The Fermi level position of a halide perovskite semiconductor can be controlled by the substrate type underneath. Earlier photoemission spectroscopy (PES) studies found that n-type substrates caused the Fermi level of the perovskite to be closer to the conduction band minimum while p-type substrates moved the Fermi level closer to the valence band maximum, which could seemingly correspond to a change of the doping type in the perovskite film.^[1][2] This effect has strong implications on the control of the band alignment between perovskite film and adjacent transport layer and could thus enable us to optimize charge carrier extraction, as well as the design of perovskite solar cells and their efficiency. 

Our recent results of ultraviolet and X-ray photoemission spectroscopy (UPS/XPS) measurements and Kelvin probe force microscopy (KPFM) of various halide perovskite films deposited on p-type NiO_x and n-type TiO₂ confirm this hypothesis. In this contribution however, we show by using PES and KPFM measurements that deviations from this energy level alignment exist, i.e. for the case of triple cation mixed halide perovskite deposited on p-type and n-type Si substrates.

KPFM is a non destructive method, does not require difficult sample preparation and allows the observation of the evolution of the perovskite’s surface properties under environmental conditions. Using a conductive tip, KPFM can be employed to record the contact potential difference (CPD) between the tip and the sample, which reflects the difference in their work function. For a given and constant work function of the tip, the local work function of the sample’s surface can then be presented in a high-resolution CPD maps, while at the same time showing the surface topography at the nanoscale. By performing the experiment under illumination and subtracting the CPD signal in the dark from the CPD signal under illumination we image the surface photovoltage (SPV). The sign of the SPV signal indicates the type of doping and surface band bending of the studied material, and is positive for n-type and negative for p-type materials ^[3]. In the complementary PES experiments we observe shifts in the spectra comparing measurements performed in the dark and with additional light bias to confirm the sign and magnitude of the SPV. With this combined approach we find that the Fermi level position and band bending in the case of perovskite films on top of doped Si substrates does not follow the trend seen for perovskites on oxide substrates.