Understanding the OER mechanism on hematite photoanodes using operando nonlinear optical characterization of interfacial field and chemical speciation

Hematite is an attractive photoanode for the oxygen evolution reaction (OER). However, insight into reaction mechanism, and thus the capability to optimize performance rationally, is limited by the challenge in relating band bending, surface state population and chemical reactivity. Conventionally the first two issues are addressed by electrochemical impedance spectroscopy (EIS) and photocurrent/voltage measurements while the third is addressed indirectly -- through measurement of Tafel slopes or current/voltage transients -- or through operando characterization of the electrode/electrolyte interface using bulk sensitive spectroscopies. Unambiguous interpretation of all of these types of measurements is challenging. Interpretation of EIS requires application of a non-unique model while bulk sensitive spectroscopic tools require exotic sample geometries or nontrivial normalization schemes to extract the desired interfacial signal. Ideally we would like an approach that allows observation of interfacial electric fields, surface state population and surface chemical speciation concurrently.

The nonlinear optical technique, Vibrational Sum Frequency (VSF) spectroscopy, is interface-specific by its symmetry selection rules and sensitive to interfacial electric fields. Here we describe our recent efforts using VSF spectroscopy to probe the potential dependent photovoltage and surface chemical speciation of a pulsed laser deposited hematite photoanode in alkalkine solution operando. We shown that the OER onset under illumination is characterized by a change in sign of the photovoltage and that this sign flip and the measured photovoltage magnitude can be reproduced by a model describing a conventional, single-hole adsorbate evolution OER mechanism with no adjustable parameters. The model results predict that that the rate limiting step in the OER is the production of interfacial Fe=O moieties. Resonant VSF measurements of the surface Fe-O spectral response confirm this prediction: we show that surface Fe=O appears only at OER onset. Recent work has argued that the OER on hematite can proceed by a multi-hole mechanism in which O-O bond formation is rate limiting. Direct spectroscopic observations of such a mechanism, however, have been limited. Our work conclusively shows that, at least for our PLD deposited hematite film, OER proceeds by a single hole mechanism.