Various factors affect the interfacial thermal resistance (ITR) between two materials, making ITR prediction a high-dimensional mathematical problem. Machine learning is a cost-effective method to address this. Here, we report ITR predictive models based on experimental data. The physical, chemical, and material properties of ITR are categorized into three sets of descriptors, and three algorithms are used for the models. Those descriptors assist the models in reducing the mismatch between predicted and experimental values and reaching high predictive performance of 96%. Over 80,000 material systems composed of 293 materials were inputs for predictions. Among the top-100 high-ITR predictions by the three different algorithms, 25 material systems are repeatedly predicted by at least two algorithms. One of the 25 material systems, Bi/Si achieved the ultra-low thermal conductivity in our previous work. We believe that the predicted high-ITR material systems are potential candidates for thermoelectric applications. This study proposed a strategy for material exploration for thermal management by means of machine learning.

Wide-bandgap gallium nitride (GaN)-based semiconductors offer significant advantages over traditional Si-based semiconductors in terms of high-power and high-frequency operations. As it has superior properties, such as high operating temperatures, high-frequency operation, high breakdown electric field, and enhanced radiation resistance, GaN is applied in various fields, such as power electronic devices, renewable energy systems, light-emitting diodes, and radio frequency (RF) electronic devices. For example, GaN-based high-electron-mobility transistors (HEMTs) are used widely in various applications, such as 5G cellular networks, satellite communication, and radar systems. When a current flows through the transistor channels during operation, the self-heating effect (SHE) deriving from joule heat generation causes a significant increase in the temperature. Increases in the channel temperature reduce the carrier mobility and cause a shift in the threshold voltage, resulting in significant performance degradation. Moreover, temperature increases cause substantial lifetime reductions. Accordingly, GaN-based HEMTs are operated at a low power, although they have demonstrated high RF output power potential. The SHE is expected to be even more important in future advanced technology designs, such as gate-all-around field-effect transistor (GAAFET) and three-dimensional (3D) IC architectures. Materials with high thermal conductivities, such as silicon carbide (SiC) and diamond, are good candidates as substrates for heat dissipation in GaN-based semiconductors. However, the thermal boundary resistance (TBR) of the GaN/substrate interface is a bottleneck for heat dissipation. This bottleneck should be reduced optimally to enable full employment of the high thermal conductivity of the substrates. Here, we comprehensively review the experimental and simulation studies that report TBRs in GaN-on-SiC and GaN-on-diamond devices. The effects of the growth methods, growth conditions, integration methods, and interlayer structures on the TBR are summarized. This study provides guidelines for decreasing the TBR for thermal management in the design and implementation of GaN-based semiconductor devices.