Enhancing Superconductor Critical Temperature Prediction: A Novel Machine Learning Approach Integrating Dopant Recognition

• DOI: 10.1021/acsami.4c11997
• Publication Date: 2024-10-25T15:55:17Z
• Authors: Chengquan Zhong; Yuelin Wang; Yanwu Long; Jiakai Liu; Kailong Hu; Jingzi Zhang; Junjie Chen; Xi Lin

Doping plays a crucial role in determining the critical temperature (Tc) of superconductors, yet accurately predicting its effects remains significant challenge. Here, we introduce novel doping descriptor that captures complex influence dopants on superconductivity. By integrating with elemental and physical features within Mixture Experts (MoE) model, achieve remarkable R2 0.962 for Tc prediction, surpassing all published prediction models. Our approach successfully identifies optimal levels Bi2–xPbxSr2Ca2–yCuyOz system, predictions closely aligning experimental results. Leveraging this screen compounds from Inorganic Crystal Structure Database employ generative to explore new doped superconductors. This process reveals 40 promising candidates high superconductivity among existing hypothetical materials. explicitly accounting effects, our method offers powerful tool guiding discovery potentially accelerating progress high-temperature research opening avenues material design.