Thesis (Ph. D.)--University of Texas at Austin, 1999. / Vita. Includes bibliographical references (leaves 140-145). Available also in a digital version from Dissertation Abstracts.

The use of SiGe/Si heterostructures in the fabrication of electronic devices results in an improvement of the device performances with respect to bulk silicon. Ion implantation has been proposed as one of the possible technologies to produce these structures and, thus, the aim of this work is to develop an ion beam technology to fabricate strained SiGe heterostructures. The formation of extended defects in SiGe alloy layers formed by high dose Ge+ ion implantation followed by Solid Phase Epitaxial Growth (SPEG) has been investigated by transmission electron microscopy. Rutherford backscattering spectroscopy has also been used to determine the chemical composition and the crystalline quality of the synthesised structures. In addition, X-ray diffraction has been used to evaluate the strain level in selected samples. Two different structures have been studied in this project. The first consisted of "all-implanted" layers, where the Ge+ implants were followed in some cases by additional implants of Si+ and/or C+ ions, prior to SPEG, to investigate methods to inhibit defect formation. The second was achieved by capping the ion beam synthesised SiGe alloy layer by the deposition of a thin film of silicon, in order to realise structures compatible with device dimensions. Single crystal device worthy SiGe alloy layers have been achieved by implantation of Ge+ ions at energies ranging from 70 keV to 400 keV, where the only extended defects observed are EOR defects at a depth correspondent to the a/c interface formed during the Ge+ implant. In some cases, "hairpin" dislocations have also been observed in the vicinity of the EOR defects and extending up to the surface. Both types of defects are annihilated after post-amorphisation with 500 keV Si+ and replaced with dislocation loops at a depth of about 1 fj,m. For each Ge+ implantation energy a critical value of the peak germanium concentration exists above which the structures relax through the formation of stacking faults or "hairpin" dislocations nucleated in the vicinity of the peak of the germanium concentration depth profile and extending up to the surface. A critical value of the elastic energy stored in the structures (~300 mJ/m2) has been determined above which ion beam synthesised SiGe alloys relax, independently of the implantation energy. This empirical approach has been found to successfully account for the results obtained in this work as well as in many other studies reported in the literature. "Hairpin" dislocations formed under different experimental conditions have been investigated by plan view TEM and have been found to have the same crystallographic orientation () and Burgers vector (b= a ). Their formation has been explained within a "strain relaxation model". For a regrowth temperature of 700° C, all samples investigated by XRD have been found to be almost fully strained, including samples containing relaxation-induced defects, indicating that, under these conditions, the energy transferred to the defects is very low. C+ co-implantation has been successfully used to reduce both relaxation-induced defects and EOR dislocation loops. It is noted that a mixed technology entailing both layer deposition and ion implantation to produce the Si/SiGe/Si device structures requires extra process steps to control surface contaminations, pre cleaning and/or native oxide formation, resulting in increased fabrication costs. In this work an " all-implanted" route to the synthesis of Si/SiGe/Si device structures is therefore described, which exploits all of the advantages given by ion implantation.

Based on first principles Density Functional Theory calculations, we have investigated for possible paths for engineering electronic band structure of Transition Metal Dichalco- genides (TMDs). We have considered two approaches which have shown to be promising for engineering electronic bands of TMDs: substitutional chemical doping and heterostruc- turing. All the calculations are done using first principles Density Functional Theory as it is implemented in Quantum Espresso package. Two possible substitutional doping meth- ods for MoS2 are considered in our calculations; cation doping where Mo is substituted by metal atoms and anion doping where Nitrogen and halogen group atoms take the posi- tion of S-sites. We observe the n-type characteristics for halogen group doping and p-type characteristics for Nitrogen group doping at S site. Similarly, we observe these bipolar characteristics when substituted by the transition metal elements (4d elements in the peri- odic table) at Mo site. Our results on doping monolayer MoS2 are in agreement with those results obtained by Dolui et al. for similar systems. Our work is extended to explore the effect of substitutional doping in bilayer MoS2. We observe the promising bipolar char- acteristics on doping while the magnitude of the band gap decreases upon the controlled S-site doping with F and As. In the second part, we considered two types of heterostructuring; Van der Waals heterostructures, and lateral heterostructures. In Van der Waals heterostructures, a direct band gap is observed with a physical separation of charges into two layers from orbital isosurface plots. We present a brief overview of the folding of energy bands in supercell approach. Using heterogeneous supercell approach, we studied the electronic properties of a mixed system of MoS2 -WS2 . The separation of the charges into the two sections shows that our MoS2 -WS2 in-plane heterostructure shows a potential for a pn junction. These systematic studies of the doped and heterostructures of TMDs can be useful for device applications.

Recent advances in experimental and computational methods have opened up new directions in graphene fundamental studies. In addition to understanding the basic properties of this material and its quasi-one dimensional structures, significant efforts are devoted to describing their long ranged dispersive interactions. Other two-dimensional materials, such as silicene, germanene, and transition metal dichalcogenides, are also being investigated aiming at finding complementary to graphene systems with other "wonder" properties. The focus of this work is to utilize first principles simulations methods to build our basic knowledge of structure-interaction relations in two-dimensional materials and design their properties. In particular, mechanical folding and extended defects in zigzag and armchair graphene nanoribbons can be used to modulate their electronic and spin polarization characteristics and achieve different stacking patterns. Our simulations concerning zigzag silicene nanoribbons show width-dependent antiferromagnetic-ferromagnetic transitions unlike the case of zigzag graphene nanoribbons, which are always antiferromagnetic. Heterostructures, build by stacking graphene, silicene, and MoS$_2$, are also investigated. It is found that hybridization alters the electronic properties of the individual layers and new flexural and breathing phonon modes display unique behaviors in the heterostructure compositions. Anchored to SiC substrate graphene nanoribbons are also proposed as possible systems to be used in graphene electronics. Our findings are of importance not only for fundamental science, but they could also be used for future experimental developments.

No description available.

The ever-increasing demand for superior devices with a smaller footprint in electronics calls for research on novel materials as a potential replacement of or integration to the existing silicon-based technology. The emergence of two-dimensional semiconductors paved a promising path in this direction. Easy isolation of atomically thin and flat layers with dangling bond free surfaces enables these materials to not only form 2D vertical heterostructures with novel properties but also facilitates advanced transistor, diode, and tunnel-device design with characteristics such as unprecedented gate-control of the channel, extremely high mobility of charge carriers, high current density, and high on-off ratios. However, like any other technology at the early development phase, 2D semiconductor research also faces numerous challenges which are needed to be addressed. In this work, we address two such challenges in the field–modeling of vertical electrostatics in these complex novel devices which enables better understanding and prediction of their characteristics and overcoming the contact resistance issue in a promising 2D semiconductor, WSe2, which enables the advancement of these devices towards near-deal characteristics. To predict and analyze the electrical characteristics of 2D vertical heterostructures, we need to develop solid understanding of the potential landscape, charge distribution, and energy band diagrams in these devices. Conventional modeling approaches and simulation tools that have been used so far to simulate the transport characteristics obscure our intuition as the devices get more arbitrary and complex. Here, we developed a circuit equivalent model to simulate the vertical electrostatics in these novel and arbitrary heterostructures in a simple and intuitive manner. In our model, all the parameters of the energy band diagram are represented by equivalent circuit elements involving capacitors and voltage sources. We also provide an elegant approach to solve these circuits by using Gauss law in electrostatics and charge-neutrality conditions in quasi-equilibrium. With a computationally efficient algorithm developed to solve these structures, we further built an opensource tool 2dmatstack on nanohub.org that enables researchers to predict and analyze the characteristics of novel heterostructures to maximize research output. In the next section, we focus on a major bottleneck in realizing these vertical devices experimentally. Fermi-level pinning and process-induced surface damage cause large Schottky barriers between metal contacts and these ultrathin 2D semiconducting layers resulting in large contact resistance and poor, non-ideal device performance. The solution to this problem is much more developed in the most widely studied n-type candidate, MoS2, compared to the common the p-type candidate, WSe2. In this work, we develop a UV-ozone-based oxidation technique that transforms the top layer of WSe2 into a nonstoichiometric oxide, TOS, that degenerately dopes the layers underneath p-type. This high hole-doping decreases the Schottky barrier width at the contacts and has resulted in the lowest p-type contact resistance to ultrathin WSe2 reported thus far. We show that this doping is stable in the ambient, remains active at low temperatures, repeatable, robust, and area selective for contact-doping without altering the channel properties. The high-performance ohmic contacts we demonstrate not only sets us in the path to realize near-ideal channel-dominated devices but also is pivotal to understand these devices better by eliminating the effect of contacts from the gate-controlled channel characteristics.

Graphene, a monolayer of carbon atoms, has gained prominence to augment existing chip-scale photonic and optoelectronic applications, especially for sensing in optical radiation, owing to its distinctive electrical properties and bandgap as well as its atomically thin profile. As a building block of photodetection, graphene has been co-integrated with mature silicon technology to realize the on-chip, high-performance photo-detecting platforms with broad spectral response from the deep-ultraviolet (UV) to the mid-infrared (MIR) regime. The recent state-of-the-art graphene-based photodetectors utilizing the combination of colloidal quantum dots (QDs) and graphene have been intensively studied, where QDs function as the absorber and the role of graphene is as a fast carrier recirculating channel. With such a configuration, an ultrahigh sensitivity can be achieved on account of the photogating mechanism; however, the response time is slow and limited to the millisecond-to-second range. To achieve balance between high sensitivity and fast response time, we have demonstrated a new photodetector that is based on graphene/two-dimensional heterostructures. The homogeneous thickness and the large contact of the heterostructure give rise to fast carrier transporting between the thin absorber layer and the graphene, leading to a fast response time. This thesis carefully investigates the optimization of fabrication as well as optoelectronic characterization of photodetectors based on graphene/semiconductor heterostructures field-effect transistors (GFETs). GFETs with different architectures were demonstrated and systematically studied under optical illumination ranging from deep-UV to MIR at varying optical powers. Noise behaviors have been studied under different device parameters such as device structure, area and gate-bias. Results show that the flicker noise of graphene-based devices can be explained by the McWhorter model in which the fluctuation of carrier numbers is the dominant process of noise in low frequencies; thus, it can be scaled down by reducing the number of introduced charged carriers with optimized fabrication. Besides, the impact of absorber on top of graphene and the bottom substrate has been comprehensively explored through various experimental techniques including current-voltage (IV), photo-response dynamics, and noise characterization measurements. With our configuration, the high sensitivity and fast response time of photodetectors can be obtained at the same time. In addition to this, the study of the bottom substrate with different doping levels suggests a concept of dual-photogating effect which is induced by the top absorbent material and the photoionization of the doped silicon substrate. In summary, this thesis showcases novel device architecture and fabrication procedures of GFETs photodetectors, optimizes device structure, quantifies the performance and evaluates the effect of various absorbent materials and substrate. It provides insight into the improvement of possible routes to achieve a broadband photo-detecting system with higher sensitivity, faster response time and lower noise level. / Doctor of Philosophy / The rapid expansion of networked devices and the development of the Internet of Things have given rise to an internet traffic and data explosion. Since conventional electrical interconnects are unable to rise to the occasion of the ever-growing demands of information technology and communication networking, next-generation alternative interconnects with higher performance and lower loss are attractive alternatives as the chip-scale optical interconnection. Among various optical interconnects, photodetectors play significant roles by converting optical input into electrical signal output. Sensing of light has a great impact in daily applications such as telecommunications, night vision, biomedical imaging and biochemical sensing. Graphene, belonging to the class of 2-dimensional materials, shows enormous potential as a building block of photodetection owing to its outstanding optical and electrical properties. One possible route to develop a sensitive and fast-operating on-chip photodetector is to integrate graphene into silicon photonics platforms since the latter has been widely studied and driven to maturity. In this thesis, graphene-based photodetectors with novel architectures have been fabricated, demonstrated and systematically investigated. Various measurements have been taken to quantify the performance of photodetectors in a wide detecting range from deep ultraviolet to mid-infrared.

<p>Technological advancements in electronics industry are driven by innovations in device fabrication techniques and development of novel materials. Halide perovskites are one of the latest additions to the semiconductor family. The performance of solid-state devices based on halide perovskites is now competing with other well-established semiconductors like silicon and gallium arsenide. However, the intrinsic instability of three-dimensional (3D) perovskites poses a great challenge in their widespread commercialization. The soft crystal lattice of hybrid halide perovskites facilitates anionic diffusion which impacts material stability, optoelectronic properties, and solid-state device performance.</p> <p>Two-dimensional (2D) halide perovskites with organic capping layers have been used for improving the extrinsic stability as well as suppressing intrinsic anionic diffusion. Nevertheless, a fundamental understanding of the role of compositional tuning, especially the impact of organic cations, in inhibiting anionic diffusion across the perovskite-ligand interface is missing. In our research, we first developed a library of atomically sharp and flat 2D heterostructures between two arbitrarily determined phase-pure halide perovskite single crystals. This platform was then used to perform a systematic investigation of anionic diffusion mechanism and quantify the impact of structural components on anionic inter-diffusion in halide perovskites. </p> <p>Stark differences were observed in anionic diffusion across 2D halide perovskite lateral and vertical heterostructures. Halide inter-diffusion in lateral heterostructures was found to be similar to the classical Fickian diffusion featuring continuous concentration profile evolution. However, vertical heterostructures show a “quantized” layer-by-layer diffusion behavior governed by a local free energy minimum and ion-blocking effects of the organic cations. For both lateral and vertical migrations, halide diffusion was found to be faster in perovskites with larger inorganic layer thickness. The increment becomes less apparent as the inorganic layer thickness increases, akin to the quantum confinement effect observed for band gaps. Furthermore, we found that bulkier and more rigid π-conjugated organic cations inhibit halide inter-diffusion much more effectively compared to short chain aliphatic cations. These results offer significant insights into the mechanism of anionic diffusion in 2D perovskites and provide a new materials platform for heterostructure assembly and device integration.</p>

OPTICAL CHARACTERIZATION OF INGAN HETEROSTRUCTURES FOR BLUE LIGHT EMITTERS AND VERTICAL CAVITY LASERS: EFFICIENCY AND RECOMBINATION DYNAMICS By Serdal Okur, Ph.D. A thesis submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy at Virginia Commonwealth University. Virginia Commonwealth University, 2014. Major Director: Ümit Özgür, Associate Professor, Electrical and Computer Engineering This thesis explores radiative efficiencies and recombination dynamics in InGaN-based heterostructures and their applications as active regions in blue light emitters and particularly vertical cavities. The investigations focus on understanding the mechanism of efficiency loss at high injection as well as developing designs to mitigate it, exploring nonpolar and semipolar crystal orientations to improve radiative efficiency, integration of optimized active regions with high reflectivity dielectric mirrors in vertical cavity structures, and achieving strong exciton-photon coupling regime in these microcavities for potential polariton lasing. In regard to active regions, multiple double heterostructure (DH) designs with sufficiently thick staircase electron injection (SEI) layers, which act as electron coolers to reduce the overflow of hot electrons injected into the active region, were found to be more viable to achieve high efficiencies and to mitigate the efficiency loss at high injection. Such active regions were embedded in novel vertical cavity structure designs with full dielectric distributed Bragg reflectors (DBRs) through epitaxial lateral overgrowth (ELO), eliminating the problems associated with semiconductor bottom DBRs having narrow stopbands and the cumbersome substrate removal process. Moreover, the ELO technique allowed the injection of carriers only through the high quality regions with substantially reduced threading dislocation densities compared to regular GaN templates grown on sapphire. Reduced electron-hole wavefunction overlap in polar heterostructures was shown to hamper the efficiency of particularly thick active regions (thicker than 3 nm) possessing three-dimensional density of states needed for higher optical output. In addition, excitation density-dependent photoluminescence (PL) measurements showed superior optical quality of double heterostructure (3 nm InGaN wells) active regions compared to quantum wells (2 nm InGaN wells) suggesting a minimum limit for the active region thickness. Therefore, multiple relatively thin but still three dimensional InGaN active regions separated by thin and low barriers were found to be more efficient for InGaN light emitters. Investigations of electroluminescence from light emitting diodes (LEDs) incorporating multi DH InGaN active regions (e.g. quad 3 nm DH) and thick SEIs (two 20 nm-thick InGaN layers with step increase in In content) revealed higher emission intensities compared to LEDs with thinner or no SEI. This indicated that injected electrons were cooled sufficiently with thicker SEI layers and their overflow was greatly reduced resulting in efficient recombination in the active region. Among the structures considered to enhance the quantum efficiency, the multi-DH design with a sufficiently thick SEI layer constitutes a viable approach to achieve high efficiency also in blue lasers. Owing to its high exciton binding energy, GaN is one of the ideal candidates for microcavities exploiting the strong exciton-photon coupling to realize the mixed quasiparticles called polaritons and achieve ideally thresholdless polariton lasing at room temperature. Angle-resolved PL and cathodoluminescence measurements revealed large Rabi splitting values up to 75 meV indicative of the strong exciton-photon coupling regime in InGaN-based microcavities with bottom semiconductor AlN/GaN and a top dielectric SiO2/SiNxDBRs, which exhibited quality factors as high as 1300. Vertical cavity structures with all dielectric DBRs were also achieved by employing a novel ELO method that allowed integration of a high quality InGaN cavity active region with a dielectric bottom DBR without removal of the substrate while forming a current aperture through the ideally defect-free active region. The full-cavity structures formed as such were shown to exhibit clear cavity modes near 400 and 412 nm in the reflectivity spectrum and quality factors of 500. Although the polar c-plane orientation has been the main platform for the development of nitride optoelectronics, significant improvement of the electron and hole wavefunction overlap in nonpolar and semipolar InGaN heterostructures makes them highly promising candidates for light emitting devices provided that they can be produced with good crystal quality. To evaluate their true potential and shed light on the limitations put forth by the structural defects, optical processes in several nonpolar and semipolar orientations of GaN and InGaN heterostructures were investigated. Particularly, stacking faults were found to affect significantly the optical properties, substantially influencing the carrier dynamics in nonpolar (1-100), and semipolar (1-101) and (11-22)GaN layers. Carrier trapping/detrapping by stacking faults and carrier transfer between stacking faults and donors were revealed by monitoring the carrier recombination dynamics at different temperatures, while nonradiative recombination was the dominant process at room temperature. Although it is evident that nonpolar (1-100)GaN and semipolar (11-22)GaN require further improvement of material quality, steady-state and time-resolved PL measurements support that (1-101)-oriented GaN templates and InGaN active regions exhibit optical performance comparable to their highly optimized polar c-plane counterparts, and therefore, are promising for vertical cavities and light emitting device applications.

Modern semiconductor fabrication techniques allow for the fabrication of semiconductor heterostructures which host electron transport with a minimum of scattering sites. In such devices, electrons populate a two-dimensional electron gas (2DEG) in which electrons propagate in exactly two dimensions, and may be further confined by potential barriers to form electron billiards. At sub-Kelvin temperatures, electron trajectories are determined largely by reflections from the billiard walls, while net conduction through the device depends on quantum mechanical wave interference. Measurements of magnetoconductance fluctuations (MCF) serve as a probe of dynamics within the electron billiard. Many prior studies have utilized heterostructures employing the modulation doping architecture, in which the 2DEG is spatially removed from the donor atoms to minimize electron scattering. Theoretical studies have claimed that MCF will be fractal when the confinement potential defining the billiard is soft-walled, regardless of the presence of smooth potentials within the billiard such as those introduced by remote ionized donors. The small-angle scattering sites resulting from these potentials are often disregarded as negligible; we use MCF measurements to investigate such claims. To probe the effect of remote ionized donor scattering on the phase space in electron billiards, we compare MCF measured on billiards in a modulation-doped heterostructure to those measured on billiards in an undoped heterostructure, in which this potential landscape is believed to be absent. Fractal studies are performed on these MCF traces, and we find that MCF measured on the undoped billiards do not exhibit measurably different fractal characteristics than those measured on the modulation-doped billiards. Having confirmed that the potential landscapes in modulation-doped heterostructures do not affect the electron phase space, we then investigate the effect of these impurities on the distribution of electron trajectories through the billiards. By employing thermal cycling experiments, we demonstrate that this distribution is highly sensitive to the precise potential landscape within the billiard, suggesting that modulation-doped heterostructures do not support fully ballistic electron transport. We compare our MCF correlation data with the dynamics of charge transfer within heterostructure systems to make qualitative conclusions regarding these dynamics.