ICFgainAI: A Deep Learning Framework for Predicting Fluorescence Lifetime and Stimulated Emission Cross‐Section of Inertial Confinement Fusion Laser Gain Materials

ABSTRACT Nd‐based laser gain materials are essential components for target ignition in laser‐driven inertial confinement fusion (ICF). Fluorescence lifetime and stimulated emission cross‐section are two essential spectroscopic properties to achieve high‐energy, high‐repetition‐rate laser systems. Here, we construct an Nd‐based laser gain material dataset by systematically text‐mining 44,206 research articles, comprising 1,376 records of fluorescence lifetimes and stimulated emission cross‐sections data for 902 materials from diverse host families. We develop a deep learning framework ICFgainAI, to predict fluorescence lifetime and stimulated emission cross‐section directly from chemical composition and dopant information. ICFgainAI effectively captures the influence of chemical compositions on spectroscopic properties, which demonstrates high predictive accuracy and strong generalization across different material families. By ICFgainAI prediction, we identify and experimentally validate two major systems of laser materials with different structures. We discover two alkaline‐earth laser gain crystals, 0.1Y‐:0.005Nd and (0.05Y, 0.05Gd):0.005Nd, exhibiting long fluorescence lifetimes exceeding 500 µs accompanied with moderate emission cross‐sections of around . We also discover that the rare‐earth fluoride 0.05Sr‐:0.008Nd crystal exhibits a long fluorescence lifetime of 670 µs. These results establish the ICFgainAI as a powerful tool for predicting spectroscopic properties and accelerating the rational design of laser gain materials for ICF applications.