Table of contents

Using X-ray methods capable of detecting changes in lattice parameter of a few parts per million, the residual damage in proton-bombarded gallium arsenide has been detected. The lattice strain in the bombarded region has been determined at approximately 200 ppm for a dose of 2*10¹⁵ cm^-2 at 600 keV. The lattice dilation per defect, following the bombardment, has also been estimated at 10-20%.

Voltage-switching in evaporated weak-link junctions was found to take two forms, that is, either a snapback or a jump in voltage depending on the arrangement of the potential probes. The present observation is consistent with a heating model previously proposed by Iwanyshyn et al. (1970).

The coexistence of the ferrimagnetic and the antiferromagnetic phases has been observed in the DyFeO₃ orthoferrite. The observations were made in a polarization microscope by means of the Faraday effect. Several features of the spin reorientation which shed some light on the mechanism of the transition, have been observed.

A new approach to deriving effective dielectric constants for two-component systems is developed. For metal-dielectric mixtures three classes of problem are discussed: two where one constituent exists as particles embedded in the other, and the third where both are in particle form.

The dependence of the apparent zero-gap intercept, obtained by extrapolation of the direct breakdown voltage versus gap length characteristics, on the presence of strongly insulating oxide films on the electrode surfaces has been studied. The intercepts were obtained from tests carried out using etched aluminium, anodized aluminium and stainless steel electrodes on both degassed and oxygen-saturated mineral oil. The results indicate that although a large intercept is associated with the presence of a highly insulating barrier on one of the electrodes, the presence of such a barrier does not necessarily lead to such an intercept. The results are interpreted by taking into account the space charges present at the electrode surfaces prior to extrapolation to zero gap.

The thermoluminescence glow curve of CaF₂:Dy TLD-200 extends from room temperature up to about 450 degrees C. Dy³⁺ is the main emitting centre for all the glow peaks. However, in addition to Dy³⁺ emission, the glow peaks above 325 degrees C show four extrabands in the region 250-600 nm. Three of these are attributed to Tb³⁺ and the fourth is possibly associated with the host lattice. The observations lead to the conclusion that all of the luminescent centres are not distributed randomly with respect to the traps. Some of these exist in the form of aggregates of traps and luminescent centres. The thermoluminescence resulting from the localized recombination within the aggregate centres leads to distinct spectral changes from one glow peak to another.

Methods of minimizing the damage are discussed and direct comparisons made between implantation and diffusion techniques in terms of defects in the final devices and the electrical performance of the devices. Defects are produced in the silicon lattice during implantation but they are annealed to form secondary defects even at room temperature. The annealing can be at a low temperature (<900 degrees C) when migration of defects in silicon is generally small, or at a high temperature when they can grow well beyond the implanted region. The defect structures can be complicated by impurity atoms knocked into the silicon from surface layers by the implantation. Defects can also be produced within layers on top of the silicon and these can be very important in device fabrication. In addition to affecting the electrical properties of the final device, defects produced during fabrication may influence the chemical properties of the materials. The use of these properties to improve devices is discussed as well as the degradation they can cause.

Experimental results on LEDs show that a number of light-degrading mechanisms can operate which probably include migration of impurities or point defects. Dislocations when present in high densities also give rise to degradation in LEDs. Extrapolated lives to half intensity of between 10⁵ and 10⁷ h have been commonly projected for these devices, and this rather slow rate of degradation appears to have been achieved by minimizing impurity contamination, reducing strain and minimizing the dislocation density in the carrier recombination region. Relatively long lives have also been predicted for high-radiance, high-current-density lamps from thermally accelerated life-tests. Dark-spot and dark-line defects have been observed. In one case, the dark structure has been shown to result from the formation of tangles of dislocations generated during fabrication. High-current-density lamps which are free from dark structure also have projected lives of between 10⁵ and 10⁷ h. Generally, laser diodes can now be made which have lives in excess of 10⁴ h. Details of the dark structure resulting from dislocations is reviewed and consideration is given to the remaining causes of degradation in lasers.

The present understanding of the operation of green-emitting GaP LEDs is reviewed. All existing visible LED devices which are made in III-V compound semiconductors are inefficient. In green-luminescent GaP this inefficiency is becoming understood, and quantitative analyses of the important mechanisms are described. The radiative processes include free-exciton and bound-exciton recombination which are important in p-type and n-type material with and without nitrogen doping, but in all materials the recombination is dominated by non-radiative processes which have proved to be elusive and difficult to eliminate. Particular emphasis is therefore placed on recent advances in the quantitative classification of these dominant non-radiative processes in n-type material. These are: (1) recombination at deep defect levels positioned approximately 0.75 eV from the valence band, (2) diffusion-limited recombination at dislocations, and (3) surface and interface recombination.

The diffusion characteristics of some of the individual dopants involved (phosphorus, arsenic, boron and gallium) are first briefly surveyed, and then interactions between sequential diffusions are described in detail. The effects discussed include 'emitter-push' or 'push-out', which is the enhanced penetration of a base dopant, such as boron, directly beneath an emitter diffusion, and 'base retardation' which is where the diffused base-collector junction away from the emitter advances more rapidly than that directly beneath it. Also discussed are effects associated with buried marker layers. It is suggested that anomalies connected with phosphorus diffusion, namely the rapidly diffusion phosphorus 'tail', the 'push-out' effect and the movement of buried marker layers are all linked and are consistent with the generation of a high supersaturation of point defects. Another general conclusion emerging from the review is that a 'depletion' or 'dip' is found in the base profile (boron or gallium) under either a phosphorus or arsenic emitter diffusion, and that most observations are consistent with that caused by the electric field generated by the diffusing emitter dopant.

A novel electrical analogue technique is described for the approximate solution of problems that have axial symmetry and are governed by the biharmonic equation. The approximation is good for problems where the radius of curvature of streamlines and vorticities is not too small. The technique uses two plane sheets of conducting paper linked by an array of resistors whose resistance is proportional to the distance from the axis of symmetry. The method has been satisfactorily applied to the flow of fluids at Reynold's number below 0.1 and it is shown how values of fluid velocity components and vorticity can be readily determined from the analogue. The flow of fluid parallel to a uniformly spaced array of fibres has been studied, and the Kozeny coefficient deduced shows good agreement with other theoretical and experimental values.

The characteristics of the decay of the surface potential of corona-charged untreated polyethylene films 25 mu m thick in air have been investigated using a well-defined circular charging area and a high-resolution 80 mu m diameter induction probe. Distinct differences between the decay characteristics of negative and positive surface charges emerge. In the former case, the nature of the decay is shown to depend not only on the magnitude of the surface potential (700-3000 V) but also on the presence of the corona light. Corona light incident on the centre of the exposed area during charging subsequently leads to a much more rapid decay of charge in this region. In contrast, corona light has no effect on the subsequent decay of positive charges which appears to decay very much in accordance with known theories. The corona light is shown to be responsible for the so-called 'cross-over' phenomena which hitherto has been observed, notably by Ieda and co-workers (1967, 1968) after positive charging.

An attempt has been made to investigate the changes in bulk and interfacial microstructure that occur when Ge-ZnSe heterojunctions are subject to high-temperature annealing treatments. The structure of the ZnSe has been investigated with X-ray diffractometry and transmission electron microscopy and diffraction. The effect of contamination upon the developments that occur with the interface during annealing has also been examined.

The complex refractive index spectrum of a solid material in the far infrared region ( approximately 5-500 cm^-1) is determined from a single dispersive interferometric measurement of a plane parallel specimen. This method uses essentially the measured (frequency-dependent) transmission of electromagnetic radiation through the specimen and is based on the interferometric formalism developed by the authors. The method is especially suited for transparent solids, which are used as window materials in far-infrared spectroscopic experiments. Results on hyperpure silicon and on crystal quartz are presented as an illustration of this new method.

The mechanical behaviour of some metals has been investigated experimentally for compressive, tensile and torsional loading cycles of about 120 and 370 mu s duration and for steady-state loading in each of the three modes of deformation. For the dynamic measurements the split Hopkinson bar technique has been used in the three modes between two rods and deformed under the action of a transient stress wave induced by dynamically loading the free end of one of the rods. The experimental results show that for the range of high strain rates investigated, the dynamic stress-strain curve is independent of strain rate in all three modes of deformation. For compression and tension, the dynamic stress levels for a given value of the strain exceed the corresponding effectively static stress levels. In marked contrast, only a comparatively small difference has been detected between the corresponding dynamic and static stress levels in torsion over a range of strain rates from approximately 10^-3 to approximately 10³ s^-1.

For pt.I see ibid., vol.10, no.4, p.519 (1977). The Swift (1947) effect, observed here as the permanent lengthening of a tubular specimen along the principal axis about which it is being examined. Permanent length changes in twisted tubes are considered in relation to the Poynting effect and incremental theories of plasticity. It is shown that the increasing body of experimental evidence, including the present measurements on aluminium, copper, iron and an aluminium alloy, cannot be accounted for by an incremental theory of plasticity which assumes incompressibility and interprets the displacement gradients in terms of small strain components. This is considered to arise from the fact that incremental theories based on small strain components do not consider a component of axial strain which is entirely geometric in origin, an observation which is independent of whether the twisted tube is to be regarded as isotropic or anisotropic. Satisfactory correlation of the observed length changes of twisted tubes with the shear strain is obtained with an incremental-type constitutive equation for an isotropic incompressible strain-hardening elastic-plastic solid with large deformation.

For pt.II see ibid., vol.10, no.4, p.533 (1977). The correlation of the stress-strain behaviour in simple compression, tension and torsion is considered for various metals in relation to the Swift (1947) effect, here observed as the permanent lengthening of a tubular specimen along the principal axis about which it is being twisted. The proposed correlation is based on an incremental-type constitutive equation for an isotropic incompressible strain-hardening elastic-plastic solid with large deformation. Formulation of the proposed constitutive equation involves a modified Mises yield condition and the assumption that the total deformation gradient can be decomposed into a matrix product of the local elastic and plastic deformation tensors. Since the Swift effect appears to be absent at high strain rates, the metals of present interest should appear stronger at high strain rates, a conclusion in accord with the observation that the dynamic yield stress is frequently observed to be higher than the static yield stress in all three simple modes of deformation.

The authors calculate the thermal conductivity of composite materials containing thinly dispersed spheroidal inclusions by solving the associated problem of heat flow across one such inclusion embedded in an infinite matrix in the presence of thermal boundary resistance. The theory shows good agreement with experimental data for different composites containing inclusions of various substance, size and geometry, at temperatures ranging from 2 to 300K. By taking thermal boundary resistance into account in a treatment closely paralleling that of Meredith and Tobias (1960), they also extend the calculation to composites with densely packed spherical inclusions, which is again found to be in close agreement with experiment.

The shape of the contact profile of a rubber cylinder loaded against a glass plate in the presence of a viscous fluid has been studied as a function of contact time. The shape of the contact and the absolute film thickness were measured using optical interferometry following the procedure outlined by Roberts and Tabor (1972). The effect of normal load and fluid viscosity have been investigated. The data are compared with the results of a theoretical analysis already reported by Herreburgh (1970). At large film thicknesses (greater than 2 mu m), the comparison between the theory and the experiment is reasonable but at smaller film thicknesses significant differences are observed.

Steel-core/steel-tube duplex reinforcing members have been studied and it has been observed that suitably designed core members can accept peak stress levels in the region of 1 GN m^-2 in zero-tension load cycling for more than 10⁶ cycles. Observations have been made of the growth of fatigue cracks in metal plates reinforced with elements of this type. Two experimental configurations have been investigated in which the reinforcing members have been arranged to be perpendicular to the length of the crack and parallel to the direction of the applied load. An edge crack in a brass sheet has been subjected to a cyclic crack opening force. A central crack was placed in a stainless steel plate, to the surface of which non-fracturing reinforcing members were attached. Studies show that it is feasible to prevent the growth of cracks of appreciable dimensions in stainless steel plates when operating at high stress values under cyclic loading. These characteristics are obtainable in composite systems which contain relatively small volume fractions of steel duplex fibres. Possible engineering applications of these composite systems are outlined.

The authors discuss the stress field in the neighbourhood of an indentation made by a steel ball in poly(methyl methacrylate), the special characteristics of which have been described by Puttick (1973). These results suggest that the deformation in the interior of the solid approximates to that round a spherical cavity expanded by internal pressure in a plastic-elastic medium, as proposed for other materials both by Marsh (1964) and by Johnson (1970). The approximations required to apply ideal plastic-elastic theory to polymers are discussed, and Hill's solution (1950) of the equilibrium equation for the expanded cavity is modified to take into account the sensitivity of the yield stress of PMMA to strain rate and hydrostatic pressure; the result gives the indentation pressure in terms of the radius of the plastic zone. The latter dimension is experimentally determined by measuring the indentation pressure, so that the scale of the stress field in the interior is known. The state of plane stress in the surface is interpreted in terms of the plastic-elastic expansion of a hole in a plate. The tensile component of the stress field thus derived appears to explain well the orientation, shape and size of the stable cracks initiated at such indentations.