A composite semiconductor electrode with the structure "n-Si/p-CuI/ITO/n-i-p a-Si/n-p GaP/ITO/RuO 2 " was fabricated for the purpose of achieving efficient solar water splitting. The electrode showed a stable photoanodic current due to oxygen evolution with a large negative photoshift (V p ) of about 2.2 V from an anodic current at a RuO 2 electrode. The photoshift was large enough for full water splitting. A photoelectrochemical (PEC) cell, composed of the composite electrode, a Pt counter electrode, and 0.10 M Na 2 SO 4 (pH 6.3), generated a photocurrent density of 1.88 mA cm -2 under simulated solar illumination (AM 1.5 G, 100 mW cm -2 ), yielding a solar to chemical conversion efficiency of 2.3% as calculated from the photocurrent value. The result has shown that the combination of "crystalline Si/a-Si/GaP" is suitable for efficient solar water splitting. It is shown that the efficiency can be increased by use of GaP with a well-regulated p-n junction.

The flatband potential (UFB) and solar cell characteristics for surface-alkylated and Pt nanodotted n-Si(111) electrodes have been studied in normalI3−∕normalI− redox electrolytes using various alkyls as the surface-terminating group. It is found that the UFB for the surface-methylated and Pt nanodotted n-Si(111) electrodes shifts toward the negative with increasing I− concentration in the electrolyte, in parallel to the equilibrium redox potential Ueq(I3−∕I−) , and thus the open-circuit photovoltage Voc remains nearly constant among various normalI3−∕normalI− redox electrolytes with different Ueq(I3−∕I−) . The constant Voc is observed only for the normalI3−∕normalI− redox couples with varied I− concentrations and not for other redox couples, indicating that it is not caused by the Fermi level pinning via a surface state. The UFB shift with the I− concentration is explained in terms of the I− adsorption in the form of a Si–I⋯normalI− complex at surface Si–I bonds, which are formed at nonmodified (naked) Si sites. The n-Si electrodes modified with long-chain alkyls show similar negative shifts in the UFB by iodine adsorption. The UFB measurements in the dark and under illumination have also shown that the Pt nanodots act as an efficient catalyst (gate) for interfacial electron transfer.

We have studied solar water splitting with a composite semiconductor electrode, composed of an n-i-p junction amorphous silicon (a-Si, E g 1.7 eV) layer, an indium tin oxide (ITO) layer, and a tungsten trioxide (WO 3 , E g 2.8 eV) particulate layer. The n-i-p a-Si layer, which had more accurately a structure of n-type microcrystalline ( c) 3C-SiC:H (25 nm)/i-type a-Si:H (400 nm)/p-type a-SiC x :H (25 nm), was prepared on a TiO 2 -covered F-doped SnO 2 (FTO)/glass plate by a Hot-Wire CVD method. The ITO layer (100 nm thick) was deposited on the p-type a-Si by the DC magnetron sputtering method, and the WO 3 particulate layer was formed by a doctor-blade method, using a colloidal solution of commercial WO 3 powder of 10-30 nm in diameter. The composite electrode thus prepared was finally heat-treated at 300 C for 1 h. The anodic (water oxidation) photocurrent for the composite electrode in 0.1 M Na 2 SO 4 yielded an IPCE (incident photon to current efficiency) of about 6 % at 400 nm and was stable for more than 24 h. Besides, the onset potential lay a little (by about 0.05 V) more negative than the equilibrium hydrogen evolution potential, indicating a possibility of solar water splitting with no external bias. A preliminary result for the water photooxidation with an "nGaP/p-Si/Pt dot" electrode is also reported briefly.