Semiconductor Electrodes: XLIX . Evidence for Fermi Level Pinning and Surface‐State Distributions from Impedance Measurements in Acetonitrile Solutions with Various Redox Couples

Abstract

Capacitance-voltage (C-V) measurements were made for the single crystal semiconductors n-TiO2, n-CdS, n-InP, p-Si, p-GaAs, n-and p-WSe2, and n-MoSe2 in acetonitrile containing a number of redox couples whose potentials (Vredox) spanned a potential regime much wider than the bandgaps. The flatband potential (VrB) evaluated from capacitance-potential (C-V) measurements (Mott-Schottky plots) exhibited three types of behavior with varying solution redox potentials: (i) VFB varied monotonically with Vredo • for p-Si, p-GaAs, and n:InP; (ii) for n-TiO2 and n-CdS, VFB did not shift for couples located negative of the midgap potential, but varied monotonically for couples positive of this value; (iii) for the layer-type, compounds (MoSe2, WSe~),VF, was almost independent of Vredo• These differences were ascribed to differences in surface-state densities. For n-TiO2 crystals, (001) face etched with molten KHSO4 and reduced, evidence for surface states at two different potentials was obtained from the in-phase component of the total admittance. Tentative assignment of these states is to lattice defects. The states closer to the conduction band are assigned to oxygen vacancies and the deeper states to Ti (III). The densities of surface states (N~) evaluated from GJ(o vs. (0 plots for TiO2 and p-Si are around 101~ and 101~ cm-2, respectively. These two values represent different situations, i.e., while the former value of N~s is not sufficient for pinning the Fermi level, the latter value is sufficiently high for the occurrence of Fermi level pinning.

Photochemical and Photoelectrochemical Reduction of CO2

“…For ideal semiconductor/liquid junctions, no photoreduction of the redox species above the conduction and below the valence band edge should exist. It has been shown before, however, that the photoreduction of species with redox potentials more negative than the conduction band edges of p-type semiconductors (Ge, Si, InP, GaAs) is feasible (109)(110)(111)(112). This phenomenon is explained by Fermi-level pinning and/or unpinning of the band edges, which causes the photovoltage observed for a p-type semiconductor/liquid junction to become independent of the redox potentials of the electroactive species.…”

“…It can be concluded that as long as the reduction potential of the molecular catalyst is above the valence band edge of the p-type semiconductor, the photoreduction of the molecular electrocatalyst is feasible. Fermi-level pinning also places a limitation on the maximum photovoltage/open circuit voltage for a particular semiconductor/liquid junction to one-half the band gap of the semiconductor (109)(110)(111)(112). The stability of the semiconductor photocathode can be further enhanced by surface modification, either by covalent attachment or coating the surface with polymer films of the molecular electrocatalyst.…”

Photoinduced transformations in semiconductor­metal nanocomposite assemblies

“…The reduction potential increases from -0.96 V vs. SCE for anthraquinone to 1.26 V vs. SCE for thianthrene (see ref. [72]). The open cicrcuit photopotential corresponds to the difference between the two energy levels, viz., E′ F -E 0 redox .…”

A Photoelectrochemical Cell Based on an Apparently Supra‐Band‐Edge Reaction:

“…This band-edge shift with redox potential is a typical behavior of Fermi level pinning (34,36). Such a phenomenon was demonstrated with a very similar interface, p-Si immerged in a solvated electron solution (K + + e%(NH3)n) in liquid ammonia (23) where the dissolved K concentration could be controlled.…”