fig1

Figure 1. Flowchart for screening promising SWIR photodetector materials via an integrated high-throughput screening, ML, and DFT approach. Starting from ternary compounds retrieved from The MP database, this workflow first implements high-throughput screening incorporating seven specific space groups. Through predefined selection criteria, the pool is refined to yield infrared detection material candidates and corresponding initial feature parameters. The resulting dataset is then processed via ML for feature optimization to identify key representative features, followed by ML-driven material prediction to obtain potential infrared detection candidates. Finally, comprehensive DFT calculations - including bandgap computation, structural optimization, formation energy analysis, optical absorption, mechanical properties, effective mass, and DOS calculations - are performed to complete the final screening and identify the most promising SWIR photodetector materials. SWIR: Short-wave infrared; ML: machine learning; DFT: density functional theory; MP: Materials Project; DOS: density of states.