Plasma discharge is used to alter the surface potential in undoped GaAs. The results determine the origin and the magnitude of photomodulation in photoreflectance from undoped GaAs. We examine photoreflectance for 30–70-kV/cm surface electric fields in undoped GaAs grown on heavily doped n- or p-type GaAs underlayers, ${\mathit{UN}}^{+}$ and ${\mathit{UP}}^{+}$ samples, respectively. Photoreflectance for 2–500-μW/${\mathrm{cm}}^2$ laser modulation intensity was analyzed using separate electro-optic functions to determine the magnitude of the built-in electric field in the dark and the laser illuminated samples. Before plasma discharge, photomodulation of the built-in electric field in ${\mathit{UP}}^{+}$ samples was about three times that in ${\mathit{UN}}^{+}$ samples at 0.2 mW/${\mathrm{cm}}^2$. The large modulation in ${\mathit{UP}}^{+}$ samples appears to have come from surface modulation effects, possibly from poor pinning of the surface Fermi energy. A hydrogen plasma discharge lasting 4 s produced long term effects in ${\mathit{UP}}^{+}$ samples. It increased the magnitude of the electric field in ${\mathit{UP}}^{+}$ samples from ∼40 kV/cm to ∼60 kV/cm, at 0.2 mW/${\mathrm{cm}}^2$. The discharge more than doubled the amplitude of ${\mathit{UP}}^{+}$ photoreflectance signal and increased the number of Franz-Keldysh oscillations. After discharge, the photomodulation voltage for ${\mathit{UP}}^{+}$ samples was reduced from 0.32 to 0.14 V at 0.2 mW/${\mathrm{cm}}^2$, and became comparable with bulk photovoltage from (p-i-n) structures and ${\mathit{UN}}^{+}$ structures.

The contributions of soliton-antisoliton (SS¯) configurations to the linear and third-order nonlinear optical response are investigated for conjugated polymers having a degenerate ground state. We treat the direct photoproduction of charged solitons as a nonlinear Franck-Condon problem and obtain analytical expressions for the linear optical susceptibility ${\mathrm{\chi }}^{\mathrm{}\left(1\right)\mathrm{}}$(ω). With the help of the oscillator strength sum rule, we decompose the π-${\mathrm{\pi }}^{\mathrm{*}}$ oscillator strength into two parts; a contribution where the final states are charged SS¯ pairs and a contribution where the final states are free electron-hole pairs (as in the noninteracting rigid lattice). The linear optical coefficients calculated from ${\mathrm{\chi }}^{\mathrm{}\left(1\right)\mathrm{}}$(ω) are in general agreement with optical data obtained from trans-polyacetylene. The results imply that approxiamtely 25% of the integrated oscillator strength of trans-polyacetylene arises from the direct photoproduction of solitons.

A parallel treatment of the generalized third-order nonlinear optical susceptibility ${\mathrm{\chi }}^{\mathrm{}\left(3\right)\mathrm{}}$(${\mathrm{\omega }}_{\mathrm{\sigma }}$) is presented, demonstrating that, for any third-order process, contributions arising from neutral SS¯ pair configurations as intermediate states are one to two orders of magnitude larger than the corresponding rigid-lattice contribution. This mechanism for ${\mathrm{\chi }}^{\mathrm{}\left(3\right)\mathrm{}}$ is enabled by nonlinear zero-point motion which provides a finite Franck-Condon overlap between the ground and SS¯ excited state lattice wave functions. The large contribution to ${\mathrm{\chi }}^{\mathrm{}\left(3\right)\mathrm{}}$ from the SS¯ intermediate states results from the large transition dipole moment between the free electron-hole pair excited states of ${\mathit{B}}_{\mathit{u}}$ symmetry and the ${\mathit{A}}_{\mathit{g}}$ symmetric neutral SS¯ excited state. This enhanced transition dipole moment is a consequence of the large virtual shifts of oscillator strength associated with the localized SS¯ electron-lattice configuration. The third-harmonic conversion efficiency ${\mathrm{\chi }}^{\mathrm{}\left(3\right)\mathrm{}}$(3ω) is further enhanced by a condition unique to degenerate-ground-state systems, simultaneous two- and three-photon resonance.

The screened Coulomb Keldysh propagator is evaluated in the time-domain for ultra-short-pulse laser excitation of a semiconductor using the nonequilibrium Green-function technique. The buildup of screening in the time interval during and shortly after the pulse excitation is treated. A time-dependent plasmon pole approximation is derived and shown to give an excellent description of the dynamics of screening after a resonant femtosecond pulse.

We have studied the paramagnetic resonance signal (PRS, Δm=+1) of the 3${\mathit{d}}^5$ spin of ${\mathrm{Mn}}^{2+}$ in ${\mathrm{Cd}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Mn}}_{\mathit{x}}$Te epilayers and, less extensively, in a quantum-well structure by means of Raman spectroscopy in an external magnetic field of 6 T at low temperatures. The exciting laser or the scattered light were in resonance with electronic transitions (free exciton transitions) between the valence and the conduction band. The resonances have been observed by continuously tuning the wavelengths of the exciting dye laser in Voigt and Faraday backscattering geometries. For the off-resonance case the PRS is only expected in Voigt geometry with crossed polarizations of incident and scattered light (single resonance case, either ingoing or outcoming). Contrary to that, in resonance, the transition appears with large intensities as double resonances in the ‘‘forbidden’’ parallel polarizations and also in Faraday configuration. These results can be interpreted by considering a spin flip of a nucleus with I≥1/2 ${(}^{111}$Cd${,}^{113}$Cd${,}^{125}$Te${,}^{55}$Mn) simultaneously with the reorientation of the electronic spin of the 3d shell.

Quasiparticle bandstructure calculations for semiconductors treat the exchange-correlation self-energy operator within the GW approximation. In this paper, a detailed study of the energy dependence of the self-energy due to dynamical screening is presented. Spectral functions for near-band-gap excitations in silicon and diamond are calculated and the quasiparticle peak and satellite structures are discussed.

We have observed a series of lines in the absorption and luminescence spectra of ${\mathrm{BiI}}_3$ crystals which are strained by a bending stress. The magnetic-field effects on these lines were investigated in pulsed high magnetic fields up to 47 T. The magneto-optical spectra were well explained by the cationic exciton model. The origin of the lines was concluded to be an exciton state confined around a stacking disorder, which was formed by deformation. We discuss the quantum size effect of the excitons in the specific region of the stacking disorder.

A two-component evolution equation of reaction-diffusion type derived recently to describe experimental results on the self-organization of current-density filaments in silicon p-n-p-n diodes has been solved numerically. By using physically reasonable parameters and applying different boundary conditions, good qualitative agreement with experimental results is obtained. In particular, the following bifurcation scenario obtained by increasing the external driving voltage in the experiment can be reproduced in the calculations: stationary homogeneous current-density distribution leads to static current filament leads to rocking current filament leads to traveling current filament. Furthermore, it is shown that a period-doubling cascade of integral system variables, as, e.g., the device voltage, strongly correlates with the spatiotemporal filament motion which undergoes the same period-doubling route. Even quantitative agreement between numerical and experimental results could be achieved to some extent.

Scanning tunneling microscopy and low-energy electron diffraction have been used to study the surface reconstructions of Ag on Si(001) as a function of the Ag coverage up to one monolayer (ML), and annealing temperatures above 450 °C. Ag grows in the Stranski-Krastanov mode. An ordered 2×3 structure is observed in the first two-dimensional (2D) layer grown, and the 2D 2×3 layer is completed at 0.5 ML of Ag. Atomic-resolution images of this ordered surface structure are presented. Above 0.5-ML saturation coverage, an unusual growth transition from 2D to 3D is seen, which results from the subsurface accommodation of Ag accompanied by rearrangement of substrate Si.

We have investigated Ge/SiGe strained-barrier quantum-well structures using photoreflectance spectroscopy. On the basis of the Γ-point transition energies associated with the Ge quantum well, the band offset at the heterojunction between Ge and SiGe has been found to vary linearly with the germanium composition in the SiGe barrier layer. The conduction-band-offset ratio ${\mathit{Q}}_{\mathit{c}}$[=Δ${\mathit{E}}_{\mathit{c}}$/(Δ${\mathit{E}}_{\mathit{c}}$+Δ${\mathit{E}}_{\mathrm{vh}}$)] at the Γ point is estimated to be 0.68±0.08. From the intrinsic linewidth of the quantum-well-related transitions, interface roughness has been characterized in this system and is estimated to be ±1 monolayer in our samples.

The temperature dependence of the diagonal conductivity, ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}$(T), at integer and fractional quantum Hall effect (FQHE) minima was measured in a sample at various densities. We find ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}^{\mathit{c}}$, the 1/T→0 extrapolated value of ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}$(T) from Arrhenius plots, is different for different densities. While a reasonable (1/q${)}^2$ scaling of ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}^{\mathit{c}}$ at the filling factors ν=p/q is observed at lower densities, the scaling is not seen in the highest density data. We explain this loss of scaling by a breakdown of the assumption for a simple activated formula caused by a crossover between the extended state width, Γ, and T for the measurement. For kT>Γ, the scaling of the 1/T intercept is recovered by plotting ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}$(T)×T vs 1/T and fitting to ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}$(T)=(${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}^{\mathrm{*}\mathit{c}}$/T)exp(-ΔE/kT). We attribute the (1/q${)}^2$ scaling in ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}^{\mathit{c}}$ and ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}^{\mathrm{*}\mathit{c}}$ observed at each density to a (1/q${)}^2$ scaling in the T=0 conductivity. This supports the assertion of Clark et al. that the charge ${\mathit{e}}^{\mathrm{*}}$ of the quasiparticle excitation from the FQHE ground state at ν=p/q can be determined from ${\mathrm{\sigma }}_{\mathit{x}\mathit{x}}$(T) and the charge is ${\mathit{e}}^{\mathrm{*}}$=e/q.

Infrared laser interferometry was used to measure the temperature dependence, β(T), of the refractive index of Si, GaAs, and InP at λ=1.15, 1.31, 1.53, and 2.39 μm. Semiconductor wafer samples that had been polished on both sides were either heated or cooled while measuring the sample temperature and the transmitted or reflected intensity of an infrared laser beam. The changing optical path length within the material causes alternating constructive and destructive interference between reflections off the front and back surfaces of the wafer. By subtracting the contribution of thermal expansion, α(T), which is small and accurately known, β(T) was obtained. Representative values of β (293 K) at 1.53 μm are 5.15×${10}^{\mathrm{-}5}$ ${\mathrm{K}}^{\mathrm{-}1}$, 6.65×${10}^{\mathrm{-}5}$ ${\mathrm{K}}^{\mathrm{-}1}$, and 5.95×${10}^{\mathrm{-}5}$ ${\mathrm{K}}^{\mathrm{-}1}$ for Si, GaAs, and InP. Polynomial expressions are presented for Si, GaAs, and InP, yielding values of β(T) that are accurate to within ±5%. $beta (T)— increases with increasing temperature and decreases with increasing wavelength. There is a large resonance enhancement of β(T) in direct-gap semiconductors as the photon energy ${\mathit{E}}_{\mathit{h}\nu }$ approaches the band-gap energy ${\mathit{E}}_{\mathit{g}}$. Absolute values and temperature dependences of β calculated from published theory agree reasonably well with the measurements. The extreme accuracy in β needed for interferometric thermometry, however, cannot be met by these theoretical calculations, and so requires the experimental measurements.

We investigate the classical properties of electrons moving in a superposition of a uniform and periodically oscillating magnetic field. The most interesting dynamics occurs in the case where the uniform and periodic fields are of comparable order of magnitude and the periodic component originates from a planar arrangement of spins. For small energies almost the complete phase space is regular. With increasing energy the fraction of irregular orbits increases and eventually the phase space becomes completely chaotic. For higher energies we observe the appearance of a ballistic mode which allows the electrons to travel with high velocity through the magnetized spin lattice. This regular ballistic mode might be of relevance for transport processes in solid-state physics.

Comparison of experimentally determined optical spectra of ${\mathrm{Zn}}_{1\mathrm{-}\mathit{x}}$${\mathrm{Cd}}_{\mathit{x}}$Se/ZnSe quantum wells with theoretical results for fundamental optical absorption and exciton linewidths is used to determine valence-band offset and properties of exciton features as a function of well width and temperature. The heavy-hole valence-band offset is found to be approximately 90 meV for x=0.23. The exciton linewidths are analyzed in terms of Gaussian inhomogeneous broadening attributable to expected sample inhomogeneities involving well widths and alloy concentrations and homogeneous broadening associated with exciton-optical-phonon interactions. The linewidths are found to be smaller and larger than the bulk value for narrow (30 Å) and wide (200 Å) widths, respectively, in agreement with experimental results.

The Luttinger-Kohn canonical transformations are used to derive the effective band Hamiltonian for a quantum well. The transformations produce δ-function potentials in the effective-mass approximation. The boundary conditions on envelope functions, obtained from this effective Hamiltonian, differ from the BenDaniel-Duke ansatz. The effective Hamiltonian depends on the electronic coherence length. It also has a velocity-dependent term that is proportional to the sum of band lineup energies.

Thermalization of near-band-gap excited carriers in an intrinsic GaAs quantum well is studied using ensemble Monte Carlo simulation. Carrier-carrier interaction is treated like two-particle collisions due to the screened Coulomb interaction. The indistinguishability of two colliding electrons (holes) is included by taking into account the exchange electron-electron (hole-hole) scattering between the electrons (holes) of like spin. The model is modified to the case of spin-polarized carriers. We show that the thermalization of the spin-polarized carriers is several times slower than the thermalization of the spin-randomized ones when the carrier density exceeds ${10}^{11}$ ${\mathrm{cm}}^{\mathrm{-}2}$. The effect is due to the exchange scattering, which significantly weakens the direct Coulomb scattering in the dense carrier gas. For small densities the exchange becomes negligible. Possibilities to observe the exchange experimentally are discussed.

Transmission phenomena in quantum waveguide structures are studied by examining the transmission amplitude in the complex-energy plane. We find that, similar to double-barrier resonant tunneling, there are transmission poles in the complex-energy plane for quantum waveguide structures which contain quasibound states in attached t-stub resonators. In contrast to double-barrier resonant tunneling, however, we find that the quantum-wire networks also possess transmission zeros (antiresonances), which always occur on the real-energy axis. The existence of transmission zeros is a characteristic feature of a quantum waveguide system with attached resonators, but is absent for double-barrier resonant tunneling, which contains the resonant cavity as part of the transmission channel. We demonstrate that each quasibound state of the resonantly coupled quantum waveguide system leads to a zero-pole pair of the transmission amplitude in the complex-energy plane. The previously noted resonance-antiresonance behavior of the transmission probability, which leads to its sharp variation as a function of energy, can be understood in terms of these zero-pole pairs.

Electronic properties of three donors in three quantum dots are examined as we change the donor configurations (maintaining one donor per dot), vary the thickness of the barriers between the dots, impose an anisotropic confinement, and apply an external dc electric field. A scaled Kohn-Sham formalism is applied to treat the electron-electron interaction self-consistently within the local-density approximation. We compare the binding energies of the centered, molecular, and random donor configurations and find that the binding energy is the largest for the centered donor configuration due to the modulation by the boundary condition. We also find that dipole matrix elements can be enhanced if confinement or an external electric field increases the overlap between the initial and final states, and sustains their atomic characters of opposite parity. In addition, we demonstrate the importance of a self-consistent treatment of the electron-electron interaction in determining the ground-state binding energies and charge distributions of the donor states.

We perform a numerical self-consistent simulation of single-electron charging effects in an experimentally fabricated quantum-dot nanostructure. We use an iterative approach based on an extraction-orthogonalization method for solving the stationary Schrödinger equation. All relevant quantized regions in the structure are incorporated along with a numerical treatment of the Coulomb blockade and exchange correlation. Transport properties of the structure are evaluated in the adiabatic approximation using an interacting form of the Landauer formula. The theoretically calculated conductance data exhibit good agreement with experiment with respect to peak periodicity but show a strong discrepancy with regard to peak amplitude. This suggests that a high-order effect such as interface disorder may play a large role in determining electronic transmission through tunnel barriers.

Angle-resolved photoemission studies of the Si 2p and Mo 4p levels have been carried out on the (001) and (110) surfaces of a ${\mathrm{MoSi}}_2$ single crystal using synchrotron radiation. Two surface-shifted Si 2p components were observed on both surfaces. The shifts were determined to be -0.26 and -0.68 eV for the (001) surface and -0.29 and -0.72 eV for the (110) surface. The reactivity upon oxygen exposure was investigated and the (110) surface was found to be most reactive. No shifted Mo 4p component could be detected either in the clean or in the oxidized spectra from any of these surfaces.

The structural properties of pseudomorphic Si/Ge(001) heterostructures are investigated using a newly developed model in which the interface is viewed as a perturbation with respect to an infinite periodic system (the virtual crystal). The use of density-functional perturbation theory allows us to treat composition profiles of virtually any complexity at a negligible numerical cost and with an accuracy which is very close to that of state-of-the-art self-consistent calculations. We find that—at variance with recent measurements from x-ray photoelectron diffraction and ion scattering experiments—the macroscopic theory of elasticity accurately predicts the interplanar spacings of the heterostructure starting already at 2–3 planes away from the interface whereas a sizeable overstrain occurs in the region close to the interface. We suggest that the discrepancy between our findings and the experimental results could be due to atomic interdiffusion across the interface which would affect the composition of nominally pure Si (Ge) slabs, thus reducing the stress acting on the epilayer.