Thermoelectric properties in topological insulator Bi2Se3 are explored with multifaceted strategies, i.e., hybrid functional with strain and artificial intelligence methodology. The assessment with the experimental band gap values recognizes the limitations of conventional functional and the effectiveness of screened hybrid functionals. A thorough investigation into the impact of biaxial and uniaxial strain on thermoelectric parameters uncovers distinctive behaviors in n‐type and p‐type Bi2Se3, providing insights into optimal strain conditions for improved performance. Furthermore, the studies on the role of topologically non‐trivial surface states (TNSS) in thermoelectric properties reveal that TNSS significantly dominate electronic transport. Dual scattering time approximation elucidates the segregation of thermoelectric transport contributions from bulk and surface states, highlighting the importance of controlling the relaxation time ratio for enhanced thermoelectric performance. Additionally, the prediction of thermoelectric properties using Random Forest and Neural Networks models showcase impressive agreement with density functional theory predictions across varying temperatures, offering a powerful tool for understanding complex temperature‐dependent trends in thermoelectric properties. In summary, this interdisciplinary study presents a unique approach to advancing the understanding and optimization of thermoelectric properties in Bi2Se3. It provides a comprehensive framework for tailoring material behavior for diverse thermoelectric applications.

中文翻译：

---

Bi2Se3 热电性能评估：来自混合功能研究、应变工程和机器学习方法的见解


通过多方面的策略探索拓扑绝缘体 Bi2Se3 中的热电性能，即应变和人工智能方法的混合泛函。使用实验带隙值进行评估认识到常规泛函的局限性和筛选的混合泛函的有效性。对双轴和单轴应变对热电参数影响的深入研究揭示了 n 型和 p 型 Bi2Se3 的独特行为，为提高性能的最佳应变条件提供了见解。此外，关于拓扑非平凡表面态 （TNSS） 在热电特性中的作用的研究表明，TNSS 在电子传输中占主导地位。双散射时间近似阐明了本体和表面状态的热电传输贡献的分离，强调了控制弛豫时间比以增强热电性能的重要性。此外，使用随机森林和神经网络模型对热电特性的预测与不同温度下的密度泛函理论预测具有令人印象深刻的一致性，为理解热电特性中复杂的温度相关趋势提供了强大的工具。总之，这项跨学科研究提出了一种独特的方法来促进对 Bi2Se3 热电特性的理解和优化。它为各种热电应用定制材料行为提供了一个全面的框架。