Phase II year
2023
(last award dollars: 1723938463)
Phase II Amount
$1,500,000
This project will establish the clear feasibility of bonded GaN-GaAs heterostructures to enable wide-bandgap heterojunction bipolar transistors (HBTs) with an unprecedented combination of high power, speed, and linearity at high frequencies (>10 GHz). Over the past decade, GaN-based high electron mobility transistors (HEMTs) have matured in reliability and performance and set the standard in high-power RF applications, with widespread adoption for commercial cellular base stations and as military applications. This performance has been achieved with numerous innovations, including the use of field plates to engineer the electric field profile, and to mitigate the trapping related anomalies in the GaN HEMTs which cause current collapse, dispersion, and current transients. Unfortunately, these field plates also inevitably significantly degrade the bandwidth of the HEMTs. In contrast, GaN HBTs, with their very large Gm and absence of trap coupling to the current conduction offer similar high power to the GaN HEMT, but greatly superior linearity (>20dB improvement in OIP3/Pdc). These advantages of the GaN HBT are further described in the following section. Realization of GaN HBT devices has been limited by the absence any well-behaved, shallow p-dopants for GaN. Bonded GaAs/GaN heterojunctions can overcome this GaN limitation by leveraging the effective, C-doped GaAs base in the well-established, InGaP-GaAs material system used in the cell phone industry. Combining this standard emitter-base structure with an N-GaN collector and sub-collector via wafer bonding should produce a high power HBT IC technology with truly excellent RF and mm-wave performance. The proposed project will be a focused effort to advance the state-of-the-art by systematically investigating several promising approaches to achieve well-behaved pn junctions through p+GaAs/N-GaN bonding. These junctions will become the high-power BC junction of the proposed GaAs/GaN HBT. A bonded GaAs-GaN HBT device can be fabricated using slight modifications to KeysightÂ’s existing, well-established GaAs HBT IC process. Of KeysightÂ’s 99 existing processes required to produce their existing GaAs HBT ICs, only 6 will need to be redeveloped for the proposed GaAs-GaN HBT. This shortens the development process of the proposed HBT substantially and creates a simple path to mass production with minimal capital investment other than the bonding tool required. Once the bonding process is developed, MicroLink and Keysight will leverage existing experience and existing equipment that have been used for HBT devices for the past two decades to ramp up production of these devices quickly. The high efficiency and high linearity power amplifiers will be delivered to system integrators for use in X-Band Radar systems.
