Application of Exsolved Structures as a Route to More Robust Anodes for Improved Biogas Utilisation in SOFCs
Biogas is mainly a mixture of carbon dioxide and methane, however the ratio of the two gases in many cases is not enough to prevent carbon formation if used directly in a nickel based cermet anode. This is compounded by the presence of hydrogen sulphide as a significant impurity in biogas, which is of course well known to poison SOFC Ni cermets. To look to tackle these issues University of St Andrews has been part of a collaborative UK-India research programme jointly funded by EPSRC (UK ) and DST (India) on improving bigas utilisation in SOFC, other partners are Imperial College London and University of Strathclyde from the UK and CGCRI (Kolkata) and IMMT (Bhubaneswar) from India. In this paper some of the recent activities at St Andrews to improve anode robustness through catalyst impregnation and exsloution are discussed.
Two approaches are described here, in the case of the impregnation, a proton conducting oxide, BaZr0.1Ce0.7Y0.1Yb0.1O3-δ (BCZYYb) was impregnated into the Ni-YSZ anode of a thin electrolyte, anode supported cell at various levels up to 1.6wt%. For the exsloution based structures, both Sr and Ca doped lanthanum titanate perovskites were doped with various catalyst materials such as Ni, Ce, Mg, and Rh. These perovskites were A-site deficient so that on reduction the catalyst materials were exsolved onto the surface as fine metallic particles with a typical size in the 10’snm. All anodes were tested on a standard reformed biogas mixture developed as part of the project. This represented the product of a 63:37 methane:CO2 input biogas exposed to 25% recirculation of an 80% utilised fuel to result in a gas mixture of 36% CH4, 26% CO2, 20% H2O, 4% H2 and 4% CO. For impurity testing H2S was added to this mixture at levels of 4-10ppm. Catalyst function was assessed by both cell testing and reforming activity.
For the impregnated specimens, microstructures revealed coarser, more concentrated distribution of impregnated particles on the surface of the anode as the level of impregnation increased. All of the BCZYYb impregnations improved cell performance, however the optimum performance was at the lowest level of impregnation (0.6wt%) (<1.5Acm-2 at 0.8V at 800°C) with performance dropping towards the non impregnated value (around 0.6 Acm-2 at 0.8V at 800°C) as impregnation levels increased to 1.6wt%. Short term steady state durability was good with initial performances proving stable over the first 50 hours of operation at 1.25 Acm-2 at 0.8V at 750°C. However poisoning effects were observed on the introduction of H2S with a performance drop of the order of 60-70% over 20 hours of exposure. Recovery was observed on return to a non-poisoned biogas mixture, however perfomance was still significantly lower than initial levels. Changing to steam/hydrogen mixtures helped bring performances back to initial levels.
The doped perovskites showed various levels of metallic particle exsolution depending on the dopant type and level. These have been tested for reforming activity, with all showing activity for reformation of the standard biogas mixture, however they also all exhibited rapid degradation on introduction of H2S. Only the Rh catalyst retained some catalytic activity in the presence of hydrogen suphide, although all specimens showed recovery when the this was removed from the gas stream. A comparison was carried out between exsolved and impregnated nickel, with the former showing greater resistance to initial particle coarsening and carbon deposition. This demonstrates that it is not just the catalyst material itself which is important but also the nature and morphology of the catalyst particle. These results show that both approaches continue to be of great interest in the continued development of robust anodes for challenging fuel environments.