No description available.

<p>Light emitting diode (LED)-based solid state displays (SSD) have attracted growing interest due to their advantages in terms of contrast ratio, brightness, viewing angle, and response time compared to liquid crystal displays. GaN based III-nitride thin film materials are suitable materials for SSD due to their wide and tunable bandgaps. However, the large size and costly manufacturing process of commercially available GaN-based LED chips limit the potential uses of LEDs as the pixels of SSD.</p> <p>In this work, tiny single crystal beta-phase (111) oriented SiC whiskers 2 microns in diameter and 18 microns in length are proposed as the substrates for GaN growth due to their small lattice constant mismatch (3%) with GaN, their conductive nature and their small size for potential use in SSD pixels. Aligned SiC whiskers with (111) planes exposed in an alumina matrix prepared by a precise manipulation and alignment method of SiC whiskers including a series of steps was developed in this work. The alignment degree of whiskers achieved in this work is higher than conventional extrusion methods, and a sintering approach capable of forming an aligned alumina/SiC composite was developed and understood using a self-limiting oxidation reaction mechanism.</p> <p>To take advantage of the potential versatility, scalability and cost effectiveness of sputtering for SSD manufacturing, a reactive sputtering system was built for a detailed investigation of GaN thin film growth nucleation and subsequent growth behavior on SiC. 6H-SiC single crystal substrates were chosen as a reference substrate for SiC whiskers. An XRRC indicates that a high quality single crystalline GaN thin film was successfully grown epitaxially on 6H-SiC by sputtering. Two-dimensional X-ray diffraction and scanning transmission electron microscopy results demonstrated that the epitaxial growth of GaN thin films relies on the short range order and/or crystalline area of the native oxide layer in GaN/SiC interface for the first time.</p> / Doctor of Philosophy (PhD)

Gallium Nitride (GaN) technology is capable of very high power levels but suffers from high non-linearity. With the advent of 5G technologies, high linearity is in greater demand due to complex modulation schemes and crowded RF (Radio Frequency) spectrum. Because of the non-linearity issue, GaN power amplifiers have to be operated at back-off input power levels. Operating at back-off reduces the efficiency of the power amplifier along-with the output power. This research presents a technique to linearize GaN amplifiers. The linearity can be improved by splitting a large device into multiple smaller devices and biasing them individually. This leads to the cancellation of the IMD3 (Third-order Intermodulation Distortion) components at the output of the FETs and hence higher linearity performance. This technique has been demonstrated in Silicon technology but has not been previously implemented in GaN. This research work presents for the first time the implementation of this technique in GaN Technology. By the application of this technique, improvement in IMD3 of 4 dBc has been shown for a 0.8-1.0 GHz PA (Power Amplifier), and 9.5 dBm in OIP3 (Third-order Intercept Point) for an S-Band GaN LNA, with linearity FOM (IP3/DC power) reaching up to 20. Large-signal simulation and analysis have been done to demonstrate linearity improvement for two parallel and four parallel FETs. A simulation methodology has been discussed in detail using commercial CAD software. A power sampler element is used to compute the IMD3 currents coming out of various FETs due to various bias currents. Simulation results show by biasing one device in Class AB and others in deep Class AB, IMD3 components of parallel FETs can be made out of phase of each other, leading to cancellation and improvement in linearity. Improvement up to 20 dBc in IMD3 has been reported through large-signal simulation when four parallel FETs with optimum bias were used. This technique has also been demonstrated in simulation for an X-Band MMIC PA from 8-10 GHz in GaN technology. Improvements up to 25-30 dBc were shown using the technique of biasing one device with Class AB and other with deep class AB/class B. The proposed amplifier achieves broadband linearization over the entire frequency compared to state-of-the-art PA's. The linearization technique demonstrated is simple, straight forward, and low cost to implement. No additional circuitry is needed. This technique finds its application in high dynamic range RF amplifier circuits for communications and sensing applications. / Doctor of Philosophy / Power amplifiers (PAs) and Low Noise Amplifiers (LNAs) form the front end of the Radio Frequency (RF) transceiver systems. With the advent of complex modulation schemes, it is becoming imperative to improve their linearity. Through this dissertation, we propose a technique for improving the linearity of amplifier circuits used for communication systems. Meanwhile, Gallium Nitride (GaN) is becoming a technology of choice for high-power amplifier circuits due to its higher power handling capability and higher breakdown voltage compared with Gallium Arsenide (GaAs), Silicon Germanium (SiGe) and Complementary Metal-Oxide-Semiconductor (CMOS) technologies. A circuit design technique of using multiple parallel GaN FETs is presented. In this technique, the multiple parallel FETs have independently controllable gate voltages. Compared to a large single FET, using multiple FETs and biasing them individually helps to improve the linearity through the cancellation of nonlinear distortion components. Experimental results show the highest linearity improvement compared with the other state-of-the-art linearization schemes. The technique demonstrated is the first time implementation in GaN technology. The technique is a simple and cost-effective solution for improving the linearity of the amplifier circuits. Applications include base station amplifiers, mobile handsets, radars, satellite communication, etc.

As the oil industry continues to drill deeper to reach previously untapped wells, the operating environments for electronic systems become harsher, especially due to high temperatures. It is essential to design electronic circuits and systems to be able to withstand these extreme temperatures. The proposed power amplifier (PA) has been designed for a downhole communication system operating at an ambient temperature of 230oC. GaN technology was chosen primarily due to its ability to function at a high junction temperature. The proposed PA was designed with Qorvo's T2G6003028-FL HEMT as it operates reliably at a high junction temperature (T_J) and also the package offers low junction to case thermal resistance . The proposed PA can operate reliably up to an ambient temperature of 230oC using passive cooling opposed to active cooling. At 230 C it operates in class A with a peak PAE of 25.03%, maximum output power of 1.66 W, peak gain of 24.5 dB, center frequency of 255 MHz with 1dB ripple in the passband over a 60 MHz bandwidth, 1dB output compression of approximately 32 dBm, and OIP3 of 37.9dBm. CW measurements were taken for all parameters. / Master of Science

Yes / Much attention has been paid to making 5G developments more en-ergy efficient, especially in view of the need for using high data rates with more complex modulation schemes within a limited bandwidth. The concept of the Doherty power amplifier for improving amplifier efficiency is explained in addi-tion to a case study of a 70W asymmetrical Doherty power Amplifier using two GaN HEMTs transistors with peak power ratings of 45W and 25W. The rationale for this choice of power ratio is discussed. The designed circuit works in the 3.4GHz frequency band with 200 MHz bandwidth. Rogers RO4350B substrate with dielectric constant εr=4.66 and thickness 0.035 mm is used. The perfor-mance analysis of the Doherty power amplifier is simulated using AWR MWO software. The simulated results showed that 54-64% drain efficiency has been achieved at 8 dB back-off within the specified bandwidth with an average gain of 10.7 dB.

As the oil industry continues to drill deeper to reach new wells, electronics are being required to operate at extreme pressures and temperatures. Coupled with substantial real-time data targets, the need for robust high speed electronics is quickly on the rise. This paper presents a high temperature wideband low noise amplifier (LNA) with zero temperature coefficient maximum available gain (ZTCMAG) biasing for a downhole communication system. The proposed LNA is designed and prototyped using 0.25μm GaN on SiC RF transistor technology, which is chosen due to the high junction temperature capability. Measurements show that the proposed LNA can operate reliably up to an ambient temperature of 230°C with a minimum noise figure (NF) of 2.0 dB, gain of 16.1 dB, and P1dB of 19.1 dBm from 230.5MHz — 285.5MHz. The maximum variation with temperature from 25°C to 230°C is 1.53dB for NF and 0.65dB for gain. / Master of Science

Recent emerging gallium nitride (GaN) high electron mobility transistor (HEMT) is expected to be a promising candidate for high frequency power conversion techniques. Due to the advantages of the material, the GaN HEMT has a better figure of merit (FOM) compared to the state-of-the-art silicon (Si) power metal oxide silicon field effect transistor (MOSFET), which allows the GaN HEMT to switch with faster transition and lower switching loss. By applying the GaN HEMT in a circuit design, it is possible to achieve high frequency, high efficiency, and high density power conversion at the same time. To characterize the switching performance of the GaN HEMT, an accurate behavior-level simulation model is developed in this thesis. The packaging related parasitic inductance, including both self-inductance and mutual-inductance, are extracted based on finite element analysis (FEA) methods. Then the accuracy of the simulation model is verified by a double-pulse tester, and the simulation results match well with experiment in terms of both device switching waveform and switching energy. Based on the simulation model, detailed loss breakdown and loss mechanism analysis are made. The cascode GaN HEMT has high turn-on loss due to the body diode reverse recovery of the low voltage Si MOSFET and the common source inductance (CSI) of the package; while the turn-off loss is extremely small attributing to the cascode structure. With this unique feature, the critical conduction mode (CRM) soft switching technique are applied to reduce the dominant turn on loss and increase converter efficiency significantly. The switching frequency is successfully pushed to 5MHz while maintaining high efficiency and good thermal performance. Traditional packaging method is becoming a bottle neck to fully utilize the advantages of GaN HEMT. So an investigation of the package influence on the cascode GaN HEMT is also conducted. Several critical parasitic inductors are identified, which cause high turn on loss and high parasitic ringing which may lead to device failure. To solve the issue, the stack-die package is proposed to eliminate all critical parasitic inductors, and as a result, reducing turn on loss by half and avoiding potential failure mode of the cascode GaN device effectively. Utilizing the proposed stack-die package and ZVS soft switching, the GaN HEMT high frequency, high efficiency, and high density power conversion capability can be further extended to a higher level. / Master of Science

Recent advancements in generative models have resulted in the improvement of hyper- realistic synthetic images or "deepfakes" at high resolutions, making them almost indistin- guishable from real images from cameras. While exciting, this technology introduces room for abuse. Deepfakes have already been misused to produce pornography, political propaganda, and misinformation. The ability to produce fully synthetic content that can cause such mis- information demands for robust deepfake detection frameworks. Most deepfake detection methods are trained in a supervised manner, and fail to generalize to deepfakes produced by newer and superior generative models. More importantly, such detection methods are usually focused on detecting deepfakes having a specific type of content, e.g., face deepfakes. How- ever, other types of deepfakes are starting to emerge, e.g., deepfakes of biomedical images, satellite imagery, people, and objects shown in different settings. Taking these challenges into account, we propose NoiseLearner, an unsupervised and content-agnostic deepfake im- age detection method. NoiseLearner aims to detect any deepfake image regardless of the generative model of origin or the content of the image. We perform a comprehensive evalu- ation by testing on multiple deepfake datasets composed of different generative models and different content groups, such as faces, satellite images, landscapes, and animals. Further- more, we include more recent state-of-the-art generative models in our evaluation, such as StyleGAN3 and probabilistic denoising diffusion models (DDPM). We observe that Noise- Learner performs well on multiple datasets, achieving 96% accuracy on both StyleGAN and StyleGAN2 datasets. / Master of Science / Images synthesized by artificial intelligence, commonly known as deepfakes, are starting to become indistinguishable from real images. While these technological advances are exciting with regards to what a computer can do, it is important to understand that such technol- ogy is currently being used with ill intent. Thus, identifying these images is becoming a growing necessity, especially as deepfake technology grows to perfectly mimic the nature of real images. Current deepfake detection approaches fail to detect deepfakes of other content, such as sattelite imagery or X-rays, and cannot generalize to deepfakes synthesized by new artificial intelligence. Taking these concerns into account, we propose NoiseLearner, a deep- fake detection method that can detect any deepfake regardless of the content and artificial intelligence model used to synthesize it. The key idea behind NoiseLearner is that it does not require any deepfakes to train. Instead, NoiseLearner learns the key features of real images and uses them to differentiate between deepfakes and real images – without ever looking at a single deepfake. Even with this strong constraint, NoiseLearner shows promise by detecting deepfakes of diverse contents and models used to generate them. We also explore different ways to improve NoiseLearner.

The wide-bandgap (WBG) devices, like gallium nitride (GaN) and silicon carbide (SiC) devices have proven to be a driving force of the development of the power conversion technology. Thanks to their distinct advantages over silicon (Si) devices including the faster switching speed and lower switching losses, WBG-based power converter can adopt a higher switching frequency and pursue higher power density and higher efficiency. As a trade-off of the advantages, there also exist the high-frequency-oriented challenges in the adoption of the GaN HEMT under research, including narrow safe gate operating area, increased switching overshoot, increased electromagnetic interference (EMI) in the gate loop and the power stages, the lack of the modules of packages for high current application, high gate oscillation under parallel operation. The dissertation is developed to addressed the all the challenges above to fully explore the potential of the GaN HEMTs. Due to the increased EMI emission in the gate loop, a small isolated capacitor in the gate driver power supply is needed to build a high-impedance barrier in the loop to protect the gate driver from interference. A 2 W dual-output gate driver power supply with ultra-low isolation capacitor for 650 V GaN-based half bridge is presented, featuring a PCB-embedded transformer substrate, achieving 85% efficiency, 1.6 pF isolation capacitor with 72 W/in3 power density. The effectiveness of the EMI reduction using the proposed power supply is demonstrated. The design consideration to build a compact 650 V GaN switching cell is presented then to address the challenges in the PCB layout and the thermal management. With the switching cell, a compact 1 kW 400 Vdc three-phase inverter is built and can operate with 500 kHz switching frequency. With the inverter, the high switching frequency effects on the inverter efficiency, volume, EMI emission and filter design are assessed to demonstrate the tradeoff of the adoption of high switching frequency in the motor drive application. In order to reduce the inverter CM EMI emission above 10 MHz, an active gate driver for 650 V GaN HEMT is proposed to control the dv/dt during turn-on and turn-off independently. With the control strategy, the penalty from the switching loss can be reduced. To build a high current power converter, paralleling devices is a normal approach. The dissertation comes up with the switching cell design using paralleled two and four 650 V GaN HEMTs with minimized and symmetric gate and power loop. The commutation between the paralleled HEMTs is analyzed, based on which the effects from the passive components on the gate oscillation are quantified. With the switching cell using paralleled GaN HEMTs, a 10 kW LLC resonant converter with the integrated litz-wire transformer is designed, achieving 97.9 % efficiency and 131 W/in3 power density. The design consideration to build the novel litz-wire transformer operated at 400 kHz switching frequency is also presented. In all, this work focuses on providing effective solutions or guidelines to adopt the 650 V GaN HEMT in the high frequency, high power density, high efficiency power conversion and demonstrates the advance of the GaN HEMTs in the hard-switched and soft-switched power converters. / Ph. D. / Silicon (Si) -based power semiconductor has developed several decades and achieved numerous outstanding performances, contributing a fast development of the power electronics. While the theatrical limit of the silicon semiconductor is almost reached limiting the progress speed to purse the high-efficiency, high-density high-reliability power conversion, the new material, including gallium-nitride (GaN) and silicon-carbide (SiC), based semiconductor, becomes the driven force to retain the development. Compared with Si-based device, GaN and SiC device own a faster switching speed and a lower on-resistance, enabling the adoption of high switching frequency and the possibility to increase the efficiency, power density and dynamic response. The GaN-based semiconductor is explored to be an even promising game changer than SiC device thanks to a higher theoretical ceiling. However, to adopt GaN-based semiconductors and fully utilize its benefits with high switching frequency, there are numerous high-frequency-oriented challenges, including high frequency oscillation at device termination, increased electromagnetic interference (EMI), the lack of the modules of packages for high current application, high frequency oscillation under parallel operation. The dissertation is developed to address the key high-frequency-oriented challenges to adopt GaN-based semiconductors in the power conversion and come up with the novel design strategy and analysis for high-switching-frequency power conversion using GaN devices. To the reduce the increased EMI emission in the gate loop, a novel PCB-embedded transformer structure is proposed to maintain a low isolation capacitor in the gate driver power supply for the GaN phase leg. With the proposed technique, the dual-output gate driver power supply can achieve high efficiency (85%), ultra-low isolation capacitor (1.6 pF) with high power density (72 W/in³ ). To reduce the high frequency oscillation at the GaN device termination, the strategy to layout GaN devices and its gate driver is proposed with corresponding thermal management. A compact structure for three-phase inverter is then presented, operating with a very high switching frequency (500 kHz). Within the inverter, the high switching frequency effects on the inverter performances are assessed to demonstrate the tradeoff and bottle neck to adopt high switching frequency in the motor drive application. In order to reduce the inverter EMI emission at high frequency ( >10 MHz), an active gate driver for GaN device is proposed for the active dv/dt control strategy. To build a high current power converter, the strategy to parallel GaN devices is proposed in the dissertation with the analysis on the commutation between the paralleled GaN devices. A high-frequency high-current litz-wire transformer structure for LLC resonant converter is presented with modeling and optimization. With the technique, a 10 kW LLC resonant converter achieves high efficiency (97.9 %) and high power density (131 W/in³).

In recent years, there is a clear trend toward increasing the demand for electric power in high-power applications. High-power converters are making major impacts on these high-power applications. Recent breakthroughs in Silicon Carbide (SiC) materials and fabrication techniques have led to the development of high-voltage, high-frequency power devices, which are at the heart of high-power converters. SiC metal-oxide semiconductor field-effect transistors (MOSFETs) have advantages over silicon (Si) devices due to their higher breakdown voltage, higher thermal capability, and lower on-state resistance. However, their fast switching frequency and high blocking voltage bring challenges to the gate-drive circuit design. The gate driver of SiC-MOSFETs requires a power supply that provides a high-voltage, high-density design, a low input-output capacitance (CI/O) transformer design, good voltage regulation, as well as good resilience to faults to enable safe and fast operation. In this thesis, a power supply that supplies multiple gate drivers for 10 kV SiC MOSFETs is presented. A transformer design approach with a single turn at the primary side is proposed. A 20 kV insulation is achieved by the primary HV cable insulation across a toroid transformer core. The CI/O is designed less than 2 pF to mitigate the Common-Mode (CM) noise. A circuit topology analysis is performed and the inductor/capacitor/capacitor/inductor (LCCL) – inductor/capacitor (LC) circuit is selected. This circuit allows Zero-Voltage Switching (ZVS) at full operation range. A Resonant-Current-Bus (RCB) is built at the transformer primary side to achieve load-independence. / Master of Science / Wide-bandgap semiconductor devices have attracted widespread attention due to their superior performance compared to their silicon devices counterpart. To utilize its full benefits, this thesis presents a complete design and optimization of a gate-drive power supply that supplies multiple gate drivers for high-voltage, high-speed semiconductor devices. Four objectives, including high density at high voltage, good noise mitigation, fair voltage regulation, resilience to faults have been achieved. During the design procedure, different topology candidates are introduced and compared, after which a resonant topology is selected. The wide-bandgap semiconductor devices are utilized to reduce the size and losses. Hardware assembly is shown and experimental testing results are provided in the end to verify the design.