Table of contents

Recent work has shown that the growth of ionization currents in air and nitrogen up to high values of the parameter pd (where p is the gas pressure and d the gap distance) follows the well-known generalized Townsend equation and that the observed breakdown potentials are in agreement with those calculated from the Townsend breakdown criterion. Having established the nature of the general mechanism the next stage is to determine the ionization coefficients α and ω/α as accurately as possible. The present paper gives an analysis of the errors in the coefficients arising from experimental errors in the measurement of the ionization currents, and on the basis of the analysis defines the experimental conditions necessary for the accurate determination of the coefficients at high values of pd in hydrogen. The apparatus designed to satisfy these conditions is described, and the values of the ionization coefficients, determined for a range of pd from about 200 to 900 mm Hg cm in hydrogen, are given. Investigation of the effect of the state of the cathode surface on the secondary ionization coefficient showed that even at the comparatively high value of pd similar 350 mm Hg cm in hydrogen, the cathode played a significant role in the growth of ionization currents and thus in the setting of the breakdown criterion.

In removing an electron from a metal, work has to be done against image-type forces. The classical image potential outside projections on an irregular surface is less than outside a plane surface but it is shown that the resulting lowering of the work function is entirely compensated by a patch field. Exact calculations cannot be made even for the single idealized surface projection for which image calculations are available. However, the axial patch potential for this case has been evaluated using an electrolytic tank and this shows that the form of the surface barrier is not affected by surface microgeometry.

It is shown that already at fairly low electronic densities (of the order 10¹⁴ cm^-3) an electron exchanges energy with other electrons quicker than with lattice vibrations, mainly because collisions with the latter are nearly elastic. The notation of an electronic temperature T, which in an external field is higher than the lattice temperature, is then introduced and it is shown that a stationary state becomes impossible if the field is above a critical strength F_c the breakdown strength. For nonpolar and for polar substances F_c can be expressed in terms of the electronic mobility in weak fields, and of other measurable quantities.

The hypothesis is made that in ultra-high-frequency breakdown where the drift movement imposed on an electron by the field is very restricted, the localization of the electron contributes to breakdown. An experimental system is described in which an auxiliary unidirectional or relatively lowfrequency field is imposed at right angles to a microwave field to destroy the localization. When the auxiliary field is unidirectional, or of relatively low frequency (0.86 Mc/s) its application raises the breakdown stress, but when the frequency is raised to 9.7 Mc/s the two fields appear to act quite independently in producing breakdown. A partial explanation is advanced. The gases studied include air, oxygen, nitrogen, hydrogen and neon.

The production of ring cracks in silicate glass, by the impact of steel balls and by the application of a static load to spherical indenters, has been studied for a large range of indenter curvature. It has been found that in impact tests a greater height of fall is required to produce ring cracks than predicted by a critical stress criterion of fracture. In static experiments, two different relations between the force required to produce fracture and indenter size have been found. Below a radius of curvature of approximately 3.5 cm the force is proportional to the radius and above this value to the square of the radius. Impact tests have also been carried out on a specimen of an organic glass-like material, with results similar to those of the impact experiments on silicate glass.

A number of independently found empirical laws for the occurrence of Hertzian ring cracks are all equivalent and can be put in the form of an energy scaling law. It is suggested that this is the true explanation of the observed size effects, which would then lie in the dynamical similarity relations of geometrically similar stress fields and fracture surfaces.

By means of the electro-optical Kerr effect, distortion by space charge of the uniform potential gradient between electrodes in purified chlorbenzene has been observed. Evidence is presented suggesting that the conduction process occurs in filaments. The modified distribution of field strength is in qualitative agreement with a previous theory.

The diffuse reflection spectrum of a powdered semiconductor is characterized by an increase in the diffuse reflectivity as the individual particles become relatively transparent. This increase has a definite linear region of greatest slope which is attributed to an exponential drop in the absorption coefficient. The onset of this exponential drop is suggested as a more universal method of determining absorption edges, and when applied to the interpretation of diffuse reflection spectra gave the following energy gaps (at room temperature) amorphous selenium 1.86 eV, metallic selenium 1.74 eV, silicon 1.20 eV, and germanium 0.69 eV.

The theory of impurity scattering in semiconductors has been developed, and the combination of impurity and lattice scattering has been considered for the general case of any degree of degeneracy of the charge carriers. A comparison is made between theory and experimental results obtained with germanium and indium antimonide.

Magnitude and mass distributions are deduced from radio echo data for sporadic meteors and for various meteor showers. The frequency of sporadic meteors is found to increase by 2.5 times per unit increase in magnitude over the range 0^m-10^m. This corresponds to an inverse square law for the differential mass distribution. Results obtained for shower meteors in the magnitude ranges 0^m to 2^m and 5^m to 7^m cannot in general be represented simply by an inverse power law of meteor masses with constant exponent s. The Arietids are found to have s>2 over the whole range of magnitudes investigated whereas the Perseids, Quadrantids and Geminids appear to have s<2 for the faint and s greater, similar 2 for the bright meteors. We conclude that the greater part of the total meteoric mass is contributed by the following ranges of magnitudes: Arietids, fainter than 7^m; Perseids, Quadrantids and Geminids, 2^m to 5^m. In the case of the Geminids, the concentration of meteoric mass into this magnitude range appears to increase during the course of activity of the shower. Good agreement is found between radio, telescopic and visual data.

Observations have been made of radio echoes from meteor trails at 55 Mc/s in order to test the theoretical estimates by Kaiser and Closs of plasma resonance effects. The aerial array consisted of two Yagis mounted with mutually perpendicular planes of polarization on a common axis. The polarization effects observed are in good agreement with prediction for both short and long duration echoes, but some unexpected results were obtained for meteor trails with line densities in the transition region of 10¹² electrons cm^-1. Some calculations have been made of the effect of the finite meteor velocity upon observed polarization effects, and are in good agreement with observation.

In addition to the polarization measurements, the relationship between the amplitude A and duration T of long duration echoes was investigated. An empirical relation of the form A is proportional to T^0.3 was found, in contrast to the predicted law A is proportional to T^3/16. It is suggested that the discrepency is a result of diffusion by turbulence in the atmosphere.