Skip to content
Open access

Active learning and explainable machine intelligence for lead-free chalcogenide perovskite photovoltaics: Gaussian process Bayesian optimization and SHAP interpretability applied to CaZrSe 3/CdZnS solar cells

Aug 2026 · Materials Research Express · Vol 13, pp. 176201 · 0 citations · 43 references
Physics

TL;DR

A Gaussian process (GP) surrogate model, guided by the expected improvement acquisition function, is integrated into the SCAPS-1D simulation workflow for CaZrSe 3/CdZnS heterojunction solar cells, providing fabrication-targeted design rules grounded in quantitative physical reasoning rather than single-variable parametric intuition.

Abstract

The global transition toward lead-free photovoltaics has intensified interest in chalcogenide perovskites with the formula ABX 3, where calcium zirconium triselenide (CaZrSe 3) stands out for its direct bandgap of approximately 1.35 eV, chemical stability under ambient conditions, and complete absence of toxic elements. Yet the device physics of CaZrSe 3 heterojunction solar cells remain poorly mapped, and the conventional approach of exhaustive parametric sweeps in SCAPS-1D, while rigorous, places a prohibitive computational demand on research groups seeking rapid, systematic optimization. This work addresses this bottleneck by integrating a Gaussian process (GP) surrogate model, guided by the expected improvement acquisition function, into the SCAPS-1D simulation workflow for CaZrSe 3/CdZnS heterojunction solar cells. A Latin hypercube initial sample of 403 simulations conditioned the GP; 201 sequential Bayesian optimization iterations then identified the global power conversion efficiency (PCE) optimum of 10.4927%, with Voc=0.5388 V, Jsc=27.43 mA cm −2, and FF = 70.99%, using 604 total evaluations, representing a 95.0% reduction relative to the full 12 100-point reference grid. Concurrently, an ensemble of random forest, XGBoost, and artificial neural network surrogates trained on the complete 12 100-point dataset achieves R2>0.997 for all performance targets, and SHapley Additive exPlanations analysis decomposes the PCE response surface to reveal that absorber acceptor doping ( NA) is the primary determinant of device performance, while second-order interaction values quantify the coupled recombination dynamics of NA and ETL donor density ND. The optimal device specification an absorber thickness of 0.70 μm, NA=5.00×1015 cm −3, an electron transport layer (ETL) thickness of 0.10 μm, and an ETL donor density of ND=1.00×1018 cm −3 provides fabrication-targeted design rules grounded in quantitative physical reasoning rather than single-variable parametric intuition.

Read PDF

Similar papers

Open access Oct 2026

DFT and Machine Learning Analysis of Eco‐Friendly Formamidinium Tin Iodide Perovskite Solar Cells: Optimizing Charge Transport Layers Through SCAPS‐1D Simulations

Tin‐based solar cells are the eco‐friendly alternative to the toxic lead‐based cells that are currently in use. In this context, Formamidinium Tin Iodide (FASnI 3 ) emerges as a promising candidate due to its superior performance, exhibiting higher efficiency compared to other options. FASnI 3 is a lead‐free ti...

Rabeya Khan, Md. Rasidul Islam, I. K. G. G. Apurba et al. · 1 citation
Open access 2026

Machine Learning-Assisted Optimization of Lead-Free Cs₂AgBiBr₆ Perovskite Solar Cells via ETL/HTL Selection: A Combined SCAPS-1D and Random Forest Approach

The performance of lead-free Cs2AgBiBr6 double perovskite solar cells (PSCs) was investigated using SCAPS-1D simulation, followed by machine learning (ML) assisted modelling of the optimised device. Four device architectures employing ZnO and WS2 as electron transport layers (ETLs) and Cu2O and CNTs as hole transport l...

N. L. Tirupathamma, Tarun Vasamsetti, Saswat Kumar Ram et al. · 0 citations
Open access Oct 2026

Machine‐Learning Assisted Optimized Photovoltaic Performance of Sb2S3/Sb2Se3 Heterojunction Using DFT and SCAPS‐1D

Antimony chalcogenides have earned significant attention as potential absorber materials owing to their suitable bandgaps, high absorption coefficients, and Earth abundance. This work presents a multiscale computational framework combining density functional theory (DFT), SCAPS‐1D simulations, and machine learning (ML)...

Anshu Dewangan, Arundhati Sharma, Ishi Pathariya et al. · 0 citations
Oct 2026

Integrated first-principles calculation and machine learning-assisted optimization of high-efficiency lead-free Sb 2 Se 3 /Cs 2 TiI 6 double-absorber solar cells

Perovskite solar cells (PSCs) are widely investigated due to their low fabrication cost, compatibility with lowtemperature solution processing and the continual improvement in power conversion efficiency (PCE). The efficiency of singlejunction architectures is restricted by the Shockley–Queisser limit, which constrai...

Qiao-Chu Sun, Yu-Hang Zhang, Jiang Zhao · 0 citations
Open access Oct 2026

Physics-grounded machine learning for inverse design of TiO2/CuInGaS2/rGO–CuxS quantum-dot solar cells using surrogate modeling and SCAPS-based validation

The inverse design of multilayer heterojunction solar cells is difficult because device efficiency depends on coupled structural and electronic parameters, while extensive SCAPS-1D parameter sweeps become costly as the design space grows. This work develops a physics-grounded machine-learning framework for TiO 2...

Ali Eftekhari, Peyman Eftekhari, Hossein Abdollahzadeh et al. · 1 citation

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.