A novel electrode architecture for passive direct methanol fuel cells

Transcription

1 Electrochemistry Communications 9 (27) A novel electrode architecture for passive direct methanol fuel cells R. Chen, T.S. Zhao * Department of Mechanical Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong SAR, China Received 13 October 26; received in revised form 4 November 26; accepted 6 November 26 Abstract The supply of cathode reactants in a passive direct methanol fuel cell (DMFC) relies on naturally breathing oxygen from ambient air. The successful operation of this type of passive fuel cell requires the overall mass transfer resistance of oxygen through the layered fuel cell structure to be minimized such that the voltage loss due to the oxygen concentration polarization can be reduced. In this work, we propose a new membrane electrode assembly (MEA), in which the conventional cathode gas diffusion layer (GDL) is eliminated while utilizing a porous metal structure for transporting oxygen and collecting current. We show theoretically that the new MEA enables a higher mass transfer rate of oxygen and thus better performance. The measured polarization and constant-current discharging behavior showed that the passive DMFC with the new MEA yielded better and much more stable performance than did the cell having the conventional MEA. The EIS spectrum analysis further demonstrated that the improved performance with the new MEA was attributed to the enhanced transport of oxygen as a result of the reduced mass transfer resistance in the fuel cell system. Ó 26 Elsevier B.V. All rights reserved. Keywords: Fuel cell; Passive DMFC; Metal foam; Mass transfer resistance; Cell performance; Oxygen transport 1. Introduction The direct methanol fuel cell (DMFC) has been recognized as the most promising power source for portable electronic devices because this type of fuel cell offers the advantages of higher energy density as a result of the use of liquid methanol, as well as simpler and more compact structure. Over the past decade, extensive effort [1 8] has been made to the study of the active DMFC with the fuel fed by a liquid pump and oxidant supplied by a gas compressor. Nevertheless, these auxiliary devices not only make the fuel cell system complex but also decrease the achievable energy density and power density due to the parasitic power losses. It may not be practical to operate the DMFC under such conditions for powering portable electronic devices. Therefore, as a potential portable power source, it is essential to eliminate some auxiliary devices in order to decrease the volume and weight and to increase the energy efficiency of the DMFC. For this reason, * Corresponding author. Tel.: ; fax: address: metzhao@ust.hk (T.S. Zhao). various DMFC systems that operate under the passive conditions, i.e., air breathing and passive methanol solution supply, have been proposed and studied [9 2]. This type of passive DMFC not only offers the advantage of simple and compact systems but also makes it possible to eliminate the parasitic power losses for powering ancillary devices required in the active DMFC. Because of these advantages, the passive DMFC has received much more attention in the area of small fuel cells. Chu and Jiang [1] investigated the effect of operating conditions on the performance and energy efficiency of a small passive DMFC. Liu et al. [11] studied sintered stainless steel fiber felt as the gas diffusion layer in an air-breathing DMFC. The effect of methanol concentration was also studied in this work. Bae et al. [12] investigated the effect of methanol concentration, catalyst loading, fuel and oxidant supply modes on the performance of a passive DMFC. Shimizu et al. [13] reported their activities regarding the research and development of DMFCs that operated passively at room temperature. Recently, Liu and Zhao et al. [18,19] found that the main reason for a higher methanol concentration leading to the improvement of the performance is /$ - see front matter Ó 26 Elsevier B.V. All rights reserved. doi:1.116/j.elecom

2 R. Chen, T.S. Zhao / Electrochemistry Communications 9 (27) because of the increased methanol permeation rate, which increases the operating temperature and thus improves the electrochemical kinetics of both methanol oxidation and oxygen reduction reactions. Previous studies showed that the passive DMFC usually has to be operated at relatively higher methanol concentration, as diffusion is the main mechanism of methanol transport from a built-in fuel reservoir to the anode catalyst layer. Moreover, the higher methanol concentration can increase the energy density of the fuel cell system, which is desired in passive DMFCs. However, at present, the increase in methanol concentration is limited by the problem of methanol crossover. Therefore, there is plenty of room up to pure methanol, meaning that methanol transport on the anode is actually not a problem. In contrast, the transport of oxygen on the cathode of the passive DMFC is a challenging problem, because the supply of oxygen in this type of fuel cell relies on naturally breathing oxygen from ambient air. As a result, the passive DMFC often operates under oxygen-starving and water-flooding conditions. With the constraint without any external means of air movement, it is critical to design and optimize the cathode architecture of the passive DMFC to ensure a higher oxygen transfer and water removal rate. Typically, the cathode of the conventional MEA, as depicted in Fig. 1a, is composed of a cathode catalyst layer and a gas diffusion layer (GDL) made of carbon paper or carbon cloth with a coated micro porous layer (MPL). Oxygen is transported to the catalyst layer via the cathode current collector and the GDL from ambient air, in which oxygen reacts with the migrated protons and the incoming electrons or directly reacts with the permeated methanol to produce water and heat. The generated water is then transported backward to ambient. As a result, the mass transport in the GDL causes the main mass transfer resistance on the cathode. It is essential to reduce the mass transfer resistance in the GDL so as to enhance the oxygen transport and water removal on the air-breathing cathode and thus improve the cell performance. Our previous work [2] showed that the porous current collector for the passive DMFC can dramatically enhance the oxygen transport and water removal rate compared with the conventional perforated-plate current collector. Hence, in this work, we propose a new membrane electrode assembly (MEA), in which the conventional cathode GDL is eliminated while utilizing a porous metal structure for transporting oxygen and collecting current. Both theoretical and experimental results show that this type of passive DMFC can not only provide a higher oxygen transfer rate but also a more effective water removal. As a consequence, the cell with the new MEA yielded higher performance and more stable operation than did the cell with the conventional MEA. 2. Theoretical Consider the transport process of oxygen in the conventional MEA, as shown in Fig. 1a, in which oxygen is a Methanol solution reservoir b Methanol solution reservoir transported to the catalyst layer through the current collector and the GDL from ambient air. The oxygen flux N O2 to the catalyst layer can be expressed as: N O2 ¼ C1 C CL 1 h þ lccc D eff;ccc C CH 3 OH+ C CH 3 OH+ 1-Anode current collector 3-Anode catalyst layer 5-Cathode catalyst layer 7-Cathode current collector þ l gdl D eff;gdl where C 1 and C CL represent the oxygen concentration in the ambient air and catalyst layer; D eff;ccc, D eff;gdl and h denote the effective diffusivity of oxygen in the current collector and GDL, the mass transfer coefficient at the current collector surface, respectively. In the cathode catalyst layer, oxygen is not only electrochemically reduced but also directly reacts with the permeated methanol. As a result, the oxygen flux can be related to the current by Faraday s law: N O2 ¼ i þ i p ð2þ 4F where i is the current density, i p is the parasitic current density corresponding to the flux of methanol crossover N cross- l gdl l ccc h h 2-Anode gas diffusion layer 4-Proton exchange membrane 6-Cathode gas diffusion layer Fig. 1. Schematic of the passive DMFC with (a) conventional MEA and (b) new MEA. ð1þ

3 72 R. Chen, T.S. Zhao / Electrochemistry Communications 9 (27) over, i.e., i p =6FN crossover. Tafel equation with considering the effect of methanol crossover is employed to describe the cathodic electrochemical kinetics as: C CL i þ i p ¼ i ref exp af C ref RT g c ð3þ where i ref is the exchange current density of oxygen, C ref is the reference concentration of oxygen and g c is the cathode overpotential. Combining Eqs. (1) (3), we can obtain: g c ¼ RT af ln i ref C 1 iþip 4F i þ i p C ref O 2 1 h þ lccc D eff;ccc þ l cdl D eff;gdl 1 C A Eq. (4) indicates that the cathode overpotential increases with the mass transfer resistance in each layer, including 1 h at the current collector surface, in the GDL, and l ccc D eff;ccc l gdl D eff;gdl ð4þ in the current collector. Apparently, eliminating the GDL will result in a lower overall mass transfer resistance, thus lowering the cathode overpotential. In line with this idea, we propose a new design of MEA, as shown in Fig. 1b, in which the conventional GDL is eliminated while utilizing a porous metal structure for transporting oxygen and collecting current. As will be demonstrated experimentally in subsequent sections, this new architecture allows for a higher mass transfer rate of oxygen and thus better performance. 3. Experimental 3.1. Membrane electrode assembly The membrane electrode assembly (MEA) having an active area of 2. cm 2. cm was fabricated in-house employing a Nafion 115 membrane and two electrodes. The employed Nafion 115 membrane with a thickness of 125 lm was pretreated in this work. The pretreatment procedures included boiling the membrane in 5 vol.% H 2, washing in DI water, boiling in.5 M H 2 SO 4 and washing in DI water for 1 h in turn. The pretreated membranes were kept in the DI water prior to the fabrication of MEAs. A single-side ELAT electrode from ETEK was used in the anode, where carbon cloth (E-TEK, Type A) was used as the backing support layer with 3 wt% PTFE wet-proofing treatment. The catalyst loading on the anode side was 4. mg/cm 2 with 8 wt% PtRu (1:1 a/o) on optimized carbon. Furthermore,.8 mg/cm 2 dry Nafion Ò ionomer was coated onto the surface of the electrode. On the cathode, the catalyst layer was fabricated in-house by the decal method [21]. The well-mixed catalyst ink was sprayed onto the Teflon blank to form a catalyst layer. The catalyst layer was then transferred onto the membrane by hot pressing the catalyst coated Teflon blank and the anode electrode on the each side of the membrane at 135 C and 4 MPa for 3 min. The cathode catalyst loading was about 2.3 mg/cm 2 using 6 wt% Pt on Vulcan XC-72 with 15 wt% Nafion as the bonding agent. Carbon paper TGPH-9 with 2 wt% PTFE, on which a MPL was coated, was put on the cathode as the GDL. The decal method described above can ensure that the passive DMFCs with the new and conventional MEAs be investigated for the same cathode catalyst layer, membrane and the anode Single cell fixture The MEA mentioned above was sandwiched between an anode and a cathode current collector. The entire cell setup was then held together between an anode and a cathode fixture, both of which were made of transparent organic glass. A 5.-mL methanol solution reservoir was built in the anode fixture. Methanol was transferred into the catalyst layer from the built-in reservoir, while oxygen, from the surrounding air, was transferred into the cathode catalyst layer through the opening of the cathode fixture. The cell temperature was measured by a thermocouple (Type T), which was installed on the outer surface of the anode gas diffusion layer. The anode current collector was made of a perforated 316L stainless steel plate of 1.5 mm in thickness. A plurality of 2.6-mm circular holes was drilled in the anode current collector, serving as the passages of fuel, which resulted in an open ratio of 47.8%. A 2-nm platinum layer was sputtered onto the surface of the anode current collector to reduce the contact resistance with the electrode. Porous current collector for the cells with both the new and conventional MEAs was fabricated and tested, as shown in Fig. 2. The porous current collector was made Fig. 2. The current collector made of metal foam.

4 R. Chen, T.S. Zhao / Electrochemistry Communications 9 (27) of a Ni Cr alloy metal foam plate of 1. mm in thickness. The Ni Cr alloy metal foam supplied by the RECEMAT Ò International offers over 95% porosity and the estimated average pore diameter of.4 mm. A 2-nm gold layer was sputtered onto the surface of the porous current collectors to reduce the contact resistance Electrochemical instrumentation and test conditions An Arbin BT2 electrical load interfaced to a computer was employed to control the condition of discharging and record the voltage current curves. The temperature of the cell was measured by the Arbin BT2 built-in function. All the experiments of the passive DMFCs were performed at room temperatures ranging from 18.3 to 18.6 C and the relative humidity of 68 73%. Prior to the performance test, the MEA was installed in an active cell fixture and activated at 7 C about 24 h. During the activation period, 1. M methanol was fed at 1. ml min 1, while oxygen was supplied under atmospheric pressure at a flow rate of 5 ml min Results and discussion Prior to the polarization and constant-current discharging tests, we measured the internal cell resistances of the passive DMFCs with the new and conventional MEAs at 2. M and 4. M methanol solution. The results are presented in Table 1. It is seen from this table that the internal cell resistance with the new MEA is higher than that with the conventional MEA. The increased internal cell resistance was attributed to the poor direct contact between the cathode catalyst layer and the porous current collector, due primarily to the gap in pore size between the catalyst layer and the porous metal foam. It is also found that the measured internal cell resistance at 4. M is lower than that at 2. M as a result of the higher operating temperature enhancing the proton transport in the Nafion membrane. The cell performance of the passive DMFC with 4. M methanol concentration is shown in Fig. 3. It can be found from Fig. 3 that the new MEA exhibited slightly higher voltages at zero and low current densities than did the conventional MEA; with increasing current density, the increment of the cell performance became significantly larger. This is because the demand of oxygen on the cathode is Table 1 The measured internal cell resistances and temperatures Methanol concentration (M) Internal cell resistance (mohm cm 2 ) New MEA Conventional MEA Cell temperature ( C) Cell voltage (V) with IR correction with IR correction Current density (ma/cm 2 ) Fig. 3. Comparison in cell performance between the new and conventional MEAs. Anode: 4. M methanol; room temperature: 18.4 C; relative humidity: 69%. larger due to the higher rate of methanol crossover at the higher methanol concentration, which consumes the additional oxygen. The cell with the conventional MEA may not provide the sufficient oxygen on the cathode under this situation. However, this is not the case for the cell with the new MEA. The increased oxygen transfer rate as a result of the lowered overall mass transfer resistance improves the electrochemical kinetics of oxygen reduction and thus yields a slightly higher OCV and higher voltages at low current densities. As the current density increases, the demand of oxygen on the cathode increases. Although the cell with the new MEA has a higher internal cell resistance, the improved electrochemical kinetics as a result of the increased oxygen transfer rate not only compensated the decreased voltage due to higher internal cell resistance but also yielded higher voltages than did the cell with the conventional MEA. We also made IR corrections to the measured voltages, which are also shown in Fig. 3. It is seen that the cell with the new MEA yielded much better performance than did the cell with the conventional MEA as a result of the increased oxygen transfer rate. Additionally, the cell operating temperature was measured and compared in Fig. 4. It is seen that the cell operating temperature is almost the same for the both MEAs. This fact further confirms that the improved performance is mainly caused by the enhanced oxygen transfer rate with the new MEA. In summary, the lowered overall mass transfer resistance of the cell with the new MEA is the major reason that yielded the better performance at high methanol concentration. To investigate the operation stability of the passive DMFC with the new MEA, we also performed the longterm operation tests. It should be noted that for the passive DMFC the methanol consumption due to the electrochemical reaction and methanol crossover will cause a decrease in methanol concentration in the fuel reservoir. Therefore, the discharging behavior for the both MEAs with the Power density (mw/cm 2 )

5 722 R. Chen, T.S. Zhao / Electrochemistry Communications 9 (27) Cell temperature ( o C) Time (min) Fig. 4. Variation in cell operating temperature. Anode: 4. M methanol; room temperature: 18.4 C; relative humidity: 69%. passive operation will actually be caused by the combined effect of the time-dependent methanol concentration, oxygen transport and cell operating temperature. In order to demonstrate the improved performance of the new MEA was due solely to the enhanced mass transport of oxygen instead of the other factors, we did the long-term test for the same DMFC with methanol solution supplied by a pump while keeping the cathode at the air-breathing operation mode and maintaining a fixed cell operating temperature. To this end, the anode current collector for the passive operation mode was replaced by a current collector made of a 316L stainless steel plate with a thickness of 1. mm, in which a single serpentine flow field, consisting of a flow channel with a cross sectional area of mm 2, was formed. Similarly, to reduce the contact resistance between the anode current collector and electrode, a 2-nm platinum layer was sputtered onto the surface of the anode current collector. The methanol solutions of 2. and 4. M were supplied with the flow rate controlled by a digital HPLC micro-pump (Series III). To ensure that all the experiments were performed with the same mass transfer rate of methanol and with the same heat removal from the cell by the methanol stream, methanol was supplied at a flow rate of 1. ml min 1. And the cell operating temperature was controlled at 4 C by a temperature controller. The transient discharging voltages at a constant-current density of 5 ma cm 2 between the cells with new and conventional MEAs under the condition of the cathode air-breathing operation but the active methanol feed are compared in Figs. 5 and 6. It should be mentioned that the voltages were IR corrected to demonstrate the sole effect of the oxygen transport on the performance. It is seen that the new MEA yielded a higher voltage than did the conventional one at both 2. M and 4. M. The voltage with the conventional MEA dropped sharply in 2 min because of more serious liquid water flooding as a result of the larger overall mass transfer Cell voltage (V) Time (min) Fig. 5. Transient discharging voltage at a constant-current density (5 ma cm 2 ) with the active methanol feed between the new and conventional MEAs. Temperature: 4 C; anode: 2. M methanol, 1. ml min 1 ; room temperature: 18.4 C; relative humidity: 71%. Cell voltage (V) Time (min) Fig. 6. Transient discharging voltage at a constant-current density (5 ma cm 2 ) with the active methanol feed between the new and conventional MEAs. Temperature: 4 C; anode: 4. M methanol, 1. ml min 1 ; room temperature: 18.4 C; relative humidity: 71%. resistance on the cathode, whereas the new MEA exhibited a slower decrease. It is also seen that the cell voltages for the both MEAs at 2. M are higher than those at 4. M as a result of a higher methanol crossover at 4. M, leading to a higher mixed overpotential on the cathode and thus a lower cell voltage. Particularly, it can be found that the voltage difference between the new and conventional MEAs at 4. M is much larger than at 2. M, because a higher rate of methanol crossover resulted in a larger demand of oxygen on the cathode. Meanwhile, a higher rate of methanol crossover caused a larger amount of liquid water generated on the cathode, which also contributed to a larger difference in the cell voltages. All these facts reveal that the improved performance with the new MEA is because of the enhanced oxygen transport and the

6 R. Chen, T.S. Zhao / Electrochemistry Communications 9 (27) increased water removal rate as a result of the lowered overall mass transfer resistance in such a MEA structure. The above polarization and constant-current discharging behaviors have showed that the improved performance of the passive DMFC with the new MEA is mainly attributed to the enhanced oxygen transfer rate on the cathode. To further elaborate this point, we also performed electrochemical impedance spectra (EIS) measurements. The EIS analysis allows us to resolve the frequency domain into the individual contributions of the various factors including ohmic, kinetic and mass transport that result in the total voltage losses. We measured the EIS for the passive DMFC with new and conventional MEAs at the cell voltage of.45 V and.3 V, respectively, with the 4. M methanol solution. The experimental results with three arcs corresponding to the high, medium and low frequency are shown in Fig. 7. As indicated, the high and medium frequency arcs are related to the internal cell resistance and the charge transfer reaction kinetics, including the methanol oxidation and oxygen reduction reactions, and removal a 1 the (CO) ads coverage [22]. It can be found from Fig. 7 that, at the both discharging voltages of.45 V and.3 V, the new MEA yielded a larger arc in the high frequency range because of the higher internal cell resistance as a result of the poorer contact between the catalyst layer and porous current collector. When the discharging voltage of the cell was decreased from.45 V to.3 V, the diameter of the medium frequency arcs shrunk for both the cells as shown in Fig. 7a and b, implying that decreasing the discharging voltage can facilitate the MOR and ORR, leading to the gradually shrunk medium frequency arcs for both the MEAs. The experimental results are in agreement with the data reported in the literature [23]. More importantly, it is clear from this figure that the conventional MEA exhibited a larger impedance spectrum than did the new MEA at low frequencies. At low frequencies, the impedance spectra are related to the mass transfer resistance. The smaller impedance spectra at low frequencies for the cell with the new MEA implied that the overall mass transfer resistance was lower than the cell with the conventional MEA, meaning that the oxygen transfer rate is enhanced. Hence, the above EIS results further demonstrated that the improved cell performance with the new MEA is attributed to the enhanced oxygen transfer rate as a result of eliminating the conventional GDL by using the porous current collector. -Z '' (ohm) b -Z '' (ohm) Z '' (ohm) Z '' (ohm) Fig. 7. Nyquist plots of the passive DMFC impedance spectra. Anode: 4. M methanol; room temperature: 18.6 C; relative humidity: 73%. (a) Cell voltage:.45 V; (b) cell voltage:.3 V. 5. Concluding remarks The use of porous metal current collectors in the passive DMFC makes it possible to eliminate the cathode GDL, resulting in a new structure of MEA consisting of an anode diffusion layer, an anode catalyst layer, a membrane and a cathode catalyst layer. It has been demonstrated theoretically that this new MEA allows for a higher mass transfer rate of oxygen, enabling a reduction in the overpotential due to the oxygen concentration polarization. The experimental results revealed that although the passive DMFC with the new MEA yielded a higher internal cell resistance, the new MEA exhibited better performance than did the conventional MEA, particularly with high methanol concentration operation. It was demonstrated that the improved performance with the new MEA was attributed to the enhanced oxygen transport on the cathode. The constant-current discharging tests showed that the passive DMFC with the new MEA yielded much better performance and lower water flooding degree than did the cell with the conventional MEA. In addition, the EIS results showed that the passive DMFC with the new MEA exhibited larger impedance at high frequencies and smaller impedance at low frequencies, further demonstrating that the increased oxygen transport on the cathode is the major reason for the improved cell performance with the new MEA. Our future work will be focused on reducing the internal cell resistance of the passive DMFC with the new MEA to further upgrade cell performance.

7 724 R. Chen, T.S. Zhao / Electrochemistry Communications 9 (27) Acknowledgements The work described in this paper was fully supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (Project No ). The authors also thank RECEMAT Ò International for supplying the metal foam samples in this work. References [1] C.K. Dyer, J. Power Sources 16 (22) [2] Z. Qi, A. Kaufman, J. Power Sources 11 (22) [3] X. Ren, T.E. Springer, T.A. Zawodzinski, S. Gottesfeld, J. Electrochem. Soc. 147 (2) [4] Q. Ye, T.S. Zhao, H. Yang, J. Prabhuram, Electrochem. Solid-State Lett. 8 (25) A52 A54. [5] J. Prabhuram, T.S. Zhao, C.W. Wong, J.W. Guo, J. Power Sources 134 (24) 1 6. [6] K. Scott, W.K. Taama, P. Argyropoulos, J. Power Sources 79 (1999) [7] R.Z. Jiang, D. Chu, J. Electrochem. Soc. 151 (24) A69 A76. [8] Q. Ye, T.S. Zhao, Electrochem. Solid-State Lett. 8 (25) A211 A214. [9] J.J. Hwang, S.D. Wu, L.K. Lai, C.K. Chen, D.Y. Lai, J. Power Sources 161 (26) [1] D. Chu, R.Z. Jiang, Electrochim. Acta 51 (26) [11] J.G. Liu, G.G. Sun, F.L. Zhao, G.X. Wang, G. Zhao, L.K. Chen, B.L. Yi, Q. Xin, J. Power Sources 133 (24) [12] B. Bae, B.K. Kho, T.H. Lim, et al., J. Power Sources 158 (26) [13] T. Shimizu, T. Momma, M. Mohamedi, et al., J. Power Sources 137 (24) [14] G.G. Park, T.H. Yang, Y.G. Yoon, W.Y. Lee, C.S. Kim, Int. J. Hydrogen Energy 28 (23) [15] J. Han, E.S. Park, J. Power Sources 112 (22) [16] Weimin Qian, David P. Wilkinson, Jun Shen, et al., J. Power Sources 154 (26) [17] B.K. Kho, B. Bae, M.A. Scibioh, J. Lee, H.Y. Ha, J. Power Sources 142 (25) [18] J.G. Liu, T.S. Zhao, R. Chen, C.W. Wong, Electrochem. Commun. 7 (25) [19] R. Chen, T.S. Zhao, J. Power Sources 152 (25) [2] R. Chen, T.S. Zhao, Fourth International ASME Conference on Fuel Cell Science, Engineering and Technology, June 19 21, 26, Irvine, California, [21] X. Chao, T.S. Zhao, Q. Ye, Electrochim. Acta 51 (26) [22] J.T. Mueller, P.M. Urban, J. Power Sources 75 (1998) [23] J.T. Muller, P.M. Urban, W.F. Holderich, J. Power Sources 84 (1999)