Machine Learning Augmented Discovery of Chalcogenide Double Perovskites for Photovoltaics

Abstract

Hybrid organic inorganic perovskite solar cells based on CH 3 NH 3 PbI 3 have drastically increased in efficiency over the past several years and are competitive with decades-old photovoltaic materials such as CdTe. Despite this impressive increase, significant issues still remain due to the intrinsic instability of CH 3 NH 3 PbI 3 which degrades into carcinogenic PbI 2 . Recently, double halide perovskites which use a pair of 1 + -3 + cations to replace Pb 2+ , such as Cs 2 InSbI 6 , and chalcogenide perovskites, such as BaZrS 3 , have been explored as potential replacements. In this work, double chalcogenide perovskites are explored to identify novel photovoltaic absorbers that can replace CH 3 NH 3 PbI 3 . Due to the large space of possible compounds, machine learning methods are used to classify materials as potential photovoltaic absorbers using data from the periodic table, eliminating wasteful computation. A random forest algorithm achieves a cross-validation accuracy of 86.4% on the constructed data set. Over 450 possible replacements are identified via traditional and statistical methods with

A Critical Review of Machine Learning of Energy Materials

“…More accurate bandgaps can be obtained computationally using more expensive methods such as hybrid functionals (HSE06) or GW calculations . Agiorgousis et al have calculated 220 double chalcogenide perovskites (A2BB′X6) with the screened hybrid HSE06 functional, followed by training RF and SVM classifiers using this data set. The percentage error of classifying whether the bandgap of given perovskite falls between 0.7 to 2.0 eV was 13.80 % for RF, and 31.63% for SVM.…”

Towards photoferroic materials by design: recent progress and perspectives

“…In the last few decades, there have been multiple studies on in silico design of novel materials for PV devices as well as new ferroelectrics. Recent examples include novel light-harvesting materials based on inorganic and metal-organic perovskites, oxynitride and sulfide materials, and known materials from existing databases [20][21][22][23][57][58][59][60][61][62][63][64]. In the case of ferroelectrics there has recently been some progress in obtaining design rules that can readily be applied to computational discovery of new materials exhibiting switchable spontaneous polarization.…”

Realization of BaZrS3 chalcogenide perovskite thin films for optoelectronics

“…Using machine learning, Agiorgousis et al isolated Ba2AlNbS6, Ba2GaNbS6, Ca2GaNbS6, Sr2InNbS6, and Ba2SnHfS6, out of 450 chalcogenide double perovskites, as the most promising photovoltaic materials [34]. Guided by computational screening of ternary sulfides, Kuhar et al…”