Major Contributions
 

1. A fine-tuned CHGNet machine-learning interatomic potential (MLIP) combined with DFT/AIMD and three-stage Monte Carlo sampling enables large-scale, near-DFT-accuracy investigation of O3 vs. P2 stacking phase stability in high-entropy and low-entropy layered oxide SIB cathodes.

2. High-entropy oxides exhibit stronger Na–TMO₂ interactions, broader O–TM bond distributions, and smaller interlayer distance ratios, revealing conformational entropy as a decisive factor favoring O3-phase stabilization where conventional ionic-potential descriptors are insufficient.

3. Jahn–Teller distortions of Mn are mitigated in high-entropy oxides, enhancing structural stability, and the MLIP framework provides an efficient platform for rational design and computational screening of next-generation SIB cathode materials.

主要貢獻
 

1. 精調 CHGNet 機器學習原子間勢（MLIP）結合 DFT/AIMD 與三階段蒙特卡洛取樣，以接近 DFT 精度大規模研究高熵與低熵層狀氧化物鈉離子電池正極的 O3 相與 P2 相穩定性。

2. 高熵氧化物呈現更強的 Na-TMO₂ 相互作用、更寬的 O-TM 鍵長分佈與更小的層間距比率，揭示構象熵在傳統離子勢描述符不足的情況下是決定 O3 相穩定性的關鍵因素。

3. 高熵氧化物中 Mn 的 Jahn-Teller 畸變得到抑制，提升結構穩定性，MLIP 框架為下一代鈉離子電池正極材料的理性設計與計算篩選提供高效平台。