CHEMSUSCHEM FULL PAPERS DOI: 10.1002/cssc.201402015
Modulated Ionomer Distribution in the Catalyst Layer of Polymer Electrolyte Membrane Fuel Cells for High Temperature Operation Min-Ju Choo,[a] Keun-Hwan Oh,[b] Hee-Tak Kim,*[a] and Jung-Ki Park*[a, b] Ionomer distribution is an important design parameter for high performance polymer electrolyte membrane fuel cells (PEMFCs); however, the nano-scale modulation of the ionomer morphology has not been intensively explored. Here, we propose a new route to modulate the ionomer distribution that features the introduction of poly(ethylene glycol) (PEG) to the cathode catalyst layer and the leaching the PEG phase from the catalyst layer using a water effluent during operation. The key concept in the approach is the expansion of the ionomer
thin film through the PEG addition. We demonstrate that the modulated ionomer distribution increases the electrochemical active area and proton transport property, without loss in oxygen transport, at a fixed ionomer content. At a high temperature of 120 8C, the power performance at 0.6 V is increased by 1.73-fold with the modulated ionomer distribution as a result of 1.25-fold increase in the electrochemical active area and two-fold increase in the proton transport rate in the catalyst layer.
Introduction PEMFC technology is almost ready for large-scale commercialization, yet the technology sector perceives that the current devices are not sufficiently durable, have limited operational flexibility, and are not cost competitive. In particular, the automotive industry is demanding a membrane electrode assembly (MEA) that endows fuel cells with a cost-to-power ratio of $30 kW1 and high operation temperatures above 90 8C. From the perspective of catalyst layer design, ionomers have a vital function in meeting these needs, as recently discussed by Holdcroft.[1] The ionomers in the catalyst layers function as proton conductors to expand the electrochemically active region, binding materials to impart mechanical stability, and hydrophilic agents to retain moisture and prevent membrane dehydration. In all optimization processes for fuel cell development, the ionomer content is a vital design parameter. If the amount of ionomer is insufficient to form a three-dimensional network, the protons cannot access every part of the catalyst layer. Therefore, only parts of the catalyst can be utilized as active sites for electrochemical reactions. In contrast, if a MEA contains too much ionomer, the electronic conduction paths and gas transport [a] M.-J. Choo, Prof. H.-T. Kim, Prof. J.-K. Park Department of Chemical & Biomolecular Engineering Korea Advanced Institute of Science and Technology (KAIST) Daejeon 305-701 (Republic of Korea) Fax: (+ 82) 42-350-3910 E-mail: [email protected] [email protected] [b] K.-H. Oh, Prof. J.-K. Park Graduate school of EEWS (Energy Environment Water Sustainability) (WCU) Korea Advanced Institute of Science and Technology (KAIST) Daejeon 305-701 (Republic of Korea) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201402015.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
channels (pores) in the catalyst layers will be blocked by either ionomer material or flooded water inside the more hydrophilic pores, particularly at a high current density.[2] In the same way that platinum loading and its dispersion are key factors in the design of catalysts,[3] the ionomer content and its distribution are essential for obtaining high efficiency in the electrochemical conversion by providing a good proton transport rate in the catalyst layer without increasing the mass-transport resistance.[4] From this perspective, the ionomer distribution of the catalytic layer must be controlled in order to achieve high performance. Research on the ionomer distribution has raised questions regarding how to expand the ionomer coverage and reduce the proton transport resistance, while not increasing the oxygen transport resistance, and how to modulate the ionomer distribution from molecular-scale to macro-scale throughout the catalyst layer for better performance. In this regard, Eikerling et al.[5] provided valuable insights into how microphase segregation occurs in ink and how the dielectric properties of the dispersant influence the size of the Pt/C and ionomer aggregates during solidification. Their simulations indicated that the ionomer aggregates interact with the preformed carbon agglomerates and form the interfaces between the two distinct phases. Despite the recent understanding of the morphological evolution of the catalyst layer,[5, 6] the method of controlling the distributions of the ionomer, catalyst, and water is evolving slowly. Here, we propose a new approach to modulate the ionomer distribution through the introduction of poly(ethylene glycol) (PEG) to the cathode catalyst layer and leaching of the PEG phase using the water generated during the fuel cell operation. The PEG, which induces nanophase separation with ionomer, expands the ionomer phase in the catalyst layer through ChemSusChem 0000, 00, 1 – 8
&1&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
dilution, which results in a higher ionomer surface coverage on the Pt/C matrix. Furthermore, the ionomer phase in the catalyst layers becomes more connected due to the expansion. Then, the water soluble PEG is removed during the break-in, reopening the reaction sites and mass transport channels that are blocked by the PEG. Overall, the key concept in the strategy is to increase the ionomer coverage at fixed ionomer content. This leads to an enlarged electrochemical surface area and lowered proton transport resistance without a decrease in mass transport. We employed a short side-chain (SSC) perfluorinated sulfonic acid (PFSA) as a model ionomer instead of the long side-chain (LSC) conventional PFSA in consideration of automotive applications. The LSCPFSA ionomer loses proton conductivity above 90 8C and under low relative humidity (RH);[7] however, the SSC-PFSA ionomers can provide ionic conductivity at high temperatures due to its Figure 1. SEM images of the cross-section of a) PEG0 and b) PEG30, TEM images of PEG30 at c) low magnification higher ion exchange capacity.[8] and d) high magnification, and e) illustration of the evolution of the morphology for the catalyst layers without and with the PEG addition. Various amounts of PEG were incorporated in the catalyst layer, and the morphological changes ed polymer phase to carbon of 69.1 %, 92.6 %, 104.4 %, and were identified. The electrochemical properties and resulting 116.2 %, respectively. Because the ionomer coverage of the catpower performance at automotive operating conditions of alyst surface increases with the ionomer content,[6, 9] a higher 120 8C were evaluated and are discussed. degree of polymer coverage is expected at a higher polymerto-carbon ratio. Figure 1 a and b present the scanning electron Results and Discussion microscopy (SEM) images of the cross-section of the PEG0 and In order to modulate the ionomer distribution in the catalyst PEG30 catalyst layers. These catalyst layers with various PEG/ layer, we varied the weight ratio of the PEG and Aquivion ionionomer ratios did not exhibit apparent differences in their omer from 0 to 0.4. It should be noted that the PEG/ionomer SEM images as shown in Figure 1 a, b, and S3 of the Supportratios higher than 0.4 lead to losses of mechanical integrity in ing Information. The typical porous structure of the Pt/C based the ionomer phase after the PEG removal. The optical transparcatalyst layer was maintained with the PEG addition, which inency of the blend films in these PEG/ionomer ranges (Figdicates that the PEG does not interrupt the evolution of the ure S1, Supporting Information) indicates compatibility besecondary pores. tween the two polymers. The ratio of Pt/C to ionomer was According to the TEM images of the PEG30 catalyst layer fixed at 7:3 irrespective of the PEG content, as maximum (Figure 1 c and d) prepared from diluted catalyst ink, PEG dopower performance was obtained at this ratio for the catalyst mains with a sizes of 50 ~ 100 nm were generated in the catalayers with zero PEG content (Figure S2, Supporting Informalyst layer. For guidance, the borders of the PEG phase and Pt tion). The resulting catalyst layers are denoted as PEG0, PEG20, particles are highlighted in the image. The lattice spacing of PEG30, and PEG40 for the PEG/ionomer ratio of 0, 0.2, 0.3, and the SSC-ionomer phase observed in the TEM image of PEG30 0.4, respectively. The addition of PEG increases the volume of (Figure 1 d and S4 b of the Supporting Information) is 3.4 the blended polymer phase; the PEG/ionomer ratios of 0, 0.2, which can be assigned to the intermolecular spacing of the 0.3, and 0.4 correspond to a percent volume ratio of the blendmain chain, referring to the previously reported value (5 ) for 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&2&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
the LSC-ionomers.[10] The intermolecular spacing of the main tion and H-desorption on Pt surfaces are clearly detected in chain is increased with the decrease in equivalent weight for the potential range of 0.07 V to 0.30 V. The peak at 0.82 V in the LSC-ionomers,[10a] which indicates that the side chain influthe anodic scan and the peak at 0.80 V in the cathodic scan ences the intermolecular spacing of the main chain. In this correspond to the Pt-oxidation and the Pt-oxide-reduction, reregard, the shorter intermolecular spacing for the SSC-ionomer spectively. The values for the ECSA and Cdl can be determined probably originates from the shorter side chain length. The from the amount of charge for the H-desorption and the caPEG domains can be differentiated from the Pt particles by the pacitive charging current at 0.4 V, respectively.[11] As shown in absence of the characteristic metallic lattice of Pt. In particular, Figure 2 and listed in Table 1, in the PEG/ionomer range of 0 to the lattice distance of 0.23 nm indicated in Figure 1 d and Figure S4d of the Supporting InforTable 1. Electrochemical parameters of the catalyst layers with various PEG/ionomer ratios. mation denotes the interplanar spacing of (111)Pt. The PEG doPt utilization Cdl Rcl Power density j0 Tafel slope Sample ECSA [%] [mF cm2] [W cm2] at 0.6 V [mW cm2] [mA cm2] [mV dec1] [m2 g1] mains appear to be embedded in the ionomer phase; the ionPEG0 14.1 46.2 28.5 0.37 108.8 3.61 104 68.9 69.0 PEG20 17.0 55.7 30.5 0.29 172.6 4.86 104 omer phase that covers the PEG 4 68.1 PEG30 17.6 57.8 39.5 0.16 188.5 5.87 10 domain is identified, as illustratPEG40 16.3 53.4 34.1 0.18 149.6 2.35 104 65.4 ed in Figure 1 d. More images that support the existence of the PEG domain are presented in Figure S4 of the Supporting Information. The PEG, which is 0.3, both ECSA and Cdl increased with the PEG/ionomer ratio easily soluble in water, leached out during the break-in and and they decreased with a further increase of the ratio up to cell operation. The disappearance of the PEG phase in the TEM 0.4. The percent ratio of the ECSA to the actual Pt surface image after soaking in water at room temperature for 24 h (30.5 m2 g1) was 46.2 %, 55.7 %, 57.8 %, and 53.4 % for PEG0, supports a facile PEG removal (Figure S4). For simplification PEG20, PEG30, and PEG40, respectively. The values for Cdl are and clarity, the morphology evolution for the PEG/ionomer catin good agreement with those for ECSA, which confirms the alyst layers is schematically represented in Figure 1 e. In the abmodulated ionomer distribution. It should be noted that, for sence of PEG, the ionomer aggregate and catalyst agglomerate PEG30, the ionomer coverage could be enhanced by 25 % partially covered the catalyst layer with an ionomer film, based on the ECSA and by 38 % based on the Cdl without alterwhereas, in the presence of PEG, the ionomer/PEG blend coving the ionomer content. This is conclusive evidence for the ered a larger portion of the catalyst layer and the removal of modulation of the ionomer distribution using this approach. the PEG left an expanded and more connected ionomer film. The existence of an optimum PEG content is not unexpected As the first evidence for the modulation of the ionomer disbecause the amount of Pt surface in contact with the PEG tribution, the PEG content-dependent variations of the electrodomain increases with the PEG/ionomer ratio, which negatively chemically active surface area (ECSA) and double layer capaciinfluences the ionomer coverage. tance (Cdl), which are known to be sensitive to the ionomer Further evidence for the modulated ionomer distribution is the change in the proton transport through the catalyst layer. coverage of Pt/C surface, are provided using the cyclic voltamIt is expected that the ionomer phases may be more connectmetry (CV) technique. As presented in Figure 2, the H-adsorped when the ionomers are expanded through the PEG addition. Therefore, the impedances under a H2/N2 (anode/cathode) atmosphere were measured at 120 8C and RH 40 % in order to eliminate the effect of the Faradaic reaction.[12] As shown in Figure 3 a, the linear region of 458, which corresponds to the ionic migration through the ionomer phases in the catalyst layer, was observed at high frequencies for all catalyst layers. The following inclined spikes at low frequencies were attributed to the total capacitance and resistance of the catalyst layer. The value for the proton transport resistance through the catalyst layer (Rcl) could be determined from the length of the Warburg-like region (Rcl/3) projected onto Z’-Rohm, when the low frequency line had a phase angle of 908. The low frequency lines slightly deviated from the 908 line for these catalyst layers as shown in Figure 3 a, which was attributed to a non-homogeneous distribution of ionic resistance across the catalyst layer.[12, 13] For simplicity, an approximate value for Rcl was obtained from the intercept on Z’ of an asymptotic line extending Figure 2. CV curves for the catalyst layers with various PEG/ionomer ratios from the low frequency line, as indicated in Figure 3 a. The measured in a H2/N2 (anode/cathode) atmosphere at 120 8C. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&3&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org phological features of the pore and ionomer film. This consideration was the motivation for conducting a limiting current test that provides a quantitative measure of the oxygen transport resistance in the catalyst layer.[15] The flux of oxygen (NO2), is proportional to the concentration difference of oxygen in the gas channel (CO2) and at the Pt surface (CO2,Pt), and it can be related to the current density using Faraday’s law as follows: N O2 ¼
j 1 C CO2 ;Pt ¼ 4F rtotal O2
ð1Þ
where F is the faraday constant. At the limit current density (jlim), the O2 concentration at the Pt surface is zero. Consequently, the total oxygen transport resistance (rtotal), which is the reciprocal of the total oxygen transport coefficient, can be evaluated from the jlim using Equation (2), as follows: rtotal ¼ CO2
4F jlim
ð2Þ
The rtotal is composed of the transport resistance through the macropores of diffusion media and catalyst layer (rmacro), through the mesopores in the catalyst layer (rmeso), and through the ionomer film (rionomer) covering the Pt catalyst surfaces, as follows: r total ¼ r macro þ r meso þ r ionomer
Figure 3. a) Nyquist plots of the impedances for the catalyst layers with various PEG/ionomer ratios measured under a H2/N2 (anode/cathode) atmosphere at 120 8C, and b) plot of Rcl and proton conductivity of the catalyst layer as a function of the PEG/ionomer ratios.
values for Rcl were 0.37, 0.29, 0.16, and 0.18 W cm2 for PEG0, PEG20, PEG30, and PEG40, respectively. The analysis results are summarized in Figure 3 b, and these indicate a two-fold increase in the proton transport rate for PEG30 compared with that for PEG0. The proton conductivity of the catalyst layer calculated from the Rcl and thickness of the catalyst layer is also illustrated in Figure 3 b. This explains the evolution of a more connected ionomer morphology with the addition of PEG. It is well known that the proton conductivity increases as the ionomer content of the catalyst layer is increased, but the catalyst sites become blocked and gas cannot pass through the catalyst layer.[9b, 14] Thus, the modulation of the ionomer distribution in this paper is meaningful because the ion conductivity of the catalyst layer was improved although the ionomer content was fixed. The structures of the pore and ionomer film generally govern the oxygen transport property in the catalyst layer, and PEG addition should be accompanied by changes in the mor 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ð3Þ
Then, rmacro is determined from the difference in the rtotal on switching the balance gas from N2 to He, as described in previous studies.[15b, 16] The oxygen transport resistance of the agglomerate of catalyst and ionomer (rmeso + rionomer) can be quantified by subtracting rmacro from rtotal. As demonstrated in Figure 4, the oxygen transport resistance of the agglomerate dominates the total transport resistance. The comparison of rmeso + rionomer for the catalyst layers (Figure 4)
Figure 4. Plots of rtotal (crossed symbols), rmacro (open symbols), and rionomer + rmeso (filled symbols) as a function of PEG/ionomer ratio.
ChemSusChem 0000, 00, 1 – 8
&4&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS clearly demonstrates that the modulated ionomer distribution does not retard the oxygen transport in the agglomerate at 120 8C. The retarded oxygen transport that results from the expanded area of ionomer film is compensated by the fast oxygen transport through the pores generated by the PEG removal, which leads to the weak dependency of the oxygen transport property on PEG content. This is strong evidence for the PEG removal during fuel cell operation, because the PEG phases in the catalyst layer function as an additional barrier for the oxygen transport, which leads to an increase in rionomer if they are not removed. Figure 5 a presents the IV polarizations for the catalyst layers obtained at 120 8C, 150 kPa, and RH 40 %. The operation at 120 8C is significant for automotive applications because it enables the use of radiators similar to those used in modern com-
www.chemsuschem.org an increase in the power density for the PEG/ionomer ratios of 0 to 0.3, followed by a decrease for ratios of 0.3 to 0.4. This result strongly suggests that the difference in activation polarization significantly contributes to the difference in the power performance. For a quantitative assessment of the influence of the activation polarization on the power performance, the IRcorrected cell voltages (EiRfree) were examined. By correcting the raw voltage data with the values of ohmic resistance (Rohm), which were determined from the impedance at 1 kHz, the ohmic polarization can be eliminated, as given in Equation (4). At very low current densities at which the activation polarization dominates the overall reaction, EiRfree can be expressed in terms of the reversible cell voltage (Er), exchange current density (j0), and hydrogen crossover current (jx), as follows:
EiRfree ¼ E þ jRohm
Figure 5. a) IV polarization curves and b) Tafel plots of the MEAs with various PEG/ionomer ratios (temperature: 120 8C, humidity: RH 40 % for both cathode and anode feeds, back pressure: 150 kPa for both cathode and anode).
bustion engines, whereas operation at 80 8C requires double the volume of the radiator.[17] In the automotive operation condition, the power performance was improved with the PEG addition in all current densities investigated. The power performance at 0.6 V was used to evaluate the cell performance (Table 1). Interestingly, the power performance of PEG30 was 73 % higher than that of PEG0. In accordance with the trend observed for the ECSA and Cdl, the IV measurements revealed 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2:303RT j þ jx ¼ Er log anF j0
ð4Þ
where R, T, and F have standard meanings, and n and a are the number of electrons transferred and the charge transfer coefficient, respectively. The values for jx were determined from the linear sweep voltammetry measurement under a H2/ N2 (anode/cathode) atmosphere. Figure 5 b presents the IR-corrected cell voltage versus the hydrogen crossover-corrected current density (jeff = j + jx, jx = 1.5–1.7 mA cm2). Departure from the Tafel slope, which is indicative of a mass transport polarization for the oxygen reduction reaction, was observed above 0.06 A cm2. To avoid the effect of mass transport on Tafel analysis, we calculated the Tafel slope and exchange current density from the data in the range of 0.003 ~ 0.02 A cm2. In the activation polarization region (E > 0.80 V), the apparent Tafel slopes for all MEAs were similar (65–69 mV dec1), which implies the same inherent electrode kinetics toward the oxygen reduction reaction. The values for j0, which were also determined from Figure 5 b, are listed in Table 1. In accordance with the variation of the ECSA with the PEG/ionomer ratio, j0, which is proportional to ECSA,[9b, 14] increased with the PEG/ionomer ratio up to 0.3, but decreased with further increases in the PEG/ionomer ratio. Considering all data for the j0, proton transport and oxygen transport, the increased power performance with the PEG addition up to a PEG/ionomer ratio of 0.3 was attributed to the increased ECSA and facilitated proton transport through the catalyst layer. The impedance analysis also exhibited significant reductions in the charge-transfer resistance in a wide range of current density with the PEG addition, and the appearance of a minimum in the charge-transfer resistance at the PEG/ionomer ratio of 0.3 (Figure S5); this is in good agreement with the results of the polarization curves. Therefore, the appearance of the maximum power performance in PEG30 results from the maximum ECSA and proton conductivity of the catalyst layer at the PEG/ionomer ratio of 0.3. Because the proton transport and oxygen transport properties of PEG40 were comparable to those of PEG30 as indicated by the limiting current measurements, the lower power performance of ChemSusChem 0000, 00, 1 – 8
&5&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS PEG40 compared with that of PEG30 should originate from the lower ECSA.
Conclusions We have demonstrated the modulation of the ionomer distribution through the addition of PEG in the catalyst layer and the corresponding improvement in fuel cell performance. The modulation enlarges the electrochemical active surface and enhances the proton transport properties of the catalyst layer, without a loss in mass transport. In particular, the incorporation of PEG at a PEG/ionomer ratio of 0.3 results in a 73 % increase in the power density at 0.6 V under automotive operation conditions in conjunction with a 24 % increase in the ionomer coverage on the Pt catalyst and a 100 % increase in proton conductivity of the catalyst layer. Therefore, the modulated ionomer distribution offers a new method for optimizing the performance of polymer electrolyte membrane fuel cells.
Experimental Section Preparation and morphological characterization of the catalyst layers For the catalyst ink using the SSC ionomer Aquivion as an electrode ionomer, a carbon-supported platinum catalyst [TEC10 V50E (45.8 wt % Pt), Tanaka Kikinzoku Kogyo, Japan] was dispersed in a binary solvent of isopropyl alcohol (IPA)/water (1:1 in volume); predetermined amounts of Aquivion dispersion D79–20BS (20 wt %, equivalent weight of 790 g eq1, Solvay Solexis, Italy) and PEG (MW 2000, Aldrich) were added to the dispersion. In this work, we employed the highly stable Pt/C catalyst with large Pt size of 9.2 nm in diameter to prevent rapid catalyst degradation by sintering or Pt dissolution at the high temperature. The weight ratio of the Pt/C and Aquivion ionomer was fixed at 7:3 in a dry state. The weight ratios of PEG to Aquivion were varied: 0, 0.2, 0.3, and 0.4 for the both anode and cathode catalyst layers. The mixtures underwent sonication for 1 h and were cast onto Kapton film (Dupont, USA) supported on a glass plate using a doctor blade (gap: 0.3 mm) at a constant speed of 15 mm s1, followed by drying at 50 8C for 24 h. The Pt loading level of the catalyst layer was carefully controlled to be within 0.42 0.02 mg cm2 for the cathode and anode catalyst layers. The cathode and anode catalyst layers (effective area of 25 cm2) coated on the Kapton film were transferred to each side of the Aquivion E87–05SX membrane (50 um, EW 870 g eq1) at a temperature of 160 8C and pressure of 1500 psi for 3 min in order to form the catalyst-coated membrane (CCM). The Kapton films were peeled from the laminate after cooling. The ionomer morphologies of the catalyst layers were investigated using scanning electron microscopy (SEM; Sirion, FEI, USA) and transmission electron microscopy (TEM; Tecnai F30, FEI, USA) operating at 300 kV. The time duration of beam exposure was carefully minimized to prevent the sample from beam damage.
www.chemsuschem.org Korea). The cells were typically operated at a total gas pressure of 150 kPa, a cell temperature of 120 8C, and a relative humidity (RH) of 40 % for both electrodes. The hydrogen gas was supplied to the anode at 500 cm3 min1 and air to the cathode at 1500 cm3 min1, which corresponds the stoichiometry of H2/air = 14.9/18.0 at 0.2 A cm2. The polarization curves were obtained by stepping up the current density, allowing the cell to stabilize, and measuring the cell voltage. One minute was spent at each current density. After daily measurements of three IV curves for successive two days as a break-in, the IV curves for detailed analysis were obtained. The IV curves were not varied after the second day. The H2 crossover current density was measured using a linear sweep voltammetry from 0.05 to 0.70 V at a scan rate of 1 mV s1 with a H2 (anode) and N2 (cathode) feed of 500 cm3 min1.
Electrochemical impedance spectroscopy and cycling voltammetry In order to measure the ionic resistance for the catalyst layer, electrochemical impedance spectroscopy (EIS) was conducted in the frequency range of 105 to 101 Hz using a Solartron 1470E potentiostat and a 1400 frequency analyzer under H2/N2 (anode/cathode = 500/1500 cm3 min1). The amplitude of the sinusoidal voltage signal was 10 mV. The cathode was used as a working electrode, and the anode as a counter electrode. It was assumed that the anode (H2 electrode) had negligible impedance relative to the cathode. The CV of the single cells under H2/N2 (anode/cathode = 50/ 0 cm3 min1) was conducted using a Biologic HCP-803 potentiostat in the potential range from 0.07 to 1.20 V vs. DHE at 50 mV s1. It was conducted in hydrogen purging for the anode and nitrogen purging for the cathode for 30 min prior to the CV measurements in order to remove any trace of oxygen in the cell.
Limiting current density measurement Linear sweep voltammetry (LSV) was performed in order to investigate the oxygen transport through the cathode at the cell temperature of 120 8C. The hydrogen gas (40 % RH) was supplied to the anode (reference electrode) and diluted O2 (1 % O2 in the balance of N2 or He, 40 % RH) was supplied to the cathode (working electrode) at a flow rate of 100 cm3 min1 and at atmospheric back pressure. The oxygen reduction current was recorded at a scan rate of 10 mV s1 from 1.2 to 0.1 V. The limiting current density was determined using an extrapolation of the currents in the range of 0.1 to 0.2 V to zero voltage.
Acknowledgements This work was supported by the New&Renewable Energy R&D Program (grant no. 20113020030020) under the Ministry of Knowledge Economy (MKE), Republic of Korea. This paper was also supported as a project of the Global PhD Fellowship supported by the National Research Foundation of Korea conducts from 2011.
IV polarization of the single cells A single cell was assembled using the CCM, a pair of gas diffusion media (39BC, SGL), a pair of gaskets, and a pair of graphite blocks with triple serpentine flow fields. The single cells were connected to a test station equipped with an electronic load (Won-A Tech, 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Keywords: catalyst layer · electrochemistry · fuel cells · ionomer distribution · PEMFCs [1] S. Holdcroft, Chem. Mater. 2014, 26, 381 – 393.
ChemSusChem 0000, 00, 1 – 8
&6&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS [2] a) E. Passalacqua, F. Lufrano, G. Squadrito, A. Patti, L. Giorgi, Electrochim. Acta 2001, 46, 799 – 805; b) G. Sasikumar, J. W. Ihm, H. Ryu, Electrochim. Acta 2004, 50, 601 – 605; c) D. Song, Q. Wang, Z. Liu, M. Eikerling, Z. Xie, T. Navessin, S. Holdcroft, Electrochim. Acta 2005, 50, 3347 – 3358; d) M. Uchida, Y. Aoyama, N. Eda, A. Ohta, J. Electrochem. Soc. 1995, 142, 4143 – 4149. [3] R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan, N. Garland, D. Myers, M. Wilson, F. Garzon, D. Wood, P. Zelenay, K. More, K. Stroh, T. Zawodzinski, J. Boncella, J. E. McGrath, M. Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi, S. Miyata, A. Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K. Kimijima, N. Iwashita, Chem. Rev. 2007, 107, 3904 – 3951. [4] a) C. S. Kong, D.-Y. Kim, H.-K. Lee, Y.-G. Shul, T.-H. Lee, J. Power Sources 2002, 108, 185 – 191; b) X. Cheng, B. Yi, M. Han, J. Zhang, Y. Qiao, J. Yu, J. Power Sources 1999, 79, 75 – 81; c) J. M. Song, S. Suzuki, H. Uchida, M. Watanabe, Langmuir 2006, 22, 6422 – 6428. [5] a) K. Malek, M. Eikerling, Q. Wang, T. Navessin, Z. Liu, J. Phys. Chem. C 2007, 111, 13627 – 13634; b) K. Malek, T. Mashio, M. Eikerling, Electrocatalysis 2011, 2, 141 – 157. [6] T. Soboleva, K. Malek, Z. Xie, T. Navessin, S. Holdcroft, ACS Appl. Mater. Interfaces 2011, 3, 1827 – 1837. [7] a) K. Epping Martin, J. P. Kopasz, Fuel Cells 2009, 9, 356 – 362; b) K. A. Mauritz, R. B. Moore, Chem. Rev. 2004, 104, 4535 – 4586. [8] a) K. D. Kreuer, M. Schuster, B. Obliers, O. Diat, U. Traub, A. Fuchs, U. Klock, S. J. Paddison, J. Maier, J. Power Sources 2008, 178, 499 – 509; b) A. S. Aric, A. D. Blasi, G. Brunaccini, F. Sergi, V. Antonucci, P. Asher, S. Buche, D. Fongalland, G. A. Hards, J. D. B. Sharman, A. Bayer, G. Heinz, R. Zuber, M. Gebert, M. Corasaniti, A. Ghielmi, D. J. Jones, ECS Trans. 2009, 25, 1999 – 2007; c) M. Gebert, A. Ghielmi, L. Merlo, M. Corasaniti, V. Arcella, ECS Trans. 2010, 26, 279 – 283; d) C. Lei, D. Bessarabov, S. Ye, Z. Xie, S. Holdcroft, T. Navessin, J. Power Sources 2011, 196, 6168 – 6176.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org [9] a) T. Soboleva, X. Zhao, K. Malek, Z. Xie, T. Navessin, S. Holdcroft, ACS Appl. Mater. Interfaces 2010, 2, 375 – 384; b) P. Gode, F. Jaouen, G. Lindbergh, A. Lundblad, G. Sundholm, Electrochim. Acta 2003, 48, 4175 – 4187. [10] a) T. D. Gierke, G. E. Munn, F. C. Wilson, J. Polym. Sci. Polym. Phys. Ed. 1981, 19, 1687 – 1704; b) N. A. Nazir, H. Kudo, T. Nishikubo, T. Kyu, J. Mater. Sci. 2012, 47, 7269 – 7279. [11] L. Su, W. Jia, C.-M. Li, Y. Lei, ChemSusChem 2014, 7, 361 – 378. [12] M. C. Lefebvre, R. B. Martin, P. G. Pickup, Electrochem. Solid-State Lett. 1999, 2, 259 – 261. [13] D. Malevich, B. R. Jayasankar, E. Halliop, J. G. Pharoah, B. A. Peppley, K. Karan, J. Electrochem. Soc. 2012, 159, F888 – F895. [14] G. Li, P. G. Pickup, J. Electrochem. Soc. 2003, 150, C745 – C752. [15] a) A. Nanjundappa, A. S. Alavijeh, M. El Hannach, D. Harvey, E. Kjeang, Electrochim. Acta 2013, 110, 349 – 357; b) N. Nonoyama, S. Okazaki, A. Z. Weber, Y. Ikogi, T. Yoshida, J. Electrochem. Soc. 2011, 158, B416 – B423; c) T. A. Greszler, D. Caulk, P. Sinha, J. Electrochem. Soc. 2012, 159, F831 – F840. [16] D. R. Baker, D. A. Caulk, K. C. Neyerlin, M. W. Murphy, J. Electrochem. Soc. 2009, 156, B991 – B1003. [17] a) M. Mathias, H. Gasteiger, R. Makharia, S. Kocha, T. Fuller, T. Xie, J. Pisco, Prepr. Pap.-Am. Chem. Soc,. Div. Fuel Chem. 2004, 49, 471 – 474; b) J. Li, X. Li, Y. Zhao, W. Lu, Z. Shao, B. Yi, ChemSusChem 2012, 5, 896 – 900; c) J. Yang, D. Aili, Q. Li, L. N. Cleemann, J. O. Jensen, N. J. Bjerrum, R. He, ChemSusChem 2013, 6, 275 – 282.
Received: February 5, 2014 Revised: February 24, 2014 Published online on && &&, 0000
ChemSusChem 0000, 00, 1 – 8
&7&
These are not the final page numbers! ÞÞ
FULL PAPERS M.-J. Choo, K.-H. Oh, H.-T. Kim,* J.-K. Park* && – && Modulated Ionomer Distribution in the Catalyst Layer of Polymer Electrolyte Membrane Fuel Cells for High Temperature Operation
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Ionomer distribution modulation: A series of new catalyst layers for high temperature operation was fabricated through a simple poly(ethylene glycol) addition and leaching process. The ionomer domain was expanded by poly(ethylene glycol) addition, resulting in an increase in the electrochemical active area and proton transport without loss in oxygen transport, at a fixed ionomer content.
ChemSusChem 0000, 00, 1 – 8
&8&
These are not the final page numbers! ÞÞ
Modulated Ionomer Distribution in the Catalyst Layer of Polymer Electrolyte Membrane Fuel Cells for High Temperature Operation Min-Ju Choo,[a] Keun-Hwan Oh,[b] Hee-Tak Kim,*[a] and Jung-Ki Park*[a, b] Ionomer distribution is an important design parameter for high performance polymer electrolyte membrane fuel cells (PEMFCs); however, the nano-scale modulation of the ionomer morphology has not been intensively explored. Here, we propose a new route to modulate the ionomer distribution that features the introduction of poly(ethylene glycol) (PEG) to the cathode catalyst layer and the leaching the PEG phase from the catalyst layer using a water effluent during operation. The key concept in the approach is the expansion of the ionomer
thin film through the PEG addition. We demonstrate that the modulated ionomer distribution increases the electrochemical active area and proton transport property, without loss in oxygen transport, at a fixed ionomer content. At a high temperature of 120 8C, the power performance at 0.6 V is increased by 1.73-fold with the modulated ionomer distribution as a result of 1.25-fold increase in the electrochemical active area and two-fold increase in the proton transport rate in the catalyst layer.
Introduction PEMFC technology is almost ready for large-scale commercialization, yet the technology sector perceives that the current devices are not sufficiently durable, have limited operational flexibility, and are not cost competitive. In particular, the automotive industry is demanding a membrane electrode assembly (MEA) that endows fuel cells with a cost-to-power ratio of $30 kW1 and high operation temperatures above 90 8C. From the perspective of catalyst layer design, ionomers have a vital function in meeting these needs, as recently discussed by Holdcroft.[1] The ionomers in the catalyst layers function as proton conductors to expand the electrochemically active region, binding materials to impart mechanical stability, and hydrophilic agents to retain moisture and prevent membrane dehydration. In all optimization processes for fuel cell development, the ionomer content is a vital design parameter. If the amount of ionomer is insufficient to form a three-dimensional network, the protons cannot access every part of the catalyst layer. Therefore, only parts of the catalyst can be utilized as active sites for electrochemical reactions. In contrast, if a MEA contains too much ionomer, the electronic conduction paths and gas transport [a] M.-J. Choo, Prof. H.-T. Kim, Prof. J.-K. Park Department of Chemical & Biomolecular Engineering Korea Advanced Institute of Science and Technology (KAIST) Daejeon 305-701 (Republic of Korea) Fax: (+ 82) 42-350-3910 E-mail: [email protected] [email protected] [b] K.-H. Oh, Prof. J.-K. Park Graduate school of EEWS (Energy Environment Water Sustainability) (WCU) Korea Advanced Institute of Science and Technology (KAIST) Daejeon 305-701 (Republic of Korea) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201402015.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
channels (pores) in the catalyst layers will be blocked by either ionomer material or flooded water inside the more hydrophilic pores, particularly at a high current density.[2] In the same way that platinum loading and its dispersion are key factors in the design of catalysts,[3] the ionomer content and its distribution are essential for obtaining high efficiency in the electrochemical conversion by providing a good proton transport rate in the catalyst layer without increasing the mass-transport resistance.[4] From this perspective, the ionomer distribution of the catalytic layer must be controlled in order to achieve high performance. Research on the ionomer distribution has raised questions regarding how to expand the ionomer coverage and reduce the proton transport resistance, while not increasing the oxygen transport resistance, and how to modulate the ionomer distribution from molecular-scale to macro-scale throughout the catalyst layer for better performance. In this regard, Eikerling et al.[5] provided valuable insights into how microphase segregation occurs in ink and how the dielectric properties of the dispersant influence the size of the Pt/C and ionomer aggregates during solidification. Their simulations indicated that the ionomer aggregates interact with the preformed carbon agglomerates and form the interfaces between the two distinct phases. Despite the recent understanding of the morphological evolution of the catalyst layer,[5, 6] the method of controlling the distributions of the ionomer, catalyst, and water is evolving slowly. Here, we propose a new approach to modulate the ionomer distribution through the introduction of poly(ethylene glycol) (PEG) to the cathode catalyst layer and leaching of the PEG phase using the water generated during the fuel cell operation. The PEG, which induces nanophase separation with ionomer, expands the ionomer phase in the catalyst layer through ChemSusChem 0000, 00, 1 – 8
&1&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
dilution, which results in a higher ionomer surface coverage on the Pt/C matrix. Furthermore, the ionomer phase in the catalyst layers becomes more connected due to the expansion. Then, the water soluble PEG is removed during the break-in, reopening the reaction sites and mass transport channels that are blocked by the PEG. Overall, the key concept in the strategy is to increase the ionomer coverage at fixed ionomer content. This leads to an enlarged electrochemical surface area and lowered proton transport resistance without a decrease in mass transport. We employed a short side-chain (SSC) perfluorinated sulfonic acid (PFSA) as a model ionomer instead of the long side-chain (LSC) conventional PFSA in consideration of automotive applications. The LSCPFSA ionomer loses proton conductivity above 90 8C and under low relative humidity (RH);[7] however, the SSC-PFSA ionomers can provide ionic conductivity at high temperatures due to its Figure 1. SEM images of the cross-section of a) PEG0 and b) PEG30, TEM images of PEG30 at c) low magnification higher ion exchange capacity.[8] and d) high magnification, and e) illustration of the evolution of the morphology for the catalyst layers without and with the PEG addition. Various amounts of PEG were incorporated in the catalyst layer, and the morphological changes ed polymer phase to carbon of 69.1 %, 92.6 %, 104.4 %, and were identified. The electrochemical properties and resulting 116.2 %, respectively. Because the ionomer coverage of the catpower performance at automotive operating conditions of alyst surface increases with the ionomer content,[6, 9] a higher 120 8C were evaluated and are discussed. degree of polymer coverage is expected at a higher polymerto-carbon ratio. Figure 1 a and b present the scanning electron Results and Discussion microscopy (SEM) images of the cross-section of the PEG0 and In order to modulate the ionomer distribution in the catalyst PEG30 catalyst layers. These catalyst layers with various PEG/ layer, we varied the weight ratio of the PEG and Aquivion ionionomer ratios did not exhibit apparent differences in their omer from 0 to 0.4. It should be noted that the PEG/ionomer SEM images as shown in Figure 1 a, b, and S3 of the Supportratios higher than 0.4 lead to losses of mechanical integrity in ing Information. The typical porous structure of the Pt/C based the ionomer phase after the PEG removal. The optical transparcatalyst layer was maintained with the PEG addition, which inency of the blend films in these PEG/ionomer ranges (Figdicates that the PEG does not interrupt the evolution of the ure S1, Supporting Information) indicates compatibility besecondary pores. tween the two polymers. The ratio of Pt/C to ionomer was According to the TEM images of the PEG30 catalyst layer fixed at 7:3 irrespective of the PEG content, as maximum (Figure 1 c and d) prepared from diluted catalyst ink, PEG dopower performance was obtained at this ratio for the catalyst mains with a sizes of 50 ~ 100 nm were generated in the catalayers with zero PEG content (Figure S2, Supporting Informalyst layer. For guidance, the borders of the PEG phase and Pt tion). The resulting catalyst layers are denoted as PEG0, PEG20, particles are highlighted in the image. The lattice spacing of PEG30, and PEG40 for the PEG/ionomer ratio of 0, 0.2, 0.3, and the SSC-ionomer phase observed in the TEM image of PEG30 0.4, respectively. The addition of PEG increases the volume of (Figure 1 d and S4 b of the Supporting Information) is 3.4 the blended polymer phase; the PEG/ionomer ratios of 0, 0.2, which can be assigned to the intermolecular spacing of the 0.3, and 0.4 correspond to a percent volume ratio of the blendmain chain, referring to the previously reported value (5 ) for 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&2&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org
the LSC-ionomers.[10] The intermolecular spacing of the main tion and H-desorption on Pt surfaces are clearly detected in chain is increased with the decrease in equivalent weight for the potential range of 0.07 V to 0.30 V. The peak at 0.82 V in the LSC-ionomers,[10a] which indicates that the side chain influthe anodic scan and the peak at 0.80 V in the cathodic scan ences the intermolecular spacing of the main chain. In this correspond to the Pt-oxidation and the Pt-oxide-reduction, reregard, the shorter intermolecular spacing for the SSC-ionomer spectively. The values for the ECSA and Cdl can be determined probably originates from the shorter side chain length. The from the amount of charge for the H-desorption and the caPEG domains can be differentiated from the Pt particles by the pacitive charging current at 0.4 V, respectively.[11] As shown in absence of the characteristic metallic lattice of Pt. In particular, Figure 2 and listed in Table 1, in the PEG/ionomer range of 0 to the lattice distance of 0.23 nm indicated in Figure 1 d and Figure S4d of the Supporting InforTable 1. Electrochemical parameters of the catalyst layers with various PEG/ionomer ratios. mation denotes the interplanar spacing of (111)Pt. The PEG doPt utilization Cdl Rcl Power density j0 Tafel slope Sample ECSA [%] [mF cm2] [W cm2] at 0.6 V [mW cm2] [mA cm2] [mV dec1] [m2 g1] mains appear to be embedded in the ionomer phase; the ionPEG0 14.1 46.2 28.5 0.37 108.8 3.61 104 68.9 69.0 PEG20 17.0 55.7 30.5 0.29 172.6 4.86 104 omer phase that covers the PEG 4 68.1 PEG30 17.6 57.8 39.5 0.16 188.5 5.87 10 domain is identified, as illustratPEG40 16.3 53.4 34.1 0.18 149.6 2.35 104 65.4 ed in Figure 1 d. More images that support the existence of the PEG domain are presented in Figure S4 of the Supporting Information. The PEG, which is 0.3, both ECSA and Cdl increased with the PEG/ionomer ratio easily soluble in water, leached out during the break-in and and they decreased with a further increase of the ratio up to cell operation. The disappearance of the PEG phase in the TEM 0.4. The percent ratio of the ECSA to the actual Pt surface image after soaking in water at room temperature for 24 h (30.5 m2 g1) was 46.2 %, 55.7 %, 57.8 %, and 53.4 % for PEG0, supports a facile PEG removal (Figure S4). For simplification PEG20, PEG30, and PEG40, respectively. The values for Cdl are and clarity, the morphology evolution for the PEG/ionomer catin good agreement with those for ECSA, which confirms the alyst layers is schematically represented in Figure 1 e. In the abmodulated ionomer distribution. It should be noted that, for sence of PEG, the ionomer aggregate and catalyst agglomerate PEG30, the ionomer coverage could be enhanced by 25 % partially covered the catalyst layer with an ionomer film, based on the ECSA and by 38 % based on the Cdl without alterwhereas, in the presence of PEG, the ionomer/PEG blend coving the ionomer content. This is conclusive evidence for the ered a larger portion of the catalyst layer and the removal of modulation of the ionomer distribution using this approach. the PEG left an expanded and more connected ionomer film. The existence of an optimum PEG content is not unexpected As the first evidence for the modulation of the ionomer disbecause the amount of Pt surface in contact with the PEG tribution, the PEG content-dependent variations of the electrodomain increases with the PEG/ionomer ratio, which negatively chemically active surface area (ECSA) and double layer capaciinfluences the ionomer coverage. tance (Cdl), which are known to be sensitive to the ionomer Further evidence for the modulated ionomer distribution is the change in the proton transport through the catalyst layer. coverage of Pt/C surface, are provided using the cyclic voltamIt is expected that the ionomer phases may be more connectmetry (CV) technique. As presented in Figure 2, the H-adsorped when the ionomers are expanded through the PEG addition. Therefore, the impedances under a H2/N2 (anode/cathode) atmosphere were measured at 120 8C and RH 40 % in order to eliminate the effect of the Faradaic reaction.[12] As shown in Figure 3 a, the linear region of 458, which corresponds to the ionic migration through the ionomer phases in the catalyst layer, was observed at high frequencies for all catalyst layers. The following inclined spikes at low frequencies were attributed to the total capacitance and resistance of the catalyst layer. The value for the proton transport resistance through the catalyst layer (Rcl) could be determined from the length of the Warburg-like region (Rcl/3) projected onto Z’-Rohm, when the low frequency line had a phase angle of 908. The low frequency lines slightly deviated from the 908 line for these catalyst layers as shown in Figure 3 a, which was attributed to a non-homogeneous distribution of ionic resistance across the catalyst layer.[12, 13] For simplicity, an approximate value for Rcl was obtained from the intercept on Z’ of an asymptotic line extending Figure 2. CV curves for the catalyst layers with various PEG/ionomer ratios from the low frequency line, as indicated in Figure 3 a. The measured in a H2/N2 (anode/cathode) atmosphere at 120 8C. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&3&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS
www.chemsuschem.org phological features of the pore and ionomer film. This consideration was the motivation for conducting a limiting current test that provides a quantitative measure of the oxygen transport resistance in the catalyst layer.[15] The flux of oxygen (NO2), is proportional to the concentration difference of oxygen in the gas channel (CO2) and at the Pt surface (CO2,Pt), and it can be related to the current density using Faraday’s law as follows: N O2 ¼
j 1 C CO2 ;Pt ¼ 4F rtotal O2
ð1Þ
where F is the faraday constant. At the limit current density (jlim), the O2 concentration at the Pt surface is zero. Consequently, the total oxygen transport resistance (rtotal), which is the reciprocal of the total oxygen transport coefficient, can be evaluated from the jlim using Equation (2), as follows: rtotal ¼ CO2
4F jlim
ð2Þ
The rtotal is composed of the transport resistance through the macropores of diffusion media and catalyst layer (rmacro), through the mesopores in the catalyst layer (rmeso), and through the ionomer film (rionomer) covering the Pt catalyst surfaces, as follows: r total ¼ r macro þ r meso þ r ionomer
Figure 3. a) Nyquist plots of the impedances for the catalyst layers with various PEG/ionomer ratios measured under a H2/N2 (anode/cathode) atmosphere at 120 8C, and b) plot of Rcl and proton conductivity of the catalyst layer as a function of the PEG/ionomer ratios.
values for Rcl were 0.37, 0.29, 0.16, and 0.18 W cm2 for PEG0, PEG20, PEG30, and PEG40, respectively. The analysis results are summarized in Figure 3 b, and these indicate a two-fold increase in the proton transport rate for PEG30 compared with that for PEG0. The proton conductivity of the catalyst layer calculated from the Rcl and thickness of the catalyst layer is also illustrated in Figure 3 b. This explains the evolution of a more connected ionomer morphology with the addition of PEG. It is well known that the proton conductivity increases as the ionomer content of the catalyst layer is increased, but the catalyst sites become blocked and gas cannot pass through the catalyst layer.[9b, 14] Thus, the modulation of the ionomer distribution in this paper is meaningful because the ion conductivity of the catalyst layer was improved although the ionomer content was fixed. The structures of the pore and ionomer film generally govern the oxygen transport property in the catalyst layer, and PEG addition should be accompanied by changes in the mor 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ð3Þ
Then, rmacro is determined from the difference in the rtotal on switching the balance gas from N2 to He, as described in previous studies.[15b, 16] The oxygen transport resistance of the agglomerate of catalyst and ionomer (rmeso + rionomer) can be quantified by subtracting rmacro from rtotal. As demonstrated in Figure 4, the oxygen transport resistance of the agglomerate dominates the total transport resistance. The comparison of rmeso + rionomer for the catalyst layers (Figure 4)
Figure 4. Plots of rtotal (crossed symbols), rmacro (open symbols), and rionomer + rmeso (filled symbols) as a function of PEG/ionomer ratio.
ChemSusChem 0000, 00, 1 – 8
&4&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS clearly demonstrates that the modulated ionomer distribution does not retard the oxygen transport in the agglomerate at 120 8C. The retarded oxygen transport that results from the expanded area of ionomer film is compensated by the fast oxygen transport through the pores generated by the PEG removal, which leads to the weak dependency of the oxygen transport property on PEG content. This is strong evidence for the PEG removal during fuel cell operation, because the PEG phases in the catalyst layer function as an additional barrier for the oxygen transport, which leads to an increase in rionomer if they are not removed. Figure 5 a presents the IV polarizations for the catalyst layers obtained at 120 8C, 150 kPa, and RH 40 %. The operation at 120 8C is significant for automotive applications because it enables the use of radiators similar to those used in modern com-
www.chemsuschem.org an increase in the power density for the PEG/ionomer ratios of 0 to 0.3, followed by a decrease for ratios of 0.3 to 0.4. This result strongly suggests that the difference in activation polarization significantly contributes to the difference in the power performance. For a quantitative assessment of the influence of the activation polarization on the power performance, the IRcorrected cell voltages (EiRfree) were examined. By correcting the raw voltage data with the values of ohmic resistance (Rohm), which were determined from the impedance at 1 kHz, the ohmic polarization can be eliminated, as given in Equation (4). At very low current densities at which the activation polarization dominates the overall reaction, EiRfree can be expressed in terms of the reversible cell voltage (Er), exchange current density (j0), and hydrogen crossover current (jx), as follows:
EiRfree ¼ E þ jRohm
Figure 5. a) IV polarization curves and b) Tafel plots of the MEAs with various PEG/ionomer ratios (temperature: 120 8C, humidity: RH 40 % for both cathode and anode feeds, back pressure: 150 kPa for both cathode and anode).
bustion engines, whereas operation at 80 8C requires double the volume of the radiator.[17] In the automotive operation condition, the power performance was improved with the PEG addition in all current densities investigated. The power performance at 0.6 V was used to evaluate the cell performance (Table 1). Interestingly, the power performance of PEG30 was 73 % higher than that of PEG0. In accordance with the trend observed for the ECSA and Cdl, the IV measurements revealed 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2:303RT j þ jx ¼ Er log anF j0
ð4Þ
where R, T, and F have standard meanings, and n and a are the number of electrons transferred and the charge transfer coefficient, respectively. The values for jx were determined from the linear sweep voltammetry measurement under a H2/ N2 (anode/cathode) atmosphere. Figure 5 b presents the IR-corrected cell voltage versus the hydrogen crossover-corrected current density (jeff = j + jx, jx = 1.5–1.7 mA cm2). Departure from the Tafel slope, which is indicative of a mass transport polarization for the oxygen reduction reaction, was observed above 0.06 A cm2. To avoid the effect of mass transport on Tafel analysis, we calculated the Tafel slope and exchange current density from the data in the range of 0.003 ~ 0.02 A cm2. In the activation polarization region (E > 0.80 V), the apparent Tafel slopes for all MEAs were similar (65–69 mV dec1), which implies the same inherent electrode kinetics toward the oxygen reduction reaction. The values for j0, which were also determined from Figure 5 b, are listed in Table 1. In accordance with the variation of the ECSA with the PEG/ionomer ratio, j0, which is proportional to ECSA,[9b, 14] increased with the PEG/ionomer ratio up to 0.3, but decreased with further increases in the PEG/ionomer ratio. Considering all data for the j0, proton transport and oxygen transport, the increased power performance with the PEG addition up to a PEG/ionomer ratio of 0.3 was attributed to the increased ECSA and facilitated proton transport through the catalyst layer. The impedance analysis also exhibited significant reductions in the charge-transfer resistance in a wide range of current density with the PEG addition, and the appearance of a minimum in the charge-transfer resistance at the PEG/ionomer ratio of 0.3 (Figure S5); this is in good agreement with the results of the polarization curves. Therefore, the appearance of the maximum power performance in PEG30 results from the maximum ECSA and proton conductivity of the catalyst layer at the PEG/ionomer ratio of 0.3. Because the proton transport and oxygen transport properties of PEG40 were comparable to those of PEG30 as indicated by the limiting current measurements, the lower power performance of ChemSusChem 0000, 00, 1 – 8
&5&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS PEG40 compared with that of PEG30 should originate from the lower ECSA.
Conclusions We have demonstrated the modulation of the ionomer distribution through the addition of PEG in the catalyst layer and the corresponding improvement in fuel cell performance. The modulation enlarges the electrochemical active surface and enhances the proton transport properties of the catalyst layer, without a loss in mass transport. In particular, the incorporation of PEG at a PEG/ionomer ratio of 0.3 results in a 73 % increase in the power density at 0.6 V under automotive operation conditions in conjunction with a 24 % increase in the ionomer coverage on the Pt catalyst and a 100 % increase in proton conductivity of the catalyst layer. Therefore, the modulated ionomer distribution offers a new method for optimizing the performance of polymer electrolyte membrane fuel cells.
Experimental Section Preparation and morphological characterization of the catalyst layers For the catalyst ink using the SSC ionomer Aquivion as an electrode ionomer, a carbon-supported platinum catalyst [TEC10 V50E (45.8 wt % Pt), Tanaka Kikinzoku Kogyo, Japan] was dispersed in a binary solvent of isopropyl alcohol (IPA)/water (1:1 in volume); predetermined amounts of Aquivion dispersion D79–20BS (20 wt %, equivalent weight of 790 g eq1, Solvay Solexis, Italy) and PEG (MW 2000, Aldrich) were added to the dispersion. In this work, we employed the highly stable Pt/C catalyst with large Pt size of 9.2 nm in diameter to prevent rapid catalyst degradation by sintering or Pt dissolution at the high temperature. The weight ratio of the Pt/C and Aquivion ionomer was fixed at 7:3 in a dry state. The weight ratios of PEG to Aquivion were varied: 0, 0.2, 0.3, and 0.4 for the both anode and cathode catalyst layers. The mixtures underwent sonication for 1 h and were cast onto Kapton film (Dupont, USA) supported on a glass plate using a doctor blade (gap: 0.3 mm) at a constant speed of 15 mm s1, followed by drying at 50 8C for 24 h. The Pt loading level of the catalyst layer was carefully controlled to be within 0.42 0.02 mg cm2 for the cathode and anode catalyst layers. The cathode and anode catalyst layers (effective area of 25 cm2) coated on the Kapton film were transferred to each side of the Aquivion E87–05SX membrane (50 um, EW 870 g eq1) at a temperature of 160 8C and pressure of 1500 psi for 3 min in order to form the catalyst-coated membrane (CCM). The Kapton films were peeled from the laminate after cooling. The ionomer morphologies of the catalyst layers were investigated using scanning electron microscopy (SEM; Sirion, FEI, USA) and transmission electron microscopy (TEM; Tecnai F30, FEI, USA) operating at 300 kV. The time duration of beam exposure was carefully minimized to prevent the sample from beam damage.
www.chemsuschem.org Korea). The cells were typically operated at a total gas pressure of 150 kPa, a cell temperature of 120 8C, and a relative humidity (RH) of 40 % for both electrodes. The hydrogen gas was supplied to the anode at 500 cm3 min1 and air to the cathode at 1500 cm3 min1, which corresponds the stoichiometry of H2/air = 14.9/18.0 at 0.2 A cm2. The polarization curves were obtained by stepping up the current density, allowing the cell to stabilize, and measuring the cell voltage. One minute was spent at each current density. After daily measurements of three IV curves for successive two days as a break-in, the IV curves for detailed analysis were obtained. The IV curves were not varied after the second day. The H2 crossover current density was measured using a linear sweep voltammetry from 0.05 to 0.70 V at a scan rate of 1 mV s1 with a H2 (anode) and N2 (cathode) feed of 500 cm3 min1.
Electrochemical impedance spectroscopy and cycling voltammetry In order to measure the ionic resistance for the catalyst layer, electrochemical impedance spectroscopy (EIS) was conducted in the frequency range of 105 to 101 Hz using a Solartron 1470E potentiostat and a 1400 frequency analyzer under H2/N2 (anode/cathode = 500/1500 cm3 min1). The amplitude of the sinusoidal voltage signal was 10 mV. The cathode was used as a working electrode, and the anode as a counter electrode. It was assumed that the anode (H2 electrode) had negligible impedance relative to the cathode. The CV of the single cells under H2/N2 (anode/cathode = 50/ 0 cm3 min1) was conducted using a Biologic HCP-803 potentiostat in the potential range from 0.07 to 1.20 V vs. DHE at 50 mV s1. It was conducted in hydrogen purging for the anode and nitrogen purging for the cathode for 30 min prior to the CV measurements in order to remove any trace of oxygen in the cell.
Limiting current density measurement Linear sweep voltammetry (LSV) was performed in order to investigate the oxygen transport through the cathode at the cell temperature of 120 8C. The hydrogen gas (40 % RH) was supplied to the anode (reference electrode) and diluted O2 (1 % O2 in the balance of N2 or He, 40 % RH) was supplied to the cathode (working electrode) at a flow rate of 100 cm3 min1 and at atmospheric back pressure. The oxygen reduction current was recorded at a scan rate of 10 mV s1 from 1.2 to 0.1 V. The limiting current density was determined using an extrapolation of the currents in the range of 0.1 to 0.2 V to zero voltage.
Acknowledgements This work was supported by the New&Renewable Energy R&D Program (grant no. 20113020030020) under the Ministry of Knowledge Economy (MKE), Republic of Korea. This paper was also supported as a project of the Global PhD Fellowship supported by the National Research Foundation of Korea conducts from 2011.
IV polarization of the single cells A single cell was assembled using the CCM, a pair of gas diffusion media (39BC, SGL), a pair of gaskets, and a pair of graphite blocks with triple serpentine flow fields. The single cells were connected to a test station equipped with an electronic load (Won-A Tech, 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Keywords: catalyst layer · electrochemistry · fuel cells · ionomer distribution · PEMFCs [1] S. Holdcroft, Chem. Mater. 2014, 26, 381 – 393.
ChemSusChem 0000, 00, 1 – 8
&6&
These are not the final page numbers! ÞÞ
CHEMSUSCHEM FULL PAPERS [2] a) E. Passalacqua, F. Lufrano, G. Squadrito, A. Patti, L. Giorgi, Electrochim. Acta 2001, 46, 799 – 805; b) G. Sasikumar, J. W. Ihm, H. Ryu, Electrochim. Acta 2004, 50, 601 – 605; c) D. Song, Q. Wang, Z. Liu, M. Eikerling, Z. Xie, T. Navessin, S. Holdcroft, Electrochim. Acta 2005, 50, 3347 – 3358; d) M. Uchida, Y. Aoyama, N. Eda, A. Ohta, J. Electrochem. Soc. 1995, 142, 4143 – 4149. [3] R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan, N. Garland, D. Myers, M. Wilson, F. Garzon, D. Wood, P. Zelenay, K. More, K. Stroh, T. Zawodzinski, J. Boncella, J. E. McGrath, M. Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi, S. Miyata, A. Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K. Kimijima, N. Iwashita, Chem. Rev. 2007, 107, 3904 – 3951. [4] a) C. S. Kong, D.-Y. Kim, H.-K. Lee, Y.-G. Shul, T.-H. Lee, J. Power Sources 2002, 108, 185 – 191; b) X. Cheng, B. Yi, M. Han, J. Zhang, Y. Qiao, J. Yu, J. Power Sources 1999, 79, 75 – 81; c) J. M. Song, S. Suzuki, H. Uchida, M. Watanabe, Langmuir 2006, 22, 6422 – 6428. [5] a) K. Malek, M. Eikerling, Q. Wang, T. Navessin, Z. Liu, J. Phys. Chem. C 2007, 111, 13627 – 13634; b) K. Malek, T. Mashio, M. Eikerling, Electrocatalysis 2011, 2, 141 – 157. [6] T. Soboleva, K. Malek, Z. Xie, T. Navessin, S. Holdcroft, ACS Appl. Mater. Interfaces 2011, 3, 1827 – 1837. [7] a) K. Epping Martin, J. P. Kopasz, Fuel Cells 2009, 9, 356 – 362; b) K. A. Mauritz, R. B. Moore, Chem. Rev. 2004, 104, 4535 – 4586. [8] a) K. D. Kreuer, M. Schuster, B. Obliers, O. Diat, U. Traub, A. Fuchs, U. Klock, S. J. Paddison, J. Maier, J. Power Sources 2008, 178, 499 – 509; b) A. S. Aric, A. D. Blasi, G. Brunaccini, F. Sergi, V. Antonucci, P. Asher, S. Buche, D. Fongalland, G. A. Hards, J. D. B. Sharman, A. Bayer, G. Heinz, R. Zuber, M. Gebert, M. Corasaniti, A. Ghielmi, D. J. Jones, ECS Trans. 2009, 25, 1999 – 2007; c) M. Gebert, A. Ghielmi, L. Merlo, M. Corasaniti, V. Arcella, ECS Trans. 2010, 26, 279 – 283; d) C. Lei, D. Bessarabov, S. Ye, Z. Xie, S. Holdcroft, T. Navessin, J. Power Sources 2011, 196, 6168 – 6176.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org [9] a) T. Soboleva, X. Zhao, K. Malek, Z. Xie, T. Navessin, S. Holdcroft, ACS Appl. Mater. Interfaces 2010, 2, 375 – 384; b) P. Gode, F. Jaouen, G. Lindbergh, A. Lundblad, G. Sundholm, Electrochim. Acta 2003, 48, 4175 – 4187. [10] a) T. D. Gierke, G. E. Munn, F. C. Wilson, J. Polym. Sci. Polym. Phys. Ed. 1981, 19, 1687 – 1704; b) N. A. Nazir, H. Kudo, T. Nishikubo, T. Kyu, J. Mater. Sci. 2012, 47, 7269 – 7279. [11] L. Su, W. Jia, C.-M. Li, Y. Lei, ChemSusChem 2014, 7, 361 – 378. [12] M. C. Lefebvre, R. B. Martin, P. G. Pickup, Electrochem. Solid-State Lett. 1999, 2, 259 – 261. [13] D. Malevich, B. R. Jayasankar, E. Halliop, J. G. Pharoah, B. A. Peppley, K. Karan, J. Electrochem. Soc. 2012, 159, F888 – F895. [14] G. Li, P. G. Pickup, J. Electrochem. Soc. 2003, 150, C745 – C752. [15] a) A. Nanjundappa, A. S. Alavijeh, M. El Hannach, D. Harvey, E. Kjeang, Electrochim. Acta 2013, 110, 349 – 357; b) N. Nonoyama, S. Okazaki, A. Z. Weber, Y. Ikogi, T. Yoshida, J. Electrochem. Soc. 2011, 158, B416 – B423; c) T. A. Greszler, D. Caulk, P. Sinha, J. Electrochem. Soc. 2012, 159, F831 – F840. [16] D. R. Baker, D. A. Caulk, K. C. Neyerlin, M. W. Murphy, J. Electrochem. Soc. 2009, 156, B991 – B1003. [17] a) M. Mathias, H. Gasteiger, R. Makharia, S. Kocha, T. Fuller, T. Xie, J. Pisco, Prepr. Pap.-Am. Chem. Soc,. Div. Fuel Chem. 2004, 49, 471 – 474; b) J. Li, X. Li, Y. Zhao, W. Lu, Z. Shao, B. Yi, ChemSusChem 2012, 5, 896 – 900; c) J. Yang, D. Aili, Q. Li, L. N. Cleemann, J. O. Jensen, N. J. Bjerrum, R. He, ChemSusChem 2013, 6, 275 – 282.
Received: February 5, 2014 Revised: February 24, 2014 Published online on && &&, 0000
ChemSusChem 0000, 00, 1 – 8
&7&
These are not the final page numbers! ÞÞ
FULL PAPERS M.-J. Choo, K.-H. Oh, H.-T. Kim,* J.-K. Park* && – && Modulated Ionomer Distribution in the Catalyst Layer of Polymer Electrolyte Membrane Fuel Cells for High Temperature Operation
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Ionomer distribution modulation: A series of new catalyst layers for high temperature operation was fabricated through a simple poly(ethylene glycol) addition and leaching process. The ionomer domain was expanded by poly(ethylene glycol) addition, resulting in an increase in the electrochemical active area and proton transport without loss in oxygen transport, at a fixed ionomer content.
ChemSusChem 0000, 00, 1 – 8
&8&
These are not the final page numbers! ÞÞ
