This thesis presents a study of the electron-phonon interaction in two dimensional electron gases (2DEGs), by measuring of the acoustic phonon emission from a sequence of n-type doped GaAs/(AlGa)As quantum wells. Previous studies of emISSIon from 2DEGs confined in GaAs heterojunctions (Chin et al., 1984) have shown a surprising absence of longitudinal acoustic (LA) mode phonon emission, in contrast with theoretical studies (Vass, 1987) which predict that deformation potential coupled LA mode emission should dominate the energy relaxation processes. This may be attributed to the finite width of the quasi-2D sheet, which imposes a restriction on the maximum emitted phonon wavevector component perpendicular to the 2DEG, leading to a suppression of the emission (the "1Iao cutoff') at smaller phonon wavevectors than predicted by the earlier theory. By using the quantum well width w as a means of modulating the thickness of the 2DEG, the dependence of the 1Iao cutoff on the phonon emission can be directly measured. In the present work, significant LA phonon emission from the quantum well samples is observed. To complement the experimental measurements, the theory of emission from a 2DEG has been modelled in detail using computer simulation techniques. Calculations of the electron-phonon interaction, including matrix element anisotropy and dynamic screening, as well as phonon focusing effects, can be combined to produce accurate predictions of the experimentally detected phonon emission energy spectra.

Thesis (Ph.D)--University of Macquaire, 2003. / Also published on CD-ROM. Includes bibliographical references.

Thesis (M.S.)--Virginia Polytechnic Institute and State University, 1993. / Vita. Abstract. Includes bibliographical references (leaves 116-125). Also available via the Internet.

Thesis (Ph. D.)--University of Oregon, 2002. / Typescript. Includes vita and abstract. Includes bibliographical references (leaves 665-706). Also available for download via the World Wide Web; free to University of Oregon users.

Thesis (Ph. D.)--Fletcher School of Law and Diplomacy, 2004. / Vita. Includes bibliographical references (leaves 571-601).

Dynamic Nuclear Polarization (DNP) is a method for signal enhancement in NMR, with numerous applications ranging from medicine to spectroscopy. Despite the success of applications of DNP, the understanding of the underlying theory is still limited. Much of the work on the theory of DNP has been carried out on small spin systems; this is a restriction due to the exponential growth of the Liouville space in quantum simulations. In the work described in this thesis, a methodology is presented by which this exponential scaling can be circumvented. This is done by mathematically projecting the DNP dynamics at resonance onto the Zeeman subspace of the density operator. This has successfully been carried out for the solid effect, cross effect and recently for the Overhauser effect in the solid state (see appendix A.4). The results are incoherent state-dependent dynamics, resembling classical behaviour. Such form of effective dynamics allows the use of kinetic Monte Carlo algorithms to simulate polarization dynamics of very large spin systems; orders of magnitude larger than has previously been possible. We verify the accuracy of the mathematical treatment of SE-DNP and CE-DNP, and illustrate the insight large spin-system simulations provide into the mechanism of DNP. For SE-DNP the mechanism of polarization to the bulk of spin systems is determined to be spin diffusion, and we carried out studies into the efficiency and performance of radicals, with an outlook on radical design. We also show that the Zeeman projection can be applied to heteronuclear spin systems if the nuclear species are close in frequency, and we present a formalism for simulating C-13 nuclear spin systems based on a linear rate approach, enabling simulations of thousands of spins in a matter of minutes. A study into the scaling of the kinetic Monte Carlo algorithm error, and the simulation run time, with respect to an increasing number of spins is also presented. For CE-DNP the error analysis led to establishing a parameter regime in which the effective dynamics are accurate. We show that spin diffusion is the mechanism of transfer of polarization to bulk nuclei. We also show how the effective rates for CE-DNP can be used to understand the efficiency of bi-radicals, point to optimisation possibilities, and hold a potential to aid in bi-radical design. We finally show large scale simulations for CE-DNP bi-radical systems with improved parameters; leading to very rapid build-up of nuclear polarization.

Increased signal to noise ratio in ultra high field Magnetic Resonance Imaging (MRI) has allowed the development of quantitative imaging techniques and new contrast mechanisms, such as Chemical Exchange Saturation Transfer (CEST) to be probed. The development of CEST contrast imaging has involved overcoming a number of technical challenges associated with ultra high field MRI. The B1 transmit field was, and still is, a major challenge. Presented in this thesis, the B1 transmit field in regions of low B1 are improved with the use of dielectric pads and a simulation study shows that the overall B1 transmit field homogeneity is significantly improved when multi-transmit slice-selective RF spokes pulse sequences are used. Multiple methods have been developed to quantify the chemical exchange from slow exchanging proton pools seen in CEST contrast imaging. However, magnetisation transfer (MT) from the macromolecular bound pool contaminates current quantification methods, and presented in this thesis is a method whereby the CEST and MT are simultaneously saturated using dual frequency saturation pulses, allowing the CEST contrast in z-spectra to be separated from the MT and to enhance visualisation of the CEST effects. Despite the considerable interest in CEST, only one study has probed the CEST effects in blood, and interestingly high levels of CEST signals can be observed from the superior sagittal sinus. To investigate these effects, z-spectra from ex vivo blood samples considering the effects of oxygenation, haematocrit levels and cell structure were quantified. Quantification shows that the main source of the CEST signals was from the cells within the blood.

Dynamic Nuclear Polarisation (DNP) produced molecules with spin polarisation levels that are up to three orders of magnitude larger than their thermal equilibrium values. Most DNP mechanisms work at temperatures of 2 K and lower, meaning that the sample is stored in the solid-state. Combining DNP with a rapid temper- ature jump to room temperature allows liquid-state NMR analysis with a signal that is ve orders of magnitude higher than observed with thermal polarisation. However, the information obtained during a dissolution experiment is limited by the intrinsic liquid-state longitudinal relaxation of the spins of interest. is thesis looks to increase the information acquired in a number of ways. First, by devel- oping a new dissolution system for the dual iso-centre magnet it was possible to reproducibly perform enhanced NMR acquisition 600 ms following sample disso- lution. is has allowed the observation of hyperpolarised 13C spins with T1 times as low as 200 ms. Complimentary information can be obtained following sample dissolution by observing multiple spin species simultaneously. 13C and 15N spins are both polarised by microwave irradiation of the same frequency, so both can be analysed during a single dissolution DNP experiment. A novel probe has been used that contains six individual 13C microcoils. ese coils are separated in space and operate independently. is probe, in conjunction with dissolution DNP, can be used for observing dynamic molecular information on the time scale of 200 ms, however with further development this time scale should drop to less than 100 ms while maintaining a required minimum spectral resolution. Initial tests have been performed with both thermally polarised and hyperpolarised samples.

This thesis studies non-equilibrium open quantum systems where the dissipation is crucial to the achievement of novel physical regimes. We focus on atomic systems which allow for the coupling of a ground state to a Rydberg state, relying on the strong interactions between Rydberg atoms to produce the collective behaviour that we aim to investigate. For atoms in an optical lattice undergoing standard dissipation forms, e.g. loss and dephasing, we find these simple settings allow for the production of models contained in the non-equilibrium realm. We start by looking at a system with engineered pair dissipation on a one-dimensional lattice. When the dissipation is strong relative to a tunnelling process it creates a quantum Zeno effect which projects the system onto a Zeno-subspace. This subspace is found to contain complexes which experience a binding due to the dissipation. The properties of these complexes are found to feature spin-orbit coupling and, in certain instances, a flat band. We then study what kinetically constrained models (KCMs) can be reproduced in a lattice system. KCMs are models which typically feature trivial steady states, but a complex relaxation dynamics. These models appear in the fields of glasses and soft matter physics. We find a general framework for the consideration of a quantum Hamiltonian and a classical potential with strong dephasing noise. We then focus on a model mimicking volume excluded KCMs and find characteristic constrained behaviour, such as ergodicity breaking. We apply this framework to the decay of a many-body localised state in an open system with interactions in which we find the decay to be classical in the two interaction limits. For weak interactions, it follows a stretched exponential form due to pair relaxation, while for strong interactions the decay follows a compressed exponential, now being modelled as an Avrami process due to the correlated relaxation. We also find that on-site loss only affects the strong interacting limit. We then move on to the study of universal non-equilibrium behaviour in the directed percolation (DP) class. We consider on-site atomic loss and gain as a substitute for the standard decay channel. We show that this replaces the absorbing state with an enlarged absorbing space, leading to a loss of the DP transition at lower average densities. This class of DP-like systems has received little study, and we present a method of experimentally realising it in current set-ups. We finish with a look at a quantum DP model, where we consider its quantum and classical limits. We find that the transition changes from first to second order as the system becomes more classical, featuring a bi-critical point. We then numerically demonstrate that the same transitions are visible in idealised and Rydberg models.

The research work in this thesis is based on strongly interacting quantum lattice systems. The biggest part of research was conducted using state-of-the-art tensor network simulations. Tensor networks provide efficient and highly accurate representations of quantum states when any simply connected patch of the quantum state is slightly entangled. Matrix Product States (MPS) is a tensor network representation for quantum states which is quasi-exact for one-dimensional systems when entanglement entropy of any bipartition of the state follows ``area-law". Projected Entangled Pair States (PEPS) is the extension of MPS for higher dimensional systems, where entanglement entropy area-law is non-trivial. These formalisms and their relevant ground state optimization techniques, Density Matrix Renormalization Group (DMRG) for MPS and Simple Update (SU) with Tensor Renormalization Group (TRG) for PEPS are thoroughly analysed. Entanglement entropy area-law is usually obeyed by the ground state of local Hamiltonians, while generic highly excited states follow a volume-law (they span a finite part of the Hilbert space). Recently, a class of interacting system was shown to undergo a dynamical phase transition (Many Body Localization) where the entropy of every eigenstate follows an area-law. This transition is achieved when the system is highly disordered and the quantum many body state becomes localized similarly to the free particle Anderson localization. The area-law property makes highly excited eigenstates of Many Body Localized (MBL) systems efficiently represented by the MPS ansatz. We develop a highly optimized algorithm (eDMRG) which goes beyond ground state optimization and successfully target eigenstates in any part of the spectrum. This algorithm is used together with analytical calculations, based on local integrability of MBL systems, to identify the universal behaviour of the entanglement spectrum of highly excited eigenstates in MBL systems. In the second part we study interacting bosons in 2D optical lattices. Strongly interacting bosons are simulated using the Bose-Hubbard (BH) model when the interactions are strictly local and the Extended Bose-Hubbard (EBH) model, when additional dipolar interactions are present. The ground states of BH/EBH Hamiltonians in a hexagonal lattice are studied. The phase spectrum includes various insulating and critical phases which are studied in detail using infinite-PEPS ansatz. Additional results on entanglement entropy scaling with respect to the filling of the lattice are presented. Finally, a strongly interacting Harper-Hofstadter Hamiltonian is realized by combining synthetic fields with the BH model. The homogeneity of the system is then broken by a parabolic trapping potential, similar to the ones used in cold atom experiments. Using time-dependent Gutzwiller ansatz (GA) the expansion dynamics of the cloud in large square lattices are studied. In contrast to the expansion dynamics of the BH model, it is found that the synthetic fields generate a self-trapping effect. Using phenomenology and simulations the dynamics are studied in the hard-core and soft-core limit.