Assessment of Thermoelectric Properties of Bi2Se3: Insights from Hybrid Functional Studies, Strain Engineering, and Machine Learning Methodology

Abstract

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.