Machine Learning-Optimized Electromagnetic Wave Absorption in Metal/C Nanocomposites

Traditional carbon-based electromagnetic wave absorbers suffer from limited tunability due to the intricate coupling of multiple synthesis parameters, hindering the rational design of high-performance materials. Herein, we apply a genetic algorithm (GA) to optimize electromagnetic wave absorption (EWA) performance in Metal/C Nanocomposites. Over three generations of GA evolution, five key synthesis parameters-carbon precursor type, metal type, molar ratio of carbon precursor to metal ions, carbonization temperature, and filler loading ratio (wt.%)-are simultaneously tuned. Progressive optimization enhances the Enhanced Absorption Band (EAB) from an initial average of 1.24 to 4.08 GHz, while the minimal reflection loss (RL_min) improves from -20.29 to -41.9 dB. The champion sample achieves a remarkable RL_min of -25.9 dB at 7.04 GHz with an EAB of 7.56 GHz. Random Forest and XGBoost models further quantify parameter importance, consistently identifying carbon precursor type (32.5% and 31.4%) and filler loading ratio (33% and 38.4%) as the dominant factors-validating the GA-driven optimization pathway. This work demonstrates the potential of evolutionary algorithms in materials design and provides a transferable framework for high-performance EWA materials.