Machine-Learning-Guided Prediction and Statistical Analysis of Electronic Bandgaps in Double-Perovskite Oxides for Photovoltaic and Optoelectronic Design
4 affiliations
- 1Department of Industrial and Production Engineering (IPE
- 2Bangladesh Army University of Science and Technology
- 3Department of Materials and Metallurgical Engineering
- 4Bangladesh University of Engineering and Technology
Abstract
Double-perovskite oxides (A2BB′O6) are attractive for photovoltaic and optoelectronic applications, but their bandgap depends on a compositional space too large for routine DFT screening. In this study, we build composition-only machine-learning models on 1,306 DFTcomputed (GLLB-SC) double perovskites, using 72 descriptors derived from elemental electronegativity, ionic/atomic radius, group and period position, Mendeleev number, and first ionization energy. Six regression algorithms were compared under 10-fold cross-validation where a tuned XGBoost model gave the best results (test R2 = 0.975, MAE = 0.188 eV, RMSE = 0.245 eV), outperforming all others with statistical confidence (Friedman χ2 = 45.14, p =1.4×10-7; Wilcoxon p < 0.005 throughout). SHAP analysis pointed to A-site group number and B-site electronegativity and radius as the main drivers of bandgap, matching the mechanisms reported in the original DFT study. These results demonstrate that interpretable, composition-only machinelearning models can reproduce DFT-level bandgap trends at negligible computational cost, providing a practical high-throughput screening tool for lead-free double-perovskite discovery.
Related
- Molybdenum cluster iodides for improved CsPbBr 3 perovskite photoelectrodes: a superatomic approach to stability, passivation, and optoelectronic performance
- Predicting the photovoltaic and optoelectronic potentials of eco-friendly kesterites Ag2PdMX4 (M = Sn, Ge; X = S, Se) through first-principles calculations
- Composition-engineered intrinsic stability in 2D perovskite optoelectronic synapses for neuromorphic vision
- Crystal structure of the rhombohedral vacancy-ordered perovskite (CH₃NH₃)₂[SnCl₆] for deep-UV optoelectronic applications.
- Solubility Limit and Optoelectronic Evolution of Heterovalent Bi3+ Doping in MAPbBr3 Perovskite Powders: A Study from Electrochemical Impedance Spectroscopy (EIS)
BibTeX
@misc{abid2026machinelearningguided,
title = {{Machine-Learning-Guided Prediction and Statistical Analysis of Electronic Bandgaps in Double-Perovskite Oxides for Photovoltaic and Optoelectronic Design}},
author = {Md Ragib Abid and Md Shafayatul Islam},
howpublished = {ChemRxiv},
year = {2026},
month = oct,
doi = {10.26434/chemrxiv.15010932/v1},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010932/v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Machine-Learning-Guided Prediction and Statistical Analysis of Electronic Bandgaps in Double-Perovskite Oxides for Photovoltaic and Optoelectronic Design
AU - Md Ragib Abid
AU - Md Shafayatul Islam
PY - 2026
DA - 2026/10/23
PB - ChemRxiv
DO - 10.26434/chemrxiv.15010932/v1
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15010932/v1
AB - Double-perovskite oxides (A2BB′O6) are attractive for photovoltaic and optoelectronic applications, but their bandgap depends on a compositional space too large for routine DFT screening. In this study, we build composition-only machine-learning models on 1,306 DFTcomputed (GLLB-SC) double perovskites, using 72 descriptors derived from elemental electronegativity, ionic/atomic radius, group and period position, Mendeleev number, and first ionization energy. Six regression algorithms were compared under 10-fold cross-validation where a tuned XGBoost model gave the best results (test R2 = 0.975, MAE = 0.188 eV, RMSE = 0.245 eV), outperforming all others with statistical confidence (Friedman χ2 = 45.14, p =1.4×10-7; Wilcoxon p < 0.005 throughout). SHAP analysis pointed to A-site group number and B-site electronegativity and radius as the main drivers of bandgap, matching the mechanisms reported in the original DFT study. These results demonstrate that interpretable, composition-only machinelearning models can reproduce DFT-level bandgap trends at negligible computational cost, providing a practical high-throughput screening tool for lead-free double-perovskite discovery.
ER -