Data-Driven Exploration of NaXTe2 (X=Al, Ga, In): From High-Throughput Screening to Tailored Optoelectronic Functionalities

Abstract

The rational design of semiconductor materials with tunable properties for targeted optoelectronic applications remains a significant challenge, particularly in moving beyond traditional paradigms limited by d-electron localization and instability. To address this, we introduce and execute a closed-loop, data-driven research strategy. This paradigm begins with the machine-learning-assisted construction and high-throughput screening of a proprietary library of 723 Na/K-based ABC 2 compounds, which efficiently identifies the ternary telluride NaXTe 2 (X = Al, Ga, In) family as a promising candidate embodying the "alkali metal genetic engineering" concept. Employing first-principles calculations within this workflow, we systematically decode the structure-property relationships in NaXTe 2 . Our results reveal that simple cation substitution (Al → Ga → In) serves as a powerful single variable for synergistic, cross-dimensional performance tuning. This is achieved by modulating the interplay of lattice strain and bonding character, which continuously shifts the electronic bandgap from a direct 1.28 eV (NaAlTe 2 ) to a direct 0.21 eV (NaGaTe 2 ), while concurrently evolving the mechanical properties from conventional toughness to the high toughness and pronounced anisotropy characteristic of NaInTe 2 . Comprehensive calculations confirm the thermodynamic, dynamic, and thermal stability of all three compounds, underpinning their experimental viability.The distinct properties map to specific applications: NaAlTe 2 (optimal direct bandgap) for photovoltaics; NaGaTe 2 (ultra-narrow direct bandgap) for infrared detection; and NaInTe 2 (high mechanical compliance and anisotropy) for flexible and polarization-sensitive optoelectronics. Beyond introducing NaXTe 2 as a tunable platform, this work validates a computational paradigm that integrates data-driven discovery with mechanistic understanding, offering a blueprint for accelerated design of next-generation functional semiconductors.