Determining the Temperature of GaN based power semiconductors based on Temperature Sensitive Parameters (TSEP)

• Applicant: Professor Dr.-Ing. Mark-Matthias Bakran
• Subject Area: Electronic Semiconductors, Components and Circuits, Integrated Systems, Sensor Technology, Theoretical Electrical Engineering ; Electrical Energy Systems, Power Management, Power Electronics, Electrical Machines and Drives
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 545448757

Both for the electrical design and mechanical dimensioning with regard to load cycle endurance, the temperature of a power semiconductor is an extremely important parameter. Methods are required for determining the temperature during electrical operation, where direct conventional measurement methods (e.g., optical methods) are not accessible. In the case of commonly used power semiconductors based on Si or SiC, various methods are known, where the temperature coefficient of the pn-junction is utilized as a reference. However, in GaN-based semiconductors as in the SPP GaNius, there is currently no established method for temperature determination, and the structure of the components does not have a classical pn-junction in the power path. At the same time, GaN components are often housed in packages that do not allow direct optical access. Therefore, there is a need to infer the semiconductor temperature based on temperature-sensitive electrical parameters (TSEP). The project begins with a review of known methods and an assessment of their applicability to GaN components, focusing on common structures such as GaN-enhancement mode, depletion-mode (cascode) GaN-GIT. In the next stage, sensitivities and cross-sensitivities (e.g., with respect to current and voltage) will be determined. Based on this, achievable accuracies will be evaluated. Simultaneously, a theoretical modeling will be carried out for the effects, which will be compared with measurements. The most promising methods will be selected from these preliminary investigations. Using these methods, their application in a pulsed power converter will be examined. This aims to investigate the effects of GaN characteristics such as "dynamic RDS-on" on the measurement methods. In this context, modeling will also be conducted to create a "digital twin" that can parallelly simulate the measurement and allow for comparison. Alternatively, an artificial neural network will be considered for approximating the nonlinear relationships. The project will conclude with the development of a setup that enables online temperature monitoring during a load cycle test. Alternatively to this last workpacke the effects under parallel condition of several GaN chips will be analyzed.

• DFG Programme: Priority Programmes
• Subproject of: SPP 2312: Energy Efficient Power Electronics "GaNius"
• Major Instrumentation: Luftbasiertes Temperiergerät für Leistungshalbleiter
• Instrumentation Group: 8430 Lufterhitzer, Luftheizgeräte (transportabel), Infrarotheizgeräte und -anlagen)