published_or_final_version / abstract / Physics / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / Physics / Doctoral / Doctor of Philosophy

AlGaN/GaN high electron mobility transistors (HEMTs) have been widely used for high power and high frequency RF communications due to their fast switching and large current handling capabilities. The reliability of such devices is strongly affected by the junction temperature where the highest magnitude occurs in a local region on the drain side edge of the gate called the hotspot. Thus, thermal management of these devices remains a major concern in the design and reliability of systems employing AlGaN/GaN HEMTs. Due to the large power densities induced in these devices locally near the drain side edge of the gate, it is clear that moving thermal management solutions closer to the heat generation region is critical in order to reduce the overall junction temperature of the device. In this work, we explore the use of embedded microchannel cooling in the substrate of AlGaN/GaN HEMTs made on Si and SiC substrates and compare them to passive cooling techniques using Si, SiC, and diamond substrates. In addition, the impact of cooling fluids and harsh environmental conditions were considered. The study was performed using a combination of CFD and finite volume analysis on packaged AlGaN/GaN HEMTs. Active cooling using embedded microchannels were shown to have a significant impact on the heat dissipation over the passive cooling methods, approaching or exceeding that of diamond cooled devices. For vertical power devices (IGBT), embedded microchannels in the power electronics substrates were explored. In both the power devices and lateral AlGaN/GaN HEMTs, the use of embedded microchannels with nonlinear channel geometries was shown to be the most effective in terms of reducing the device junction temperature while minimizing the pumping power required.

Silicon based semiconductor devices are rapidly approaching the theoretical limit of operation and are becoming unsuitable for future military requirements. The scope of semiconductor devices has been expanded by wide bandgap devices such as gallium nitride (GaN) to include the possibility for high power and high frequency operation. A new generation of high speed â high frequency devices is required to meet current and future military needs. The Gallium Nitride High Electron Mobility Transistor (HEMT) is showing great promise as the enabling technology in the development of military radar systems, electronic surveillance systems, communications systems and high voltage power systems. Typically, sapphire or silicon carbide is utilized as the substrate material in most HEMT designs. This thesis explores the possibility of utilizing a diamond substrate to increase the power handling capability of the AlGaN/GaN HEMT. Diamond offers increased thermal property parameters that can be simulated in the commercially available Silvaco software package. A complete electrical and thermal analysis of the model was conducted and compared to actual device characteristics. The results of the software simulation and measurements on the test devices indicate diamond substrates will enable the HEMT to be operated at a higher power than traditional sapphire substrate HEMTS.

The objective of this dissertation is to understand the structural and optoelectronic properties of group III-nitride materials grown by High-Pressure Metal Organic Chemical Vapor Deposition (HP-MOCVD) and Migration Enhanced Plasma Assisted MOCVD by FTIR reflectance spectroscopy, Raman spectroscopy, X-ray diffraction, and Atomic Force Microscopy. The influence of the substrates/templates (Sapphire, AlN, Ga-polar GaN, N-polar GaN, n-GaN, and p-GaN) on the free carrier concentration, carrier mobility, short-range crystalline ordering, and surface morphology of the InN layers grown on HP-MOCVD were investigated using those techniques. The lowest carrier concentration of 7.1×1018 cm-3 with mobility of 660 cm2V-1s-1 was found in the InN film on AlN template, by FTIR reflectance spectra analysis. Furthermore, in addition to the bulk layer, an intermediate InN layers with different optoelectronic properties were identified in these samples. The best local crystalline order was observed in the InN/AlN/Sapphire by the Raman E2 high analysis. The smoothest InN surface was observed on the InN film on p-GaN template. The influence of reactor pressures (2.5–18.5 bar) on the long-range crystalline order, in plane structural quality, local crystalline order, free carrier concentration, and carrier mobility of the InN epilayers deposited on GaN/sapphire by HP-MOCVD has also been studied using those methods. Within the studied process parameter space, the best material properties were achieved at a reactor pressure of 12.5 bar and a group-V/III ratio of 2500 with a free carrier concentration of 1.5x1018 cm-3, a mobility in the bulk InN layer of 270 cm2 V-1s-1 and the Raman (E2 high) FWHM of 10.3 cm-1. The crystalline properties, probed by XRD 2θ–ω scans have shown an improvement with the increasing reactor pressure. The effect of an AlN buffer layer on the free carrier concentration, carrier mobility, local crystalline order, and surface morphology of InN layers grown by Migration-Enhanced Plasma Assisted MOCVD were also investigated. Here, the AlN nucleation layer was varied to assess the physical properties of the InN layers. This study was focused on optimization of the AlN nucleation layer (e.g. temporal precursor exposure, nitrogen plasma exposure, and plasma power) and its effect on the InN layer properties.

This dissertation details the study of band bending in n- and p-type GaN samples with a Kelvin probe utilizing different illumination geometries, ambients (air, oxygen, vacuum 10-6 mbar), and sample temperatures (77 – 650 K). The Kelvin probe, which is mounted inside an optical cryostat, is used to measure the surface potential. Illumination of the GaN surface with band-to-band light generates electron-hole pairs, which quickly separate in the depletion region due to a strong electric field caused by the near-surface band bending. The charge that is swept to the surface reduces the band bending and generates a surface photovoltage (SPV). Information about the band bending can be obtained by fitting the SPV measurements with a thermionic model based on the emission of charge carriers from bulk to surface and vice versa. The band bending in freestanding n-type GaN templates has been evaluated. The Ga-polar and N-polar surfaces exhibit upward band bending of about 0.74 and 0.57 eV, respectively. The surface treatment also plays a major role in the SPV behavior, where the SPV for mechanical polished surfaces restores faster than predicted by a thermionic model in dark. When measuring the photoluminescence (PL) signal, the PL from mechanically polished surfaces was about 4 orders of magnitude smaller than the PL from chemically mechanically polished surfaces. The PL and SPV behaviors were explained by the presence of a large density of defects near the surface, which quench PL and aid in the restoration of the SPV via electron hopping between defects. Temperature-dependent SPV studies have also been performed on doped n- and p-type GaN samples. In Si-doped n-type GaN, the estimated upward band bending was about 1 eV at temperatures between 295 and 500 K. However, in p-type GaN, the downward band bending appeared to increase with increasing temperature, where the magnitude of band bending increased from 0.8 eV to 2.1 eV as the temperature increased from 295 to 650 K. It appears that heating the p-type GaN samples allows for band bending values larger than 1 eV to fully restore. Pre-heating of samples was of paramount importance to measure the correct value of band bending in p-type GaN. The slope of the dependence of the SPV on excitation intensity at low temperatures was larger than expected; however, once the temperature exceeded 500 K, the slope began to reach values that are in agreement with a thermionic model.

AlGaN/GaN and InAlN/GaN-based heterojunction field effect transistors (HFETs) have demonstrated great high power and high frequency performance. Although AlGaN/GaN HFETs are commercially available, there still remain issues regarding long-term reliability, particularly degradation and ultimately device failure due to the gate-drain region where the electric field peaks. One of the proposed degradation mechanisms is the inverse-piezoelectric effect that results from the vertical electric field and increases the tensile strain. Other proposed mechanisms include hot-electron-induced trap generation, impurity diffusion, surface oxidation, and hot-electron/phonon effects. To investigate the degradation mechanism and its impact on DC, microwave, and noise performance, comprehensive stress experiments were conducted in both un-passivated and passivated AlGaN/GaN HFETs. It was found that degradation of AlGaN/GaN HFETs under reverse-gate-bias stress is dominated by inverse-piezoelectric effect and/or hot-electron injection due to gate leakage. Degradation under on-state-high-field stress is dominated by hot-electron/phonon effects, especially at high drain bias. Both effects are induced by the high electric field present during stress, where the inverse-piezoelectric effect only relates to the vertical electric field and the hot-electron effect relates to the total electric field. InAlN/GaN-based HFETs are expected to have even better performance as power amplifiers due to the large 2DEG density at the InAlN/GaN interface and better lattice-matching. Electrical stress experiments were therefore conducted on InAlN/GaN HFETs with indium compositions ranging from 15.7% to 20.0%. Devices with indium composition of 18.5% were found to give the best compromise between reliability and device performance. For indium compositions of 15.7% and 17.5%, the HFET devices degraded very fast (25 h) under on-state-high-field stress, while the HFET devices with 20.0% indium composition showed very small drain. It was also demonstrated that hot-electron/phonon effects are the major degradation mechanism for InAlN/GaN HFETs due to a large 2DEG density under on-state operations, whereas the inverse-piezoelectric effect is very small due to the small strain for the near lattice-matched InAlN barrier. Compared to lattice-matched InAlN/GaN HFETs, AlGaN/GaN HFETs have much larger strain in the barrier and about half of the drain current level; however, the hot electron/hot phonon effects are still important, especially at high drain bias.

The Micro-Raman scattering technique has been used for the study of GaN and ZnO. Capabilities of the Raman technique and existing literature on Raman spectroscopy in GaN and ZnO are reviewed. About 50 GaN and ZnO samples with a wide range of properties are studied. From the analysis of positions of the E2H and A1(LO) phonon modes, biaxial stress and plasmon coupling of the Al(LO) mode are observed and compared to a bulk GaN sample. The stress-related shift rate for the AI(LO) mode in hexagonal GaN is established to be 2.7 ± 0.4 cm-1/GPa through series of GaN with low free carrier concentration. Bulk ZnO and ZnO layers grown on sapphire have been studied, and no biaxial stress is found in ZnO layers. Doping and impurity modes resulted in disorder-activated scattering in ZnO. The choice of the laser for study of GaN and ZnO layers on sapphire substrate is discussed.

This thesis uses the techniques of atomic force microscopy (AFM) and conductiveAFM (CAFM) to study defect sites on GaN films. In particular, these defect sites demonstrate current leakage under reverse-bias conditions that are detrimental to device fabrication. Two growth techniques that were used to improve this leakage behavior for samples in this study included: epitaxial lateral overgrowth (ELO) and nano-ELO using a Si3N4 film. Both techniques decrease defects such as threading dislocations by controlling the nucleation and growth behavior of the GaN films. The EL0 technique uses a patterned dielectric film to laterally grow micron-wide regions (referred to as 'wings') that minimize dislocation defects. Our CAFM studies indicate that ELO films have no detectable leakage sites in these wing regions; however, between these regions the films have typical leakage site densities seen for standard films on the order of 107cm-3. The nano-ELO technique utilizes a porous Si3N4 film to reduce defects over the entire film, and CAFM data indicate nearly a factor of ten reduction in leakage site densities. The nano-ELO technique is therefore optimal for an overall improvement in film quality, whereas the ELO technique is suitable for device fabrication in patterned regions with optimized film quality.

Master of Science / Department of Electrical and Computer Engineering / William B. Kuhn / The design of power amplifiers in any semi-conductor process is not a trivial exercise and it is often encountered that the simulated solution is significantly different than the results obtained. Oscillatory phenomena occurring either in-band or out of band and sometimes at subharmonic intervals can render a design useless. Other less apparent effects such as jumps, hysteresis and continuous spectrum, often referred to as chaos, can also invalidate a design. All of these problems might have been identified through a more rigorous approach to stability analysis. Designing for stability is probably the one area of amplifier design that receives the least amount of attention but incurs the most catastrophic of effects if it is not performed properly. Other parameters such as gain, power output, frequency response and even matching may have suitable mitigation paths. But the lack of stability in an amplifier has no mitigating path. In addition to the loss of the design there are the increased production cycle costs, costs involved with investigating and resolving the problem and the costs involved with schedule slips or delays resulting from it. The Linville or Rollett stability criteria that many microwave engineers follow and rely exclusively on is not sufficient by itself to ensure a stable and robust design. It will be shown that the belief that unconditional stability is obtained through an analysis of the scattering matrix S to determine if K>1 and [delta][supscript]s<1 can fail and other tools must be used to validate circuit stability. With the emphasis being placed on stability, a 1W pulse modulated S-band power amplifier is designed using a battery of analysis tools in addition to the standard Linville or Rollett criteria to rigorously confirm the stability of the circuit. Test measurements are then presented to confirm the stability of the design and illustrate the results. The research shown contributes to the state of the art by offering a detailed approach to stability design and then applying the techniques to the design of a 1W pulse modulated S-band power amplifier demonstrating the first with 20 nanosecond pulse width switching and single digit nanosecond rise and fall times at 1 Watt power levels.