High performance AlGaN/GaN based high electron mobility transistors in Si
AlGaN/GaN high electron mobility transistors (HEMTs) have attracted a lot of attention for the development of high power and high frequency transistors that find a range of applications in base station amplifiers, radars and power conditioning. Spontaneous and piezoelectric polarization present in nitride devices grown in the c-crystallographic direction induces high density of 2-dimensional electron gas (2DEG). Furthermore, high carrier saturation velocities and high breakdown voltage of the III-nitride material system allows development of the state of the art transistors.
In our lab, we have adopted a holistic approach to design high performance AlGaN/GaN HEMTs on Si substrate by combining density functional theory calculation to understand behavior of Ga, Al, and N adatom diffusion on GaN and AlN surfaces, energy band engineering and device simulation, development and employment of novel growth techniques (adatom migration enhanced MOCVD), optimization of the fabrication process and the design of dedicated masks for integrating various device design techniques.
Cs-free polarization-engineered III-Nitride photocathode
III-Nitride based photocathodes have been the subject of much research in photoemissive devices for ultraviolet detection in astronomy, planetary or defense applications. In order to achieve high quantum efficiency (QE), negative electron affinity (NEA) is necessary. NEA is conventionally reached via surface cesiation which requires special in-situ fabrication steps, including cleaning and activation in vacuum and sealed-tube packaging, due Cesium’s high chemical activity. Photocathodes using this technology have been reported to suffer from chemical instability and QE degradation over time.
Recent work has been performed to eliminate Cs-based surface treatments and improve device efficiency by taking advantage of the polarization exhibited by III-Nitrides in order to achieve NEA. Previously, we developed and reported a novel Cs-free GaN photocathode based on Si-delta-doping.
We have proposed and investigated novel Cs-free III-Nitride photocathodes that show permanent NEA without the use of delta-doping and thus show potential for higher QE. By replacing Cs-based surface treatments with polarization engineering to affect surface properties, we allow the device to be air stable. Additionally, polarization engineering is used as an alternative to impurity-based doping in order to allow high free carrier concentrations without a decrease in mobility. Devices have been studied via both simulation and photoemission measurements of devices grown via MOCVD.
AlGaN based solar blind avalance photodetector
UV photodetectors have applications in missile plume detection, UV astronomy, astrophysics and non-line of sight communication systems. Currently, microchannel plate (MCP) sealed tubes are used extensively in UV instruments for photomultiplication purposes because they can be used in photon counting mode, are solar blind and don’t require cooling. However, this technology is bulky, fragile and expensive, which restricts its use to a very specific set of applications.
Solid state UV detectors offer considerable benefits compared to MCP technology. The wide bandgap of AlGaN-based detectors can be tuned such that they are intrinsically solar blind, allowing for operation of AlGaN APDs in broad daylight without significant background radiation. AlGaN APDs also promise high efficiency, high gain and low noise characteristics which can be exploited for photon counting. Furthermore, these devices can operate at higher temperatures and are radiation hard, making AlGaN the optimal choice for UV photodetectors.
There have been significant challenges in realizing reliable, solar-blind (sub-280nm) devices based on AlGaN. A large lattice mistmatch caused by heteroepitaxy results in a high defect density of the film, and poor film morphology can lead to premature and permanent breakdown. There are a number of difficulties that result from doping AlGaN in high concentrations, and Mg doping for p-type AlGaN has proven particularly difficult.
This project explores innovative methods to realize solar-blind AlGaN-based APDs, using growth, doping and processing methods developed in-house to limit premature breakdown and achieve high sensitive, high efficiency solar-blind devices. Techniques such as pulsed AlGaN, delta-doping, and novel heterostructures (SAM and nano-SAM) are being implemented in these efforts, and processes such as implantation isolation are being developed to improve performance of AlGaN APDs.
Fundamental studies into stress evolution in III-Nitrides
The heteroepitaxial nature of AlxGa1-xN films, particularly grown on Si substrates, brings about challenges like defect formation and cracking for thick films due to a large amount of stress in the films caused by lattice mismatch and thermal expansion mismatch. To reduce these effects, a number of techniques such as migration enhanced epitaxy, growth of superlattices, selective area epitaxy and lateral epitaxial overgrowth (LEO) have been adopted. To compensate for this tensile stress, intrinsic compressive stress can be built into the layers during growth that has proven to be an effective method but often involves detailed knowledge of layer stacks such as in superlattices, and interlayers. In one approach, step graded AlGaN films are grown in succession, building compressive stress in each layer. To achieve the necessary stress compensation optimization of the growth parameters, number of AlGaN layers, layer thicknesses and compositions are necessary. Such commonly used approach requires a great amount of experimentation. This is because the process of building compressive stress and stress relaxation in low mobility III-Nitride is not yet fully understood but is thought to be driven by a process known as dislocation inclination.
In order to get an depth understanding of the stress evolution and this dislocation inclination process, it is required to abserve the stress in the films during the deposition. Using in-situ stress evolution measurement, the curvature of the film is measured and the stress is calculated using Stoney equation. When growth AlGaN films on an AlN buffer, curvature drops initially with the deposition of each AlGaN layer, which indicates that the film is under compressive stress and over time its rate of decrease slows down, and change into tensile stress at higher thickness. The in-situ stress evolution observation of AlGaN film growth confirms that stress relaxation is not linear and complex in nature. Work is being conducted to further understand the stress evolution process and the factors that influence it. The goal is to combine experimental data with ex-situ characterization and theoretical models to develop a predictive model for stress evolution.
Growth of III-Nitrides is typically performed along the vertical c-axis of the wurtzite crystallographic structure, which results in the material exhibiting a spontaneous and piezoelectric polarization field along that axis. This polarization field is detrimental to the performance of LEDs, for which this material system plays a critical role in home and commercial lighting. Much work over the last decade has optimized growth along alternate directions to mitigate the effect of this field on devices, however many of these techniques result in lower material quality which diminishes the benefit of this growth orientation. Another way to access these facets of the crystal is to grow the material laterally from pillars formed in conventionally grown material. These structures yield a higher surface area of the non-polar facet than a planar material would in the same footprint, while laterally-grown III-Nitrides are known to have a lower density of defects within the material. Very recently, this work has been translated to silicon, whose benefits include lower cost and compatibility with current manufacturing.
Our work in nanostructures furthers these current efforts on silicon using our pulsed MOCVD technique, which allows us to achieve precise control of the evolving planes of the nanostructures. In our lab, we have developed the proper processes to yield uniform arrays of nanostructures with nonpolar planes suitable for the fabrication of LEDs and other types of devices. Utilizing these, quantum well structures on these facets have been grown yielding a 398nm violet photoluminescence using In0.08Ga0.92N.