Understanding the origin of photoelectrode performance enhancement by probing surface kinetics† †Electronic supplementary information (ESI) available: Calculations of Gärtner currents, JV curves of films, further IMPS data. See DOI: 10.1039/c5sc04519c

Probing the surface kinetics of different hematite electrodes with and without surface passivations.


IMPS Fitting to Steady State Currents
Mott-Schottky analysis was used in order to determine the flat band potential and the doping density of aH electrodes. The measurements were made in the dark with a CH Instruments CHI604C potentiostat, in 1 M NaOH while stirring the solution. Both 100 Hz and 1000 Hz were sampled from 0.5 V to 1.1 V. The data set obtained was plotted using the following relation: A linear line was found and the donor density of aH was determined to be 2.65 x 10 19 cm 3 . For the flat band potential a value of 0.73 V RHE was obtained, consistent with other aH electrodes tested and previous values reported for aH.
S3 Figure S2. A Mott-Schottky plot of aH at 1000 Hz.
We next calculated the width of the space charge region using the following equation: Where ε = 60, ε 0 = 8.85 x 10 12 F/m, V app -V fb is the magnitude of the space charge region assuming that all applied potential contributes to increased band bending, where V fb = 0.73 V vs.
RHE, and N D = 2.65 x 10 19 cm -3 for the aH sample used here. With the space charge region we were now able to fully calculate the Gartner current in our aH electrodes. Where the Gartner current is defined by: Here J is the Gartner current, J 0 = 2.   In order to more closely fit our model, we next corrected for the potential drop across the Helmholtz layer across the studied potential range, 0.5 -1.2 V vs. RHE. This can be done by calculating the change in potential across the Helmholtz layer with the following equation:  Figure S4. IMPS measurements of all six samples tested for this study. The high frequency crossing point with the real photocurrent axis was used to normalize the imaginary and real axes. This allows for an easy comparison between samples at different applied potential, and for the transfer efficiency to be easily determined. The black trace represents a 0.4 V RHE applied potential for each graph with 0.1 V steps up to 1.4 V RHE . The order of the traces goes, black, red, green, blue, cyan, magenta, dark yellow, olive green, orange, purple, pink. Each color represents the same applied potential in each graph.

IMPS at Varying Light Intensities
At lower light intensities it is expected that the quasi Fermi level splitting in light will be reduced. This will lead to lower rate constants in a system with Fermi level pinning. In addition, since the band bending will be determined by the applied potential, the surface recombination should remain constant. Figure S6 shows that this is indeed the case, with figure S7 providing a schematic of the system at different light intensities. Figure S6. IMPS measurements made at 5 -90% of the LED intensity, where 90% is 134 mW/cm 2 . The samples tested here are bare electrodes. The figure on the left shows the transfer rate constant, which increases monotonically with light intensity. The figure on the right shows the recombination which remains nearly constant at all light intensities probed. All measurements were made on rgh2 at 0.6 V vs RHE, which is slightly after the onset of the photocurrent for these photoelectrodes. Figure S7: A schematic of the bands with different light intensities. When the light intensity is greater, the Fermi level splitting is greater, reducing the activation energy for water oxidation, as seen in Figure S6. The applied potential for the low intensity and high intensity is the same, resulting in same degree of band bending (Φ B ), resulting in similar rate constants for recombination, also as seen in Figure S6.