Table of contents

The existence of a boundary layer of different impurity concentration on the solid-melt interface during crystal growth into impure unstirred melts has been inferred by previous authors. The problem of how rapidly this layer attains its final form in the case of growth into an infinite unstirred melt, hitherto treated approximately, is shown to be capable of exact solution for all values of the segregation constant. The impurity distribution in the resultant solid is calculated and graphs are given for various values of the segregation constant.

Experiments have been made with a radioactive, annealed steel pin rubbing against a hardened steel ring to compare the amount of wear with the amount of metal transferred from one surface to the other by welding. It was found that a layer of transferred matter from the pin was formed on the surface of the ring and that an equilibrium condition was reached in which the loss of wear debris from this layer was balanced by further transfer. By using a combination of radioactive and inactive test-pieces the rate of transfer to the ring in the equilibrium condition could be estimated and was found to be the same as the wear rate of the pin. There was no evidence that any loose wear particles were formed by direct removal of material from the pin. The transferred layer became oxidized due to frictional heating and it is concluded that the wear process consisted of three stages: transfer of metal, oxidation of the transferred layer, and the subsequent removal of the oxide to form a loose wear-product.

Analysis of three years' observation (1949-1951) of the ionized aurora indicate that it displays a marked diurnal distribution having a strong maximum at 2100 L.S.T. and having 86% of the total occurrences between 1700 and 2400 L.S.T. Auroral activity begins to increase between 1200 and 1500 L.S.T., attaining a nocturnal maximum about six hours later. The diurnal variation does not seem to be dependent upon geomagnetic control in the geomagnetic latitude interval 50-60° N.

At temperatures below 1°K the magnetic susceptibility of cerium ethyl sulphate is highly anisotropic, the molar Curie constant being 1.35 parallel to the crystal axis and less than 0.01 perpendicular to the axis. Demagnetization experiments using fields along the axis show that the magnetic susceptibility in this direction passes through a maximum at low temperatures, and that the specific heat is too large to be accounted for by magnetic dipole interaction between the magnetic ions. This behaviour is attributed to antiferromagnetic coupling of the ions, possibly due to the atomic electric quadrupole-quadrupole coupling suggested by Finkelstein and Mencher.

It is shown that after demagnetization to a low temperature it is possible to apply a magnetic field in a direction of low susceptibility without greatly raising the temperature, so that magnetic measurements can be made below 1°K. This technique has been applied to the determination of the critical fields of superconducting zinc and lead at temperatures below 1°K.

New determinations have been made of the dielectric constants of benzene, cyclohexane, dichlorethane and nitrobenzene with the object of providing values of a higher accuracy than has hitherto been available for four standard liquids covering the range of values commonly required. Liquids of specially high purity were obtained with the co-operation of Professor J. Timmermans of Brussels and Dr. Herington of the Chemical Research Laboratory, Teddington, and the measurements were made by the a.c. bridge technique developed at the National Physical Laboratory for precision capacitance measurement and also by a cavity resonator method.

Part I of the paper discusses the general problem and the special features of the capacitor designed to receive the liquids and surveys the results obtained. Part II gives details of the measurements made by capacitance methods at frequencies from 50 c/s to 10 kc/s, and Part III describes the measurements made at microwave frequencies.

The values obtained confirm estimates of the most reliable values and their probable accuracy published in 1951 by A. A. Maryott and E. R. Smith as a result of a survey of existing data made at the National Bureau of Standards, Washington. It is considered that the new values narrow the range of uncertainty by a factor varying from 3 to 10 for the different liquids.

The accuracy obtained for benzene at audio frequencies was within about 0.01%; at 9000 Mc/s the accuracy was a trifle worse but the results show clearly that although the dielectric constant exceeds the square of the optical refractive index by about 1%, no measurable diminution in dielectric constant occurs in passing from 50 c/s to 9000 Mc/s. A considerable increase in power absorption is however noticeable at 9000 Mc/s and an even greater increase at 25 000 Mc/s, and it is shown that this absorption is not entirely dependent on the water that most samples of benzene contain and that can only be removed by weeks of drying treatment. The necessity of applying such treatments is a serious disadvantage to the use of benzene as a standard liquid for measurements of dielectric constant, and cyclohexane, which is free from this disadvantage, is recommended instead.

A method for measuring the injection ratio of metal-semiconductor contacts is described. It is suitable for measurements at low emitter currents. The current through the contact is applied as a series of constant current pulses and the change in conductance of the specimen is measured as a function of time. Measurements have been carried out for tungsten point contacts on n-type and p-type germanium and results of injection ratio as a function of emitter current are given.

The frequency and intensity of the sound produced by flow past circular cylinders was investigated experimentally in the range of Reynolds number in which a periodic shedding of vortices occurs. The cylinders were placed in a wind tunnel or on a whirling device.

The sound resulting from the production of a vortex street was found to possess a dipole field which consisted of a predominant fundamental frequency accompanied by harmonics. The investigation was confined to the examination of the intensity and frequency of the fundamental. When this frequency was expressed as a Strouhal number Nd/U agreement was obtained between the acoustic frequency and that found, by other workers, in the wake itself. The sound intensity was found to vary with Reynolds number in a similar way to the pressure coefficient at the rear of the cylinder.

The author's vibrating plate method of forming powder ridges on a flat surface has now been used to study the phenomenon with graded powders of different densities and at different frequencies and intensities of vibration of the sound field. It is found that the distances between the ridges increase roughly as the square root of the particle diameters, a result which is in agreement with Bagnold's observations with powders under oscillating water, but needs still to be taken into account of in connection with Kundt powder ridges. `Granular' ridges, one particle wide, are nearer together than powder ridges, the spacings of which increase with the amount of powder which is present. Much larger intensities are required to move dense powders than light powders; the resulting spacings are, however, approximately the same. The variables which have most influence on the spacing are the diameters of the grains and the frequency of the vibration, not the intensity. These matters are discussed in relation to the phenomenon of `wavy' ridges. Experiments in which larger objects, as well as powders, are introduced into the field are also described.