Surface-modified Nafion membrane by trioctylphosphine-stabilized palladium nanoparticles for DMFC applications

Surface-modified Nafion membrane by trioctylphosphine-stabilized palladium nanoparticles for DMFC applications

ARTICLE IN PRESS Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212 Contents lists available at ScienceDirect Journal of Physics and Che...

448KB Sizes 6 Downloads 54 Views

ARTICLE IN PRESS Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212

Contents lists available at ScienceDirect

Journal of Physics and Chemistry of Solids journal homepage:

Surface-modified Nafion membrane by trioctylphosphine-stabilized palladium nanoparticles for DMFC applications Ai Hua Tian a,b, Ji-Young Kim a, Jin Yi Shi a,b, Kwangyeol Lee a, Keon Kim a, a b

Department of Chemistry, Korea University, Seoul 136-701, South Korea Jilin Institute of Chemical Technology, Jilin 132022, China

a r t i c l e in fo


Article history: Received 18 February 2008 Received in revised form 29 June 2009 Accepted 4 July 2009

Trioctylphosphine (TOP)/Pd composites have been synthesized and used as a methanol-barrier material to modify the surface of Nafion 115. The TOP/Pd composites have been applied to the surface of Nafion instead of being incorporated into the Nafion matrix, to provide the best chance of maintaining the inherent proton conductivity of Nafion. The properties of the TOP/Pd-modified membrane, in terms of its conductivity and methanol permeability, as well as the performance of the membrane electrode assembly (MEA) in direct methanol fuel cell (DMFC), have been analyzed and compared with those of bare Nafion. The DMFC performance of the TOP/Pd-modified membrane is somewhat better than that of the bare Nafion one at methanol concentration of 2 M and significantly better at a high concentration of 5 M. The TOP/Pd-modified membrane is able to operate the DMFC using a high concentration of methanol, which can satisfy the requirement to reduce the reactant volumes for portable applications as well as to achieve high performance. In contrast to bare Nafion, the TOP/Pd-modified membrane with its well-adhering and crack-free modified surface shows effect on reducing the methanol loss. & 2009 Published by Elsevier Ltd.

Keywords: A. Surfaces A. Thin films B. Chemical synthesis

1. Introduction Direct methanol fuel cells (DMFCs) have attracted enormous interest as a primary power source for use in portable electronic devices and transportation [1,2]. In direct methanol fuel cells, methanol can be fed directly for the purpose of converting chemical energy to electrical energy. The fact that methanol is liquid makes it specially attractive, and has several advantages when compared with hydrogen: availability at low cost, ease of handling, transport and storage [3]. However, to use DMFCs commercially, there are two points that should be addressed to achieve high efficiency. The first is the development of highly active anode electrocatalytic activity. The second is the prevention of methanol crossover [4–6]. The crossover of methanol from the anode to the cathode through the polymer electrolyte membrane causes a significant reduction in both cathode potential and fuel utilization, due to the chemical short-circuit reaction of the crossed-over methanol with oxygen at the cathode. Consequently, it is necessary to eliminate or at least reduce the loss of fuel across the cell to improve the performance of DMFCs. Two different strategies can be used to

 Corresponding author. Tel.: +82 2 953 1172.

E-mail address: [email protected] (K. Kim). 0022-3697/$ - see front matter & 2009 Published by Elsevier Ltd. doi:10.1016/j.jpcs.2009.07.005

solve this problem. The first one is to develop new membranes [7–11], because the present perfluorosulfonic polymer membranes, which are now commonly used, are known to be quite permeable to methanol. The second one is to modify the existing perfluorosulfonic polymer membranes to improve their performance [12–14]. This study will focus on the second approach, that of modifying the commercial Nafion membranes. As shown in Fig. 1, we modified the polymer electrolyte membrane by applying organic–inorganic catalytic barrier materials to decrease the methanol crossover and increase the catalytic activity. The organic part of these materials can not only act as a methanol barrier but also increase the compatibility between the inorganic particles and the polymer. The inorganic part plays a catalytic role. Pd, which has similar characteristics to Pt, helped in the electrocatalysis of the electrode. The methanol that is not consumed within the anode catalyst layer would be expected to be further consumed at the catalytic palladium sites before crossing over to the cathode. Pd and Pd alloys are also quite widely used as electrodes for hydrogen diffusion [15,16]. Also, as a methanol-impermeable proton-conducting material [17,18], palladium is a very promising material for use in DMFCs. Herein, we modified the anode side of Nafion by brushing a thin film of trioctylphosphine (TOP)-stabilized palladium onto it. The resulting crack-free modified membranes reduced the methanol crossover, while maintaining a certain degree of proton conductivity, thus providing good performance.


A.H. Tian et al. / Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212

Fig. 1. Schematic illustration of modified membrane.

2. Experimental

2.3. Methanol permeability

2.1. Membrane preparation and modification

A custom-built diffusion cell (Fig. 2) was used for the measurement of the membrane’s methanol permeability. The cell essentially consists of two reservoirs, one of which (reservoir A, VA ¼ 23.70 cm3) was filled with a solution of methanol (8 vol%) and 1-butanol (0.9 vol%), and the other (reservoir B, VB ¼ 23.70 cm3) was filled with only 1-butanol (0.9 vol%) solution. The 0.9 vol% 1-butanol solution in both reservoirs was used as an internal standard. The two reservoirs had a transport channel between them with the membrane acting as a separation barrier. Both reservoirs were vigorously stirred during the permeation experiments. The methanol flux was established across the membrane, driven by the concentration difference between the reservoirs. The change of methanol concentration with time in reservoir B was measured using a capillary gas chromatographic instrument (HP5890). The methanol permeability was calculated using the following equation, which has already been applied elsewhere [22,23]:

All membranes in this study were made of Nafion 115 (Du Pont, USA). They were pretreated according to the standard procedure: 1 h in 3 vol% H2O2 solution at 80 1C, 1 h in distilled water at 80 1C, 1 h in 0.5 M H2SO4 solution at 80 1C and finally 1 h in distilled water at 80 1C again [19]. Then the membranes were dried naturally in air for not less than 4 h. The barrier materials were synthesized by slowly heating a reaction mixture containing 0.1 g of Pd(acac)2 (0.33 mmol) and 7 mL of TOP (16 mmol) from room temperature to 300 1C. After heating for 0.5 h, the resulting dispersion was left to cool down to room temperature, the nanoparticles were retrieved by centrifugation and the excess TOP ligand was removed by washing with methanol [20]. After the removal of the supernatant, the particles were re-dispersed in 5 mL toluene to obtain the methanol barrier materials. Then, they were brushed onto the anode side of Nafion 115 by means of a camel hair brush, with the loading of the Pd nanoparticles on Nafion being about 0.015 mg cm2.

2.2. Characterization of membrane Transmission electron microscopy (TEM, JEOL JEM-2010) was used to examine the morphology of the synthesized barrier materials and field emission scanning electron spectroscopy (FE-SEM, Hitachi S-4300) was employed to obtain information on the quality of the palladium applied. The proton conductivity of the membranes was determined from the impedance data taken after an appropriate equilibrium period, over the frequency range 100 mHz–100 kHz with AC perturbation 5 mV using an impedance analyzer (IM6-Zahner Co.). The conductivity (s) was determined from the bulk resistance (Rb) obtained from the complex impedance plot (Nyquist plot) [21] according to the following equation:

L A  Rb

where L and A are the thickness and area of the polymer electrolyte membrane, respectively.


dcB ðtÞ DK ¼A ðcA  cB ðtÞÞ dt L


where cA and cB are the methanol concentrations in reservoirs A and B, respectively, A and L are the cross-sectional area and thickness of the membrane, respectively, VB is the volume of reservoir B, D the methanol diffusivity and K is the partition coefficient between the membrane and the adjacent solution. Eq. (1) is valid provided that cB5cA, a condition fulfilled in all of the permeation experiments carried out in this work. Here, we assumed that D and K are independent of methanol concentration at the beginning of the experiment. The equation above can be solved to give   cB ðtÞ PA ¼ ðt  t0 Þ VB ln 1  cA L where P is defined as the product, DK the membrane permeability and t0 the time lag. The methanol permeability was calculated from the relation of the methanol concentration and permeation time. 2.4. Membrane electrode assembly (MEA) fabrication In order to test the membranes in a fuel cell station, MEAs were prepared with home-made electrodes according to the following

ARTICLE IN PRESS A.H. Tian et al. / Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212


Fig. 2. Experimental setup for membrane methanol permeability measurement.

procedure: a mixture of catalyst (anode: 40%Pt and 20%Ru on Vulcan XC-72R Carbon, ca. atomic ratio 1:1; cathode: 60%Pt on Vulcan XC-72R Carbon, 3 mg cm2 from Johnson Matthey), distilled water, Nafion solution (5 wt% in a mixture of aliphatic alcohols, Sigma Aldrich) and isopropanol was sprayed onto a commercially available carbon gas diffusion layer (GDL30BC, SGL) by air-brushing until the desired catalyst loading (anode: 3 mg cm2 and cathode: 3 mg cm2) was reached [24]. The electrodes were assembled with the TOP/Pd-modified membrane by hot-pressing at 160 1C and 40 kgf cm2 for 5 min. Similarly, an MEA using Nafion 115 was fabricated for a comparative test. All the MEAs had a size of 2.0 cm  2.0 cm. The procedures were carefully controlled so that their fabrication conditions were identical. 2.5. Unit DMFC test The MEA was coupled with gas-sealing gaskets and placed in a single-cell test station. The cell consisted of two bipolar plates on which channels with a parallel geometry (1 mm wide, 1 mm deep with ridges 1 mm wide) were machined. The active area of the cell was 2.89 cm2. Gold-plated metallic bolts were screwed into the blocks to allow for electrical contact. Methanol was fed into the anode from a methanol container with the aid of a pump (MASTERFLEX C/L) at a flow rate of 0.6 mL min1. Non-humidified oxygen was supplied to the cathode at a flow rate of 100 mL min1 under ambient pressure. The aqueous methanol solution is fed in a single flow through the cell without re-circulation into a reservoir, in order to maintain a constant methanol concentration. The cells were operated at temperatures 30 and 60 1C. The unit cell performances were characterized by the polarization curves, which were obtained using an electronic loader by varying the current. All experimental results were obtained under steadystate conditions. Additionally, 5 M methanol was used to evaluate the possibility of using high methanol concentrations in conjunction with the DMFC.

3. Results and discussions 3.1. Characterization of membranes Fig. 3 shows the TEM image of the palladium nanoparticles stabilized by TOP ligand. The spherical palladium particles are dispersed with a size of 7–8 nm. These Pd nanoparticles were used to modify the Nafion membrane in order to restrict the methanol crossover. The TOP-stabilized Pd nanoparticles are characterized by their distinct agglomeration, indicating that the concentration

Fig. 3. TEM image of TOP ligand-stabilized palladium.

of the TOP ligand is not high enough to prevent the agglomeration of the Pd nanoparticles. It is proposed that the appropriate agglomeration of the Pd nanoparticles may be beneficial for the electrocatalytic activity of the Pd nanoparticle–Nafion membrane interface [25]. Appropriate agglomeration could also increase the loading and coverage of the Pd nanoparticles on the Nafion membrane surface. The morphologies of the surface of the TOP/Pd-modified Nafion membrane were investigated by SEM. Depositing Pd or Pd alloy on the surface of the Nafion membrane by the sputtering technique, which has been studied by other researchers, was an effective method of modifying Nafion [17,18,26,27]. Most of these modified membranes exhibited significantly decreased methanol crossover and improved the performance of fuel cells. Unfortunately, there were many cracks in the Pd films, making them unstable and easily delaminated from the membrane surface and, consequently, the methanol crossover was still very severe. This will affect the longtime operational performance of the cell. Moreover, swelling due to the hydration of the membrane during its operation in the DMFC results in the formation of cracks in the vacuum-deposited Pd layer. Similarly, some cracks were found on the surface of the electroless-plated Pd membrane [28], and to avoid crack formation, the membrane had to be kept in deionized water to keep it fully hydrated [29]. From Fig. 4, we can see clearly that the TOP-stabilized Pd film showed no cracks or exfoliation and had no critical interface, which would result in significant resistance. The TOP-stabilized Pd film was well attached to the Nafion membrane even after hydrating and drying, due to the


A.H. Tian et al. / Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212

Table 1 Conductivity and methanol permeability of Nafion and TOP/Pd-Nafion membranes. Conductivity (S cm1)


Nafion115 TOP/Pd-Nafion

Methanol Concentration (vol%)


30 1C

60 1C

80 1C

Methanol permeability (cm2 min1) (25 1C)

0.061 0.049

0.074 0.056

0.078 0.061

1.124  104 9.454  105

Nafion TOP/Pd-Nafion

2.0 1.5 1.0 0.5 0.0 0


100 150 Time (min)



Fig. 5. Methanol concentration in compartment B as a function of time during methanol permeation.

0.45 0.40

Fig. 4. SEM images of (a) Nafion and (b) TOP/Pd-modified membranes.

Nafion Linear fit of Data_Nafion TOP/Pd-Nafion Linear Fit of Data_TOP/Pd-Nafion

increased compatibility between the palladium nanoparticles and Nafion under the stabilization of the TOP ligand.

3.2. Conductivity and methanol permeability Table 1 compares the proton conductivity and methanol permeability of different samples. The proton conductivity was measured at 30, 60 and 80 1C. The Pd nanoparticles would be expected to enhance the water uptake of Nafion and reduce its methanol uptake [30] and, typically, in a proton-conducting polymer electrolyte, the proton conductivity depends on the mobility of the water molecules. Increasing the water uptake of Nafion will enhance its proton conductivity. However, as reported by other researchers, applying barrier materials to modify proton exchange membranes always decreases the methanol crossover and proton conductivity simultaneously [17,30–32], and our TOP/ Pd-modified membrane was no exception, showing lower proton conductivity than Nafion. The concentration of methanol, which permeated through the clamped membrane, is represented in Fig. 5 as a function of time. Linear behavior is observed in Fig. 6 and methanol permeability can be obtained from the slope of the straight lines. The permeability values at room temperature are listed in Table 1. The methanol permeability varied from 1.124  104 cm2 min1 for Nafion to 9.454  105 cm2 min1 for the membrane surface

-ln (1-cB (t)/cA)

0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 0






Time (min) Fig. 6. ln(1(cB(t)/cA)) vs. time of methanol permeation.

modified with TOP-stabilized palladium nanoparticles, which amounts to a reduction of about 20%.

3.3. Cell performance After installing the MEA in the fuel cell fixture, the cell was activated by feeding it with H2O at 80 1C for at least 30 min, followed by 0.5 M methanol and oxygen at 80 1C for 12 h to

ARTICLE IN PRESS A.H. Tian et al. / Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212

achieve steady-state performance. Afterwards, the cell was operated under different conditions. Fuel cell tests were carried out using both low (2 M) and high (5 M) concentrations of methanol. The current density (I) vs. voltage (V) and power density (W) curves of the MEA were obtained by increasing the current stepwise using an electronic loader. All the cell tests were performed without humidification. Fig. 7 shows the polarization curves for the MEA employing a modified membrane fed with 2 M methanol and oxygen at 30 and 60 1C. These results were compared with those of an MEA based on commercial Nafion 115 operated under the same conditions. The modified membrane gave slightly better performance than the bare Nafion. Considering its reduced conductivity compared with the unmodified Nafion, the use of the Nafion membrane whose surface was modified by TOP-stabilized palladium improved the performance of the DMFC single cell by reducing the methanol permeability. Most DMFC tests are performed with relatively dilute aqueous methanol solutions, due to the high methanol crossover of commercial Nafion membranes. The TOP/Pd-modified membrane was tested in a DMFC at a methanol feed concentration of 5 M. Fig. 8 shows the polarization curves for the TOP/Pd-modified membrane vs. bare Nafion, when 5 M methanol was fed into the unit cell. The modified membrane showed higher performance than the bare Nafion. It was also found that the maximum power density of the TOP/Pd-modified membrane at 30 and 60 1C 0.7

Voltage (V)



2M methanol & oxygen Nafion 30°C Nafion 60°C TOP/Pd-Nafion 30°C TOP/Pd-Nafion 60°C





Power Density (mW/cm2)

80 0.6

0.2 0 0.1 -50



100 150 200 250 300 350 400 Current Density (mA/cm2)




120 100

Voltage (V)


80 0.4

5M methanol & oxygen Nafion 30°C Nafion 60°C Nafion 80°C TOP/Pd-Nafion 30°C TOP/Pd-Nafion 60°C TOP/Pd-Nafion 80°C

0.3 0.2 0.1

60 40 20

Power Density (mW/cm2)

Fig. 7. Comparison of performances of TOP/Pd-Nafion and bare Nafion115 (0.6 mL min1 of 2 M methanol, 100 mL min1 of non-humidified oxygen).

0 0








increased with increasing concentration of methanol, while the maximum power density of the bare Nafion fed with 5 M methanol was the same as that obtained when it was fed with 2 M methanol. The TOP/Pd-modified Nafion membrane has the potential to be used with a high concentration of methanol to reduce the volume of the system and the usage of fuel. Because a higher methanol concentration leads to a higher methanol permeation rate from the anode to the cathode, the power output of the cell tended to decrease with increasing methanol concentration. Modifying the surface of Nafion improved the cell performance by reducing the methanol crossover. However, the TOP/Pd-modified Nafion did not completely eliminate the methanol crossover and higher methanol permeability occurred at higher operation temperatures. When using the TOP/ Pd-modified membrane, the maximum power output obtained using 5 M methanol was increased by 35.9% at 30 1C, but only by 14.3% at 60 1C compared with those obtained using 2 M methanol. And the effect of TOP/Pd addition at higher temperature around 80 1C is not prominent. It showed that the porous barrier was not as effective at high temperature as it was at relatively low temperature because of higher methanol crossover at high temperature. The well-adhering and crack-free modified surface should be optimized to be denser to prohibit more methanol crossover. In addition, the competing effects of the reduced proton conductivity due to the incorporated organic materials and increased proton conductivity due to the incorporated palladium must be considered. There should be an optimum ratio of the organic content to the inorganic content, which provides best performance with low methanol crossover and relative high proton conductivity. This will be studied in further works.

4. Conclusions To make use of palladium’s hydrogen-permeable and methanol-impermeable properties, TOP-stabilized palladium has been synthesized and applied on the surface of Nafion in order to decrease the methanol crossover and improve the performance of fuel cells. With the help of the organic content in the methanol barrier materials, which can increase the compatibility between the palladium particles and Nafion, the modified surface is crackfree and adhered to Nafion very well. This provides more stable performance than the pure Pd-modified membranes, which are characterized by severe cracks. In the case where 2 M methanol is used, the modified membrane shows only slightly better performance than the bare Nafion. However, when fed with 5 M methanol, while the maximum power output of the unit cell using Nafion is the same as that using 2 M methanol, the cell using the TOP/Pd-modified membrane shows an increase in its maximum power output. To conclude, the TOP/Pd-modified Nafion membrane shows reduced methanol loss and represents a good alternative approach for DMFC applications.

Acknowledgements This work was supported by Korea Research Foundation Grant funded by the Korea Government (MOEHRD) (KRF-2005211-C00069), as well as the Korea Science and Engineering Foundation through the Research Center for Energy Conversion & Storage.


Current Density (mA/cm2) Fig. 8. Comparison of performances of TOP/Pd-Nafion and bare Nafion115 (0.6 mL min1 of 5 M methanol, 100 mL min1 of non-humidified oxygen).


References [1] B.D. McNicol, D.A.J. Rand, K.R. Williams, J. Power Sources 83 (1999) 15. [2] C.K. Dyer, J. Power Sources 106 (2002) 31.


[3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]

A.H. Tian et al. / Journal of Physics and Chemistry of Solids 70 (2009) 1207–1212

V. Ramani, H.R. Kunz, J.M. Fenton, J. Membr. Sci. 232 (2004) 31. S. Wasmus, A. Kuver, J. Electroanal. Chem. 461 (1–2) (1999) 14. X. Ren, P. Zelenay, S. Thomas, J. Power Sources 86 (1-2) (2000) 111. A. Heinzel, V.M. Barragan, J. Power Sources 84 (1999) 70. S.N. Nunes, B. Ruffmann, E. Rikowski, S. Vetter, K. Richau, J. Membr. Sci. 203 (2002) 215. B. Ruffmann, H. Silva, B. Schulte, S.P. Nunes, Solid State Ionics 162 (2003) 269. D.H. Jung, Y.B. Myoung, S.Y. Cho, D.R. Shin, D.H. Peck, Int. J. Hydrogen Energy 26 (2001) 1263. Y. Woo, S.Y. Oh, Y.S. Kang, B. Jung, J. Membr. Sci. 220 (2003) 31. J.S. Wainright, J.T. Wang, D. Weng, R.F. Savinell, M. Litt, J. Electrochem. Soc. 142 (1995) L121. P. Dimitrova, K.A. Friedrich, B. Vogt, U. Stimming, J. Electroanal. Chem. 532 (2002) 75. D.H. Jung, S.Y. Cho, D.H. Peck, D.R. Shin, J.S. Kim, J. Power Sources 106 (2002) 173. T. Yamaguchi, F. Miyata, S. Nakao, J. Membr. Sci. 214 (2003) 283. P.L. Cabot, M. Centellens, L. Segarra, J. Casado, J. Electrochem. Soc. 144 (1997) 3749. C. Pu, W. Huang, J. Electrochem. Soc. 142 (7) (1995) 119. S.R. Yoon, G.H. Huang, W.I. Cho, I.-H. Oh, S.-A. Hong, H.Y. Ha, J. Power Sources 106 (2002) 215. W.C. Choi, J.D. Kim, S.I. Woo, J. Power Sources 96 (2001) 411.

[19] X. Ren, T.E. Springer, T.A. Zawodzinski, S. Gottesfeld, J. Electrochem. Soc. 147 (2) (2000) 466. [20] Seung Uk Son, Youngjin Jang, Ki Youl Yoon, Eunae Kang, Taeghwan Hyeon, Nano Letters 4 (6) (2004) 1147. [21] M.H. Woo, O. Kwon, S.H. Choi, M.Z. Hong, H.-W. Ha, K. Kim, Electrochim. Acta 51 (2006) 6051. [22] X.P. Qiu, W.Q. Li, S.C. Zhang, H.Y. Liang, W.T. Zhu, J. Electrochem. Soc. 150 (2003) A917. [23] V. Tricoli, J. Electrochem. Soc. 145 (1998) 3798. [24] Ai Hua Tian, Ji-Young Kim, Jin Yi Shi, Keon Kim, Kwangyeol Lee, J. Power Sources 167 (2007) 302. [25] Zengcai Liu, San Ping Jiang, J. Power Sources 159 (2006) 55. [26] Z.Q. Ma, P. Cheng, T.S. Zhao, J. Membr. Sci. 215 (2003) 327. [27] J. Prabhuram, T.S. Zhao, Z.X. Liang, H. Yang, C.W. Wong, J. Electrochem. Soc. 157 (7) (2005) A1390. [28] H. Sun, G. Sun, S. Wang, J. Liu, X. Zhao, G. Wang, H. Xu, S. Hou, Q. Xin, J. Membr. Sci. 259 (2005) 27. [29] T. Hejze, B.R. Gollas, R.K. Sauerbrey, M. Schmied, F. Hofer, J.O. Besenhard, J. Power Sources 140 (2005) 21. [30] Y.-M. Kim, K.W. Park, J.-H. Choi, I.-S. Park, Y.-E. Sung, Electrochem. Commun. 5 (2003) 571. [31] Y.J. Kim, W.C. Choi, S.I. Woo, W.H. Hong, Electrochim. Acta 49 (2004) 3227. [32] H. Tang, M. Pan, S. Jiang, Z. Wan, R. Yuan, Colloids Surf. A: Physicochem. Eng. Aspects 262 (2005) 65.