First-principles investigations of III-nitride bulk and surface properties
- Author(s): Dreyer, Cyrus Eduard
- Advisor(s): Van de Walle, Chris G
- et al.
The III-nitride semiconductors, including AlN, InN, GaN, and BN have been demonstrated as technologically exciting materials for a wide range of device applications. With band gaps that span the visible range, GaN, InN, and InGaN alloys are used for high efficiency light emitting diodes for general lighting, as well as laser diodes for optical storage. The wide gaps, large band offsets, and polarization fields in AlN, GaN, and AlGaN alloys are promising for high-frequency, high-power transistors with applications in power conversion and radio frequency amplifiers.
Despite the plethora of attractive material parameters of the III-nitride materials there are several issues that significantly limit the efficiency of devices and range of possible applications. In this study, we use first-principles electronic structure calculations to explore several of these properties relevant to understanding growth, processing, and device design.
Arguably the most detrimental issue in this material system is the lack of widely available, cost-effective substrates for the growth of films and devices. Heteroepitaxy, as well as the lattice mismatch between the layers of different III-nitride alloys in heterostructures, results in residual stresses in films and devices. Such stress will alter the electronic structure of the materials, so it is necessary for device design to be able to quantify these effects. We explore the influence of strain on the effective mass of carriers in GaN and AlN, a parameter that is tied to the conductivity. In addition, films under tensile strain can crack if the strain energy is sufficient. We explore the propensity for AlN, GaN, and AlGaN to crack on different crystallographic planes.
There has been significant work done to overcome the issue of residual strains in III-nitride films, both through the growth of bulk crystals for substrates, and the growth of structures such as nanowires that avoid many of the thickness and alloy-content limitations of epitaxial layers. Also, selective area growth followed by lateral overgrowth techniques is used to produce high-quality GaN for laser applications. In all of these cases, growth involves multiple growth planes, and the resulting morphologies will be determined by the relative growth rates of the planes present. One of the fundamental parameters that determines the stability and growth rate on a given plane is the absolute surface energies of that plane. Calculations of such surface energies are extremely challenging; we have succeeded in obtaining such values, for the first time, for the most technologically important planes in GaN.
Finally, residual stresses contribute to a piezoelectric polarization moment in the III-nitrides in the most common growth direction, and even in the absence of such strain, the wurtzite crystal structure has a spontaneous polarization moment. Polarization differences between layers of heterostructures allow for strong carrier confinement and the formation of two dimensional electron gas (2DEG) with high densities at AlGaN/GaN interfaces, exploited in high electron mobility transistors (HEMTs). The effect of polarization can also be detrimental, for example causing the quantum confined Stark effect in quantum wells of light emitting diodes (LEDs). In both HEMT and LED devices, accurate values of the magnitudes of the spontaneous and piezoelectric polarization constants are required for a fundamental understanding and design of III-nitride devices. We have discovered that the set of spontaneous polarization coefficients that has been widely used for device design and analysis suffers from inconsistencies in the choice of reference structure, and have derived a new reliable and accurate set of constants to describe the polarization properties of III-nitrides. We discuss the results in the context of experimental observations.