Table of contents

A computer-based theoretical model has been developed for pre-breakdown current growth for the low pressure side of the Paschen minimum. This model shows how the re-entrant section of the Paschen curve obtained in Hg vapour can take place, as well as giving reasons why a continuous curve is obtained in most gases. The theory has been confirmed by new measurements of pre-breakdown current at voltages as high as 6-8 kV, an order of magnitude higher than previous values and well within the observed inflexion.

An outward travelling-wave current distribution can be produced on a centre-fed dipole, by loading it with nonreflecting loading at the ends of each element. The radiation resistance of such an aerial is evaluated in an isotropic, collisionless, warm plasma, using hydrodynamic theory. It is shown that this aerial is less affected by the presence of plasma than the usual unloaded dipole aerial, which has a sinusoidal current distribution, and that it can usefully be employed in satellite transmission.

The various criteria which determine the degree of a polynomial to be used in curve fitting have been assessed by applying them to eighty sets of data. The most reliable truncating criteria are found to involve the use of the statistically based t or F tests. An ad hoc criterion involving the standard errors of the polynomial coefficients is found to be satisfactory on 90% of the occasions it was used.

The structure of planar detonation waves propagating in a narrow rectangular channel, of cross section 3×¼ in², has been studied for a mixture of stoichiometric oxyhydrogen diluted with 60% argon. It is confirmed that at an initial pressure of 80 torr the waves generated are self-sustaining and possess either two or three Mach interactions along the broad wall of the tube. Multi-spark schlieren and interferometric records, together with the trajectories of the triple-points obtained from smoked-foil records, enable the flow fields associated with the transverse wave system to be described.

It is found that the pressure and density ratios across the transverse shock that are computed from observed wave velocities and also from the direct measurement of density changes, are in reasonable accord with oblique shock analysis in which the Mach stems are assumed to be stationary with respect to the triple-point. For the particular Mach configuration which is examined the strength of the transverse wave is found to vary in a continuous manner over the experimentally observed range of flow entrance angle. From the measured induction reaction times behind the leading shock front the position of the recombination zone in the doubly-shocked gas behind the transverse wave is deduced; this is found to be in agreement with its observed position on the interferograms.

A coaxial cable can be regarded as a screened surface waveguide and, for wall impedances of any given value, propagation in the dominant mode, not subject to cut-off, is examined.

The effect of loading a supporting surface is seen to reduce dispersion and the proximity of the outer surface to the inner is shown to affect the rate of decay of the field in the radial plane.

The details of the study of spectral distribution and colour of the daytime north sky at Bombay are presented. The results are discussed in relation to the results at Delhi and other places outside India. Individual measured curves for Bombay are compared with the typical distributions derived from Delhi data and are found to agree fairly well though there are still differences at both ends of the spectrum. The possible reasons for the disagreement in the ultraviolet noticed between Bombay and Delhi curves and in the curves of other workers are also discussed.

Non-uniformity and anisotropy of the conductivity of a liquid flowing through an electromagnetic flowmeter affects its sensitivity. This effect is analysed for a circular meter with small electrodes and rectilinear flow with an axisymmetric velocity profile. The magnetic field is assumed invariant in the flow direction. The conductivity tensor is assumed to have principal axes in the radial and azimuthal directions with the principal conductivities σ_r, σ_θ depending only on the radius. The sensitivity is found analytically for an arbitrary plane magnetic field in the case σ_rσ_θ=constant. When the field is uniform the condition on σ_r and σ_θ necessary to make the sensitivity independent of the shape of the velocity profile is found. The sensitivity change is deduced for small perturbations from this conductivity distribution, for an arbitrary axisymmetric velocity profile. The effect of blood vessel walls having large conductivity variations of a simple type is also found. The results agree well with the numerical calculations of Baker in 1970 and satisfactorily with deductions made from the experimental results of Phibbs in 1968 and Dennis and Wyatt in 1969.

Measurements are made of the electrical parameters of high current discharges in capillaries and the results compared with a recent theory put forward by Wheeler in 1970. There is excellent agreement between the experimental electrical characteristic and the theoretical prediction which has the form EαI^2/5. The capillary voltage drop is found to depend on the nature of the initial gas filling. This is unexpected and leads to a thorough investigation of the process and consequence of wall evaporation. It is shown that the discharge plasma is almost exclusively composed of ionized wall material with an average ionic charge of the order of 3. However, it is suggested that the initial phase of the discharge, which is controlled by the gas filling, influences the subsequent rate of evaporation of wall material.

Spatial distribution of the conduction current in a full scale mercury arc converter valve has been studied with a segmented anode. Below 80 A peak (at 50 Hz) the distribution was asymmetric and diffuse. Between 80 and 600 A it changed to a number of constricted filaments which moved in the grid apertures, their velocities increasing linearly with current. Single half-cycles of 1 to 8 kA peak were non-uniformly distributed and irreproducible. The constricted filament mode of conduction is tentatively explained in terms of self-magnetic pinching in the grid apertures. Practical implications of the observed current distributions are discussed.

A diagnostic method suitable for studying temperature and composition distributions in hot gases based on using an optical method for measuring the local velocity of sound has been developed. To avoid excessive interaction between the ultrasonic beam (about 100 mW cm⁻²) and the flowing gas (up to about 100 cm s⁻¹) it was necessary to photograph the acoustic wavefronts within as short a time as possible (5 ms or less) after switching on the ultrasonic beam. The limitations of the method, which has been tested up to about 1000 K, are discussed. It seems probable that, by using higher frequencies and acoustic power densities, it should be possible to apply the method to gases up to about 3000 K.

The normal approach of a rubber sphere to a glass plate through thin separating films of different liquids has been studied by optical interference over a thickness range of 50 to 100 000 Å. The formation and collapse of liquid `bell' entrapments produced by the elastic indentation of the rubber surface is described, as is the role of repulsive electrical forces that can offer resistance to this collapse. Under certain conditions these forces are able to support the normal load so that after some time of squeezing an equilibrium film about 200 Å thick is reached.

The thinning of a liquid film with time provides a measure of its viscosity. Measurements made upon water films (containing 0·3% soap) thinning from 2000 down to 200 Å showed that their effective viscosity over this whole range was constant and nearly the same as the bulk viscosity. This suggests that for our system no structure is induced into thin water films by the solid surface.

Optically smooth rubber pressed against glass provides a microconforming contact between which a thin uniform film of liquid can be sandwiched. The contact largely overcomes the problem of asperity interaction between surfaces. The rubber surface could be slid relative to the glass plate under controlled conditions and in this way it proved possible to study the shear properties of liquid films as thin as 40 Å. Results for water show that down to film thicknesses of 70 Å there is no increase in shear strength of the liquid. Below this thickness asperity interaction increasingly confuses the interpretation of measurements.

If the water film contains a surface active agent, monolayer protection does much to reduce the friction between surfaces when the film is less than 70 Å. In the absence of monolayers the friction rapidly rises with decreasing film thickness towards the dry value.

Experimental values of the thermal conductivity of molten salts are reviewed and found to deviate appreciably from values predicted by semi-empirical theories to the thermal conductivity of liquids. The author shows that for univalent molten salts a simple correlation exists between the thermal conductivity at the melting point and the mean ionic weight of a univalent molten salt. Mixtures of molten salts have a thermal conductivity which is less than the proportional mean of the constituent salts.

A systematic study of the pre-breakdown conduction current pulses in transformer oil using static electric stresses showed that these pulses are very sensitive to test conditions. The pulses were found to depend on gap length, applied stress, temperature, electrode material and hydrostatic pressure. The distribution of the pulse heights was exponential, indicating the presence of a multiplication process in the liquid. It is suggested that the pulses are due to the ionization of microscopic gas bubbles by electrons emitted at the cathode.

The electron density of an argon plasma jet is evaluated from probe measurements using several different theoretical treatments of probe behaviour in a dense plasma and the results are compared with those from spectroscopic measurements. It is concluded that none of the theories tested is adequate to evaluate the plasma density from the probe measurements.

Low energy electron diffraction, Auger electron spectroscopy and photoemission studies have been carried out on surfaces of molybdenite prepared by cleavage. Surfaces cleaved in ultra high vacuum are extremely inert. There is some evidence that surfaces containing adsorbed oxygen have a weak affinity for water vapour and the significance of the result for some of the mechanical properties of MoS₂ is discussed. Low energy electron diffraction measurements are consistent with a surface structure not differing greatly from that of the bulk, and the energy analysis of back scattered electrons confirms the existence of two valence bands in MoS₂.

In the longitudinal heat flow method for the measurement of thermal conductivity, the surface losses present a very serious source of error, specially at high temperatures. A mathematical procedure has been proposed to eliminate this source of error. It is shown that if the temperature is measured at three points instead of two as is usual, the surface losses can be accounted for.

Negative resistance and memory effects occurring in formed thin insulating films are described and new experimental results for zinc sulphide thin film devices are presented. Contrary to previous work the voltage for maximum current is shown not to be a constant. The switching time of the devices is interpreted in terms of an activation energy. Previous theories are shown to be unsatisfactory and a theory based on the thermal rupture of conducting filaments and applicable to all such formed devices is presented.

The dependence of forward current upon the height of the drain potential step in a metal-semiconductor-metal structure was investigated using the current-voltage characteristics of Al-p-type Ge-Au diodes in forward and reverse bias at various temperatures. It was observed in forward bias that the current was two-carrier type when ϕ_h>ϕ_e and one-carrier type when ϕ_h<ϕ_e where ϕ_h and ϕ_e are the drain (Au) potential steps for holes and for electrons respectively. The two-carrier current was much smaller than the one-carrier current.