Table of contents

Radiation damage of biological specimens in the transmission electron microscope is commonly investigated by electron diffraction. Since this method has several drawbacks, it is proposed that additional information should be used by computer evaluation of successive images taken under conditions of low exposure. A measure of dissimilarity and a radial distribution function in Fourier space are defined, which can be used to characterize the history of structural changes caused by radiation damage in a way that is independent of the particular features of the object.

Photoelastic measurements of stresses in the vicinity of holes drilled in fused quartz with a CW CO₂ laser are reported. Stresses exist over a volume whose dimensions are defined by the thermal diffusion length for the heat pulse used. Incipient devitrification was noted in the vicinity of the hole, with large stresses often leading to fracture present at the boundary of the fused and devitrified phases.

The techniques available for oxygen diffusion measurements are summarized and the usefulness of using nuclear techniques is illustrated.

In the present work, the mass-18 isotope of oxygen is used as the diffusing species and its concentration distribution is subsequently explored using the 1·763 MeV resonance in the ¹⁸O(p, α)¹⁵N reaction. The yield from the resonance gives a measure of ¹⁸O concentration, and depth is explored by measuring at different energies of the incident protons. Variations in the technique have been used. Various methods of analysing the data have been explored and the preferred one is that which makes allowance for the various spreading factors which contribute to the observed α-particle spectrum. Measurements are quoted for the ¹⁸O self-diffusion coefficient in a TiO₂ specimen showing that they have good reproducibility and are consistent with those of other observers.

The present understanding of the influence of structural variables on the tensile flow strength and work-hardening behaviour of cubic metals is outlined in terms of the separate effects of crystal structure and orientation, stacking-fault energy, grain size, solution hardening, and dispersed-particle strengthening. Typical forms of uniaxial tensile stress/strain relationships obtained with different structural conditions are compared with commonly used empirical descriptions of the relationships which assume power-law work hardening. It is shown that the simple empirical equations cannot be relied upon to give precise representation of the stress/strain relationships of a wide variety of alloys because there are too many independent factors which can influence flow strength and the way in which work hardening develops. Nevertheless, simplified descriptions of work-hardening behaviour can be very useful when applied with discrimination. In this connection, knowledge of the underlying relationships between microstructure and work-hardening behaviour is helpful in guiding application.

Conventional strength and strain-hardening parameters have been derived for idealized true-stress/true-strain curves obeying the Hollomon equation σ=K_pⁿ, where K and n have values typical of real metals. All stress parameters are proportional to the constant K. The true tensile strength is almost independent of n, but the stress at 0·2% plastic strain is strongly dependent on n. The strain-hardening rate dσ/d_p is significantly affected by n only when _p<0·01; then dσ/d_p increases with increasing n when n<0·2 and decreases with increasing n when n>0·2. The strain-hardening rate is not easily related to the parameters nσ_max, nK or the 0·2% proof-stress/true-tensile-strength ratio. The magnitude of the strain hardening, given by Δσ = (σ₂ - σ₁), also has a maximum between n=0·1 and 0·3. With these results, assumptions and conclusions in the published literature are discussed and some are shown to be incorrect. It is concluded that for maximum strength and strain hardening in materials obeying the Hollomon equation, large values of K and n values between 0·1 and 0·3 are required.

Perhaps the area in which a detailed knowledge of stress/strain relations is most important is that of sheet forming. It is true that the precise form of the stress/strain equation should be known to assist analyses of sheet-forming processes. However, in the present context, the proposition is that such data can be obtained from a tensile test. This contribution examines that proposition especially under the circumstance likely to prevail in sheet forming that the material is anisotropic. The anisotropy may be crystallographic in origin or may otherwise arise from the pre-strain history. It is concluded that the differences between any of the empirical stress/strain equations represent the least significant of the errors in describing the actual stress/strain behaviour of an element of material under sheet-forming conditions. The importance of considering the total states of stress and strain is illustrated by analysing an example from the literature due to Richards.

The regime of sheet metal deformation extends from pure shear, ₁= - ₂, through plane strain, ₂=0, to balanced biaxial tension, under which ₁=₂. During deformation a material passes into different states, termed (1) uniform straining, (2) diffuse straining, and (3) localized straining, respectively. Depending on the strain ratio, (1) alone, (1) & (2) & (3), (1) & (3) or (1) & (2) can be experienced by a sheet between zero strain and fracture. In this work the deformation behaviour, in tensile and hydraulic bulge testing, of steel, aluminium and 70-30 brass has been studied. The onset of state (3) has been correlated with strain measurements, taken from a cine film of the deformation, indicated by a grid electrochemically marked on the sheet surface. Various theories predict the major and minor strains at the initiation of these states, and the theoretical and experimental results are compared for the various materials; also the implications with respect to forming-limit diagrams are discussed.

Measurements have been made using a flash thermal diffusivity technique from 550 to 2500 K on a range of well-characterized materials. Data are reported on specimens of five different porosity levels, samples in which porosity has been redistributed, and specimens, initially of two different porosity levels, oxidized to three different hyperstoichiometries. The data are discussed in the light of current theories, and it is shown that the porosity correction term is temperature-dependent. The conductivity λ_P at any porosity P over the temperature (T) range 500-1600 K is given by λ_P (W m⁻¹ K⁻¹) = [0·0433+0·355P + (2·01+1·47P) × 10⁻⁴T]⁻¹. This temperature-dependent porosity correction is shown to be related to total porosity, and it is suggested that this is due to additional vacancy scattering. An expression is derived for the additional phonon scattering due to interstitial oxygen atoms up to small hyperstoichiometry levels. Finally, the deviation from linearity at high temperatures is shown to be due to an electronic contribution to the total thermal conductivity.

An exact expression is obtained for the principal dielectric constants of a macroscopically homogeneous and axially isotropic two-phase composite material. This is compared with the result obtained in the isotropic case. Upper and lower bounds are given. They are shown to be attainable and are found to depend upon the direction of the applied field. Approximate power-law formulae are derived and their applicability discussed in terms of the internal geometry of the sample. It is suggested that these approximate formulae may be usefully applied to polymer-polymer mixtures. It is further shown that except for the Looyenga third-power formula they are not applicable to isotropic samples.

Equations given by Hirsch and Humphreys (1970) and by Spencer et al (1972) for describing the contrast in electron-channelling patterns (ECP) are generalized by taking into account the energy losses which the fast electrons experience in the specimen. In this way a theory is established which considers diffraction, backscattering and energy losses. This theory is used and calculations are made of the dependence of the contrast in ECP on the energy of the backscattered electrons. Unlike earlier theories for ECP the present work gives the correct behaviour of the backscattering coefficient for large foil thicknesses. It is shown that the visibility can be strongly enhanced by using an energy-analysing detector, in agreement with experimental results. The signal-to-noise ratio is evaluated to find the energy window which gives optimum contrast. For Si, Cu and Au these windows are approximately E₀ > E > E₀ − 700eV, E₀ > E > E₀ − 250 eV and E₀ > E > E₀ − 400 eV respectively. Calculations of the mean escape depth as a function of energy show that the electrons are emitted from a thin surface layer for small energy losses. The resolution can therefore be significantly improved in the backscattered mode by using an energy-analysing detector.

The magnetic properties of 3% grain-oriented silicon-iron were measured when the material was subjected to circular bending. Loss per cycle, magnetization waveforms and the visual observation of surface domains were recorded at increasing curvatures. Surface observation indicated a break-up of the parallel domain pattern when the material was subjected to a longitudinal compressive stress of greater than about 5 MN m⁻², while the domain pattern under tension remained virtually unchanged. A rise in both loss per cycle and magnetizing force was measured for increasing curvature, and this was considered to be the result of the change in domain structure due to compression. The magnetization loops confirmed the existence of freely moving walls on the tension side, while the re-orientation of the transverse structure due to compression required a relatively large magnetizing force.

A detailed analysis of the non-reciprocal transmission property of artificial dielectrics containing ferromagnetic metal particles has been made assuming a tensor permeability which involves the Landau-Lifshiftz form of damping for these particles. Corrections have also been applied to the results for the case when the absorption curve departs from lorentzian lineshape due to a field-dependent loss term. It is shown that the optimum non-reciprocal behaviour for these materials can be realized using metals with high saturation magnetization and low ferromagnetic resonance loss. A steady Faraday rotation as high as 32 deg cm⁻¹ at an average figure of merit of 6 deg dB⁻¹ may be realized over the entire X-band with a suitable choice of particle size. The ratio of the attenuation constants appropriate for the two circular polarizations is found to have a maximum value of 11 for metals like nickel, permalloy, etc which have a relaxation time of the order of 10⁻⁹ s. A higher value of the ratio would be feasible for metals with a larger relaxation time.

The dependence on the excitation dose of the maximum thermoluminescence intensity as well as of the peak temperature are investigated theoretically. It is shown that certain irregularities of the dose dependences can be explained by assuming the existence of a trapping level, the transition into which competes with the retrapping and recombination of the free carriers. By numerical solution of the appropriate equations, it is demonstrated that the maximum thermoluminescence intensity may depend super-linearly on the excitation dose. The power of the dose dependence was found to be 2 under certain circumstances at low doses, and reached even higher values before saturation of the competing level. The maximum temperature sometimes behaved in an unusual way; namely, it increased with increasing dose. The relation between the area under a glow peak and its maximum intensity is also studied; it is shown that the latter can usually serve as a measure for the former. This finding is of practical importance, especially in thermoluminescent dosimetry, since the evaluation of the maximum intensity is obviously more convenient than that of the area.