$In\sb{1-{\rm x}}Ga\sb{\rm x}As\sb{\rm y}P\sb{1-{\rm y}}$ epilayers lattice matched to InP and $In\sb{1-{\rm x}}Ga\sb{\rm x}As\sb{\rm y}P\sb{1-{\rm y}}/InP$ Multiple Quantum Wells (MQWs) grown by Chemical-Beam Epitaxy (CBE) are being studied systematically using the Photovoltaic (PV) effect. At first, the Schottky barriers on the interfaces (metal-semiconductor, metal-insulator-semiconductor) are determined as an important factor for the electrical and optical properties of the samples. Samples with identical Schottky contact deposition but with an insulating layer on the front surface, have shown much smaller leakage current and yield enhanced barrier heights. The photovoltaic signal in the temperature interval 4-300K has maximum amplitude at about 150-180K for the MQW samples and at about 190K for the epilayer. An applied electric field changes the integrated intensity and spectrally shifts the allowed and forbidden transitions observed in bias dependent PV spectra of various InGaAsP/InP MQWs. The combined effect of two external factors, the thermal ionization and the electric field on the shape and magnitude of the 11H exciton peak, are discussed in terms of exciton binding energy and field ionization. The optically induced changes and energy shifting of the 11H/ exciton peak are observed, when excitation dependent double beam experiments are conducted on the $In\sb{0.72}Ga\sb{0.28}As\sb{0.68}P\sb{0.32}/InP$ MQWs. The photomodulation of the internal fields through carrier transport results in observing effective nonlinearities at milliwatt power levels. The experimentally measured transition energies for the MQWs show good agreement with the envelope wave function calculations. The observed Schottky barrier heights and band gap energies are consistent with the interpolation scheme estimations. (Abstract shortened by UMI.)

This thesis describes a study of reentrant metal-insulator transitions observed in the inversion layer of extremely high mobility Si-MOSFETs. Magneto-transport measurements were carried out in the temperature range 20mK-4.2 K in a $\sp3$He/$\sp4$He dilution refrigerator which was surrounded by a 15 Tesla superconducting magnet. Below a melting temperature $(T\sb{M}\sim500$ mK) and a critical electron density $(n\sb{s}\sim9\times10\sp $ cm$\sp{-2}),$ the Shubnikov-de Haas oscillations in the diagonal resistivity enormous maximum values at the half filled Landau levels while maintaining deep minima corresponding to the quantum Hall effect at filled Landau levels. At even lower electron densities the insulating regions began to spread and eventually a metal-insulator transition could be induced at zero magnetic field. The measurement of extremely large resistances in the milliKelvin temperature range required the use of very low currents (typically in the $10\sp{-12}$ A range) and in certain measurements minimizing the noise was also a consideration. The improvements achieved in these areas through the use of shielding, optical decouplers and battery operated instruments are described. The transport signatures of the insulating state are considered in terms of two basic mechanisms: single particle localization with transport by variable range hopping and the formation of a collective state such as a pinned Wigner crystal or electron solid with transport through the motion of bound dislocation pairs. The experimental data is best described by the latter model. Thus the two dimensional electron system in these high mobility Si-MOSFETs provides the first and only experimental demonstration to date of the formation of an electron solid at zero and low magnetic fields in the quantum limit where the Coulomb interaction energy dominates over the zero point oscillation energy. The role of disorder in favouring either single particle localization or the formation of a Wigner crystal is explored by considering a variety of samples with a wide range of mobilities and by varying the ratio of the carrier density (controlled by the applied gate voltage) to the impurity density (fixed during sample growth). A phase diagram showing the boundaries between the two dimensional electron gas, the Wigner solid, and the single particle localization induced insulator is established in terms of carrier density and sample mobility.

Ideal annite has structural formula $\rm\{K\sp{+}\}\lbrack Fe\sbsp{3}{2+}\rbrack \langle Al\sp{3+}Si\sbsp{3}{4+}\rangle O\sbsp {2}(OH\sp{-})\sb2$ where $\{\ \},$ () and $\langle\rangle$ represent the interlayer, octahedral, and tetrahedral sites, respectively. Simple crystal chemical considerations suggest that ideal annite should exist as do other end-members having Fe replaced by Mg, Ni, Co, etc. However, due to structural constraints involving long range layer misfit and short range bond matching, ideal annite is never synthesized. Instead, the constraints are satisfied by both non-destructive partial oxydation (via H-loss) of the Fe$\sp{2+}$ and $\rm Fe\sp{3+}-Al\sp{3+}$ exchange. Consequently, real annite must actually contain small fractions of both (Fe$\sp{3+}\rbrack$ and $\rm\langle Fe\sp{3+}\rangle$ with an associated amount of hydrogen deficiency. The observed stoichiometry of real annite (assuming no octahedral vacancies) is: $$\eqalign{\{K\sp{+}\}\lbrack Fe\sbsp{1-x-y}{2+}Fe\sbsp{x}{3+}Al\sbsp{y}{3+}\rbrack\sb3 \langle & Al\sbsp{1-3y}{3+}Fe\sbsp{3y}{3+}Si\sbsp{3}{4+} \rangle\cr & O\sbsp{10+3x+3y}{2-} (OH\sp{-})\sb{2-3x-3y}\cr}$$where $x={{\lbrack Fe\sp{3+}\rbrack}\over{Fe\sb{tot}}}$ and $y={{\langle Fe\sp{3+}\rangle}\over {Fe\sb{tot}}}.$ We measured Fe$\sp{3+}$ (x and y) and Fe$\sp{2+}$ site populations in nine annite samples synthesized using the C-CH$\sb4$ and Ni-NiO buffers, and various equilibration temperatures. We obtained site population measurements with precisions of 0.2-1% of Fe$\sb{tot},$ whereas the generally accepted limit is 1-5% of Fe$\sb{tot}.$ By using temperatures over the entire stability range for annite, we have traced the evolution of $\rm\lbrack Fe\sp{3+}\rbrack$ and $\langle\rm Fe\sp{3+}\rangle$ site populations and have obtained an experimental lower bound on ${{Fe\sp{3+}}\over{Fe\sb{tot}}}.$ Temperature is seen to have a systematic effect that is understood in terms of differential thermal expansion between the octahedral and tetrahedral layers. To explain this effect, we propose a structural misfit model for annite mica that gives definite predictions on the $\rm\lbrack Fe\sp{3+}\rbrack$ and $\langle Fe\sp{3+}\rangle$ site populations and includes a cation size dependant octahedral flattening and the effect of octahedral/tetrahedral sheet differential thermal expansion. (Abstract shortened by UMI.)

Infrared radiation may be absorbed by a vibrational mode of an adsorption system at resonance. The homogeneous IR absorption line shape of a vibrational mode of an adsorbate is determined by its anharmonic interactions with other localized and delocalized modes of the adsorbate and substrate. These anharmonic interactions include both the adsorption potential between an adsorbed molecule or atom and the surface atoms, and the anharmonic couplings among the substrate atoms themselves at the adsorption site. Temperature dependent effects of substrate phonons and local mode self interactions on the line broadening and frequency shift are studied in a finite temperature field theory for both the internal stretching mode and admolecule-substrate stretching mode of a diatomic admolecule at an ontop site, and adatom-substrate stretching mode of an adatom absorbed at a bridge site. Modified phonon emission and elastic phonon scattering processes are described in a unified formalism. For the adsorption of an admolecule or adatom at an ontop site, a reduction factor is found which reduces all substrate phonon contributions induced by the anharmonicity of the adsorption potential on the line shift and line width of the low frequency admolecule-substrate stretching or adatom-substrate stretching mode. For the adsorption of an admolecule or adatom at a bridge site, reduction factors are found which reduce only some of the substrate phonon contributions induced by the same anharmonicity on the line shift and line width of the low frequency adatom-substrate stretching mode. A bulk Debye model and a semi-infinite elastic continuum model for the substrate are employed to describe the harmonic substrate behavior. The theory is applied to the extensively studied adsorption systems CO/Pt(111) and O/Cu(110). Qualitative agreement with experimental data is found in considering the partial contributions and simplicity of the model.

Excitonic transport measurements in ultra-pure semiconducting Cu$\sb2$O crystals were performed over a wide range of exciton densities at liquid Helium temperatures. The excitons were created with a high power pulsed Nd:YAG laser in the Cu$\sb2$O over a 3 mm$\sp2$ area and in a thin layer a few micrometers thick. The excitons then migrated over several millimeters through the crystal and were measured at the opposite sample face in time resolve using the exciton mediated photovoltaic effect. At low exciton density the transport is diffusive as expected and the results are well fitted by a diffusion equation. However at high exciton densities the transport becomes ballistic and a reduction in the size of the exciton distribution becomes apparent. Furthermore, the density at which there is a crossover from diffusive to ballistic behaviour follows a $T\sp{3/2}$ temperature dependence and is of the same order of magnitude as the predicted critical density for Bose-Einstein condensation. Pump-probe experiments with two laser pulses separated by a variable time delay allowed the formation of two exciton packets separated by the distance traveled by the first packet during the time delay. The experiments confirmed the attractive nature of a Bose condensate with other condensates and normal Bose particles, and revealed the formation of a soliton condensate. We conclude that the results presented in this work are consistent with the condensation and superfluid transport of excitons in Cu$\sb2$O.

Formulae for the magnetic moment $\vec\mu$ of anisotropic platelets of high $T\sb{c}$ superconductors developed by Gyorgy et al and Peterson are frequently exploited by these and other researchers to estimate $j\sbsp{c}{c}$ and $j\sbsp{c}{ab}$, the critical current densities along the c axis and in the ab plane taken to be independent of the magnetic flux density B. These formulae were derived using the basic definition, $\ = ( -\ \mu\sb0H\sb{a})/\mu\sb0$ and ignoring end effects, (i.e. any demagnetizing fields), hence implied that the aspect ratio along the magnetizing field $H\sb{a}$ is large. This approximation is inappropriate for platelets penetrated by $H\sb{a}$. We develop these formulae using the alternative basic definition of a magnetic moment, $\vec\mu = 1/2\int(\vec{R}\times \vec{j})dV$. Now however, for the approach to be valid, $\vec{j} = j\sbsp{c}{c}$ or $\vec{j} = j\sbsp{c}{ab}$ must be independent of B (Bean approximation) and fill the entire volume of the specimen (i.e. a saturated critical state must be established). We show that these formulae are correct under these restrictions regardless of the configuration of $\vec{B}(x,y,z)$ and the neglect of end effects and attendant demagnetizing fields. Pursuing this framework and the latter definition we develop formulae for $\vec\mu$ for isotropic parallelepipeds of various aspect ratios as a function of their inclination $\theta$ with respect to the magnetizing field $\vec H\sb{a}$. We maintain throughout the critical assumption that the induced persistent currents circulate transverse to $\vec H\sb{a}$. The graphs of computations with these formulae are useful in identifying the role of geometry on the magnitude of $\vec\mu$. We also envisage two simple but basic regimes of anisotropy of the critical current densities. (Abstract shortened by UMI.)

Using isothermal-isochore and isoenthalpic-isobaric molecular dynamics simulations, with either periodic boundary conditions or free boundary conditions, we explore in detail the vacancy annealing mechanisms in monolayers with or without substrate. A new mechanism, named dislocation mediated annealing (DMA), is observed. In this mechanism, vacancies condense rapidly into dislocation dipoles, with an associated shear modulus collapse, and anneal out of the system. External pressure, mobility of the vacancies and increased range of the interactions all favour DMA. We find that the other mechanism observed, annealing of the vacancies by the formation of voids, is a slower process, by at least an order of magnitude. We clarify the concept of reference configuration and volume in the calculation of the elastic constants in "equilibrium" fluctuation methods and give the general expressions. We derive the correct fluctuation formulae for elastic constants in uniform dilation systems. We show that the traditional thermodynamic potentials cannot be used in stability problems for an anisotropically stressed system. We derive general expressions for the mechanical stability criteria of a stressed material. We find that for a system under isotropic initial stress, the elastic stiffness coefficients which govern stress-strain relations can be used as stability criteria. However, for a system under anisotropic initial stress, stability criteria are different from either elastic constants or elastic stiffness coefficients. We show that the stability conditions in the constant pressure ensemble are stronger than in the constant volume ensemble. Exact solutions for perfect systems at zero temperature with three types of interactions of different range are found to be consistent with those obtained from long-wavelength expansion and computer simulations. We study the rupture of models of solid membranes for several interparticle interactions. We show that rupture at zero temperature occurs at the mechanical instability point. We verify that in general rupture at finite temperature takes place before the mechanical instability point. Two regimes are observed in the variation of the critical rupture pressure ($P\sb{c}$) with temperature (T) for both one and two dimensional (2D) systems; $P\sb{c}$ drops very fast at low T, but relatively slowly at high T. In ideal 2D membranes under isotropic tension, a linear decrease of $P\sb{c}$ was observed at high T. The kinetics in non defected systems, involve, just before rupture, the creation of gliding dislocation dipoles. In-grown vacancies reduce the influence of dislocations and can lead to direct cavitation.

Four different aspects of the properties of dislocations in monolayer and semiconductors have been investigated: (i) Using atomic relaxation techniques, dislocation dipoles of various sizes and orientations have been studied for monolayers with the Lennard-Jones potential (LJP) and the nearest-neighbour piecewise linear force (PLF) interactions. In the WP system the lower energy vacancy dipoles have over a wide range of angles an energy which is mainly a function of the vacancy content of the dipole. There is a competition between the elastic forces and the topological constraints which favour a five-fold coordinate vacancy (FCV) at the centre of each core. For the short range PLF system the lattice usually compresses upon the introduction of a dislocation, a consequence of the soft core of the interaction potential, and interstitial dipoles are lower in energy. For the long range LJP system the dislocations are mobile whereas for the PLF system they are pinned. The relevance of these results to existing theories of melting are discussed. (ii) Using generalized stacking-fault (GSF) energies obtained from first-principles density-functional calculations, a zero-temperature model for dislocations in silicon is constructed within the framework of a Peierls-Nabarro (PN) model. Core widths, core energies, PN pinning energies, and stresses are calculated for various possible perfect and imperfect dislocations. Both shuffle and glide sets are considered. 90$\sp\circ$ partials are shown to have a lower Peierls stress (PS) than 30$\sp\circ$ partials in accord with experiment. (iii) We have also studied by atomic relaxation techniques the properties of dislocations in silicon, modelled by the empirical potential of Stillinger and Weber. In order to compare with the preceding calculation no reconstruction is allowed. We find no evidence of dissociation in the shuffle dislocations. Within this model shuffle dislocations glide along their slipping planes. On the other hand, glide sets are shown to glide only in dissociated form. The dislocation displacement fields are essentially planar. The PS is found to be isotropic within the (111) plane. In other words the minimum stress at 0K required to move the dislocation in any direction with in that plane has the same projection unto the Burgers vector, the PS of the dislocation. Our PS are in good agreement with those from (ii). (iv) Using a simple two dimensional UP model, relaxation mechanisms of the epitaxial strain layers (ESL) have been simulated for various misfits and layer thickness. In this model, the relationship of two competing relaxation mechanisms is found. At small misfit, strain is released by nucleating misfit dislocations from the edges of system. This process is more favourable for the thicker layer. At large misfit, stress is relaxed through surface instability, allowing easy generation of misfit dislocations from the surface. Those results are qualitatively in agreement with experiments.

The magnetic properties of fcc Fe-Ni alloys are studied by Mossbauer spectroscopy and Monte Carlo (MC) simulations. Both macroscopic (magnetization, paraprocess susceptibility, Curie points, etc.) and microscopic properties (hyperfine fields) are used to test simple local moment models under various assumptions. A non-linear composition dependence of the average hyperfine field is observed by Fe-57 Mossbauer spectroscopy. A microscopic vector hyperfine field model is proposed and used to model the measured average hyperfine fields and hyperfine field distributions (HFDs) in the collinear ferromagnetic Fe-Ni alloys (y $\le$ 0.45 in Fe$\sb{y}$Ni$\sb{1-y}).$ Modeling the liquid helium temperature average hyperfine fields and HFDs resolves the coupling parameters in the proposed hyperfine field model:$$\langle\vec H\sb{k}\rangle\sb{T}=A\langle\vec\mu\sb{k}\rangle\sb{T}+ B\sum\sb{j}\langle\vec\mu \sb{j}\rangle\sb{T}.$$To the extent that chemical short range order can be neglected in our rapidly quenched samples, the coupling parameters are $\rm A=A\sb0+A\sb1y\ (A\sb0=89$ kOe/$\mu\sb{B},$ A$\sb1={-}20$ kOe/$\mu\sb{B})$ and B = B$\rm\sb0=B\sb1y\ (B\sb0=4.4$ kOe/$\mu\sb{B},$ B$\sb1=3.2$ kOe/$\mu\sb{B}).$ MC simulations show the success and the limits of a simple local moment model, in characterizing the bulk magnetic properties of Fe-Ni. A new approach for simulating HFDs is developed. It combines MC simulation for the spin structure and the above phenomenological hyperfine field model for the site-specific hyperfine field values. Using this method, we calculated spin structures and HFDs in Fe-Ni alloys at different compositions and temperatures. Finally, interplay between the magnetic and the atomic ordering phenomena is studied in FeNi$\sb3,$ FeNi and Fe$\sb3$Ni, by considering the magnetic and chemical interactions simultaneously using MC simulations. Serveral new features that are not predicted by mean-field theory or MC simulations with chemical interactions only arise: (1) chemical order can be induced where using chemical interactions only leads to the prediction of no chemical order (2) chemical segregation can be induced where using chemical interactions only leads to the prediction of no chemical segregation, (3) FeNi$\sb3$ and Fe$\sb3$Ni are found to have significantly different chemical ordering temperatures where chemical interactions only lead to equal ordering temperatures, (4) chemical ordering temperatures are significantly shifted from their chemical interactions only values, even when the chemical ordering temperature is larger than the magnetic ordering temperature, (5) abrupt steps can occur in the spontaneous magnetization at the chemical ordering temperature, when the latter is smaller than the magnetic ordering temperature, and (6) nonlinear relations arise between the chemical ordering temperature and the chemical exchange parameter U $\equiv$ 2U$\sb{FeNi}-{\rm U}\sb{FeFe}-{\rm U}\sb{NiNi},$ where the U$\sb{ij}$ are the near-neighbour pair-wise chemical bonds.

We have studied samples of CuO and Cu$\sb2$O to pressures of 60 GPa and 35 GPa, respectively, using energy dispersive X-ray diffraction. Samples were studied using a gasketed diamond anvil cell, energy dispersive X-ray diffraction at the Cornell High Energy Synchrotron Source, with a Mo or Al$\sb2$O$\sb3$:Cr$\sp{3+}$ pressure gauge. In the case of Cu$\sb2$O, the two phase transitions, (cubic to hexagonal to another hexagonal) as expected were seen. In CuO no definite phase transition was seen, although indications are given for CuO at 60 GPa converting to a distorted rocksalt-type structure from the initial monoclinic structure. The volume data obtained from the pressure experiments on CuO were best fitted with the third-order Birch-Murnaghan equation of state with B$\sb0$ = 110 GPa and B$\sb0\sp\prime$ = 5.5.