Machine learning prediction of luminescence concentration quenching of Tb3+-doped inorganic phosphors
Year: 2026
Type: article
Abstract: Luminescent concentration quenching (LCQ) is a critical factor limiting the performance of rare-earth-doped inorganic phosphors, yet its prediction remains challenging due to the reliance on extensive experimental characterization. In particular, determining the optimal dopant concentration that maximizes emission intensity is both time-consuming and resource-intensive. In this work, we develop a machine learning framework based on Support Vector Regression (SVR) with a radial basis function (RBF) kernel to predict LCQ in Tb 3+ -doped inorganic phosphors using compositional and structural descriptors. The model was trained on a dataset of 122 experimentally reported compounds and optimized through feature selection and hyperparameter tuning. Despite the limited dataset size, the model achieves a coefficient of determination of R 2 = 0.72 on the test and R 2 = 0,75 set, indicating robust predictive performance. The trained model was then applied to estimate LCQ values for 162 previously unexplored host materials. Additionally, predictions for a subset of five compounds were validated against experimental data, showing good agreement. These results demonstrate that the proposed SVR-based approach provides physically meaningful predictions and can significantly accelerate the design and optimization of phosphor materials by reducing experimental trial-and-error.
Source: Journal of Luminescence
Authors: João P.S. Afonso, Otávio M. Neto, Pedro I.A. Ferreira, Marcos V. dos S. Rezende
Institution: Universidade Federal de Sergipe
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,850
Awards: G1516471320, G7802646218
OpenAlex ID: W7163575601