CVD Catalyzed PEM Fuel Cell and Electrolyzer Electrodes

Abstract

Chemical vapor deposition (CVD) is commonly used to deposit metals from volatile precursors, typically by thermal decomposition processes. Papandrew has demonstrated the use of CVD to deposit platinum onto solid acid1 and perfluoroionomer2 electrolytes at relatively modest temperatures, suggesting that acid functionality facilitates the deposition process. Here, the possibility that the electrolyte acidity could provide deposition selectivity over other components such as PTFE and carbon in a typical fuel cell catalyst layer is being investigated. Figure 1 illustrates the original proof-of-concept experiment to demonstrate the preferential deposition of the catalyst on ionomer in favor of the other materials composing a gas diffusion layer (GDL). Here, a spiral of Nafion ionomer initially applied to the microporous side of a GDL is XRF imaged in green by the sulfur K emission line originating from the sulfonate functionality. In the experiment, the microporous side of the spiral impregnated GDL was then masked off and platinum acetylacetonate vapor was introduced to the opposite (uncoated) side of the GDL. As shown in the yellow and red Pt L & M scanning XRF images, the platinum subsequently CVD’d is concentrated at the ionomer locations, indicating the preferential deposition of platinum on ionomer over the other GDL materials. By introducing the Pt precursor vapor by the same pathway that reactants would travel in the fuel cell, in principle the deposited platinum is exclusively in locations with both reactant and ionomer access, conceivably maximizing catalyst utilization. A further benefit of the CVD process is that XRD and TEM both indicate that the Pt particles are uniformly in the 2-3 nm range, thus seemingly promising high mass activities. In practice, the CVD process results in the Pt particles seeded into the ionomer, and with low particle densities many are isolated from one another. Consequently, much of the catalyst in 0.1 mg Pt/cm2 fuel cell cathodes is electronically stranded and the electrochemical surface areas (ECSA’s) are disappointingly low. Experimentation with CVD directly onto membranes indicates that the acidity of the electrolyte is indeed a key factor in the deposition process and selectivity. For example, little to no deposition occurs on membranes in the Na+ form, a feature that may be exploitable in limiting the range of deposition. Physical strategies can also be employed to limit the depth of penetration, such as minimizing ionomer film thickness or by introducing permeation barriers. Further, the CVD conditions are observed to affect the deposition profile. Through various strategies, improvements are being made in the fuel cell catalyst utilization. If the amount of precursor is increased for CVD’ing directly onto membranes, Pt particle density and connectivity likewise increase and eventually silvering of the membrane surface occurs. Because of the interpenetrating nature of the catalyst and ionomer networks, it may be possible that highly effective electrodes can be achieved with such configurations in cases where gaseous reactant access is not key, such as in water electrolysis. Consequently, efforts are underway to see if it is possible to prepare iridium and platinum catalyzed membrane electrodes with “low” loadings (< 1 mg/cm2) that also provide sufficient particle density and connectivity to yield high electrolyzer catalyst utilizations. Figure 1. Scanning XRF images showing the signatures of an ionomer spiral (Sulfur K) impregnated into the microporous layer of a GDL, and of the subsequently CVD’d platinum (L & M lines). Acknowledgments Funding was provided by the U.S. Department of Energy, Fuel Cell Technologies Office, and by the Los Alamos National Laboratory LDRD (Laboratory Directed R&D) Program. References 1A. B. Papandrew, C.R.I. Chisholm, S. K. Zecevic, G.M. Veith, and T.A. Zawodzinski, J. Electrochem. Soc., 160, (2), F175-F182 (2013) 2S. Komini Babu, R.W. Atkinson, A.B. Papandrew, and S. Lister, Chem. ElectroChem., 2, 1752-1759 (2015) Figure 1