Modeling Thickness Dependencies of Electrical Parameters and Nanostructure Formation in Vapor-Phase Condensates of LAST Compounds Using Machine Learning

Abstract

The article examines the modeling of thickness dependencies of the electrical parameters of thin films based on LAST compounds (Pb-Ag-Sb-Te) using machine learning methods. The aim of the work is to optimize the process of vapor-phase condensate deposition to improve the thermoelectric properties of materials. The focus is on studying the effect of film thickness and nanocrystal size on electrical conductivity and charge carrier mobility. For the first time, machine learning methods are applied in the article to predict electrical parameters based on experimental data. The XGBoost model was used to predict the behavior of electrical conductivity and other parameters depending on changes in film thickness, which contributes to improving the efficiency of the film formation process. The research results show that proper optimization of deposition parameters can significantly improve the thermoelectric characteristics of materials, which is important for applications in energy and electronic devices. Thus, the article demonstrates the potential of machine learning as a tool for improving technological processes in the production of nanostructured films of LAST compounds.