Speaker
Description
ARPES(ARPES : angle-resolved photoemission spectroscopy) is a powerful tool for observing electronic structures in solid-state materials, widely used in characterizing quantum materials. Spatially resolved ARPES (Nano-ARPES) allows measurements on relatively inhomogeneous surfaces, due to its sub-micrometer beam size. However, challenges remain due to the complexity of the surface, particularly in positioning the area of interest，which only relies on the scanning photoemission microscopy (SPEM). However, the SPEM data is quite heavy, which needs to be well analyzed in order to distinguish the desired regions. This is important but time consuming.
In recent years, unsupervised clustering method has shown strong capabilities in automatically categorizing the ARPES spatial mapping dataset[1,2]. However, it is only for real space currently and usually has limited ability in distinguishing subtle differences caused by complex and variable scenarios, such as different layers and substrates. Here, we propose a novel method called High-order Unsupervised Clustering Approach (HUCA). Using the K-means clustering results/metrics for real space in different energy-momentum windows as the input of the second round K-means clustering for momentum space, the energy-momentum windows that exhibit subtle inhomogeneity in real space will be highlighted. It recognizes different types of electronic structures both in real space and momentum space in spatially resolved ARPES dataset. Moreover, some subtle band differences, such as band shift or splitting, can still be pointed out by HUCA. Our results demonstrate that the clustering accuracy and identification limit can be significantly improved by HUCA. This method can be used to quickly capture the areas of interest, and is especially suitable for unknown samples, extremely small areas and areas with dispersed distribution.
Furthermore, by combining HUCA with the ARPES data acquisition system, it will achieve online fine clustering and band structure extraction, opening an era of intelligent ARPES experimental data collection and promoting the widespread application of machine learning in high-dimensional data clustering processing.
[1] H.~Iwasawa et,al, "Unsupervised clustering for identifying spatial inhomogeneity on local electronic structures",npj Quantum Mater. 7, 24 (2022)
[2] C.~N et al,"K-means-driven Gaussian Process data collection for angle-resolved photoemission spectroscopy", Mach. Learn.: Sci. Technol. 1 045015 (2020)
| Abstract publication | I agree that the abstract will be published on the web site |
|---|
