The use of semiconductors, as photoelectrodes in electrolytic cells for the electrolysis of water, is described and the results reported in the literature for various semiconductors are reviewed. The most important properties of the semiconductor are shown to be the band-gap energy E g, and the flat-band potential U fb . The semiconductor absorbs photons that are more energetic than the band-gap energy and creates electronhole pairs. These charge carriers can be separated before recombination by the electric field at the semiconductor-electrolyte interface. For electrolysis to proceed, the potential corresponding to the band gap must appreciably exceed the standard potential for the electrolysis of water, 1.23 volts. In addition, the flat-band potential must be more negative than the hydrogen potential or an external bias voltage is required. The semiconductor must not corrode under the operating conditions and must permit transfer of the minority carrier to the electrolyte. The current theories of charge transfer and reaction mechanism are discussed. Many semiconductive oxides have been tested as photoanodes and found to be stable. However, only those compounds that have band gaps of ca. 3 eV have been found to have flat-band potentials that are more negative than the hydrogen potential. These compounds will electrolyse water, without additional bias voltage, but are inefficient absorbers of solar energy. Experiments with p-type photocathodes, p and n combinations and various special configurations are also described. The paper concludes with a general discussion of the practical prospects of photoelectrolysis in comparison with solid-state solar cells. © 1979 Springer-Verlag.
