Bifunctional catalysis of Mo/HZSM-5 in the dehydroaromatization of methane to benzene and naphthalene XAFS/TG/DTA/ MASS/FTIR characterization and supporting effects

Article abstract of DOI:10.1006/jcat.1998.2310

The direct conversion of methane to aromatics such as benzene and naphthalene has been studied on a series of Mo-supported catalysts using HZSM-5, FSM-16, mordenite, USY, SiO₂, and Al₂O₃ as the supporting materials. Among all the supports used, the HZSM-5-supported Mo catalysts exhibit the highest yield of aromatic products, achieving over 70% total selectivity of the hydrocarbons on a carbon basis at 5-12% methane conversion at 973 K and 1 atm. By contrast, less than 20% of the converted methane is transformed to hydrocarbon products on the other Mo-supported catalysts, which are drastically deactivated, owing to serious coke formation. The XANES/EXAFS and TG/DTA/mass studies reveal that the zeolite-supported Mo oxide is endothermally converted with methane around 955 K to molybdenum carbide (Mo₂C) cluster (Mo-C, C.N. = 1,R = 2.09 A; Mo-Mo, C.N. = 2.3-3.5; R = 2.98 A), which initiates the methane aromatization yielding benzene and naphthalene at 873-1023 K. Although both Mo₂C and HZSM-5 support alone have a very low activity for the reaction, physically mixed hybrid catalysts consisting of 3 wt% Mo/SiO₂ + HZSM-5 and Mo₂C + HZSM-5 exhibited a remarkable promotion to enhance the yields of benzene and naphthalene over 100-300 times more than either component alone. On the other hand, it was demonstrated by the IR measurement in pyridine adsorption that the Mo/HZSM-5 catalysts having the optimum SiO₂/Al₂O₃ ratios, around 40, show maximum Broensted acidity among the catalysts with SiO₂/Al₂O₃ ratios of 20-1900. There is a close correlation between the activity of benzene formation in methane aromatization and the Broensted acidity of Mo/HZSM-5, but not Lewis aciditiy. It was found that maximum benzene formation was obtained on the Mo/HZSM-5 having SiO₂/Al₂O₃ ratios of 20-49, but substantially poor activities on those with SiO₂/Al₂O₃ ratios smaller and higher than 40. The results suggest that methane is dissociated on the molybdenum carbide cluster supported on HZSM-5 having optimum Broensted acidity to form CH_x (x > 1) and C₂-species as the primary intermediates which are oligomerized subsequently to aromatics such as benzene and naphthalene at the interface of Mo₂C and HZSM-5 zeolite having the optimum Broensted acidity. The bifunctional catalysis of Mo/HZSM for methane conversion towards aromatics is discussed by analogy with the promotion mechanism on the Pt/Al₂/O₃ catalyst for the dehydro-aromatization of alkanes.

Full text of DOI:10.1006/jcat.1998.2310

Journal of Catalysis 181, 175–188 (1999)
Article ID jcat.1998.2310, available online at http://www.idealibrary.com on
Bifunctional Catalysis of Mo/HZSM-5 in the Dehydroaromatization
of Methane to Benzene and Naphthalene XAFS/TG/DTA/
MASS/FTIR Characterization and Supporting Effects
Shetian Liu,1 Linsheng Wang,2 Ryuichiro Ohnishi, and Masaru Ichikawa3
Catalysis Research Center, Hakkaido University, Kita-Ku, N-11, W-10, Sapporo 060, Japan
Received March 30, 1998; revised September 12, 1998; accepted October 19, 1998
Mo/HZSM for methane conversion towards aromatics is discussed
by analogy with the promotion mechanism on the Pt/Al2O3 catalyst
The direct conversion of methane to aromatics such as benzene
and naphthalene has been studied on a series of Mo-supported cata-
lysts using HZSM-5, FSM-16, mordenite, USY, SiO2, and Al2O3
as the supporting materials. Among all the supports used, the
HZSM-5-supported Mo catalysts exhibit the highest yield of aro-
matic products, achieving over 70% total selectivity of the hydro-
carbons on a carbon basis at 5–12% methane conversion at 973 K
and 1 atm. By contrast, less than 20% of the converted methane is
transformed to hydrocarbon products on the other Mo-supported
catalysts, which are drastically deactivated, owing to serious coke
formation. The XANES/EXAFS and TG/DTA/mass studies reveal
that the zeolite-supported Mo oxide is endothermally converted
with methane around 955 K to molybdenum carbide (Mo2C) cluster
c
for the dehydro-aromatization of alkanes.
1999 Academic Press
Key Words: bifunctional role; methane conversion; benzene; Mo/
ZSM-5; methane aromatization; Mo/HZSM-5 catalysts; XAFS/
TG/DTA/IR characterization.
1. INTRODUCTION
The catalytic conversion of methane to petrochemical
feed stocks such as ethylene and benzene has been of cur-
rent importance and industrial interest in the effective uti-
lization of carbon resources of natural gas. Several one-step
catalytic processes converting methane to valuable prod-
ucts have been studied during the past 20 years, including
the oxidative coupling of methane (OCM) to ethene (or
ethane), selective oxidation of methane to methanol/form-
aldehyde, and the homologation of methane to lower hy-
drocarbons (1–6). Nevertheless the industrial interests in
these processes have been limited so far because of low
product selectivity at higher conversion of methane. During
the past five years, the nonoxidative conversion of methane
to benzene has been reported by Wang, Xu, and Guo (7, 8),
Solymosi (9–12) , and Lunsford (16, 17) on HZSM-5 zeolite-
supported Mo catalysts in conjunction with their reac-
tivities, reaction mechanism, and catalyst characterization
by XPS (10, 12), XRD, and NMR (13, 14). Moreover,
there have been subsequent works by Solymosi (9–12)
and Lunsford (16, 17) on the role of the acidity sites of
HZSM-5 and the carburization mechanism of Mo oxide
on HZSM-5 with methane to form Mo carbide which is
proposed to be an active species for the dehydroconden-
sation of methane towards benzene on Mo/HZSM cata-
lysts. Recent works by Lunsford (17) and Ichikawa (18, 19)
demonstrated that methane is actually dehydrocondensed
into various aromatics, not only benzene, but also naphtha-
lene on Mo/HZSM-5and the Fe/Co promoted Mo/HZSM-5
at 873 K–973 K. Since the pioneering works by Wang, Liu,
and Xu in 1993 (7), a number of previous papers have
˚
˚
(Mo-C, C.N. = 1, R = 2.09 A; Mo-Mo, C.N. = 2.3–3.5; R = 2.98 A),
which initiates the methane aromatization yielding benzene and
naphthalene at 873–1023 K. Although both Mo2C and HZSM-5
support alone have a very low activity for the reaction, physically
mixed hybrid catalysts consisting of 3 wt% Mo/SiO2 + HZSM-5 and
Mo2C + HZSM-5 exhibited a remarkable promotion to enhance the
yields of benzene and naphthalene over 100–300 times more than
either component alone. On the other hand, it was demonstrated by
the IR measurement in pyridine adsorption that the Mo/HZSM-5
catalysts having the optimum SiO2/Al2O3 ratios, around 40, show
maximum Bro¨nsted acidity among the catalysts with SiO2/Al2O3
ratios of 20–1900. There is a close correlation between the activity
of benzene formation in methane aromatization and the Bro¨nsted
acidity of Mo/HZSM-5, but not Lewis aciditiy. It was found that
maximum benzene formation was obtained on the Mo/HZSM-5
having SiO2/Al2O3 ratios of 20–49, but substantially poor activi-
ties on those with SiO2/Al2O3 ratios smaller and higher than 40.
The results suggest that methane is dissociated on the molybdenum
carbide cluster supported on HZSM-5 having optimum Bro¨nsted
acidity toform CHx (x > 1) and C2-species asthe primaryintermedi-
ates which are oligomerized subsequently to aromatics such as ben-
zene and naphthalene at the interface of Mo2C and HZSM-5 zeolite
having the optimum Bro¨nsted acidity. The bifunctional catalysis of
1 On leave from Hebei University of Science and Technology, Shijia-
zhuang 050018, China.
2 On leave from Dalian Institute of Chemical Physics, Dalian, China.
3 Corresponding author. E-mail: michi@cat.hokudai.ac.jp.
175
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

176
LIU ET AL.
been concerned only about the methane conversion to ben-
The hybrid catalysts of 3 wt% Mo/SiO2 + HZSM-5 were
zene on Mo/HZSM-5. Regarding the quantitative study on prepared by mechanically mixing and grinding both compo-
methane aromatization, the authors are concerned how to nents in pyrex-glass tubing and Mo2C + HZSM-5 was pre-
reasonably evaluate the product yields and the coke for- pared by the coimpregnation of both component powders
mation in this process, and they found, using the improved from the slurry solution, which were pressed in a pellet,
analytic system of on-line GC/MS and the high-molecular crushed into sizes of 20–42 mesh for catalytic tests.
TOF-mass analyzer, that coke formation in the methane
aromatization reaction is much more serious than had been
expected and the benzene formation is not so selective on
Mo/HZSM-5. Actually, there are a number of various aro-
2.2. Catalytic Measurements and Product Analysis
matic compounds and undetected hydrocarbons formed
during the reaction.
The catalytic tests were carried out under atmospheric
pressure of methane in a continuous flow system with a
quartz reactor of 8-mm ID which is charged with 0.30 g of
catalyst pelleted and sized to 20–42 mesh. The outlet pipe
line from the reactor and the on-line sampling valve were
kept at higher than 500 K to prevent the condensation or
strong adsorption of the higher hydrocarbon products, as
shown in Fig. 1. The feed gas mixture of 98% CH4 (99.9%
purity) and 2% Ar (Daido Hoxan Inc., 99.9% purity) as the
internal standard for analysis was introduced into the re-
actor at 7.2 mL/min, controlled with a mass flow controller
(Brooks Co.) after flushing with He at 973 K for 40 min. The
Shimadzu GC-14A gas chromatograph with a flame ioniza-
tion detector (FID) is equipped with a six-way valve heat-
ing at 260 C for sampling and separation of hydrocarbon
productsincludingnonvolatile compounds, such asbenzene
derivatives and naphthalene on a 4 mm 1 m Porapak-P
column. Another GC of Shimadzu GC-8A was employed
for on-line product analysis of H2, Ar, CO, CH4, and CO2
on a 4 mm 2 m activated-carbon column with a thermal
conductivity detector (TCD). Hydrocarbons produced in
this reaction were C2-C4 alkanes (and/or alkenes) and C6-
In this report, we study carefully the activities and selec-
tivities of methane conversion to aromatics and coke for-
mation on the Mo-supported catalysts by varying the
supporting materials. The Mo/HZSM-5 catalysts are char-
acterized by EXAFS, TG/DTA/mass, and TPR (temp-
rature-programmed reaction) studies, including the carbur-
ization of Mo oxide with methane towards Mo2C. The
acidity measurement of Mo/HZSM-5 having different Si/Al
ratios has been conducted by IR spectroscopy in the pyri-
dine adsoprtion. The bifunctional role of molybdenum car-
bide and HZSM-5 zeolite is discussed in conjunction with
the active states of Mo and the Bro¨nsted and Lewis acidity
function of HZSM-5 supports having different Si/Al ratios
which affect the activity and selectivity of methane con-
version towards benzene and naphthalene, as well as coke
selectivity in the reaction.
2. EXPERIMENTAL
2.1. Materials and Catalysts Preparation
The Mo-supported catalysts were prepared using ei- C12 condensable materials such as such as benzene, toluene,
ther an incipient wetness or a conventional impregnation xylene, naphthalene, and methyl-naphthalene which were
method. The materials used are: (NH4)6Mo7O24 4H2O identified by using on-line GC and GC-MS (Perkin-Elmer,
(Kanto Chem. Co.); NH4ZSM-5 (SiO2/Al2O3 = 20 1900; Auto System GC with 910 Q-Mass) spectrometer.
surface area, 780–925 m2/g (Toso Co. and CRI Zeolyst, Inc.,
The total gas flow rate at the inlet and outlet of the reac-
as purchased); Al2O3 (Japan Aerosil Co., 280 m2/g), SiO2 tor can be described as shown in Eq. [1] since the mass flow
(Japan Aerosil Co., 350 m2/g), FSM-16 (gift from Toyota rate of Ar as an internal standard gas should be constant
Central Research Lab.; pore sizes = 4.7 nm and 2.7 nm; through the reaction progress. By use of Eq. [1], conversion
SiO2/Al2O3 = 320;980–1020m2/g), and H-Mordenite (SiO2/ of methane (Conv.) and selectivity of each product con-
Al2O3 = 44; Cat. & Chem. Ind.. Co. Ltd.). The modified taining a carbon atom, including CO and CO2 on carbon
carbon
FSM-16 having Si/Al ratios of 15 and 20 were prepared base (S
) is calculated from Eqs. [2] and [3], respec-
product
by CVD technique with AlCl3. The Al-contents in the tively. Thus, the selectivity for coke formation on a car-
HZSM-5, H-mordenite, and FSM-16 supports were ana- bon basis is derived from (1 the sum of product selectiv-
lyzed by ICP measurements of the samples dissolved with ity). Coke consists of all undetected carbon products such
HF solution. The 3–6 wt% Mo loading HZSM-5 cata- as amorphous and graphic inert carbons, including higher
lysts were prepared by impregnation of NH4ZSM-5 with boiling compounds which are difficult to detect and ana-
(NH4)6Mo7O24 4H2O (Kanto Co., as purchased) from an lyze by the FID and TCD gc used. In these equations, F,
aqueous solution and an incipient wetness technique. The X, and Ncarbon represent the total gas flow rate, mole frac-
resulting materials are dried at 393 K and calcined at 773 K tion, and carbon number in a molecule, respectively. Simi-
for 6 h, similarly as reported in the literature (18). Powdered larly, selectivity for the formation of a hydrogen-containing
Mo carbide (Mo2C; 1.8- m size, surface area of 1–2 m2/g) product on a hydrogen base can be calculated and, thus,
was purchased from Japan New Metal Co. and used as it.
the ratio of hydrogen to carbon in the coke formed can be

BIFUNCTIONAL CATALYSIS OF Mo/HZSM-5
177
FIG. 1. Schematic flow system: (1) mass flow controller (MFC); (2) heating 6-way valve; (3) quartz cell containing a catalyst; (4) gas chromatograph
(FID, TCD).
estimated:
a TE-150multiple channelmassspectrometer (ThermoLab,
VG Gas) at m/e = 2, 28, 44, 26, 30, 78, and 128, respectively.
The reaction temperature was raised from room tempera-
ture to 973 K with a ramping rate of 10 K/min and held
there for 30 min. Before the reaction started, the catalyst
was heated in a He stream at 873 K for 30 min and, in some
cases, was further pretreated with H2, methane, or ethane
at 873 or 973 K for 30 min.
inlet
Ar
outlet
Ar
Finlet
X
= Foutlet
X
.
[1]
The rate of methane consumption is expressed by Finlet
inlet
CH4
outlet
CH4
X
Foutlet
X
. Thus, methane conversion is esti-
mated:
inlet
methane
outlet
Finlet
X
Foutlet
X
methane
Conv. =
inlet
Finlet
X
methane
2.4. XAFS Measurement and Analysis
outlet
methane
inlet
methane
inlet
X
X
Ar
The powdered and disk samples were charged under N2
in an in-situ XAFS cell with a Kapton film window (500 m)
to prevent exposure of the sample to air. Mo K-edge XAFS
(X-ray absorption fine structure) measurements were con-
ducted on the samples of Mo/HZSM-5 after impregnation,
calcination at 873 K, and reaction with methane at 873 K
and 973 K at BL-10B at the Photon Factory of the National
Laboratory for High Energy Physics (KEK-PF, Tsukuba,
Japan). The energy and the current of the electron (or
positron) were 2.5 GeV and 250 mA, respectively. A Si(311)
channel-cut monochromator was used. The XANES (X-
ray absorption near edge structure) and EXAFS (extended
X-ray absorption fine structure) spectra were analyzed by
a computer program supplied by Technos Co. LTD [20].
The k3-weighted EXAFS function was Fourier transformed
= 1
[2]
[3]
outlet
X
Ar
X
outlet
product
carbon
Foutlet
X
N
product
carbon
product
S
=
=
inlet
methane
outlet
Finlet
X
Foutlet
X
methane
inlet
Ar
outlet
product
carbon
X
X
N
product
.
outlet
inlet
methane
inlet
Ar
outlet
X
X
X
X
methane
Ar
2.3. TG/DTA/Mass Experiment for Carburization
of Mo/HZSM-5
TG/DTA/Mass study was performed using TG/DTA/
MASS System (Mac Science Co., TG-DTA2020S) under
a methane/He stream (flow rate: CH4 = 15 ml/min and
He = 150 ml/min) and H2/He (170 ml/min); 40–100 mg of
Mo/HZSM-5catalyst were mounted in a fused alumina boat
and the products developed, such as H2, CO, CO2, C2H4,
C2H6, C6H6, and C10H8 were continuously monitored with
into R-space using the k-range from 3.5 to 18 A 1. The
˚
Hanning function used was = 0.5 A 1. The phase shift was
˚
not corrected for the preliminary Fourier transformation.

178
LIU ET AL.
The inverse Fourier transform was calculated to obtain activation of the catalyst. The major hydrocarbons pro-
a filtered EXAFS function. The R range of an inverse duced on Mo/HZSM-5 consist of benzene, naphthalene,
˚
Fourier transformation was taken from 1.09 to 3.31 A. The and toluene as the aromatics and C2 products such as ethene
˚
Hanning function of 1 = 0.05 A was used as a window func- and ethane. The other carbon-containing products, such as
tion. The fitting EXAFS parameters of the sample were de- CO, propene, butene, methyl-naphthalene, phenanthrene,
termined for coordination number (C.N.) and interatomic and anthracene were only produced in trace amounts (to-
distance R, to minimize and correction of threshold en- taling a few percentages of selectivity). Figure 2b shows the
ergy (1Eo). The backscattering amplitudes and phase shift selectivity of hydrocarbon products such as benzene, naph-
functions of Mo-Mo, Mo-C, and Mo-O bondngs are cor- thalene, and C2 species and the coke formation with time
rected using those of Mo foil (Mo-Mo), Mo2C (Mo-Mo, on-stream of methane at 973 K. The total selectivity of gas
Mo-C), and MoO3 (Mo-Mo, Mo-O). The EXAFS param- phase hydrocarbon products increases from 55% at the ini-
eters for discussion were fixed around the optimum value, tial stage of the reaction and then keeps constant (70% )
and the residual factor R was calculated to offer the opti- at the steady state of the reaction, while the selectivity of
mum values for the other parameters. The real value was coke formation changes in a reverse way from 43 to 32% .
estimated using a residual factor R which is smaller than The selectivity of CO is less than 1% throughout the reac-
twice the optimum value.
tion. This may suggest that hydrocarbon products and coke
share a common intermediate such as CHx leading in part
to C2 species and hydrocarbon products and in part to coke
formation.
2.5. IR Measurement in Pyridine Adsorption
Each powder of 3 wt% loading Mo/HZSM-5, having a
different silica to alumina ratio, SiO2/Al2O3 = 20–1900, was
pressed into a self-supporting wafer (2.0 mm ID, 15–18 mg/
cm2) with a pressure of 500 kgf/cm2. The wafer was mounted
in an in-situ IR cell equipped with CaF2 window. IR spectra
were recorded using a step-scan Fourier transform infrared
spectrometer (BIO-RAD FTS-60A/896) with resolution
of 4 cm 1, and a co-accumulation of 64 interferograms
to improve the signal/noise ratios of the IR spectra. The
IR measuremnt was conducted by exposure of the disk of
HZSM-5 after evacuating at 723 K for 2 h to the pyridyne
vapor of 1.2 kPa at 300 K.
Figure 2c shows the compositions in the gas phase hy-
drocarbon products on 3% Mo/HZSM-5 catalyst with the
time on stream of the methane reaction. At the initial stage
of the reaction C2 species, benzene, toluene, and naphtha-
lene are produced in 4% , 50% , 2% , and 45% on a carbon
basis, among the produced hydrocarbons, respectively. Af-
ter running the reaction 24 h the percentage of naphtha-
lene among the hydrocarbon products decreases from 45%
to 12% , while the percentage of C2 and benzene increase
to about 20% and 68% at the expense of the decrease of
naphthalene. This may suggest that naphthalene is appar-
ently produced consecutively from C2 species and benzene.
As shown in Fig. 2d, the rates of benzene and naphtha-
lene formation in the methane aromatization reaction de-
creased monotonously from the maximum rates of 98 to
50 nmmol/s g-cat after 30 h, possibly by catalyst poisoning
due to the irreversible coke deposition. The molar ratios for
toluene to benzene and methylnaphthalene to naphthalene
are almost held constant at 3–5% during the entire reaction
process. Further substituted aromatic compounds are ob-
tained in negligible amounts on the Mo/HZSM-5, partly
because of the higher activity of hydrogenolysis and molec-
ular shape selectivity due to the limited size of micropores
of HZSM-5 that are accessible only to benzene and toluene.
Using 2% Ar as the internal standard for the reaction
analysis, the present work has been conducted to carefully
evaluate the yields and selectivities of benzene formation in
the reaction. The results we obtained show that the benzene
selectivities, based on the methane consumed, did not ex-
ceed 48% as the maximum value and that of coke formation
is 32–43% , which differs greatly from the estimates of Wang
et al. (16, 17) and others (7, 9, 13, 15). This large discrepancy
may be mainly due to the analytic accuracy of the methane
concentration consumed using the internal standard of
Ar (18) and N2 (16, 17) in a different concentration and
3. RESULT AND DISCUSSION
3.1. The Reaction of Methane on 3% Mo/HZSM-5
and Product Analysis
The catalytic performance of 3% Mo/HZSM-5 for me-
thane aromatization at 973 K and 1 atm pressure of me-
thane has been described in our previous study (18). Hydro-
gen is usually produced in large amounts (H2/benzene =
9–18 mol/mol) by the dehydrocondensation of methane.
CO, CO2, and H2O are evolved at the initial stage of the
reaction above 873 K, due to the reduction of MoO3 with
methane, being converted to a lower valent state species
or molybdenum carbide (9, 16–18). Methane conversion
and the product formation rates with reaction time on-
stream are illustrated in Fig. 2a. The conversion of methane
(ca 12% ) usually decreases rapidly at the initial stage and
moderately at the later stage of the reaction, down to 4%
in 30 h of the time on-stream. The suppression of methane
conversion may be attributable to coke formation upon
the admission of methane to the fresh Mo/HZSM-5. This
suggests that the deposited carbon becomes less reactive
with time on-stream, which also leads to the gradual de-

BIFUNCTIONAL CATALYSIS OF Mo/HZSM-5
179
FIG. 2. Catalytic performances in methane aromatization on 3 wt% Mo/HZSM-5 at 973 K: (a) methane conversion vs time on-stream; (b) selec-
tivities (% ) on carbon base for hydrocarbon products, coke, and CO vs time on-stream; (c) composition (% ) in hydrocarbon products; and (d) rates of
product formation for benzene, naphthalene, and C2 species (nmol/s g-cat).
neglecting the higher polycondensed hydrocarbons such aromatic compounds are difficult to analyze by the on-line
as naphthalene, methylnaphthalene, and anthracene as the GC system due to their high boiling point and strong ad-
products in the reaction (7–16). To confirm this, we have sorption on the separation column. The total mount of such
analyzed all of the products in the effluent gas which are polycondensed products is estimated to be not over 10%
condensed by using the ethanol–dry ice trap. The collected of the detected hydrocarbons, such as benzene and naph-
reaction products were analyzed by the on-line GS/Mass thalene. The most distinct feature is that the conversion
with a capillary column (OV-1; 25 m) up to 553 K and of methane usually decreases monotonously with time on-
MALDI-TOF (time-of flight reflect-mass spectrometer; stream, while the total selectivity of all the hydrocarbon
VOYAGER; m/e = 1–2500; PerSeptive Biosystems Co.) products keeps constant.
mass sepctromtery. Table 1 shows a list of aromatic com-
pounds, such as benzene and naphtalene as the major prod-
ucts, including other polycondensed aromatics such as an-
3.2. Reaction of Methane on 3% Mo/FSM-16
thracene, pyrene, tetracene, and their deriviates which were
Figure 3 shows the time dependencies of methane con-
collected as products of the methane reaction on 3 wt% version (3a), the selectivity of hydrocarbon on a carbon
Mo/HZSM-5 at 973 K. The latter part of the polycondensed basis (3b), the composition (% ) of benzene devided by 2

180
LIU ET AL.
TABLE 1
Distribution of Aromatic Products Formed in the Methane Con-
version on the 3 wt% Loading Mo/HZSM-5 Catalyst at 973 K
Percentage composition
Aromatic product
Benzene
on carbon basis
67.6
3.5
Toluene
p-Xylene, m-xylene
C9 aromatics (e.g., methylene,
1,2,4-trimethylbenzene)
Naphthalene
Methylnaphthalene
Dimethylnaphthalene (2,6-)
0.1
0.2
24.6
0.7
0.2
C
13 aromatics
0.3
(anthracene, phenanthrene)
Other polycondensed aromatics
Pyrene, Teracene, etc.
Less than 2%
FIG. 3. Catalytic performances in methane aromatization on
3 wt% Mo/FSM-16 (2.8 nm) at 973 K: (a) methane conversion vs time
Note. The aromatic products were condensed by the cold-trap in the
methane reaction for 5 h. Methane pressure = 1 atm, SV = 1440 ml/h/
g-cat.
–
–
on stream ( ᭿ ); (b) selectivity (% ) of hydrocarbon products on carbon
–
–
basis vs time on stream ( ᭺ ); (c) composition (% ) of benzene devided
–
–
by 2 in hydrocarbon products ( ᭹ ), and (d) rates of product formation
–
–
for benzene ( ᮀ ) (nmol/s/g-cat).
on a carbon basis (3c), and the formation rate of benzene
(3d) in the methane reaction at 973 K on 3 wt% Mo loading
catalysts prepared by impregnating (NH4)6Mo7O24 with the
mesoprous material of FSM-16 (2.7 nm; Si/Al = 320). Al-
though benzene and C2 species with a minor yield of toluene
and naphthalene were produced in the methane reaction (1
atm, SV = 1440 ml/h/g-cat) at 973 K, the methane conver-
sion decreases rapidly from 14% at the initial stage to 3%
within 6 h, possibly owing to serious coke formation. The
formation rate of benzene reaches a maximum value of
10.5 nmol/s/g-cat at 250 min of time on-stream and then de-
creases while the formation rates of ethene and ethane rel-
atively increase. The steady-state distributions of benzene
and C2 species among the hydrocarbon products (selectiv-
ities of the hydrocarbons decreased from 38 to 5% in the
time on-stream of the reaction) were 71–74% and 24–27%
on a carbon basis, respectively. It is of interest to find that
the catalytic performances of Mo-loading catalysts of FSM-
16 are not affected with their pore sizes between 2.7 and
4.7 nm, as shown in Table 2. Although the compositions of
the hydrocarbon products on the 3% Mo/FSM-16 catalaysts
are idential to those on 3% Mo/HZSM-5 as illustrated in
TABLE 2
Catalytic Performances of 3 wt%-loading Mo Catalysts (0.3 g) using Various Supports for CH4
Aromatization Reaction at 973 K
Selectivity (% )
Coke
HC distribution (% )a
Conv.
(% )
Catalysts
HC
CO
C2
C6H6
C7H8
C10H8
3% Mo/HZSM-5 (Si/Al = 79)
HZSM-5 (Si/Al = 79)b
3% Mo/Mordenite
3% Mo/USY
3% Mo/Al2O3
9.4
0.7
7.3
6.4
7.4
5.3
6.8
6.6
5.9
6.2
62.8
12.1
7.8
13.2
6.2
11.5
9.0
8.4
36.1
80.8
83.1
84.3
88.9
86.6
87.7
89.0
87.5
86.1
1.1
7.1
9.1
2.4
4.9
1.9
3.3
2.5
4.8
2.6
4.5
82.1
44.4
16.9
33.8
18.6
24.8
20.0
21.9
24.0
65.2
17.9
53.2
79.8
63.8
75.8
73.4
77.3
75.4
73.6
3.3
0.0
0.0
3.3
2.4
3.2
1.8
2.8
2.7
2.4
26.9
0.0
0.0
0.0
0.0
2.4
0.0
0.0
0.0
0.0
3% Mo/SiO2
3% Mo/FSM-16 (2.7 nm)
3% Mo/FSM-16 (2.7 nm, Si/Al = 20)
3% Mo/FSM-16 (2.7 nm, Si/Al = 15)
3% Mo/FSM-16 (4.7 nm)
7.8
11.2
Note. Methane 1 atm, flow rate = 7.6 ml/min.
a The data for methane conversion and product selectivity were taken at the initial stage of the reaction in 40 min
after the 1 atm methane flow passed through the catalyst bed (less than 3-mm thickness) at 973 K. C2 = C2H4 + C2H6;
C7H8 = toluene; C10H8 = naphthalene.
b The data were obtained on HZSM-5 support in the methane reaction at 973 K.

BIFUNCTIONAL CATALYSIS OF Mo/HZSM-5
181
Fig. 1 and Table 2, the catalyst deactivation was more dom-
TABLE 3
inant than those of the HZSM-5-supported one. In fact, on
3% Mo/FSM-16 having Si/Al ratios of 15–320 the selectiv-
ity of coke formation is substantially high (over 80% ), in
comparison with that of 30–32% on 3% Mo/HZSM-5 for
the methane reaction under similar reaction conditions.
Catalytic Performances of 3 wt% Mo/HZSM-5 Catalysts with
Different Silica to Alumina Ratio for Methane Aromatization Re-
action at 973 K and 1 atm Pressure
Hydrocarbon products
Selectivity (% )
distribution (% )
CH4
Si/Al2 Conv. (% ) HC
Coke
C2 Benzene Toluene Naph.
3.3. Effects of the Support of the Catalysts for the Reaction
23.8
26.0
39.5
53.0
73.4
216
800
1900
8.8
8.0
10.2
7.2
8.3
7.5
5.1
6.3
55.2
54.1
62.8
48.1
35.6
16.7
30.8
19.7
41.6
43.6
35.3
47.3
60.9
79.3
61.4
76.9
4.7
5.6
3.9
7.3
7.1
21.5
18.3
26.9
65.4
67.0
59.4
62.9
66.5
59.0
64.0
57.0
3.4
3.7
3.1
3.3
3.5
4.0
3.6
4.2
26.5
23.4
33.2
26.3
22.8
14.5
13.5
10.6
The catalytic performances of other 3 wt% loading Mo-
supported catalysts using various sorts of supporting mate-
rials have been studied in the methane reaction, and the re-
sultsare listed in Table 2. Comparingwith 3% Mo/HZSM-5,
although the methane conversions were comparable or rel-
atively lower (5.3–7.4% ), the selectivities of the hydrocar-
bon products (HC), such as C2-hydrocarbons (C2) and ben-
zene, on a carbon basis are substantially low due to serious
coke formation on the supported catalysts using FSM-16,
Mordenite, USY, SiO2, and Al2O3 as the supports. The se-
lectivities of hydrocarbon products in the gas phase were
less than 13% (less than 10% for benzene) on a carbon
basis, while the coke formation is higher than 80% , as pre-
sented in Table 2. The main products for hydrocarbons on
these catalysts are ethene, ethane, and benzene. Naphtha-
lene is only produced in a trace amount on these supported
catalysts. In the early stage of the reaction, the distribu-
tion of C2 species among the hydrocarbon products range
from 20% to 40% which is higher than for those on 3%
Mo/HZSM-5 (less than 5% ), and for prolonged reaction
the C2 products increase to more than 60% , in spite of the
lower methane conversion of less than 1% . The composi-
tions of the hydrocarbon products with reaction times on
SiO2, Mordenite, USY, and Al2O3-supported Mo catalysts
have a trend similar to those on Mo/FSM-16 having differ-
ent Si/Al ratios (15, 20, and 320). Regarding the methane
conversion on Mo-supported catalysts, the results suggest
that the HZSM-5 zeolite having Si/Al = 40 has been so far
a better support for the Mo-supported catalysts exhibit-
ing the highest methane conversion and selectivities for the
aromatic products with a lower coke formation, probably
owing to their acidity and their unique pore structures of
0.6-nm size.
Note. Data are taken at 40 min of reaction time on-stream.
ing the maximum activities of methane aromatization for
the HZSM-5 supports having SiO2/Al2O3 ratio of 30–45. By
contast, asshown in Table 3the Mo/HZSM-5havingsmaller
and higher SiO2/Al2O3 ratios than the optimum values
of 30–45 exhibited poor activities for the benzene forma-
tion, smaller by one or two orders of magnitude and more
serious coke formation than those on the Mo/HZSM-5
catalyst having a SiO2/Al2O3 ratio of 40. The Bro¨nsted and
Lewis acidity of the samples have been determined by pyri-
dine adsorption using FT-IR spectroscopy. Figure 4a shows
the IR spectra of pyridine adsorption on 3% Mo/HZSM-5
catalysts with different SiO2/Al2O3 ratios. Upon pyridine
adsorption at room temperature and outgassing at 423 K,
absorption bands associated with the chemisorbed pyridine
were observed at 1547 cm 1, 1491 cm 1, and 1454 cm 1. Ac-
cording to previous reports on the IR studies in the pyridine
chemisorption (21, 23), the pyridine band at 1547 cm is
assignable to pyridinum ions and the band at 1454 cm
to Lewis acid-coordinated pyridine, respectively. The band
at 1491 cm consists of bands due to the Bro¨nsted and
Lewis sites. All the spectra were normalized using the peak
intensities of ZSM-5 support at around 1873 cm as the
standard, and the intensities of bands at 1547 cm for
Bro¨nsted acid and 1454 cm for Lewis acid sites were
1
1
1
1
1
1
used for probing the acidity on the Mo/HZSM-5 catalysts
having different SiO2/Al2O3 ratios. It is found as shown in
Fig. 4b that there is not an appreciable difference in the
IR bands at 1547 cm 1 due to the Bro¨nsted acidity on the
3.4. Influence of Silica to Alumina Ratios in HZSM-5
on Bro¨nsted Acidity and Methane Aromatization
Activity on Mo/HZSM-5
A series of 3% Mo/HZSM-5 catalysts was prepared us- samples of 3 wt% Mo/HZSM-5 and the HZSM-5 support
ing different SiO2/Al2O3 ratios. The performaces in the itself after being calcined at 773 K. This implies that the im-
methane conversion at 973 K have been studied on the cata- pregnated Mo species does not change the Bro¨nsted acid-
lysts in conjunction with the SiO2/Al2O3 ratio of the support ity of the HZSM-5 support. By contrast, the band around
HZSM-5 used as shown in Table 3. Actually, it is of interest 1440 cm 1, assignable to the Lewis acid site was relatively
to find that the ratesofhydrocarbon product formation such increased on the Mo/HZSM-5 after being calcined at 773 K
as benzene on the Mo/HZSM-5 catalysts substantially de- and even after the methane aromatization reaction at 973K.
pend on the SiO2/Al2O3 ratios of the HZSM-5 used, show- There is a corelation between the Lewis acidity and the

182
LIU ET AL.
FIG. 4. (a) IR spectra of adsorbed pyridine on 3 wt% Mo/HZSM-5 having various SiO2/Al2O3 ratios; Pyridine of 1.2 kPa was adsorbed at 300 K and
evacuated at 423 K for 2 h prior to the IR observation. (b) The change of IR spectra in the pyridine adsorption on the 3 wt% Mo/HZSM-5 (SiO2/Al2O3
ratio of 39.5) before and after the methane reaction at 973 K, compared with that on the HZSM-5 support itself. The samples were calcined at 773 K
for 2 h prior to the IR study in pyridine adsorption. The IR spectra were recorded in the pyridine chemisorption after the samples were evaculated at
423 K for 2 h.
lattice-substituted metal ions, such as Al3+, and higher va- by the Lewis acid ones of the ZSM-5 support. Moreover,
lent Mo species, such as Mo6 +
.
as shown in Fig. 5, it was demonstrated that the coke se-
In Fig. 5, the variations of Bro¨nsted and Lewis acidity lectivities in the methane conversion on the Mo/HZSM-5
on the Mo/HZSM-5 and those of the benzene formation are minimized (less than 30% ) for the HZSM-5 having the
rates and coke selectivities in the methane aromatization optimum SiO2/Al2O3 ratio of 39.5, whereas the coke forms
on Mo/HZSM-5 catalysts are illustrated against the func- seriously up to over 85% on the catalysts using the HZSM-
tion of the SiO2/Al2O3 ratios of the HZSM-5 supports used. 5 with the lower or higher SiO2/Al2O3 ratio deviated from
The Bro¨ nsted acidity of HZSM-5 showed a sharp maximum 35–45. Moreover, the acidity measurements of HZSM-5
for the HZSM-5 having the SiO2/Al2O3 ratio of 39.5, which and 3% Mo/HZSM-5 with SiO2/Al2O3 ratio of 39.5 be-
is identical to that of benzene formation rates. Neverthe- fore and after the methane aromatization reaction also
less, there is not a particular relationship between the ben- supply us some additional evidence. It is worthy to note
zene production and the Lewis acidity of the Mo/HZSM- that for the 3% Mo/HZSM-5 the intensity of IR band at
5 which monotonously decreases with the increase of the 1547 cm 1 due to the pyridine adsorption on Bro¨nsted acid
SiO2/Al2O3 ratio for the HZSM-5 used.
sites was greatly suppressed after the methane aromatiza-
According to the close correlation between the Bro¨nsted tion reaction at 973 K for 8 h, as shown in Fig. 4b. This
acidityofthe Mo/HZSM-5catalystsand the benzene forma- suppression of Bro¨nsted acidity on the used Mo/HZSM-5
tion rates in methane aromatization reaction, it is reason- catalyst might be related to the coke deposition in the reac-
ably suggested that the aromatic products such as benzene tion with methane at 973 K, which plugs the internal chan-
are most probably promoted by the Bro¨nsted acid sites, not nels of HZSM-5, or thermal reaction of methane with acidic

BIFUNCTIONAL CATALYSIS OF Mo/HZSM-5
183
FIG. 5. Dependencies of benzene formation rates (nmol/s/g-cat) and coke selectivitty (% ) in the methane reaction at 973 K and those of
Bro¨nsted and Lewis acidities deduced by the IR study in pyridine adsorption on the 3 wt% Mo/HZSM-5 upon the SiO2/Al2O3 ratios of the HZSM-5
used.
silanol groups on the Mo/HZSM-5 above 773 K. The latter or molybdenum carbide because of the higher dispersion
effect on the Bro¨nsted acidity of HZSM-5 is similar to the of both Mo crystallites on the HZSM-5 supports as sug-
preferential decrease of bands at 3611 cm 1 due to Bro¨nsted gested previously (12, 16, 18). The unsupported Mo2C sam-
acidic OH of HZSM-5 by heating above 673 K, as recently ple showed peaks corresponding to Mo2C of the hexagonal
reported by Lunsford et al. (17).
close-packed crystal structure with normal lattice parame-
˚
˚
The higher Bro¨nsted acidityofthe HZSM-5supportsmay ters of a = 2.99 A and c = 4.72 A. To characterise the active
bifunctionally promote not only the methane activation structure of the Mo/HZSM-5 which is incorporated into
(CH4 ⇒ CHx + nH2) on the Mo carbide sites of Mo/HZSM- the methane aromatization, the Mo K-edge XAFS studies
5 catalysts, but also a further oligomerization of the have been conducted using the 10B line at Photon Factory
dissociative CHx surface species towards the primary prod- of National Laboratory for High Energy Physics (KEK-PF;
ucts such as ethene, which are effectively converted into Tsukuba, Japan). Figure 6 represents some XANES spectra
benzene and naphthalene on the Mo/HZSM-5 catalysts.
of 6 wt% loading Mo/HZSM-5 calcined, treated at increas-
ing temperatures in flowing methane at 973 K for 1 and 24 h.
The spectra of a MoO3, Mo foil, and a Mo2C powder were
observed as references. All the spectra (Figs. 6c and 6d)
of Mo/HZSM-5 after the carburization with methane at
973 K showed identical edge position, similar to Mo car-
3.5. Mo Carbide Formation in the Methane Conversion
on Mo/HZSM-5 Characterized by TG/DTA/Mass
and EXAFS Studies
The X-ray diffraction patterns of calcined or carburized bide (Fig. 6e) and Mo foil (Fig. 6f) as the first inflection
supported samples (Mo/HZSM-5) with up to 6% load- point in the rapidly rising portion of the edge. This posi-
ing showed no peaks corresponding to molybdenum oxide tion was assigned an energy of 20003.9 eV. It is known that

184
LIU ET AL.
reference samples, respectively. These distances should be
˚
˚
corrected for phase shifts of 0.34 A for Mo-Mo and 0.5 A
for Mo-C distances. After the calcination of the 6 wt% Mo
loading sample, (NH4)6Mo7O24/HZSM-5 at 973 K prior to
the methane reaction, the Mo K edge Fourier transform
functions showed that the calcined sample (Fig. 7b) consists
˚
of the intense peak at 1.42 A, possibly due to Mo-O bond
of highly dispersed Mo oxide species similar to NaMoO4
and [Mo(=O)4] in NaY as reported by Ozin et al. (25).
The FT function spectrum (Fig. 7b) is different from that
of MoO3 crystal (Fig. 7a). The EXAFS data suggest that
Mo oxide species in the calcined sample of Mo/HZSM-5
is highly dispersed in the internal channels of HZSM-5.
The FT function of the sample reacted with the methane at
973 K for 1 h (Fig. 7c) and at 973 K for 24 h (Fig. 7d) showed
FIG. 6. Mo K edge XANES spectra of (a) MoO3 crystal; (b) 6 wt%
Mo/HZSM-5 (SiO2/Al2O3 ratio of 39.5) after calcined at 873 K for 2 h;
(c) after (b) treated with methane at 973 K for 1 h; (d) after (b) treated
with methane at 973 K for 24 h; (e) Mo2C powder; and (f) Mo foil as the
reference.
edge features move to lower energy as the oxidation state of
Mo is reduced (21), as shown for the Mo/HZSM-5 sample
after the reaction with methane at 973 K, compared with
those of the calcined sample (Fig. 6b) and MoO3 (Fig. 6a).
The only notable difference between Mo foil and Mo2C
XANES spectra was a separation of the two broad peaks
at 20.04 KeV and 20.03 keV above the absorption edge.
The XANES spectrum of Mo foil showed a larger sepa-
ration than the Mo2C spectrum (Fig. 6e). The XAFS of
the Mo/HZSM-5 sample reacted with methane at 973 K
for 1 and 24 h (Figs. 6c and 6d) look basically identical to
the Mo2C reference spectrum (Fig. 6e), not to the Mo foil
(Fig. 6f).
On the other hand, the corresponding Fourier transform
functions are shown in Fig. 7. In Fig. 7, the three dominant
peaks on each spectrum represent the Mo-C distance in
FIG. 7. Mo K edge Fourier transform function spectra of (a) MoO3
crystal; (b) 6 wt% Mo/HZSM-5 after calcined at 873 K for 2 h; (c) after (b)
treated with methane at 973 K for 1 h; (d) after (b) treated with methane
˚
˚
Mo2C (1.58 A), Mo-Mo distance in Mo2C (2.62 A), and
˚
the Mo-Mo distance in in metallic Mo (2.38 A) as the at 973 K for 24 h; (e) Mo2C powder; and (f) Mo foil as the reference.

BIFUNCTIONAL CATALYSIS OF Mo/HZSM-5
185
and 6 wt% Mo/HZSM-5 after the calcination at 873 K con-
sists of an isolated Mo oxide, which might be highly dis-
TABLE 4
EXAFS Parameters for the 3 wt% Mo/HZSM-5 and 6 wt%
Mo/HZSM-5 after the Reaction with Methane (1 atm) at 973 K,
in Comparison with Those of Mo Foil and Mo2C
˚
persed in the internal channels of HZSM-5 (6 A diameter).
After the reaction with methane at 973 K for 1 h and 24 h,
the Mo oxide species is converted with methane to Mo2C
˚
Mo-Mo
Mo-C(O)
cluster (Mo-Mo: C.N. = 2.1–2.5, R = 2.96–2.98 A; Mo-C:
˚
C.N. = 0.8–0.9, R = 2.12–2.14 A), in comparison with those
Samplesa
C.N.
2.1
R (A)
C.N.
0.8
R (A)
b
˚
˚
˚
of the Mo2C reference (Mo-Mo: C.N. = 12, R = 2.97 A;
3 wt% Mo/HZSM-5
973 K—1 h
3 wt% Mo/HZSM-5
973 K—24 h
6 wt% Mo/HZSM-5
973 K—4 h
2.97
2.98
2.96
0.084
0.087
0.095
2.10
2.14
2.11
0.047
0.055
0.028
˚
Mo-C: C.N. = 3, R = 2.10 A) (23, 24). The FT function spec-
tra ofthe Mo/HZSM-5retained a partialcontribution ofMo
oxide even after the prolonged carburization with methane
at 973 as indicated in Fig. 7. These results imply that the
Mo oxide species dispersed in the HZSM-5 framework may
migrate at the external surface of the HZSM-5 and be con-
verted with methane to Mo carbide which is highly spread
on the support surface and which is active for methane
aromatization at 973 K.
2.3
2.5
0.9
0.6
Mo foil
Mo2C powder
8
12
2.72
2.97
0.064
0.071
—
3
—
2.09
—
0.047
a The samples were stored at 300 K in the EXAFS cell under N2 atmo-
sphere after the carburization at 973 K.
b
The temperature programed reaction (TPR) with
methane was conducted on 3 wt% Mo/HZSM-5 using TG/
= Debye–Waller factor.
peaks at the same positions as those for the Mo2C reference DTA/mass by flowing methane while heating from 300 K
(Fig. 7e). The spectra of the two samples (Fig. 7c and Fig. 7d) to 973 K. As shown in Fig. 8, the evolution of CO
˚
are identical in every detail, except for the peak at 2.23 A, (m/e = 28) and H2 (m/e = 2) occurred at 923–973 K,
possibly due to molybdenum oxycarbide. The structural in- where the molybdenum oxide is converted to Mo2C. This
formation abstracted from the curve-fitting of the FT func- initiates the methane conversion to C2 products (ethene
tion for the samples of Mo/HZSM-5 after the carburization (m/e = 26) + ethane (m/e = 30)), benzene (m/e = 78) and
with methane at 973 K are listed in Table 4. The results sug- naphthalene (m/e = 128), accompanied with the hydrogen
gest that the impregnated sample of 3 wt% Mo/HZSM-5 (m/e = 2) evolution. On the other hand, the DTA pattern
FIG. 8. TPR spectra for the product formation evolved in the reaction with methane on 3 wt% Mo/HZSM-5 (SiO2/Al2O3 ratio of 39.5) by heating
from 300 K to 973 K and holding at 973 K for 30 min. Flow rate: CH4/He = 15/150 (ml/min); ramping rate of temperature: 5 K/min.

186
LIU ET AL.
3.6. The Hybrid Catalysts of Molybdenum
Carbide and HZSM-5 Supports
From the above discussion it has been known that
HZSM-5 as the support is superior to the other materials
such as FSM-16 and SiO2 in methane aromatization
reaction, and its Bro¨nsted acididy is an important factor
for methane activation and conversion to higher hydrocar-
bons. But the role of HZSM-5 support and molybdenum
species in methane aromatization reaction is still unclear.
To discuss on this aspect, two series of hybrid cata-
lysts, 3 wt% Mo/SiO2 + HZSM-5 (SiO2/Al2O3 = 40) and
Mo2C + HZSM-5 (SiO2/Al2O3 = 40), were prepared and
their catalytic activities were evaluated. In this case, the
temperature of the catalyst was elevated to 973 K in helium
flow instead of air to prevent the oxidation of Mo2C. Only
trace amounts of higher hydrocarbons such ethene and
benzene were produced by flowing methane at 973 K on
pure Mo2C and HZSM-5 (Table 5). The formation rate
of benzene in the methane conversion on the physically
mixed Mo2C + HZSM-5 by varying the dilution of Mo2C,
as shown in Fig. 10a. The physically mixed samples were
exposed to hydrogen at 723 K for 2 h before testing their
effects in the methane conversion. It was demonstrated
that a maximum promotion of benzene formation was
obtained when Mo2C was diluted with HZSM-5 in a 5 wt%
loading Mo2C in HZSM-5. Considering the poor mobility
FIG. 9. DTA (a) and TG (b) spectra of 3 wt% Mo/HZSM-5 (SiO2/
Al2O3 ratio of 40) in the methane reaction by programmed heating (5 K/
min) from 300 K to 973 K and holding at 973 K for 30 min. DTA spectrum
(c) of 3 wt% Mo/HZSM-5 (SiO2/Al2O3 ratio of 40) by the programmed
heating (5/min) from 300 K to 973 K in a flow of H2/He (170 ml/min). Flow
rate: CH4/He = 15/150 (ml/min) and H2/He = 20/150 (ml/min).
(Fig. 9) in the methane stream on the Mo/HZSM-5 shows
the distinct endothermic peak at 955 K due to the Mo
carbide formation, whereas there is no particular peak in
the hydrogen flow (c). The TG pattern in methane flow (a)
offers a sharp weight loss at 957 K due to the Mo carbide
formation, accompanied by a gradual weight increase
above 960 K, possibly owing to the coke formation on the
Mo/HZSM-5 in the reaction with methane. These data in
XAFS and TG/DTA/mass studies indicate that Mo oxide
on HZSM-5 is carburized with methane at 923–963 K to
form Mo2C¸ cluster in high dispersion on HZSM-5 which
initiates the methane aromatization to benzene and C2
hydrocarbons at 973 K.
of Mo2C on HZSM-5, the Mo2C particles (100–500
m
size) are expected to maximize mutual contact of Mo2C
particles at the vicinity of the external surface of HZSM-5.
The marked promotion for the benzene yield and methane
conversion (Table 5) may be associated with the synergetic
effect between Mo2C and HZSM-5. It is plausible that
the CHx species, due to the dissociative activation of CH4,
may be converted to ethene on the external surface of Mo
carbide particle, which migrates and forms benzene and
naphthalene at the interface HZSM-5 support having the
proper Bro¨nsted acidity.
FIG. 10. Variation of the rates of benzene formation ( mol/Mo mol/s) in the methane reaction (SV = 1440 ml/h/g-cat) at 973 K on the hybrid
catalysts consisting of [Mo2C] + [HZSM-5 (SiO2/Al2O3 ratio of 40)] (a) and [Mo/SiO2] + [HZSM-5 (SiO2/Al2O3 ratio of 40)] (b), depending upon the
dilution of Mo components (Mo/SiO2 and Mo2C) in the physical mixture systems. Methane 1 atm; flow rate 7.6 ml/min, 973 K. The value of dilution is
calculated by the weight of each component in a physical mixture.

BIFUNCTIONAL CATALYSIS OF Mo/HZSM-5
187
primary intermeidates such as CHx and ethene to higher
hydrocarbons and aromatics such as benzene, but also for
the reduction of carbon deposition. The two hybrid catalytic
systems mentioned above exhibit different catalytic perfor-
mances for CH4 aromatization reaction; the formation rate
of benzene on 3 wt% Mo/SiO2 + HZSM-5 with time on-
stream has a long induction period before reaching the max-
imum, similar to that on 3 wt% Mo/HZSM-5, whereas on
Mo2C + HZSM-5 the benzene formation usually decreases
at the initial stage and then reaches a steady-state value.
These different performances could be explained by the dif-
ferent pathway for the methane activation between Mo2C
and MoO3. MoO3 needs to be reduced and carburized with
methane to form molybdenum carbide or oxycarbide be-
fore it becomes catalytically active for methane aromatiza-
tion reaction. Compared with Mo2C + HZSM-5, the higher
activity of 3 wt% Mo/SiO2 + HZSM-5 is related to a higher
exposure of Mo/SiO2 at the interface of the HZSM-5, as
has been discussed in Section 3.5.
TABLE 5
Promotion Effect of Physical Mixing of Mo2C and 3 wt% Mo/SiO2
with HZSM-5 (SiO2/Al2O3 Ratio of 39.5) for Benzene Production
from Methane at 973 K
CH4 conversiona
(% )
Benzene formation
rate ( mol/mol-Mo/s)
Catalystb
HZSM-5 (Si/Al2 = 39.5)
Mo2C
Mo2C + HZSM-5 (Si/Al2 = 39.5)
(5 wt% Mo loading)
3 wt% Mo/SiO2
0.7
0.8
5.7
0.42c
0.3
15
5.3
6.2
40
200
3 wt% Mo/SiO2 + HZSM-5
(Si/Al2 = 39.5)
(1 : 1 volume mixture)
3 wt% Mo/HZSM-5
8.5
316
a The reaction condition: methane 1 atm; Flow rate = 7.6 ml/min. The data
for methane conversion and product selectivity are taken at the initial stage
of the reaction in 40 min after the methane flow passes through the catalyst
bed at 973 K.
b
The physically mixed samples of 0.30 g were amounted in a fixed bed
flow-mode pyrex-glassreactor. HZSM-5(SiO2/Al2O3 ratio of39.5) and Mo2C
(1.8- m size, 99% purity, surface area 1–2 m2/g) were used.
The rate of benzene formation based on mol/g-HZSM-5/s was taken
in the methane conversion under the same reaction condition: 973 K, 1 atm
3.7. Bifunctional Role of Mo2C and HZSM-5 for the
Selective Aromatization of Methane towards Benzene
and Naphthalene on Mo/HZSM-5 Catalyst
c
of methane, SV = 1440 ml/h/g-cat.
From the TG/DTA/mass for the temperature-pro-
grammed reaction with methane and XAFS studies on the
HZSM-5-supported Mo, it is reasonably suggested that the
precarburization of Mo oxide on HZSM-5 above 945 K is
essential for the formation of highly dispersed Mo carbide
(hcp Mo2C) on HZSM-5, which is remarkably active for
the dehydrocyclization of methane to form benzene and
naphthalene at the prevailing reaction conditions. As we
have disscussed in Sections of 3.4. and 3.6., the HZSM-5
support having optimum Bro¨nsted acidity (SiO2/Al2O3
ratio of 40) provides the Mo carbide-supported catalysts,
exhibiting the highest activity and selectivity for the
benzene formation and minimizing the coke formation
in the methane aromatization. In fact, methane is dehy-
drogenated on the active Mo carbide (in TG/DTA/mass)
to offer the surface carbon species CHx(0 < x > 3) and
H2 in the gas phase. It is conceivable that the active CHx
and a coupled C2 species as the primary products which
are oligomerized and dehydrocylized to form aromatics
such as benzene and naphthalene at the interface of the
Bro¨nsted acidic HZSM-5 as presented in Scheme 1. There
is a common expectation for the HZSM-5 support to
proceed to a further oligomerization reaction of carbon
The combination of 3 wt% Mo/SiO2 with HZSM-5
(SiO2/Al2O3 = 40) also promotes the methane conversion
and products such as benzene formation (Table 5), and the
benzene formation rate with composition of the 3 wt%
Mo/SiO2 + HZSM-5 (SiO2/Al2O3 = 40) is shown in Fig. 10b.
The 1 : 1 mixture of 3 wt% Mo/SiO2 with HZSM-5 shows
the highest formation rate of benzene, based on per gram
of catalyst, and the benzene formation rate, based on per
mole Mo metal, increases monotonously with the decrease
of 3 wt% Mo/SiO2 content (by increasing dilution of the
Mo carbide) in the physical mixture. In methane arom-
atization, a higher methane conversion was obtained on
3 wt% Mo/SiO2, especially at the early stage of the reac-
tion, but the selectivitytowardsbenzene formation wasvery
low. Most of the converted methane was deposited as sur-
face carbon on the catalyst (above 85% coke selectivity).
It is interesting to find that the marked increase in ben-
zene formation on 3 wt% Mo/SiO2 + HZSM-5, compared
to that on 3 wt% Mo/SiO2, is due to the dilution with
HZSM-5in the physicalmixture. The primaryproductssuch
as CHx, formed on Mo/SiO2 are converted to C2-species
such as ethene. C2-species could diffuse and migrate to the
interface of HZSM-5 and be converted into benzene and
naphthalene as the aromatic products.
From the above discussion, we suggest the bifunctional
catalysis of Mo/HZSM-5 catalysts, consisting of the Mo
carbide sites and HZSM-5 having higher Bro¨nsted acidity.
Methane is activated on the Mo carbide sites, while HZSM-
SCHEME 1. (a–b) On an Mo site of carbide or oxycarbide; (c) on
5 is responsible not only for the further conversion of the HZSM-5.

188
LIU ET AL.
species to benzene and benzene derivatives by analogy
(5) This bifunctional catalysis of Mo2C/HZSM-5 for a
with the catalytic aromatization of methanol and alkanes methane aromatization reaction is discussed by analogy
at 523–623 K, referred to as “Mobil process by HZSM-5” with the synergetic promotion on Pt/Al2O3 catalyst for the
(26). Nevertheless, it is a controversial problem that such dehydroaromatization of alkanes.
an acidic HZSM-5 remains active for the aromatization of
alkanes at the extremely high temperatures above 923 K.
The evidences of synergitic promotion in the physical mix-
ACKNOWLEDGMENTS
ture of Mo/SiO2 + HZSM-5 and Mo2C + HZSM-5 implies
This work was supported by the Proposal-Based New Industry Creative
Type Technology R&D Promotion Program from the New Energy and
Industrial Technology Development Organization (NEDO) of Japan. The
authors thank Dr. Inagaki and Dr. Fukushima, Toyota Central Research
Laboratory, for their gifts of the mesoporous materials of FSM-16, and
helpful discussion on the bifunctional role of the supports for the methane
that the active carbon species may migrate to HZSM-5 in
the vicinity of Mo carbide sites via the formation of C2 prod-
ucts such as acetylene, ethene, and ethane, which are con-
verted to aromatics such as benzene and naphthalene. In
this sense, the bifunctional role of Mo carbide and HZSM-5
for the methane aromatization has a close analogy with the
synergetic promotion on Pt/alumina catalysts for the alkane
reforming process, consisting of the dehydrogenation of
alkanes to olefins catalyzed on Pt sites which migrate to the
acidic alumina support, resulting in the dehydrocyclization
to benzene derivatives. In addition, Boudart et al. (24) have
reported previously that Mo carbide is catalytically active
for the dehydrogenation towards olefins and hydrogenol-
ysis of alkanes, similar to the typical noble metals such as
Pt and Ir.
dehydroaromatization on the Mo catalysts.
REFERENCES
1. Fox III, J. M., Catal. Rev.-Sci. Eng. 35(2), 169 (1993).
2. Brown, M. J., and Parkyns, N. D., Catal. Today 8, 305 (1991).
3. Kuo, J. C. W., Kresge, C. T., and Palermo, R. E., Catal. Today 4, 463
(1989).
4. Krylov, O. V., Catal. Today 18, 209 (1993).
5. Guczi, L., van Santen, R. A., and Sayma, K. V., Catal. Rev.-Sci. Eng.
38, 249 (1996).
6. Lunsford, J., Catal. Today 6, 235 (1990).
7. Wang, L., Tao, L., Xie, M., Xu, G., Huang, J., and Xu, Y., Catal. Lett.
21, 35 (1993).
8. Xu, Y., Liu, S., Wang, L., Xie, M., and Guo, X., Catal. Lett. 30, 135
(1995).
4. CONCLUSION
The present work may be summerized by the following
conclusions;
9. Solymosi, F., Erdo¨helyi, A., and Szo¨ke, A., Catal. Lett. 32, 43
(1995).
10. Solymosi, F., Szo¨ke, A., and Csere´nyi, J., Catal. Lett. 39, 157 (1996).
11. Szo¨ke, A., and Solymosi, F., Appl. Catal. A: General 142, 361 (1996).
12. Solymosi, F., Csere´nyi, J., Szo¨ke, A., Ba´nsagi, T., and Oszko´, A.,
J. Catal. 165, 156 (1997).
13. Wang, S.-T., Xu, Y., Wang, L., Liu, S., Li, G., Xie, M., and Guo, X.,
Catal. Lett. 38, 39 (1996).
14. Xu, Y., Shu, Y., Liu, S., Huang, J., and Guo, X., Catal. Lett. 35, 233
(1995).
15. Liu, W., Xu, Y., Wong, S.-T., Qiu, J., and Yang, N., J. Mol. Catal. A:
Chemical 120, 257 (1997).
16. Wang, D., Lunsford, J. H., and Rosynek, M. P., J. Catal. 169, 347
(1997).
17. Wang, D., Rosynek, M. P., and Lunsford, J. H., Topics Catal. 3, 289
(1996).
18. Liu, S., Dong, Q., Ohnishi, R., and Ichikawa, M., J.C.S. Chem. Com-
mun., 1455 (1997).
(1) Among the supports such as FSM-16, Mordenite,
USY, alumina, and SiO2, 3 wt% Mo-loading catalysts us-
ing HZSM-5 offer the highest activities and selectivities for
aromatic products such as benzene and naphthalene in the
methane conversion at 973 K and 1 atm.
(2) The TG/DTA/mass and EXAFS studied reveal that
the Mo oxide supported on HZSM-5 is endothermically
converted at 955 K with methane in evolving H2 and CO to
molybdenum carbide Mo2C, which initiates the methane
conversion around 960 K, yielding benzene, naphtha-
lene, and C2 hydrocarbons such as ethene as hydrocarbon
products.
(3) The IR measurements in pyridine adsorption reveal
that Mo/HZSM-5 catalysts having the optimum SiO2/Al2O3
ratio between 20 and 70 show the maximum Bro¨nstead
19. Liu, S., Dong, Q., Ohonishi, R., and Ichikawa, M., Chem. Commun.,
1217 (1998).
20. User Manual of Technos EXAFS analyzing Program.
acidity among the catalysts with the SiO2/Al2O3 ratio of 21a. Basila, M. R., Kanter, T. R., and Rhee, K. H., Nature 68, 3197
(1964).
20–1900. There is a close correlation between the benzene
formation in the methane conversion and their Bro¨nsted
acidity, regardless of the Lewis acidity of the Mo/HZSM-5
21b. Poncelet, G., and Dubru, M. L., J. Catal. 52, 321 (1978).
22. Mirodetos, C., and Barthomeuf, D., J. Catal. 57, 136 (1979).
23. Lee, J. S., and Boudart, M., Catal. Lett. 20, 97 (1993).
catalysts. The highest benzene formation and minimizing
of coke formation are obtained on the Mo/HZSM-5 having
the SiO2/Al2O3 ratio of 40.
(4) A remarkable synergetic promotion in benzene for-
mation exists for the hybrid catalysts consisting of 3 wt%
Mo/SiO2 + HZSM-5 and Mo2C + HZSM-5, whereas each
component by itself shows marginal activity in the reaction.
24. Lee, J. S., Locatelli, S., Oyama, S. T., and Boudart, M., J. Catal. 125,
157 (1990).
25. Ozin, G. A., Prokorowitcz, P. A., and Ozkar, S., J. Am. Chem. Soc.
114, 8957 (1992).
26a. Olson, D. H., Lago, R. M., and Haag, W. O., J. Catal. 61, 390 (1980).
26b. Lago, R. M., Haag, W. O., Mikovsky, R. J., Olson, D. H., Hellring,
S. D., Scmitt, K. D., and Kerr, G. T., Stud. Surf. Sci. Catal. 28, 661
(1986).