Hole Extraction Beyond the Depletion Layer in Ge-Doped Hematite Photoanode

Abstract

As extensively discussed in the literature, photoelectrochemistry water splitting has enormous potential for the generation of green hydrogen. However, to obtain an efficient and economically viable photoelectrochemistry cell, it is necessary to use efficient semiconductor materials from the photogenerated charge separation point of view. In n-type semiconductors, based on transition metal oxides, where the diffusion length (Lp) is small compared to the absorption depth (α-1), i.e., αLp<< 1, the charge separation process will occur in the depletion layer region. Since this layer is typically tens of nanometers in size, only a fraction of the material exposed to light will contribute to the water photo-oxidation process, thus drastically restricting the efficiency of these semiconductor oxides. Hematite (α-Fe2O3) is an n-type semiconductor oxide widely explored as a photoanode for water photo-oxidation and is known to be a material dominated by the charge separation process in the depletion layer, following the Gartner-Butler model. In this way, we will only have photocurrent in Vappl>Vfb (Vappl is the applied potential, and Vfb is the flat band potential). Here we will explore the possibility of extracting holes at potentials smaller than Vfb in Ge-doped hematite photoanodes through rigorous measurements of Vfb and absorbed photon-to-current conversion efficiency (APCE) as a function of Vappl. We determined Vfb based on the Gartner–Butler analysis in the presence of a sacrificial reagent (H2O2) under different wavelength excitation and observed a dependence of the Vfb with the wavelength. The APCE measurements as a function of Vappl-Vfb clearly show the existence of photocurrent at potentials below Vfb, strongly suggesting that we are extracting holes even before the formation of the depletion layer (before the band bending). We estimate that the holes arrive from a region approximately 3nm close to the electrolyte-semiconductor interface. This phenomenon was only observed for Ge-doped hematite and not for pristine hematite.