Ternary chalcogenide-based photoelectrochemical cells. 6. Is there a thermodynamic explanation for the output stability of CuInS₂ and CuInSe₂ photoanodes?

Using recently determined thermodynamic data for CuInS₂ and CuInSe₂, we estimated the Gibbs free energy changes for several decomposition and exchange reactions for these semiconductors in aqueous polysulfide and aqueous polyiodide, and, where possible, calculated their corresponding electrode potentials. Exchange reactions for In₂O₃, believed to be the native oxide on these semiconductors, are included, as well as chemical conversion reactions for their preparation from the corresponding indium sesquichalcogenides. With the aid of measured flat-band potentials, the band edges of the semiconductors (or semiconductor/surface oxide systems) are determined on the electrochemical scale and the quasi-Fermi levels are estimated from literature data. When these are combined with the calculated decomposition potentials and the measured redox potentials, (in)stability electron energy diagrams that are strictly valid only near open circuit are constructed. On the basis of these the absence of Se/S exchange on n-CuInSe₂ in polysulfide, the protective function of an indium oxide surface layer on CuInSe₂ in polyiodide, the stabilizing effect, on CuInSe₂, of adding copper ions to polyiodide solutions, and the stability of CuInSe₂ against reductive decomposition can be understood. Also, the greater likelihood of In₂X₃ (X = S or Se) formation (in polysulfide) on CuInS₂ than on CuInSe₂, the possibility of CuInSe₂ formation on the surface of In₂Se₃, immersed in Cu-containing solutions, and the photoanodic decomposition of n-CuInSe₂ under illumination in acetonitrile can be rationalized. The data do not explain the difference between CdSe and CuInSe₂, with respect to chemical Se/S exchange, the absence of InI₃ formation in In₂O₃ in polyiodide, the absence of photoanodic decomposition of CuInX₂ into copper and indium sulfides in polysulfide, and the lack of Cu₂Se formation from photoanodic decomposition of CuInSe₂ in polyiodide. In these latter cases kinetic factors, which may be related to the large number of reactants and charge carriers that are involved in some of the suggested decomposition reactions, appear to be dominant. Because of the possibility of very large deviations from steady state and the occurrence, at least initially, of products in states very different from their standard state, the free energy changes for virtually all decomposition reactions are expected to be shifted sufficiently to make these and other semiconductors thermodynamically unstable toward initial decomposition.