A. E. Wickenden

Current collapse and the role of carbon in AlGaN/GaN high electron mobility transistors grown by metalorganic vapor-phase epitaxy

The two deep traps responsible for current collapse in AlGaN/GaN high electron mobility transistors grown by metalorganic vapor-phase epitaxy have been studied by photoionization spectroscopy. Varying the growth pressure of the high resistivity GaN buffer layer results in a change in the deep trap incorporation that is reflected in the observed current collapse. Variations in the measured trap concentrations with growth pressure and carbon incorporation indicate that the deepest trap is a carbon-related defect, while the mid-gap trap may be associated with grain boundaries or dislocations.

Observation of Deep Traps Responsible for Current Collapse in GaN Metal-Semiconductor Field-Effect Transistors

Deep traps responsible for current collapse phenomena in GaN metal–semiconductor field-effect transistors have been detected using a spectroscopic technique that employs the optical reversibility of current collapse to determine the photoionization spectra of the traps involved. In the n-channel device investigated, the two electron traps observed were found to be very deep and strongly coupled to the lattice. Photoionization thresholds for these traps were determined at 1.8 and at 2.85 eV. Both also appear to be the same traps recently associated with persistent photoconductivity effects in GaN.

Influence of MOVPE growth conditions on carbon and silicon concentrations in GaN

Abstract Impurity incorporation is studied as a function of metalorganic vapor phase epitaxy growth conditions. The same GaN growth conditions were used initially, resulting in films with approximately the same dislocation density, after which a single growth parameter was varied and the impurity concentrations measured using SIMS. The C concentrations were found to decrease with increasing growth temperature, pressure, and ammonia flow, and to increase with increasing H 2 carrier and trimethylgallium flow. The Si concentrations for both unintentionally doped (UID) and intentionally doped (ID) films increased with increasing growth pressure. The UID and ID Si concentrations varied inversely with the GaN growth rate, suggesting an independent source for UID Si within the reactor. Moreover, the NH 3 flow rate influenced the Si-doping concentration, even though the GaN growth rate remained constant. A H 2 /NH 3 etching mechanism is proposed to explain the growth parameter influence on the observed C and Si concentrations. The reduction in the ID Si concentrations at high NH 3 flows is explained by NH 3 site blocking, similar to that proposed for increased Ga vacancies at high NH 3 flows.

GROWTH MODEL FOR GAN WITH COMPARISON TO STRUCTURAL, OPTICAL, AND ELECTRICAL PROPERTIES

A kinetic model is presented to explain the metal organic vapor phase epitaxy (MOVPE) growth of GaN. The model is based upon measured desorption rates and assumptions on the precursor dissociation and sticking probabilities. The model shows how the growth temperature and V/III ratio are linked for the growth of high quality GaN films. From a comparison of growth conditions cited in the literature to the quality of GaN produced, optimal film growth appears to occur when the V/III ratio is chosen to be slightly larger than the N to Ga desorption ratio. The relationship between the growth temperature, V/III ratio, and GaN quality are explained in terms of how the growth parameters influence the incorporation of Ga and N atoms into the growing film. The Ga and N diffusion lengths are estimated to be 2–20 nm and <1 nm at 1050 °C, respectively, for practical MOVPE growth rates. Growth conditions for smooth (0001) surface morphology are described in terms of the growth model, as well as possible origins for defe...

The growth and properties of Al and AlN films on GaN(0001)–(1×1)

The growth, structure, and annealing behavior of Al films, formed by in situ vapor deposition on GaN(0001)–(1×1) near 25 °C, have been studied using Auger, electron energy loss, x ray and ultraviolet photoemission spectroscopies and low‐energy electron diffraction. Film growth occurs by a Stranski–Krastanov process with reaction at the immediate interface leading to metallic Ga. Annealing at ≳800 °C leads to release of N, which reacts with Al to form a (1×1)‐ordered layer of AlN, possibly alloyed with a small amount of Ga. The AlN layer has been characterized using the various spectroscopies, and the work function, band bending, and electron affinity of GaN and of the AlN overlayer have been obtained. The Al/GaN Schottky barrier height has been measured and compared with previous results for Ni/GaN.

Persistent photoconductivity in n-type GaN

Persistent photoconductivity has been observed in n-type GaN:Si. The effect is seen at room temperature in both nonoptimally grown films as well as in device quality channel layers. The relaxation dynamics are found to agree with a stretched exponential model of recovery. A comparison between different samples, based upon stretched exponential parameters, Hall measurements, and photoluminescence data is made. The data suggest that the cause of persistent photoconductivity is the same among the different samples and that there is a transition in the relaxation dynamics between room temperature and 130 °C.

Photoionization spectroscopy of traps in GaN metal-semiconductor field-effect transistors

Measurements of the spectral and intensity dependences of the optically-induced reversal of current collapse in a GaN metal-semiconductor field-effect transistor (MESFET) have been compared to calculated results. The model assumes a net transfer of charge from the conducting channel to trapping states in the high-resistivity region of the device. The reversal, a light-induced increase in the trap-limited drain current, results from the photoionization of trapped carriers and their return to the channel under the influence of the built-in electric field associated with the trapped charge distribution. For a MESFET in which two distinct trapping centers have been spectrally resolved, the experimentally measured dependence upon light intensity was fitted using this model. The two traps were found to have very different photoionization cross-sections but comparable concentrations (4×1011 cm−2 and 6×1011 cm−2), suggesting that both traps contribute comparably to the observed current collapse.

Enhanced GaN decomposition in H2 near atmospheric pressures

GaN decomposition is studied at metallorganic vapor phase epitaxy pressures (i.e., 10–700 Torr) in flowing H2. For temperatures ranging from 850 to 1050 °C, the GaN decomposition rate is accelerated when the H2 pressure is increased above 100 Torr. The Ga desorption rate is found to be independent of pressure, and therefore, does not account for the enhanced GaN decomposition rate. Instead, the excess Ga from the decomposed GaN forms droplets on the surface which, for identical annealing conditions, increase in size as the pressure is increased. Possible connections between the enhanced GaN decomposition rate, the coarsening of the nucleation layer during the ramp to high temperature, and increased GaN grain size at high temperature are discussed.

Current collapse induced in AlGaN/GaN high-electron-mobility transistors by bias stress

Current collapse is observed to be induced in AlGaN/GaN high-electron-mobility transistors as a result of short-term bias stress. This effect was seen in devices grown by both metalorganic chemical vapor deposition (MOCVD) and molecular-beam epitaxy (MBE). The induced collapse appears to be permanent and can be reversed by SiN passivation. The traps responsible for the collapse have been studied by photoionization spectroscopy. For the MOCVD-grown devices, the same traps cause the collapse in both unstressed and stressed devices. These effects are thought to result from hot-carrier damage during stress.

Doping of gallium nitride using disilane

The silicon doping of n-type GaN using disilane has been demonstrated for films grown on sapphire substrates by low pressure organometallic vapor phase epitaxy. The binding energy of an exciton bound to a neutral Si donor has been determined from low temperature (6K) photoluminescence spectra to be 8.6 meV. Nearly complete activation of the Si impurity atom in the GaN lattice has been observed.