Bandgap Prediction of Perovskite Solar Cell Using Multiple Regression Model Towards Higher Efficiency

EasyChair Preprint 9440

Abstract

Perovskite solar cells have been emerged as most promising third generation solar cell technology. In past few years, the efficiency of PSCs has increased drastically from 3.8% to 25.6% for lab scale devices in single junction architecture. The traditional way to develop materials is usually based on trial and error, continuous synthesis methods which are time consuming and costly. This motivates the use of autonomous experimentation toolkits like linear or multiple regression (MR) or various machine learning algorithms. A dataset containing 100 plus data points are collected from various published papers and analyzed using multiple regression algorithm in excel. The multiple regression (MR) model is used to predict the bandgap of the perovskite with the formula Cs_aFA_bMA_(1-a-b)Pb(Cl_xBr_yI_(1-x-y))₃ which takes into account the compositional engineering of cations and halide anions for predicting the optimum bandgap. The bandgap predicted by this model has a R square error of 0.96 which indicates the power of this model in prediction of the bandgap of the perovskites from their compositions.

Keyphrases: Band gap Optimization, High Efficiency Perovskite Solar Cell, Multiple Regression

@booklet{EasyChair:9440,
  author    = {Debmalya Sadhu and Devansh Dattatreya and Arjun Deo and Debasis De},
  title     = {Bandgap Prediction of Perovskite Solar Cell Using Multiple Regression Model Towards Higher Efficiency},
  howpublished = {EasyChair Preprint 9440},
  year      = {EasyChair, 2022}}