Browsing by Author "Dr. Mark Johnson, Committee Member"

Now showing 1 - 7 of 7

Characterization and Modeling of III-N MOS-HFETs for High Frequency Applications
(2006-11-20) Dandu, Krishnanshu; Dr. Doug Barlage, Committee Chair; Dr. Mark Johnson, Committee Member; Dr. Leda Lunardi, Committee Member; Dr. Kevin Gard, Committee Member
This research focuses on the characterization and modeling of AlGaN⁄InGaN MOS-HFETs. DC and small signal S parameter measurements were used to characterize these FETs and the associated AlGaN⁄InGaN MOS heterojunction varactors. An equivalent circuit model was developed for the AlGaN⁄InGaN MOS varactor. The model accurately represents measured S-parameters of the device from 45 MHz to 10 GHz over the entire operating range of the device (gate bias varying from -8V to 6V). The extracted gate capacitance indicated the presence of an accumulation layer in the AlGaN barrier layer in forward bias (Vg > 3V). A physics based large signal model has been developed for the varactor. The model utilizes the triangular well approximation to describe charge control at the heterointerface and takes polarization into account. Free carrier generation and neutralization of donor atoms in the AlGaN barrier layer. The model accurately fits the extracted gate capacitance in HFET mode of operation and exhibits the real space transfer of free charge into the barrier layer. The second part of this work focused on the small signal characterization and modeling of the FETs. A direct extraction technique has been developed to extract the small signal components of a FET in presence of bias dependent series resistances. This method was applied to the extraction of small signal equivalent circuits for the AlGaN/InGaN MOS-HFET which exhibited varying source and drain resistances with bias. In the third part, a large signal model has been developed for these devices using the linear charge control equation while taking polarization effects into account. The intrinsic device model uses a quasi two dimensional solution of the Poisson equation in the channel to take velocity saturation effects into account. The capacitances are modeled by developing analytical expressions for the channel charge partitioned between the source and drain. The model has been implemented in Agilent ADS using Verilog-A and is compared to measured DC IV and small signal parameters. The limitations of this model are discussed and methods to enhance it are proposed. Finally, work done on characterization of GaN n-i-n structures utilizing re-grown source drain contacts is presented and discussed. The devices were modeled using expressions developed for devices with a uniform trap distribution in the bandgap.
Gallium Nitride Ultraviolet Optical Modulators
(2005-02-28) Oberhofer, Andrew Edward; Dr. John F. Muth, Committee Chair; Dr. Richard T. Kuehn, Committee Member; Dr. Salah M. Bedair, Committee Member; Dr. Dennis M. Maher, Committee Member; Dr. Mark Johnson, Committee Member
In narrower band gap semiconductors researchers have exploited the ability to manipulate the exciton resonance via the Quantum Confined Stark Effect to make a variety of different types of optical modulators at infrared wavelengths. In this thesis, the large exciton binding energy of Gallium Nitride is used as the basis for ultraviolet optical modulators without the need for quantum confinement. A 5x5 array of UV optical modulators at 360 nm was fabricated. The modulators operated in a transverse geometry and consisted of a GaN active layer surrounded by transparent AlGaN insulating and electrical contact layers. The typical thickness of the GaN layer was 0.4 um so the effects of the electric field on the exciton resonance could be directly observed. A hydrogenic model for the bulk exciton was assumed. The applied electric field opposed the attractive coulomb potential between the electron and hole and broadens the exciton resonance. This results in more or less light through the device depending on the spectral position. To understand the magnitude of the applied field within the device structure a 1D Poisson Solver was used. Spontaneous polarization and piezoelectric effects due to lattice strain between the AlGaN and GaN layers were included in the model and were found to have influence on the device at lower operating voltages. In the electric field modulated devices a contrast ratio of about 20 percent was obtained. In thermally modulated devices, at low frequencies less than 200 Hz large shifts in the band edge led to large contrast ratios as expected. The temperature dependence of the device followed the Varshni relationship and allowed the magnitude of the temperature shift to be quantified. At higher frequencies from 1kHz to 120 kHz an optical modulation of ~ 5 percent was readily observed and was attributed to electronic effects. The limitation of 100 kHz was equipment related and it is conjectured that the modulation bandwidth would extend into the MHz.
Growth and Fabrication of GaN and InxGa1-xN Based Optoelectronic Devices
(2008-08-18) Berkman, Erkan Acar; Dr. Salah M. Bedair, Committee Co-Chair; Dr. Nadia A. El-Masry, Committee Chair; Dr. Mark Johnson, Committee Member; Dr. Jerry J. Cuomo, Committee Member
In this study, heteroepitaxial growth of III-Nitrides was performed by metalorganic chemical vapor deposition (MOCVD) technique on (0001) Al2O3 substrates to develop GaN and InxGa1-xN based optoelectronic devices. Comprehensive experimental studies on emission and relaxation mechanisms of InxGa1-xN quantum wells (QWs) and InxGa1-xN single layers were performed. The grown films were characterized by x-ray diffraction (XRD), Hall Effect measurements, photoluminescence measurements (PL) and transmission electron microscopy (TEM). An investigation on the effect of number and width of QWs on PL emission properties of InxGa1-xN single QWs and multi-quantum wells (MQW) was conducted. The experimental results were explained by the developed theoretical bandgap model. The study on the single layer InxGa1-xN films within and beyond critical layer thickness (CLT) demonstrated that thick InxGa1-xN films display simultaneous presence of strained and (partially) relaxed layers. The In incorporation into the lattice was observed to be dependent on the strain state of the film. The findings on InxGa1-xN QWs and single layers were implemented in the development of InxGa1-xN based LEDs and photodiodes, respectively. The as-grown samples were fabricated using conventional lithography techniques into various optoelectronic devices including long wavelength LEDs, dichromatic monolithic white LEDs, and p-i-n photodiodes. Emission from InxGa1-xN⁄GaN MQW LEDs at wavelengths as long as 625nm was demonstrated. This is one of the longest peak emission wavelengths reported for MOCVD grown InxGa1-xN MQW structures. Dichromatic white emission in LEDs was realized by utilizing two InGaN MQW active regions emitting at complementary wavelengths. InGaN p-i-n photodiodes operating at various regions of the visible spectrum tailored by the i-layer properties were developed. This was achieved by the novel approach of employing InxGa1-xN in all layers of the p-i-n photodiodes, enabling nearly-lattice matched growth. The photodiodes displayed zero-bias responsivity values as high as 0.037A⁄W, and the peak responsivity wavelength of the photodiodes ranged between 416nm and 466nm. To the author's best knowledge, the latter value remains the longest peak detection wavelength among InxGa1-xN based p-i-n photodiodes.
Investigation of Surface States in Gallium Nitride Devices using a New High Frequency Measurement Technique
(2006-10-05) Ramachandran, Ramya; Dr. Douglas Barlage, Committee Chair; Dr. Robert Kolbas, Committee Member; Dr. Mark Johnson, Committee Member
Surface states place a limitation on the high-frequency behavior of Gallium Nitride devices by causing RF dispersion. They are also a source of undesirable 1⁄f noise. This thesis specifically aims to explore techniques to address known experimental observations of dispersion phenomena of unknown origin in the 1-30 GHz frequency range. A new method to investigate these at high frequencies is proposed. It involves the measurement of S-parameters between the drain and source of a GaN nin structure in the GHz frequency range. The basis of the proposed technique is the assumption that the behavior of surface states and other dispersion phenomena can be isolated by subtracting the behavior of the device from the measured Y-parameters.
A Physics-based Large-signal Analytical Model for AlGaN/GaN HFETs
(2008-08-26) Yin, Hong; Dr. Griff L. Bilbro, Committee Chair; Dr. Robert J.Trew, Committee Co-Chair; Dr. Doug Barlage, Committee Member; Dr. Mark Johnson, Committee Member
Planar Edge Defined Alternate Layer Process (PEDAL) - An Unconventional Technique for Fabricatinon of Wafer Scale Sub-25 nm Nanowires and Nanowire Template
(2006-12-08) Sonkusale, Sachin Ramrao; Dr. Mark Johnson, Committee Member; Dr. Veena Misra, Committee Member; Dr. Gerald Iafrate, Committee Member; Dr. Paul D. Franzon, Committee Chair
As defined by the US national science foundation, "nanofabrication is the process of making functional structures with arbitrary patterns having minimum dimensions less than 100 nm". Nanofabrication, a key step in nanotechnology, has applications not only in conventional semiconductor devices but also in sensors, memory, nanofluidics, cross-bar logic architecture and nanoelectrical mechanical systems. In this research I have proposed and successfully demonstrated an unconventional lithographic technique called Planar Edge Defined Alternate Layer (PEDAL) to fabricate wafer scale sub 25 nm nanowire template. Good dimensional control and wafer scale uniformity of this process is shown by uniformity analysis of the width and spacing of an array of sixteen line-width structures with approximately 42 nm pitch and twenty four line-width structures with approximately 23 nm pitch. Results on routing capability of this process along with results of palladium nanowires obtained by PEDAL lift-off process done on the template with 42 nm pitch is also reported. In the case of template with array of sixteen lines, the average pitch of array across the 4 inch wafer was measured to be 40.83 nm with the standard deviation of 2.29 nm where as the average pitch of the lines in an array was found to be 41.5 nm with the standard deviation of 4.64 nm. After Pd lift-off the average pitch in nanowire array was measured to be 41.88 nm with standard deviation of 1.83 nm, close to the values obtained for the template. In the case of array of twenty four line-widths, average pitch of array across the 4 inch wafer was measured to be 21.1 nm with the standard deviation of 5 A where as the average pitch of the line in an array was found to be 22.6 nm with the standard deviation of 9 A. Other than experimental analysis, results from numerical simulations to find processing conditions to get good dimensional control in PEDAL process by taking process variations into account are also presented in this thesis.
Schottky Barrier GaN FET Model Creation and Verification using TCAD for Technology Evaluation and Design.
(2008-04-25) Ozbek, Ayse Merve; Dr. Mark Johnson, Committee Member; Dr.Doug Barlage, Committee Chair; Dr. Mehmet C. Ozturk, Committee Member