Design of III-Nitride Hot Electron Transistors
III-Nitride based devices have made great progress over the past few decades in electronics and photonics applications. As the technology and theoretical understanding of the III-N system matures, the limitations on further development are based on very basic electronic properties of the material, one of which is electron scattering (or ballistic electron effects). This thesis explores the design space of III-N based ballistic electron transistors using novel design, growth and process techniques. The hot electron transistor (HET) is a unipolar vertical device that operates on the principle of injecting electrons over a high-energy barrier (ϕBE) called the emitter into an n-doped region called base and finally collecting the high energy electrons (hot electrons) over another barrier (ϕBC) called the collector barrier. The injected electrons traverse the base in a quasi-ballistic manner. Electrons that get scattered in the base contribute to base current. High gain in the HET is thus achieved by enabling ballistic transport of electrons in the base. In addition, low leakage across the collector barrier (IBCleak) and low base resistance (RB) are needed to achieve high performance. Because of device attributes such as vertical structure, ballistic transport and low-resistance n-type base, the HET has the potential of operating at very high frequencies. Electrical measurements of a HET structure can be used to understand high-energy electron physics and extract information like mean free path in semiconductors.
The III-Nitride material system is particularly suited for HETs as it offers a wide range of ΔEcs and polarization charges which can be engineered to obtain barriers which can inject hot-electrons and have low leakage at room temperature. In addition, polarization charges in the III-N system can be engineered to obtain a high-density and high-mobility 2DEG in the
base, which can be used to reduce base resistance and allow vertical scaling.
With these considerations in mind, III-N HETs had been explored in our research group earlier and gave us encouraging common base IV characteristics. Common emitter transistor operation was, however, not observed due to high RB and IBCleak. This thesis discusses several design and process challenges associated with the HET in general and specific to the III-N system. Many of these challenges like RB, IBCleak, and high energy injection were solved using novel combinations of hetero-structure and polarization engineering, device fabrication, and growth. Common-Emitter operation (with current gain ~ 0.1) was demonstrated in III-N HETs for the first time using injection and collector barriers induced by AlGaN and InGaN polarization-dipoles. In order to improve current gain, different parts of the III-N HET base which contribute to scattering, were identified. A novel base contact methodology using selective etching of GaN with respect to AlN was developed to enable base scaling. Aggressive scaling of all parts of the base was then used to increase current gain. A maximum gain of ~3.5 was demonstrated using a 1.5nm AlN layer as the emitter, 2nm GaN base and 2nm In0.2Ga0.8N as the collector P-D. This is the highest reported DC current gain in III-N HETs to date. The III-N HET structure was also used to extract the mean free path of hot-electrons (λmfp = 6nm) in GaN. The extracted value of mean free path has significant implications for any scaled devices which use ballistic or quasi-ballistic electron transport. We believe that the work presented in this dissertation provides a pathway for high gain in III-N HETs and eventual realization of their high frequency potential.