Taking into consideration the inaccurate temperature predictions in traditional thermal models of power devices, we undertook a study on the temperature rise characteristics of heterojunction bipolar transistors (HBTs) with a two-dimensional cross-sectional structure including a sub-collector region. We developed a current-adjusted polynomial electro-thermal coupling model based on investigating floating heat sources. This model was developed using precise simulation data acquired from SILVACO (Santa Clara, CA, USA). Additionally, we utilized COMSOL software (version 5.6) to simulate the temperature distribution within parallel power cells, examining further impacts resulting from thermal coupling. The research findings indicate that the rise in current induces modifications in the local carrier concentration, thereby prompting variations in the local electric field, including changes in the heat source’s peak location and intensity. The device’s peak temperature exhibits a non-linear trend regulated by the current, revealing an error margin of less than 1.5% in the proposed current-corrected model. At higher current levels, the drift of the heat source leads to an increase in the heat dissipation path and reduces the coupling strength between parallel devices. Experiments were performed on 64 GaAs (gallium arsenide) HBT-based power cells using a QFI infrared imaging system. Compared to the traditional temperature calculation model, the proposed model increased the accuracy by 6.84%, allowing for more precise predictions of transistor peak temperatures in high-power applications.
In this study, an accurate temperature prediction model is proposed for GaAs HBT, which considers both the bias voltage and current rather than power consumption only. The increase in temperature is closely related to the heat source property, which leads to a complex interaction between the lattice vibration and the uneven distribution of the electric field and current density. To improve the accuracy and stability of the temperature prediction model, a machine learning method of Extreme Learning Machine (ELM) optimized with an Atomic Search Algorithm (ASO) is introduced. The validity of the model is verified by comparing it with experimental observations by the QFI InfraScope TM temperature mapping system. The predicted temperatures for an 8 × 8 HBT power cell fabricated with 2 μm GaAs technology show good agreement with the measurement results, with a ±2 °C error and a relative error deviation below 3%. This demonstrates the superior performance of the proposed model in accurately predicting the temperature of GaAs HBT.