Charge Transfer across the n-Gallium Phosphide(100) Photoanode/Electrolyte Interface during Photoelectrochemical Water Splitting

Abstract

A detailed understanding of charge transfer mechanisms at the electrode/electrolyte interface is important for the development of electrodes for efficient photoelectrochemical (PEC) processes. We have studied the charge transfer processes across the n-GaP(100) photoanode/electrolyte interface. In a 0.02 M HCl electrolyte, high photoanodic currents from the n-GaP(100) photoanodes related to photolytic water splitting were measured at low anodic potentials but these photocurrents diminished at cathodic potentials and high anodic potentials. Electrochemical impedance spectroscopy (EIS) was carried out for n-GaP(100) photoanodes at different potentials to analyze the relevant charge transfer processes. Our EIS results suggest that the adsorption of hydroxide on metal-like surface Ga and their subsequent oxidative transformations – formation of these surface species is most favorable at low anodic potentials – is the driving force for high photoanodic currents at low anodic potentials. III-V semiconductors are prone to corrosion during PEC water splitting processes with corrosion-related decrease of efficiency. Our n-GaP(100) photoanodes were surface-conditioned via oxidizing at 0.8 V vs RHE and subsequently hydrogenated to passivate the defects in the oxide film. After this preparation, water splitting was observed at potentials between 0 and 0.3 V. The Nyquist plots derived from our EIS measurements for the n-GaP(100) photoanode consist of three semicircles indicating three distinct charge transfer mechanisms. The plots were analyzed with two different equivalent electrical circuits for the electrode/electrolyte system. These equivalent electrical circuits allowed to evaluate the relevance of different possible charge transfer pathways from the n-GaP(100) photoanode to the electrolyte. For each equivalent electrical circuit, the potential dependence of the resistances and capacitances including constant phase elements were determined from the fit and then compared with the potential variation of the current in the cyclic voltammogram. In a theoretical study Bertoluzzi et al. discussed the PEC implications of charge transfer from an n-doped semiconductor to the electrolyte that takes place via two different pathways1: by a direct charge transfer from the valence band of the semiconductor to the electrolyte, and by an indirect charge transfer from the valence band to the electrolyte via surface states (or defect states at the surface). The latter gives rise to a pronounced maximum of the interface- or defect-related capacitance as a function of potential. Our EIS results suggest that at low anodic potentials, the effective charge transfer process occurs from the valence band of the n-GaP(100) photoanode to defect states present in a thin Ga2O3-like layer at the surface, and subsequently, from these defect states to the electrolyte. This indirect charge transfer pathway is derived from a pronounced maximum of the capacitance as a function of applied potential. Our experimental results also suggest that subsequent charge transfer from these defect states leads to the formation of a molecular hydroxide layer the surface of metal-like Ga regions at the n-GaP(100) photoanode. Metal-like Ga at the surface of the photoanode was identified by XPS. A tafel slop of 63 mV/decade measured at the onset potential of the photoanodic current strongly points to the presence of adsorbed metal hydroxide (GaOH) species and subsequent oxidation products GaO and GaOOH as surface intermediates for the oxygen evolution which are finally catalyzed to molecular oxygen2. Our electrochemical impedance spectroscopy results also suggest that at low anodic potentials, overpotentials for the formation of hydroxide absorbates and their subsequent oxidation are lowest whereas at higher anodic potentials these hydroxide absorbates are stable and do not oxidize further to produce oxygen. At cathodic potentials, on the other hand, these surface hydroxides are reduced. As a consequence, large PEC currents associated with oxygen formation are possible only at low anodic potentials. By analyzing and understanding the charge transfer processes and the effect of different electrochemical parameters on these processes, we were able to optimize the PEC conditions which finally led to photolytic water splitting without applied voltage or addition of catalysts. Bertoluzzi, L; Lopez-Varo, P; Tejada, J. A. J; Bisquert, J., J. Mater. Chem. A., 2016, 4, 2873.Antoine, O; Butel, Y; Durand, R., J. Electroanal. Chem. 2001, 499, 85.