Proceedings of Online nanoGe Fall Meeting 20 (OnlineNFM20)
Publication date: 4th October 2020
The combination of functional scanning probe microscopy with advanced signal processing techniques in recent years has enabled new discoveries in a wide range of energy materials. These “big data” methods merge the availability of affordable data storage with the advances of the broader data science community to extract hidden information from gigabytes of raw time-dependent cantilever response data. In this talk, we will discuss our work using data-driven scanning probe methods, in particular time-resolved electrostatic force microscopy, for analyzing mixed organic-inorganic halide perovskites in situ in response to illumination. Through signal processing of the raw cantilever deflection signal during photoinduced charging, it is possible to extract the photoresponse of materials at microsecond timescales via analysis of the instantaneous frequency or the reconstructed electrostatic force. Importantly, we show that in layered (n=1) perovskites it is possible to observe photovoltage dynamics with timescales comparable to ion motion or trap-mediated carrier motion, in contrast to device-level studies. Furthermore, these timescales exhibit strong spatial dependence, with grain centers showing faster response compared to grain boundaries. This result is confirmed by general mode scanning Kelvin probe microscopy as well as by unsupervised clustering methods like k-means. These data indicate that layered perovskite materials may be more defect-prone than previously thought. Lastly, we discuss our work on analysis of hyperspectral photoinduced force microscopy for studying the spatial distribution of components in layered perovskites.