Virtual-Alloy Physics-Aware Semantic System for extrapolative prediction of perovskite solar cells
Abstract
Existing machine-learning models for perovskite solar cells struggle to predict devices made from new materials. Most treat device layers as arbitrary categorical labels and represent perovskite absorbers either through featureless composition vectors or through rigid crystal graphs that require a pre-solved unit cell. As a result, they cannot generalize to unrecorded alloys or unfamiliar cell architectures. We present the Virtual-Alloy Physics-Aware Semantic System (V-PaSS), an architecture built around the physical continuity of perovskite solid solutions. V-PaSS represents arbitrary mixed-cation, mixed-halide absorbers by interpolating relaxed crystal graphs of 23 pure ABX3 endpoint compounds via Vegard's law with learned bandgap bowing. For the device stack, it couples layer sequence and cell polarity (n-i-p or p-i-n) with interfacial band offsets and encodes chemical descriptions with a scientific language model. Evaluated under an institution-grouped protocol where test devices come exclusively from unseen laboratories, an untuned default configuration of V-PaSS achieves the lowest error among fourteen tested models (2.861% MAE, a 5.2% relative reduction over a gradient-boosted baseline at 3.017%, and outperforming the best tuned benchmark at 2.912%). V-PaSS pulls furthest ahead where test devices differ most from training: reducing error by 15.7-19.0% on unseen absorber alloys, by 16.1% on mixed absorbers from new laboratories, and widening its lead over the baseline by 3.6× on devices with unobserved transport layers. Combined with tree ensembles, V-PaSS sets a new state of the art across all five benchmark splits.
Related
- Physics-statistics decoupled operators for ensemble-averaged intensity prediction in turbulent laser propagation
- Physics-Informed Latent Neural Operator for Three-Dimensional Compressor Cascade Flow Prediction
- SteelAgent: LLM-Orchestrated, Physics-Informed Steel Property Prediction with Audited Generalization and Calibrated Uncertainty
- AI-EXAFS: Physics-Informed Prediction of EXAFS from Atomic Structure.
- A virtual physicist framework integrating dose prediction and large language models for automated treatment planning
BibTeX
@misc{han2026virtualalloy,
title = {{Virtual-Alloy Physics-Aware Semantic System for extrapolative prediction of perovskite solar cells}},
author = {Qiwei Han and Chi Zhou},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.18949},
eprint = {2610.18949},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.18949},
}
RIS
TY - UNPB
TI - Virtual-Alloy Physics-Aware Semantic System for extrapolative prediction of perovskite solar cells
AU - Qiwei Han
AU - Chi Zhou
PY - 2026
DA - 2026/10/13
PB - arXiv
DO - 10.48550/arxiv.2610.18949
UR - https://arxiv.org/abs/2610.18949
AB - Existing machine-learning models for perovskite solar cells struggle to predict devices made from new materials. Most treat device layers as arbitrary categorical labels and represent perovskite absorbers either through featureless composition vectors or through rigid crystal graphs that require a pre-solved unit cell. As a result, they cannot generalize to unrecorded alloys or unfamiliar cell architectures. We present the Virtual-Alloy Physics-Aware Semantic System (V-PaSS), an architecture built around the physical continuity of perovskite solid solutions. V-PaSS represents arbitrary mixed-cation, mixed-halide absorbers by interpolating relaxed crystal graphs of 23 pure ABX_3 endpoint compounds via Vegard's law with learned bandgap bowing. For the device stack, it couples layer sequence and cell polarity (n-i-p or p-i-n) with interfacial band offsets and encodes chemical descriptions with a scientific language model. Evaluated under an institution-grouped protocol where test devices come exclusively from unseen laboratories, an untuned default configuration of V-PaSS achieves the lowest error among fourteen tested models (2.861% MAE, a 5.2% relative reduction over a gradient-boosted baseline at 3.017%, and outperforming the best tuned benchmark at 2.912%). V-PaSS pulls furthest ahead where test devices differ most from training: reducing error by 15.7-19.0% on unseen absorber alloys, by 16.1% on mixed absorbers from new laboratories, and widening its lead over the baseline by 3.6× on devices with unobserved transport layers. Combined with tree ensembles, V-PaSS sets a new state of the art across all five benchmark splits.
ER -