Professor Fuqiang Huang’s team published a research paper titled "Exceptional layered cathode stability at 4.8 V via supersaturated high-valence cation design" in Nature Energy. Through sodium (Na) assistance, the study achieved a highly enriched Ti4+ population in LiNi0.8Co0.1Mn0.1O2, thereby markedly enhancing cycling stability at high voltages.
To address the design challenges of high-entropy oxide cathode materials for sodium-ion batteries, Prof. Fuqiang Huang’s team proposed a Hybrid-Flow Machine Learning (HFML) framework that integrates ensemble learning, unsupervised learning, and Bayesian optimization, enabling efficient screening of stable O3-type high-entropy oxide cathodes from over 2.25 million candidate structures.
The team led by Fuqiang Huang collaborated with Jinjin Li's group to address the performance-prediction challenges arising from data scarcity and structural complexity in high-entropy materials by developing the CGformer model, which integrates Transformer architectures with crystal graph networks and incorporates unsupervised clustering and transfer learning.