Table of contents

Epitaxial growth of silicon carbide on α-SiC substrates has been carried out at 1650°C in carbon-saturated silicon solutions. The quality of growth is found to be strongly dependent on the inclination of the substrate in the melt, an angle of 40° to the horizontal resulting in the growth of uniform layers of the same polytype as the substrate. With this technique p-n junctions have been prepared using nitrogen, aluminium and boron impurities, and their diode and electroluminescent properties examined.

By comparing the etch patterns on matched pairs of BaF₂ crystals it has been observed that, although correspondence exists in the number and position of the etch pits, the shape of the pits on heated cleavages differs. The pits produced on non-heated cleavages are triangular in shape, while those on the cleavages heated at a temperature above 600°C are rod-shaped and strictly crystallographically oriented in <110> directions. When a heated cleavage was repeatedly polished and etched and the inclinations of the two types of pits were measured, it was found that (i) the rod-shaped pits nucleated at the intersections of platelets by the plane of observation and (ii) the platelets made an angle of nearly 70°, which agreed with the inclination of the dislocations to the cleavage planes. It is conjectured that the rod-shaped pits are caused by the precipitation of impurities at the dislocations principally in octahedral planes of the crystal.

X-ray analysis and infra-red absorption spectra support the formation of Ba(OH)₂ and BaO in heat-treated crystals. The implications are discussed.

The Johnsen-Rahbek effect is briefly described and the existing theories of the phenomenon are discussed. A model is derived expressing the attractive force between the contacting materials as a function of the voltage applied across them. The model, which takes into account the presence of surface irregularities, is obtained by an assessment of the effect of field emission on the electrostatic capacitor forces between the contacting surfaces. Good agreement with existing experimental results is obtained.

General expressions are developed for the dependence of acoustoelectric gain and current on the electron density distribution associated with an acoustic wave propagating through a piezoelectric semiconductor. The results are valid for nearly-periodic travelling waves with arbitrary harmonic content. It is shown that the frequency and magnitude of the peak gain can both tend to zero as the acoustic wave amplitude increases and the electrons bunch more strongly.

Analytical techniques for calculating the propagation of Gunn-effect domains in non-uniform specimens are discussed. Operational characteristics for domains running between concentric electrodes are computed using a simple domain shape. Hence the variation in the frequency of current oscillations with voltage applied between the electrodes is deduced as a function of the ratio of the electrode radii and of the circuit loading.

The equilibrium magnetization distributions of various types of domain walls in the prism plane of cobalt have been calculated. The effects of the first- and second-order anisotropy constants are included when calculating the minimum energy of both Bloch and Néel walls.

The initial permeability of high-permeability alloy sheet rolled from sintered compacts is substantially increased by sintering for times longer than are currently used. Complete alloying of copper does not take place in a compact using a conventional sintering time; sheet rolled from such a compact has a banded microstructure so that the material is slightly inhomogeneous. An upper limit to the initial permeability of such sheet is probably set by internal strains caused by local variations in the lattice parameter.

In experimental investigations of the surface magnetic properties of ferromagnetic materials using the Kerr magneto-optical effect, difficulties can arise in the interpretation of signals if the polarizer transmission axis is set at an arbitrary angle to the plane of incidence. A rigorous method of interpretation is given here on the basis of a combination of the transverse and longitudinal Kerr effects. Particular experimental arrangements which allow of simplified interpretation are also briefly discussed.

Thermally stimulated current peaks, corresponding to second-order thermo-luminescence when the recombination lifetime is a rapid function of temperature, are investigated. By numerical calculation, values of the maximum and the half intensity temperatures are calculated. A method for calculating the activation energy by use of the maximum and half intensity temperatures is found empirically in analogy with similar first- and second-order methods.

The initial-rise method for calculating the activation energy is also shown to hold true in this case. The value of the parameter μ_g characterizing the geometrical shape of the peak is found to be around 0·8 for various values of the activation energy and the frequency factor. This, in comparison with the value of about 0·4 for first-order and 0·5 for second-order peaks, indicates very slow decay of this phenomenon at high temperatures.

Recent measurements on iron single crystals of 99·999% purity, containing less than 0·05 p.p.m. of carbon, show that the strong increase of the flow stress with decreasing temperature in the range 10-300°K is largely independent of small concentrations of interstitial impurities; it appears to be due to the Peierls force. Theoretical considerations employing the double-kink mechanism of dislocation propagation do not account satisfactorily for the observations. A new, simple, model is developed in which dislocation movement is induced by thermally activated formation of curved segments of dislocations trapped in Peierls troughs, the curvature being determined essentially by the effective shear stress. The predominant role of the shear stress in the determination of the curvature of the dislocation segment in the saddle-point configuration is the principal feature differentiating the present model from earlier double-kink treatments. Measurements made on iron and other body-centred cubic metals are successfully explained on the basis of the model.

The range of solid solution has been investigated in alloys of PbTe-GeS, PbSe-GeS, PbSe-GeTe and PbSe-GeSe. The solubility of GeS in PbTe and PbSe was limited to about 5·7 and 28 mol.% GeS respectively, and of GeTe and GeSe in PbSe to about 16·5 and 40 mol.% respectively.

Measurements of the variation of the optical energy gap have been made for the lead-germanium chalcogenide systems as a function of both composition and temperature. The gap was found to increase with germanium chalcogenide concentration, to increase with temperature in the PbTe-GeS, PbSe-GeS and PbSe-GeTe alloys, but to decrease with temperature in the PbTe-GeSe, PbSe-GeSe and PbTe-GeTe alloys.

When the concentration of GeTe in PbTe reached 40 mol.% the energy gap decreased rapidly. This could suggest that the gap of 1 ev measured for GeTe is in fact the second minimum band separation and not the first.

The basic equations of neutron-wave and sound-wave propagation are deduced by appealing directly to the non-linear Boltzmann transport equation. The approximations inherent in these linear equations are pointed out and the properties of the respective neutron and gas-atom scattering kernels are elucidated.

The low-frequency behaviour of both types of wave is studied and explanations are given for the small attenuation coefficient of sound waves as compared with neutron waves of the same frequency.

The source boundary conditions are discussed and certain approximate conditions are suggested for the sound-weave problem that may reduce the mathematical complications.

The effect on the dispersion law of finite transverse dimensions is explored in some detail. It is found that although a dispersion law does not exist for neutron waves in a sufficiently small system, one does exist for sound waves. However, the frequency range over which it exists becomes smaller as the transverse dimensions become smaller.

The pulsed-neutron method has been used to measure the temperature dependence of both the diffusion constant and the extrapolated end point in water. The measured temperature dependence of the diffusion constant was found to be in good agreement with the theoretical values predicted by the Nelkin kernel.

Measurements of the extrapolated end point in a cylinder of diameter 7 in. and height 7 in. gave a value of 0·323 ± 0·013 cm, with a similar result for a 7 in. cubical system. For an 8 in. diameter cylinder 4 in. high, however, the value was 0·465 ± 0·039 cm, a change which would not be expected theoretically. The extrapolated end point was also measured at temperatures up to 250°C. For a 4 in. high cylinder no definite temperature dependence was observed, but in a 16 in. high cylinder there was an increase with temperature from 0·340 ± 0·030 cm at 18°C to 0·469 ± 0·021 cm at 250°C.

Measurements of the electrode temperatures for arcs operating in room air over a current range of 5-30 A are given. A distinct difference in the peak temperatures of the radial distributions for graphite and carbon cathodes has been observed, the values being 4150 and 3500°K respectively. Using the Richardson-Dushman equation approximate values of the corresponding peak cathode electron current densities have been calculated and these are given as 4000 and 700 A cm⁻². Radial temperature profiles for the graphite anode are also given for the same current range.

Measurements of the expansion rates of the cylindrical shock waves produced in the early stages of spark-channel formation are given. Comparison with predicted rates, based on an improved analytical treatment, are satisfactory.

The model of the coaxial plasma gun presented in Part I of this paper has been extended by consideration of the detailed structure of the flowing gas layer preceding the driving current sheet. It is found that to a good degree of approximation the gas is uniform across the layer. The conservation of energy equation then reduces to Bernoulli's equation, and allows calculation of the state of the gas in the layer. The range of validity of the model is examined.

A technique for measuring the rate of release of a primary trapped ion species from tungsten when subjected to a second ion species is described. Preliminary data show that gas release rates are generally much larger than would be anticipated from a mechanism based upon erosion of the tungsten with attendant gas emission and that the second ion species releases the primary gas by some form of efficient energy transfer process.

The variation of the apparent coefficient of secondary emission of molybdenum bombarded by Cl⁺, Cl⁻, O⁺, O⁻, H⁺ and H⁻ ions was determined as a function of the cleanliness of the surface and the energy and incidence of the ions. These measurements, performed at a pressure lower than 1 ntorr, show that the kinetic emission is the same in the case of positive and negative ions of the same element, in agreement with the theory of Parilis and Kishinevskii. Finally the high value of the apparent emission under bombardment of H⁻ ions is interpreted by taking account of the phenomenon of detachment.

Electric field distributions have been measured in nitrobenzene using the Kerr electro-optical effect. Both metal and composite electrodes have been used to investigate their charge-injection properties. From the field-distribution measurements it has been shown that positive injection readily occurs when two metal electrodes are used with a space charge near the cathode. When two dissimilar electrodes are used (composite and metal), injection from the composite electrode is always suppressed in comparison with the injection from the metal electrode, this phenomenon being independent of polarity. From the space-charge thickness, estimates of mobility are 1·33 cm²v⁻¹s⁻¹ for the positive carriers and 0·28 cm²v⁻¹s⁻¹ for the negative carriers.

An estimate has been made of the significance of coagulation in barium-oxide-argon suspensions postulated as working fluids for closed-cycle magneto-hydrodynamic generators. Assuming an initial particle diameter of 0·05 μm and a barium oxide to argon weight ratio of 0·1, coagulation would cause the mean particle diameter to increase to at least 0·26 μm during a single circuit round an MHD steam turbine power station with a net output of 500 MW. According to previous theory the electrical conductivity of the suspension would then fall to an unacceptable value.

The problem would be expected to occur also with suspensions of other solid thermionic emitting particles.

An expression is derived for the brightness or directional beam intensity of an electron beam, for a general velocity distribution at the emitting cathode. The result is then specialized to the case of the brightness on the axis in the direction of the axis for a cylindrically symmetric system. In this case the Langmuir law B/J=eV/πkT is shown to hold for emission in the Schottky region even though the velocity distribution is not Maxwellian; and an analogous expression B/J=eV/πd, where d is a function only of the applied field, is shown to hold for field-emitted beams. It is shown that detailed ray tracing is necessary to evaluate the brightness for points not on the axis and directions not in the axial direction.

The theory of heat conduction through polyatomic gas mixtures given by Saxena, Saksena, Gambhir and Gandhi has been extended for those mixtures where each component may have more than one internal degree of freedom. It is concluded that for mixtures consisting of complex molecules, the deviations from the above theory will be quite pronounced.

Experimental results are presented which show that surface contamination due to the method of preparation and bulk trapping centres can have a first-order effect on semiconductor impurity concentration measurements using Schottky barriers. A criterion is established which allows the reliability of the measurements to be assessed, and a method is given for measuring the impurity concentration in the presence of traps.

A method for determining the optical constants of a metal covered by a thin film using ellipsometric data is given. This is very useful when it is difficult to obtain the bare surface of the metal. The method is illustrated by determination of the optical constants of chromium from ellipsometric data, obtained with a chromium surface covered with a film of Cr₂O₃.

Some special factors involved in probe measurements which could lead to a gross over-estimation of the electron temperature in flames are discussed. The importance of the proper positioning of the probes with respect to secondary reaction zones is demonstrated for atmospheric flames. An attachment mechanism, operative below 373°K, is proposed to account for the high values of electron temperature measured in hydrocarbon flames by earlier workers using water-cooled probes.

The effect of irradiation by several spectral bands in the visible region on the charge transferred by the pulses occurring in the Siemens' type ozonizer discharge through hydrogen due to d.c. excitation has been studied. It has been observed that the charge transferred per minute is reduced under irradiation. This reduction is found to be independent of the potential applied to the discharge tube in the range 800-1200 v. On the other hand, the reduction in the charge transferred increases with the intensity and the frequency of radiation. The plots between the charge transferred and the logarithm of intensity are linear; the slope of these lines reduces with the increase in the wavelength of radiation. These plots suggest that there may be a long-wavelength limit beyond which the Joshi effect would not be observed. Any such limiting wavelength exceeds 600 nm under the present experimental conditions. The results can be explained on the basis of photoelectric emission from the adsorption-like boundary layers on the ozonizer, as postulated by Joshi.