Candidate: Yijing Feng
Date: August 1, 2024
Time: 12:00pm
Location: online (Teams)
Supervisor: Lan Wei and Radhakrishna Ujwal
All are welcome!
Abstract:
SiC and GaN are actively employed in Power Electronics (PE) circuits, due to their superior material properties such as wide bandgap and high melting point. GaN has a high electron mobility due to the 2DEG at AlGaN/GaN interface, making it popular for HV/HF applications. SiC instead has a high melting point and thermal conductivity, making it suitable for HV/HP applications. Besides using materials with superior properties, there has also been innovative device structures proposed to further improve power device performance, such as super-junction, multi-channel (stacked 2DEG) and FinFETs. They could further enhance breakdown voltage (BV) and reduce ON-resistance (RON) compared with one-dimensional structure.
In order to facilitate circuit design using SiC/GaN transistors with innovative device architectures, there is a growing demand of accurate, scalable, robust and standardized compact models. Using an innovative modular approach, this thesis first proposes a compact model for multi-channel AlGaN/GaN Schottky Barrier Diode (SBD) for kilo-volt applications. The modeling approach, detailed model formation as well as model evaluation against an experimental of SBD with five stacked 2DEG channels are explained in detail. Using a similar modular formulation method, the thesis also proposes Waterloo Virtual-source SiC Compact Model (WAVSiC), a comprehensive and user-friendly physics-based compact model for SiC MOSFETs. Accuracy, flexibility and scalability are demonstrated for WAVSiC via benchmarking against measurement of commercial devices. The WAVSiC model is also validated for computational efficiency and robustness, circuit simulator compatibility and ready for Process-Design-Kit (PDK) integration.
In summary, by using a modular approach, the thesis introduces two innovative physics-based compact models for wide-bandgap power electronic devices with accuracy, scalability, and computational efficiency, enabling efficient circuit simulation and PDK development based on these device technologies.