www.elsevier.com/locate/jcice

Available online at www.sciencedirect.com

l Engineers 39 (2008) 129–136

Journal of the Chinese Institute of Chemica

Effects of system parameters on the performance of CO2-selective

WGS membrane reactor for fuel cells

Jin Huang a,1, W.S. Winston Ho a,b,*a Department of Chemical and Biomolecular Engineering, The Ohio State University, 2041 College Road, Columbus, OH 43210-1180, USA

b Department of Materials Science and Engineering, The Ohio State University, Columbus, OH 43210-1180, USA

Abstract

Performing water gas shift (WGS) reaction efficiently is critical to hydrogen purification for fuel cells. In our earlier work, we proposed a CO2-

selective WGS membrane reactor, developed a one-dimensional non-isothermal model to simulate the simultaneous reaction and transport process and

verified the model experimentally under an isothermal condition. Further modeling investigations were made on the effects of several important system

parameters, including inlet feed temperature, inlet sweep temperature, feed-side pressure, feed inlet CO concentration, and catalyst activity, on

membrane reactor performance. The synthesis gases from both autothermal reforming and steam reforming were used as the feed gas. As the inlet feed

temperature increased, the required membrane area reduced because of the higher WGS reaction rate. Increasing the inlet sweep temperature decreased

the required membrane area more significantly, even though the required membrane area increased slightly when the inlet sweep temperature exceeded

about 160 8C. Higher feed-side pressure decreased the required membrane area as a result of the higher permeation driving force and reaction rate. A

potentially more active catalyst could make the membrane reactor more compact because of the enhanced reaction rate. The modeling results have

shown that a CO concentration of less than 10 ppm is achievable from syngases containing up to 10% CO.

# 2007 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: CO2-selective membrane; Membrane reactor; Water gas shift reaction; Reforming syngas; Modeling; Catalyst activity; Fuel cell

1. Introduction

As an efficient and environmentally friendly energy

conversion device, fuel cell has attracted worldwide interest

for both transportation and stationary power generation in the

recent years (Acres, 2001; Barnett and Teagan, 1992; Cropper

et al., 2004). Hydrogen is the most common fuel for the fuel

cell. Due to the lack of large capacity storage systems,

hydrogen used in an automotive fuel cell is suggested to be

generated by on-site reforming reactions of the commonly

available fuels, such as methanol, natural gas, gasoline, diesel,

and bio fuel. Steam reforming (SR), partial oxidation (POX)

and autothermal reforming (ATR) are three major reforming

processes (Brown, 2001; Ghenciu, 2002).

For transportation fuel cell applications, an important

concern is the purity of hydrogen. The platinum electrocatalyst

* Corresponding author at: Department of Chemical and Biomolecular

Engineering, The Ohio State University, 2041 College Road, Columbus, OH

43210-1180, USA. Tel.: +1 614 292 9970; fax: +1 614 292 3769.

E-mail address: ho@chbmeng.ohio-state.edu (W.S.W. Ho).1 Current address: OSIsoft Inc., 777 Davis Street, Suite 250, San Leandro, CA

94577-6923, USA.

0368-1653/$ – see front matter # 2007 Taiwan Institute of Chemical Engineers.

doi:10.1016/j.jcice.2007.12.004

in polymer electrolyte membrane fuel cells would be poisoned

severely and irreversibly by even a very small amount of CO in

the hydrogen, e.g., >10 parts per million (ppm) (Ahmed and

Krumpelt, 2001). In addition, in order to fit into the restricted

space in a vehicle, the whole fuel processor needs to be light and

compact. Many studies have been conducted to explore CO-

free fuel processing methods. Typically, water gas shift (WGS)

reaction is used to convert CO in the synthesis gas from the

reformer and generate additional H2.

COðgÞ þ H2OðgÞ $ CO2ðgÞ þ H2ðgÞ DHr¼ �41:1 kJ=mol

(1)

After that, a methanation or CO preferential oxidation step

can be used to decrease the CO concentration further to about

10 ppm.

The WGS catalysts have been studied for a long time. A two-

stage process is usually used in the industry: Fe3O4/Cr2O3

catalyst is used for the high-temperature shift (HTS), and Cu/

ZnO/Al2O3 catalyst is used for the low-temperature shift (LTS).

After the development of Cu/ZnO/Al2O3 during 1960s, the

main industrial application of WGS reaction has been for the

generation of H2 for ammonia production and petroleum

Published by Elsevier B.V. All rights reserved.

Nomenclature

h thickness of catalyst layer (cm)

J permeation flux (mol/cm2 s)

KT reaction equilibrium constant (1/atm2)

‘ membrane active layer thickness (cm)

n molar flow rate (mol/s)

p pressure (atm)

P permeability (Barrer)

r volumetric reaction rate (mol/cm3 s)

R ideal gas constant (atm cm3/mol K)

T temperature (8C)

w width of reactor (cm)

x feed-side molar fraction

y sweep-side molar fraction

z axial position along the length of reactor (cm)

Greek symbols

a CO2/H2 selectivity

rb catalyst bulk density (g/cm3)

Subscripts

0 initial

f feed-side

i species

s sweep-side

t total

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136130

refining (Choi and Stenger, 2003; Rhodes et al., 1995). The

issues of an automotive application are fundamentally different

from those of industrial use. The suitable catalysts should be

very active and poison-resistant materials, which would result

in a small reactor volume, short start-up time and good stability

under transient or steady state conditions.

There are many studies focusing on catalyst preparation and

the reaction mechanism in the literature for both classical and

novel LT-WGS catalysts. The traditional Cu/ZnO/Al2O3 catalyst

has shown good activity at about 180–240 8C as the result of the

optimal adsorption heat of CO on Cu (about 80 kJ/mol)

(Grenoble et al., 1981). A number of studies on the reaction

kinetics have been published (Amadeo and Laborde, 1995;

Campbell, 1977; Fiolitakis et al., 1980; Keiski et al., 1993; Moe,

1962; Salmi and Hakkarainen, 1989). The main drawbacks of

this catalyst are that it is pyrophoric and highly sensitive to sulfur

poisoning. The Novel WGS catalysts active at low temperatures

are normally based on platinum or gold. For many years, Pt/CeO2

has been proven to be a very active WGS catalyst especially at

high temperatures (Hilaire et al., 2004; Mendelovici and

Steinberg, 1985; Swartz et al., 2001). However, the activity

and stability of these ceria-supported catalysts need to be

improved by optimizing the preparation conditions and

formulations. Newly developed Pt-containing catalysts have

shown improved durability over a wide range of temperatures

(200–500 8C) (Ghenciu, 2002).

Gold-based catalyst for WGS reaction has been a new

alternative investigated for the past decade (Andreeva, 2002;

Andreeva et al., 2002; Fu et al., 2001; Tabakova et al., 2004). It

has been reported that Au/Fe2O3 has higher activity than Cu/

ZnO/Al2O3 starting from 120 8C (Andreeva, 2002). Addition-

ally, improved stability has been obtained from Au/CeO2 (Fu

et al., 2001). The activity and stability of these catalysts,

however, are sensitive to the preparation techniques, gold

particle size and specific gold-support interaction. Recent

research by Fu et al. has shown that the active sites in these

catalysts are gold or platinum ions strongly associated with the

surface oxygen of ceria instead of gold or platinum metal

nanoparticles (Fu et al., 2003). This finding might significantly

decrease the requirement for the precious metal.

In addition to the advancement of the WGS catalyst, a

membrane reactor is another promising approach to enhance

the CO removal and decrease the reactor size. The general

principle of the membrane reactor is to improve the reaction

performance with the in situ separation of one or more products.

Several studies have been done on H2-selective membrane

reactors, mainly based on palladium membranes and using

high-temperature WGS catalysts (Armor, 1998; Basile et al.,

1996; Criscuoli et al., 2000; Ma and Lund, 2003; Tosti et al.,

2003; Uemiya et al., 1991; Xue et al., 1996). By using a novel

type of membrane (Huang et al., in press; Ho, 2000; Zou and

Ho, 2006; Zou et al., 2007), the CO2-selective WGS membrane

reactor is more advantageous than the H2-selective membrane

reactor because (1) a H2-rich product is recovered at high

pressure (feed gas pressure) and (2) air can be used to sweep the

permeate, CO2, on the low-pressure side of the membrane to

obtain a high driving force for the separation.

In an earlier paper (Huang et al., 2005), we proposed a CO2-

selective membrane reactor and developed a one-dimensional

non-isothermal model to simulate the simultaneous reaction

and transport process in the countercurrent WGS hollow-fiber

membrane reactor. The modeling results have showed that a CO

concentration of less than 10 ppm, a H2 recovery of greater than

97%, and a H2 concentration of greater than 54% (on the dry

basis) are achievable from autothermal reforming syngas

derived from gasoline using air as the oxygen source. If steam

reforming syngas is used as the feed gas, H2 concentration can

be as high as 99.64% (on the dry basis). The model was later

verified experimentally using a rectangular WGS membrane

reactor with an autothermal reforming syngas at 150 8C (Zou

et al., 2007). For further investigation in this paper, we studied

the effects of several important system parameters, including

inlet feed temperature, inlet sweep temperature, feed-side

pressure, inlet feed CO concentration, and catalyst activity, on

the performance of the membrane reactor by using the

established model. Both autothermal reforming syngas and

steam reforming syngas were used as the feed gases.

2. Calculation description

As described in the earlier paper (Huang et al., 2005), the

WGS membrane reactor was configured to be a hollow-fiber

membrane module with catalyst particles packed inside the

fibers (Fig. 1). The hollow-fiber module was assumed to be

composed of CO2-selective facilitated transport membrane.

Fig. 1. Schematic diagram of water gas shift hollow-fiber membrane reactor.

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136 131

The CO2/H2 selectivity, expressed in Eq. (2) (Ho and Sirkar,

1992), was set as 40.

a ¼yCO2

=yH2

xCO2=xH2

(2)

The CO2 permeability of the membrane was set as 4000

Barrers (1 Barrer = 10�10 cm3 (STP) cm/cm2 s cm Hg), which

was fixed in the calculations and did not change with

temperature variation in the module. It should be noted that

the values of the CO2/H2 selectivity and CO2 permeability were

based on the experimental data obtained from CO2-selective

membranes containing amines in crosslinked poly(vinyl

alcohol) (Huang et al., in press; Zou and Ho, 2006; Zou

et al., 2007).

The catalyst packed was assumed to be the commercial Cu/

ZnO catalyst for low-temperature WGS reaction. Because of

the similarity of our operation conditions to those used by

Keiski et al. and the fact that steam is in excess in most of the

membrane reactors, their first reaction rate expression was

chosen for this work (Keiski et al., 1993). The reaction rate is

given by the following equation:

ri ¼ 1:0� 10�3 rb pf

ntRT f

exp

�13:39� 5557

T f

�nCO

�1� nf;H2

nf;CO2

KTnf;COnf;H2O

�(3)

Table 1

The compositions of autothermal reforming syngas and steam reforming syngas

Synthesis gas CO (%) H2O (%)

Autothermal reforming syngas 1 9.5

5 13.5

10 18.5

Steam reforming syngas 1 18.2

5 22.2

10 27.2

where the expression for KT (Moe, 1962) is as follows:

KT ¼ exp

�� 4:33þ 4577:8

T f

�(4)

Based on the volume element of the countercurrent

membrane reactor, the molar and energy balances were

performed on both feed (lumen) and sweep (shell) sides,

respectively. The detailed equations and boundary conditions

were shown in our earlier publication (Huang et al., 2005). By

using the bvp4c solver in Matlab1, resulting differential

equations of the boundary value problem were solved. In the

calculation, the hollow-fiber number was adjusted to satisfy

the constraint of feed exit CO concentration, i.e., <10 ppm.

For autothermal reforming syngas and steam reforming

syngas, the feed mole flow rate was 1 and 0.635 mol/s,

respectively. With the compositions given in Table 1, these

flow rates would hence provide a sufficient H2 molar flow rate

to generate a power of 50 kW via the fuel cell for a five-

passenger car (Brown, 2001).

3. Comparison with experimental data

Inspired by the modeling results, we built a flat-sheet

rectangular permeation cell for membrane reactor experiments

(Zou et al., 2007). This cell had a width of 17.5 cm and an

active membrane area of 342.7 cm2 and was specially designed

to make the gas flows on both the feed side and the sweep side

as ideal plug-flows. The catalyst used in the membrane reactor

experiments was Cu/ZnO/Al2O3 low-temperature WGS cata-

lyst (C18-AMT-2) obtained from Sud-Chemie Inc. (Louisville,

KY, USA). To facilitate the comparison with the experimental

results, the one-dimensional model was revised to reflect the

actual reactor dimension and process conditions. The tem-

perature was set to be a constant value of 150 8C, and the molar

balances for the gas species i on the feed side and the sweep side

were expressed as Eqs. (5) and (6), respectively.

Feed side :dnfi

dz¼ wh ri � w Ji

�Ji ¼ pi

D pi

‘

�(5)

Sweep side :dnsi

dz¼ �w Ji (6)

Fig. 2 shows the results obtained from both modeling and

experimental studies on the rectangular WGS membrane

reactor (Zou et al., 2007). In the calculation, we assumed that

the membrane had a CO2 permeability of 6500 Barrers, a

H2 (%) CO2 (%) N2 (%) CH4 (%)

41 15 33.5 0

37 11 33.5 0

32 6 33.5 0

65.1 15.5 0 0.2

61.1 11.5 0 0.2

56.1 6.5 0 0.2

Fig. 2. Retentate CO concentration vs. feed flow rate in the rectangular WGS

membrane reactor (feed gas: 1% CO, 17% CO2, 45% H2, 37% N2, T = 150 8C,

pf = 2.0 atm, ps = 1.0 atm, sweep-to-feed molar flow rate ratio = 1 (on the dry

basis)) (Zou et al., 2007).

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136132

CO2/H2 selectivity of 40, and negligible N2 and CO

permeation, which were reasonable based on our previous

results (Huang et al., in press; Zou and Ho, 2006; Zou et al.,

2007). In the experiments at 150 8C, the synthesis gas feed

gas was 1% CO, 17% CO2, 45% H2, 37% N2, the sweep-to-

feed molar flow rate ratio was 1 (on the dry basis), and the

feed and sweep pressures were 2 and 1 atm, respectively. As

shown in this figure, for the various feed flow rates from 20 to

70 cm3/min, the CO concentration on the feed side was

reduced from 1% to less than 10 ppm (on the dry basis) from

the experiments, which was equivalent to nearly 100% CO

conversion. As also shown in this figure, the data agreed well

with the prediction by the isothermal mathematical model

described earlier based on the material balances, membrane

permeation, and the WGS reaction kinetics for the Cu/ZnO/

Al2O3 catalyst. As the feed flow rate increased, the retentate

CO concentration slightly increased owing to the reduced

residence time. The good agreement between the experi-

mental data and the modeling prediction proved the validity

of the modeling assumptions.

Fig. 3. The effect of inlet feed temperature on required membrane area for

autothermal reforming syngas.

4. Results and discussion

4.1. Autothermal reforming syngas

A reference case was chosen with the inlet sweep-to-feed

molar flow rate ratio of 1, the membrane thickness of 5 mm,

52,500 hollow fibers (a length of 61 cm, an inner diameter of

0.1 cm, and a porous support with a porosity of 50% and a

thickness of 30 mm), both inlet feed and sweep temperatures of

140 8C, and the feed and sweep pressures of 3 and 1 atm,

respectively. For this case, the calculated CO concentration and

H2 recovery were 9.82 ppm and 97.38%, respectively. Based on

the reference case, the effects of inlet feed temperature, inlet

sweep temperature, feed-side pressure, inlet feed CO con-

centration, and catalyst activity on the reactor performance

were investigated.

4.1.1. Effect of inlet feed temperature

The membrane areas required for the exit feed CO

concentration of <10 ppm in the H2 product were calculated

with seven different inlet feed temperatures ranging from 80 to

200 8C, while the other parameters for the reference case were

kept constant. As shown in Fig. 3, the required membrane area

or hollow-fiber number decreased as the inlet feed temperature

increased. It gradually approached an asymptotic value. The

feed-side temperature profiles for different inlet feed tempera-

tures are presented in Fig. 4. The feed-side temperature

increased as inlet feed temperature increased, especially at the

entrance section. The higher feed-side temperature brought a

higher WGS reaction rate, and thus a less reactor size or catalyst

volume was required. The unfavorable WGS equilibrium at

high temperatures was compensated by the simultaneous CO2

removal.

4.1.2. Effect of inlet sweep temperature

In order to study the impact of inlet sweep temperature on

the membrane reactor performance, Ts0 = 80, 100, 120, 140,

160, 180 and 200 8C were applied in the model while the other

parameters for the reference case were kept constant. As

demonstrated in Fig. 5, the required membrane area or hollow-

Fig. 4. Feed-side temperature profiles along the length of membrane reactor for

autothermal reforming syngas with different inlet feed temperatures.

Fig. 5. The effect of inlet sweep temperature on required membrane area for

autothermal reforming syngas.

Fig. 7. The effect of feed-side pressure on required membrane area for

autothermal reforming syngas.

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136 133

fiber number dropped rapidly as the inlet sweep temperature

increased from 80 to 160 8C. But, beyond 160 8C, it increased

slightly. Fig. 6 depicts the feed-side temperature profiles along

the membrane reactor with different inlet sweep temperatures.

Increasing the inlet sweep temperature increased the feed-side

temperature significantly over a longer reactor length in

comparison with increasing the inlet feed temperature as shown

in Fig. 4. The higher feed-side temperature resulted in the

higher WGS reaction rate. When the inlet sweep temperature

exceeded about 160 8C, the WGS reaction equilibrium became

less favorable for CO conversion, and the overall system

became more mass transfer controlled. Hence, more membrane

area was needed to remove the generated CO2 to achieve

<10 ppm CO in the H2 product.

4.1.3. Effect of feed-side pressure

Fig. 7 gives the required membrane area for different feed-

side pressures. As the pressure increased from 1 to 2 atm, the

required hollow-fiber number decreased drastically from

390,000 to 89,000. Beyond that, the hollow-fiber number

continued to decrease gradually, but much slowly. Obviously,

the increasing feed-side pressure brought a higher CO2

permeation driving force, but the pressure of 1 atm would be

too small to provide enough driving force. Additionally, for a

gas phase reaction, the WGS reaction rate is proportional to the

Fig. 6. Feed-side temperature profiles along the length of membrane reactor for

autothermal reforming syngas with different inlet sweep temperatures.

feed-side pressure, pf, which was evidenced by Eq. (3).

Therefore, the higher feed-side pressure would not only

increase the mass transfer driving force, but also enhance the

reaction rate.

4.1.4. Effect of inlet feed CO concentration

The composition of autothermal syngas varies with the

different fuel type and reforming conditions. The autothermal

syngases containing higher feed inlet CO concentrations, e.g.,

5% and 10%, were used as the feed gases in the calculation.

Fig. 8 shows the feed-side CO mole fraction profiles for 1%, 5%

and 10% CO. The required hollow-fiber numbers for 5% and

10% CO were 51,500 and 52,000, respectively, which were

close to 52,500 for the 1% CO feed gas in the reference case.

This could be explained due to the fact that the module was

assumed to be adiabatic. A higher CO amount in the feed gas

could require a larger reactor volume to convert CO completely,

but it also generated more WGS reaction heat in the adiabatic

module and increased the feed-side temperature. Increasing the

feed-side temperature decreased the reactor size as described

earlier. The balance between these two opposite effects resulted

in similar membrane area requirements for these three syngases

with different CO concentrations.

4.1.5. Effect of catalyst activity

The effect of catalyst activity on the required membrane area

was studied by assuming several WGS reaction kinetics based

Fig. 8. Feed-side CO mole fraction profiles along the length of membrane

reactor for autothermal reforming syngases with different inlet feed CO

concentrations.

Fig. 9. The effect of catalyst activity on required membrane area for auto-

thermal reforming syngas.Fig. 11. Feed-side temperature profiles along the length of membrane reactor

for steam reforming syngas with different inlet feed temperatures.

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136134

on the Cu/ZnO kinetics equation proposed by Keiski et al.

(1993). In Fig. 9, the number on the horizontal x-axis indicates

the reaction kinetic rate in terms of the Cu/ZnO kinetics, e.g., 1

represents the Cu/ZnO kinetics, 2 represents a kinetics two

times the reaction rate of the Cu/ZnO kinetics, etc. As

illustrated in this figure, increasing catalyst activity decreased

the required membrane area significantly. The higher catalyst

activity resulted in a higher reaction rate, which also increased

the CO2 permeation rate because of a higher CO2 partial

pressure on the feed-side and thus a higher driving force across

the membrane. Hence, with the development of more active

WGS catalysts, the membrane reactor would become more

compact.

4.2. Steam reforming syngas

For the steam reforming syngas, we chose a reference case

with the same system parameter values as those for the

autothermal reforming syngas but with 31,000 hollow fibers

with the same dimensions as described earlier. The reduced

number of hollow fibers for the steam reforming syngas was

due to the fact that this syngas had a lower flow rate than the

autothermal reforming syngas because of a higher H2

concentration. For this case, the calculated CO concentration

and H2 recovery were 9.82 ppm and 97.38%, respectively.

Fig. 10. The effect of inlet feed temperature on required membrane area for

steam reforming syngas.

Similarly, the effects of inlet feed temperature, inlet sweep

temperature, feed-side pressure, inlet feed CO concentration,

and catalyst activity on the reactor behavior were investi-

gated for the steam reforming syngas, based on the reference

case.

4.2.1. Effect of inlet feed temperature

The inlet feed temperatures of 80, 100, 120, 140, 160, 180

and 200 8C were applied in the model to study the impact of

inlet feed temperature on the membrane reactor performance

while the other parameters for the reference case were kept

constant. As demonstrated in Fig. 10, the curve for the required

membrane area showed the consistent trend with that for the

autothermal reforming syngas. The required membrane area

reduced as the inlet feed temperature increased. Fig. 11 shows

the feed-side temperature profiles for different inlet feed

temperatures. As explained earlier, the higher feed-side

temperature from the higher inlet feed temperature increased

the WGS reaction rate and decreased the membrane area

requirement.

4.2.2. Effect of inlet sweep temperature

Seven different inlet sweep temperatures ranging from 80 to

200 8C were used in the calculation while other parameters for

Fig. 12. The effect of inlet sweep temperature on required membrane area for

steam reforming syngas.

Fig. 13. Feed-side temperature profiles along the length of membrane reactor

for steam reforming syngas with different inlet sweep temperatures.

Fig. 15. Feed-side CO mole fraction profiles along the length of membrane

reactor for steam reforming syngases with different inlet feed CO concentra-

tions.

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136 135

the reference case were kept constant. The effect of inlet sweep

temperature on the required membrane area for the exit feed CO

concentration of<10 ppm is presented in Fig. 12. Similar to the

autothermal reforming syngas, the required hollow-fiber

number reduced significantly as the inlet sweep temperature

increased from 80 to 170 8C, and a minimum value existed at

about 170 8C. As shown in Fig. 13, the higher inlet sweep

temperature increased the feed-side temperature significantly

over most of the reactor length, which increased the WGS

reaction rate. When the inlet sweep temperature exceeded about

170 8C, the WGS reaction equilibrium became less favorable,

and the overall system became more mass transfer controlled.

Hence, more membrane area was needed to decrease the feed

exit CO concentration to less than 10 ppm.

4.2.3. Effect of feed-side pressure

The feed-side pressures of 1, 2, 3, 4, 5, and 6 atm were

used in the calculation while the other parameters for the

reference case were kept constant. Fig. 14 demonstrates the

required membrane area for different feed-side pressures.

Similar to the autothermal reforming syngas, the required

hollow-fiber number reduced significantly as the feed

pressure increased from 1 to 2 atm. Beyond 2 atm, the

hollow-fiber number continued to decrease gradually, but

much slowly. As described earlier, the higher feed pressure

not only enhanced the CO2 permeation but also increased the

WGS reaction rate.

Fig. 14. The effect of feed-side pressure on required membrane area for steam

reforming syngas.

4.2.4. Effect of feed inlet CO concentration

To investigate the effect of CO concentration on the

performance of membrane reactor, steam-reforming syngases

with feed inlet CO concentration of 5% and 10% were also used

in the calculation. The feed-side CO mole fraction profiles are

illustrated in Fig. 15. The hollow-fiber numbers required for the

5% and 10% CO syngases were 31,000 and 31,500, respectively,

which were very similar to 31,000 for the 1% CO feed gas in the

reference case. As explained earlier, a higher CO amount in the

feed gas could require a larger reactor volume, but it also

generated more reaction heat in the adiabatic module to increase

the feed-side temperature. These balanced effects resulted in

similar membrane area requirements for these three syngases.

4.2.5. Effect of catalyst activity

Fig. 16 illustrates the effect of catalyst activity on the

required membrane area to satisfy the constraint of less than

10 ppm exit CO concentration. As illustrated in this figure, the

membrane area requirement decreased significantly with the

improved catalyst activity, a similar trend to the autothermal

reforming syngas. As explained earlier, a higher catalyst

activity increased the WGS reaction rate and enhanced the CO2

permeation because of a higher driving force. Therefore, the

higher catalyst activity resulted in a small amount of catalyst or

a small reactor size.

Fig. 16. The effect of catalyst activity on required membrane area for steam

reforming syngas.

J. Huang, W.S.W. Ho / Journal of the Chinese Institute of Chemical Engineers 39 (2008) 129–136136

5. Conclusions

Based on the model that we have developed and verified in

our earlier work, the effects of several important system

parameters on the performance of CO2-selective WGS

membrane reactor were studied. Both autothermal reforming

and steam reforming syngases showed similar trends with

respect to the five system parameters: inlet feed temperature,

inlet sweep temperature, feed-side pressure, WGS catalyst

activity, and CO inlet feed concentration. As the inlet feed

temperature increased, the membrane area requirement reduced

because of the higher WGS reaction rate. The increase of inlet

sweep temperature resulted in more significant reduction of the

required membrane area because the feed-side temperature was

shifted over a longer reactor length. However, overly high

temperatures of greater than about 160 8C were slightly

unfavorable to the exothermic and reversible WGS reaction.

Higher feed-side pressure decreased the required membrane

area as a result of the higher permeation driving force and

reaction rate. Increasing catalyst activity increased the WGS

reaction rate, enhanced the permeation driving force, and

decreased the membrane area requirement. The modeling

results have shown that a CO concentration of less than 10 ppm

is achievable from syngases containing up to 10% CO.

Acknowledgements

We would like to thank the National Science Foundation, the

Office of Naval Research via DJW Technology, LLC (grant no.

10020182), the US Department of Energy (grant/contract no.

DE-FC36AL68510), and The Ohio State University for the

financial support of this work. Part of this material is based

upon work supported by the National Science Foundation under

grant no. 0625758.

References

Acres, G. J. K., ‘‘Recent Advances in Fuel Cell Technology and its

Applications,’’ J. Power Sources, 100, 60 (2001).

Ahmed, S. and M. Krumpelt, ‘‘Hydrogen from Hydrocarbon Fuels for Fuel

Cells,’’ Int. J. Hydrogen Energy, 26, 291 (2001).

Amadeo, N. and M. Laborde, ‘‘Hydrogen Production from the Low Tempera-

ture Water–Gas Shift Reaction: Kinetics and Simulation of the Industrial

Reactor,’’ Int. J. Hydrogen Energy, 20, 949 (1995).

Andreeva, D., ‘‘Low Temperature Water Gas Shift over Gold Catalysts,’’ Gold

Bull., 35, 82 (2002).

Andreeva, D., V. Idakiev, T. Tabakova, L. Ilieva, P. Falaras, A. Bourlinos, and A.

Travlos, ‘‘Low-Temperature Water–Gas Shift Reaction over Au/CeO2

Catalysts,’’ Catal. Today, 72, 51 (2002).

Armor, J. N., ‘‘Applications of Catalytic Inorganic Membrane Reactors to

Refinery Products,’’ J. Membr. Sci., 147, 217 (1998).

Barnett, B. M. and W. P. Teagan, ‘‘The Role of Fuel Cells in our Energy

Future,’’ J. Power Sources, 37, 15 (1992).

Basile, A., A. Criscuoli, F. Santella, and E. Drioli, ‘‘Membrane Reactor for

Water Gas Shift Reaction,’’ Gas Sep. Purif., 10, 243 (1996).

Brown, L. F., ‘‘A Comparative Study of Fuels for On-Board Hydrogen

Production for Fuel-Cell-Powered Automobiles,’’ Int. J. Hydrogen Energy,

26, 381 (2001).

Campbell, J. S., ‘‘Influences of Catalyst Formulation and Poisoning on the

Activity and Die-off of Low Temperature Shift Catalysts,’’ Ind. Eng. Chem.

Proc. Des. Develop., 9, 588 (1977).

Choi, Y. and H. G. Stenger, ‘‘Water Gas Shift Reaction Kinetics and Reactor

Modeling for Fuel Cell Grade Hydrogen,’’ J. Power Sources, 124, 432

(2003).

Criscuoli, A., A. Basile, and E. Drioli, ‘‘An Analysis of the Performance of

Membrane Reactors for the Water-Gas Shift Reaction Using Gas Feed

Mixtures,’’ Catal. Today, 56, 53 (2000).

Cropper, M. A. J., S. Geiger, and D. M. Jollie, ‘‘Fuel Cells: A Survey of Current

Developments,’’ J. Power Sources, 131, 57 (2004).

Fiolitakis, E., U. Hoffmann, and H. Hoffmann, ‘‘Application of Wavefront

Analysis for Kinetic Investigation of Water–Gas Shift Reaction,’’ Chem.

Eng. Sci., 35, 1021 (1980).

Fu, Q., A. Weber, and M. Flytzani-Stephanopoulos, ‘‘Nanostructured Au–CeO2

Catalysts for Low-Temperature Water–Gas Shift,’’ Catal. Lett., 77, 87 (2001).

Fu, Q., H. Saltsburg, and M. Flytzani-Stephanopoulos, ‘‘Active Nonmetallic Au

and Pt Species on Ceria-Based Water–Gas Shift Catalysts,’’ Science, 301,

935 (2003).

Ghenciu, A. F., ‘‘Review of Fuel Processing Catalysts for Hydrogen Production

in PEM Fuel Cell Systems,’’ Curr. Opin. Solid State Mater. Sci., 6, 389

(2002).

Grenoble, D. C., M. M. Estadt, and D. F. Ollis, ‘‘Chemistry and Catalysis of the

Water Gas Shift Reaction. 1. The Kinetics over Supported Metal Catalysts,’’

J. Catal., 67, 90 (1981).

Hilaire, S., X. Wang, T. Luo, R. J. Gorte, and J. Wagner, ‘‘A Comparative Study

of Water–Gas-Shift Reaction over Ceria-Supported Metallic Catalysts,’’

Appl. Catal. A: Gen., 258, 271 (2004).

Ho, W.S.W., Sirkar, K.K. (Eds.), Membrane Handbook, Chapman and Hall,

New York, U.S.A. (1992).

Ho, W. S. W., ‘‘Membranes Comprising Aminoacid Salts in Polyamine Poly-

mers and Blends,’’ U.S. Patent, 6, 099, 621 (2000).

Huang, J., L. El-Azzami, and W. S. W. Ho, ‘‘Modeling Of CO2-Selective

Water–Gas-Shift Membrane Reactor for Fuel Cell,’’ J. Membr. Sci., 261, 67

(2005).

Huang, J., J. Zou, and W. S. W. Ho, ‘‘Carbon Dioxide Capture Using a CO2-

Selective Facilitated Transport Membrane,’’ Ind. Eng. Chem. Res. (in press).

Keiski, R. L., O. Desponds, Y. F. Chang, and G. A. Somorjai, ‘‘Kinetics of the

Water–Gas Shift Reaction over Several Alkane Activation and Water–Gas

Shift Catalysts,’’ Appl. Catal. A: Gen., 101, 317 (1993).

Ma, D. and C. R. F. Lund, ‘‘Assessing High-Temperature Water–Gas Shift

Membrane Reactors,’’ Ind. Eng. Chem. Res., 42, 711 (2003).

Mendelovici, L. and M. Steinberg, ‘‘Methanation and Water–Gas Shift Reac-

tion over Pt/CeO2,’’ J. Catal., 96, 285 (1985).

Moe, J., ‘‘Design of Water–Gas Shift Reactors,’’ Chem. Eng. Progr., 58 (3), 33

(1962).

Rhodes, C., G. J. Hutchings, and A. M. Ward, ‘‘Water–Gas Shift Reaction:

Finding the Mechanistic Boundary,’’ Catal. Today, 23, 43 (1995).

Salmi, T. and R. Hakkarainen, ‘‘Kinetic Study of Low-Temperature Water–Gas

Shift Reactor over a Copper–Zinc Oxide Catalyst,’’ Appl. Catal., 49, 285

(1989).

Swartz, S. L., M. M. Seabaugh, C. T. Holt, and W. J. Dawson, ‘‘Fuel Processing

Catalysts Based on Nanoscale Ceria,’’ Fuel Cell Bull., 4 (30), 7 (2001).

Tabakova, T., F. Boccuzzi, M. Manzoli, J. W. Sobczak, V. Idakiev, and D.

Andreeva, ‘‘Effect of Synthesis Procedure on the Low-Temperature WGS

Activity of Au/Ceria Catalysts,’’ Appl. Catal. B: Environ., 49, 73 (2004).

Tosti, S., A. Basile, G. Chiappetta, C. Rizzello, and V. Violante, ‘‘Pd–Ag

Membrane Reactors for Water Gas Shift Reaction,’’ Chem. Eng. J., 93, 23

(2003).

Uemiya, S. N., H. Sato, E. Ando, and Kikuchi, ‘‘The Water Gas Shift Reaction

Assisted by a Palladium Membrane Reactor,’’ Ind. Eng. Chem. Res., 30, 585

(1991).

Xue, E., M. O’Keeffe, and J. R. H. Ross, ‘‘Water–Gas Shift Conversion Using a

Feed with a Low Steam to Carbon Monoxide Ratio and Containing

Sulphur,’’ Catal. Today, 30, 107 (1996).

Zou, J. and W. S. W. Ho, ‘‘CO2-Selective Polymeric Membranes Containing

Amines in Crosslinked Poly(Vinyl Alcohol),’’ J. Membr. Sci., 286, 310

(2006).

Zou, J., J. Huang, and W. S. W. Ho, ‘‘CO2-Selective Water Gas Shift Membrane

Reactor for Fuel Cell Hydrogen Processing,’’ Ind. Eng. Chem. Res., 46,

2272 (2007).