Mesoporous Platinum Prepared by Electrodeposition for Ultralow Loading Proton Exchange Membrane Fuel Cells

The porosity and utilization of platinum catalysts have a direct impact on their performance within proton exchange membrane fuel cells. It is desirable to identify methods that can prepare these catalysts with the desired features, and that can be widely implemented using existing and industrially scalable techniques. Through the use of electrodeposition processes, fuel cell testing, and electron microscopy analyses before and after fuel cell testing, we report the preparation and performance of mesoporous platinum catalysts for proton exchange membrane fuel cells. We found that these mesoporous platinum catalysts can be prepared in sufficient quantities through techniques that also enable their direct incorporation into membrane electrode assemblies. We also determined that the mesoporous catalysts achieved a high porosity, which was retained after assembly and utilization within fuel cells. In addition, these mesoporous platinum catalysts exhibited an improved platinum mass specific power over catalysts prepared from commercially available platinum nanocatalysts.


Substrate Preparation
Mesoporous Pt catalysts were prepared by electrodeposition onto a variety of substrates to characterize the deposition process. The electrochemical activities of these nanostructures for the oxygen reduction reaction (ORR) were compared to commercial catalysts of Pt nanoparticles. The performance of these catalysts were evaluated through a series of tests in electrochemical rotating disc electrode (RDE) and proton exchange membrane fuel cell (PEMFC) setups.
A planar conductive Pt substrate was initially used to identify appropriate surfactants for use in the electrodeposition process, and to assess the current densities required for the formation of mesoporous Pt. A p-type <001> Si wafer was coated with a 5-nm thick Cr adhesion layer and a 150-nm thick layer of Pt, which was used as a planar substrate for the electrodeposition processes.
These Cr and Pt metal films were sequentially deposited by thermal evaporation, without breaking vacuum, using a custom built physical vapor deposition (PVD) system in 4D LABS (Simon Fraser University).
Mesoporous Pt was also deposited onto glassy carbon electrodes that were coated with a layer of Vulcan XC-72 (Cabot Corporation, United States) carbon particles and ionomer (Nafion ® DE2020, Dupont, United States)referred to as the C and ionomer films in the main text. This layer of catalyst free C and ionomer film mimicked was prepared similarly to the reference CCLs.
The similarities in CL preparation and layer thickness allowed a fairer comparison of the electrochemical performance of these mesoporous Pt catalysts to conventional CCLs prepared with Pt nanoparticles. The glassy carbon electrodes (5-mm in diameter) were inserted into a ChangeDisk electrode holder (Pine Research Instrumentation; NC, United States) and used as the working electrodes to support the Pt nanostructures during their electrochemical analyses by RDE S-3 techniques. The glassy carbon electrodes were each sequentially polished using a suspension of 300-nm and 50-nm diameter alumina particles (Buehler, IL, United States) to create a mirror-like finish. The polished carbon was coated with an ~10-µm thick layer of Vulcan XC-72 carbon nanoparticles mixed with a 30 wt % loading of ionomer (Nafion ® DE2020, Dupont, United States).
This layer of carbon particles and ionomer was prepared by spin casting this ink mixture at 500 rpm for 5 min prior to Pt electrodeposition to mimic the structure and composition of conventional CCLs in PEMFCs. Prior to spin coating, the suspension of carbon and ionomer were dispersed in a 3:1 (w/w) mixture of 2-propanol (Anachemia, Canada, ACS grade) and deionized (DI) water to achieve a 0.3 w/w % (solid/liquid) solution. All DI water, used to prepare the necessary solutions and for rinsing of the electrodes, was filtered with a Barnstead DIamond TM deionizing water system with an output of 18 MΩ·cm. An aliquot of 20 mL of the dilute ink solution was placed into a glass vial. The vial containing the ink solution was chilled by placing it into an ice filled container and sonicated using a process of 1 s "on" and 3 s "off" for a total duration of 1 h (and a total of 15 min of sonication during the collective "on" time) using a probe sonicator (Fisher Scientific, United States, Sonic Dismembrator 500). A 12-mm diameter sonication probe was immersed ~1 cm into the solution, and operated at 40 % of the maximum power (the maximum power was ~500 Watts).
Large area glassy carbon electrodes were also coated with the layer of carbon particles and ionomer upon which the electrodeposited mesoporous Pt were prepared and transferred to membrane electrode assemblies (MEAs) for fuel cell testing. To prepare these films containing a layer of mesoporous Pt, a 3-mm thick and 5-cm by 5-cm wide square glassy carbon plate (SPI Supplies, United States) was first spin coated (at 500 rpm for 30 min) with the layer of carbon particles and ionomer, and subsequently loaded into a custom electrodeposition cell (see Figure   S-4 S11 for more details). The electrodeposition of mesoporous Pt were performed on planar Pt electrodes, C and ionomer coated glassy carbon RDE electrodes, and C and ionomer coated large area glassy carbon electrodes to determining the optimal electrodeposition conditions, to analyze the electrochemical characteristics of the mesoporous Pt catalyst, and to prepare mesoporous Pt for incorporation into the MEAs for PEMFC analyses.

Electrodeposition of Mesoporous Pt
Planar Pt electrodes, C and ionomer coated glassy carbon RDE electrodes, and C and ionomer coated large area glassy carbon electrodes were each separately used as the working

Electron Microscopy Characterization
Electron microscopy techniques, such as scanning electron microscopy (SEM) and transmission electron microscopy (TEM), were used to verify the surface coverage, morphology, and crystalline nature of the electrodeposited Pt. The SEM analyses were performed using an FEI Helios SEM/ focused ion beam (FIB) dual beam system operating at 10 kV. This system was also equipped with an EDAX energy dispersive X-ray spectroscopy (EDS) detector for elemental mapping of the materials. The surface coverage of mesoporous Pt over the layers of C particles and ionomer were calculated from a series of SEM images by using image segmentation with assistance from ImageJ (ImageJ 1. Furthermore, the number of electrons involved in the ORR were investigated using a Koutecký-Levich plot (current density versus the inverse square root of the rotational speed in rad/s) generated from the series of RDE experiments performed for each of the electrodes. A representative analysis is presented in Figure S10.
A total of five CV scans at 50 mV/s were performed, from 0 to 0.8 V versus RHE, between each consecutive LSV measurement as part of the ORR analysis to ensure a consistency of the Pt catalysts. The CV profiles were used to assess the Aecsa of the Pt for each of the electrodes (e.g., Figure 4a). The Aecsa of Pt was calculated from the area under the hydrogen desorption peaks assuming that the charge associated with the formation of a monolayer of hydrogen on the Pt surfaces is approximately 210 µC/cm 2 . [2][3][4] The Aecsa per gram of Pt was determined following the series of electrochemical analyses. This series of electrochemical tests, which were performed over a continuous period of at least 12 h, were followed by X-ray fluorescence spectroscopy (XRF) and inductively coupled plasma mass spectrometry (ICP-MS). This order of experiments was adopted due to the destructive nature of the ICP-MS analyses.

Preparation of Catalyst Coated Membranes
A series of MEAs were prepared and tested to characterize the electrodeposited mesoporous Pt for its performance in PEMFCs. To compare these materials with conventional cathode catalysts for PEMFCs, MEAs were prepared using cathode catalysts containing a mixture of commercially available Pt nanoparticles coated C particles (TEC10V50E, Tanaka Kikinzoku Kogyo, Japan), and ionomer (Nafion ® DE2020). A series of MEAs were also prepared using mesoporous Pt electrodeposited onto films of carbon particles (Vulcan ® XC-72) mixed with S-9 ionomer. The standard anode and cathode catalyst inks were prepared according to the same procedures outlined in the section providing details on the preparations of the substrates. The same procedures outlined therein for preparing the solutions of C particles and ionomer were also used to create supporting materials for the electrodeposition of mesoporous Pt.
These anode and cathode catalysts were hot-bonded with a proton exchange membrane to form catalyst coated membranes (CCMs) by a process of decal transfer. 5  min with a use of a Carver ® 25 ton hydraulic unit (Carver Inc., United States). The reference samples were each prepared by the same hot-bonding conditions for both the cathode catalysts and anode catalysts, which were prepared using the Pt NP containing ink coated onto ETFE decals.
The different components used in the hot-bonding process are depicted in Figure S11. A sheet of S-10 0.05-mm thick Teflon ® was placed on top of the layered CCMs, followed by a sheet of 2-mm thick polyurethane rubber prior to hot bonding to ensure an even distribution of the bonding pressure.
After 5 min of bonding, the layered CCMs were immediately removed from the Teflon ® sheet and allowed to cool to room temperature for at least 10 min. The ETFE and glassy carbon plate were gently removed by hand after the CCM assembly was sufficiently cooled. The MEAs were prepared by hot bonding the CCMs with of macroporous layer (MPL) coated gas diffusion layer (GDL) (Sigracet ® gas diffusion layers, SGL Group, Germany) and adding a piece of G10 gasket on either side of the CCM. The hot bonding procedure was identical to the one outlined above for the preparation of the CCMs except that the bonding time was reduced from 5 min to 2 min.
Additional Mylar ® gaskets (DuPont Teijin Films, UK) were used to adjust the overall thickness of the gaskets as necessary to induce a 25 % compression of the MPL/GDL materials when assembled for testing in the fuel cell.  Figure S8. The MEAs were deemed to be stable when at least three consecutive LSV experiments exhibited an overall deviation of < 0.5 mA/cm 2 . After the LSV measurements had reached stabilization, 3 additional LSV experiments (100 mV/s; 0 to 1.2 V versus RHE) were performed, the results averaged, and analyzed to assess the fuel cell performance for each of the MEAs.

Analysis of Elemental Composition
Quantitative elemental analysis techniques were used to compare the Pt loadings of the samples. The Pt loading of the samples were initially verified by XRF spectroscopy (Thermo Scientific, Niton XL3t, United States; operating at 50 kV), which was calibrated with a series of Pt thin film standards (ranging from 10 µgPt/cm 2 to 400 µgPt/cm 2 in 50 µgPt/cm 2 increments) made   S-25