dispersed MoO3,
area (ꢂ1.1 m2/g). The use of a more advanced method in is the active site for the cleavage of the C H bond in the
–
the preparation of Mo2C resulted in a more active catalyst methane molecule, but it is far less active for the further
toward CH4 reactions. However, the dominant process was decomposition of CH3 and CH2 species produced. As a re-
still the complete decomposition of methane to give H2, C, sult, these radicals have a certain lifetime to recombine into
and a small amount of ethane (C2H6/H2 ratio was between C2H6 and C2H4 either on the catalyst surface or in the gas
2.0 ꢄ 10ꢀ4 and 6.0 ꢄ 10ꢀ4). The initial catalytic activity of phase. Further reactions, the oligomerization and aroma-
Mo2C was greatly enhanced when the excess carbon de- tization of ethylene, occur on the ZSM-5 support (12–15,
posited during the preparation was removed by H2 treat- 27, 28).

AROMATIZATION OF METHANE OVER Mo-BASED CATALYSTS
161
5. CONCLUSIONS
3. Belgued, M., Pareja, P., Amariglio, A., and Amariglio, H., Nature 352,
789 (1991).
4. Solymosi, F., Kutsa´n, Gy., and Erdo¨helyi, A., Catal. Lett. 11, 149
(1991).
5. Solymosi, F., Erdo¨helyi, A., and Csere´nyi, J., Catal. Lett. 16, 399
(1992).
6. Erdo¨helyi, A., Csere´nyi, J., and Solymosi, F., J. Catal. 141, 287 (1993).
7. Solymosi, F., Erdo¨helyi, A., Csere´nyi, J., and Felve´gi, A., J. Catal. 147,
272 (1994).
8. Solymosi, F., and Csere´nyi, J., Catal. Today 21, 561 (1994).
9. Solymosi, F., and Csere´nyi, J., Catal. Lett. 34, 343 (1995).
10. Wang, L., Tao, L., Xie, M., Xu, G., Huang, J., and Xu, Y., Catal. Lett.
21, 35 (1993); Xu, Y., Liu, S., Wang, L., Xie, M., and Guo, X., Catal.
Lett. 30, 135 (1995).
1. MoO3 and MoO2 interact strongly with methane at
973 K to give H2O and CO2 with trace amounts of ethylene
and ethane. The reaction of methane on metallic Mo pro-
duces H2 with a very small amount of ethane. The catalyst
is soon poisoned due to carbon deposition.
2. Mixing of these compounds with ZSM-5 changes the
product distribution, and the formation of ethylene, ethane,
and benzene comes into prominence. High selectivity to
benzene was attained when these Mo compounds were dis-
persed on ZSM-5.
3. XPS results suggest that the major part of molybdenum
metaland oxidesistransformed into Mo carbidesduringthe
high temperature reaction with methane.
4. Unsupported Mo2C and MoC(1ꢀx) catalyzed the de-
composition of methane to hydrogen, carbon, and to a mi-
nor amount of ethane.
11. Solymosi, F., Erdo¨helyi, A., and Szo¨ke, A., Catal. Lett. 32, 43 (1995).
12. Solymosi, F., and Szo¨ke, A., 39, 157 (1996).
13. Solymosi, F., Szo¨ke, A., and Csere´nyi, J., Appl. Catal. A 142, 361 (1996).
14. Wang, D., Lunsford, J. H., and Rosynek, M., paper presented at the
14th Meeting of the North American Catalysis Society, Snowbird,
1995; Lunsford, J. H., Rosynek, M. P., and Dingjun, W., paper pre-
sented at the 4th Intern. Natural Gas Conversion, Kruger National
Park, South Africa, 1995.
15. Chen, L., Lin, L., Xu, Z., Li, X., and Zhang, T., J. Catal. 157, 190
(1995).
16. Burwell, R. L., and Brenner, A., J. Mol. Catal. 1, 77 (1975/76).
5. Preparation of Mo2C on ZSM-5 in high dispersion pro-
duced catalyst which is active and selective in the aroma-
tization of methane (selectivity to benzene is 80–85% at a
conversion of 6.0–7.0% ). It appears that Mo2C is respon- 17. Bowman, R. G., and Burwell, R. L., J. Catal. 63, 463 (1980).
18. Lee, J. S., Oyama, S. T., and Boudart, M., J. Catal. 106, 125 (1987).
19. Lee, J. S., Locatelli, S., Oyama, S. T., and Boudart, M., J. Catal. 125,
sible for the production of ethylene from methane which
is converted to benzene on the acidic sites of the ZSM-5
support.
157 (1990).
20. Volpe, L., and Boudart, M., J. Solid State Chem. 59, 348 (1985).
21. Oyama, S. T., and Haller, G. L., in “Catalysis, Specialist Periodical
ACKNOWLEDGMENTS
Reports” (G. C. Bond and G. Webb, Eds.), Vol. 5, p. 333. The Chemical
Society, London, 1981.
This work was supported by the Hungarian Academy of Sciences and
by the OTKA 2038.
22. Oyama, S. T., Catal. Today 15, 179 (1992).
23. Spevack, P. A., and McIntyre, N. S., J. Phys. Chem. 97, 11020 (1993).
24. Leclercq, L., Provost, M., Pastor, H., Grimblot, J., Hardy, A. M.,
Gengembre, L., and Leclercq, G., J. Catal. 117, 371 (1989).
25. Ledoux, M. J., Huu, C. P., Guille, J., and Dunlop, H., J. Catal. 134, 383
(1992).
REFERENCES
1. van Santen, R. A., de Koster, A., and Koerts, T., Catal. Lett. 7, 1 (1990);
Koerts, T., and van Santen, R. A., “Proc. 10th Int. Congr. on Catalysis 26. Bonzel H. P., and Krebs, H. J., Surf. Sci. 91, 499 (1980).
(Budapest 1992)” (L. Guczi, F. Solymosi, and P. Te´te´nyi, Eds.), p. 1065, 27. Anderson, J. R., Foger, K., Mole, T., Rajadhyaksha, R. A., and Sanders,
(Akade´miai Kiado´, Budapest, 1993).
2. Koerts, T., Deelen, M. J. A. G., and van Santen, R. A., J. Catal. 138,
101 (1992).
J. V., J. Catal. 58, 14 (1979).
28. Chu, P., U.S. Pat. 4120910 (1978); Chester, A. W., and Chu, Y. F., U.S.
Pat. 4350835 (1982).