Table of contents

An expression is found for the distribution of light in the image of two small apertures which are illuminated by a condenser of numerical aperture equal to s times that of the imaging objective. The resolving power of the system is written ρ=Kλ/(N.A.), and K is found as a function of s. It is shown that precisely the same result obtains for both critical and Kohler illumination.

A direct photoelectric method allowed the reflectivity R and the transmissivity T of partially reflecting films to be measured to an accuracy of 0.3% of the incident light intensity. For a large number of silver films, deposited on glass by evaporation in vacuo, the values of R and T were measured for visible light of five different wavelengths, in the range of high reflectivities (R>0.7). The change of R as a function of time was also studied. With only moderate control of the conditions of surface cleaning and evaporation, the values of R were nearly always found to he close to a `standard' curve giving R as a function of T, the wavelength, and the age of the film. Possible causes of scatter from this standard curve were investigated T was found not to obey an exponential law as a function of the thickness.

On the basis of these results, resolving power and brightness of etalon fringes were plotted as functions of T. The conclusion is reached that, in most practical cases, the resolving power of etalons for visible light is limited by imperfections of the surfaces rather than by the reflectivity of the silver films, if their thickness is properly chosen.

Recent interest in lead sulphide, selenide and telluride as photoconductors makes information on their absorption spectra valuable. It is found that the absorption spectrum of each material is characterized by strong absorption in the visible and ultraviolet region, with a further band in the infra-red. The effects of temperature and oxidation are also studied.

In PbS the forbidden energy region between the filled and empty bands is found to be about 1.3 ev. in width, not 0.4 ev. as previously suggested. Hence photoconductivity in PbS cannot be associated with the main lattice. The corresponding energy gaps in PbSe and PbTe are found to be about 1.05 ev. and 0.9 ev. respectively.

An analysis has been made of a large number of hot cathode low voltage arc-type discharges in mercury vapour and hydrogen at low pressure by the cold probe methods of Langmuir and Mott-Smith. It has been verified that the space potential in the plasma near the filament is positive relative to anode potential, and it has been shown that consistent data for the space potential and concentrations of the slow (ultimate) electrons are obtainable from analyses of probe characteristics in both retarding and accelerating fields for electrons. The way in which the fast (primary) electrons develop a random motion has been traced by progressively confining them within the main discharge, firstly by using a cylindrical gauze anode with open ends, secondly by using a solid anode with open ends and lastly by using a solid anode with end caps.

A method is described for the measurement of the viscosity of molten metals at temperatures up to about 900° C. The metal is contained in a cylindrical crucible which is allowed to oscillate about its axis under the control of the torsion in a suspension wire. The theory of the method is considered.

The viscosity of zinc has been determined in the temperature range 430° C. to 480°C. and the results are given.

Methods are described for the measurement of resistive and inductive permeabilities over a range of frequencies from 150 to 10,000 Mc/s. The measurements of the resistive permeabilities were carried out by comparing the attenuation of a coaxial transmission line with the ferromagnetic as the inner conductor with that of a coaxial line having a non-ferromagnetic as inner conductor The inductive permeabilities were obtained from the wavelength in a coaxial line having the ferromagnetic as the inner conductor. The results show that at all frequencies the resistive permeability is greater than the inductive, and both decrease to unity at very high frequencies. The effective permeability, which is assumed to be complex, is deduced as a function of resistive and inductive permeabilities and plotted against frequency for a variety of materials. The experimental values are compared with theoretical values derived from two formal treatments.

An economical way of producing a pulsed magnetic field is to use condensers as the energy source and an air-cored magnet coil which, with the storage condensers, forms an oscillatory circuit By means of a switch the coil is connected to the charged condensers and then disconnected precisely at the end of one cycle. In this way all the stored energy is used to produce the field, while most of it is recovered at the end of the pulse, the only appreciable losses being due to the resistance of the coil and switch.

The design of coils for use in the way described above is considered in detail. Expressions and curves are presented which enable the field produced by a specified coil and condenser combination to be readily calculated. Conversely, the data can be used to design the most satisfactory coil for any practical application An example is given of the design and construction of an air-cooled coil.

A theory of contact noise is described in which the low-frequency noise is attributed to the random movement of adsorbed ions on the surface of a semiconductor from which an electron current is being drawn. Emission of electrons is assumed to take place only at localized patches on the surface and the adsorbed ions are assumed to give rise to a Schottky barrier layer, in which the potential maximum is linearly related to the concentration of ions. Diffusion of the ions over the surface gives rise to random fluctuations in the concentration of ions in a patch which results in random fluctuations in the height of the potential barrier and the emission current. It is shown that for a circular patch the spectral power density of the noise current varies with mean current J0 and frequency f as j0²f^x over a small range of frequency and that x varies monotonically from -0.75 at the lowest frequencies to -1.125 at the highest frequencies. It is also shown that for a long thin rectangular patch in index x varies monotonically from -0.5 to -1.5 as the frequency is increased from zero. The dependence of the noise power density on temperature is also discussed.

In order to provide further evidence in support of the structural change which takes place in a liquid when it is supercooled, measurements of dielectric loss at high frequencies and of dielectric constant have been made on various polar liquids. The evidence from the dielectric loss measurements is inconclusive but for the four liquids tested a significant discontinuity in the dielectric constant-temperature curve has been established at the melting point in each case.