UCLA

Heterojunction bipolar transistors (HBT) are sought as a building block for implementation of modern broadband electronics, defense applications, and other mm-wave electronic systems demanding in high-speed and high-power performance. The push for greater frequency performance, higher power densities and reliability have pushed research towards wide-bandgap materials such as Gallium Nitride (GaN) given its superior intrinsic material properties, providing for higher breakdown voltage, and enabling higher power performance. However, GaN faces a fundamental limitation with p-type doping, limiting its adoption in RF power electronics. Attempts at a wide bandgap HBT to-date has exclusively relied on epitaxial growth using metal-organic chemical vapor deposition or molecular beam epitaxy (MBE) to fabricate a GaN-based transistor; however, none have yielded a device with both sufficient current gain and transition frequency - both key measures of performance in a bipolar transistor. Alternative materials for the different HBT layers have been considered but are limited by significant lattice mismatch. To bypass p-type GaN limitations and improve the base-collector junction with minimal interface trap density, a nanomembrane interlayer device transfer is proposed as an alternative for fabrication of an HBT. There are two aims in this dissertation: to demonstrate an experimental HBT with a sufficient current gain above 20 and demonstrate using computer-aided modeling that sufficient frequency performance can be achieved – both to demonstrate that a wide-bandgap HBT is both possible and worth further exploration for RF electronics. In Chapter 1, the landscape of research into wide-bandgap bipolar transistors is presented. In Chapter 2, multiple methods of integration for different diode pairs within the HBT are evaluated for probability of success in the overall device, where a GaAs-GaN base-collector diode is demonstrated to have the best performance using nanomembrane layer transfer. Additionally, an MBE-grown AlGaAs-GaAs film stack was transferred and demonstrated an emitter-base structure can be transferred with no degradation in performance. In Chapter 3, fabrication of the HBT is demonstrated using the best methods selected from Chapter 2, where a AlGaAs-GaAs-GaN HBT was demonstrated to have a current gain greater than 70. In Chapter 4, technology computer-aided design simulations were developed to validate the DC results shown in Chapter 2 and 3, and simulated transition frequencies of at least 60 GHz for the experimental structure fabricated in Chapter 3. Finally, in Chapter 5, next steps are outlined for exploration beyond this initial proof-of-concept device. Taken altogether, this dissertation serves to demonstrate a device structure with the potential to vastly exceed existing solutions that can be applied to a vast array of wide-bandgap materials without being limited by its p-type analogues, enabling performance for a wide array of applications in the next generations of electronics.