Contactless measurement of the photovoltage in BiVO4 photoelectrodes

The power output of photoelectrochemical devices for solar energy-to-fuel conversion is determined by the photovoltage of the junction under illumination. In the presence of fast redox couples, the photovoltage can be obtained directly from current-voltage measurements of the device. However, for slow redox couples (H+/H2, O2/H2O) used in solar fuel photoelectrodes, photovoltage measurements are not straightforward, due to the kinetic overpotentials during charge transfer. Here we show that the photovoltage of BiVO4 electrodes in contact with fast electron donors KI, Na2SO3 or H2O2 or K4Fe(CN)6 can be measured in a contactless way with vibrating Kelvin probe surface photovoltage (SPV) spectroscopy. The photovoltage varies with illumination wavelength and intensity and matches the open circuit potential of the electrodes, obtained separately from electrochemical measurements. Plots of the photovoltage versus irradiance can be used to predict the oxidizing power of each electrode under zero applied bias. Except for K4Fe(CN)6, which causes shunting in the BiVO4 electrode, photovoltage values correlate well with the built-in potential of each junction. The ability to obtain photovoltage information through contactless SPV measurements will be useful in the search for solid-liquid junctions with superior energy conversion properties.

Methods
Surface photovoltage (SPV) data was obtained with the vibrating Kelvin probe technique, using a semi-transparent (60%) 3.0 mm diameter gold Kelvin probe (Kelvin Probe S, Delta PHI Besocke) and a Besocke Kelvin Control.  Measurements were conducted in a custom-made chamber under vacuum (≤ 2 × 10-4 mbar) or in H2O-saturated N2 gas. Samples were coated  with 10-15 mL of liquid electrolytes using a micropipette and covered with a glass cover slip (Fisher Scientific, 0.17 to 0.25 mm thickness). For the acquisition of full spectra, n-GaP was illuminated through the Kelvin probe using light from a 300 W tungsten-halogen lamp filtered through an Oriel Cornerstone 130 monochromator (1-10 mW cm-2). Scans were performed from 9600 cm-1 to 40000 cm-1. The contact potential difference (CPD) data were corrected for drift effects by subtracting a fitted logarithmic curve of a dark scan from the spectral scan. Intensity-dependent measurements were performed with an air-cooled 400 nm LED array connected to a DC power supply. The voltage was regulated between 33 and 43.2 V to produce irradiances of 5x10-4 – 60 mW/cm2 at the sample surface, as measured by a photometer equipped with a GaAsP UV-Vis detector (International Light Technologies, Inc), and after correction using a 60% transmission value for the Kelvin probe.

Linear sweep voltammograms were recorded in an electrochemical cell controlled by a Gamry Reference 600 potentiostat. Working electrodes and electrolytes were prepared as described below. All electrolytes were purged with N2for 20 mins before and during the measurements, except for H2O2 solutions, which were purged with air.  

A calomel electrode (3.5 M KCl) was used as the reference and a platinum wire as the counter electrode. The cell was calibrated using the standard reduction of potential of hexacyanoferrate (0.358 V vs NHE) and potentials were then adjusted to the RHE scale using the formula VRHE = VNHE + 0.0592 × pH. A 400 nm LED was used for illumination. Measurements were taken at 20 mV/s intervals with light chopped approximately every 3 s while stirring the electrolyte.

Open circuit potential (OCP) measurements were conducted using a two-electrode system consisting of the sample as working electrode and a calomel electrode (3.5 M KCl) as reference and counter electrode. A monochromatic 400 nm LED with an intensity range of 0.0009 − 305 mW/cm2 at the working electrode was employed for illumination. 

Chronoamperometry was performed in a 2-electrode set-up, with the sample as the working electrode and a Pt wire as both the counter and reference electrode. A 0 V bias was applied and a 400 nm LED (250 mW/cm2) was used as the light source.