The impact of Kelvin probe force microscopy operation modes and environment on grain boundary band bending in perovskite and Cu(In,Ga)Se2 solar cells

TL;DR

This study demonstrates that AM-KPFM measurements on rough polycrystalline absorbers can produce misleading grain-boundary contrasts due to cantilever–sample height variations, whereas FM-KPFM in ultra-high vacuum provides a more faithful map of surface potential, revealing that GB band bending is negligible in non-air-exposed samples and that facet-related contrasts dominate. By comparing MAPI perovskite, epitaxial CISe, and polycrystalline CIGSe across ambient and UHV environments, the work highlights the critical roles of surface contamination, sample history, and measurement mode in interpreting GB physics. The findings urge stringent environmental control and the use of FM-KPFM to obtain artifact-free GB insights, informing passivation and device strategies. Collectively, the results suggest that GBs may not be the dominant source of band bending in high-quality polycrystalline absorbers, with surface facets and adsorbates playing major roles in the observed electrostatics.

Abstract

An in-depth understanding of the electronic properties of grain boundaries (GB) in polycrystalline semiconductor absorbers is of high importance since their charge carrier recombination rates may be very high and hence limit the solar cell device performance. Kelvin Probe Force Microscopy (KPFM) is the method of choice to investigate GB band bending on the nanometer scale and thereby helps to develop passivation strategies. Here, it is shown that amplitude modulation AM-KPFM, which is by far the most common KPFM measurement mode, is not suitable to measure workfunction variations at GBs on rough samples, such as Cu(In,Ga)Se2 and CH3NH3PbI3. This is a direct consequence of a change in the cantilever-sample distance that varies on rough samples. Furthermore, we critically discuss the impact of different environments (air versus vacuum) and show that air exposure alters the GB and facet contrast, which leads to erroneous interpretations of the GB physics. Frequency modulation FM-KPFM measurements on non-air-exposed CIGSe and perovskite absorbers show that the amount of band bending measured at the GB is negligible and that the electronic landscape of the semiconductor surface is dominated by facet-related contrast due to the polycrystalline nature of the absorbers.

Figures (19)

• Figure 1: (a) Contact potential differences $CPD_\mathrm{GB}$ – $CPD_\mathrm{Grain}$ at the GBs and at the grains for MAPI and CIGSe absorbers as a function of the reported PCE. (b) Schematic representation of the scanning mechanism during KPFM measurements on GBs, illustrating that for a constant probe-sample distance $z$, the cantilever-sample distance changes when entering the GB. Blue, red and green arrows depict the cantilever, cone and probe apex of the SPM probe with $H$ defining the cone height. The distances between the cantilever and the sample are marked with blue lines, $h_{G}$ and $h_{GB}$ represent the distance when the probe apex is at the grain and at the GB, respectively. The GB depth is defined as $\Delta h = h_{G} - h_{GB}$. (c) 3D illustration of the band diagram showing the vacuum level ($VL$), conduction band ($CB$), Fermi level ($E_\mathrm{F}$) and valence band ($VB$).
• Figure 2: KPFM data for a MAPI showing the topography (a-c) with the respective CPD maps (d-f) measured under UHV (FM-KPFM), ambient (AM-KPFM) and UHV (FM-KPFM) after exposing it to air. Topography and CPD maps are shown with an optimum contrast illustrated with a scale bar in the respective image. Facet-related contrast completely vanished after exposing the sample to ambient conditions.
• Figure 3: (a) topography and (b) workfunction map of a MAPI absorber measured with AM-KPFM under ambient conditions. (c) calculated AM-KPFM image taking into account only the changes in the electrostatic force caused by the surface roughness (details see text). (d) line profiles from the topography, from the measured and calculated AM-KPFM image.
• Figure 4: Ambient AM-KPFM, UHV FM-KPFM and simulated AM-KPFM acquired on an epitaxially grown CISe absorber. (a,b) AM-KPFM under ambient conditions, where (a) shows the topography and (b) shows the surface potential images. (c,d) FM-KPFM under UHV conditions, where (c) shows the topography and (d) shows the surface potential. (e) measured and (f) calculated AM-KPFM (details see text). (g) line profiles of the topography, of the measured and calculated AM-KPFM measurement.
• Figure 5: FM-KPFM results on an epitaxial CISe sample exposed to different levels of surface contamination. (a-d) topography and (e-h) workfunction maps. (a) and (e) represent the as-grown and non-air-exposed condition. (b), (f) represent the same non-air-exposed sample that was kept in the UHV environment for 20 months. (c) and (g) show topography and KPFM measurements after UHV annealing of the aged sample ((b),(f)). (d) and (h) depict the topography and workfunction map after UHV annealing of another piece of the same sample that was exposed to ambient conditions for 6 months and then annealed.