Machine Learning-Guided Conductivity Prediction in 2D Organic Metal Chalcogenides for Accelerated Electromagnetic Wave Absorber Design

The rational design of electromagnetic wave (EMW) absorbers relies on precise conductivity control, yet conventional trial-and-error methods fail to efficiently explore the multidimensional synthesis parameter space of two-dimensional organic metal chalcogenides (2D OMCs). Here, we propose a machine learning framework that deciphers the nonlinear relationships between synthetic parameters and electrical conductivity in 2D OMCs, enabling quantitative predictions for targeted EMW absorber design. The trained model achieves 86% accuracy in three-level conductivity classification (I: <10^–6 S/m, II: 10^–6–10^–2 S/m, III: >10^–2 S/m), significantly outperforming empirical approaches. Notably, it demonstrates robust extrapolation capability by correctly predicting 12 out of 15 novel OMCs beyond the training data set. Guided by conductivity predictions, the EMW absorption performance of specific 2D OMCs can be efficiently ranked, accelerating material design with experimentally validated accuracy. This machine learning-assisted strategy reveals the complex relationship between synthesis parameters and conductivity, expediting the design and synthesis of materials with optimized EMW absorption performance.