Return to search

Parasitics and Current-Dispersion Modeling of AlGaN/GaN HEMTs Fabricated on Different Substrates Using the Equivalent-Circuit Modeling Technique

Electrical equivalent circuit modeling of active components is one of the most important approaches for modeling high-frequency high-power devices. Amongst the most used microwave devices, AlGaN/GaN HEMTs demonstrated their superior performance, making them highly suitable for 5G, wireless and satellite communications. Despite the remarkable performance of AlGaN/GaN HEMTs, these devices reside on substrates that invoke limitations on the operating-frequency, power-efficiency, and current dispersion phenomenon. Also, there is a limitation in present parameters extraction techniques being not able to consider both the substrate effect (Silicon, Silicon Carbide, and Diamond) and the asymmetrical GaN HEMT structure. In this thesis work, a single extrinsic parameters extraction technique using a single small-signal topology takes into account both the asymmetrical GaN HEMT structure and the different substrate types with their parasitic conduction will be developed and studied for the first time. Moreover, large-signal modeling using Quasi-Physical Zone Division technique has been applied to both GaN/D and GaN/SiC to model the isothermal-trapping free drain current, and combined with a new simple technique for comparing performance between active devices in terms of current-dispersion. The models were verified by simulating the small-signal S-parameters, large-signal IV characteristics, and single-tone load-pull. High accuracy was achieved compared to the measurement data available in the technical literature and obtained from fabricated devices.

Identiferoai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/40706
Date06 July 2020
CreatorsAlsabbagh, Mohamad
ContributorsPark, Jeongwon, Yagoub, Mustapha
PublisherUniversité d'Ottawa / University of Ottawa
Source SetsUniversité d’Ottawa
LanguageEnglish
Detected LanguageEnglish
TypeThesis
Formatapplication/pdf

Page generated in 0.0024 seconds