Understanding How Interfacial Properties Affect the Photovoltage of Metal-Insulator-Semiconductor (MIS) Photoelectrodes

A metal-insulator-semiconductor (MIS) structure is a highly promising photoelectrode architecture to promote photoelectrochemical (PEC) reactions, such as solar water splitting, for the storage of solar energy in chemical fuels that can be transported and used on demand. The semiconductor absorbs photons, creating electron-hole pairs; the insulator facilitates the separation of these excited electron-hole pairs; and the metal collects the desired charge carrier and facilitates its use in the fuel-forming reaction. The ultrathin insulator also stabilizes the semiconductor by serving as a physical barrier that mitigates corrosion by the electrolyte. Despite these attractive features, the MIS fuel-formation rates are significantly limited by the reaction driving force, i.e. , the MIS photovoltage. Previous work has shown that the MIS photovoltage can be improved substantially by tuning the insulator properties ( e.g. , thickness and band structure) and by using metal nanoparticles instead of a planar metal film. These studies explain that tuning the interfacial properties affect the selectivity of carrier transmission into the metal, but there lacks a clear understanding of how carrier selectivity affects the MIS photovoltage. To design MIS structures with enhanced photovoltages, it is critical to understand how interfacial properties affect the MIS photovoltage. Such a fundamental understanding is only possible with a multiphysics continuum model of MIS photoelectrodes. This talk will present our recent efforts to simulate MIS photoelectrodes performing PEC water splitting by using a comprehensive continuum model that accounts for photon absorption, carrier and electrolyte transport, interfacial tunneling, and reaction kinetics. Our simulations have identified how the insulator properties modulate carrier tunneling rates and how these rates impact the MIS photovoltage. Specifically, tuning the insulator properties affects the probability of tunneling, which alters the carrier concentrations at the semiconductor surface. Increasing the carrier concentrations at the semiconductor surface increases the ideal photovoltage ( i.e. , the difference in quasi-Fermi levels at the semiconductor surface) because of a reduction in the magnitude of the electric field and an increase in the quasi-Fermi level of the minority carrier. This understanding of how insulator properties modulate carrier buildup and, thereby, affect the MIS photovoltage helps identify optimal insulator properties. These predicted optimal properties of the insulator are a large bandgap (resulting in a high band offset with the semiconductor), a high dielectric constant, and a thickness of ~ 1.2 nm. Insulating materials that have a large band offset with the semiconductor are predicted to have high MIS photovoltages by creating an appreciable interfacial tunneling resistance that promotes carrier buildup at the semiconductor surface, and the moderate insulator thickness allows for unimpeded tunneling of the desired carrier. Increasing the insulator thickness beyond its optimal value results in hindered tunneling and a very large concentration of carriers at the semiconductor surface. This extremely large carrier concentration leads to a very large density of charges trapped at interfacial defect states, which are neutralized by a buildup of image charge in the metal. The accumulation of charges across the insulator results in a significant drop in potential that reduces the MIS photovoltage. We find this deleterious potential drop is significantly reduced by using metal nanoparticles, rather than a planar film, because the electrolyte neutralizes most of the trapped charges. This electrolyte charge screening phenomenon reduces the image charge buildup in the metal and, thereby, the drop in potential between the semiconductor and the metal. Thus, MIS structures that use metal nanoparticles (np-MIS) exhibit greater photovoltages than MIS structures with a planar metal film primarily because of a reduction in the potential drop across the insulator. In addition, because of electrolyte charge screening, the optimal insulator thickness in np-MIS structures is larger than that in MIS structures, which enables even further improvements in the photovoltage and provides greater corrosion resistance. Our model predicts the phenomenon of electrolyte charge screening can be further leveraged by using a low coverage of nanoparticles on the insulator (~ 10%), which can be achieved using low loadings and/or small nanoparticles. In all, the results of our study provide significant insights into how interfacial properties affect the MIS photovoltage and provide predictions of the optimal values for the MIS properties. Such understanding is vital for the development of next-generation MIS photoelectrodes that can enable attainment of high rates of solar-to-chemical energy conversion.