Defect formation in CsSnI$_3$ from Density Functional Theory and Machine Learning

TL;DR

This paper combines density functional theory and machine learning to predict defect formation energies in CsSnI$_3$, identifying effective dopants to suppress p-type self-doping and improve optoelectronic performance.

Contribution

It introduces a novel integrated approach using DFT data and machine learning to predict defect energetics in Sn-based perovskites, highlighting key descriptors and effective dopants.

Findings

Y, Sc, Al, Nb, Ba, Sr are effective dopants

ML model accurately predicts defect formation energies

DFT-ML approach offers new insights into defect energetics

Abstract

Sn-based perovskites as low-toxic materials are actively studied for optoelectronic applications. However, their performance is limited by $p$-type self-doping, which can be suppressed by substitutional doping on the cation sites. In this study, we combine density functional theory (DFT) calculations with machine learning (ML) to develop a predictive model and identify the key descriptors affecting formation energy and charge transition levels of the substitutional dopants in CsSnI$_{3}$. Our DFT calculations create a dataset of formation energies and charge transition levels and show that Y, Sc, Al, Nb, Ba, and Sr are effective dopants that pin the fermi level higher in the band gap, suppressing the $p$-type self-doping. We explore ML algorithms and propose training the random forest regression model to predict the defect formation properties. This work shows the predictive capability of combining DFT with machine learning and provides insights into the important features that determine the defect formation energetics.