Proceedings of MATSUS Spring 2025 Conference (MATSUSSpring25)
DOI: https://doi.org/10.29363/nanoge.matsusspring.2025.346
Publication date: 16th December 2024
Extracting relevant material properties from experimental measurement is challenging, especially in the field of organic semiconductors. The models used to fit and reproduce experimental results are complex with multiple correlated parameters, which render the use of such model to extract relevant material properties very complicated. To overcome such limitations, we consider in this work the use of Bayesian inference for parameter estimation. Bayesian inference is a powerful tool to extract parameters distribution considering the experimental observations and the models considered [1].
In this study, we apply Bayesian inference techniques to analyze temperature-dependent photoluminescence spectra obtained from organic solar cells. We model the photoluminescence spectra of organic semiconductor films using a semi-classical marcus-levich-jortner expression [2, 3]. We model the spectra under different temperature and reproduce the change in spectral shape and relative intensity. Using the model and a Bayesian inference approach, we extract distributions for the different relevant properties of interest such as: Energy of the first excited state, the static disorder in energy, the reorganization energies (low and high frequency) as well as the dominant frequency mode. Our approach provides robust parameter estimation and quantifies uncertainties, enabling more accurate characterization of organic semiconductor materials. The results demonstrate the effectiveness of Bayesian inference in unraveling complex material properties and guiding future research in renewable energy applications.