Table of contents

The information from which the optical constants are determined is produced on a photographic plate as a vertical band of variable intensity for each wavelength. This is compared microphotometrically with an adjacent band of graded intensity produced by the same wavelength.

Two photoelectric methods of analysis of elliptically polarized light have been devised. One is a development of that of Kent and Lawson in which a rotating analyser yields a zero signal only when the light is circularly polarized We use a double image prism after a rotating analyser and the two resulting modulated beams are in antiphase only when the axes of the double image prism coincide with the axes of the vibration ellipse. In our second method, the light is modulated by chopping, and a double image prism selects two components in phase In principle these are equal only when the axes of the double image prism are at 45° to the axes of the vibration ellipse Thus the directions of the major and minor axes are found, and if the outputs from the two photocells which receive the two components are balanced on an electrical bridge the axial ratio may be obtained. The double image prism plays here the part of the half shadow device in visual observations, but as its half shadow angle is 90° instead of only a few degrees, the photocells are used under optimum conditions, i.e the signal-to-noise ratio is maintained as large as possible and the potential accuracy of the method greatly exceeds the limitations imposed by the rest of the apparatus. By slight modifications the method may be made partially self-recording.

Some examples are given of measurements made on metallic films undergoing surface changes.

Noble metals show absorption in the blue or ultra-violet; to this is due the colour of copper and gold. As was pointed out by Mott, the main absorption seems due to internal photoelectric effect, with excitation of a d electron. This absorption could be analysed - as for the alkalis - as a transition between two bands, described by Bloch orbitals, here d and s. But we must take into account the correlation between the hole created in the nd shell and the Fermi electrons. We can assume that the hole is screened by a (n+1)s electron bound to it, so that absorption should begin for the frequency of formation of the excited nd⁹(n+1)s configuration of the monovalent ion. Agreement with experiment obtained using this model is satisfactory both in the pure metals and their alloys The same model may be used for x-rays, and optical absorption energy is related to the energy separation between the s and d parts of K and L_II,III emission bands An order of magnitude for the van der Waals interaction energy between d shells could perhaps be deduced from the structure on the short wavelength side of the absorption peak.

The optical properties of evaporated films of silver, gold, tin and indium have been determined by a photoelectric method The influence of factors concerned with the evaporation procedure has been investigated in the case of the silver films. The experimental results are accounted for by the agglomerated state of the films, and the application of Maxwell Garnett's theory is shown to lead to qualitative agreement with the experimental values.

When the diameter of the focal spot of electrons on an X-ray target is reduced to the order of a micron, the conditions of heat dissipation become so favourable that a greatly increased specific loading is possible. The use of an electron lens to obtain such a small spot introduces a limitation on the rate of supply of energy, as its high spherical aberration requires insertion of a very small aperture, the size of which must be reduced pari passu with the size of spot desired. For magnetic lenses of usual design the rate of energy supply is then proportional to the radius of the spot. Comparison with the rate of heat dissipation shows that, for spots of the order of a micron, the limitation on X-ray intensity is not the thermal properties of the target, as in tubes with large focus, but the emission obtainable from the cathode. Specific loadings of the order of 10⁷ watts cm^-2 can be tolerated in a spot of 1 micron; the corresponding emission current density would be of the order of 100 amp cm^-2.

The optical properties of liquid selenium in the near infra-red have been measured for a temperature range up to about 400° C. The absorption edge moves out to longer wavelengths as the temperature increases by an amount corresponding to a change of energy gap of 1.4×10^-3 eV/C°. Absorption is very small at low frequencies. For long wavelengths the refractive index decreases as the temperature increases by an amount corresponding to the thermal expansion.

Two methods have been used to detect contamination in evaporated films of silver and germanium by the boats or filaments used as sources. It is shown that a limit of a few parts in ten thousand of tungsten or molybdenum is attainable by microchemical methods. The much lower limit of one or two parts in 10⁷ is attained by the radioactive tracer method. The extent of contamination of films of silver and germanium by tungsten and tantalum boats has been studied using the tracer technique and the minimum contamination attainable found to be of the order of a few parts per million. The possibilities of transfer of the boat material by direct evaporation or by reaction with the residual gas in the system have been minimized; the results suggest that contamination arises by the solution of the boat materials by the molten silver or germanium.

An analytical determination of the self-absorption of a spectral line emitted by a source is not normally possible if account is taken of the spatial variations in the source of the excited atom concentrations and spectral line profiles. The problem is simpler for uniformly excited sources, for example sources in thermal equilibrium, but not for high-pressure arc sources normally assumed to be in thermal equilibrium, because of the existence of temperature gradients. In the paper an analysis is developed which shows that if such arcs are treated as uniformly excited, an error of less than 5% is introduced into the determination of the self-absorption of spectral lines, emitted by transitions between an upper state s and a lower absorbing state t, which satisfy the limits hν_sq/kT₀ gt-or-equal, slanted 10 and ν_st/ν_tq less-than-or-eq, slant 0.5, where q is the ground state and T₀ the axial source temperature. The theory is then applied to the experimental data for a high-pressure mercury arc discharge and to new measurements of the line widths at 5461 and 4358 Å. It is shown that the calculated value of the self-absorption at 5461 Å agrees with the experimental value only when the line has a resonance profile The existence of this form of profile is also indicated by considerations of the broadening processes in the discharge

The calculation of ring shims of the kind suggested by Rose has been extended to include polecaps in the form of truncated cones Design data are presented in graphical form Actual and predicted performances are compared for one magnet.

New observations of growth and etch features on haematite (Fe₂O₃) crystals are reported Of the many growth patterns observed, the simplest are the elementary ` molecular growth spirals ' in accordance with the theory of Burton, Cabrera and Frank. The step height between the successive arms of an elementary spiral has been measured by the application of multiple beam interference fringes and is shown to be equal to the x-ray value of the repeat distance. More complex growth patterns are illustrated and discussed. The etch figures reported are also of molecular thicknesses and consist of a very large number (~ 10⁷/cm² of the crystal surface) of similarly oriented small triangular figures, and some more complex shapes.

The concept of a kink-band is introduced It is a thin plate of sheared material in a crystal, transverse to a slip direction, bounded by opposite ` tilt walls ' of dislocations. It is shown that if the angle of kinking shear γ exceeds a critical value (estimated to be 3½°) the stress concentration at the edges of the band is sufficient for the creation of new dislocation pairs or loops, and that if the applied stress exceeds a certain value, inversely proportional to the square root of the diameter of the kink-band, the kink-band will grow by this process. The stress field of a kink-band is similar to that of a slipband, a plate of deformation twin, or a crack under tangential stress. In suitable circumstances these may act as sources for each other. The successive generation of kink-bands, inside each other, from the same source, will produce a kink of the kind experimentally observed.