Concepts for Ultra-High Power Density Solid Oxide Fuel Cells (SOFC)

Abstract

Over the past two decades, considerable work has been reported on solid oxide fuel cells (SOFC) [1-3]. While progress has been made, there has not been significant further gain in performance (at 800oC), beyond that achieved about fifteen years ago (~2 Wcm-2), despite using many different cathodes. Thus in many cases, the cathode may not be limiting the performance. To achieve further improvements in cell performance beyond what has been discussed in previous studies [4,5], it will require accurate identification of various polarization losses, their sources, and dependence on material and microstructural parameters, atmosphere and temperature. The majority of the modeling studies do not provide the required guidance as they rely on a number of often not so easily measurable parameters thus necessitating a somewhat arbitrary selection of such parameters. The objective of this work is to estimate various polarization losses based on out-of-cell measurements made using micro-fabricated electrodes, microstructural measurements on cells, and cell electrochemical performance measurements using a parametric model. Such an approach may identify where the dominant losses are and how to further improve cell performance. The minimum number of independent parameters necessary to model cell performance for an anode-supported SOFC with five distinct layers are 9 and they are: Deff1 O2-N2 and Deff2 O2-N2 , the effective binary diffusivities through the cathode current collector and the cathode functional layer, respectively; Deff1 H2-H2Oand Deff2 H2-H2O , the effective binary diffusivities through the anode support and the anode functional layer, respectively; cathode exchange current density (i0c ); anode exchange current density ((i0a ), cathode transfer coefficient (αc), anode transfer coefficient (αa), and ohmic resistance Ri. The parametric model is given by equation (1). V(i)=E0-iRi-(RT/2F)arcsinh(i/2i0 c)-(RT/2F)arcsinh(i/2i0 a)+( RT/2F)ln(pH2'(i)pH2Oo /( pH2o pH2O')+ ( RT/4F)ln(pO2'(i)/ pO2o ) (1) where pO2o and pO2'(i) are the oxygen partial pressures just outside the cathode current collector and close to the cathode functional layer/electrolyte interface, respectively; pH2o and pH2'(i) are the hydrogen partial pressures in the fuel just outside the anode support and close to the anode functional layer/electrolyte interface, respectively; pH2Oo and pH2O' are the water vapor partial pressures in the fuel just outside the anode and close to the anode functional layer/electrolyte interface, respectively. In equation (1), both transfer coefficients are assumed to be 0.5. As seen in Fig. 1, at low current density, the most dominant contribution is cathodic activation polarization for a cathode grain size of 2 mm. However, beyond about 1.3 A/cm2, the ohmic contribution is the dominant one, and underscores its importance. In Figs. 2a to 2b, the cathode functional layer grain size is varied from 2 μm to 0.2 μm. This increases the maximum power density by 0.3 W/cm2. However, if the ohmic resistance can be somehow reduced from 0.1 to 0.05 Ωcm2, the maximum power density increases by 1.2 W/cm2 (Figs. 2b and 2c) even in a thin electrolyte, anode-supported cell. This shows the profound role of ohmic contribution on cell performance. By contrast, a change from LSM/YSZ to Pt/YSZ only leads to a modest increase in maximum power density of about 0.4 W/cm2 (Figs. 2c and 2d). This is consistent with the observation that cells with several very different cathodes exhibit similar performance. Further work is thus required in lowering the ohmic contribution. Many cathodes that have been developed over the years appear quite satisfactory and not limiting cell performance for operation at 800oC. Funded by DOE EFRC Grant Number DE-SC0001061 as a flow-through from the University of South Carolina.