Search

Three large (4–7 mm) natural diamonds, each half brown and half white, from the Argyle Diamond Mine in Western Australia, have been studied for growth history and crystal perfection. Scanning electron microscopy and X-ray single-crystal topography showed the diamonds to have good octahedral morphologies but poor internal perfection. X-ray double-crystal topography quantified this lack of perfection, rocking-curve widths taking unusually large values of 300 and 75 s of arc for the brown and white regions, respectively, in a diffraction geometry for which 3 s of arc would be expected for a perfect diamond. Fourier transform infrared spectroscopy revealed significant differences in nitrogen concentration between the brown and white regions of the diamonds. The white regions, with 400 to 600 p.p.m. nitrogen, contained over 1.6 times more nitrogen than the brown regions. It is concluded that the extra nitrogen (in A and B forms) in the white regions stiffens the lattice against distortion by natural plastic deformation.

The periodicities, profiles and contrast of stacking-fault fringes observed in high-resolution synchrotron X-ray section topographs have been comprehensively studied with controlled changes of diffraction parameters such as wavelength and polarization mode, which allowed observations under diffraction conditions inaccessible with conventional sources of characteristic radiation. Comparisons with simulated patterns and computed fringe profiles are presented. The fringe patterns were generated by a single intrinsic stacking fault of area 1.3 mm2 contained in a polished parallelepiped specimen of natural diamond unusually free from lattice defects producing long-range strains. The principal X-ray wavelengths employed were 0.08, 0.10, 0.12 and 0.15 nm, providing values of μ0t0 ranging from 0.25 to 1.5 (μ0 = normal photoelectric absorption coefficient, t0 = specimen thickness). Fringe orders up to about 100 were observed at the longer wavelengths. Controlled fractional changes of Pendellosung interference order were obtained by wavelength fine-tuning (Δλ ≃ 10−3 nm). Matching of simulations with the resultant patterns is discussed in detail. Fringe patterns were recorded with pure σ, pure π and mixed polarization modes. Change of polarization mode was employed in order to vary the ratio of anomalous to normal absorption without having to change other diffraction variables. Particular attention was devoted to exploring the range of diffraction conditions under which fault vector sign could be experimentally determined from the sign of `first-fringe' contrast, and to examining how this range can be extended with the aid of comparison with pattern simulations.

Computer-simulated X-ray section topograph images of a specimen containing a single large-area stacking fault are compared with the experimental image to illustrate useful improvements in simulation realism with regard to both image geometry and stacking-fault fringe contrast. The computer-program developments include exact treatment of the X-ray polarization factor so that anomalous transmission and the periodic fadings of fringe visibility occurring when unpolarized or partially polarized sources are used can be correctly computed. A procedure is described for photographing a cathode-ray-tube display to obtain records with acceptable resolution, contrast and freedom from distortion.

The intrabranch interference effects expected under Laue-Bragg diffraction conditions [Saka, Katagawa & Kato (1972). Acta Cryst. A28, 102-113, 113-120] and first studied experimentally under high-absorption conditions by Borrmann & Lehmann [Crystallography and Crystal Perfection (1963), edited by