Interplay of Device Design and Carbon-Doped GaN Buffer Parameters in Determining Dynamic in AlGaN/GaN HEMTs

• Published in: IEEE transactions on electron devices Vol. 69; no. 11; pp. 6035 - 6042
• Main Authors: Joshi, Vipin, Gupta, Sayak Dutta, Chaudhuri, Rajarshi Roy, Shrivastava, Mayank
• Format: Journal Article
• Language: English
• Published: New York The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 01.11.2022
• Subjects: Aluminum gallium nitrides; Buffers; Carbon; Design parameters; Doping; Electric fields; Gallium nitrides; High electron mobility transistors; Optimization; Passivity; Piezoelectricity; Polarization; Thickness; Trapping
• ISSN: 0018-9383; 1557-9646
• DOI: 10.1109/TED.2022.3209635

Using a well-calibrated computational framework, we reveal a complex interplay between the device design and the epi-stack parameters, which determines the electron trapping in the carbon-doped GaN buffer, leading to dynamic ON resistance ([Formula Omitted]) in AlGaN[Formula Omitted]GaN HEMTs. The parameters being considered here are surface trap concentration, passivation thickness, field plate length, unintentionally doped (UID) GaN channel thickness, strain-induced piezoelectric polarization in the AlGaN layer, buffer traps, and carbon–Si co-doping of the GaN buffer. The role of surface traps in determining the extent of electron injection and trapping in the GaN buffer is revealed. Furthermore, its dependence on piezoelectric polarization in the AlGaN layer and implications on dynamic [Formula Omitted] is discussed. Correlation among the passivation thickness, the field plate length, and the UID channel thickness affecting the channel electric field profile is explored, which, in turn, determines the extent of electron injection into the GaN buffer and eventually the extent of the dynamic [Formula Omitted] degradation. This work also develops detailed physical insights explaining the mechanisms responsible for the disclosed complex interplay. This allowed us to discuss buffer-doping optimization to minimize electron trapping in the GaN buffer and resulting dynamic [Formula Omitted] while maximizing the breakdown voltage of the device. These new findings are expected to provide guidelines to design dynamic [Formula Omitted] resilient HEMTs and also to explain experimental trends associated with dynamic [Formula Omitted] behavior as a function of the device and epi-stack parameters.