Continuous production of H2 at low temperature from methane decomposition over Ni-containing catalyst followed by gasification by steam of the carbon on the catalyst in two parallel reactors operated in cyclic manner

Article abstract of DOI:10.1006/jcat.2000.3135

CO-free H₂ in close to stoichiometric amounts was produced continuously at a constant H₂ production rate by the stepwise steam reforming of methane at low temperature (500°C) in two parallel catalytic reactors operated in cyclic manner. The process involved two simultaneous reactions: first, catalytic decomposition of methane to H₂ and carbon (deposited on the catalyst), and second, gasification of the carbon deposited on the catalyst by steam to H₂ and CO₂. The two reactions were carried out separately in two parallel reactors (both containing the same Ni-containing catalyst), operated in cyclic manner by switching a methane-containing feed and a steam-containing feed between the two reactors at a predecided interval of time. The process shows best performance at an optimum value of the feed switchover time. Among the different Ni-containing metal oxide (ZrO₂, MgO, ThO₂, CeO₂, UO₃, B₂O₃ or MoO₃) and zeolite [HZSM-5, Hβ, HM, NaY, Ce(72)NaY, or Si-MCM-41] catalysts, the Ni/ZrO₂ and Ni/Ce(72)NaY showed promising results for the cyclic process.

Full text of DOI:10.1006/jcat.2000.3135

Journal of Catalysis 198, 136–141 (2001)
doi:10.1006/jcat.2000.3135, available online at http://www.idealibrary.com on
Continuous Production of H2 at Low Temperature from Methane
Decomposition over Ni-Containing Catalyst Followed by
Gasification by Steam of the Carbon on the Catalyst in
Two Parallel Reactors Operated in Cyclic Manner
Vasant R. Choudhary,1 Subhabrata Banerjee, and Amarjeet M. Rajput
Chemical Engineering Division, National Chemical Laboratory, Pune 411008, India
Received September 21, 2000; revised November 24, 2000; accepted November 27, 2000; published online February 1, 2001
(>800 C) and produce CO (which is converted by a water–
gasshift reaction, CO + H2O → CO2 + H2, in a downstream
CO-free H2 in close to stoichiometric amounts was produced
continuously at a constant H2 production rate by the stepwise steam
reforming of methane at low temperature (500 C) in two paral-
lel catalytic reactors operated in cyclic manner. The process in-
volved two simultaneous reactions: first, catalytic decomposition
of methane to H2 and carbon (deposited on the catalyst), and sec-
ond, gasification of the carbon deposited on the catalyst by steam
to H2 and CO2. The two reactions were carried out separately in
two parallel reactors (both containing the same Ni-containing cata-
lyst), operated in cyclic manner by switching a methane-containing
feed and a steam-containing feed between the two reactors at a
predecided interval of time. The process shows best performance
at an optimum value of the feed switchover time. Among the dif-
ferent Ni-containing metal oxide (ZrO2, MgO, ThO2, CeO2, UO3,
B2O3 or MoO3) and zeolite [HZSM-5, H , HM, NaY, Ce(72)NaY,
or Si-MCM-41] catalysts, the Ni/ZrO2 and Ni/Ce(72)NaY showed
c
processing step) as one of the major byproducts. Hence,
efforts are being made for producing CO-free hydrogen
from low-temperature catalytic stepwise steam reforming
of methane (1–3), involving the following reactions in con-
secutive steps:
CH4 → C (on catalyst) + 2H2 [1]
C (on catalyst) + 2H2O → CO2 + 2H2.
[2]
Zhang and Amiridis (1) have carried out the above two
reactions alternatively (one after another) over Ni/SiO2.
Choudhary and Goodman (2) have passed alternative
pulses of methane (5% CH4 in He) and water (2 l) over
Ni/ZrO2 at 375 C to obtain CO-free hydrogen from the
above reactions. Very recently, Choudhary and Goodman
(3) have studied the two steps of the process by carrying out
the two reactions above alternatively as a function of tem-
perature and surface coverage of carbon. They observed
that the removal of surface carbon in reaction [2] becomes
increasingly difficult at higher temperatures and surface
coverage of carbon. In all of the above investigations, the
two reactions were carried out one after another in cyclic
promising results for the cyclic process.
2001 Academic Press
KeyWords:low-temperatureCO-freehydrogen production;cata-
lytic methane decomposition; gasification of carbon on catalyst by
steam; Ni-containing catalysts; parallel catalytic reactors operated
in cyclic manner; feed switchover between two reactors.
INTRODUCTION
The demand for hydrogen is ever increasing due to its manner in the same reactor and there was no continuous de-
use in various hydrotreating processes in petroleum in- composition of methane and/or gasification of carbon from
dustries and also for hydrogen fuel cells. Hydrogen is a the catalyst.
nonpolluting fuel and its use as fuel in fuel cells used in
For commercial application, both reactions need to be
automobiles is increasing day by day. However, the hydro- carried out continuously without disturbing feed for reac-
gen fuel cells require CO-free hydrogen to avoid deactiva- tions [1] and [2] for the continuous production of H2 from
tion of its costly noble metal catalyst. Apart from water, methane and water. In this paper, we report a continuous
methane is a preferred source of hydrogen due to its high production of H2 from methane and water in two steps, re-
hydrogen-to-carbon ratio. The conventional steam reform- action [1] followed by reaction [2], by carrying out the two
ing and autothermal reforming processes for the produc- reactions simultaneously and continuously in cyclic man-
tion of H2 from methane are high-temperature processes ner over a Ni-containing catalyst in two parallel fixed bed
reactors at a low temperature (500 C) by switching two
1 To whom correspondence should be addressed. Fax: +91-20-5893041.
E-mail: vrc@ems.ncl.res.in.
feeds—first, a methane–N2 mixture, and second, a steam–
N2 mixture—between the two reactors at a fixed interval of
136
0021-9517/01 $35.00
c
Copyright
2001 by Academic Press
All rights of reproduction in any form reserved.

LOW-TEMPERATURE CONTINUOUS PRODUCTION OF H2 FROM CH4 AND H2O
137
TABLE 1
Results of the Cyclic Stepwise Steam Reforming of Methane over Different Ni-Containing
Metal Oxide (Ni/M = 1.0) Catalysts (Prereduced by H2 at 500 C for 4 h)
Pressure drop
(atm)
H
2 produced per
mole of CH4
Surface area
No. of feed
switchovers
TOF
(mmol g
1
1
1
Catalyst
(m2
g
)
h
)
converted (mol)
Initiala
Finalb
Ni/ZrO2
Ni/MgO
Ni/ThO2
Ni/UO3
Ni/CeO2
Ni/B2O3
Ni/MoO3
Ni–Co/ZrO2
62.8
33.0
59.0
6.0
43.0
37.4
11.4
31.1
36
12
10
12
10
12
12
14
11.3
7.9
9.5
8.7
8.8
7.7
6.4
8.0
3.8
3.6
2.9
3.1
3.0
3.2
3.5
3.7
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.21
0.16
0.18
0.08
0.04
0.03
c
1
Note. Reaction conditions: Feed A = 18.2 mol% CH4 in N2, GHSV of Feed A = 7100 cm3
g
h
1; Feed B = 20.5
mol% steam in N2, GHSV of Feed B = 7290 cm3 g 1 h 1; temperature = 500 C and feed switchover time = 10 min.
a,b Pressure drop across the catalyst bed for the methane decomposition step at the start of the first cycle and at the
end of the last cycle, respectively.
c Ni :Co :Zr = 1 :1 :2.
time. A number of Ni-containing catalysts were evaluated nitrate, uranyl acetate, ammonium molybdate, or boric acid
for their performance in the process.
while grinding the mixed salts in the presence of water
just enough to form a thick paste, which was then dried
and calcined as above. Ni/HZSM-5, Ni/Si-MCM-41, Ni/H-
Mordenite, Ni/H , Ni/NaY, and Ni/Ce(72)NaY catalysts
EXPERIMENTAL
Ni/ZrO2, Ni/ThO2, and Ni-Co/ZrO2 catalysts (Table 1) (Table 2) were prepared by impregnating nickel nitrate on
were prepared by coprecipitating the corresponding metal HZSM-5 (Si/Al = 40), Si-MCM-41, H-Mordenite (Z-900 H,
hydroxides from their aqueous solution containing respec- Norton Co., USA), H , NaY, or CeNaY (degree of Ce+3
tive metal nitrates using sodium hydroxide at pH 9.0 at exchange = 72% ), using the incipient wetness technique,
room temperature. The precipitate was thoroughly washed, followed by drying and calcining in air at 500 C for 4 h. All
dried, and then calcined at 600 C for 2 h under static air. the above catalysts were powdered, pressed, and crushed
Ni/MgO (Table 1) catalyst was also prepared by the similar to particles of 30–52 mesh size.
coprecipitation method except that the precipitating agent
The methane decomposition and carbon gasification re-
used was sodium carbonate. Ni/CeO2, Ni/UO3, Ni/MoO3, actions over the catalysts were carried out in two parallel
and Ni/B2O3 catalysts (Table 1) were prepared by thor- stainless steel fixed bed reactors, each containing a 0.4-g
oughly mixing nickel nitrate with ammonium cerium (IV) catalyst. The two had different feeds but a common outlet.
TABLE 2
Results of the Stepwise Steam Reforming of Methane over Different Nickel (10 wt%) Impregnated
Zeolite Catalysts (Prereduced at 500 C for 4 h)
Pressure drop
H
2 produced per
mole of CH4
(atm)
Surface area
No. of feed
switchovers
TOF
(mmol g
1
1
1
Catalyst
(m2
g
)
h
)
converted (mol)
Initiala
Finalb
Ni/H-ZSM-5
Ni/Si-MCM-41
Ni/H-Mordenite
Ni/H-beta zeolite
Ni/NaY
266.0
786.0
319.7
408.0
450.0
341.8
12
12
18
10
12
12
9.6
8.7
6.9
10.8
9.6
10.7
3.7
3.7
3.6
3.4
3.3
3.6
0.03
0.03
0.03
0.03
0.03
0.03
0.04
0.04
0.04
0.15
0.20
0.05
Ni/Ce(72)NaY
Note. Reaction conditions: same as those given in Table 1.
a,b Pressure drop across the catalyst bed for the methane decomposition step at the start of the first cycle and at the
end of the last cycle, respectively.

138
CHOUDHARY, BANERJEE, AND RAJPUT
A flow switchover valve was connected to inlets of the two rate (35 cm3 min 1) were passed through the TCD and FID
reactors to switch the two different gaseous feeds between detectors connected in series (without any gas chromato-
the two reactors. One feed (called Feed A) was a mixture of graphic column) using nitrogen as a carrier gas, for continu-
methane and nitrogen, and the other (called Feed-B) was a ously measuring the concentrations of hydrogen (by TCD)
mixture of steam and nitrogen. The two feeds were switched and methane (by FID) present in the product gases as a
between the two reactors at a fixed interval of time (called function of reaction time (time-on-stream).
feed switchover time). The experimental arrangement is
The gas hourly space velocity (GHSV) of both the feeds
shown schematically in Fig. 1. Before carrying out the reac- was measured at 0 C and 1 atm pressure. Under the reac-
tions, the catalyst in both reactors was reduced in a flow of tion conditions employed in the present investigation, no
H2–N2 mixture (50 mol% hydrogen) at 500 C for 4 h. The formation of carbon monoxide in the cyclic process over all
reaction temperature in both reactors was controlled by a the catalysts was observed.
Cr–Al thermocouple located in the catalyst bed. The reac-
tor effluent gases were cooled to about 2 C using a coiled
condenser immersed in the ice-water slurry to remove the
RESULTS AND DISCUSSION
water from the product gases. The product gases were col-
Results of a representative experiment for the cyclic
lected in a collapsible plastic bag for a reaction period of 2 h
stepwise steam reforming of methane (at 500 C) over the
and then analyzed by a gas chromatograph with TCD detec-
Ni/ZrO2 catalyst in the two parallel reactors operated in
tor, using Porapak Q and Spherocarb columns. Also in the
cyclic manner with a feed switchover time of 10 min are
separate experiments, the product gases at a constant flow
presented in Fig. 2. In this case, the combined feed streams
(without steam) and the combined gaseous product streams
(after the removal of water from them) are directly passed
through the TCD and FID detectors connected in series.
The frontal chromatograms obtained by the FID (Figs. 2a,
2b) show the relative concentration of methane in the com-
bined feed streams and in the combined product streams.
The chromatogram (Fig. 2b) indicates a steady conversion
of methane (19.6% , after correcting for a change in the gas
flow rate due to the reaction) in the process. The observed
downward spikes, which are due to a momentary change
in flow rate, correspond to the switchover of Feed A and
Feed B between the two reactors (Figs. 1 and 2). The frontal
chromatograms obtained by the TCD (Figs. 2c, 2d) indi-
cate the formation of hydrogen in the process; the detector
response for the product stream is higher because of the
higher thermal conductivity of H2 present in the products,
as compared to the components in the feed. Because of the
detector’s high sensitivity toward a small change in the gas
flow rate, which results from the feed switchover between
the two reactors, upward spikes and up-and-down steps in
the chromatogram (Fig. 2d) are observed. Nevertheless, the
results (Fig. 2) clearly show that H2 can be produced con-
tinuously almost at the same rate and also without affecting
the feed flow rates by carrying out the methane decompo-
sition (reaction [1]) and the carbon gasification by steam
(reaction [2]) simultaneously in two separate reactors op-
erated in parallel, while switching the methane-containing
feed (Feed A) and the steam-containing feed (Feed B) be-
tween the two reactors at a predecided interval of time. Un-
der the experimental conditions, the conversion of methane
FIG. 1. Schematic experimental arrangement for carrying out cata-
lytic methane decomposition and carbon gasification reactions simulta-
neously in two separate parallel catalytic reactors in cyclic manner by
switching two different feeds (Feed A and Feed B) between the two re-
at thermodynamic equilibrium is estimated to be 59.4% .
However the observed methane conversion is less than the
equilibrium one.
The amount of carbon formed in the methane decom-
position for the total reaction period of 6 h (36 cycles) is
actors after an interval of time in order to produce H2 continuously at a
constant rate by the cyclic stepwise steam reforming of methane.

LOW-TEMPERATURE CONTINUOUS PRODUCTION OF H2 FROM CH4 AND H2O
139
FIG. 2. Representative frontal chromatograms obtained with FID and TCD for the combined feed streams and the combined product streams of
the stepwise steam reforming of methane carried out in two parallel reactors, each containing the same catalyst (Ni/ZrO2), at 500 C and with a feed
switchover time of 10 min. Feed A = 18.2 mol% CH4 in N2; Feed B = 20.5 mol% steam in N2; GHSV of Feed A = 7100 cm3 g 1 h 1; GHSV of Feed
1
B = 7290 cm3 g 1 h
.
estimated as 0.82 g per gram of catalyst, whereas the car- was increased continuously from its initial value of 0.02 to
bon leftover on the catalyst after the reaction (at the end 0.12 atm. A steady methane conversion in the decomposi-
of 36th cycle) was 0.044 g per gram catalyst. This shows tion of methane for a long period was also observed earlier
that almost 95% of the carbon formed in the process was for the Ni/SiO2 catalyst (1) and Ni (or Co)/Al2O3 (4) cata-
gasified in the steam treatment cycles and hence there was lyst. However, a need for the intermittent removalofcarbon
no increase in the pressure drop across the catalyst bed in from the catalyst by its gasification by steam is necessary to
both parallel reactors. However, when the carbon formed avoid the undesirable large pressure drop across the cata-
in the methane decomposition was not removed intermit- lyst due to formation of excessive carbon between the cata-
tently by its gasification with steam, there was an appre- lyst particles. Moreover, the removal of carbon from the
ciable and continuous increase in the pressure drop with catalyst is more and more difficult at higher and higher sur-
an increase in the time-on-stream, particularly about the face coverage of carbon (3). The aging of carbon during the
reaction period of 30 min.
methane decomposition process is also expected to cause a
It may also be noted that the catalyst did not show a sign change in the nature and/or form of the carbon deposited
of deactivation in the process for a sufficiently long period on the catalyst, making removal of the carbon increasingly
(Fig. 2). However, when we carried out the methane de- difficult. Moreover, the aging may lead to a permanent de-
composition alone (without the cyclic operation) over the activation of the catalyst or reduce drastically the catalyst
same catalyst under similar conditions continuously for a regenerability.
period of 2 h, we observed a steady conversion of methane
Results showing the influence of the feed switchover
(19.6 0.2% ), even though there was a continuous built up time on the process performance are presented in Fig. 3.
ofcarbon (estimated as0.27gper gram at the end ofthe run) Both the methane conversion and the H2 productivity, and
on the catalyst. The pressure drop across the catalyst bed consequently the CO2 produced in the process, are passed

140
CHOUDHARY, BANERJEE, AND RAJPUT
switchover time (10 min) is unusual. At the very short feed
switchover time (1.0 min), the carbon deposited on the cat-
alyst is expected to be much smaller and hence a higher
methane conversion is expected. The observed higher
methane conversion at the intermediate feed switchover
time indicates a beneficial effect of the presence of carbon
species at an optimum concentration on the surface of the
catalyst for it to be more active in the methane decom-
position. A further detailed investigation is required for
understanding the observed optimum feed switchover time
in the process of this investigation. Also longer duration
tests (>100 h) are necessary to assess deactivation and to
determine steady-state carbon deposition on the catalyst.
Results on a few more Ni-containing metal oxides and ze-
olite catalysts for the stepwise steam reforming of methane
(at 500 C) operated in the cyclic manner in the two parallel
reactors with a feed switchover time of 10 min for a time-
on-stream of 2 h are presented in Tables 1 and 2. TOF was
estimated as the amount of methane converted per unit
mass of catalyst per unit time. The initial and final pres-
sure drops across the catalyst bed were measured for the
methane decomposition step in the cyclic process.
Among the Ni-containing metal oxide catalysts (Table 1),
the Ni/ZrO2 showed the best performance in the cyclic
process, whereas among the Ni-containing zeolite catalysts
(Table 2), the Ni/Ce(72)NaY showed the best performance.
Although the Ni/H showed methane conversion activity
comparable to that of Ni/Ce(72)NaY, the former produced
less H2 per mole of methane converted and also a larger
amount of carbon accumulated on the catalyst, resulting in
an undesirable pressure drop across the catalyst bed. The
Ni/UO3, Ni/ThO2, Ni/CeO2, Ni/B2O3, and Ni/NaY catalysts
also showed a high pressure drop across the catalyst bed;
the pressure drop was found to increase exponentially with
increasing reaction period (or number of cycles). The high
pressure drop observed for these catalysts resulted mainly
due to the formation of filamentous carbon blocking the
interparticle voids in the catalyst bed. This was confirmed
by observing a drastic reduction in the pressure drop, ap-
proaching its initial value, due to an air oxidation of the
catalyst at 500 C. Because of the larger amount of carbon
accumulated on these catalysts, the H2 produced per mole
of the methane converted was appreciably lower than the
expected theoretical value (4.0). For the cyclic process, the
Ni/ZrO2 catalyst, however, showed the best performance—
highest TOF (methane conversion activity), highest moles
of H2 (3.8) produced per mole of CH4 converted, and al-
most no pressure drop across the catalyst bed, even when
operated for a much larger number of cycles.
FIG. 3. Effect of feed switchover time on the process performance
using Ni/ZrO2 (Ni/Zr = 1.0) catalyst. Feed A = 20 mol% CH4 in N2; Feed
B = 80.9 mol% steam in N2; GHSV = 3225 (feed A) and 6770 (feed B)
1
cm3
g
h
1; temperature in both reactors = 500 C.
through a maximum (at a feed switchover time of about
10 min) with increasing feed switchover time from 1.0 to
25.0 min. The amount of H2 produced per mole of the
methane converted was 3.9 0.05, which is very close to
the stoichiometric amount (4.0) expected to be formed in
the cyclic process. The results clearly show that there is an
optimum value for the feed switchover time; it is 10 min
for the Ni/ZrO2 catalyst under the specified reaction con-
ditions. However, the optimum value may vary with the
reaction conditions for the same catalyst and it may also
differ from catalyst to catalyst. Further studies are required
for knowing the carbon retained on the catalyst after each
successive cycle (i.e., the methane decomposition followed
by the carbon gasification) and its dependence on the dif-
ferent process conditions.
The observed lower methane conversion and conse-
quently the lower H2 productivity, at the high switchover
time (25 min), is certainly due to a larger amount of carbon
deposited on the catalyst. However, the observed lower
methane conversion for the shorter feed switchover time
(1 and 5 min) as compared to that at the optimum
In their earlier cyclic pulse studies, Choudhary and
Goodman (2, 3) obtained 1.0 to 1.3 mol H2 per mole of
methane consumed in the methane decomposition step
over Ni/ZrO2 catalyst at 375 C. However, in the present
cyclic process the H2 produced is almost close to the

LOW-TEMPERATURE CONTINUOUS PRODUCTION OF H2 FROM CH4 AND H2O
141
stoichiometric amounts. This is expected most probably be- by the cyclic stepwise steam reforming of methane at a low
cause of the fact that, unlike the cyclic pulse process, the temperature (500 C), by carrying out catalytic methane de-
present process operates under or very close to a steady composition and gasification of carbon deposited on the
state and also involves the reforming by steam of partially catalyst during the methane decomposition separately in
hydrogenated carbon species (produced in the methane de- two parallel catalytic reactors containing the same cata-
composition step) into CO2 and H2 in the steam treatment lyst, operated in cyclic manner by switching the methane-
step.
containing feed and the steam-containing feed between the
Apart from its low-temperature operation, another im- two reactors at a known interval of time.
portant advantage of the present process is that the CO
(ii) The feed switchover time has an optimum value for
conversion by the water gas-shift reaction, which is an in- achieving the best performance in the process.
tegral part of the H2 production processes practiced at (iii) Ni/ZrO2 and Ni/Ce(72)NaY are promising catalysts
present, is not required, thus eliminating the low- and high- for the cyclic stepwise steam reforming of methane to hy-
temperature water-gas shift reactors for the downstream drogen and CO2 at a low temperature (500 C).
processing of the product stream. The commercial feasi-
bility of the process will be increased by achieving higher
methane conversion close to that at equilibrium through op-
ACKNOWLEDGMENT
timizing process parameters and developing a better cata-
lyst for the process.
S. Banerjee is grateful to the Council of Scientific and Industrial Re-
search, New Delhi, India, for awarding him a Senior Research Fellowship.
CONCLUSIONS
REFERENCES
The following important conclusions have been drawn
from this investigation:
1. Zhang, T., and Amiridis, M. D., Appl. Catal A: Gen. 167, 161 (1998).
2. Choudhary, T. V., and Goodman, D. W., Catal. Lett. 59, 93 (1999).
3. Choudhary, T. V., and Goodman, D. W., J. Catal. 192, 316 (2000).
4. Avdeeva, L. B., Kochubey, D. I., and Shaikhutdinov, Sh. K., Appl. Catal.
A: Gen. 177, 43 (1999).
(i) It is possible to produce continuously hydrogen at
a constant rate and also close to stoichiometric amounts