Supporting Information

Transcription

1 Supporting Information Wiley-VCH Weinheim, Germany

2 Efficient Oxygen Reduction Fuel Cell Electrocatalysis on Voltammetrically De-alloyed Pt-Cu-Co Nanoparticles Ratndeep Srivastava, Prasanna Mani, Nathan Hahn, and Peter Strasser* Pt 20 Cu 60 Co 20 Pt 20 Cu 20 Co 60 Intensity (a.u) Pt 20 Cu 40 Co Figure S1: X-ray diffraction profiles of the Pt-Cu-Co ternary nanoparticle alloy precursors. Synthesis conditions: 600 C, 7h

3 0.5 a Mass Activity, IM (A/mg Pt ) PtCuCo 20:60:20 PtCuCo 20:20:60 PtCuCo 20:40:40 PtCo 25:75 Pt

4 500 b Specific Activity I S (ma/cm2 ) Pt PtCuCo 20:60:20 PtCuCo 20:20:60 PtCuCo 20:40:40 PtCo 25:75 Pt

5 120 c 100 ECSA (m 2 /g Pt ) PtCuCo 20:60:20 PtCuCo 20:20:60 PtCuCo 20:40:40 PtCo 25:75 Pt Figure S2: Electrochemical characteristics of voltammetrically de-alloyed Pt-Cu-Co catalysts (a) Mass activities at 0.9V cell potential (b), Surface area normalized (specific) activities at 0.9V cell potential (c), and electrochemical surface areas (ECSA) determined after activity testing.

6 Normalized Surface Area Pt 20 Cu 60 Co 20 Pt 20 Cu 20 Co 60 Pt 20 Cu 40 Co 40 Pt Number of Cycles Figure S3: Relative Electrochemical Surface Area (ECSA) loss during 10,000 and 30,000 potential cycles within V/RHE with a scan rate of 100 mv/s. Pt 20 Cu 60 Co 20 was cycled within the range V /RHE.

7 Experimental Details Preparation of the catalysts PtCuCo alloy catalysts were synthesized via a liquid precursor impregnation method. Cu,Co precursor solutions were prepared by dissolving appropriate amount of Cu (NO 3 ) 2 6H 2 O and Co(NO 3 ) 2.6H 2 O(ACS reagent, Sigma Aldrich) in de-ionized water (>18.2MO, Milli-Q gradient system, Millipore Inc.). The precursor solutions were then added to weighted amount of commercially viable high surface area carbon supported Pt (30 % Pt by weight). The mixtures were ultra sonicated for 1 min (Branson Sonifier 150) until thick slurries were formed. The catalyst synthesis mixture was subsequently frozen in liquid nitrogen for 15min, and then freeze-dried in vacuum (50 mtorr) overnight until the temperature of the samples reached 25 o C. The resulting powder was loaded as a thin powder film on the bottom of a rectangular alumina boat (50 ml) with a partially covering lid. The powders were then annealed to a maximum temperature of 600 o C for 7 h ( 10 K/min heating rate) in a flow furnace (Lindberg Blue) under a flowing 4% hydrogen atmosphere (Ar balance). The PtCuCo/C samples were then stored in a Nalgene desiccator with continuous flow of N 2 through the desiccator for future use. Catalyst Characterization by XRD XRD analysis of the electrocatalysts was performed by a Siemens D5000 (? / 2?) diffractometer equipped with a Braun Position Sensitive Detector (PSD) with an angular range of 8 o. The Cu Ka source was operating at a potential of 35kV and a current of 30mA. 2? diffraction angles ranged from 21 deg to 70 deg, using step scans of 0.02 deg / step and a holding time of 10 s per step. Advanced X ray Solution (X-ray commander, Bruker AXS) software was used to control the diffractometer with a desktop computer. XRD patterns were analyzed by JADE 8.1 software (MDI): Peak profiles of individual reflections were obtained by a non-linear least-square fit of the Cu Ka 2 corrected data. Rotating Disk Electrode (RDE) Film Preparation Prior to electrode preparation, a 5 mm diameter glassy carbon Rotating Disk Electrode (RDE) was polished to a mirror finish using 0.5 and 0.05 µm alumina suspension (Buehler Inc.). A catalyst ink was prepared by mixing measured amount of the alloy precursor catalysts in 20 ml of an aqueous solution containing 5wt% Nafion solution (Sigma, #274704) and ultrasonicated for 15min. A 10 µl aliquot was dispensed onto the Rotating Disk electrode resulting in a typically Pt loading of about 14 µg Pt/cm 2 geometric surface area. The ink was then dried for 10 min in air.

8 Electrochemical Characterization using RDE The electrochemical RDE cell was a custom-made, three-compartment cell. The working electrode was a commercial glassy carbon rotating disk electrode of 5 mm fixed diameter. The reference electrode was a mercury-mercury sulphate electrode. All electrode potentials were subsequently converted into and are reported with respect to the reversible hydrogen electrode (RHE) scale. The counter electrode was a piece of platinum gauze to ensure large surface area. A commercial rotator from Pine Instrument was used to conduct the rotating disk experiment. The electrolyte used was 0.1 M HClO 4, prepared by diluting 70 % redistilled HClO 4 (Sigma #311421) with de-ionized water. The disk potential was controlled with a potentiostat, BiStat (Princeton Applied Research, Ametek). All measurements were conducted at room temperature. At the beginning of electrochemical measurements, electrocatalysts were immersed into the electrolyte under potential control and held at 0.06 V / RHE until the measurements commenced. Cyclic voltammetric (CV) measurements were conducted in de-aerated electrolyte, under N 2 atmosphere. The initial three cyclic voltammograms were recorded between 0.06V and 1.2 V at 100 mv/s in order to monitor the early stage of the de-alloying process. Then, the electrocatalysts were pretreated using 200 fast CV scans between 0.06V and 1.2 V at a scan rate of 1000 mv/s. Thereafter, the potential was again scanned at 100 mv/s from 0.06 V to 1.2 V and back to 0.06 V in order to determine the platinum electrochemical surface area (Pt-ECSA). The Pt-ECSA of the catalyst was determined from the mean integral charge of the hydrogen adsorption and desorption areas after double-layer correction, using 210 µc cm 2 Pt as the conversion factor. Linear sweep voltammetry (LSV) measurements were conducted by sweeping the potential from 0.06 V anodically to the open circuit potential (around 1.0 V) at the scan rate of 5 mv/s. Mass and specific activities were established at 900 mv /[RHE], at room temperature. The electrochemical behavior (CV and LSV) of the Pt-Cu catalysts was compared to a 30 wt% reference platinum electrocatalyst supported on a high surface area carbon support. MEA fabrication Catalyst ink was prepared by initially blending the Pt/C or Pt-Cu-Co/C cathode catalyst with cold isopropyl alcohol (HPLC grade, Sigma Aldrich) with a magnetic stirrer. Then it is sonically dispersed (Model 75D, VWR) for 1h. Appropriate amount of 5wt% 1100 eq wt Nafion was added to the dispersed catalyst for H + conduction and binding of the catalyst particles in the catalytic layer and with a membrane. This mixture is again sonicated for 1h. Catalyst coated membranes (CCMs) were prepared by a robotic spray machine (PVA Inc.). Precise control of the catalyst ink flow to the atomizing nozzle was fetched by a peristaltic pump (Masterflex, Cole-Parmer Inc.). Nafion NRE212 (dispersion cast membrane) was supported by placing the membrane between two high density polyethylene sheet with a desired window size on the sheets for catalyst coating which is then clamped together and kept in the spray pattern area. Several spray pattern

9 cycles was repeated for desired catalyst loading before drying the CCM/MEA in a conventional oven at 80 o C. Electrochemical De-alloying A separate 10 cm 2 fuel cell assembly has been separated for in-situ electrochemical dealloying, since base metal dissolution contaminates the flow fields. We have used S. S bottles and diverted the outlet vent lines from PEMFC into the bottles to collect, contaminated water (base metal dissolved). After MEA preparation the cell is assembled with a GDM (SGL Carbon Inc.,) and gasket (Teflon) (completion of STEP 1 of the MEA dealloying procedure outlined in Fig. 1) Cell temperature is fixed at 30 o C, cathode and anode humidifier are kept at 50 o C. Hydrogen is fed onto the anode side and nitrogen on the cathode side, when the OCV (Open circuit potential) becomes stable a CV is measured using potentiostat (GAMRY reference 600) between 0.05V and 1.2V, with a scan rate of 20 mv/sec. This CV is taken to depict that as the MEA is prepared with a higher base metal content, no clear hydrogen desorption peak is observed. Now the cell temperature is increased to 80 o C with humidifiers at 100% RH. Oxygen is fed onto the cathode side, and when stable OCV is reached, the cell is put into the potentiostatic mode at a constant voltage of 0.6 V. The cell is allowed to hydrate for 4 hrs, after which the cell temperature is brought down to 30 o C, and cathode and anode at 50 o C. Again nitrogen is fed onto the cathode side. When stable OCV is reached, process of electrochemical leaching is begun, where the potential is cycled between 0.5V 1.0V repeatedly (about times) until the hydrogen desorption peak has become time stable and similar to a pure Pt catalyst (completion of STEP 2 of the MEA dealloying procedure outlined in Fig. 1). The cell is now dissembled and the metal ion containing MEA is removed and soaked in 1 M sulfuric acid at 80 o C for 1 hr, followed by washing with DI water. This process is again repeated once more I order to remove any SO 4-2 ions presence (completion of STEP 3 of the MEA dealloying procedure outlines in Fig. 1). The MEA is further dried in an oven at 80 o C overnight. Fuel cell testing The electrochemically de-alloyed CCMs/membrane electrode assemblies (MEAs) with active area of 10cm 2 were then assembled with gas diffusion media, GDL 10BC (Sigracet, SGL Carbon Inc.), 3-channel serpentine flow fields (Poco graphite blocks) and gold coated current collectors supplied from Fuel Cell Technologies Inc. It is then held at a compression of 75in-lb with a digital torque range (ComputorQ3, CDI torque) and tested in a dual channel test station (Fuel Cell Technologies Inc.). The relative humidity of the feed streams for anode/cathode = 100 %. The MEAs were activated under 150kPa abs pressure at both electrodes initially potentiostatically (0.6V), once the current is stabilized the MEAs were then conditioned for 24h galvanostatically (equivalent current at 0.6V) before recording i-v by LabView software. During polarization stoichiometric flows of? = 2/10 (H 2 /O 2 ) for i = 0.2 and 0.2A/cm 2 flows for i < 0.2A/cm 2. The performance of the fuel cell was monitored by a data acquisition system from start to stop. IR correction was done by correcting the cell voltage with measured

10 ohmic resistance of the fuel cell by an inbuilt AC impedance analyzer and cell current. The frequency of the AC sine wave was set to 1kHz and the peak to peak amplitude of the fuel cell was optimized to be 50mA in the entire range of the fuel cell DC to reduce noise/harmonics. MEA characterization and performance evaluation To accurately measure the activation polarization of the catalyst, it is essential to determine H 2 crossover current in the fuel cell determined by linear sweep voltammetry (LSV) by sweeping the potential from 0.05V to 0.6V (20 mv/s) at same temperature and pressure after i-v recording. At >0.4V the resulting hydrogen oxidation current is purely limited by H 2 permeation (cross over) rate. The H 2 anode served as a counter & reference, while the N 2 purged cathode is the working electrode with gas flow of H 2 /N 2 = 160mL/min. All the voltages reported here are in RHE scale. Electrochemical Surface Area Characterization Pt surface area is determined by cyclic voltammogram, the anode and cathode side of the cell is fed with H 2 and N 2. The cell temperature is brought to 30 o C, and humidifiers are kept at 50 o C at atmospheric pressure. Cyclic voltammogram are performed between V at a scan rate of 20 mv/s, the area of H 2 adsorption peak between 0.05V and 0.4 V is calculated after double layer correction.