A rapid and effective method for evaluating the initial activity of Mo/HZSM-5 catalyst in the methane dehydroaromatization reaction at severe conditions

Article abstract of DOI:10.1016/j.catcom.2010.08.017

The very rapid deactivation behavior of Mo/HZSM-5 catalyst in the non-oxidative CH₄ dehydroaromatization needs its cyclic or continuous regeneration operation in a practical reactor system. Design and operation of such a system requires knowledge of the kinetics of catalyst deactivation and its regeneration. The present work reports an easy, rapid and quantitative approach to follow the dynamic variation of the activity of Mo/HZSM-5 in the very initial stage of CH₄ dehydroaromatization at severe conditions. This approach comprises an on-line sampling and storing of a reacted effluent into a number of volume-known loops at intervals from seconds to minutes, and an off-line analysis of the stored samples. With it a nearly continuous dynamic variation in outlet benzene concentration was observed in the first tens of minutes of the reaction for all tested catalysts with Mo loadings ranging from 1.0 to 10 wt.%. Quantitative comparison of the variation with that in the benzene formation rate measured using conventional on-line GC analysis further revealed that the variation actually exhibited the dynamic behavior of the benzene formation activity of catalyst. The approach was subsequently applied successfully to follow the intrinsic maximum activity of Mo/HZSM-5 at severe conditions and in periodic reaction-regeneration operation mode to collect kinetic data for the reactor design.

Full text of DOI:10.1016/j.catcom.2010.08.017

Catalysis Communications 12 (2010) 127–131
Contents lists available at ScienceDirect
Catalysis Communications
journal homepage: www.elsevier.com/locate/catcom
A rapid and effective method for evaluating the initial activity of Mo/HZSM-5 catalyst
in the methane dehydroaromatization reaction at severe conditions
⁎
Jiangyin Lu, Yuebing Xu, Yoshizo Suzuki, Zhan-Guo Zhang
National Institute of Advanced Industrial Science and Technology (AIST), Onogawa 16-1, Tsukuba 305-8659, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 25 May 2010
Received in revised form 26 July 2010
Accepted 21 August 2010
Available online 26 September 2010
The very rapid deactivation behavior of Mo/HZSM-5 catalyst in the non-oxidative CH4 dehydroaromatization
needs its cyclic or continuous regeneration operation in a practical reactor system. Design and operation of
such a system requires knowledge of the kinetics of catalyst deactivation and its regeneration. The present
work reports an easy, rapid and quantitative approach to follow the dynamic variation of the activity of Mo/
HZSM-5 in the very initial stage of CH4 dehydroaromatization at severe conditions. This approach comprises
an on-line sampling and storing of a reacted effluent into a number of volume-known loops at intervals from
seconds to minutes, and an off-line analysis of the stored samples. With it a nearly continuous dynamic
variation in outlet benzene concentration was observed in the first tens of minutes of the reaction for all
tested catalysts with Mo loadings ranging from 1.0 to 10 wt.%. Quantitative comparison of the variation with
that in the benzene formation rate measured using conventional on-line GC analysis further revealed that
the variation actually exhibited the dynamic behavior of the benzene formation activity of catalyst. The
approach was subsequently applied successfully to follow the intrinsic maximum activity of Mo/HZSM-5 at
severe conditions and in periodic reaction–regeneration operation mode to collect kinetic data for the
reactor design.
Keywords:
Methane dehydroaromatization
Mo/HZSM-5
Maximum activity
Benzene
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
reactor system requires knowledge of the kinetics of catalyst
deactivation and its regeneration at practical operation conditions.
In the past two decades concerted efforts have been made to
investigate the fundamental aspects of the non-oxidative methane
dehydroaromatization over Mo/HZSM-5 [1–9] as well as the industrial
applicability of this new reaction [10–15]. In Japan, a combination of
the technologies of adding a small amount of CO2 to methane to
suppress coking [16] and of using H2 to recover the initial activity of
Mo/HZSM-5 [17] had led to two successful national process
developing projects [14,18].
Case studies conducted in the projects have revealed that this
reaction has to be processed at temperatures not lower than 1073 K
and space velocities not smaller than 10,000 ml/g-cat/h to obtain
performance-acceptable high CH4 conversion and hourly benzene
yield simultaneously. At such severe conditions, however, Mo/HZSM-
5 catalyst could lose its activity to a great degree in a few minutes.
Thus, the problem arises of how to realize cyclic regeneration of
deactivated Mo/HZSM-5 catalyst at short intervals or its continuous
regeneration in a practical reactor system. Fluidized bed technology
allows an easy, high-rate circulation of solid particles between
reactors and therefore can be certainly a solution to the problem
[10]. Design and operation of a dual bed circulating fluidized bed
Therefore, there is a practical need to pursue the dynamic variation of
the activity of Mo/HZSM-5 in the first few minutes of the reaction at
severe conditions. Use of the conventional on-line GC analysis
method, however, is unable to accomplish this task because of its
too long analysis period.
Here we report for the first time an easy, rapid and quantitative
approach to follow dynamic variation of the activity of Mo/HZSM-5
catalyst in the very initial stage of CH4 dehydroaromatization. This
approach consisted of sampling and storing of a reacted effluent into
15 loops of 100 μL on an on-line 16-port sampling valve at intervals
from seconds to minutes immediately after the start of the reaction,
and analysis of the stored samples by an off-line GC. Therefore, it
allowed a nearly continuous monitoring of the rapid variation of
outlet benzene concentration in the initial stage of the reaction, which
was verified further in the study to exhibit the dynamic variation of
the initial activity of Mo/HZSM-5 catalyst itself.
2. Experimental
2.1. Catalyst preparation
A commercial HZSM-5 zeolite (Zeolyst International, Si/Al atomic
ratio=11.5, BET surface area=350 m2/g) was used to prepare the
four tested Mo/HZSM-5 catalysts with 1, 3, 5, and 10 wt.% Mo
⁎
Corresponding author. Tel.: +81 29 861 8091, 8409; fax: +81 29 861 8209.
E-mail address: z.zhang@aist.go.jp (Z.-G. Zhang).
1566-7367/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.catcom.2010.08.017

128
J. Lu et al. / Catalysis Communications 12 (2010) 127–131
Temp. and flow profile
T.C.
503K
MF
He
10
10 Ar/CH4
973 or 1073K
He
Vent
Ar/CH4
He
923K
P
MF
R.T.
Sampling pattern
10%Ar/CH4
0
0
10
20
31
41min
16 ports valve
Vent
FID
100
200
300 sec
Time on stream / min (and sec)
Fig. 1. The schematic of experimental set up, temperature and flow profile and two sampling patterns used.
contents, respectively. The preparation procedure consisted of Mo
impregnation on HZSM-5 with an aqueous solution of ammonium
heptamolybdate, drying at 383 K for 1 h of the impregnated sample
and calcination in air at 773 K for 5 h of the dried sample. The obtained
powdery catalysts were pressed, crushed and sieved to particles in
sizes of 250–500 μm for use. The Mo contents of the four catalysts
were confirmed by ICP analysis to be 0.9, 2.4, 3.9 and 10.3 wt.%,
respectively. Their surface areas were also measured using BET
method to be 350, 340, 290 and 270 m2/g, respectively. Furthermore,
FT-IR measurement of the catalysts was performed to confirm the
dispersion of Mo onto the Bronsted acid sites in the zeolite channels.
Additionally, a H2–CH4–H2 switch blank test was conducted to
confirm the dead volume of the reactor system. The outlet CH4
concentration pattern recorded revealed that a 12 second flow of
50 mL/min CH4 after a H2 flow at room temperature was enough to
purge H2 out of the reactor completely. This implies that any sample
taken 12 s after a flow switching from H2 to CH4 at the used flow rate
will certainly be of the characteristic of the reacted stream.
3. Results and discussion
3.1. Monitoring of dynamic variation of the benzene concentration using
the proposed approach
2.2. Catalyst test
Fig. 2 compares the variations of the outlet benzene concentration
with time on stream measured over all used catalysts at 973 K and
3500 mL/g-cat/h in the proposed approach. In 41 min of the
reaction 15 samples were taken and analyzed and therefore the
smooth curves were obtained. Considering that only one or two points
of data could be obtained in this short reaction period when the
conventional on-line gas analysis technique was applied, one has to
recognize the effectiveness of the proposed approach in following the
rapid variation of outlet benzene concentration in the very initial
stage of the reaction.
Tests were conducted in a flow type fixed-bed quartz reactor
(8 mm i.d.) at atmospheric pressure and two different temperatures
of 973 and 1073 K. Fig. 1 shows the schematic of experimental set up
with the detailed configuration of sampling valves, the flow and
temperature profile and two sampling patterns used. Briefly, 300 mg
of a fresh catalyst packed in the reactor was first carburized in a
17.5 mL/min CH4 stream at 923 K for 30 min, heated in a He stream to
973 or 1073 K, and then at the temperature subjected to a 10% Ar/CH4
flow to start the reaction. Subsequently, the effluent from the reactor
was sampled at pre-designed intervals from seconds to minutes into
15 loops of 100 μL on an on-line 16-port sampling valve held at 503 K.
At the end of the sampling a He stream was flowed through the bed to
stop the reaction. Simultaneously, an analysis of the stored samples
was carried out by an off-line GC/FID (Shimazu GC-14A) with a
Chemipak PH packed column used for the analysis of aromatics in the
samples. The samples were introduced to the GC one by one through
an on-line 6-port valve between the 16-port valve and GC (Fig. 1). The
analysis was performed at a constant column temperature of 423 K
and it took in total no longer than 2 h. It was confirmed that no sample
leak occurred during the sample-stored periods of time for all 15
loops. The confirmation was made via comparing two sets of analyses
of 15 stored gas samples (10% Ar/CH4), one conducted soon after
sampling and the other after the samples were stored in the loops at
503 K for 3 h, and finding little difference in the CH4 peak areas
between the two sets of analyses for each loop. Additionally, absolute
amounts of benzene and naphthalene in each of stored 100 μL
samples were estimated using external calibration method.
It is clear from the figure that the variation of outlet benzene
concentration depends strongly on Mo loading. At 1 wt.% of Mo
loading the benzene concentration reaches the maximum of about
150
120
1 wt% Mo/HZSM-5
3 wt%
5 wt%
90
60
30
a
10 wt%
b
c
0
10
20
30
40
To confirm the reliability and effectiveness of the approach, a set of
comparative tests was also conducted with the simultaneous
measurement of methane conversion, benzene formation rate and
benzene selectivity using the conventional on-line GC analysis
method [11].
Time on stream /min
Fig. 2. Variation of outlet benzene concentration with time in the catalytic methane
dehydroaromatization at 973 K and 3500 mL/g-cat/h, measured in the proposed
approach.

J. Lu et al. / Catalysis Communications 12 (2010) 127–131
129
52 nmol-C/100 μL at the second data point (2 min after the start of the
reaction) and then decreases slowly with time until the reaction was
stopped at 41 min. For the other three catalysts with the Mo loadings
of 3, 5 and 10 wt.% respectively, the benzene concentrations, on the
other hand, show a similar rapid increase in the first 2 to 6 min,
followed by a similar slow increase to their respective maximums. The
period of time needed for the catalysts to reach their respective
maximum activities became longer with increasing Mo loading,
suggesting a longer period of time for the full activation of Mo as its
content increased.
activities of different Mo/HZSM-5 catalysts in the methane
dehydroaromatization.
The cause of the 1 wt.% Mo/HZSM-5 showing the lowest activity
was pursued using NH3-TPD and FT-IR techniques. Obtained NH3-TPD
patterns and IR spectra [19] suggested that this catalyst had more Mo-
unoccupied Bronsted acid sites than the others (fewer Mo sites), and
thus it was concluded that its insufficiently dispersed Mo sites, over
which the activation of CH4 to CHx and then C2 intermediates took
place, must be responsible for its low benzene formation activity [11].
Dispersion of MoO3 onto the Bronsted acid sites in the zeolite
channels was confirmed by Pyridine chemisorption on the tested
catalysts followed by FT-IR measurements. Obtained spectra revealed
a gradual decrease in the intensity of the band 1545 cm−1 with the Mo
loading increasing from 1 to 10 wt.% [19]. This thus suggests that, as
Mo loading increased, there were more Mo species, dispersing into
the deep inside of the zeolite channels to create more active sites for
CH4 activation and benzene formation and consequently to lead to an
increased maximum benzene formation activity. If a difference in the
increase of stream volume through the reaction was assured
negligible for the tested catalysts, the variations observed here in
the outlet benzene concentration with time would exhibit the
dynamic variations of the benzene formation activities of the tested
catalysts. This implies that the proposed approach actually allows a
quantitative comparison of the very initial benzene formation
3.2. Evaluation of catalytic performance in conventional on-line GC
analysis method
To verify the above deduction another set of tests with longer time
was performed at the same temperature and space velocity but with
reacted effluents being analyzed via the conventional on-line analysis
method [11]. Fig. 3 shows the obtained time-courses of CH4
conversion, benzene formation rate and benzene selectivity, respec-
tively. The maximum benzene formation rates are in the order of 10%
Mo/HZSM-5N5% Mo/HZSM-5N3% Mo/HZSM-5 Mo≫1% Mo/HZSM-5,
which is in good agreement with that of the maximum outlet benzene
concentration observed in Fig. 2. This indicates that monitoring of the
variation of outlet benzene concentration in the initial stage of Mo/
HZSM-5 catalyzed methane dehydroaromatization is actually mea-
suring the dynamic variation of the initial activity of the catalyst.
To make more quantitative the above comparison, the data at 4, 24
and 44 min, which were labeled as a, b and c in Figs. 2 and 3(b),
respectively, were further analyzed. Concretely, the relative benzene
concentrations at 4, 24 and 44 min in Fig. 2 and the relative benzene
formation rates at the same time points in Fig. 3(b) were estimated,
respectively, for all the tested catalysts. As can be seen from Table 1, at
each comparative point the difference between the relative benzene
concentration and the relative benzene formation rate is not greater
than 2% in all cases. This excellent agreement strongly suggests that
the outlet benzene concentration curves in Fig. 2 exhibit essentially
the dynamic variations of the initial benzene forming activities of the
tested catalysts in the methane dehydroaromatization reaction, and
thus verify the effectiveness of the proposed approach in quantitative
evaluation of the initial activity of Mo/HZSM-5 catalyst.
15
1 wt% Mo/HZSM-5
(a)
3 wt%
12
5 wt%
10 wt%
9
6
3
3
(b)
3.3. Measurement of the maximum benzene forming activity at 1073 K
and 10,000 mL/g/h
2
Because of its variable, short sampling intervals the proposed
approach can certainly be used to pursue the intrinsic maximum
benzene formation activity of Mo/HZSM-5 catalyst at any severe
condition. As shown in Fig. 4(a), at 1073 K and 10,000 mL/g-cat/
h there appeared no stable activity period for serious coking and the
maximum benzene concentration was observed at approximately
6 min after the start of the reaction. With the conventional on-line
analysis method, on the other hand, the maximum benzene formation
rate was falsely measured to be at the first sampling point (4 min after
the start of the reaction, Fig. 4(b)). By varying the timing for the first
sampling one can find out the true maximum benzene formation
activity via the conventional approach, but accomplishing this task is
1
a
b
c
0
100
(c)
80
60
40
20
0
Table 1
The relative benzene concentration (Rconc.) and the relative benzene formation rate
(Rrate) estimated from Figs. 2 and 3(b) and at the times of 4, 24 and 44 min, respectively.
Mo
loading
a (4 min)
Rconc.
b (24 min)
Rconc.
c (44 min)
Rconc.
0
40
80
120 160 200 240 280 320
Time on stream /min
Rrate
51.2
91.2
96.6
Rrate
29.3
75.5
88.4
Rrate
25.0
71.2
89.2
1%
3%
5%
10%
50.4
92.2
97.8
100
30.1
76.2
90.2
100
25.5
72.1
90.4
100
Fig. 3. Effect of Mo loading on the catalytic performance of Mo/HZSM-5 in the methane
dehydroaromatization at 973 K and 3500 mL/g-cat/h, evaluated using conventional on-
line analysis method.
100
100
100

130
J. Lu et al. / Catalysis Communications 12 (2010) 127–131
240
200
160
120
80
exposures of three different cycles (the 2nd, 8th and 14th) in a 5 h
(a)
Benzene
periodic 5 min CH4–10 min H2 switch test. With the increase in the
number of cycles both the maximum outlet benzene concentration
and the time needed to reach the maximum became lower and
shorter, respectively. While the gradually decreased maximum
benzene concentration with the increase in the number of cycles
surely suggests a graduate deactivation of the catalyst under periodic
operation, the slightly shortened time to attain the maximum might
record dynamic variation of the deactivating process itself. The time
recorded here to reach the maximum at the three cycles reduced from
about 60 to 40 s, but they are all still 3–5 times longer than required
for a full purge of H2 out of the reactor system by CH4 at the tested
flow rate (about 12 s confirmed by the blank test described in the
Experimental). Thus it can be concluded that the data in Fig. 5 fully
exclude the influence of residual H2. That is, the rapid increase of
outlet benzene concentration in all the cases in the first minute in the
figure surely reveals the dynamic behavior of the activity after cyclic
Naphthalene
Toluene
40
0
0
10
20
30
40
15
12
9
(b)
(a) 40 min
H
2 regeneration.
In addition to provide information about when the catalyst will
6
reach its maximum activity under H2–CH4 switching operation, the
curves in Fig. 5 also indicate when the catalyst must be subjected to
regeneration. Presume that it is essential to keep a catalyst exhibiting
no less than, for example, 90% of its maximum activity in the CH4
exposure period (the reaction period) for operation of a practical
reactor system. Thus it is clear from the data for the 2nd cycle in Fig. 5
that the duration of CH4 exposure in each reaction–regeneration cycle
should not be longer than 5 min, that is, the catalyst must be
regenerated at 5 min intervals at the used condition. In a two bed
catalyst-circulating fluidized bed reactor system [10] this short period
of time actually indicates the average residence time of catalyst
particles in its CH4 converter. That is, the proposed approach can also
be used to determine the average residence time of catalyst particles
in either CH4 converter or H2-regenerator of a two bed circulating
fluidized bed reactor system, which is one of the most important
factors to be primarily considered in the reactor and process design.
3
0
0
25
50
75
100
125
Time on stream /min
Fig. 4. Time dependences of the benzene formation activity of 5% Mo/HZSM-5 in the
methane dehydroaromatization at 1073 K and 10,000 mL/g-cat/h, (a) measured in the
proposed approach and (b) measured using the conventional on-line analysis method.
certainly time-consuming. The true maximum benzene forming
activity existing in the very initial stage of the reaction could not be
easily followed by the conventional on-line GC analysis method,
indicating inapplicability of the method for the evaluation of the
catalytic activity of Mo/HZSM-5 at rapidly deactivating conditions.
3.4. Following of the maximum benzene forming activity under periodic
4. Conclusion
CH4–H2 switching operation at 1073 K and 20,000 mL/g/h
As demonstrated all above, the proposed approach enables a rapid
and quantitative evaluation of the very initial activity of Mo/HZSM-5
catalyst at severe conditions and in periodic CH4–H2 switching
operation mode. Its application to the activity evaluation of any
modified Mo/HZSM-5 catalyst at severe operation conditions, for
example, at temperatures and space velocities up to 1173 K and
60,000 mL/g/h, respectively, certainly speeds up the development of
industrially applicable catalysts for the reaction [19]. With the help of
a second on-line 16-port sampling valve, the approach can also allow
an easy acquisition of the intrinsic kinetic data of the catalytic system
at severe conditions [20], which includes CH4 conversion, maximum
benzene formation rate and catalyst deactivation rate, all primarily
necessary for the reactor and process design.
Mo/HZSM-5 catalyst shows a remarkably improved stability in
periodic H2–CH4 switching operation mode [17]. In this mode Mo/
HZSM-5 experiences cyclic exposures to H2 for its regeneration and
therefore its catalytic behavior might vary with the increasing number
of cycles. With the proposed approach such variations are also easily
pursued. Fig. 5 compares three time-courses of the concentrations of
outlet benzene and naphthalene recorded during the 5 min CH4
240
5wt%Mo/HZSM-5
Benzene
200
160
References
[1] D. Wang, J.H. Lunsford, M.P. Rosynek, J. Catal. 169 (1997) 347.
[2] F. Solymosi, J. Cserényi, A. Szöke, T. Bansagi, A. Oszko, J. Catal. 165 (1997) 150.
[3] S. Liu, L. Wang, R. Ohnishi, M. Ichikawa, J. Catal. 181 (1999) 175.
[4] R.W. Borry III, Y.H. Kim, A. Huffsmith, J.A. Reimer, E. Iglesia, J. Phys. Chem. B 103
(1999) 5787.
[5] D. Ma, Y. Shu, X. Han, X. Liu, Y. Xu, X. Bao, J. Phys. Chem. B 105 (2001) 1786.
[6] J.-P. Tessonnier, B. Louis, S. Walspurger, J. Sommer, M.-J. Ledoux, C. Pham-Huu, J.
Phys. Chem. B 110 (2006) 10390.
[7] J.-P. Tessonnier, B. Louis, S. Rigolet, M.-J. Ledoux, C. Pham-Huu, Appl. Catal. A : Gen.
336 (2008) 79.
[8] H. Liu, X. Bao, Y. Xu, J. Catal. 239 (2006) 441.
[9] J.Z. Hu, J.H. Kwak, Y. Wang, C.H.F. Peden, H. Zheng, D. Ma, X. Bao, J. Phys. Chem. C
113 (2009) 2936.
120
80
40
0
,
,
,
2rd cycle
8th
14th
Naphthalene
0
50
100
150
200
250
300
Time on stream /s
Fig. 5. Variations of outlet benzene and naphthalene concentrations with time in the
5 min CH4 exposures of three different cycles of a 5 h periodic 5 min CH4–10 min H2
switching test at 1073 K and 20,000 mL/g-cat/h.
[10] Z.G. Zhang, K. Honda, Y. Suzuki, Proc. of the 9th China–Japan Symposium on Coal
and C1 Chemistry, October 2006, Chengdu, China, 2006, pp. 229–230.
[11] K. Honda, X. Chen, Z.G. Zhang, Appl. Catal. A: Gen. 351 (2008) 122.

J. Lu et al. / Catalysis Communications 12 (2010) 127–131
131
[12] K. Skutil, M Taniewski, Fuel Proc. Technol. 84 (2006) 511.
[13] B. Cook, D. Mousko, W. Hoelderich, R. Zennaro, Appl. Catal. A: Gen. 365 (2008) 34.
[14] Y. Shudo, T. Ohkubo, Y. Hideshima, T. Kiyama, Int J. Hydrogen, Energy 34 (2009)
4500.
[17] K. Honda, T. Yoshida, Z.G. Zhang, Catal. Commun. 4 (2003) 21.
[18] NEDO Research Project on “Technological development of manufacturing clean
hydrogen and petrochemicals by direct methane reforming”, April 1, 2000–March
31, 2003, https://app 5.infoc.nedo.go.jp/disclosure/.
[15] M.P. Gimeno, J. Soler, J. Herguido, M. Menendez, Ind. Eng. Chem. Res. 49 (2010)
996.
[19] Unpublished data. Available upon request.
[20] Detailed configuration available upon request.
[16] S. Liu, Q. Dong, R. Ohnishi, M. Ichikawa, Chem. Commun (1998) 1217.