International Journal of Hydrogen Energy 32 (2007) 4350 – 4357 www.elsevier.com/locate/ijhydene

High performance PEMFC stack with open-cathode at ambient pressure and temperature conditions D.T. Santa Rosa a , D.G. Pinto a , V.S. Silva a,∗ , R.A. Silva b , C.M. Rangel b a SRE—Soluções Racionais de Energia, S.A., Polígono Industrial do Alto do Ameal, Pav. C 13, 2565-641 Ramalhal, Portugal b INETI, Unidade de Electroquímica de Materiais, DMTP, Paço do Lumiar, 22, 1649-038 Lisboa, Portugal

Received 14 November 2006; received in revised form 31 May 2007; accepted 31 May 2007 Available online 7 August 2007

Abstract An open-air cathode proton exchange membrane fuel cell (PEMFC) was developed. This paper presents a study of the effect of several critical operating conditions on the performance of an 8-cell stack. The studied operating conditions such as cell temperature, air flow rate and hydrogen pressure and flow rate were varied in order to identify situations that could arise when the PEMFC stack is used in low-power portable PEMFC applications. The stack uses an air fan in the edge of the cathode manifolds, combining high stoichiometric oxidant supply and stack cooling purposes. In comparison with natural convection air-breathing stacks, the air dual-function approach brings higher stack performances, at the expense of having a lower use of the total stack power output. Although improving the electrochemical reactions kinetics and decreasing the polarization effects, the increase of the stack temperature lead to membrane excessive dehydration (loss of sorbed water), increasing the ohmic resistance of the stack (lower performance). The results show that the stack outputs a maximum power density of 310 mW/cm2 at 790 mA/cm2 when operating at ambient temperature, atmospheric air pressure, self-humidifying, air fan voltage at 5.0 V and 250 mbar hydrogen relative pressure. For the studied range of hydrogen relative pressure (150–750 mbar), it is found that the stack performance is practically not affected by this operation condition, although a slightly higher power output for 150 mbar was observed. On the other hand, it is found that the stack performance increases appreciably when operated with forced air convection instead of natural convection. Finally, the continuous fuel flow operation mode does not improve the stack performance in comparison with the hydrogen dead-end mode, in spite of being preferable to operate the stack with hydrogen flow rates above 0.20 l/min. 䉷 2007 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. Keywords: PEMFC stack; Open-air cathode manifold stack; Forced convection air-breathing cathode

1. Introduction The proton exchange membrane fuel cell (PEMFC) has been regarded as the most adequate system as power source of many portable electric devices and the best conventional battery system substitute [1]. The PEMFC could also be a sustainable alternative for the power generation in zero-emission automotive applications as well as for stationary power stations [2,3]. The membrane electrode assembly (MEA) is usually assumed as the core of the PEMFC stack for generating electrical and thermal energy [4–7], consisting of anode and cathode ∗ Corresponding author. Tel.: +351 261 910180; fax: +351 261 911246.

E-mail address: [email protected] (V.S. Silva).

catalyst layers, ion-exchange polymer membrane (PEM) and anode and cathode electrode backing/gas diffusion layers (GDL) [8]. The GDL provide high electrical and thermal conductivities, high corrosion resistance and reactants permeability [9]. The interaction between the diffusion layers and the catalyst layers, with a thin film like structure, is expected to enable an optimum water management, in order to guarantee the humidification of the proton exchange membrane and to prevent the mass transfer paths flooding [10]. In the development of open-cathode manifolds stack, one common problem is related to the optimum flow rate requirement for air to be used as coolant as well as an oxidant. The air flow rate and humidification should be such as to provide an adequate supply of oxygen to the MEA, while preventing

0360-3199/$ - see front matter 䉷 2007 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2007.05.042

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357

the dehydration of the membrane (excessive temperature or air flow rate) and flooding of the air mass transport paths (deficient air flow rate) [11,12]. The use of an air fan is expected to enable the MEA hydration level control, varying the air flow rate according to the stack operation conditions (power output, ambient air relative humidity and temperature). Another problem with fuel cell stacks using dual-function air supply is the need of an air filter. The oxidant stream should ideally be free of dust and other particulate matter that might contaminate the MEA or block the oxidant mass transport paths [13]. On the other hand, operating conditions such as stack temperature, although improving the electrochemical reactions kinetics and decreasing the polarization effects, may lead to the membrane excessive dehydration (loss of sorbed water), increasing the ohmic resistance of the stack [14,15]. It is well known that the proton exchange membrane proton conductivity depends to a large extent on the amount of sorbed water with the proton transport being influenced by it [16–18]. In contrast, excessive accumulation of water can be detrimental because it can flood the gas diffusion paths and block the reactants mass transfer (fuel and oxidant starvation) [19,20]. The PEMFC fuel cell stack, 10SR4-A, under study in the present paper, was developed in our laboratories and is commercially available from SRE. This stack consists in an open-air cathode type and is provided with the conventional configuration that comprises two end plates, nine bipolar plates, eight MEAs and gaskets. The stack uses a low-power consumption air fan in the edge of the cathode manifolds, for combined high stoichiometric oxidant supply and stack cooling purposes, avoiding the use of costly, large, heavy and power-consuming air pumps and external cooling systems [21,22]. In comparison with natural convection air-breathing stacks, the air dual-function approach presents higher stack performances, although having a lower use of the total stack power output [23]. The air fan provides improved means for cooling the stack, using air as a coolant fluid in order to absorb and carry away the generated heat. In this work, a study of several critical operation conditions effects was carried out on the performance of an open-air cathode stack developed at our laboratories. The operation conditions were varied in order to cover several situations that could arise when the stack is used in low-power portable PEMFC power applications [24–28]. 2. Experimental 2.1. Stack assembly In the present study a 10SR4-A stack developed in our laboratories and commercially available from SRE was used (Fig. 1). The 8-cells stack was assembled using MEAs purchased from 3 M. The active electrode area was 3.8 cm2 . The stack design is such that the hydrogen stream flows through the individual cells in series. The stack uses an air fan for combined oxidant supply (as air) and cooling (Fig. 1b). Material for own design graphite bipolar plates were purchased from Schunk. Current collector plates were made of gold plated brass.

4351

Fig. 1. Open-cathode stack, 10SR4-A, used in the present study: (a) front side and (b) back side.

2.2. Characterization method for the stack performance The fuel cell was characterized implementing load duty cycles comprising open circuit, constant current and polarization curves in a step-by-step procedure. First, it was measured the open circuit voltage during 5 min, followed by the application of a constant current of 1.5 A during 10 min. The polarization curves were then measured by scanning the current ranging from 0 to 3.5 A. Every step was applied during 2 min and, at the end of this period, the steady-state individual cell voltage and power values were taken. Previously to the measurements, the stack was fully activated, implementing 10 of the above mentioned cycles. Furthermore, each of the operation conditions study was performed with three cycles. The steadystate results obtained in the last cycle are presented in this paper. The average error obtained using this procedure was 3.8% (t distribution). The temperatures at the inlet and outlet of the cathode air were also measured. Except for the hydrogen flow rate experiments, the stack was always operated in hydrogen

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357

dead-end mode. The air fan of the stack was typically used at 5.0 V, except when mentioned otherwise. The air fan power consumption was 350, 850 and 1610 mW for 3.5, 5.0 and 7.0 V voltage input, respectively. On the other hand, the air fan average flow rate was 2.74, 5.00 and 6.72 l/min for 3.5, 5.0 and 7.0 V voltage input, respectively. The experimental results were obtained using dry hydrogen with of 99.9% purity. 3. Results and discussion

400

6

300

4

200

2

100

0 0

200

400

Current Density

0.8

Cell Voltage (V)

8

Power Density (mW/cm2)

Voltage (V)

Fig. 2 shows a typical polarization curve obtained using the 10SR4-A stack. Data were obtained at 250 mbar H2 relative pressure, room temperature in a hydrogen outlet dead-end operation mode and with the air fan working at 5.0 V. From Fig. 2, it can be observed that the stack was able to output a maximum power density of 310 mW/cm2 at 790 mA/cm2 (9.4 W at 3.5 A).

0.6

0.4 150 mbar

750 mbar

0

0 800

600

250 mbar

0.2

500 mbar

0

1

(mA/cm2)

Fig. 2. Current density–voltage and power density plots of the 10SR4-A stack (250 mbar hydrogen relative pressure, hydrogen dead-end mode, room temperature, air fan supply at 5.0 V).

Voltage (V)

Fig. 3 shows the polarization curves as a function of the hydrogen inlet relative pressure. As shown, it can be observed no clear variation of the stack performance for the hydrogen relative pressures of 250, 500 and 750 mbar (variation in the same order of the experimental error). In contrast, the lowest experimented hydrogen relative pressure, 150 mbar, exhibits a slight increase in performance in comparison to that of the other experimented values (variation higher than the experimental error). For the 150 mbar hydrogen relative pressure, it was obtained the power density of 333 mW/cm2 at 790 mA/cm2 . A possible interpretation of this result is the improved water back diffusion from the cathode, through the membrane, towards the anode, for the lower hydrogen relative pressure (lower ohmic resistance due to higher hydration of the PEM). Detailed individual cells voltage at a constant current of 1.5 A is shown in Fig. 4. Clearly, it can be seen that the performance of the first cell is lower than that of the others, for the studied

2

3

4 5 Cell Number

7

8

9

Fig. 4. Individual cells voltage at the constant current of 1.5 A as a function of the hydrogen relative pressure (hydrogen dead-end mode, room temperature, air fan supply at 5.0 V).

8

400

6

300

4

200

150 mbar

2

6

100

250 mbar

Power Density (mW/cm2)

4352

500 mbar 750 mbar

0 0

200

400

600

0 800

Current Density (mA/cm2) Fig. 3. Current density–voltage and power density plots as a function of the hydrogen relative pressure obtained using a 10SR4-A stack (hydrogen dead-end mode, room temperature and air fan supply at 5.0 V).

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357

400

A

Voltage (V)

6

300 B

4

200 C

2

100

0.0 V 3.5 V 5.0 V 7.5 V

0 0

200

400

Power Density (mW/cm2)

8

4353

0 800

600

Current Density (mA/cm2) Stage C ≡ 790 mA/cm2

Stage B ≡ 395 mA/cm2 8

7

7

6

6

5 4 3

8

7.5 V 5.0 V 3.5 V 0.0 V

7 Cell Number

8

Cell Number

Cell Number

Stage A ≡ OCV

5 4 3

6 5 4 3

2

2

2

1

1

1

0

0.2 0.4 0.6 0.8 Cell Voltage(V)

1

0

0.2 0.4 0.6 0.8 Cell Voltage (V)

1

0

0.2 0.4 0.6 0.8 Cell Voltage (V)

1

Fig. 5. Current density–voltage and power density plots as a function of the air fan supply voltage (250 mbar hydrogen relative pressure, hydrogen dead-end mode, room temperature).

50 Stack Temperature (°C)

range of hydrogen relative pressures. We believe that this result is associated to the first cell lower humidification conditions due to dry hydrogen feed. This fact is not so severe for the other cells because water is being generated at 1.5 A and dragged from cell to cell due to the forced hydrogen flux by the electrochemical reaction (individual cells associated in series). As the fuel cell stack tested in this work was an open-cathode air fan type, it is convenient to investigate how the air flow rate affects its performance. From Fig. 5 it can be observed that the use of the air fan (voltage at 3.5, 5.0 and 7.0 V) increases appreciably the stack performance in comparison with the natural air convection approach. Although having the disadvantage of increasing the parasitic power consumption, the use of the dual-function air fan enables a maximum stack power output of 9.7 W (fan working at 5 V) in comparison with 2.0 W (without fan). When comparing Fig. 5 curves, corresponding to different fan voltages, it can be observed that the stack performances are similar for current densities lower than 660 mA/cm2 .We believe that this fact is due to the excessive oxygen feed stoichiometry (from 8.3 to 20.3 for 3.5–7.5 V air fan voltage supply at 660 mA/cm2 ) and enough stack cooling. However, it can be noticed that when using the air fan at 3.5 V, the stack was not able to output current densities higher than 660 mA/cm2 . The main reason for this limitation is the excessive stack temperature that leads to the dehydration of the membrane

40 30 20 10

3.5 V 5.0 V 7.5 V

0 0

200

400

Current Density

600

800

(mA/cm2)

Fig. 6. Cathode outlet air temperature as a function of the stack current density and air fan supply voltages (250 mbar hydrogen relative pressure, hydrogen dead-end mode, room temperature).

(higher ohmic losses). The detrimental effect of this phenomenon in the stack performance is more pronounced because the proton exchange membrane is very thin, Nafion䉸 111 equivalent, and therefore more sensitive to water management conditions. The cathode outlet air temperature as a function of the stack current density can be observed in Fig. 6. It can be

4354

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357

1 0.9 0.8 0.7

Cell Voltage (V)

0.9-1 0.8-0.9 0.7-0.8 0.6-0.7 0.5-0.6 0.4-0.5 0.3-0.4 0.2-0.3 0.1-0.2 0-0.1

0.6 0.5 0.4 0.3 0.2 0.1 0

2

0 Cu 26 rre nt D 53 79 ens ity (mA 105 /cm 2 )

1

3 4

6 7 132 158

er 5 mb Nu l l e

C

8

Fig. 7. Individual cells voltage as a function of the stack current density obtained using a 10SR4-A stack operated under air-breathing conditions (250 mbar hydrogen relative pressure, hydrogen dead-end mode, room temperature).

280

Voltage (V)

6 210 4 140 2 30°C

70

Power Density (mW/cm2)

350

8

35°C 40°C

0 0

200

400

600

0 800

Current Density (mA/cm2) Fig. 8. Current density–voltage and power density plots as a function of the ambient temperature (250 mbar hydrogen relative pressure, hydrogen dead-end mode, air fan supply at 5.0 V).

seen that the cathode outlet air temperature increases with the stack current density. It is observed that for higher power output regimes, the stack temperature increases above 40 ◦ C for the lowest experimented fan voltage (3.5 V) due to deficient cooling. When using the air fan at 5.0 and 7.5 V, the effect on the stack temperature is not very significant, as Fig. 6 demonstrates. Therefore, the air fan at 5.0 V seems to be the better

option because it ensures enough stack cooling and oxidant supply and prevents the detrimental excessive dehydration of the MEAs when using the air fan at 7.5 V. Fig. 5 also puts in evidence, the individual cells voltage at three selected points of the curve as a function of the air fan voltage. Stage A corresponds to the open circuit conditions and it can be observed that the fan voltage has almost no influence, except in the

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357

60 Stack Temperature (°C)

case of natural air-breathing, where all the cells but the last, present approximately the same voltage. On the other hand, in stage B, corresponding to a current density of 395 mA/cm2 , it can be observed that the first cell starts to be limiting, having a voltage loss higher than that observed for the other cells with individual voltages slightly higher for the experiments performed with the air fan working at 3.5 V (mainly cells 6, 7 and 8). We believe that this fact is due to improved humidification of the MEA when using the air fan voltage at 3.5 V (lower ohmic losses). However, as mentioned before, lower air fan flow rate leads to excessive heating of the stack for higher current densities (Fig. 6). Therefore, at stage C, corresponding to the current density of 790 mA/cm2 , the stack operating with the air fan working at 3.5 V was not able to reach this performance. From the plots corresponding to stage C, it can be also observed that the stack individual cells had improved performance when the air fan was operated at 5.0 V, except for cells 1 and 2. This observation is related with the water balance management in the stack, as explained above. In the case of natural air convection, it is clear that it is not possible to reach the stages B and C, because the cells tends to heat faster at lower current densities due to the obvious limitations in the amount of air for expel the reaction generated heat. Fig. 7 shows the individual cell voltage as a function of the current density for the stack operating under natural airbreathing conditions. It can be observed that the individual cell voltage decreases as the number of cell and current density increases. For example, the last cell voltage decreases from 0.4 V at open circuit to practically null voltage at 79 mA/cm2 . For cells ranging from 4 to 7, it is noticed an abrupt voltage drop for current densities higher than 132 mA/cm2 . Cells number 1–3 present an individual voltage higher than 0.6 V, for all the current density experimented. A possible interpretation of the results is the excessive increase of the stack temperature, which is caused by the absence of heat removal in the cells. As it can be observed, the first cells voltage (1, 2 and 3) does not decrease as that of the others. This result is due to having the feed of hydrogen at ambient temperature that prevents the excessive heating of the first cells. As we move along the stack, the hydrogen stream temperature increases and the individual cell performance starts to decrease due to excessive drying of the MEAs. On the other hand, an undersupply of oxygen in the last cells for higher current densities is also a possible interpretation of the results. Fig. 8 shows the stack performance as a function of the ambient temperature. It can be observed that the highest power density is obtained for the temperature of 30 ◦ C. It can be noticed that the stack performance in terms of maximum power density decreases with the ambient temperature. However, it should be mentioned that for current densities lower than 660 mA/cm2 , the highest performance is obtained for the ambient temperature of 35 ◦ C. This is due to improved electrochemical reaction kinetics, mass transport and polarization losses for higher operation temperatures. According to this, as the ambient temperature increases, more power output should be obtained from the stack. However, the same trend is expected for the generated heat that could also affect the stack performance. For higher

4355

50 40 30 20 30°C 35°C 40°C

10 0 0

200

400

Current Density (mA/cm

600

800

2)

Fig. 9. Cathode outlet air temperature as a function of the stack current density and ambient temperature (250 mbar hydrogen relative pressure, hydrogen dead-end mode, air fan supply at 5.0 V).

ambient temperatures, the generated reaction heat starts to dehydrate the PEM and, consequently, increases the ohmic losses (lower performance). As shown in Fig. 9, the air cathode outlet temperature for both 35 and 40 ◦ C ambient temperature exceed 50 ◦ C for higher current densities than 600 mA/cm2 , which is already very close to the limit temperature supported by this self-humidifying stack. It is referred as operation limit that the stack air cathode outlet temperature is of 60 ◦ C, because at this point the MEA gets too dry, which implicates higher ohmic resistance. As seen in Fig. 9, stack air cathode outlet temperature at a 30 ◦ C ambient, attains the 50 ◦ C limit only after the current density of 660 mA/cm2 . In the previous experiments, the fuel cell stack was operated in a hydrogen dead-end supply mode. In order to study the effect of the hydrogen flow in the stack performance, the 8-cells stack was operated with a continuous hydrogen flow. From Fig. 10, it can be observed that for the hydrogen flow rate of 0.15 l/min, the maximal current density is of 530 mA/cm2 due to hydrogen stoichiometric limitations. For the current density of 660 mA/cm2 , the stack would need the stoichiometric hydrogen flow rate higher than 0.15 l/min. The maximum power density output under open-end mode was obtained for the hydrogen flow rate of 0.20 l/min, 290 mW/cm2 at 790 mA/cm2 . On the other hand, the worst performance was obtained for the highest experimented hydrogen flow rate, 0.50 l/min. Also, the stack performance under open-end mode was always lower than that under dead-end mode, for all the studied hydrogen flow rate range. However, it should be mentioned that for the current density 530 mA/cm2 , the highest stack performance was obtained for the continuous hydrogen flow rate of 0.15 l/min. A possible interpretation for the lower performance of the stack under open-end hydrogen flow mode is the higher imposed dehydration of the PEM (dry hydrogen used as feed). Since the stack is operated under self-humidifying conditions, the continuous feed of dry hydrogen dehydrates the proton exchange membrane and, consequently, increases the voltage losses (higher ohmic resistance). In contrast, when operated under dead-end mode, the reaction imposed hydrogen flow will be continuously humidified from cell to cell and an improved water management is achieved.

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357

350

8

280

Voltage (V)

6 210 4 140 0.15 l/min 0.20 l/min 0.30 l/min 0.40 l/min 0.50 l/min dead-end

2

70

0 0

200

400

Power Density (mW/cm2)

4356

600

0 800

Current Density (mA/cm2) Fig. 10. Current density–voltage and power density plots as a function of the hydrogen flow rate. (250 mbar hydrogen relative pressure, air fan supply at 5.0 V, room temperature).

4. Conclusions An open-air cathode PEMFC stack with eight cells and an active reaction area of 3.8 cm2 was developed and studied at several critical operation conditions. Stack performance increases appreciably when operated with forced air convection instead of natural convection. When using an air fan for combined oxidant supply and stack cooling, the stack performance seems to have an optimum for the air fan voltage supply at 5.0 V. The maximum power obtained using the studied stack was 9.4 W (310 mW/cm2 ), at 250 mbar H2 relative pressure, room temperature, self-humidifying, hydrogen dead-end mode and with the air fan working at 5.0 V. The best stack performance was achieved at 150 mbar, with a maximum power of 10.1 W (333 mW/cm2 ). In contrast, lower performance variations were observed for the other hydrogen relative pressures. Obtained results showed that the stack performance is higher for temperatures as low as 30 ◦ C. Finally, the stack performance is higher when operated under dead-end hydrogen flow. For open-end mode hydrogen flow, the best performance was obtained for the lowest hydrogen flow rate used. References [1] Mao ZQ. Fuel cells. Beijing: Chemical Industry Press; 2005. [2] Ahn S-Y, Shin S-J, Ha H-Y, Hong S-A, Lee Y-C, Lim T-W. et al. Performance and lifetime analysis of the kW-class PEMFC stack. J Power Sources 2002;106:295. [3] Hottinen T, Mikkola M, Lund P. Evaluation of planar free-breathing polymer electrolyte membrane fuel cell design. J Power Sources 2004;129:68. [4] Qi Z, Kaufman A. Quick and effective activation of proton-exchange membrane fuel cells. J Power Sources 2003;114:21. [5] Zhu H, Kee R. A general mathematical model for analyzing the performance of fuel cell membrane electrode assemblies. J Power Sources 2003;117:61.

[6] Acosta M, Merten C, Eigenberger G, Class H, Helmig R, Thoben B. et al. Modeling non-isothermal two-phase multicomponent flow in the cathode of PEM fuel cells. J Power Sources 2006;159:1123. [7] Whishart J, Dong Z, Secanell M. Optimization of a PEM fuel cell system based on empirical data and generalized electrochemical semi-empirical model. J Power Sources 2006;161:1041. [8] Silva VS, Mendes A, Madeira LM, Nunes S. Membranes for direct methanol fuel cell applications: analysis based on characterization, experimentation and modeling. In: Zhang X, editors. Advances in fuel cells. 2005. p. 57–80. [9] Larminie J, Dicks A. Fuel cell system explained. 2nd ed., UK: Wiley; 2003. [10] Watanable M, Uchida H, Seki Y, Emori M. Self-humidifying polymer electrolyte membranes for fuel cells. J Electrochem Soc 1996;143. [11] Dong Z, Shen J. Solid cage fuel cell stack. Patent US 6 720 101 B1; 2004. [12] Leahy S. Active flow control of lab-scale solid polymer electrolyte fuel cells. Master thesis, Georgia Institute of Technology; 2004. [13] Yoon Y-G, Lee W-Y, Yang T-H, Park G-G, Kim C-S. Current distribution in a single cell of PEMFC. J Power Sources 2003;118:193. [14] Zawodzinski T, Springer T, Uribe F, Gottersfeld S. Characterization of polymer electrolytes for fuel cell applications. Solid State Ionics 1993;60:199. [15] Santarelli M, Torchio M. Experimental analysis of the effects of the operating variables on the performance of a single PEMFC. Energy Convers Manage 2007;48:40. [16] Fowler M, Mann R, Amphlett J, Peppley B, Roberge P. Incorporation of voltage degradation into a generalized steady state electrochemical model for PEM fuel cell. J Power Sources 2002;106:274. [17] Su A, Chiu Y, Weng F. The impact of flow field pattern on the concentration and performance in PEMFC. Int J Energy Res 2005;29:409. [18] Goldpaygan A, Ashgriz N. Effect of oxidant properties on the mobility of water droplets in the channels of the PEM fuel cell. Int J Energy Res 2005;29:1027. [19] Chen K, Hickner M, Noble D. Simplified models for predicting the onset of liquid water droplet instability at the gas diffusion layer/gas flow channel interface. Int J Energy Res 2005;29:1113. [20] Ferng Y, Sun C, Su A. Numerical simulation of thermal-hydraulic characteristics in a proton exchange membrane fuel cell. Int J Energy Res 2003;27:495.

D.T. Santa Rosa et al. / International Journal of Hydrogen Energy 32 (2007) 4350 – 4357 [21] Rodatz P, Onder C, Guzzella L. Air supply system of a PEMFC stack dynamic model. Fuel Cells 2005;1:5. [22] Baoa C, Ouyanga M, Yib B. Modelling and control of air stream and hydrogen flow with recirculation in a PEM fuel cell system. Int J Hydrogen Energy 2006;31:1879. [23] Noponen M, Mennola T, Mikkola M, Hottinen T, Luna P. Measurement of current distribution in a free-breathing PEMFC. J Power Sources 2002;106:304. [24] Fabian T, Posner J, O’Hayre R, Cha S-W, Eaton J, Prinz F. et al. The role of ambient conditions on the performance of a planar, air-breathing hydrogen PEM fuel cell. J Power Source 2006;161:168.

4357

[25] Chu D, Jiang R. Performance of polymer electrolyte membrane fuel cell stacks, part I. Evaluation and simulation of an air-breathing PEMFC stack. J Power Sources 1999;83:128. [26] Sohn Y-J, Park G-G, Yang T-H, Yoon Y-G, Lee W-Y, Yim S-D. et al. Operating characteristic of an air-cooling PEMFC for portable applications. J Power Sources 2005;145:604. [27] Liu Z, Mao Z, Wang C, Zhuge W, Zhang Y. Numerical simulation of a mini PEMFC stack. J Power Sources 2006;160:1111. [28] Knights S, Colbow KM, St-Pierre J, Wilkinson DP. Aging mechanisms and lifetime of PEFC and DMFC. J Power Sources 2004; 127:127.