A Transconductance-Based Extraction Method for Thermal Resistance in GaN HEMTs

Thermal resistance plays a crucial role in characterizing the self-heating effects or electrothermal effects in the gallium nitride high-electron-mobility transistors (GaN HEMTs). In this work, a novel extraction method for thermal resistance in GaN HEMTs is proposed. With this method, the maximum value of transconductance extracted at the drain–source voltage, which is significantly higher than the quiescent drain–source voltage, is used to establish the relationship between channel temperature and dissipated power, from which the thermal resistance can be calculated. The proposed method effectively avoids the adverse impact of current collapse (CC) effects on thermal resistance extraction while maintaining a satisfactory level of extraction accuracy. The proposed method is implemented on HEMTs and compared with existing electrical extraction methods. It is found to have satisfactory accuracy and stability at different quiescent biases. A large-signal model, which includes the extracted thermal resistance, is built to reflect and verify the importance of thermal resistance. The model exhibits good accuracy in various verifications, including dc current–voltage (DCIV), ${S}$ -parameters, and large-signal tests. We also address other commonly raised concerns such as self-heating, gate lag, and temperature sensitivity.