- Email: [email protected]
Durable Marimo-like carbon support for Platinum nanoparticle catalyst in polymer electrolyte fuel cell
Electrochimica Acta 213 (2016) 447–451
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Durable Marimo-like carbon support for Platinum nanoparticle catalyst in polymer electrolyte fuel cell Koki Babaa , Mikka Nishitani-Gamob , Toshihiro Andoc , Mika Eguchia,* a b c
Department of Biomolecular Functional Engineering, College of Engineering, Ibaraki University, 4-12-1 Nakanarusawa, Hitachi, Ibaraki, 316-8511, Japan Department of Applied Chemistry, Graduate School of Engineering, Toyo University, 2100 Kujirai, Kawagoe, Saitama, 350-8585, Japan National Institute for Material Science, 1-1 Namiki, Tsukuba, Ibaraki, 305-0044, Japan
A R T I C L E I N F O
Article history: Received 18 January 2016 Received in revised form 4 July 2016 Accepted 4 July 2016 Available online 18 July 2016 Keywords: Polymer electrolyte fuel cells Catalyst support Electrochemical oxidation Carbon nanofilaments
A B S T R A C T
This study deals with the electrochemical stability of carbon-based catalyst supports as the cathode material of a polymer electrolyte fuel cell (PEFC). The platinum catalyst on the carbon black supports (Pt/ CB) was investigated for use in the PEFC. Marimo-like carbon (MC) was used for the catalyst support. Electrochemical surface oxidation of the Marimo-like carbon was compared to that of conventional carbon black during simulated start-stop cycles. We observed a different durability of the electrocatalytic activity between the carbon black and the Marimo-like carbon based on the CV measurements and the XPS. In case of the carbon black, the electrochemical active surface area (ECA) decreased and the electric double layer current density increased with increasing number of cycles in the accelerated degradation test (ADT). These results indicated oxidation of the carbon in the Pt/CB by the ADT. In case of the Marimolike carbon, the ECA decreased less than in the case of the carbon black, and no change was observed in the XPS. These results indicated that surface oxidation of the Marimo-like carbon proceeded less than that of the conventional carbon black. ã 2016 Elsevier Ltd. All rights reserved.
1. Introduction Polymer electrolyte fuel cells (PEFCs) can function with a high efficiency for fast starting motion devices. These properties allow the PEFC to be used as the power sources for automotive, stationary and portable applications [1,2]. To construct a smart grid system using a hydrogen fuel, efficient and durable PEFC systems are required. The hydrogen PEFC required a catalyst promoting chemical reactions in anode and cathode. Platinum is generally used as a catalyst. The Pt catalyst usually consists of Pt nanoparticles dispersed on the surface of carbon black (Pt/CB). The carbon blacks are well known to be a catalyst support for the PEFCs. In recent years, fibrous carbon materials are attracting attention due to their high electric conductivity and water repellency. Carbon blacks have indefinite structures and many fine and deep pores that produce large specific surface areas. However, the carbon black does not have a durable electrochemical property under the PEFC operation conditions. It has been reported that oxygen containing groups (e.g., carboxyl, carbonyl, hydroxyl, phenol, etc.) are formed on the carbon surface at the reaction
* Corresponding author. E-mail address: [email protected] (M. Eguchi). http://dx.doi.org/10.1016/j.electacta.2016.07.022 0013-4686/ã 2016 Elsevier Ltd. All rights reserved.
temperatures (>65 C) or operating potentials (>1.0 V vs. RHE) [3– 5]. The produced oxygen containing groups cause a decreased conductivity of the carbon and detachment of the Pt nanoparticles from the carbon surface [6–8]. These phenomena decrease the catalytic activity and drastically reduce the performance of the PEFC. To endure the catalytic electrode activity, it is essential to study the stability of the catalyst and carbon support structures. Some studies have reported durability of carbon materials. Zana et al. investigated the corrosion of high surface area carbons under the start/stop PEFC conditions [9]. Shanahan et al. reported that the graphitized carbon shows a higher stability as a catalyst support [10]. Yuan et al. described a high performance PEFC using fibrous graphite “carbon nanotubes” [11]. Recently, we used a new type of carbon material, i.e., the Marimo-like carbon (Fig. 1). We found that the Marimo-like carbon supported platinum catalyst improved the PEFC performance [12]. Especially, in the high current density range, the MEA efficiently worked. It caused the reactant gases and product water to easily diffuse between the carbon nanofilaments. The optimum ionomer content for the carbon black was about 30 wt%. However, in the case of using the Marimo-like carbom, the ionomer content of 10 wt% gave the best performance because the Marimo-like carbon has a lower surface area of inner pore than the carbon black. The Marimo-like carbon consists of many carbon
448
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
influenced by the Ni catalyst. The Pt nanoparticles were loaded on the Marimo-like carbon support by the modified nanocolloidal method (Pt/MC) [17,18]. The Pt nanoparticles were directly deposited on the carbon nanofilaments of the Marimo-like carbon from the solution. NaOH (2.76 mol l1, 5 ml) was added to deionized water (80 ml), followed by the addition of the Marimo-like carbon (100 mg). The carbon-included solution was stirred for 30 min and then irradiated with an ultrasonic vibration for 30 min. Next, H2PtCl66H2O (0.36 mol l1, 250 ml) and citric acid (0.56 mol l1, 5 ml) were added to the carbon-included solution. The solution was stirred for 30 min and then irradiated with an ultrasonic vibration for 30 min. Finally, NaBH4 (167.58 mol l1, 5 mL) was added to the solution and Pt complex ions were reduced to generate Pt metallic nanocolloidal particles. The yielded Pt particles were separated from the solution by centrifugation. After the supernatant solution was removed by the centrifugation, the precipitation products composed of the Pt particles on the Marimo-like carbon. 2.2. Electrochemical measurements
Fig. 1. SEM images of (a) the Marimo-like carbon and (b) the carbon nanofilaments constructed the Marimo-like carbon.
nanofilaments [13,14]. The carbon nanofilaments were interwoven to form a spherical secondary shape. The carbon nanofilaments have a small diameter and long length that form large specific surface areas. They have few pores inside, and their surface areas are due to the outside of the fibers. The fibrous structure simultaneously produces both an increasing surface area and increasing mass transportation. The carbon nanofilaments can contact each other to form a conductive network and provide a good electric conductivity. This structure should be favorable for supplying reactant gases and removing the product water. The carbon nanofilaments have a cup-stack primary structure. The edges of the graphene sheets are exposed to the carbon nanofilaments surface and they act as anchoring sites for the Pt nanoparticle catalyst. Marimo-like carbon does not require any heat pretreatments or any mechanical pretreatments for supporting the Pt particles on their surfaces. The Marimo-like carbon has a superior stability as a support material for the PEFC. In this study, we examined the durability of the Pt nanoparticle catalyst supported on the Marimo-like carbon (Pt/MC) to investigate changes in the electrochemical property and chemical structures during simulated start/stop cycles [15,16].
Catalyst inks (5.4 gcat l1) were prepared by mixing the 8.4 mg of catalysts (prepared 20 wt% Pt/MC or commercial 20 wt% Pt/CB (TANAKA HOLDINGS Co., Ltd., TEC10E20E)) with 0.043 ml of 5 wt% Nafion solution (E.I. Du Pont de Nemours & Co., Inc., DE521), 0.5 ml of reagent grade 2-propanol (Kanto Chemical Co.) and 1 ml of deionized water under ultrasonic irradiation. 20 ml of the catalyst ink was dropped onto a platinum disk electrode, then the electrode was dried in a nitrogen atmosphere at room temperature for 30 min. The prepared electrode on a platinum disk was used as the working electrode for a triple electrode cell system consisting of a platinum electrode as the counter electrode and a Ag/AgCl as the reference electrode. To remove the influence of oxygen, all the electrochemical measurements were carried out in a deoxygenated 0.1 mol l1 H2SO4 solution.
2. Experimental 2.1. Catalyst preparation We used the Marimo-like carbon for the catalyst support. The Marimo-like carbon was synthesized by the decomposition of hydrocarbons using the thermal chemical vapor deposition method [13]. The decomposition of methane was carried out in a flow reactor. Using 100 mg of an oxidized diamond supported Ni catalyst, 30 SCCM CH4 was introduced at 823 K and heat-treated for 3 h. Since the Ni catalyst was not detected by XPS and EDX analyses, we consider that cell performance and/or durability was not
Fig. 2. Procedure for evaluating degradation of Pt/MC and Pt/CB catalysts.
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
449
2.3. Accelerated degradation test Fig. 2 illustrates the procedure for the accelerated degradation test (ADT). In step 1, the electrodes were conditioned by a potential cycle at the scanning rate 200 mV s1 and the scanning range of 50 1200 mV vs. RHE. In step 2, the cyclic voltammetry (CV) was performed to evaluate the electrochemical active surface area (ECA) and electric double layer current density (EDC). The coulombic charge for the hydrogen adsorption was used to calculate the ECA of each electrode Eq. (1): ECSA ¼
Q Had ð0:21 ½PtÞ
ð1Þ
where QHad is the charge for hydrogen adsorption (mC cm2), [Pt] is the weight of platinum loaded on the electrode (mg cm2) and 0.21 (mC cm2) is the constant factor conventionally used for a clean polycrystalline platinum. The electric double layer current density, which corresponds to the ECA of carbon, was determined at 0.5 V vs. RHE. After the first CV measurement, the electrode material was analyzed by X-ray photoelectron spectroscopy (XPS) to evaluate the chemical structures of the initial carbon surface (Step 3). In step 4, the accelerated degradation test (ADT) of each catalyst was performed by a potential sweep cycling of a symmetrical triangular wave protocol at room temperature [19]. The potential at the working electrode was swept between 1.0 V and 1.5 V vs. RHE at a 0.5 V s1 scan rate. After every 150 potential cycles, the cyclic voltammetry was performed (Step 5). After the 11,100th cycle, the final cyclic voltammogram was recorded, and then the electrode material was analyzed by XPS to evaluate the chemical structures of the final carbon surface (Step 6). 3. Results and discussion No macro morphological change was observed by SEM. A little change was observed by TEM. Fig. 3 shows TEM images of Pt/MC(a, c) and Pt/CB(b, d) at before(a, b) and after(c, d) the ADT. Before the ADT, Pt particles were highly dispersed on the carbon nanofilament of the Marimo carbon as shown in Fig. 3 (a). The Pt/MC was prepared by the nanocolloidal solution method, which was modified in our laboratory. The method can form fine Pt nanoparticles on the carbon support with a high dispersion. Very little change was observed between the Pt/MC before and after the ADT. The Pt particle density decreased after the ADT. TEM images shown in Fig. 3 (a) and (c) are well dispersed and are almost spherical in shape. For Pt/MC, the MC support corroded mildly, and few Pt particles were found to be detached to the electrolyte solution. In case of Pt/CB (Fig. 3 (b) and (d)), TEM images shown the decreasing Pt particle density and the increasing particle size (to aggregate). For Pt/CB, the carbon support corroded severely, and many Pt particles were found to be detached to the electrolyte solution and generated the aggregate. Fig. 4 shows change in the cyclic voltammograms (CVs) during the ADT. Feature of the CVs in the Pt/MC (Fig. 4(a)) did not change and hydrogen-under potential deposition/desorption waves slightly decreased with the ADT. Additionally, the electric double layer current density changed little. On the contrary, a new peak appeared and increased at 0.6 V vs. RHE in the Pt/CB (Fig. 4(b)). Furthermore, peak intensities around 0.5–1.2 and 0.5–0.8 vs. RHE decreased with the ADT. In addition, the electric double layer current density increased during the ADT cycles. The electric double layer current densities are summarized in Table 1. The values were determined for 1 g of Pt. No change was observed in the case of the Pt/MC. However, in the case of the Pt/CB, the electric double layer current density
Fig. 3. TEM images of (a) Pt/MC before ADT, (b) Pt/CB before ADT, (c) Pt/MC after ADT and (d) Pt/CB after ADT.
increased by a factor of two after the ADT. The change in the current density of the Pt/MC was ca. 1/50 of the Pt/CB. Fig. 5 shows the decay in the ECA with the ADT. The Pt/MC retained higher ECA compared to the Pt/CB in the all cycle number, indicating that the Marimo-like carbon support has a higher durability than the conventional amorphous carbon black support. Fig. 6 shows change in the normalized ECA during the ADT. The half-life (cycle number decreasing to 50% of the initial ECA) was determined from the plot. The half-life of the Pt/MC was 300,000 cycles (extrapolated value), and the half-life of the Pt/CB was 11,100 cycles (measured value). The half-life of the Pt/MC was 30 times longer than that of the Pt/CB. These results also indicate a higher durability of the marimo-carbon support for the Pt particle catalyst. This characteristic property is attributable to the carbon nanofilament of the Marimo-like carbon. The carbon nanofilament was made of sp2 carbons and its structure was similar to the graphite stacking structure. It is well known that crystalline graphite has high chemical, mechanical and thermal stabilities which caused the higher durability in the Pt/MC catalyst against the ADT. The Marimo-like carbon was composed of graphitic carbon nanofilaments. The carbon nanofilaments were not oxidized by the redox reaction during the potential cycling. The conductivity and high structural stability of the Marimo-like carbon caused the high durability of the Pt/MC in the ADT protocol. Fig. 7 shows the polarization curves of Pt/MC and Pt/CB catalyst recorded at room temperature in O2 saturated H2SO4 solution using a rotating disk electrode (RDE) at room temperature with a sweep rate of 1 mV s1 and rotation speed of 2500 rpm. The RDE voltammetry was used to evaluate the ORR catalytic activity. The
450
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
Fig. 6. Retentions of ECA of (a) Pt/MC () and (b) Pt/CB (&) with ADT.
Fig. 4. Changes in cyclic voltammograms of (a) Pt/MC and (b) Pt/CB during ADT. 0 (long dashed line), 150 (short dashed line), 750 (dotted line), 1950 (chain line), 4050 (two dotted chain line), 11100 (solid line) cycles.
Table 1 Changes in electric double layer current (EDC) and electrochemical active surface area (ECA) of Pt/MC and Pt/CB catalysts with ADT.
Fig. 7. Linear sweep voltammograms obtained at RDE coated with Pt/MC () or Pt/ CB (&) at before (dashed line) and after (solid line) ADT.
Electric double layer current/A g1 ECA/m2 g1 Initial Pt/MC 5.46 Pt/CB 6.62
Final
Ratio (Final/Initial)
Initial Final Ratio (Final/Initial)
5.55 12.9
1.02 1.95
37 30
30 15
0.81 0.50
Fig. 5. Changes in ECA of (a) Pt/MC () and (b) Pt/CB (&) with ADT.
onset potential of ORR on the Pt/MC (circle symbol) is 0.57 V vs. Ag/ AgCl, whereas it is more negative (0.02 V vs. Ag/AgCl) than that on the Pt/CB. The onset potential involved in the ORR is an important parameter for evaluating the catalyst performance. The onset potential for ORR was defined as the potential giving 10 mA. This result indicates that the catalytic activity of Pt/MC is higher
than that of the Pt/CB. This improvement can be attributed to an improved catalyst-support binding that facilitate the electron transport on the Pt/MC electrode. Moreover, the onset potential of ORR (the catalytic activity) keep the decreasing order of Pt/MC > Pt/ CB after the ADT. It is shown the higher catalyst activity for ORR and the same or more high stability of Pt/MC under the simulated cathodic conditions. Fig. 8 shows X-ray photoelectron spectra (XPS) of the Pt/MC and Pt/CB before and after the ADT. The peak areas were normalized by the main peaks at 284.5 eV. Table 2 summarized the changes in the ratio of different chemical states of carbon during the ADT. Table 2 corrects the normalized peak area ratios of the different chemical states of oxygen on the carbons, calculated from the high resolution C1 s XPS for initial and final electrodes. In the case of Pt/MC, no change was observed in the XPS before and after the ADT. The peak area ratio at 286.0 eV and 287.8 eV were kept constant after the ADT. In the case of Pt/CB, a clear change was observed in the XPS before and after the ADT. The peak area ratio at 286.0 eV increased by 3 times of the initial peak area ratio after the ADT. This result suggests that the carbon black surface was oxidized and decomposed, and consequently, structural degradation of carbon black occurred during the ADT. The oxidation of the carbon black would cause a degradation in the electrochemical characteristic. This increased the electric double layer current density and decreased the ECA. The increase in the electric double layer current density and the growth of the new peak were due to oxidation of the carbon surface with the ADT. The electric double layer current density is proportional to the level of oxidation of the carbon surface. The new peak at 0.6 vs. RHE is assigned to the quinone-hydroquinon redox coupling process
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
451
This result revealed that the Marimo-like carbon has a superior oxidation-resistant property under the ADT conditions. The Marimo-like carbon consists of carbon nanofilaments which have a graphene stacked structure. Only the edges of the graphene sheets could be oxidized while the inside of the graphene could not be oxidized. The structure was more stable for oxidation and decomposition than an amorphous structure. The stability of the structure could suppress the decrease of the ECA during the ADT. The Marimo-like carbon has a superior durability for oxidation compared to the conventional carbon black. 4. Conclusion
Fig. 8. C1 s XPS spectra of (a) Pt/MC before ADT, (b) Pt/CB before ADT, (c) Pt/MC after ADT and (d) Pt/CB after ADT.
Table 2 XPS analysis of C1 s in Pt/MC and Pt/CB catalysts before and after ADT.
The electrochemical durability of the Pt/MC was confirmed to be higher than that of the Pt/CB during the ADT under a typical potential cycle protocol. The conventional Pt/CB was changed and corroded by the ADT after 11,100 cycles. The Pt/CB produced an increase in the electric double layer current density (EDC) and a significant decrease in the electrochemical active surface area (ECA). The Pt/MC maintained almost constant values of the EDC and the ECA during the ADT. The half-life of the ECA for the Pt/MC was 30 times longer (over 300,000 cycles) than that for the Pt/CB. The XPS analysis revealed that the Marimo-like carbon is highly resistant against the oxidation, which improved stability of the Pt/ MC during the ADT. The unique structure of the Marimo-like carbon endures a superior electrochemical property. References
Peak
Pt/MC
(Binding energy/eV)
Initial
Final
Initial
Final
284.5 (CC) 286.0 (CO) 287.8 (C¼O)
1 0.2 0.21
1 0.21 0.2
1 0.46 0.34
1 1.33 0.42
Pt/MC
[9,20]. The redox peaks indicated that surface functional groups such as alcohol and carbonyl groups are formed by electrochemical oxidation of the carbon surface during ADT. The redox peak increase of the Pt/MC was smaller than the Pt/CB. These results indicate that the Marimo-like carbon has a higher resistance against the surface oxidation than the conventional amorphous carbon black in the simulated start and stop durability test. In the Pt/MC, the currents in the hydrogen adsorption/desorption potential regions (0.05–0.3 V vs. RHE) and PtOx formation region (0.7–1.2 V vs. RHE) slightly decreased with the ADT. However, in the Pt/CB, those currents significantly decreased with the ADT. The ECA evaluated with the hydrogen adsorption wave (0.05–0.3 V vs. RHE) is summarized in Table 1. After the ADT, the decrease of the ECA and in the Pt/MC and the Pt/CB catalysts were 19.0% and 47.6%, respectively, suggesting that the Marimo-like carbon suppressed the ECA decay. Since the ADT was carried out under the oxidative condition of carbon support, not under the dissolution/redeposition condition of Pt (0.6–1.0 V vs. RHE), the decrease in the ECA and current of PtOx formation region are attributed to the decrease in the electrical conductivity by the carbon surface oxidation and/or detaching of the Pt particles from the carbon support.
[1] P. Trogadas, T.F. Fuller, P. Strasser, Carbon 75 (2014) 5. [2] B. Avasarala, R. Moore, P. Haldar, Electrochimica Acta 55 (2010) 4765. [3] J.P. Meyers, R.M. Darling, Journal of The Electrochemical Society 153 (2006) A1432. [4] S.D. Knights, K.M. Colbow, J. St-Pierre, D.P. Wilkinson, J. Power Sources 127 (2004) 127. [5] L.M. Roen, C.H. Paik, T.D. Jarvi, Electrochemical and Solid-State Letters 7 (2004) A19. [6] Y.Y. Shao, G. Yin, Y. Gao, P. Shi, Journal of The Electrochemical Society 153 (2006) A1093. [7] Y. Shao, G. Yin, J. Zhang, Y. Gao, Electrochimica Acta 51 (2006) 5853. [8] F. Coloma, A. Sepulveda-Escribano, F. Rodriguez-Reinoso, Journal of Catalysis 154 (1995) 299. [9] A. Zana, J. Speder, N.E.A. Reeler, T. Vosch, M. Arenz, Electrochimica Acta 114 (2013) 455. [10] P.V. Shanahan, L. Xu, C. Liang, M. Waje, S. Dai, Y.S. Yan, Journal of Power Sources 185 (2008) 423. [11] Y. Yuan, J.A. Smith, G. Goenaga, D. Liu, Z. Luo, J. Liu, Journal of Power Sources 196 (2011) 6160. [12] K. Baba, K. Iwasawa, M. Eguchi, Y. Kobayashi, M. Kobori, M. Nishitani-Gamo, T. Ando, Japanese Journal of Applied Physics 52 (2013) 06GD06. [13] K. Nakagawa, H. Oda, A. Yamashita, M. Okamoto, Y. Sato, H. Gamo, M. NishitanGamo, K. Ogawa, T. Ando, J. Mater. Sci 44 (2009) 221. [14] T. Ando, K. Nakagawa, M. Gamo, H. Oda, U. S. Patent 0,193,764 A1 (14, August, 2008). [15] M. Eguchi, S. Yamamoto, M. Kikuchi, K. Uno, Y. Kobayashi, M. Nishitani-Gamo, T. Ando, Journal of Surface Finishing Society of Japan 62 (2011) 179. [16] M. Eguchi, A. Okubo, S. Yamamoto, M. Kikuchi, K. Uno, Y. Kobayashi, M. Nishitani-Gamo, T. Ando, Journal of Power Sources 195 (2010) 5862. [17] G. Alvarez, F. Alcaide, O. Miguel, P.L. Cabot, M.V. Martinez-Huerta, J.L.G. Fierro, Electrochimica Acta 56 (2011) 9370. [18] M. Lee, M. Uchida, K. Okaya, H. Uchida, M. Watanabe, Electrochemistry 79 (2011) 381. [19] S. Mitsushima, S. Kawahara, K. Ota, N. Kamiya, Journal of Electrochemical Society 154 (2007) B153. [20] H. Choo, T. Kinumoto, M. Nose, K. Miyazaki, T. Abe, Z. Ogumi, Journal of Power Sources 185 (2008) 740.
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
Durable Marimo-like carbon support for Platinum nanoparticle catalyst in polymer electrolyte fuel cell Koki Babaa , Mikka Nishitani-Gamob , Toshihiro Andoc , Mika Eguchia,* a b c
Department of Biomolecular Functional Engineering, College of Engineering, Ibaraki University, 4-12-1 Nakanarusawa, Hitachi, Ibaraki, 316-8511, Japan Department of Applied Chemistry, Graduate School of Engineering, Toyo University, 2100 Kujirai, Kawagoe, Saitama, 350-8585, Japan National Institute for Material Science, 1-1 Namiki, Tsukuba, Ibaraki, 305-0044, Japan
A R T I C L E I N F O
Article history: Received 18 January 2016 Received in revised form 4 July 2016 Accepted 4 July 2016 Available online 18 July 2016 Keywords: Polymer electrolyte fuel cells Catalyst support Electrochemical oxidation Carbon nanofilaments
A B S T R A C T
This study deals with the electrochemical stability of carbon-based catalyst supports as the cathode material of a polymer electrolyte fuel cell (PEFC). The platinum catalyst on the carbon black supports (Pt/ CB) was investigated for use in the PEFC. Marimo-like carbon (MC) was used for the catalyst support. Electrochemical surface oxidation of the Marimo-like carbon was compared to that of conventional carbon black during simulated start-stop cycles. We observed a different durability of the electrocatalytic activity between the carbon black and the Marimo-like carbon based on the CV measurements and the XPS. In case of the carbon black, the electrochemical active surface area (ECA) decreased and the electric double layer current density increased with increasing number of cycles in the accelerated degradation test (ADT). These results indicated oxidation of the carbon in the Pt/CB by the ADT. In case of the Marimolike carbon, the ECA decreased less than in the case of the carbon black, and no change was observed in the XPS. These results indicated that surface oxidation of the Marimo-like carbon proceeded less than that of the conventional carbon black. ã 2016 Elsevier Ltd. All rights reserved.
1. Introduction Polymer electrolyte fuel cells (PEFCs) can function with a high efficiency for fast starting motion devices. These properties allow the PEFC to be used as the power sources for automotive, stationary and portable applications [1,2]. To construct a smart grid system using a hydrogen fuel, efficient and durable PEFC systems are required. The hydrogen PEFC required a catalyst promoting chemical reactions in anode and cathode. Platinum is generally used as a catalyst. The Pt catalyst usually consists of Pt nanoparticles dispersed on the surface of carbon black (Pt/CB). The carbon blacks are well known to be a catalyst support for the PEFCs. In recent years, fibrous carbon materials are attracting attention due to their high electric conductivity and water repellency. Carbon blacks have indefinite structures and many fine and deep pores that produce large specific surface areas. However, the carbon black does not have a durable electrochemical property under the PEFC operation conditions. It has been reported that oxygen containing groups (e.g., carboxyl, carbonyl, hydroxyl, phenol, etc.) are formed on the carbon surface at the reaction
* Corresponding author. E-mail address: [email protected] (M. Eguchi). http://dx.doi.org/10.1016/j.electacta.2016.07.022 0013-4686/ã 2016 Elsevier Ltd. All rights reserved.
temperatures (>65 C) or operating potentials (>1.0 V vs. RHE) [3– 5]. The produced oxygen containing groups cause a decreased conductivity of the carbon and detachment of the Pt nanoparticles from the carbon surface [6–8]. These phenomena decrease the catalytic activity and drastically reduce the performance of the PEFC. To endure the catalytic electrode activity, it is essential to study the stability of the catalyst and carbon support structures. Some studies have reported durability of carbon materials. Zana et al. investigated the corrosion of high surface area carbons under the start/stop PEFC conditions [9]. Shanahan et al. reported that the graphitized carbon shows a higher stability as a catalyst support [10]. Yuan et al. described a high performance PEFC using fibrous graphite “carbon nanotubes” [11]. Recently, we used a new type of carbon material, i.e., the Marimo-like carbon (Fig. 1). We found that the Marimo-like carbon supported platinum catalyst improved the PEFC performance [12]. Especially, in the high current density range, the MEA efficiently worked. It caused the reactant gases and product water to easily diffuse between the carbon nanofilaments. The optimum ionomer content for the carbon black was about 30 wt%. However, in the case of using the Marimo-like carbom, the ionomer content of 10 wt% gave the best performance because the Marimo-like carbon has a lower surface area of inner pore than the carbon black. The Marimo-like carbon consists of many carbon
448
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
influenced by the Ni catalyst. The Pt nanoparticles were loaded on the Marimo-like carbon support by the modified nanocolloidal method (Pt/MC) [17,18]. The Pt nanoparticles were directly deposited on the carbon nanofilaments of the Marimo-like carbon from the solution. NaOH (2.76 mol l1, 5 ml) was added to deionized water (80 ml), followed by the addition of the Marimo-like carbon (100 mg). The carbon-included solution was stirred for 30 min and then irradiated with an ultrasonic vibration for 30 min. Next, H2PtCl66H2O (0.36 mol l1, 250 ml) and citric acid (0.56 mol l1, 5 ml) were added to the carbon-included solution. The solution was stirred for 30 min and then irradiated with an ultrasonic vibration for 30 min. Finally, NaBH4 (167.58 mol l1, 5 mL) was added to the solution and Pt complex ions were reduced to generate Pt metallic nanocolloidal particles. The yielded Pt particles were separated from the solution by centrifugation. After the supernatant solution was removed by the centrifugation, the precipitation products composed of the Pt particles on the Marimo-like carbon. 2.2. Electrochemical measurements
Fig. 1. SEM images of (a) the Marimo-like carbon and (b) the carbon nanofilaments constructed the Marimo-like carbon.
nanofilaments [13,14]. The carbon nanofilaments were interwoven to form a spherical secondary shape. The carbon nanofilaments have a small diameter and long length that form large specific surface areas. They have few pores inside, and their surface areas are due to the outside of the fibers. The fibrous structure simultaneously produces both an increasing surface area and increasing mass transportation. The carbon nanofilaments can contact each other to form a conductive network and provide a good electric conductivity. This structure should be favorable for supplying reactant gases and removing the product water. The carbon nanofilaments have a cup-stack primary structure. The edges of the graphene sheets are exposed to the carbon nanofilaments surface and they act as anchoring sites for the Pt nanoparticle catalyst. Marimo-like carbon does not require any heat pretreatments or any mechanical pretreatments for supporting the Pt particles on their surfaces. The Marimo-like carbon has a superior stability as a support material for the PEFC. In this study, we examined the durability of the Pt nanoparticle catalyst supported on the Marimo-like carbon (Pt/MC) to investigate changes in the electrochemical property and chemical structures during simulated start/stop cycles [15,16].
Catalyst inks (5.4 gcat l1) were prepared by mixing the 8.4 mg of catalysts (prepared 20 wt% Pt/MC or commercial 20 wt% Pt/CB (TANAKA HOLDINGS Co., Ltd., TEC10E20E)) with 0.043 ml of 5 wt% Nafion solution (E.I. Du Pont de Nemours & Co., Inc., DE521), 0.5 ml of reagent grade 2-propanol (Kanto Chemical Co.) and 1 ml of deionized water under ultrasonic irradiation. 20 ml of the catalyst ink was dropped onto a platinum disk electrode, then the electrode was dried in a nitrogen atmosphere at room temperature for 30 min. The prepared electrode on a platinum disk was used as the working electrode for a triple electrode cell system consisting of a platinum electrode as the counter electrode and a Ag/AgCl as the reference electrode. To remove the influence of oxygen, all the electrochemical measurements were carried out in a deoxygenated 0.1 mol l1 H2SO4 solution.
2. Experimental 2.1. Catalyst preparation We used the Marimo-like carbon for the catalyst support. The Marimo-like carbon was synthesized by the decomposition of hydrocarbons using the thermal chemical vapor deposition method [13]. The decomposition of methane was carried out in a flow reactor. Using 100 mg of an oxidized diamond supported Ni catalyst, 30 SCCM CH4 was introduced at 823 K and heat-treated for 3 h. Since the Ni catalyst was not detected by XPS and EDX analyses, we consider that cell performance and/or durability was not
Fig. 2. Procedure for evaluating degradation of Pt/MC and Pt/CB catalysts.
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
449
2.3. Accelerated degradation test Fig. 2 illustrates the procedure for the accelerated degradation test (ADT). In step 1, the electrodes were conditioned by a potential cycle at the scanning rate 200 mV s1 and the scanning range of 50 1200 mV vs. RHE. In step 2, the cyclic voltammetry (CV) was performed to evaluate the electrochemical active surface area (ECA) and electric double layer current density (EDC). The coulombic charge for the hydrogen adsorption was used to calculate the ECA of each electrode Eq. (1): ECSA ¼
Q Had ð0:21 ½PtÞ
ð1Þ
where QHad is the charge for hydrogen adsorption (mC cm2), [Pt] is the weight of platinum loaded on the electrode (mg cm2) and 0.21 (mC cm2) is the constant factor conventionally used for a clean polycrystalline platinum. The electric double layer current density, which corresponds to the ECA of carbon, was determined at 0.5 V vs. RHE. After the first CV measurement, the electrode material was analyzed by X-ray photoelectron spectroscopy (XPS) to evaluate the chemical structures of the initial carbon surface (Step 3). In step 4, the accelerated degradation test (ADT) of each catalyst was performed by a potential sweep cycling of a symmetrical triangular wave protocol at room temperature [19]. The potential at the working electrode was swept between 1.0 V and 1.5 V vs. RHE at a 0.5 V s1 scan rate. After every 150 potential cycles, the cyclic voltammetry was performed (Step 5). After the 11,100th cycle, the final cyclic voltammogram was recorded, and then the electrode material was analyzed by XPS to evaluate the chemical structures of the final carbon surface (Step 6). 3. Results and discussion No macro morphological change was observed by SEM. A little change was observed by TEM. Fig. 3 shows TEM images of Pt/MC(a, c) and Pt/CB(b, d) at before(a, b) and after(c, d) the ADT. Before the ADT, Pt particles were highly dispersed on the carbon nanofilament of the Marimo carbon as shown in Fig. 3 (a). The Pt/MC was prepared by the nanocolloidal solution method, which was modified in our laboratory. The method can form fine Pt nanoparticles on the carbon support with a high dispersion. Very little change was observed between the Pt/MC before and after the ADT. The Pt particle density decreased after the ADT. TEM images shown in Fig. 3 (a) and (c) are well dispersed and are almost spherical in shape. For Pt/MC, the MC support corroded mildly, and few Pt particles were found to be detached to the electrolyte solution. In case of Pt/CB (Fig. 3 (b) and (d)), TEM images shown the decreasing Pt particle density and the increasing particle size (to aggregate). For Pt/CB, the carbon support corroded severely, and many Pt particles were found to be detached to the electrolyte solution and generated the aggregate. Fig. 4 shows change in the cyclic voltammograms (CVs) during the ADT. Feature of the CVs in the Pt/MC (Fig. 4(a)) did not change and hydrogen-under potential deposition/desorption waves slightly decreased with the ADT. Additionally, the electric double layer current density changed little. On the contrary, a new peak appeared and increased at 0.6 V vs. RHE in the Pt/CB (Fig. 4(b)). Furthermore, peak intensities around 0.5–1.2 and 0.5–0.8 vs. RHE decreased with the ADT. In addition, the electric double layer current density increased during the ADT cycles. The electric double layer current densities are summarized in Table 1. The values were determined for 1 g of Pt. No change was observed in the case of the Pt/MC. However, in the case of the Pt/CB, the electric double layer current density
Fig. 3. TEM images of (a) Pt/MC before ADT, (b) Pt/CB before ADT, (c) Pt/MC after ADT and (d) Pt/CB after ADT.
increased by a factor of two after the ADT. The change in the current density of the Pt/MC was ca. 1/50 of the Pt/CB. Fig. 5 shows the decay in the ECA with the ADT. The Pt/MC retained higher ECA compared to the Pt/CB in the all cycle number, indicating that the Marimo-like carbon support has a higher durability than the conventional amorphous carbon black support. Fig. 6 shows change in the normalized ECA during the ADT. The half-life (cycle number decreasing to 50% of the initial ECA) was determined from the plot. The half-life of the Pt/MC was 300,000 cycles (extrapolated value), and the half-life of the Pt/CB was 11,100 cycles (measured value). The half-life of the Pt/MC was 30 times longer than that of the Pt/CB. These results also indicate a higher durability of the marimo-carbon support for the Pt particle catalyst. This characteristic property is attributable to the carbon nanofilament of the Marimo-like carbon. The carbon nanofilament was made of sp2 carbons and its structure was similar to the graphite stacking structure. It is well known that crystalline graphite has high chemical, mechanical and thermal stabilities which caused the higher durability in the Pt/MC catalyst against the ADT. The Marimo-like carbon was composed of graphitic carbon nanofilaments. The carbon nanofilaments were not oxidized by the redox reaction during the potential cycling. The conductivity and high structural stability of the Marimo-like carbon caused the high durability of the Pt/MC in the ADT protocol. Fig. 7 shows the polarization curves of Pt/MC and Pt/CB catalyst recorded at room temperature in O2 saturated H2SO4 solution using a rotating disk electrode (RDE) at room temperature with a sweep rate of 1 mV s1 and rotation speed of 2500 rpm. The RDE voltammetry was used to evaluate the ORR catalytic activity. The
450
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
Fig. 6. Retentions of ECA of (a) Pt/MC () and (b) Pt/CB (&) with ADT.
Fig. 4. Changes in cyclic voltammograms of (a) Pt/MC and (b) Pt/CB during ADT. 0 (long dashed line), 150 (short dashed line), 750 (dotted line), 1950 (chain line), 4050 (two dotted chain line), 11100 (solid line) cycles.
Table 1 Changes in electric double layer current (EDC) and electrochemical active surface area (ECA) of Pt/MC and Pt/CB catalysts with ADT.
Fig. 7. Linear sweep voltammograms obtained at RDE coated with Pt/MC () or Pt/ CB (&) at before (dashed line) and after (solid line) ADT.
Electric double layer current/A g1 ECA/m2 g1 Initial Pt/MC 5.46 Pt/CB 6.62
Final
Ratio (Final/Initial)
Initial Final Ratio (Final/Initial)
5.55 12.9
1.02 1.95
37 30
30 15
0.81 0.50
Fig. 5. Changes in ECA of (a) Pt/MC () and (b) Pt/CB (&) with ADT.
onset potential of ORR on the Pt/MC (circle symbol) is 0.57 V vs. Ag/ AgCl, whereas it is more negative (0.02 V vs. Ag/AgCl) than that on the Pt/CB. The onset potential involved in the ORR is an important parameter for evaluating the catalyst performance. The onset potential for ORR was defined as the potential giving 10 mA. This result indicates that the catalytic activity of Pt/MC is higher
than that of the Pt/CB. This improvement can be attributed to an improved catalyst-support binding that facilitate the electron transport on the Pt/MC electrode. Moreover, the onset potential of ORR (the catalytic activity) keep the decreasing order of Pt/MC > Pt/ CB after the ADT. It is shown the higher catalyst activity for ORR and the same or more high stability of Pt/MC under the simulated cathodic conditions. Fig. 8 shows X-ray photoelectron spectra (XPS) of the Pt/MC and Pt/CB before and after the ADT. The peak areas were normalized by the main peaks at 284.5 eV. Table 2 summarized the changes in the ratio of different chemical states of carbon during the ADT. Table 2 corrects the normalized peak area ratios of the different chemical states of oxygen on the carbons, calculated from the high resolution C1 s XPS for initial and final electrodes. In the case of Pt/MC, no change was observed in the XPS before and after the ADT. The peak area ratio at 286.0 eV and 287.8 eV were kept constant after the ADT. In the case of Pt/CB, a clear change was observed in the XPS before and after the ADT. The peak area ratio at 286.0 eV increased by 3 times of the initial peak area ratio after the ADT. This result suggests that the carbon black surface was oxidized and decomposed, and consequently, structural degradation of carbon black occurred during the ADT. The oxidation of the carbon black would cause a degradation in the electrochemical characteristic. This increased the electric double layer current density and decreased the ECA. The increase in the electric double layer current density and the growth of the new peak were due to oxidation of the carbon surface with the ADT. The electric double layer current density is proportional to the level of oxidation of the carbon surface. The new peak at 0.6 vs. RHE is assigned to the quinone-hydroquinon redox coupling process
K. Baba et al. / Electrochimica Acta 213 (2016) 447–451
451
This result revealed that the Marimo-like carbon has a superior oxidation-resistant property under the ADT conditions. The Marimo-like carbon consists of carbon nanofilaments which have a graphene stacked structure. Only the edges of the graphene sheets could be oxidized while the inside of the graphene could not be oxidized. The structure was more stable for oxidation and decomposition than an amorphous structure. The stability of the structure could suppress the decrease of the ECA during the ADT. The Marimo-like carbon has a superior durability for oxidation compared to the conventional carbon black. 4. Conclusion
Fig. 8. C1 s XPS spectra of (a) Pt/MC before ADT, (b) Pt/CB before ADT, (c) Pt/MC after ADT and (d) Pt/CB after ADT.
Table 2 XPS analysis of C1 s in Pt/MC and Pt/CB catalysts before and after ADT.
The electrochemical durability of the Pt/MC was confirmed to be higher than that of the Pt/CB during the ADT under a typical potential cycle protocol. The conventional Pt/CB was changed and corroded by the ADT after 11,100 cycles. The Pt/CB produced an increase in the electric double layer current density (EDC) and a significant decrease in the electrochemical active surface area (ECA). The Pt/MC maintained almost constant values of the EDC and the ECA during the ADT. The half-life of the ECA for the Pt/MC was 30 times longer (over 300,000 cycles) than that for the Pt/CB. The XPS analysis revealed that the Marimo-like carbon is highly resistant against the oxidation, which improved stability of the Pt/ MC during the ADT. The unique structure of the Marimo-like carbon endures a superior electrochemical property. References
Peak
Pt/MC
(Binding energy/eV)
Initial
Final
Initial
Final
284.5 (CC) 286.0 (CO) 287.8 (C¼O)
1 0.2 0.21
1 0.21 0.2
1 0.46 0.34
1 1.33 0.42
Pt/MC
[9,20]. The redox peaks indicated that surface functional groups such as alcohol and carbonyl groups are formed by electrochemical oxidation of the carbon surface during ADT. The redox peak increase of the Pt/MC was smaller than the Pt/CB. These results indicate that the Marimo-like carbon has a higher resistance against the surface oxidation than the conventional amorphous carbon black in the simulated start and stop durability test. In the Pt/MC, the currents in the hydrogen adsorption/desorption potential regions (0.05–0.3 V vs. RHE) and PtOx formation region (0.7–1.2 V vs. RHE) slightly decreased with the ADT. However, in the Pt/CB, those currents significantly decreased with the ADT. The ECA evaluated with the hydrogen adsorption wave (0.05–0.3 V vs. RHE) is summarized in Table 1. After the ADT, the decrease of the ECA and in the Pt/MC and the Pt/CB catalysts were 19.0% and 47.6%, respectively, suggesting that the Marimo-like carbon suppressed the ECA decay. Since the ADT was carried out under the oxidative condition of carbon support, not under the dissolution/redeposition condition of Pt (0.6–1.0 V vs. RHE), the decrease in the ECA and current of PtOx formation region are attributed to the decrease in the electrical conductivity by the carbon surface oxidation and/or detaching of the Pt particles from the carbon support.
[1] P. Trogadas, T.F. Fuller, P. Strasser, Carbon 75 (2014) 5. [2] B. Avasarala, R. Moore, P. Haldar, Electrochimica Acta 55 (2010) 4765. [3] J.P. Meyers, R.M. Darling, Journal of The Electrochemical Society 153 (2006) A1432. [4] S.D. Knights, K.M. Colbow, J. St-Pierre, D.P. Wilkinson, J. Power Sources 127 (2004) 127. [5] L.M. Roen, C.H. Paik, T.D. Jarvi, Electrochemical and Solid-State Letters 7 (2004) A19. [6] Y.Y. Shao, G. Yin, Y. Gao, P. Shi, Journal of The Electrochemical Society 153 (2006) A1093. [7] Y. Shao, G. Yin, J. Zhang, Y. Gao, Electrochimica Acta 51 (2006) 5853. [8] F. Coloma, A. Sepulveda-Escribano, F. Rodriguez-Reinoso, Journal of Catalysis 154 (1995) 299. [9] A. Zana, J. Speder, N.E.A. Reeler, T. Vosch, M. Arenz, Electrochimica Acta 114 (2013) 455. [10] P.V. Shanahan, L. Xu, C. Liang, M. Waje, S. Dai, Y.S. Yan, Journal of Power Sources 185 (2008) 423. [11] Y. Yuan, J.A. Smith, G. Goenaga, D. Liu, Z. Luo, J. Liu, Journal of Power Sources 196 (2011) 6160. [12] K. Baba, K. Iwasawa, M. Eguchi, Y. Kobayashi, M. Kobori, M. Nishitani-Gamo, T. Ando, Japanese Journal of Applied Physics 52 (2013) 06GD06. [13] K. Nakagawa, H. Oda, A. Yamashita, M. Okamoto, Y. Sato, H. Gamo, M. NishitanGamo, K. Ogawa, T. Ando, J. Mater. Sci 44 (2009) 221. [14] T. Ando, K. Nakagawa, M. Gamo, H. Oda, U. S. Patent 0,193,764 A1 (14, August, 2008). [15] M. Eguchi, S. Yamamoto, M. Kikuchi, K. Uno, Y. Kobayashi, M. Nishitani-Gamo, T. Ando, Journal of Surface Finishing Society of Japan 62 (2011) 179. [16] M. Eguchi, A. Okubo, S. Yamamoto, M. Kikuchi, K. Uno, Y. Kobayashi, M. Nishitani-Gamo, T. Ando, Journal of Power Sources 195 (2010) 5862. [17] G. Alvarez, F. Alcaide, O. Miguel, P.L. Cabot, M.V. Martinez-Huerta, J.L.G. Fierro, Electrochimica Acta 56 (2011) 9370. [18] M. Lee, M. Uchida, K. Okaya, H. Uchida, M. Watanabe, Electrochemistry 79 (2011) 381. [19] S. Mitsushima, S. Kawahara, K. Ota, N. Kamiya, Journal of Electrochemical Society 154 (2007) B153. [20] H. Choo, T. Kinumoto, M. Nose, K. Miyazaki, T. Abe, Z. Ogumi, Journal of Power Sources 185 (2008) 740.