Luminescence properties of fluoride laser crystals: a joint prediction model based on universal atomic embedded graph neural networks and transfer learning.

Predicting the luminescence properties of fluoride laser crystals in the 3-5 µm mid-infrared bands is challenging due to data scarcity and inefficient empirical methods. To address this, we propose a joint prediction model (UAEGNN-TL) that combines a Universal Atomic Embedded Graph Neural Networks with Transfer Learning. The UAEGNN component utilizes an enhanced universal atomic embedding strategy to represent the intrinsic features of constituent atoms within a crystal structure. This is synergistically combined with a transfer learning (TL) framework, which leverages knowledge from source tasks to mitigate the data scarcity limitation in the target task. Our results demonstrate that the UAEGNN architecture effectively captures the complex structure-property relationships governed by atomic interactions in fluoride crystals. Its predictive accuracy significantly outperforms mainstream models such as Crystal Graph Convolutional Neural Network (CGCNN), achieving a Mean Absolute Error (MAE) of 0.065 with a reduction of nearly 39% compared to the CGCNN model. Its coefficient of determination (R 2) reached 0.98. Building on this foundation, we incorporated a transfer learning strategy to construct the UAEGNN-TL model. Notably, the UAEGNN-TL model maintains robust predictive performance even when fine-tuned on a limited dataset of only 200 annotated fluoride laser crystals, achieving an R 2 of 0.83. Leveraging this joint model, we screened the luminescence properties of over a thousand fluoride crystals, enabling the rapid identification of promising candidates for the 3-5 µm band. Our analysis identifies dopant ions such as Er3+, Dy3+ and Ho3+ as the most favorable centers for inducing efficient mid-infrared luminescence. This work establishes a new paradigm for the efficient design and precise screening of mid-infrared fluoride laser crystals, thereby offering significant value for accelerating the development of advanced mid-infrared laser materials.