Degradation forecasting for accelerated lifetime studies of organic photovoltaics via symbolic regression

Song Q, Le Corre VM, Heumüller T, Zhang H, Tang H, Peng Z, Bornschlegl A, Du X, Lüer L, Brabec C (2026)

Publication Type: Journal article

Publication year: 2026

Journal

EES Solar Royal Society of Chemistry (RSC)

DOI: 10.1039/d6el00116e

Abstract

As organic photovoltaics (OPVs) achieve higher stability, traditional long-duration testing fails to provide timely feedback, impeding rapid optimization and commercialization. Degradation forecasting offers a solution, yet it remains fundamentally challenged by the complex, nonlinear degradation kinetics of OPV devices, which complicate forecasting. Here, we present a machine learning pipeline based on Class Symbolic Regression (CSR) that accurately extrapolates device degradation by an order of magnitude beyond the measured window, enabling day-scale stability decisions. Our workflow achieves 3–4% test error in quantitatively forecasting future degradation using the first 20 h for preliminary screening (200 h) and the first 100 h for long-term stability testing (1000 h). This approach, validated across >40 donor–acceptor systems and >100 processing conditions, demonstrates generalization across key stability metrics. Additionally, the workflow is lightweight, requiring few resources and training, making it suitable for various OPV laboratories, thereby compressing OPV stability-testing protocols from months to days. This work will accelerate decision-making and enhance the R&D process for the commercialization of OPV technology.

Authors with CRIS profile

Qizhen Song Institute Materials for Electronics and Energy Technology (i-MEET) Thomas Heumüller Institute Materials for Electronics and Energy Technology (i-MEET) Hua Tang Institute Materials for Electronics and Energy Technology (i-MEET) Zijian Peng Institute Materials for Electronics and Energy Technology (i-MEET) Andreas Bornschlegl Institute Materials for Electronics and Energy Technology (i-MEET) Larry Lüer Institute Materials for Electronics and Energy Technology (i-MEET) Christoph Brabec Institute Materials for Electronics and Energy Technology (i-MEET)

Involved external institutions

Shandong-University (SDU) CN China (CN) University of Southern Denmark / Syddansk Universitet DK Denmark (DK)

How to cite

APA:

Song, Q., Le Corre, V.M., Heumüller, T., Zhang, H., Tang, H., Peng, Z.,... Brabec, C. (2026). Degradation forecasting for accelerated lifetime studies of organic photovoltaics via symbolic regression. EES Solar. https://doi.org/10.1039/d6el00116e

MLA:

Song, Qizhen, et al. "Degradation forecasting for accelerated lifetime studies of organic photovoltaics via symbolic regression." EES Solar (2026).

BibTeX: Download