MODEL DEVELOPMENT AND VALIDATION OF PALLADIUM-BASED MEMBRANES FOR HYDROGEN SEPARATION IN PEM FUEL CELL SYSTEMS

Abstract

Selective Pd-based membranes can extract near pure H2 from hydrocarbon reformate gases in liquid-fueled standalone proton exchange membrane fuel cell (PEMFC) systems. A steady state system-level model of such a PEMFC system is built to understand trade-offs among key system performance metrics. H2 recovery fraction in the membrane purifier critically affects overall system efficiency and thermal management. Hence, the development of detailed, high-fidelity mathematical models for Pd-based H2 membranes is critical for optimizing hydrocarbon-fueled PEMFC systems relying on such membrane purification. Although interactions between Pd with reformate gases have been explored experimentally and with density functional theory models in the literature, membrane purifier design is often carried out using high-level approximations in empirical fits to experimental data with inadequate range of applicability for optimizing for system performance. Therefore, to provide a more comprehensive membrane reactor model that predicts membrane performance over a broader range of conditions, a microkinetic model that captures surface and bulk Pd-H2 and Pd0.77Ag0.23-H2 interactions is presented. A systematic procedure to estimate thermokinetic parameters is established such that model results compare favorably with experimental measurements of H solubility and fluxes as a function of H2 partial pressures and temperature. The Pd-H2 interaction model is combined with a porous media transport model in a 1-D through-the-membrane composite membrane model that is validated against experiments performed on a composite membrane. Thermokinetic parameters for CO and H2O competitive adsorption on Pd surface are estimated by fitting experimental observations and thermodynamics from literature DFT studies, and this thermochemistry is added to the pure Pd-H2 thermochemistry to assess the impact of CO and H2O poisoning for reformate purification applications. The 1-D composite membrane model is combined with a channel flow model to form a comprehensive 2-D "down-the-channel" model, which is validated through counter-current gas flow experiments on the same composite membrane. Simulations using the 2-D model show that for a tested composite membrane, the porous support resistance is often a rate-limiting process for H2 transfer. However, with an improved substrate design and a thinner membrane, other processes, such as competitive surface adsorption, can have a more significant role in limiting hydrogen fluxes as characterized by the effective drop in H chemical potential related to the process. important role, depending on the operating conditions This 2-D model is useful for carrying out parametric studies, and is a basis for further research.