Electrochemical Characterisation and Fuel Cell Testing of Bespoke, Low Loading Platinum on Carbon Black Catalysts Fabricated By Physical Vapour Deposition.

Abstract

Listen

Translate article

A critical component of achieving net zero emissions by 2050 is the continuing research and development of improved methods of producing, as well as using, green hydrogen. Hydrogen is a key component of many industrial and chemical processes, but here we focus on its role in the energy transition, specifically proton exchange membrane (PEM) fuel cells. A key part of the PEM fuel cell is the catalyst material and structure, and this work is part of a wider, multi-institutional collaboration concerning metal atoms on surfaces and interfaces (MASI).MASI’s goal is to increase the surface area to volume ratio of metal particle on support catalysts, exploring the sub-nanometre metal particle size domain (referred to henceforth as nanoclusters) by fabricating catalysts utilising physical vapour deposition (PVD.) In contrast to wet chemical methods, the size of metal particle depositing onto the surface can be better controlled. This versatile method has been applied to deposit many different metals onto a variety of morphologies of surface to date. The chemical and physical properties of these nanoclusters can then be studied and exploited where an improvement in activity on nanoparticle catalysts is observed, as well as reducing the amount of ‘wasted’ atoms buried below the active surface of the catalyst, reducing the unit cost of, for example, a fuel cell stack.In this study, platinum is deposited onto a carbon black (XC-72R) support at four different weight loadings (1.1wt%, 2.3wt%, 5.0wt% and 9.7wt% Pt/XC-72R carbon named Pt-1, Pt-2, Pt-5 and Pt-10 respectively,) and is compared to two commercially available equivalents (46.2wt% Pt/HSC TKK and 10wt% Pt/XC-72 carbon E-TEC.) The catalyst loadings were determined using ICP-OES. Imaging of the catalyst surface was performed using aberration-corrected scanning transmission electron microscopy (AC-STEM), including the generation of a size distribution of the platinum particles present. The electrochemically active surface area (ECSA) of the platinum catalyst was obtained using carbon monoxide stripping voltammetry. The kinetic properties of the catalysts were then probed using the oxygen reduction reaction (ORR) via rotating disc electrode (RDE) voltammetry, and a discussion of the factors impacting the observed onset potential (specifically kinetics and catalyst coverage) is presented. The selectivity of the ORR with respect to hydrogen peroxide production is also studied using rotating ring disc electrode (RRDE) voltammetry. Finally, single cell testing to study fuel cell performance, specifically on the anode side of the cell, is performed and analysed for future optimisation. Accelerated stress testing (AST) was also performed; ORR activity and fuel cell performance were probed as a function of number of break in cycles, up to 4000 cycles.STEM imaging revealed significant differences in the size of platinum particle across the samples studied. For example, the 2.3wt% Pt/C catalyst exhibited a platinum particle size of 0.6 ± 0.3 nm, far smaller than the 2.2 ± 0.5 nm found for TKK.Further, our study found that all PVD fabricated catalysts showed a significantly higher ECSA than the commercial equivalents, along with a shift in onset towards a higher overpotential for the ORR which is explained by kinetic differences due to size of platinum particle – platinum catalysts become worse at catalysing the ORR as the size of particle decreases. This is therefore used as further evidence of our nanoclusters’ catalytic activity. RRDE voltammetry revealed a general trend of decreasing peroxide production as weight loading increasing, explained by probable decreasing interparticle distance as well as increased particle size. AST showed all catalysts are relatively stable; increases in overpotential between initial performance and after 4000 cycles were in the region of 16 mV.Single cell testing revealed exceptional performance for the PVD catalysts tested. The 5.0wt% and 9.7wt% catalysts showed superior initial performance to TKK at a fraction of the platinum loading in all regions of the IV curve, with an associated improved peak power density. AST revealed that peak power density decreased in percentage terms more for the PVD catalysts, and the decrease was larger as weight loading increased. After 4000 cycles, the 9.7wt% catalyst still showed comparable performance to TKK, with the 5wt% just below.Future work will focus on trying to decrease the mean size of platinum nanocluster further, optimising catalyst loading and enhancing stability of the catalyst with respect to degradation. Figure 1