# Alumni Dissertations

#### Filter Dissertations By:

### A study of Optically Pumped Nuclear Magnetic Polarization on Gallium Arsenide

Author:Yunpu LiYear of Dissertation:2012Program:PhysicsAdvisor:Carlos MerilesAbstract:This thesis aims to study the dynamic process of optically pumped nuclear spin polarization on Gallium Arsenide. First of all, the time-resolved optical Faraday rotation is applied to observe the electron spin dynamics in the presence of the nuclear magnetic field. And then the optically-pumped NMR is measured on different parameters, including the dependences of light helicity, irradiation intensity, photon energy, illumination time and temperature. We report a new phenomenology at low irradiation intensity. A nuclear polarization model combining hyperfine and quadrupolar relaxation is developed, with experimental data supported. By exploiting the two competing mechanism on various photon energies, illumination intensity and NMR pulse sequences, we use high field stray-field NMR imaging to realize all-optical creation of three-dimensional patterning of positive and negative nuclear polarization on the micron length scale. Finally, the effect of nuclear spin diffusion effect is investigated. We demonstrate in a remarkable way the unambiguous evidence of diffusion, which results in a great enhancement of quadrupolar relaxation in the nanometer scale.

### SYSTEMATIC CONTROL OF SCHOTTKY BARRIER HEIGHT BY PARTISAN INTERLAYERS

Author:Yang LiYear of Dissertation:2012Program:PhysicsAdvisor:Raymond TungAbstract:The relationship between the starting surface structure and the Schottky barrier height (SBH) in metal-silicon systems has been investigated to explore the possibility of modifying the interface dipole through the insertion of an inorganic interface layer. A systematic and comprehensible way to perform this modification has been introduced as a "partisan interlayer" method and was extensively studied for a variety of interlayer elements, along with several choices for the metal. Employing elements with a larger electronegativity than that of silicon, a monolayer of As, S, or Cl was deposited on Si surfaces and processed to form stable surface structures. The electron affinities of these surfaces were measured by Kelvin probe and found to increase significantly from the clean surface, consistent with the expected charge transfer from Si to the adsorbates and also in agreement with results of ab initio density functional theory calculations. Subsequent deposition of metal on these adsorbate terminated semiconductor (ATS) surfaces led to the fabrication of metastable interface structures with the SBH successfully and significantly modified in a predictable manner from the clean Si results. The chemical stability of these surfaces that weakens the interaction with the deposited metal, likely leads to the preservation of electric dipole from such partisan interlayers. The partisan interlayer method was found to work particularly well with As-terminated Si(111) surface, on which a interface behavior parameter exceeding 0.50 was found. This exceeded the S-parameter usually observed for covalent semiconductors such as Si, ~ 0.1, and highlighted a major reason for the adjustability of the SBH by the PI method. The SBH of all interfaces studied in this work was inhomogeneous. Making use of the theory of electronic transport through inhomogeneous SBH and temperature-dependent measurements, the extent of the SBH nonuniformity was routinely characterized from the Schottky diodes. The largest adjustment in the SBH was observed for Au on S-terminated Si(100), where the n-type junction became nearly perfectly ohmic. It was demonstrated, for the first time, that quantitative information on the distribution of the SBH and the lateral size of conduction patches ("hot spots") could still be obtained from ohmic junctions. The physical basis for these analyses and the special experimental conditions which enabled these analyses were carefully explained. The implications of these results for SBH control of MS systems in general and the understanding of the formation of SBH in general are also discussed.

### Infrared and Raman Spectroscopy Study of Layered Systems

Author:Jian LiYear of Dissertation:2013Program:PhysicsAdvisor:Jiufeng TuAbstract:Optical spectroscopy studies the interaction between light (photon) and matter. During such interaction, different processes such as reflection, transmission, scattering, absorption or fluorescence can occur. Among all the optical spectroscopic techniques, infrared (IR) and Raman spectroscopy are most commonly used. In an Infrared process, photons are absorbed. The required light source emits polychromatic Infrared light and when it passes through or being reflected by the sample the light is partially absorbed. The frequency dependent absorption allows one to study the electronic and vibrational structure of the sample. On the other hand, the Raman spectroscopy is second order in nature where the photon is scattered instead of being absorbed. A monochromatic light source is used instead of a continuous spectrum. Generally, the dominate effect in an optical process is absorption and transmission but a (small) portion of photons are scattered. A small fraction of photons change their energy/wavelength during the scattering. Depending on the scale of the change in energy, those inelastic scatterings can be categorized into Brillouin scattering and Raman scattering. Although sharing the same mechanism, different energy scale require completely different experimental setups for Brillouin scattering and Raman scattering. In the study of infrared and Raman spectroscopy, group theory is a very helpful tool. The calculation of absolute intensity of an optical transition is rather difficult and sometimes infeasible, especially for crystal vibrations. Group theory is the mathematical language that describes the symmetry property of the physical system. Selection rules based on symmetry consideration had been predicted. Group theory, especially representation theory, is an important branch of condensed matter physics. Both theoretical and experimental results of my PhD research are presented. The topics being covered are: infrared study of iron based superconductor BaFe$_{1.85}$Co$_{0.15}$As$_{2}$; the study of Raman scattering with Laguerre-Gaussian (LG) beam.

### Phase locking of solid-state laser arrays

Author:LIPING LIUYear of Dissertation:2010Program:PhysicsAdvisor:Ying-Chih ChenAbstract:This thesis reports a study of phase locking in solid-state laser arrays of a variety of configurations, including a 2x2 Nd:YVO4 continuous-wave laser array, a two-element passively Q-switched Nd,Cr:YAG laser array, a two-element continuous-wave ytterbium fiber laser array, and a two-element ytterbium fiber lasers passively Q-switched by stimulated Brillouin scattering. Phase locking is accomplished by coupling the lasing elements into a common Fourier-transform resonator, with the lasing elements placed at one focal plane of a converging lens and the output mirror placed at the other focal plane. The control of the relative phase among the elements is done by placing a spatial filter in front of the output mirror to introduce different modal losses. We have succeeded in achieving highly stable phase-locked operation in the in-phase mode in continuous-wave laser arrays. The fringe visibility of the phase-locked beam is nearly 1. As the coupling strength decreases, the transition from phase locked to unlocked mode is abrupt. Phase locking of nano-second pulsed laser arrays requires a tight control of the resonance frequencies and path lengths of the individual lasing elements to ensure the pulses generated by the individual elements to occur simultaneously. As the difference in path lengths increase or the coupling strengths decrease, the transition from the phase locked to the unlocked states is characterized by a gradual loss of coincidence of the pulses from the individual elements and a reduction in the fringe contrast in the combined laser beam. The current approach of phase locking has high efficiency and is applicable to two-dimensional laser arrays containing a large number of elements.

### Systematic Tuning of Silicon Schottky Barrier Height by Atomic Interlayers with Low Electronegativities

Author:Wei LongYear of Dissertation:2012Program:PhysicsAdvisor:Raymond TungAbstract:The Schottky barrier height (SBH) is of great importance to the functionality of semiconductor devices, as it governs the carrier transport across the metal-semiconductor (MS) interface. The presence of the Fermi level (FL) pinning phenomena makes tuning the SBH a difficult goal to achieve. The technique of "partisan interlayer" (PI) was proposed recently to modify the SBH, where stable adsorbate-terminated semiconductor (ATS) surfaces were used to form SBs with subsequently applied metal. When elements with large electronegativities were used to form the ATS, the PI technique was effective in reducing the n-type SBH and increasing the p-type SBH, driven by the expected transfer of charge from the semiconductor to the adsorbates. In this thesis work, elements with electronegativities smaller than that of the semiconductor are used as surface termination. SBHs for Ag, Au and In on Si surfaces are found to increase for the n-type and decrease for the p-type interfaces, by as much as 0.25eV, when Ga, Mg and K are used to terminate the Si surfaces. The present results are thus in agreement with the expected charge transfers from elements with smaller electronegativities to silicon and illustrate the general validity of the PI technique. The chemical stability of these surfaces likely weakens the MS interaction and leads to the (partial) preservation of the surface dipole at the MS interface. However, large degrees of SBH inhomogeneity are observed for diodes on these surfaces, likely due to insufficient stability of these surfaces to completely withstand metal interaction. These results are discussed within the basic models of SBH formation and the implications of these results for SBH control of MS systems are also addressed.

### Magnetic deflagration in the molecular magnet Mn12-ac

Author:Sean McHughYear of Dissertation:2009Program:PhysicsAdvisor:Myriam SarachikAbstract:In 1995, Paulsen and Park observed abrupt spontaneous reversals of the magnetization in crystals of the molecular magnet Mn12-ac, which they dubbed ``magnetic avalanches". They suggested that the magnetic avalanches were a thermal runaway process where the reversing spins release heat stimulating further relaxation. Various exotic phenomena were proposed as an alternative explanations. In 2005, Suzuki et al. established that this spontaneous magnetic relaxation occurs as a ``front" separating regions of opposing magnetization that propagates at a constant speed through the crystal. They suggested that this propagating front is analogous to a flame in chemical deflagration and introduced the thermal relaxation process, magnetic deflagration. The analysis presented there was limited by lack of data. A more thorough comparison with the theory would require the ability to trigger avalanches in a more controlled way rather than wait for their spontaneous occurrence. The work presented in this thesis is a continuation of the program initiated by Suzuki. Significant progress experimental progress has been made allowing us to trigger avalanches over a wide range of conditions. The magnetization dynamics and the ignition temperatures are studied in detail using an array of micro-sized Hall sensors and Germanium thermometers. In addition, we report the existence of a new species of avalanches consisting only of the fast-relaxing isomers of Mn12-ac, the so-called ``minor species". We explore avalanches of both species, as well as the interaction between them. Finally, a detailed analysis is performed to compare the experiment with the theory of magnetic deflagration. We find the theory of magnetic deflagration to be consistent with the data and extract values for the key physical quantities: the thermal diffusivity and avalanche front temperatures. Agreement between our predicted values and an independent measurement of these quantities would provide compelling verification of the theory.

### Solid State NMR Studies of Materials for Energy Technology

Author:Chandana Nambukara Kodiweera ArachchilageYear of Dissertation:2010Program:PhysicsAdvisor:Steve GreenbaumAbstract:Abstract SOLID STATE NMR STUDIES OF MATERIALS FOR ENERGY TECHNOLOGY by Chandana K. Nambukara Kodiweera Arachchilage Adviser: Professor Steve Greenbaum Presented in this thesis are NMR investigations of the dynamical and structural properties of materials for energy conversion and storage devices. 1H and 2H NMR was used to study water and methanol transportation in sulfonated poly(arylene ether ketone) based membranes for direct methanol fuel cells (DMFC). These results are presented in chapter 3. The amount of liquid in the membrane and ion exchange capacity (IEC) are two main factors that govern the dynamics in these membranes. Water and methanol diffusion coefficients also are comparable. Chapters 4 and 5 are concerned with 31P and 1H NMR in phosphoric acid doped PBI membranes (para-PBI and 2OH-PBI) as well as PBI membranes containing ionic liquids (H3PO4/PMIH2PO4/PBI). These membranes are designed for higher-temperature fuel cell operation. In general, stronger short and long range interactions were observed in the 2OH-PBI matrix, yielding reduced proton transport compared to that of para-PBI. In the case of H3PO4/PMIH2PO4/PBI, both conductivity and diffusion are higher for the sample with molar ratio 2/4/1. Finally, chapter 6 is devoted to the 31P NMR MAS study of phosphorus-containing structural groups on the surfaces of micro/mesoporous activated carbons. Two spectral features were observed and the narrow feature identifies surface phosphates while the broad component identifies heterogeneous subsurface phosphorus environments including phosphate and more complex structure multiple P-C, P-N and P=N bonds.

### Quantum Rotational Effects in Nanomagnetic Systems

Author:Michael O'KeeffeYear of Dissertation:2013Program:PhysicsAdvisor:Eugene ChudnovskyAbstract:Quantum tunneling of the magnetic moment in a nanomagnet must conserve the total angular momentum. For a nanomagnet embedded in a rigid body, reversal of the magnetic moment will cause the body to rotate as a whole. When embedded in an elastic environment, tunneling of the magnetic moment will cause local elastic twists of the crystal structure. In this thesis, I will present a theoretical study of the interplay between magnetization and rotations in a variety of nanomagnetic systems which have some degree of rotational freedom. We investigate the effect of rotational freedom on the tunnel splitting of a nanomagnet which is free to rotate about its easy axis. Calculating the exact instanton of the coupled equations of motion shows that mechanical freedom of the particle renormalizes the easy axis anisotropy, increasing the tunnel splitting. To understand magnetization dynamics in free particles, we study a quantum mechanical model of a tunneling spin embedded in a rigid rotor. The exact energy levels for a symmetric rotor exhibit first and second order quantum phase transitions between states with different values the magnetic moment. A quantum phase diagram is obtained in which the magnetic moment depends strongly on the moments of inertia. An intrinsic contribution to decoherence of current oscillations of a flux qubit must come from the angular momentum it transfers to the surrounding body. Within exactly solvable models of a qubit embedded in a rigid body and an elastic medium, we show that slow decoherence is permitted if the solid is macroscopically large. The spin-boson model is one of the simplest representations of a two-level system interacting with a quantum harmonic oscillator, yet has eluded a closed-form solution. I investigate some possible approaches to understanding its spectrum. The Landau-Zener dynamics of a tunneling spin coupled to a torsional resonator show that for certain parameter ranges the system exhibits multiple Landau-Zener transitions. These transitions coincide in time with changes in the oscillator dynamics. A large number of spins on a single oscillator coupled only through the in-phase oscillations behaves as a single large spin, greatly enhancing the spin-phonon coupling.

### Wave Scattering in Random Layered Media

Author:Jongchul ParkYear of Dissertation:2010Program:PhysicsAdvisor:Azriel GenackAbstract:The scaling and statistics of the transport of waves in random media depend strongly on the dimensionality of the medium. The statistic of transmission in one dimension (1D) and quasi-1D (Q1D) have been calculated and tested. However, the statistic for other dimensions has not been established. Exploring transport in a layered system of stacks of glass cover slips with transverse nonuniformity has allowed us to study a dimensional crossover in transport from 1D towards 3D. The crossover occurs when the lateral spread of the wave become larger than the transverse coherence length in the transmitted speckle pattern as the number of layers increases. In thin samples, in which light does not spread beyond a single coherence area of the field on the output surface, the statistics of normalized intensity follow 1D statistics associated with a segment of a log-normal distribution with a sharp drop below the log-normal distribution for low values of intensity. Once the lateral spread is larger than the transverse coherence length, the probability density of intensity becomes a mixture of a mesoscopic distribution and an intensity distribution of a Gaussian field. This distribution was originally found for Q1D. Beyond 1D, the intensity statistics have a same form as Q1D statistics which is a function of a single localization parameter, the “statistical conductance”

g ’. This transition from 1D to Q1D statistics reflects a topological change in the transmitted field. In 1D, the transmitted intensity never vanishes, while beyond 1D, a speckle pattern built upon a network of phase singularities forms.### PHASE SPACE EXPLORATIONS IN TIME DEPENDENT DENSITY FUNCTIONAL THEORY

Author:ARUNA RAJAMYear of Dissertation:2010Program:PhysicsAdvisor:Neepa MaitraAbstract:Time dependent density functional theory (TDDFT) is one of the useful tools for the study of the dynamic behavior of correlated electronic systems under the influence of external potentials. The success of this formally exact theory practically relies on approximations for the exchange-correlation potential which is a complicated functional of the co-ordinate density, non-local in space and time. Adiabatic approximations (such as ALDA), which are local in time, are most commonly used in the increasing applications of the field. Going beyond ALDA, has been proved difficult leading to mathematical inconsistencies. We explore the regions where the theory faces challenges, and try to answer some of them via the insights from two electron model systems. In this thesis work we propose a phase-space extension of the TDDFT. We want to answer the challenges the theory is facing currently by exploring the one-body phase-space. We give a general introduction to this theory iv and its mathematical background in the first chapter. In second chapter, we carryout a detailed study of instantaneous phase-space densities and argue that the functionals of distributions can be a better alternative to the nonlocality issue of the exchange-correlation potentials. For this we study in detail the interacting and the non-interacting phase-space distributions for Hookes atom model. The applicability of ALDA-based TDDFT for the dynamics in strongfields can become severely problematic due to the failure of single-Slater determinant picture.. In the third chapter, we analyze how the phase-space distributions can shine some light into this problem. We do a comparative study of Kohn-Sham and interacting phase-space and momentum distributions for single ionization and double ionization systems. Using a simple model of two-electron systems, we have showed that the momentum distribution computed directly from the exact KS system contains spurious oscillations: a non-classical description of the essentially classical two-electron dynamics. In Time dependent density matrix functional theory (TDDMFT), the evolution scheme of the 1RDM (first order reduced density matrix) contains second-order reduced density matrix (2RDM), which has to be expressed in terms of 1RDMs. Any non-correlated approximations (Hartree-Fock) for 2RDM would fail to capture the natural occupations of the system. In v our fourth chapter, we show that by applying the quasi-classical and semiclassical approximations one can capture the natural occupations of the excited systems.We study a time-dependent Moshinsky atom model for this. The fifth chapter contains a comparative work on the existing non-local exchange-correlation kernels that are based on current density response frame work and the co-moving frame work. We show that the two approaches though coinciding with each other in linear response regime, actually turn out to be different in non-linear regime.