Hydrogen Electrocatalysis by Carbon Supported Pt and Pt Alloys An In Situ X - Ray Absorption Study

In situ x-ray absorption spectroscopy (XAS) in 1 M HC1O4 was used to examine the electronic and structural effects of hydrogen adsorption on carbon supported Pt (Pt/C) and Pt alloyed with first row transition metals (Cr, Mn, Fe, Co, and Ni). In the case of Pt/C, potential excursions from the double layer region (0.54 V vs. RHE) to 0.0 V caused significant changes in the XAS spectra whereas none was observed for the alloys. The L, and L2 x-ray absorption near edge structure indicated the generation of empty electronic states in the vicinity of the Fermi level due to adsorption of hydrogen, and the L, extended x-ray absorption fine structure indicated an increase in the coordination number of the first Pt-Pt shell from 9 to 11. The latter was attributed to a reversible surface restructuring process. Alloying of the Ptsuppresses both the electronic and structural effects at 0.0 V. A comparison of the electrochemical kinetics for hydrogen oxidation by these electrocatalysts in a proton exchange membrane fuel cell indicated that alloying of the Pt had insignificant effects on the kinetics. Previous work1 has shown that the most accepted mechanism of hydrogen molecule oxidation is the Tafel-Volmer sequence, with the rate-determining step being dissociation of the hydrogen molecule (Tafel, Ref. 2) followed by a fast oxidation step (Volmer reaction). Despite the facile nature of this reaction on Pt and Pt group metals, a better understanding of the electrocatalysis in terms of the electronic and geometric parameters involved is essential. The primary motivation emanates from the promise of electrocatalysts with better tolerance toward commonly encountered catalyst poisons such as CO and S '' and the eventual success of the direct methanol fuel cells.6 The electronic and geometric parameters which determine the electrocatalysis encompass several factors such as (i) the crystallite size effect, coordination numbers, bond distances, etc., and (ii) the Pt 5 d orbital vacancies. The effect of crystallite size and other geometric parameters for Pt has been extensively reviewed.7'8 Attempts to understand the mechanism and kinetics of hydrogen molecule oxidation on Pt surfaces based on different models of hydrogen chemisorption and the interplay of electronic factors have also been made.''° The x-ray absorption spectroscopy (XAS) technique with the near edge part (x-ray absorption near edge struc* Electrochemical Society Active Member. ture, XANES) and higher energy side of the spectra (extended x-ray absorption fine structure, EXAFS) offers the prospect for investigating these factors and their interplay under in situ conditions of hydrogen oxidation. Recently we have completed an in situ XAS study on a series of carbon supported Pt alloy electrocatalysts, where Pt is alloyed with the first row transition elements (Cr, Mn, Fe, Co, and Ni). 11,12 This study correlated the electronic (Pt 5 d orbital vacancy obtained from the Pt L3 and L2 XANES) and geometric parameters (bond distance and coordination numbers from the Pt L3 EXAFS ) with electrocatalysis for oxygen reduction reaction (ORR). There have been several reports on the effect of adsorbed hydrogen on the Pt L3 XANES and the Pt d band vacancies. Almost all of these are for oxide-supported Pt catalysts in gaseous hydrogen.'"4 The main effect is a broadening of the Pt L3 XANES white line on the high energy side of the peak. Similar effects have been observed for carbon supported Pt electrodes in 1 M HC1O4 in the hydrogen adsorption region.'1 Boudart has attributed this widening of the white line to transitions into Pt-H antibonding orbitals.'4 Recently Allen et al., have used in situ dispersive EXAFS to study carbon supported Pt over a wide potential range.'6 They found significant increases in the magnitude of the first peak (the first Pt-Pt shell) of the Fourier Downloaded 09 Mar 2011 to 129.10.187.55. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp 2286 J. Electroctiem. Soc., Vol. 143, No. 7, July 1996 The Electrochemical Society, Inc. transform in the hydrogen adsorption region. No explanation was given for this, but it indicates some structural changes. Tidswell et al.,'7 using x-ray reflectivity on Pt single crystals (<100>), have shown that adsorbed hydrogen can induce surface relaxation on Pt. All these studies indicate that adsorbed hydrogen can induce significant electronic and structural changes on Pt. The implication of this for the hydrogen oxidation kinetics or the susceptibility of the electrode to poisons is not known. The present study uses in situ XAS to study electronic and structural changes in carbon supported Pt and Pt alloys in the hydrogen adsorption region. Experimental Electrocatalysts, electrodes, and electrochemical characterization.—Five carbon-supported binary Pt alloys (Pt/Cr; Pt/Mn, Pt/Fe, Pt/Co, and Pt/Ni) and Pt electrocatalysts were procured from Johnson Matthey Inc. (West Deptford, NJ). Based on previous investigations" the electrocatalyst loading was chosen as 20% (by weight) metal on carbon. The gas diffusion electrode structure comprised a reaction layer (containing 40 to 50% PTFE) and a wetproofed carbon cloth substrate (Textron, MA). The electrodes were prepared with a constant Pt loading of 0.3 mg/cm2 (confirmed by atomic absorption spectroscopy). The electrode kinetic evaluation of hydrogen oxidation was conducted at a solid polymer electrolyte membrane (SPEM) interface at 95°C and 5 atm pressure. Methodologies for electrode impregnation with Nafion® solution (Aldrich), membrane electrolyte purification (Aciplex,® Asahi Chemical Co., Japan), and the fabrication of the membrane electrode assembly (MEA) are described elsewhere.20'2' The MEA was next incorporated into a single-cell test fixture with capabilities for half-cell measurements.2' Electrode kinetic measurements were carried out in a test station with provisions for temperature and pressure control, humidification of reactant gases, and gas flow measurements.2' Prior to electrode kinetic measurements, the MEA assembly was conditioned at several temperature, pressure, and humidity conditions, details of which are given elsewhere.2' Besides the electrode kinetic measurements at 95°C and 5 atm pressure, measurements were also carried out as a function of temperature (in the range 35 to 80°C) for determination of the activation energy. X-ray diffraction .—The characteristics of the crystalline structure of the supported Pt and Pt alloy electrocatalysts (formation of superlattices, etc.) were determined using xray powder diffraction (XRD). Measurements were carried out using a Sintag automated diffractometer with a Cu K radiation source. 28 Bragg angle were scanned over a range of 0 to 80°. The diffraction patterns were recorded and analyzed by comparing them with standard powder diffraction data such as the JCPDS powder diffraction patterns (NIST). Details of sample preparation, internal calibration, and analysis are given elsewhere."2 fit situ XANES and EXAFS investigation.—The x-ray absorption spectroscopic measurements were carried out at the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory (BNL), using the National Institute of Science and Technology (NIST) beam line X23A2. Details of the monochromator design and energy resolution are given elsewhere." Based on the monochromator and the overall beam line configuration, the presence of second harmonics was negligible. This was confirmed at both Pt (L, and L2 edge) and the K edge of the alloying element using absorption edges of test samples at approximately twice the energy." Prior to electrode fabrication all the electrocatalysts were soaked in either 2 M KOH or 1 M HC1O4 to remove residual oxides and unalloyed first row transition elements. The electrodes used for the in situ XAS measurements were prepared by a vacuum table paper making technique22 and comprise 76% electrocatalyst, 12% carbon fibers, and 12% PTFE (Teflon TFE-30, Du Pont). This afforded adequate step heights at both Pt L and the K edge of the alloying element in transmission and fluorescence modes. The electrodes were soaked in 1 M HC1O4 for 48 h to ensure complete wetting prior to incorporation in the spectroelectrochemical cell. Since XAS is an averaging technique, it was essential to ensure that all of the electrocatalyst was electroactive and there was a minimal amount of residual oxides and other phases (<5%). The spectro-electrochemical cell used as a part of this study allowed measurements in both transmission and fluorescence modes. The cell comprised the working electrode and an uncatalyzed carbon counterelectrode that were wetted with 1 N HC1O,. The separator used was a Nafion 117 proton exchange membrane (Du Pont). The reference was a calomel electrode, however; all potentials in the paper are reported vs. the reversible hydrogen electrode (RHE). The electrolyte was 1 N HC1O4, chosen due to nonabsorption of its anion, a property similar to that of the perfiuorinated sulfonic acid membrane (Aciplex,® Asahi Chemical Company, Japan). The cell was specially constructed to maintain all components under compression in order to avoid complications in the spectra due to random density fluctuations caused as a result of gas bubbles. Details of the cell and the data acquisition setup are given elsewhere."2 The potential control for the in situ XAS measurements was carried out using a potentiostat (Stonehart Associates, Model No. BC-1200) and a function generator (EG&G Princeton Applied Research Model No. PAR-175). XAS data were first recorded at the Pt L2 and L3 edges and then at the K edge of the respective alloying element. The measurements were first made at the opencircuit potential and then at potentials of 0.0, 0.24, and 0.54 V The potential was changed from one value to the next at a sweep rate of 1 mV/s. The K edge XAS measurements for the alloying element were then carried out at 0.0, 0.24, 0.54, 0.84, and 1.14 V to check for the stability of the alloying element. The methods used in analyzing the Pt XANES data and the determination of the Pt d band vacancies followed those of Mansour and co-workers'3'23 and are described in detail elsewhere."2 The EXAFS data were analyzed by methods described in detail in previous publications."2'24 Phase and amplitude data for the Pt-Pt and Pt-O interactions were derived from Pt foil and Na2Pt(OH), data at liquid nitrogen temperature. The FEFF program of Rehr et al.,2' (version 4.08) was used to derive the Pt-M (M denotes the first row transition element) phase and amplitude parameters. Results and Discussion X-ray diffraction—Comparison of the x-ray powder diffraction patterns with those of the standard JCPDS powder diffraction data base shows that the PtMn/C, PtCr/C, PtFe/C, PtCo/C, and PtNi/C alloys form intermetallic crystalline structures, the primary superlattice phase being of the type Pt3M (where M is the first row transition alloying element) possessing an L12-type lattice with an fcc structure. There were indications of a small contribution of a secondary phase of the type PtM possessing an Table I. Structural characteristics of the Pt and Pt alloy electrocatalysts, obtained from x-ray powder diffraction studies. Lattice parameter (Pt-Pt bonil distance) Average pticle size Electrocatalyst (A) (A)