Apr 7, 2025
2:00pm - 2:15pm
Summit, Level 4, Room 435
Jonah Shoemaker1,Robert Kaplar2,Stephen Goodnick1,Jack Flicker2,Andrew Binder2,Mihai Negoita2,Srabanti Chowdhury3
Arizona State University1,Sandia National Laboratories2,Stanford University3
Jonah Shoemaker1,Robert Kaplar2,Stephen Goodnick1,Jack Flicker2,Andrew Binder2,Mihai Negoita2,Srabanti Chowdhury3
Arizona State University1,Sandia National Laboratories2,Stanford University3
Ultra-Wide Bandgap (UWBG) semiconductors constitute the fourth generation of semiconductor technology beyond WBG materials such as SiC and GaN, and as such are at the forefront of semiconductor research today. Although the Baliga unipolar figure of merit, is useful for comparing between different semiconductor materials in power device applications, the incorporation of UWBG devices into power systems is determined by the system-level benefits derived at given operational parameters, including switching frequency, blocking voltage, current-carrying ability, and thermal management.
To compare the performance of UWBG and WBG materials, we have improved the optimization tool first reported in [1]. This optimizer takes various operational parameters and optimizes the device drift region width and doping concentration to minimize the total power dissipation from static and dynamic losses. This tool has been extended in the present talk to incorporate self-heating effects. Temperature-dependent transport parameters such as carrier mobilities, incomplete ionization due to high dopant activation energies, doping- and drift-dependent critical fields, and space-charge limited conduction are incorporated to model UWBG-specific effects more accurately.
For each material, a mobility model is used to determine carrier mobilities for a given doping level, operating temperature, electric field, and critical field. For each set of operational parameters, the tool minimizes power dissipation by altering the device drift region thicknesses and dopant concentrations subject to the constraints described in [1]. After the power dissipation is computed, a thermal resistance model is used to compute the temperature increase from thermal conduction through the active layer and substrate and thermal convection from the heat sink using liquid or gas cooling. The temperature increase is used to obtain updated mobilities and carrier densities, and the power dissipation is computed at the increased temperature. By iterating across all operating parameters, it is possible to construct a color-map in system operating space and determine which material and design combinations produce the lower power loss at any operational point.
The resulting color-maps show a strong dependence on many of the material parameters used as input. These parameters in turn are highly influenced by the operating temperature and therefore are strongly dependent on the thermal conductivities of the materials and the choices of the heat transfer coefficient and ratio of chip area to heat sink area. We compare color-maps across a range of current densities, with and without self-heating, to show the influence of UWBG-specific effects. We show that diamond’s deep dopant activation energies are mitigated by self-heating, which leads to near-complete dopant ionizations and lowers diamond’s on-state resistance while minimizing its space-charge conduction regime under low temperature increases. These results obtained from self-heating considerations can give researchers educated insight into which materials show the greatest promise at any given point in operating space, including extreme operating conditions such as ultra-high forward current densities, switching frequencies, or blocking voltages which typically produce high quantities of heat within the device.
1. Flicker, J. and R. Kaplar,
Design Optimization of GaN Vertical Power Diodes and Comparison to Si and SiC, in
2017 IEEE 5th Workshop on Wide Bandgap Power Devices and Applications (WiPDA). 2017, IEEE. p. 31-38.