Table of contents

This paper deals with the high-voltage (1·5 MV) Toulouse electron microscope and contains some unpublished information about the objective lens and the column of the instrument. Some examples are given of the advantage of working at very high voltages; in particular the increase of resolving power and the spectacular reduction of chromatic aberration are pointed out. A few micrographs give an idea of the results obtained with this microscope.

Films of ZnS were deposited on to NaCl by vacuum evaporation. The influence of the growth conditions on the structure of the films was studied by transmission electron microscopy. The most important single factor in producing the widest range of epitaxial substrate temperatures and the most perfect films was the use of focused electron-beam evaporation. Contamination raised the minimum substrate temperature for epitaxial growth. It also resulted in the incorporation of doubly positioned wurtzite-structure grains in the sphalerite-structure films. Co-evaporation of an excess of either Zn or S with ZnS produced sphalerite-structure films containing singly positioned wurtzite grains. The use of conventional hot tantalum evaporators resulted in the incorporation of 0·2 wt.% of tantalum in the films and in poor structural quality. Reproducible current-voltage characteristics were obtained in the best structural quality films. These current-voltage characteristics could be understood in terms of the effect of Schottky barriers but not of space-charge-limited currents.

Carbon fibres graphitized at high temperature have been investigated by the techniques of x-ray diffraction and electron microscopy. The particular sample discussed here contains misoriented turbostratic crystals stacked end to end with widths in the range 20-110 Å; the maximum frequency occurs at 65 Å and this is also a reasonable measure for the length of a crystal. Sharp-edged voids with a mean width of less than 10 Å separate the crystals. A perfectly stacked three-dimensional graphite phase has been discovered and is the predominant phase in recrystallized nickel-graphite fibres. Regions with relatively short turbostratic crystals are also found in the nickel-graphite fibres. The mechanism of recrystallization is discussed in terms of these structures. Extinction bands orthogonal to the crystallite direction are held to define twist boundaries between misoriented crystals. The presence of tilt boundaries is also inferred and the boundaries are assumed to comprise non-basal screw and edge dislocations.

The flow of dilute polymer solutions in water, through tubes and granular beds, has been compared with the flow of water and of glycerol solutions and common salt solutions through the same tubes and granular beds. Particular attention was paid to the value of the Reynolds number (Re)prime at which the Hagen-Poiseuille law broke down for tubes and at which Darcy's law broke down for beds. The polymer solutions used were: 0·005, 0·01, 0·025 and 0·05% (w/w) polyacrylamide in water and 0·005% (w/w) `Cellofas' (a sodium carboxymethyl cellulose) in water. Two tubes were used; one was 0·30 m long with an internal diameter of 3·68 mm and the pressure difference was measured across it; the other was 0·70 m long with an internal diameter of 4·08 mm and the pressure difference was measured across the last 0·30 m. Two granular beds were made up, one with glass beads of diameter 2·06 mm, the other with glass beads of diameter 1·15 mm. Most experiments were performed at 25°C, but some were done at 15°C and others at 35°C so that the influence of temperature on the results could be studied. Destabilization occurred in the flow of polyacrylamide solutions through the tubes and the beds; that is (Re)prime (polyacrylamide) < (Re)prime (water). However, stabilization occurred in the flow of Cellofas solutions. Drag reduction occurred in the flow of polyacrylamide solutions through the tubes but not in the flow through the beds; no drag reduction occurred in the flow of the Cellofas solutions. The destabilization can be attributed to the release of energy stored in stretched molecules in the flow, and drag reduction can be attributed to energy absorbed by the molecules thus damping down some turbulent eddies.

The Couette flow of 0·005, 0·01, 0·025 and 0·05% (w/w) solutions of polyacrylamide in water and of 5% (w/w) solutions of glycerol in water, has been compared with that of water. The coaxial cylinder apparatus used was very similar to that designed by Donnelly. At low rotational speeds of the inner cylinder, the torque on the outer cylinder was proportional to the shear viscosities of the liquid and to the speed of rotation. At a definite value of the Taylor number (Ta) = (Ta)_c instabilities occurred; then for (Ta)>(Ta)_c drag reduction in the polymer solutions was such that the torque on the outer cylinder was the same as that for the solvent (water). Furthermore (Ta)_c (polymer solutions)<(Ta)_c (water) = (Ta)_c (glycerol solution) = 3390. Equations derived to account for destabilization and drag reduction in tubes are applied to the Couette flow.

A new rheometer, called the balance rheometer, has recently been introduced. It consists essentially of two concentric spheres rotating with the same angular velocity about two axes which pass through the centre of the spheres, the angle between the axes being small. In the present paper a theory is developed which indicates that the balance rheometer can be used to determine the complex viscosity of an elasticoviscous liquid.

The yield strengths of polyethylene-stainless-steel lap joints and of polyethylene samples having the same thermal history have been measured over a range of strain rates at temperatures between 245 and 323°K. A yield mechanism based on the motion of dislocation loops through crystalline regions of the polymer is preferred to one based on the jumping of polymer segments into adjacent free volume in an amorphous region. The size of the dislocation loops may be derived from the time dependence of yield strength, and it is found that the ratio of these sizes for the polymer when in the joint and when free is constant under all conditions examined.

Results of experimental studies of adhesion between gold/gold, platinum/platinum and gold/platinum-group metal contacts, over the load range 1 mg to 10 g, are interpreted on the basis of simultaneous measurements of contact resistances. The relationship established between coefficient of adhesion and contact resistance demonstrates the importance of surface films in preventing adhesion. Adhesion between gold/gold contacts was significantly higher than between gold/platinum-group metal contacts. In gold/gold contacts adhesion increased markedly above 90°C, but with platinum/platinum contacts temperatures above 550°C are required before adhesion is accentuated.

Measurement of the high-frequency differential capacitance associated with the surface space-charge region on evaporated polycrystalline layers of both PbTe and InSb has shown that, although qualitatively similar, the surface-state distribution on PbTe is approximately an order of magnitude larger than that on InSb. The quiescent surface charge is ~ 740 nc cm⁻² on PbTe and ~96 nc cm⁻² on InSb.

Hall-voltage and resistivity measurements have been carried out on both materials and the apparent intrinsic behaviour of both p- and n-type PbTe films is shown to be the result of impurity compensation by surface states.

Single crystals of α-MoTe₂ having graphite electrodes applied to a cleavage face show a photovoltage when irradiated with light which is absorbed by the α-MoTe₂ crystal. The spectral distribution of photovoltage (at 290 and 77°K) shows maxima near the α-MoTe₂ absorption bands and the properties of the photovoltage, in crystals 150 and 1 μm thick, suggest that it arises from the thermal dissociation of excitons.

The dark electrical conductivity in crystals 150 and 1 μm thick has been measured over the temperature range 77-290°K and activation energies of conduction obtained.

Models of randomly packed hard spheres exhibit some features of the properties of simple liquids, e.g. the packing density and the radial distribution. The value of the maximum packing density of spheres can be determined from models if care is taken to ensure random packing at the boundary surfaces and if correction is made for volume errors at the boundaries. Experiments for both the random `loose' and the random close-packed densities are reported with fraction one-eighth in. plexiglass, nylon and steel balls in air, and also with steel balls immersed in oil. A series of measurements for the random close-packed density has been made with up to 80 000 steel balls and with the aid of a mechanical vibrator. A computer analysis of the results permits a one-step, two-parameter extrapolation to infinite volume. The figure so obtained for the random close-packed density is 0·6366±0·0005, which represents an improvement in precision over previous results by an order of magnitude.

Pre-breakdown ionization currents have been measured in hydrogen and nitrogen by the Townsend method for E/p₀ in the ranges 27-46 and 54-99 v cm⁻¹ torr⁻¹. The ionization coefficients have been deduced by applying both the Townsend equation and also an equation derived by Lucas in which allowance is made for the loss of secondary electrons due to loss of photons out of the gap. A digital computer technique was used for the deduction of the coefficients and the comparison of the validity of the equations.

The degree of fit of the experimentally measured currents is better by the Lucas equation than by the Townsend equation over the region studied. The values of the electron ionization coefficient α in hydrogen as a function of E/p₀ were in excellent agreement with the values summarized by Rose in 1956 when the pre-breakdown currents were analysed by the Townsend equation. This confirms the accuracy of the experimental apparatus. When the same currents were analysed by using the Lucas equation the α values were lower by no more than 7%. The main advantage of the Lucas equation is its ability to isolate the γ_I and δ_p coefficients for ions and non-resonant photons, respectively, instead of the obscure generalized secondary coefficient of Townsend. γ_I decreased to zero with decreasing E/p₀, whilst δ_p increased with decreasing E/p₀ and showed a marked pressure dependence satisfying the Stern-Volmer formula which expresses the destruction of excited states in collision with ground-state molecules.

Similar results and conclusions were obtained for the values of the coefficients α, γ_I and δ_p for nitrogen.

The motion of a charged metal particle which is itself a source of field emission has been investigated. It will be shown that the effect of electron emission from a particle may significantly affect its behaviour and may even reverse its direction of motion in the electric field within a vacuum gap. Cathode-directed particles have their energies increased by this process, while the reverse is true for anode-directed particles. These findings are in accord with the results of other investigators who have suggested that particle impact on the cathode of a vacuum gap is most likely to initiate vacuum breakdown.

Expressions are derived for the brightness and directional beam intensity of an electron beam at a general point and direction in space, in terms of the tangential velocity of emission at the cathode, for Schottky emission and field emission. An expression for the directional beam intensity at the cross-over position of an electron gun is given, in terms of the magnification of the gun, the radius of the cathode tip, the potential at the cross-over and the temperature, or applied field, at the cathode.

The variation of the detection efficiency of electron multipliers used in the pulse-counting mode is investigated as a function of the average gain per dynode, the energy of the electrons being detected and the applied voltage. It is assumed that the pulse-height distributions can be obtained from Polya statistics and that the universal yield curve for secondary emission is applicable. Results are presented for 12-stage multipliers having values of the Polya parameter b=0·0, 0·5 and 1·0.

The method of data analysis described in this paper was developed for one-phonon resonances in the coherent inelastic scattering of neutrons from a crystal, but may be applied to any resonance with relatively large counting uncertainties. Simple formulae have been derived for the position of the mean and the width at half maximum of such a resonance, and for the standard errors in these quantities. The errors are objective and invariably smaller than errors estimated by inspection. It is emphasized that the formulae can be rigorously tested by applying them to artificial resonances generated in a computer by `randomizing' exactly known peaks, and a description is given of tests that have been carried out. The derivation of a line width from the width of an observed peak and the resolution width is also described for both Lorentzian and non-Lorentzian line shapes.

A stationary mould solidification technique is described which utilizes a fixed temperature gradient. This technique has made it possible to grow large magnesium-zinc alloy single crystals previously unobtainable owing to multiple nucleation of crystallites.

Desorption of ionically pumped argon from clean and metallized glass surfaces was investigated at different temperatures (between 0° and 40°C). The residence times were found to be longest on metallized glass surfaces. Activation energies of desorption and pre-exponential factors in the rate equation, obtained for two types of adsorption site in the above temperature range, were determined and found to follow the trend of the `compensation law'.

The use of a scanning electron microscope to examine the tracks left on cathode electrodes by arcs moving with a high transverse velocity has enabled estimates to be made of the sum area of coexisting emission sites. Correlation of such areas with the instantaneous values of arc current leads to current density values in the range 1-50 MA cm⁻², in agreement with results of other recent investigations.

Expressions are derived for the equivalent permittivity of a double-layer ellipsoid. When the ellipsoids degenerate into spheres the expression reduces to a form first derived by Maxwell for the equivalent conductivity of a double-layer sphere.

G. Szego showed that the sum of the three translational virtual masses along any three mutually perpendicular directions is an invariant for a solid moving in an infinite, incompressible, frictionless and irrotational fluid medium. A similar hypothesis is presented here, which shows, through a finite approximate analysis, that this sum is also an invariant when the fluid medium is finite, irrotational and friction-less but compressible, provided that the exciting frequencies are lower than the fundamental natural frequency of the medium. As an example this invariant has been obtained in the case of a thin flat plate for different depths of submersion and different frequency parameters by using the well-known method of `finite elements'.

In a recent paper P. Feltham has proposed a dislocation bowing model to describe the flow stress of body-centred cubic metals at low temperatures. It is shown here that the good agreement with experiment obtained by Feltham is due to an error in the calculation of the saddle-point configuration. The correct expressions for the activation parameters are derived, and it is seen that this model forms the basis of other, more sophisticated theories.