Data-Driven Description of the Lattice Thermal Conductivity of Two-Dimensional Materials

Two-dimensional (2D) materials hold great promise for advanced thermal management due to their unique phonon transport properties, but 2D semiconductors with a lattice thermal conductivity (κ_L) of more than 10 W/mK remain scarce. Using high-throughput computation and first-principles calculations, we identify 18 2D materials with room-temperature κ_Lvalues exceeding 20 W/mK. Our analysis reveals a low mean atomic mass, a high Young’s modulus, and small surface corrugation as critical descriptors for enhanced κ_Lvalues in 2D materials. We further developed a machine learning-assisted model predicting a series of new 2D materials with κ_Lvalues exceeding 300 W/mK. Notably, a C₂N₂monolayer is predicted to exhibit a high room-temperature κ_Lof 1300 W/mK and a wide bandgap of 5.19 eV, while a B₄C₄monolayer achieves a balanced κ_Lof 574 W/mK and a bandgap of 0.98 eV. These findings offer robust guidance for evaluating and designing the κ_Lof 2D materials for effective thermal management in nanodevices.