Interpretation of Mott-Schottky Plots of Photoanodes for Water Splitting

TL;DR

This paper critically examines the interpretation of Mott-Schottky plots in photoanodes for water splitting, revealing that commonly reported high doping densities are likely overestimated due to measurement artifacts.

Contribution

It demonstrates that electron injection effects can mimic depletion capacitance, leading to overestimated doping densities in Mott-Schottky analysis of photoanodes.

Findings

High doping densities reported are close to the resolution limit, questioning their credibility.

Injection-induced capacitance steps can be mistaken for depletion capacitance.

Most literature doping values are likely overestimated due to measurement artifacts.

Abstract

A large body of literature reports that both bismuth vanadate and haematite photoanodes are semiconductors with an extremely high doping density between 10^18-10^21 cm^-3. Such values are obtained from Mott-Schottky plots by assuming that the measured capacitance is dominated by the capacitance of the depletion layer formed by the doping density within the photoanode. In this work, we show that such an assumption is erroneous in many cases because the injection of electrons from the collecting contact creates a ubiquitous capacitance step that is very difficult to distinguish from that of the depletion layer. Based on this reasoning, we derive an analytical resolution limit that is independent of the assumed active area and surface roughness of the photoanode, below which doping densities cannot be measured in a capacitance measurement. We find that the reported doping densities in literature lie very close to this value and therefore conclude that there is no credible evidence from capacitance measurements that confirms that bismuth vanadate and haematite photoanodes contain high doping densities.