Dislocations have been imaged in bulk specimens of Si using a FEG SEM operating at 30keV and 100keV, without using an energy filter, but by image processing of the back scattered electron signal collected by a highly efficient detector. The dislocation contrast is greater at 30 keV than at 10OkeV. Ilowever, the depth to which useful information may be obtained is greater at 1O0keV (˜210nm) than at 3OkeV (˜95nm). The final depth to which dislocations can be imaged is strongly dependent on image processing conditions.

Basing our calculations on a realistic N-body interatomic potential for Zr, we study the vacancy migration mechanism and determine the related diffusion coefficient in the bcc phase. The form of the potential energy along the nearest-neighborjump migration path is single-peaked. The vacancy jump rate determined by molecular dynamics simulations has a perfectly Arrhenian behavior and its activation energy is very close to the static value of the vacancy migration energy, both being very low (≈ 0.3 eV) . The diffusion coefficient is in very satisfactory agreement with experiments.

Analysis of the nonlinear dynamical response of a rod-like nematic liquid crystalline polymer, between parallel untreate plates, in a magnetic field is presented. Numerical solutions to the torque balance equations showthat when a uniformly aligned polymer is acted by a magnetic field normal to the initial polymer orientation an array of inversion walls develops. Stability analysis shows that the inversion walls are unstable to localized perturbations. The nonlinear resolution ofthe instability leads to the nucleation of an array of disclination line pairs of strength ±1/2. A perturbation analysis is used to obtain the evolution of the polymer structure towards equilibrium

A model predicting the number of prismatic loops dislocation punched at the ends of a cylindrical fiber by thermal mismatch stresses is presented and compared to another based on a mismatching ellipsoid. The longitudinal stress in the fiber and the interfacial shear stress are derived by adapting a shear-lag model to the plastic portion of the interface. In certain cases, the central part of the fiber is strained by plastic and elastic interfacial shear until it exhibits no mismatch with the matrix. This leads to a critical fiber length above which the number of punched loops is constant.

Pulsed nuclear magnetic resonance proved to be a complementary new technique for the study of moving dislocations in b.c.c. metals. From the motion induced part of the spin-lattice relaxation rate the mean jump distance of mobile dislocationshas been measured in Vanadium as a function of temperature. The NMR experiments are combined with transmission electron microscopic investigations to reveal the static structure of defects in the samples. The NMR experiments show that the mean jump distance is nearly constant below 230 K whereas it decreases substantially above 230 K to 300 K indicating a transition that marks two different mechanisms. NMR observations in combination with TEM support the physical picture that above that transition temperature dislocation segments are stopped between localized obstacles whereas below Tc the latticefriction controls the plastic behaviour.

Two different sizes (0.4 and 0.1 μm) of rubber particles (polybutadiene) are usually used to toughen the glassy SAN (styrene acrylonitrile) matrix in ABS copolymers. Experiments on a batch of specimens without the large particles confirmedour previous contention that cavitation of large particles within a zone of cardioid shape is an important part of crack tip plasticity which protects the crack tip from the applied stress field. Large rubber particles require a less strong applied stress field than small rubber particles to initiate crazes so that the large rubber particles help avoid brittlefracture which occurs without crazes.

We have studied a novel surface reaction in which acetylene (C2H2) cyclizes with sulfur atoms on the Pd(111) surface to form thiophene (C4H4 S). The reaction occurs at defects in the (√7 ×√7)R19° sulfur overlayer that is formed on this surface. The surfaces were prepared by decomposing H2S on the clean Pd(111) surface and annealing to temperatures in the range 600K to 1100K to produce defects i the sulfur overlayer. The reaction occurs to greatest extent on surfaces annealed to high temperatures and only on surfaces with coverages approaching saturation. The defects have been characterized by adsorption of CO using both vibrational spectroscopyand measurementsof desorption kinetics. On the sulfided surface CO is bound in a linear configuration to a single Pd atom. At defects in the sulfur overlayer CO is bound in a bridging site between two atoms, as on the clean surface.

The intercalation of fluorine into graphite introduces defects into the highly crystalline pristine fibers. These defectsare studied using temperature-dependent resistivity and magnetoresistance measurements. A logarithmic increase in resistivity at low temperature is observed, whereas the high temperature behavior is metallic. At weak magnetic fields and low temperatures, a negative magnetoresistance is observed, which becomes positive at high fields. These effects are explainedusing the two theories of weak localization and hole-hole interaction. In the light of TEM pictures of the microstructure of the fluorinated fibers, the origin of the defects in the intercalated fibers is discussed.

Low-temperature electrical conductivity and Raman scattering are studied as characterization tools for activated carbon fibers, which have a high density of defects and a huge specific surface area. The transport mechanism at low temperature is governed by variablerange hopping, as in other strongly disordered systems. From the Raman spectra obtained, we deduce that the long phenolic fibers are more disordered than the acrylic fibers and that increased specific surface area corresponds to increased disorder. The average in-plane microcrystallite size is about 20–30 Å.

A highly disordered carbon material, activated carbon fibers, is investigated through bulk conductivity and photoconductivity measurements. This material has a high density of defect states introduced by an aqtivation process that leads to a huge specific surface area of up to 2000m2/g. The conductivity increases by a factor of 4–10 with increasing temperaturefrom 30K to 290K. In contrast, the photoconductivity decreases by a factor of three with increasing temperature. The relaxation time of the photoconductivity is rather long(on the order 10−1 sec), indicating that the recombination process proceeds through localized states. flopping processes are used to interpret the transport properties of this material.

We report Raman scattering measurements on fluorine-intercalated graphite fibers (CxF) for stoichiometries x =7.8, 4.5 and 2.9. Lorentzian fits to our Raman lineshapes indicate the presence of two lines around 1600 cm−1 and a broad line around 1355 cm−1 . The 1355 cm−1 line is the disorder-induced graphite line and the ratio of the integrated intensity of this line to that of the 1600 cm−1 doublet (R) provides a measure of the intercalationinduced disorder in the CjF fibers. Both TEM and Raman studies indicate the presence of unintercalated graphitic regions in the CxF fiber. The inverse relation between the average crystal planar domain size La, and the Raman intensity ratio R yields La. ,˜52Å for C7.8F fibers and La. , ˜40Å for C4.5,F and C2.9F fibers.

The recently formulated theory of vacancy formation in CsCl-type ordered alloys has been applied to β-NiGa and β-NiAl. The observed Nisublattice vacancy concentrations in these compounds could be well described by the theory. For the first time it is also shown that structural vacancies most likely exist in Ga-rich NiGa and Al-rich NiAl at absolute zero temperature.

(200), (111) AND (311) neutron diffraction peaks were measured on polycrystalline Cu samples with varying degrees of deformation by compression: O%. 0.8%. 1.2%,, 1.5% and 2%. The first two reflections were measured with wavelength 2.358Å and the third one with wavelength 1.55Å.

Integrated intensities vary more less monotonically and they tend to reach the kinematical value. This limit was estimated by extrapolation and used to evaluate extinction. To obtain dislocation densities a simple model was proposed which divides the crystal in regions of perfect periodicity and regions perturbed by dislocations, and evaluates the integratedintensity from perfect regions by the general theory of X-Ray diffraction of Zachariasen (1967).

Dislocation densities result between 1.2×1O7 cm−2 for the sample without deformation, and approximately 5.0×1O7 cm−2 for the 2%. deformed sample. These results are in the expected range.

I summarize a large body of experimental and theoretical work, especially in Sidoped GaAs and AlxGal−xAs, which has led to our present understanding of the DX center. There is good evidence that the DX center is just the simple donor, but each donor atom can exist in either of two distinct lattice configurations, each with its own spectrum of bound electronic states. At present the model which appears to best explain this unexpected complexity is that of Chadi and Chang, which predicts a large bond-breaking distortion of the lattice, accompanied by capture of two electrons by the donor. I argue that the best evidence to date for the predicted distortion is provided by observations of alloyperturbations of the DX level. Furthermore, there is now very convincing evidence that the DX level is a two-electron state. I briefly summarize what is known about the bound state spectra of the substitutional and relaxed (DX) configurations, and then discuss the very interesting question of how the donor captures two electrons from the conduction band to theDX level. There is good, although indirect, evidence that an excited one-electron state acts as an intermediate in the thermal emission and capture process. A different excited state appears to 15e involved in photoemission. Although much of this complexity was unforeseen by Chadi and Chang, it nevertheless seems to be consistent with their model. However, I point out a number of issues which still require experimental or theoretical resolution.

The electrical properties and preferred lattice sites of Cu in GaAs were studied combining electrical and optical measurements with Particle Induced X-ray Emission (PIXE) and Channeling. For electronic characterization, Deep Level Transient Spectroscopy (DLTS), Hall Effect measurements, and Photoluminescence (PL) were used. From this comprehensive study it was determined that Cu introduces two levels in the bandgap, that the concentration of electrically active copper is considerably smaller than the total copper concentration, and that most of the Cu in GaAs is not of purely substitutional character.

Copper was precipitated on extrinsic near-surface dislocations in {100} Si wafers. Several types of precipitates were heterogeneously nucleated on low index planar arrays. Large crystalline precipitates with visible strain fields proved to bea silicide. Large weakly diffracting precipitates without strain fields proved to be voids. Small, thin, partially crystalline particles of unknown composition also precipitated on the arrays. Point defects played an important role in the precipitation reaction.

The predominance of Si-H bonding and the origin of {111} platelets in hydrogenated Si remain important unsolvedproblems in the study of H in Si.Recent theoretical and experimental results indicate that H predominately enters the Si network in pairs. A promising diatomic H configuration consists of a bond centered H closely associated with an antibonding centered H. In this work, we show that adjacent diatomic H pairs have a binding energy of 0.2 eV/2H. The binding originates from relaxation of strained Si-Si backbonds. Further clustering of the H pairs eliminates all strained bonds, forming a hydrogenated platelet oriented along the {111} plane. The binding energy of 3.95 eV/2H for the platelet is 0.15 eV lower than that for interstitial H2 molecules in c-Si. Lattice expansion makes the platelets energetically more competitivewith the lowest energy Si-H bonding confi gration at hydrogenated Si (111) surfaces. These higher level complexes explainthe formation of platelets, Raman spectra, and absence of gap states in hydrogenated c- Si as well as the clustered phaseseen in NMR and of H evolution and diffusion in hydrogenated amorphous Si.

The two most outstanding features observed for dopants in hydrogenated amorphous silicon (a-Si:H) - a shift in the Fermi level accompanied by an increase in the defect density and an absence of degenerate doping - have previously been postulated to stem from the formation of substitutional dopant-dangling bond complexes. Using firstprinciples self-consistent pseudopotential calculations in conjunction with a supercell model for the amorphous network and the ability of network relaxation from the first-principles results, we have studied the electronic and structural properties of substitutional fourfoldcoordinated phosphorus and boron at the second neighbor position to a dangling bond defect. We demonstrate that such impurity-defect complexes can account for the general features observed experimentally in doped a-Si:H.

Total yield photoelectron spectroscopy has been used to study the electronic structure change of UHV evaporated a-Ge subjected to posthydrogenation and various annealing cycles. We identify in R.T. hydrogenated a-Ge:H a new hydrogen induced defect at about Ev + 0.45eV, which can be healed upon 300°C annealing. This new defect accounts for the defect density gradient of hydrogenated amorphous semiconductors, spanning the range from ∼ 1018 cm−3 at the growing surface to 1018−1015 cm−3 in the bulk, depending on growth condition and time. The origin of this new defect is discussed.

The precipitation of copper and nickel at grain boundaries in cast polycrystalline silicon is investigated. The metals are diffused into the specimens from a surface source between 800 – 1000 °C and the precipitation after cooling is studied by TEM. Copper precipitates in form of colonies containing hundreds of particles with a size between 5–6 nm. In the grain boundary they nucleate preferentially at dislocations and steps. The distribution and size of the precipitates depend on the cooling rate after the diffusion. Nickel forms only few large (micrometer size) plate-like or three dimensional precipitates at and near grain boundaries. The main features of the results and the differences between the two elements are explained under the assumption that the precipitation requires the transport of native point defects.