Challenges\nReconciling Theory and Experiments in the\nPrediction of Lattice Thermal Conductivity: The Case of Cu-Based Sulvanites

The exploration of large chemical spaces in search of\nnew thermoelectric\nmaterials requires the integration of experiments, theory, simulations,\nand data science. The development of high-throughput strategies that\ncombine DFT calculations with machine learning has emerged as a powerful\napproach to discovering new materials. However, experimental validation\nis crucial to confirm the accuracy of these workflows. This validation\nbecomes especially important in understanding the transport properties\nthat govern the thermoelectric performance of materials since they\nare highly influenced by synthetic, processing, and operating conditions.\nIn this work, we explore the thermal conductivity of Cu-based sulvanites\nby using a combination of theoretical and experimental methods. Previous\ndiscrepancies and significant variations in reported data for Cu<sub>3</sub>VS<sub>4</sub> and Cu<sub>3</sub>VSe<sub>4</sub> are explained\nusing the Boltzmann Transport Equation for phonons and by synthesizing\nwell-characterized defect-free samples. The use of machine learning\napproaches for extracting high-order force constants opens doors to\ncharting the lattice thermal conductivity across the entire Cu-based\nsulvanite familyfinding not only materials with κ<sub><i>l</i></sub> values below 2 W m<sup>–1</sup> K<sup>–1</sup> at moderate temperatures but also rationalizing their\nthermal transport properties based on chemical composition.