Surface-State-Controlled Fluence-Dependent Apce Behavior of p-GaAs-TiO₂-Pt Water Reduction Photocathode

The demand for reform in energy production and reduction in carbon dioxide emission lays emphasis on the development and mechanistic study of photocatalytic water splitting photoelectrodes. Understanding the bottlenecks of the photoelectrochemical (PEC) conversion efficiency in photocatalytic water splitting systems is essential to impart guidance to material selection and photoelectrode design. Absorbed photon-to-current conversion efficiency (APCE), which represents the fraction of absorbed photons converted into electrons in photocurrents, is the key metric in the evaluation of PEC performance. A large portion of studies have reported an illumination fluence-independent APCE under continuous wave (CW) illumination with power densities around 10~100 mW/cm2, whereas few cases have reported the fluence-dependent APCE performance. In the present work, we observe a rarely reported fluence-dependent APCE performance in the p-GaAs/5 nm TiO2/2 nm Pt photocathode under a low-excitation intensity regime. The APCE reaches close to unity at 0.005 mW/cm2 and gradually decreases to 60% at 1 mW/cm2. However, a theoretical explanation for this fluence-dependent APCE reduction is not established. We develop a steady-state kinetic model with the least parameters to decipher such a phenomenon. In this model, the surface electrons in the GaAs/TiO2/Pt photocathode are consumed by three pathways: 1) transferred to redox active species in solution via the conduction band, 2) trapped into surface states that serve as the reaction sites, which competes with the electron-hole recombination at such surface states, and 3) directly recombining with the surface holes. The APCE decrease at higher illumination intensity is attributed to the increasing recombination between the GaAs holes and the TiO2 electrons caused by the saturation of surface sites. A detailed parametric survey shows that the interfacial electron transfer rate, ks, from the trap states to solution redox species is the only parameter that tunes the extent of the APCE reduction. With higher ks, APCE displays a smaller reduction in response to increasing fluences. Thus, the fluence-dependent feature of the APCE is an indicator that the surface reaction is the efficiency-determining step. Through this work, we establish a kinetic model for the semiconductor photoelectrode with catalyst layer and successfully explain the cause to the fluence-dependent APCE performance. These significant findings can be universally applied to address the function of catalytic layer (nanoparticle or molecular) and surface modifier in hybrid photoelectrode systems that exhibit fluence-dependent APCE.