Development of a Self-Adaptive Direct

ORIGINAL RESEARCH PAPER

DOI: 10.1002/fuce.201200236

Development of a Self-Adaptive Direct Methanol Fuel Cell Fed with 20 M Methanol J. Guo1, H. Zhang1*, J. Jiang1, Q. Huang1, T. Yuan1, H. Yang1* 1

Energy Storage Laboratory, Shanghai Advanced Research Institute, Chinese Academy of Sciences, Shanghai, 201210, P. R. China

Received January 7, 2013; accepted August 26, 2013; published online September 16, 2013

Abstract A passive and self-adaptive direct methanol fuel cell (DMFC) directly fed with 20 M of methanol is developed for a high energy density of the cell. By using a polypropylene based pervaporation film, methanol is supplied into the DMFC’s anode in vapor form. The mass transport of methanol from the cartridge to the anodic catalyst layer can be controlled by varying the open ratio of the anodic bipolar plate and by tuning the hydrophobicity of anodic diffusion layer. An effective back diffusion of water from the cathode to the anode through Nafion film is carried out by using an additive microporous layer in the cathode that consists of 50 wt.% Teflon and KB-600 carbon. Accordingly, the water

1 Introduction The direct methanol fuel cell (DMFC) has attracted broad interest as a promising power source for portable applications due to its high energy density, system simplicity, quick and easy refueling as well as availability and ease of storage of methanol [1–4]. In order to make the DMFC more competitive with the conventional Li-ion batteries, it is the best choice for the miniature DMFC to be operated in a passive mode and with highly concentrated methanol as fuel. Technically, there are several critical issues that have to be addressed before the widespread applications of the passive DMFCs. These challenges include the crossover of methanol from the anode to the cathode, sluggish kinetics of both anode and cathode reactions, the direct use of methanol of high concentration (until pure methanol) within the system to ensure the higher energy density, the management of water and heat and limited lifetime [5]. To reach the most remarkable feature of the high energy density of methanol, the DMFC is expected to operate with highly concentrated methanol. However, in a conventional DMFC structure, an increase in concentration of the fed methanol would lead to the increase in methanol crossover.

1018

© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

back diffusion not only ensures the water requirement for the methanol oxidation reaction but also reduces water accumulation in the cathode and then avoids serious water flooding, thus improving the adaptability of the passive DMFC. Based on the optimized DMFC structure, a passive DMFC fed with 20 M methanol exhibits a peak power density of 42 mW cm–2 at 25 °C, and no obvious performance degradation after over 90 h continuous operation at a constant current density of 40 mA cm–2. Keywords: Adaptability, Highly Concentrated Methanol, Passive DMFC, Vapor-Feed, Water Management

The crossover of methanol not only gives rise to so-called “mixed potential” effect that greatly reduces the output voltage of the DMFC, but also causes a waste of methanol that lowers the fuel utility [6–8]. To avoid this problem, one essential strategy is the development of a novel less-methanolpermeable proton-exchange membrane [9, 10]. Another route is the use of the oxygen reduction catalysts, which are inactive toward methanol oxidation or have a high methanol tolerance. To date, the Nafion® membranes are still commonly used in the DMFCs. Under such a condition, the dilute methanol solutions (e.g. 1–4 M) are usually used as fuels in the traditional DMFC to mitigate the side effect of methanol crossover [11]. In this case, the energy density of the DMFC system is quite low. This situation is particularly important for the passive DMFC system. As reported, to operate the passive DMFC system with highly concentrated methanol, the methanol crossover could be suppressed significantly by using porous conductive materials with excessively small pores and by employing a barrier layer at the anode to

– [*] Corresponding authors, [email protected] and [email protected]

FUEL CELLS 13, 2013, No. 6, 1018–1023

Guo et al.: Development of a Self-Adaptive Direct Methanol Fuel Cell Fed with 20 M Methanol

2 Experimental 2.1 Preparation of a Membrane Electrode Assembly The Nafion® membrane was sequentially cleaned in 3 vol.% H2O2, 0.5 M H2SO4, and de-ionized water at 80 °C for 2 h. The gas diffusion electrode (GDE) was prepared via a two-step process. Firstly, a mixture of 20 wt.% PTFE and carbon particles (Vulcan® XC-72R) was coated onto a wetproofed Toray® carbon paper (TGPH-60, Toray Inc.) to form the outer MPL. Secondly, an inner MPL, close to the cathode catalyst layer (cCL), which is composed of different PTFE contents varying from 40, 50 to 60 wt.% and different carbon materials (e.g. XC-72R, KB-600, and BP-2000), was prepared. Then, the ink comprised of Pt–Ru black (Johnson Matthey) and carbon-supported Pt (60 wt.% Pt, Johnson Matthey) mixed with 20 wt.% Nafion was blade-coated on the surface of the as-prepared diffusion layer as the catalyst layer at the anode and the cathode, respectively. Finally, the well-treated membrane was sandwiched between the two electrodes by a hot-pressing process at 130 °C and 5 MPa for 3 min to form an MEA with an active geometric area of 4 cm2. 2.2 Fuel Cell Assembly A schematic diagram of the DMFC system with a PVF structure is shown in Figure 1. A self-made PVF based on polypropylene is inserted between the fuel cartridge and anodic current collector as a methanol barrier layer to control the methanol transport. The MEA is sandwiched between two Ti based current collectors whose surface was sputtered with a thin gold layer to reduce the contact resistance between the current collector and the diffusion layer. The open ratios of anode current collector are 12.5, 25, 50, and 75%, respectively; while the open ratio for the cathode current collector is kept at 75%. 2.3 Characterization of DMFC Performance The performance of the passive DMFCs was evaluated by a fuel cell testing station (Arbin FCTS). All the tests were con-

Fig. 1 Schematic of the self-adaptive DMFC.

FUEL CELLS 13, 2013, No. 6, 1018–1023

www.fuelcells.wiley-vch.de

© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1019

ORIGINAL RESEARCH PAPER

increase the mass transfer resistance of methanol from the fuel reservoir to the anode catalyst layer (aCL) [12–16]. Therefore, in order to take advantage of the high energy density of methanol, high concentration methanol should be stored in the fuel cartridge and in the meantime, the methanol concentration should be diluted to an appropriate level at the aCL [17, 18] when the passive DMFC works. The effective management of water within the passive DMFC system is another challenge when the DMFC is fed with highly concentrated methanol [19–21]. It is well known that water is required for the methanol oxidation reaction (MOR) with the molecular ratio of water to methanol over 1:1. In addition, the proton conductivity of Nafion® membrane is also highly dependent on its water content. In the case of a passive DMFC with high concentration methanol, very little and even no water is available. It is expected that the partial water can only come from the back diffusion from the cathode to the anode with the help of new structure design of the membrane electrode assembly (MEA). To operate the passive DMFC with highly concentrated methanol and to carry out an effective management of water, in this work, we focus on the development of a passive DMFC fed with 20 M methanol, capable of self-adaptable and stable operation. The source of water required for the MOR can be sufficiently replenished via the methanol aqueous solution and the water backflow from the cathode to the anode, which could be achieved by using a highly hydrophobic microporous layer (MPL) at cathode that consists of 50 wt.% Teflon and KB-600 carbon. On the other hand, methanol is supplied to the DMFC’s anode in vaporous form driven by the vapor evaporation without consuming any extra energy with the help of a polypropylene based pervaporation film (PVF). Besides, the mass transport of methanol from the cartridge to the aCL can be controlled by varying the open ratio of the anodic bipolar plate and by tuning the hydrophobicity of anodic diffusion layer. Finally, a self-adaptive DMFC fed with 20 M methanol was successfully developed with a stable performance.

ORIGINAL RESEARCH PAPER

Guo et al.: Development of a Self-Adaptive Direct Methanol Fuel Cell Fed with 20 M Methanol ducted at ca. 25 °C and with the relative humidity (RH) of 30%. Prior to testing, the MEA was activated under ambient conditions by immersed into 2 M methanol solution for 24 h. Each experimental point was recorded after a 90 s waiting time.

3 Results and Discussion 3.1 Effect of Anodic Side Structure on DMFC Performance Figure 2 is a performance comparison of two passive DMFCs with and without PVF film, both fed with 20 M methanol. From the figure, the open circuit voltage (OCV) of the PVF based DMFC is ca. 0.74 V, which is much higher than that of the DMFC without PVF (i.e. 0.48 V), suggesting that methanol crossover can be greatly suppressed by the use of PVF since the OCV of the DMFC is a direct reflection of methanol crossover [16]. Moreover, the DMFC with PVF clearly exhibits a better polarization performance than that without PVF. With the help of PVF, the maximum discharging current density and the peak power density significantly increase from 40 to 210 mA cm–2 and from 5 to 40 mW cm–2, respectively; indicating that the methanol transport from the fuel cartridge to the aCL is effectively suppressed by the PVF structure, thus leading to the presence of a low methanol concentration at the aCL. Consequently, the DMFC with PVF shows a much higher power density when operating with 20 M methanol. Figure 3 shows the effect of PTFE content within the carbon paper as anodic backing layer on the performance of the DMFC with PVF. It is clearly seen that the OCV increases from 0.68 to 0.75 V and the maximum power density increases from 31 to 42 mW cm–2 for the PTFE content from 0 to 20 wt.%. It indicates that an increase in PTFE content within the carbon paper would result in a decrease in methanol mass transport rate from fuel cartridge to the aCL. Moreover, the additive PTFE content might help form a certain amount of hydrophobic pores, which may be beneficial for CO2 removal. To further control methanol transport from the fuel reservoir to the aCL, the effect of the open ratio of anodic current

Fig. 2 Performance of the DMFC w/o a PVF fed with 20 M methanol solution.

1020

© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

collector on cell performance was investigated. Figure 4 shows the discharging curves of four DMFCs with different open ratios of anodic current collector at a current density of 40 mA cm–2. It is clear that the highest discharging voltage can be obtained for the DMFC with an open ratio of 25%, indicating that an appropriate mass transport limitation of methanol is achieved at this open ratio. In the case of the open ratio of 12.5%, although the OCV is the highest (0.76 V), the methanol starvation would take place when discharging at high current density, leading to the rapid decline in discharging voltage. Therefore, the change in open ratio of anodic current collector could effectively control the methanol transport rate from the fuel reservoir to the aCL. 3.2 Effect of the Membrane Thickness on DMFC Performance The thickness of Nafion membrane as proton-conductive membrane has a significant influence on water backflow, cell internal resistance and methanol crossover. The thinner the membrane is, the more is the water backflow. On the other hand, the thinner the membrane is, the more is the amount of methanol crossover. There is a trade-off among water backflow, cell internal resistance, and methanol crossover within the DMFC when using different kinds of membranes. Therefore, in the following, we investigate the effect of the mem-

Fig. 3 Effect of PTFE contents within the anode backing layer on the cell performance (20 M CH3OH).

Fig. 4 Effect of the open ratio of anode current collectors on the cell performance when discharging at a constant current density of 40 mA cm–2 (20 M CH3OH).

www.fuelcells.wiley-vch.de

FUEL CELLS 13, 2013, No. 6, 1018–1023

Guo et al.: Development of a Self-Adaptive Direct Methanol Fuel Cell Fed with 20 M Methanol DMFC with a MPL structure has a higher polarization performance and OCV value. The maximum power density is significantly enhanced from 23 to 40 mW cm–2 and OCV value varying from 0.52 to 0.75 V, indicating that the amount of water backflow increased with the help of the cathode MPL. It can be inferred that the methanol concentration at the aCL for a DMFC based on the MPL was lower than that without the MPL. On the other hand, with the help of the MPL, the generated water in the cathode would be hindered to vent out of the cathode diffusion layer to the ambient air. Therefore, to suppress methanol crossover and improve water management is very critical for a passive DMFC operating at high methanol concentration by introducing such a highly hydrophobic MPL at the cathode. In addition, due to no need of extra energy to replenish water, the parasitic power loss of the fuel cell system can be minimized, which is desired for portable applications. In the MEA, the driving force of water backflow is mainly determined by the hydraulic pressure difference (DPc–a) across the membrane, which can be expressed as [23]: …DPc a † ˆ

2r × cos h r

(1†

where r is the surface tension coefficient, h the contact angel, and r the pore radius. It is concluded that the hydraulic pres-

3.3 Effect of Cathode Structure on DMFC Performance It has been reported that water content and liquid pressure in both the anode and cathode catalyst layer can be manipulated by changing the diffusion layer, MPL, and material properties of the catalyst layer such as hydrophobicity, permeability and porosity [22]. An additional highly hydrophobic MPL contributes to force water backflow from the cathode to the anode across the membrane, which not only meets the requirement of the MOR, but also dilutes the methanol concentration in the aCL, thereby mitigating methanol crossover further. Figure 6 shows the performance of DMFC w/o a MPL in the cathode fed with 20 M methanol solution. The

a)

Fig. 6 Comparison of DMFC w/o a MPL in the cathode operating with 20 M methanol solution.

b)

Fig. 5 Variations of polarization curves (a) and internal resistance (b) of DMFC using different Nafion membranes (20 M methanol; 20% PTFE within the anode backing layer; open ratio of anode current collectors at 25%).

FUEL CELLS 13, 2013, No. 6, 1018–1023

www.fuelcells.wiley-vch.de

© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1021

ORIGINAL RESEARCH PAPER

brane thickness on the DMFC performance. Figure 5a shows a performance comparison of the passive DMFCs with different Nafion® membranes. It is seen that when reducing the membrane thickness from 175 lm (Nafion117) to 125 lm (Nafion 115), the OCV and the maximum power density of DMFC increases from 0.53 to 0.75 V and from about 34.5 to 40 mW cm–2, respectively. The improved cell performance could be ascribed to two factors that the thinner membrane leads to a lower internal resistance and that the thinner membrane facilitates an effective water recovery from the cathode to the anode through the membrane. As shown in Figure 5b, the internal resistance of the DMFC decreases from about 600 to 500 mX cm2 when using the thinner Nafion 115 membrane over Nafion 117 membrane. Furthermore, a higher water recovery for the thinner membrane led to a reduced methanol crossover and possibly an intensified water backflow in the MEA, and thus improved the cell performance. However, as shown in Figure 5a, DMFC using Nafion 212 (50 lm) has a much lower peak power of 33.4 mW cm–2 than that using Nafion 115. This can be attributed to that although the thinner membrane benefits to water backflow, too thin membrane tends to exacerbate the methanol crossover in the meantime. As a result, Nafion 115 membrane achieves such a relatively good tradeoff between the reduction of methanol crossover and the water recovery for yielding the best performance under above mentioned experimental conditions.

ORIGINAL RESEARCH PAPER

Guo et al.: Development of a Self-Adaptive Direct Methanol Fuel Cell Fed with 20 M Methanol sure difference created between the aCL and cCL is dependent on the contact angle and micropore radius of the cathode MPL. Therefore, two ways can be adopted to increase the driving force of water back diffusion, i.e. one is to increase the contact angle and the other is to reduce microporous equivalent radius of the cathode MPL. It is known that the loading enhancement of PTFE in the MPL can increase the contact angle, but excessive PTFE content in the MPL would increase the contact resistance and subsequently degrade the cell performance. On the other hand, a smaller pore radius can give rise to a larger hydraulic pressure difference, while too much smaller pore radius would result in air mass transport limitation. Therefore, it is very necessary to determine the optimal value of PTFE content and the pore radius in the cathode MPL. Figure 7 shows the effect of different PTFE contents in the MPL on the cell performance. The DMFC using 50 wt.% PTFE in the cathode MPL yields the best performance with the peak power density of 42 mW cm–2. With the increase of PTFE content to 60 wt.%, the performance of DMFC obviously degrades because of premature mass transport limitation of air at the cathode. It proves that the DMFC with 50 wt.% PTFE content in the cathode MPL is beneficial for water backflow and air transport to the cCL. In order to investigate the effect of equivalent micropore radius of cathode MPL on the hydraulic pressure difference, three kinds of carbon supports including KB-600, BP-2000, and XC-72R were adopted to mix with 50 wt.% PTFE to form the cathode MPL. This structure can create very high hydraulic liquid pressure at the cathode by the formation of a certain amount of hydrophobic hole with small size. It is seen from Figure 8 that the peak power density was respectively obtained at 41 mW cm–2 using KB-600, 39 mW cm–2 using BP-2000, and 35 mW cm–2 using XC-72R, indicating that the passive water recovery is well achieved by using KB-600. This could be due to the smallest average through-pore size of the MPL composed of the mixture of KB-600 and 50 wt.% PTFE. The mean through-pore sizes of the above three-materials based MPL were 0.61 lm with KB-600, 1.29 lm with XC-72R, and 2.68 lm with BP-2000, respectively, which were measured by the capillary flow porometer (Porolux 1000, Ger-

Fig. 7 Polarization performance of the passive DMFCs as a function of different PTFE contents in the cathode MPL with 20 M methanol feed.

1022

© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

many). The smaller the pore radius is, the bigger the hydraulic pressure difference. 3.4 Evaluation of Self-Adaptive Capability for the DMFC Based on the design and optimization of structure parameters, the as-prepared passive DMFC with PVF structure, Nafion 115 membrane, the MPL in the cathode, and 25% open ratio of anode current collectors was continuously run for a long-term stability test. Figure 9 shows the evaluation of cell voltage for more than 90 h under a constant current density of 40 mA cm–2, feeding with 20 M of methanol concentration. During the continuous operation, 20 M methanol was refilled to the fuel cartridge when the cell voltage has a remarkable decrease tendency. It is clear that there is no evidence of permanent performance degradation, revealing that such a passive DMFC has a promising self-adaptive capability for stable operation.

4 Conclusion A passive DMFC directly fed with 20 M methanol aqueous solution as fuel was successfully developed with a high performance by a self-adaptive operating capability. Two key

Fig. 8 Effect of different carbon supports in the cathode MPL on the performance of DMFC fed with 20 M methanol.

Fig. 9 A continuous operation of the self-adaptive DMFC at a current density of 40 mA cm–2 and a methanol concentration of 20 M.

www.fuelcells.wiley-vch.de

FUEL CELLS 13, 2013, No. 6, 1018–1023

Guo et al.: Development of a Self-Adaptive Direct Methanol Fuel Cell Fed with 20 M Methanol

Acknowledgements This work was supported by the National Basic Research Program of China (973 Program, 2012CB932800), the Natural Science Foundation of China (21276158), Shanghai Science and Technology Committee (12ZR1431200 and 11DZ1200400), and the Knowledge Innovation Engineering of the Chinese Academy of Sciences (12406 and 124091231).

References [1] A. Faghri, Z. Guo, Int. J. Heat Mass Transfer 2005, 48, 3891. [2] D. Kim, E. A. Cho, S.-A. Hong, I.-H. Oh, H. Y. Ha, J. Power Sources 2004, 130, 172. [3] Q. Ye, T. S. Zhao, H. Yang, J. Prabhuram, Electrochem. Solid-State Lett. 2005, 8, A52.

[4] Y. H. Chan, T.S. Zhao, R. Chen, C. Xu, J. Power Sources 2008, 178, 118. [5] M. Choi, C. Han, I.-T. Kim, J.-C. An, J.-J. Lee, H.-K. Lee, J. Shim, J. Nanosci. Nanotechnol. 2011, 11, 838. [6] C. Y. Du, T. S. Zhao, W. W. Yang, Electrochim. Acta 2007, 52, 5266. [7] Y. Paik, S.-S. Kim, O. H. Han, Angew. Chem. 2008, 120, 100. [8] T. S. Zhao, C. Xu, R. Chen, W. W. Yang, Progress Energy Combustion Sci. 2009, 35, 275. [9] K. Lee, J. D. Nam, J. Power Sources 2006, 157, 201. [10] T. Yamaguchi, H. Zhou, S. Nakazawa, N. Hara, Adv. Mater. 2007, 19, 592. [11] R. Chen, T. S. Zhao, J. Power Sources 2005, 152, 122. [12] H. Kim, J. Power Sources 2006, 162, 1232. [13] H. Deligöz, S. Yilmaztürk, T. Karaca, H. Özdemir, S. N. Koç, F. Öksüzömer, A. Durmus, M. A. Gürkaynak, J. Membr. Sci. 2009, 326, 643. [14] X. Li, A. Faghri, C. Xu, J. Power Sources 2010, 195, 8202. [15] M. Ali Abdelkareem, N. Nakagawa, J. Power Sources 2006, 162, 114. [16] H. Zhang, I. M. Hsing, J. Power Sources 2007, 167, 450. [17] C. Xu, A. Faghri, X. Li, Int. J. Hydrogen Energy 2011, 36, 8468. [18] C. Xu, A. Faghri, X. Li, J. Electrochem. Soc. 2010, 157, B1109. [19] X. Li, A. Faghri, C. Xu, Int. J. Hydrogen Energy 2010, 35, 8690. [20] V. B. Oliveira, C. M. Rangel, A. M. F. R. Pinto, Int. J. Hydrogen Energy 2009, 34, 8245. [21] C. Xu, A. Faghri, Int. J. Heat Mass Transfer 2010, 53, 1951. [22] E. Peled, A. Blum, A. Aharon, M. Philosoph, Y. Lavi, Electrochem. Solid-State Lett. 2003, 6, A268. [23] G. Q. Lu, F. Q. Liu, C.-Y. Wang, Electrochem. Solid State Lett. 2005, 8, A1.

______________________

FUEL CELLS 13, 2013, No. 6, 1018–1023

www.fuelcells.wiley-vch.de

© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1023

ORIGINAL RESEARCH PAPER

issues including the use of highly concentrated methanol and effective water management have been addressed. A PVF film is applied to enable the methanol evaporation without any extra energy. The open ratio of anode current collector was optimized and 25% open ratio further controlled the methanol mass transport. On the other hand, a MPL at the cathode was introduced to enhance water backflow with the aim of meeting the requirement of the MOR at the aCL. Based on the optimization of structure parameters, a passive DMFC exhibits a very high power density of over 42 mW cm–2 at 20 M methanol operation. Especially, after discharged at 40 mA cm–2 for more than 90 h, the passive DMFC shows no evident permanent performance degradation, indicating a promising self-adaptive capability. The passive DMFC directly fed with highly concentrated methanol shows a substantial potential for portable applications.