- Email: [email protected]
Electrocatalytic performance of Pt-based trimetallic alloy nanoparticle catalysts in proton exchange membrane fuel cells
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Electrocatalytic performance of Pt-based trimetallic alloy nanoparticle catalysts in proton exchange membrane fuel cells Bin Fang, Bridgid N. Wanjala, Jun Yin, Rameshwori Loukrakpam, Jin Luo, Xiang Hu, Jordan Last, Chuan-Jian Zhong* Department of Chemistry, State University of New York at Binghamton, Binghamton, NY 13902, USA
article info
abstract
Article history:
The electrocatalytic performance of nanoengineered PtNiFe catalysts in proton exchange
Received 3 February 2011
membrane fuel cells (PEMFC) is described in this report. The membrane electrode assembly
Received in revised form
was prepared using carbon-supported PtNiFe nanoparticles treated at two different
5 May 2011
temperatures as the cathode electrocatalysts in PEMFC. The PtNiFe/C catalysts were found
Accepted 11 May 2011
to exhibit excellent fuel cell performance, much better than that of commercial Pt/C
Available online 15 June 2011
catalyst. In addition to assessing the mass activities in the kinetic current region, the fuel cell performance was also determined in the ohmic and mass transport regions. The
Keywords:
electrocatalytic and fuel cell performance are shown to depend on the thermal treatment
Nanoengineered trimetallic
temperature of the trimetallic catalysts. The higher-temperature treated catalysts showed
catalysts
a higher power density than the lower-temperature treated catalysts. The results are also
Platinum-nickel-iron nanoparticles
discussed in terms of the effect of lattice shrinking in the trimetallic alloy nanoparticles on
Electrocatalysts
the electrocatalytic activity.
Oxygen reduction reaction
Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
PEM fuel cells
1.
Introduction
Fuel cells utilizing hydrogen as fuels represent an important form of tomorrow’s energy because hydrogen is an efficient fuel and is environmentally clean. Proton exchange membrane fuel cell (PEMFC) become attractive because of high conversion efficiency, low pollution, lightweight, high power density, and a wide range of applications from power sources in automobiles and space shuttles to power grids for buildings and factories. One of the major problems for the commercialization of the fuel cell driven vehicles is the high overall manufacturing cost of PEMFCs. The cost of the catalysts counts to 30% of the overall manufacturing cost because currently platinum is the only catalyst for both anodes and
reserved.
cathodes in PEMFCs [1,2]. The lowering of Pt-loading in the catalysts, the improvement of the utilization of noble metals, and the increase of the stability of catalysts are some of the current approaches to reducing the high cost of catalysts for the ultimate commercialization of PEMFCs. The lowering of the cathode loadings to about 0.4 mgPt/cm2 is often limited by the poor activity of Pt for the oxygen reduction reaction (ORR) at the cathode. One important area of research interests has been the design and development of active and robust Ptalloy catalysts [1]. The development of Pt-based multimetallic or alloy electrocatalysts is currently one promising area of finding effective solutions to the problem [2]. While there have been numerous studies of bimetallic or trimetallic catalysts for increasing the electrocatalytic ORR activities using rotating
* Corresponding author. E-mail address: [email protected] (C.-J. Zhong). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.05.066
4628
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
disc electrode method [3e15], including our own previous work [10e14], there have been limited reports on the fuel cell performance of the multimetallic or alloy catalysts [5,6,16e20]. The preparation of most existing multimetallic or alloy catalysts were based on traditional catalyst-preparation methods such as co-precipitation and impregnation, which are often not adequate for controlling size and composition of the catalysts. We have recently demonstrated a promising pathway for nano-engineering the size, shape, composition and structural properties of multimetallic nanoparticles (e.g., AuPt, PtNi, PtCo, PtVFe, PtNiFe, PtNiCo, etc.) as active and stable electrocatalysts for ORR in fuel cell cathode reaction [21e24]. Examples of the catalysts that have been demonstrated the viability for use in PEMFCs included nanoengineered AuPt and PtVFe catalysts [25e28]. For the trimetallic Pt-based nanoparticles prepared by our nanoengineered synthesis and processing [10,11], the high electrocatalytic activity of the catalysts for ORR is a result of the introduction of a second and third metals with smaller atomic sizes than platinum which produces a combination of reduction of the lattice distance, formation of metaleoxygen bond and adsorption of hydroxide groups, and the modification of the d-band centre. In this report, we describe the results of an investigation of the nanoengineered PtNiFe electrocatalysts in PEM fuel cells. The fuel cell performance of the catalysts in PEMFCs has been evaluated in terms of the electrocatalytic activity for ORR, and also in comparison with other nanoengineered bimetallic and trimetallic catalysts in PEMFCs.
2.
Experimental
The synthesis of PtNiFe nanoparticles followed the synthesis protocol reported previously [29]. Briefly, the synthesis involved mixing three metal precursors, PtII(acac)2, NiII(acac)2, and FeII(acac)2 in controlled molar ratios, which underwent reduction by 1,2-hexadecanediol in an octyl ether solvent in the presence of a mixture of oleylamine and oleic acid as capping agents. In several of our previous reports [10,11,29], we demonstrated how our synthetic approach was more effective than those based on traditional catalyst-preparation methods such as co-precipitation and impregnation methods in terms of a better size or composition control. The as-synthesized PtNiFe nanoparticles were supported on carbon black and subsequently thermally treated. Briefly, 140 mg carbon black (Vulcan XC-72R) and 66 mg of the assynthesized PtNiFe nanoparticles were used to produce the carbon-supported PtNiFe catalysts (PtNiFe/C). The thermal treatment involved first treatment at 280 C under 20% O2 to remove the organic capping molecules from the nanoparticles, and followed by treatment at 400e800 C under 15% H2 to form alloy phase [29]. The actual loading was determined by TGA method. Glassy carbon (GC) disks (geometric area: 0.196 cm2) were polished with 0.03 mm Al2O3 powders, followed by careful rinsing with deionized water. The geometric area of the GC electrode provides a measure of the loading of catalyst on the electrode surface used for the voltammetric characterization. The electrode was coated with the catalyst layer using
modified method from a previous report [10,11]. Briefly, a typical suspension of the catalysts was prepared by suspending 1.0 mg catalysts (PtNiFe/C) in 1 mL Millipore water with diluted (5% vol.) Nafion (5wt%, Aldrich). The suspension was then quantitatively transferred to the surface of the polished GC disk. The electrodes were dried overnight at room temperature. Membrane electrode assemblies (MEAs) (5 cm2 active area) used in this study were prepared by conventional catalyst coated substrate (CCS) method. The electrocatalyst-Nafion ink was painted on a wet-proofed carbon paper (Toray EC-TP1060T). The MEAs were prepared using Pt46Ni22Fe32/C catalyst (metal loading: 27% for 400 C sample and 37% for 800 C, 0.4 mgPt/cm2) for the cathode and Pt/C catalyst (20% Pt/C, Etek, 0.4 mgPt/cm2) for the anode. For comparison, MEAs were also prepared using Pt/C (20% Pt/C, E-tek, 0.4 mgPt/cm2) catalyst for both anode and cathode. The MEAs were prepared by hot pressing the sandwich structured Nafion 212 membrane (DuPont) and catalyst coated electrodes at 120 C. The MEAs were tested in a single-cell test station (Electrochem Inc.). The testing conditions included 100% humidified H2 and O2 at a flow rate of 100 mL/min, a back pressure of 30 psi for both electrodes, and an operating temperature of 75 C. Transmission electron microscopy (TEM) was performed on Hitachi H-7000 electron microscope (100 kV) to obtain the particle size and its distribution. X-ray Powder Diffraction (XRD) was used to study the lattice constants and particle sizes of the catalysts. XRD data was collected from 20 to 90 2q with a step size of 0.5 at room temperature on a Phillips X’pert PW 3040 MPD diffractometer using Cu Ka (l ¼ 1.5418 A). Thermogravimetric analysis (TGA) was performed on a PerkineElmer Pyris 1-TGA for determining the weight of organic shell. Typical samples weighed w4 mg and were heated in a platinum pan. Samples were heated in 20% O2 at a rate of 10 C/min. Inductively Coupled Plasma - Optical Emission Spectroscopy (ICP-OES) was used to analyze the composition, which was performed using a PerkineElmer 2000 DV ICP-OES utilizing a cross flow nebulizer with the following parameters: plasma 18.0 L Ar(g)/min; auxiliary 0.3 L Ar(g)/min; nebulizer 0.73 L Ar(g)/min; power 1500 W; peristaltic pump rate 1.40 mL/min. Reported values <1.0 mg/L were analyzed using a Meinhardt nebulizer coupled to a cyclonic spray chamber to increase analyte sensitivity at the following parameters: 18.0 L Ar(g)/min; auxiliary 0.3 L Ar(g)/min; nebulizer 0.63 L Ar(g)/min; power 1500 W; peristaltic pump rate 1.00 mL/min. Elemental concentrations were determined by measuring one or more emission lines (nm) to check for interferences: Pt 214.423, 203.646; V 309.311 and Co 238.892, 228.616. Concentrations for Pt were determined only on emission line 214.423 in the presence of Co due to spectral interferences. The nanoparticle samples were dissolved in concentrated aqua regia, and then diluted to concentrations in the range of 1e50 ppm for analysis. Multi-point calibration curves were made from dissolved standards with concentrations from 0 to 50 ppm in the same acid matrix as the unknowns. Laboratory check standards were analyzed every 6 or 12 samples, with instrument re-calibration if check standards were not within 5% of the initial concentration. Method
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
detection limits were determined using 1.0 mg/L of analyte in the same acid matrix and are reported in mg/L as follows: Pt <0.040 and V, Co - <0.022. Instrument reproducibility (n ¼ 10) determined using 1 mg/L elemental solutions resulted in <2% error for all three elements.
3.
Results and discussion
The average sizes of the carbon-supported PtNiFe nanoparticles after the thermal treatment were found to depend on the thermal treatment temperature. The catalysts described in this report showed an average size of 5.3 0.4 nm for the treatment temperature of 400 C and 7.8 0.6 nm for the treatment temperature of 800 C. The composition of the assynthesized and the supported PtNiFe nanoparticles were found to depend only on the synthesis condition. The trimetallic composition was determined by ICP-OES method. The catalysts described in this reported had a composition of Pt46Ni22Fe32. The actual loading of the nanoparticles on carbon-supported was determined by loading from TGA method. The catalysts described in this reported had an actual loading of 27% for 400 C sample and 37% for 800 C. The subtle differences in the actual loading reflects the difference of carbon burning at different temperatures for the thermal treatment. The mass activity for ORR was normalized against the actual mass loading. Fig. 1 shows a representative set of CV curves for the hydrogen adsorption/desorption and the platinum oxide oxidation/reduction characteristics in 0.1M HClO4 electrolyte for the Pt46Ni22Fe32/C catalysts treated at two different temperatures. In comparison with Pt/C, the Pt46Ni22Fe32/C catalysts showed similar waves, but less-resolved features in the hydrogen adsorption/desorption and the platinum oxide oxidation/reduction regions. The CV exhibit features characteristic of the hydrogen adsorption/desorption peaks in the potential region of
Fig. 1 e CV curves for Pt/C and Pt46Ni22Fe32/C catalysts treated at 400 C and 800 C. Electrode: glassy carbon electrode inked with the catalysts; Electrolyte: 0.1 M HClO4. Scan rate: 50 mV/s.
4629
0e0.4V (vs. RHE). It is notable that in the hydrogen adsorption region (between 0 and 0.4 V), PtNiFe/C catalysts showed very broad peak comparing to Pt/C. This difference likely reflects the different particle sizes, surfaces and defects among these catalysts. The integration of charges under the hydrogen adsorption/desorption waves in the 0e0.4 region allowed us to measure the electrochemical active area (ECA) of the catalysts. In comparison with the ECA value of 90.1 m2/gPt for Pt/C, the PtNiFe/C catalysts exhibited values of 53.4 m2/gPt and 26.5 m2/gPt for 400 C and 800 C treatments. The 800 C treated catalyst showed a smaller value of ECA, which is consistent with the size increase of the catalyst particles. Based on TEM data for the Pt46Ni22Fe32/C catalysts treated at the two different temperatures, there is an increase in the average size of the particles (from 5.3 0.4 nm at 400 C to 7.8 0.6 nm at 800 C). Fig. 2 shows a representative set of RDE curves for the ORR activities at the Pt46Ni22Fe32/C catalysts in 0.1 M HClO4 electrolyte. The measurement of the kinetics current at 0.9 V (vs. RHE) showed a value of 0.43 mA for Pt46Ni22Fe32/C 400 C and 0.44 mA for Pt46Ni22Fe32/C 800 C. This value translates to a Ptmass activity of 0.38 A/mgPt for the Pt46Ni22Fe32/C 400 C catalyst and 0.27 A/mgPt for the Pt46Ni22Fe32/C 800 C, which is higher than that for Pt/C (0.22 A/mgPt), indicating that there is a higher electrocatalytic activity for this catalyst. Considering the differences in ECA, the specific activities of the Pt46Ni22Fe32/C catalysts treated at 400 C and 800 C are 0.71 mA/cm2 and 1.02 mA/cm2, respectively, which are 3e5 times of Pt/C (0.24 mA/cm2). The trimetallic catalysts were evaluated in a PEM fuel cell to determine its fuel cell performance. Fig. 3 shows a representative set of fuel cell performance data for MEAs loaded with Pt46Ni22Fe32/C and Pt/C catalysts under the loading of 0.4 mgPt/ cm2. The data between our catalysts and those using the commercial Pt/C catalyst were compared. As shown in the
Fig. 2 e RDE curves for Pt/C and Pt46Ni22Fe32/C catalysts treated at 400 C and 800 C. Electrode: glassy carbon electrode inked with the catalysts. Electrolyte: 0.1 M HClO4 saturated with O2. Scan rate: 10 mV/s, and rotating speed: 1600 rpm.
4630
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
Fig. 3 e Polarization (close circles/squares) and power density (open circles/squares) curves of MEAs with Pt46Ni22Fe32/C catalysts (400 C and 800 C) or Pt/C (triangles) catalyst in the cathode in PEMFC at 75 C. Ptloading in both anode and cathode was 0.4 mgPt/cm2 for these MEAs.
polarization curve obtained under 75 C, the MEA with Pt/C catalyst exhibited a value of 0.52 V at 1.0 A/cm2. This value is largely comparable to those reported under similar operation conditions using CCS method for the MEA fabrication [6], thus validating the quality and effectiveness of our MEA preparation for the evaluation of the fuel cell performance in comparing Pt46Ni22Fe32/C (400 C and 800 C) and Pt/C catalysts. It is evident that both the cell voltage and the power density for the fuel cell with the Pt46Ni22Fe32/C catalysts in the cathode are higher than those for the Pt/C catalyst under the same test conditions. In the activation region, it is noted that catalyst treated at 400 C showed better performance than that treated at 800 C. For example, at 0.85V, the current density for Pt is 0.04 A/cm2, while PtNiFe 400 C showed 0.33 A/ cm2 and PtNiFe/C 800 C showed 0.22 A/cm2. Both PtNiFe catalysts showed 5e8 times higher current density than the Pt/C catalysts. This result is interesting considering the fact that in the RDE result the mass activity of the PtNiFe/C catalysts only showed less than two-fold improvement than that for the Pt/C catalyst. In the ohmic and transport region, the PtNiFe/C catalyst treated at 800 C showed a higher peak power density (1.21 W/cm2) than that for the PtNiFe/C catalyst treated at 400 C (0.62 W/cm2). Peak power density for Pt/C catalyst was found to be 0.52 W/cm2. The fuel cell with the 800 C treated catalyst showed a 130% increase in peak power density in comparison with that of Pt/C. These results demonstrate the excellent performance of the PtNiFe/C catalyst in PEM fuel cell, which is better than Pt/C catalyst. The finding is also consistent with the electrocatalytic activity trend revealed by the RDE data. In the kinetic or activation-controlled region, the values of current density at 0.9 V in the fuel cell (FC) data was used to calculate the mass activities of the catalysts, which were then compared with the mass activities determined by rotating disk electrode (RDE). In Fig. 4A, the mass activities obtained
Fig. 4 e (A) Comparison of the mass activities obtained form the kinetic region at 0.900 V from IR-free fuel cell (FC) IeEcell curves and the mass activities obtained form RDE curves at 0.900 V in 0.1 M HClO4 electrolyte for Pt/C and PtNiFe/C catalysts treated at 400 and 800 C. (BeC) Comparison of the electrocatalytic performance data of the PtNiFe/C catalyst treated at 800 C with several other catalysts [25e28]: the current density values at E [ 0.900 V from the actual fuel-cell’s IeEcell curves (B), and the peak power density values from the IeP curves (C).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
form the kinetic region at 0.900 V from IR-free fuel cell IeEcell curves and the mass activities obtained form RDE curves at 0.900 V in 0.1M HClO4 electrolyte are compared for the PtNiFe/ C catalysts treated at 400 and 800 C. In general, the mass activities obtained from RDE data and the fuel cell polarization data showed similar trends. However, there are subtle differences in absolute values, indicating that there are complex factors that influence the electrocatalytic performance of the catalysts in a real fuel cell, which is a subject of our further investigation. In Fig. 4BeC, the current densities and the peak power densities of the PtNiFe/C catalyst treated at 800 C is compared with several alloy catalysts studied in our previous work, including PtVFe/C [25] and AuPt catalysts [26]. The inclusion of the comparison with those for some of our other bimetallic/ trimetallic catalysts was to establish the improved performance of this PtNiFe/C catalyst in fuel cell. Based on the current density at E ¼ 0.9 V obtained from the actual IeEcell curves and the peak power density values obtained from the IP curves, the PtNiFe/C showed an enhanced electrocatalytic performance in the PEM fuel among these catalysts. It is important to note that the trend observed from comparing the mass activities between FC and RDE data in the kinetic region and the trend from the peak power densities in the ohmic and mass transport regions are not always consistent, reflecting the complexity in evaluating the electrocatalytic and fuel cell performance of the catalysts. The 800 C treated catalyst showed a lower mass activity than the 400 C treated catalyst based on the kinetic current data from both RDE and FC IeV curves. However, the 800 C treated catalyst showed a higher performance than the 400 C treated catalyst based on the IeV data in the ohmic and transport regions for the fuel cell. This complexity has in fact been observed for the fuel cell performance of other alloy catalysts treated differently [30,20]. Work is in progress to probe the structural properties of the trimetallic catalysts, aiming at establishing the correlation between the electrocatalytic performance in the fuel cell and the structural parameters such as size, composition, surface, and phase properties. Based on XRD data for the Pt46Ni22Fe32/C catalysts treated at the two different temperatures, both catalysts displayed fcc structure. However, a shift in peak positions (e.g., (111) peak) towards higher angles with an increase of annealing temperature was observed, which is attributed to the lattice shrinking with the thermal treatment temperature. The lattice constant for the catalyst treated at 400 C (0.3798 nm) was found to be, larger than that for the catalyst treated at 800 C (0.3780 nm). Therefore, we believe that the different catalytic activities for PtNiFe/C catalysts are at least partially associated with the different lattice parameter. The trend is consistent with the enhancement of the catalytic activity due to lattice shrinking effect, as demonstrated for other alloy catalysts [21e24].
4.
Conclusion
In conclusion, the nanoengineered trimetallic PtNiFe nanoparticle catalysts have been shown to exhibit excellent electrocatalytic performance in PEM fuel cells. The thermal treatment temperature of the catalysts has been shown to
4631
play an important role in enhancing the electrocatalytic performance in PEM fuel cells. The increase in thermal treatment temperature was found to lead to a better alloy of the trimetallic catalysts. The findings, upon further structural characterizations, will have important implications to the work is in progress to evaluate the durability of the catalysts in PEM fuel cells, and to the design and engineering of alloy catalysts for a wide range of different types of fuel cells [31,32]. While the focus of this report is to demonstrate the improved performance of the PtNiFe/C catalysts in fuel cells, we also carried out studies aimed at a detailed correlation between the atomic scale structure and the electrocatalytic performance. One important finding was that the higher mass activity observed for the lower-temperature treated catalyst was found to partially reflect Pt-enrichment on the surface sites, details of which are discussed in separate reports [33].
Acknowledgements The work was supported by the National Science Foundation(CBET-0709113), and in part from NYSERDA funding.
references
[1] Gasteiger HA, Kocha SS, Sompalli B, Wagner FT. Appl Catal B Environ 2005;56:9. [2] Zhong CJ, Luo J, Njoki P, Mott D, Wanjala B, Loukrakpam R, et al. Energy Environ Sci 2008;1:454. [3] Wei YC, Liu CW, Chang YW, Lai CM, Lim PY, Tsai LD, et al. Int J Hydrogen Energy 2010;35:1864. [4] Ramos-Sa´nchez G, Yee-Madeira H, Solorza-Feria O. Int J Hydrogen Energy 2008;33:3596. [5] Cho YH, Jeon TY, Lim JW, Cho YH, Ahnb M, Jung N, et al. Int J Hydrogen Energy 2011;36:4394. [6] Seo A, Lee J, Han K, Kim H. Electrochim Acta 2006;52:1603. [7] Ramos-Sa´nchez G, Solorza-Feria O. Int J Hydrogen Energy 2010;35:12105. [8] Khosravi Mohsen, Amini Mohammad K. Int J Hydrogen Energy 2010;35:10527. [9] Fox Elise B, Colon-Mercado Hector R. Int J Hydrogen Energy 2010;35:3280. [10] Luo J, Han L, Kariuki N, Wang L, Mott D, Zhong CJ, et al. Chem Mater 2005;17:5282. [11] Luo J, Kariuki N, Han L, Wang L, Zhong CJ, He T. Electrochim Acta 2006;51:4821. [12] Luo J, Njoki PN, Lin Y, Wang L, Mott D, Zhong CJ. Electrochem Comm 2006;8:581. [13] Luo J, Wang LY, Mott D, Njoki PN, Lin Y, He T, et al. Adv Mater 2008;20:4342. [14] He T, Kreidler E, Xiong L, Luo J, Zhong CJ. J Electrochem Soc 2006;153:A1637. [15] He T, Kreidler E, Xiong L, Ding E. J Power Sources 2007;165:87. [16] Xiong L, More KL, He Ting. J Power Sources 2010;195:2570. [17] Yu P, Pemberton M, Plasse P. J Power Sources 2005;144:11. [18] Tamizhmani G, Capuano GA. J Electrochem Soc 1994; 141:L132. [19] He Q, Mukerjee S. Electrochim Acta 2010;55:1709. [20] Mani P, Srivastava R, Strasser P. J Power Sources 2011;196:666. [21] Loukrakpam R, Luo J, He T, Chen YS, Xu Z, Njoki PN, et al. J Phys Chem C 2011;115:1682.
4632
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
[22] Wanjala BN, Luo J, Fang B, Mott D, Zhong CJ. 2011, 21, 4012. [23] Loukrakpam R, Chang P, Luo J, Fang B, Mott D, Bae IT, et al. Chem Comm 2010;46:7184. [24] Wanjala B, Luo J, Loukrakpam R, Njoki PN, Mott D, Zhong CJ, et al. J Phys Chem C 2010;114:17580. [25] Fang B, Luo J, Chen Y, Wanjala BN, Loukrakpam R, Hong J, et al. Chem Cat Chem 2011;3:421. [26] Fang B, Wanjala BN, Hu X, Last J, Loukrakpam R, Yin J, et al. J Power Sources 2011;196:659. [27] Fang B, Luo J, Njoki PN, Loukrakpam R, Wanjala BN, Hong J, et al. Electrochim Acta 2010;55:8230.
[28] Fang B, Luo J, Njoki PN, Loukrakpam R, Mott D, Wanjala B, et al. Electrochem Comm 2009;11:1139. [29] Luo J, Wang LY, Mott D, Njoki PN, Kariuki N, Zhong CJ, et al. J Mater Chem 2006;16:1665. [30] Trongchuankij W, Pruksathorn K, Hunsom M. Appl Energy 2011;88:974. [31] Bianchini C, Shen PK. Chem Rev 2009;109:4183. [32] Wang CY. Chem Rev 2004;104:4727. [33] Loukrakpam R, Wanjala BN, Yin J, Fang B, Luo J, Shao M, Protsailo L, Kawamura T, Chen Y, Petkov V, Zhong CJ. ACS Catalysis 2011;1:562.
Available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/he
Electrocatalytic performance of Pt-based trimetallic alloy nanoparticle catalysts in proton exchange membrane fuel cells Bin Fang, Bridgid N. Wanjala, Jun Yin, Rameshwori Loukrakpam, Jin Luo, Xiang Hu, Jordan Last, Chuan-Jian Zhong* Department of Chemistry, State University of New York at Binghamton, Binghamton, NY 13902, USA
article info
abstract
Article history:
The electrocatalytic performance of nanoengineered PtNiFe catalysts in proton exchange
Received 3 February 2011
membrane fuel cells (PEMFC) is described in this report. The membrane electrode assembly
Received in revised form
was prepared using carbon-supported PtNiFe nanoparticles treated at two different
5 May 2011
temperatures as the cathode electrocatalysts in PEMFC. The PtNiFe/C catalysts were found
Accepted 11 May 2011
to exhibit excellent fuel cell performance, much better than that of commercial Pt/C
Available online 15 June 2011
catalyst. In addition to assessing the mass activities in the kinetic current region, the fuel cell performance was also determined in the ohmic and mass transport regions. The
Keywords:
electrocatalytic and fuel cell performance are shown to depend on the thermal treatment
Nanoengineered trimetallic
temperature of the trimetallic catalysts. The higher-temperature treated catalysts showed
catalysts
a higher power density than the lower-temperature treated catalysts. The results are also
Platinum-nickel-iron nanoparticles
discussed in terms of the effect of lattice shrinking in the trimetallic alloy nanoparticles on
Electrocatalysts
the electrocatalytic activity.
Oxygen reduction reaction
Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
PEM fuel cells
1.
Introduction
Fuel cells utilizing hydrogen as fuels represent an important form of tomorrow’s energy because hydrogen is an efficient fuel and is environmentally clean. Proton exchange membrane fuel cell (PEMFC) become attractive because of high conversion efficiency, low pollution, lightweight, high power density, and a wide range of applications from power sources in automobiles and space shuttles to power grids for buildings and factories. One of the major problems for the commercialization of the fuel cell driven vehicles is the high overall manufacturing cost of PEMFCs. The cost of the catalysts counts to 30% of the overall manufacturing cost because currently platinum is the only catalyst for both anodes and
reserved.
cathodes in PEMFCs [1,2]. The lowering of Pt-loading in the catalysts, the improvement of the utilization of noble metals, and the increase of the stability of catalysts are some of the current approaches to reducing the high cost of catalysts for the ultimate commercialization of PEMFCs. The lowering of the cathode loadings to about 0.4 mgPt/cm2 is often limited by the poor activity of Pt for the oxygen reduction reaction (ORR) at the cathode. One important area of research interests has been the design and development of active and robust Ptalloy catalysts [1]. The development of Pt-based multimetallic or alloy electrocatalysts is currently one promising area of finding effective solutions to the problem [2]. While there have been numerous studies of bimetallic or trimetallic catalysts for increasing the electrocatalytic ORR activities using rotating
* Corresponding author. E-mail address: [email protected] (C.-J. Zhong). 0360-3199/$ e see front matter Copyright ª 2011, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ijhydene.2011.05.066
4628
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
disc electrode method [3e15], including our own previous work [10e14], there have been limited reports on the fuel cell performance of the multimetallic or alloy catalysts [5,6,16e20]. The preparation of most existing multimetallic or alloy catalysts were based on traditional catalyst-preparation methods such as co-precipitation and impregnation, which are often not adequate for controlling size and composition of the catalysts. We have recently demonstrated a promising pathway for nano-engineering the size, shape, composition and structural properties of multimetallic nanoparticles (e.g., AuPt, PtNi, PtCo, PtVFe, PtNiFe, PtNiCo, etc.) as active and stable electrocatalysts for ORR in fuel cell cathode reaction [21e24]. Examples of the catalysts that have been demonstrated the viability for use in PEMFCs included nanoengineered AuPt and PtVFe catalysts [25e28]. For the trimetallic Pt-based nanoparticles prepared by our nanoengineered synthesis and processing [10,11], the high electrocatalytic activity of the catalysts for ORR is a result of the introduction of a second and third metals with smaller atomic sizes than platinum which produces a combination of reduction of the lattice distance, formation of metaleoxygen bond and adsorption of hydroxide groups, and the modification of the d-band centre. In this report, we describe the results of an investigation of the nanoengineered PtNiFe electrocatalysts in PEM fuel cells. The fuel cell performance of the catalysts in PEMFCs has been evaluated in terms of the electrocatalytic activity for ORR, and also in comparison with other nanoengineered bimetallic and trimetallic catalysts in PEMFCs.
2.
Experimental
The synthesis of PtNiFe nanoparticles followed the synthesis protocol reported previously [29]. Briefly, the synthesis involved mixing three metal precursors, PtII(acac)2, NiII(acac)2, and FeII(acac)2 in controlled molar ratios, which underwent reduction by 1,2-hexadecanediol in an octyl ether solvent in the presence of a mixture of oleylamine and oleic acid as capping agents. In several of our previous reports [10,11,29], we demonstrated how our synthetic approach was more effective than those based on traditional catalyst-preparation methods such as co-precipitation and impregnation methods in terms of a better size or composition control. The as-synthesized PtNiFe nanoparticles were supported on carbon black and subsequently thermally treated. Briefly, 140 mg carbon black (Vulcan XC-72R) and 66 mg of the assynthesized PtNiFe nanoparticles were used to produce the carbon-supported PtNiFe catalysts (PtNiFe/C). The thermal treatment involved first treatment at 280 C under 20% O2 to remove the organic capping molecules from the nanoparticles, and followed by treatment at 400e800 C under 15% H2 to form alloy phase [29]. The actual loading was determined by TGA method. Glassy carbon (GC) disks (geometric area: 0.196 cm2) were polished with 0.03 mm Al2O3 powders, followed by careful rinsing with deionized water. The geometric area of the GC electrode provides a measure of the loading of catalyst on the electrode surface used for the voltammetric characterization. The electrode was coated with the catalyst layer using
modified method from a previous report [10,11]. Briefly, a typical suspension of the catalysts was prepared by suspending 1.0 mg catalysts (PtNiFe/C) in 1 mL Millipore water with diluted (5% vol.) Nafion (5wt%, Aldrich). The suspension was then quantitatively transferred to the surface of the polished GC disk. The electrodes were dried overnight at room temperature. Membrane electrode assemblies (MEAs) (5 cm2 active area) used in this study were prepared by conventional catalyst coated substrate (CCS) method. The electrocatalyst-Nafion ink was painted on a wet-proofed carbon paper (Toray EC-TP1060T). The MEAs were prepared using Pt46Ni22Fe32/C catalyst (metal loading: 27% for 400 C sample and 37% for 800 C, 0.4 mgPt/cm2) for the cathode and Pt/C catalyst (20% Pt/C, Etek, 0.4 mgPt/cm2) for the anode. For comparison, MEAs were also prepared using Pt/C (20% Pt/C, E-tek, 0.4 mgPt/cm2) catalyst for both anode and cathode. The MEAs were prepared by hot pressing the sandwich structured Nafion 212 membrane (DuPont) and catalyst coated electrodes at 120 C. The MEAs were tested in a single-cell test station (Electrochem Inc.). The testing conditions included 100% humidified H2 and O2 at a flow rate of 100 mL/min, a back pressure of 30 psi for both electrodes, and an operating temperature of 75 C. Transmission electron microscopy (TEM) was performed on Hitachi H-7000 electron microscope (100 kV) to obtain the particle size and its distribution. X-ray Powder Diffraction (XRD) was used to study the lattice constants and particle sizes of the catalysts. XRD data was collected from 20 to 90 2q with a step size of 0.5 at room temperature on a Phillips X’pert PW 3040 MPD diffractometer using Cu Ka (l ¼ 1.5418 A). Thermogravimetric analysis (TGA) was performed on a PerkineElmer Pyris 1-TGA for determining the weight of organic shell. Typical samples weighed w4 mg and were heated in a platinum pan. Samples were heated in 20% O2 at a rate of 10 C/min. Inductively Coupled Plasma - Optical Emission Spectroscopy (ICP-OES) was used to analyze the composition, which was performed using a PerkineElmer 2000 DV ICP-OES utilizing a cross flow nebulizer with the following parameters: plasma 18.0 L Ar(g)/min; auxiliary 0.3 L Ar(g)/min; nebulizer 0.73 L Ar(g)/min; power 1500 W; peristaltic pump rate 1.40 mL/min. Reported values <1.0 mg/L were analyzed using a Meinhardt nebulizer coupled to a cyclonic spray chamber to increase analyte sensitivity at the following parameters: 18.0 L Ar(g)/min; auxiliary 0.3 L Ar(g)/min; nebulizer 0.63 L Ar(g)/min; power 1500 W; peristaltic pump rate 1.00 mL/min. Elemental concentrations were determined by measuring one or more emission lines (nm) to check for interferences: Pt 214.423, 203.646; V 309.311 and Co 238.892, 228.616. Concentrations for Pt were determined only on emission line 214.423 in the presence of Co due to spectral interferences. The nanoparticle samples were dissolved in concentrated aqua regia, and then diluted to concentrations in the range of 1e50 ppm for analysis. Multi-point calibration curves were made from dissolved standards with concentrations from 0 to 50 ppm in the same acid matrix as the unknowns. Laboratory check standards were analyzed every 6 or 12 samples, with instrument re-calibration if check standards were not within 5% of the initial concentration. Method
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
detection limits were determined using 1.0 mg/L of analyte in the same acid matrix and are reported in mg/L as follows: Pt <0.040 and V, Co - <0.022. Instrument reproducibility (n ¼ 10) determined using 1 mg/L elemental solutions resulted in <2% error for all three elements.
3.
Results and discussion
The average sizes of the carbon-supported PtNiFe nanoparticles after the thermal treatment were found to depend on the thermal treatment temperature. The catalysts described in this report showed an average size of 5.3 0.4 nm for the treatment temperature of 400 C and 7.8 0.6 nm for the treatment temperature of 800 C. The composition of the assynthesized and the supported PtNiFe nanoparticles were found to depend only on the synthesis condition. The trimetallic composition was determined by ICP-OES method. The catalysts described in this reported had a composition of Pt46Ni22Fe32. The actual loading of the nanoparticles on carbon-supported was determined by loading from TGA method. The catalysts described in this reported had an actual loading of 27% for 400 C sample and 37% for 800 C. The subtle differences in the actual loading reflects the difference of carbon burning at different temperatures for the thermal treatment. The mass activity for ORR was normalized against the actual mass loading. Fig. 1 shows a representative set of CV curves for the hydrogen adsorption/desorption and the platinum oxide oxidation/reduction characteristics in 0.1M HClO4 electrolyte for the Pt46Ni22Fe32/C catalysts treated at two different temperatures. In comparison with Pt/C, the Pt46Ni22Fe32/C catalysts showed similar waves, but less-resolved features in the hydrogen adsorption/desorption and the platinum oxide oxidation/reduction regions. The CV exhibit features characteristic of the hydrogen adsorption/desorption peaks in the potential region of
Fig. 1 e CV curves for Pt/C and Pt46Ni22Fe32/C catalysts treated at 400 C and 800 C. Electrode: glassy carbon electrode inked with the catalysts; Electrolyte: 0.1 M HClO4. Scan rate: 50 mV/s.
4629
0e0.4V (vs. RHE). It is notable that in the hydrogen adsorption region (between 0 and 0.4 V), PtNiFe/C catalysts showed very broad peak comparing to Pt/C. This difference likely reflects the different particle sizes, surfaces and defects among these catalysts. The integration of charges under the hydrogen adsorption/desorption waves in the 0e0.4 region allowed us to measure the electrochemical active area (ECA) of the catalysts. In comparison with the ECA value of 90.1 m2/gPt for Pt/C, the PtNiFe/C catalysts exhibited values of 53.4 m2/gPt and 26.5 m2/gPt for 400 C and 800 C treatments. The 800 C treated catalyst showed a smaller value of ECA, which is consistent with the size increase of the catalyst particles. Based on TEM data for the Pt46Ni22Fe32/C catalysts treated at the two different temperatures, there is an increase in the average size of the particles (from 5.3 0.4 nm at 400 C to 7.8 0.6 nm at 800 C). Fig. 2 shows a representative set of RDE curves for the ORR activities at the Pt46Ni22Fe32/C catalysts in 0.1 M HClO4 electrolyte. The measurement of the kinetics current at 0.9 V (vs. RHE) showed a value of 0.43 mA for Pt46Ni22Fe32/C 400 C and 0.44 mA for Pt46Ni22Fe32/C 800 C. This value translates to a Ptmass activity of 0.38 A/mgPt for the Pt46Ni22Fe32/C 400 C catalyst and 0.27 A/mgPt for the Pt46Ni22Fe32/C 800 C, which is higher than that for Pt/C (0.22 A/mgPt), indicating that there is a higher electrocatalytic activity for this catalyst. Considering the differences in ECA, the specific activities of the Pt46Ni22Fe32/C catalysts treated at 400 C and 800 C are 0.71 mA/cm2 and 1.02 mA/cm2, respectively, which are 3e5 times of Pt/C (0.24 mA/cm2). The trimetallic catalysts were evaluated in a PEM fuel cell to determine its fuel cell performance. Fig. 3 shows a representative set of fuel cell performance data for MEAs loaded with Pt46Ni22Fe32/C and Pt/C catalysts under the loading of 0.4 mgPt/ cm2. The data between our catalysts and those using the commercial Pt/C catalyst were compared. As shown in the
Fig. 2 e RDE curves for Pt/C and Pt46Ni22Fe32/C catalysts treated at 400 C and 800 C. Electrode: glassy carbon electrode inked with the catalysts. Electrolyte: 0.1 M HClO4 saturated with O2. Scan rate: 10 mV/s, and rotating speed: 1600 rpm.
4630
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
Fig. 3 e Polarization (close circles/squares) and power density (open circles/squares) curves of MEAs with Pt46Ni22Fe32/C catalysts (400 C and 800 C) or Pt/C (triangles) catalyst in the cathode in PEMFC at 75 C. Ptloading in both anode and cathode was 0.4 mgPt/cm2 for these MEAs.
polarization curve obtained under 75 C, the MEA with Pt/C catalyst exhibited a value of 0.52 V at 1.0 A/cm2. This value is largely comparable to those reported under similar operation conditions using CCS method for the MEA fabrication [6], thus validating the quality and effectiveness of our MEA preparation for the evaluation of the fuel cell performance in comparing Pt46Ni22Fe32/C (400 C and 800 C) and Pt/C catalysts. It is evident that both the cell voltage and the power density for the fuel cell with the Pt46Ni22Fe32/C catalysts in the cathode are higher than those for the Pt/C catalyst under the same test conditions. In the activation region, it is noted that catalyst treated at 400 C showed better performance than that treated at 800 C. For example, at 0.85V, the current density for Pt is 0.04 A/cm2, while PtNiFe 400 C showed 0.33 A/ cm2 and PtNiFe/C 800 C showed 0.22 A/cm2. Both PtNiFe catalysts showed 5e8 times higher current density than the Pt/C catalysts. This result is interesting considering the fact that in the RDE result the mass activity of the PtNiFe/C catalysts only showed less than two-fold improvement than that for the Pt/C catalyst. In the ohmic and transport region, the PtNiFe/C catalyst treated at 800 C showed a higher peak power density (1.21 W/cm2) than that for the PtNiFe/C catalyst treated at 400 C (0.62 W/cm2). Peak power density for Pt/C catalyst was found to be 0.52 W/cm2. The fuel cell with the 800 C treated catalyst showed a 130% increase in peak power density in comparison with that of Pt/C. These results demonstrate the excellent performance of the PtNiFe/C catalyst in PEM fuel cell, which is better than Pt/C catalyst. The finding is also consistent with the electrocatalytic activity trend revealed by the RDE data. In the kinetic or activation-controlled region, the values of current density at 0.9 V in the fuel cell (FC) data was used to calculate the mass activities of the catalysts, which were then compared with the mass activities determined by rotating disk electrode (RDE). In Fig. 4A, the mass activities obtained
Fig. 4 e (A) Comparison of the mass activities obtained form the kinetic region at 0.900 V from IR-free fuel cell (FC) IeEcell curves and the mass activities obtained form RDE curves at 0.900 V in 0.1 M HClO4 electrolyte for Pt/C and PtNiFe/C catalysts treated at 400 and 800 C. (BeC) Comparison of the electrocatalytic performance data of the PtNiFe/C catalyst treated at 800 C with several other catalysts [25e28]: the current density values at E [ 0.900 V from the actual fuel-cell’s IeEcell curves (B), and the peak power density values from the IeP curves (C).
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
form the kinetic region at 0.900 V from IR-free fuel cell IeEcell curves and the mass activities obtained form RDE curves at 0.900 V in 0.1M HClO4 electrolyte are compared for the PtNiFe/ C catalysts treated at 400 and 800 C. In general, the mass activities obtained from RDE data and the fuel cell polarization data showed similar trends. However, there are subtle differences in absolute values, indicating that there are complex factors that influence the electrocatalytic performance of the catalysts in a real fuel cell, which is a subject of our further investigation. In Fig. 4BeC, the current densities and the peak power densities of the PtNiFe/C catalyst treated at 800 C is compared with several alloy catalysts studied in our previous work, including PtVFe/C [25] and AuPt catalysts [26]. The inclusion of the comparison with those for some of our other bimetallic/ trimetallic catalysts was to establish the improved performance of this PtNiFe/C catalyst in fuel cell. Based on the current density at E ¼ 0.9 V obtained from the actual IeEcell curves and the peak power density values obtained from the IP curves, the PtNiFe/C showed an enhanced electrocatalytic performance in the PEM fuel among these catalysts. It is important to note that the trend observed from comparing the mass activities between FC and RDE data in the kinetic region and the trend from the peak power densities in the ohmic and mass transport regions are not always consistent, reflecting the complexity in evaluating the electrocatalytic and fuel cell performance of the catalysts. The 800 C treated catalyst showed a lower mass activity than the 400 C treated catalyst based on the kinetic current data from both RDE and FC IeV curves. However, the 800 C treated catalyst showed a higher performance than the 400 C treated catalyst based on the IeV data in the ohmic and transport regions for the fuel cell. This complexity has in fact been observed for the fuel cell performance of other alloy catalysts treated differently [30,20]. Work is in progress to probe the structural properties of the trimetallic catalysts, aiming at establishing the correlation between the electrocatalytic performance in the fuel cell and the structural parameters such as size, composition, surface, and phase properties. Based on XRD data for the Pt46Ni22Fe32/C catalysts treated at the two different temperatures, both catalysts displayed fcc structure. However, a shift in peak positions (e.g., (111) peak) towards higher angles with an increase of annealing temperature was observed, which is attributed to the lattice shrinking with the thermal treatment temperature. The lattice constant for the catalyst treated at 400 C (0.3798 nm) was found to be, larger than that for the catalyst treated at 800 C (0.3780 nm). Therefore, we believe that the different catalytic activities for PtNiFe/C catalysts are at least partially associated with the different lattice parameter. The trend is consistent with the enhancement of the catalytic activity due to lattice shrinking effect, as demonstrated for other alloy catalysts [21e24].
4.
Conclusion
In conclusion, the nanoengineered trimetallic PtNiFe nanoparticle catalysts have been shown to exhibit excellent electrocatalytic performance in PEM fuel cells. The thermal treatment temperature of the catalysts has been shown to
4631
play an important role in enhancing the electrocatalytic performance in PEM fuel cells. The increase in thermal treatment temperature was found to lead to a better alloy of the trimetallic catalysts. The findings, upon further structural characterizations, will have important implications to the work is in progress to evaluate the durability of the catalysts in PEM fuel cells, and to the design and engineering of alloy catalysts for a wide range of different types of fuel cells [31,32]. While the focus of this report is to demonstrate the improved performance of the PtNiFe/C catalysts in fuel cells, we also carried out studies aimed at a detailed correlation between the atomic scale structure and the electrocatalytic performance. One important finding was that the higher mass activity observed for the lower-temperature treated catalyst was found to partially reflect Pt-enrichment on the surface sites, details of which are discussed in separate reports [33].
Acknowledgements The work was supported by the National Science Foundation(CBET-0709113), and in part from NYSERDA funding.
references
[1] Gasteiger HA, Kocha SS, Sompalli B, Wagner FT. Appl Catal B Environ 2005;56:9. [2] Zhong CJ, Luo J, Njoki P, Mott D, Wanjala B, Loukrakpam R, et al. Energy Environ Sci 2008;1:454. [3] Wei YC, Liu CW, Chang YW, Lai CM, Lim PY, Tsai LD, et al. Int J Hydrogen Energy 2010;35:1864. [4] Ramos-Sa´nchez G, Yee-Madeira H, Solorza-Feria O. Int J Hydrogen Energy 2008;33:3596. [5] Cho YH, Jeon TY, Lim JW, Cho YH, Ahnb M, Jung N, et al. Int J Hydrogen Energy 2011;36:4394. [6] Seo A, Lee J, Han K, Kim H. Electrochim Acta 2006;52:1603. [7] Ramos-Sa´nchez G, Solorza-Feria O. Int J Hydrogen Energy 2010;35:12105. [8] Khosravi Mohsen, Amini Mohammad K. Int J Hydrogen Energy 2010;35:10527. [9] Fox Elise B, Colon-Mercado Hector R. Int J Hydrogen Energy 2010;35:3280. [10] Luo J, Han L, Kariuki N, Wang L, Mott D, Zhong CJ, et al. Chem Mater 2005;17:5282. [11] Luo J, Kariuki N, Han L, Wang L, Zhong CJ, He T. Electrochim Acta 2006;51:4821. [12] Luo J, Njoki PN, Lin Y, Wang L, Mott D, Zhong CJ. Electrochem Comm 2006;8:581. [13] Luo J, Wang LY, Mott D, Njoki PN, Lin Y, He T, et al. Adv Mater 2008;20:4342. [14] He T, Kreidler E, Xiong L, Luo J, Zhong CJ. J Electrochem Soc 2006;153:A1637. [15] He T, Kreidler E, Xiong L, Ding E. J Power Sources 2007;165:87. [16] Xiong L, More KL, He Ting. J Power Sources 2010;195:2570. [17] Yu P, Pemberton M, Plasse P. J Power Sources 2005;144:11. [18] Tamizhmani G, Capuano GA. J Electrochem Soc 1994; 141:L132. [19] He Q, Mukerjee S. Electrochim Acta 2010;55:1709. [20] Mani P, Srivastava R, Strasser P. J Power Sources 2011;196:666. [21] Loukrakpam R, Luo J, He T, Chen YS, Xu Z, Njoki PN, et al. J Phys Chem C 2011;115:1682.
4632
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 3 7 ( 2 0 1 2 ) 4 6 2 7 e4 6 3 2
[22] Wanjala BN, Luo J, Fang B, Mott D, Zhong CJ. 2011, 21, 4012. [23] Loukrakpam R, Chang P, Luo J, Fang B, Mott D, Bae IT, et al. Chem Comm 2010;46:7184. [24] Wanjala B, Luo J, Loukrakpam R, Njoki PN, Mott D, Zhong CJ, et al. J Phys Chem C 2010;114:17580. [25] Fang B, Luo J, Chen Y, Wanjala BN, Loukrakpam R, Hong J, et al. Chem Cat Chem 2011;3:421. [26] Fang B, Wanjala BN, Hu X, Last J, Loukrakpam R, Yin J, et al. J Power Sources 2011;196:659. [27] Fang B, Luo J, Njoki PN, Loukrakpam R, Wanjala BN, Hong J, et al. Electrochim Acta 2010;55:8230.
[28] Fang B, Luo J, Njoki PN, Loukrakpam R, Mott D, Wanjala B, et al. Electrochem Comm 2009;11:1139. [29] Luo J, Wang LY, Mott D, Njoki PN, Kariuki N, Zhong CJ, et al. J Mater Chem 2006;16:1665. [30] Trongchuankij W, Pruksathorn K, Hunsom M. Appl Energy 2011;88:974. [31] Bianchini C, Shen PK. Chem Rev 2009;109:4183. [32] Wang CY. Chem Rev 2004;104:4727. [33] Loukrakpam R, Wanjala BN, Yin J, Fang B, Luo J, Shao M, Protsailo L, Kawamura T, Chen Y, Petkov V, Zhong CJ. ACS Catalysis 2011;1:562.