On the growth-sectorial dependence of defects in natural diamonds

Abstract

Certain rare diamonds have had epochs of growth during which they were bounded by two surface forms: normal {111} facets together with non­-flat surfaces of mean {100} orientation (‘cuboid’ surfaces). Differences of lattice imperfection content in octahedral and cuboid growth sectors (i. e. in regions that had crystallized respectively on {111} facets or on cuboid surfaces) have been studied by birefringence, ultramicroscopy, cathodoluminescence and various X-ray topographic techniques in­cluding X-ray topographic recording of the intensity of the anomalous ‘spike’ X-ray reflexions which are due to {100}-orientation-platelet pre­cipitates. Dissimilar manifestations of combined octahedral and cuboid growth are illustrated in the growth histories of three specimens in­ tensively examined. The sequence of relative development of octahedral and cuboid growth is found by mapping X-ray topographically both the differences in texture presented by the X-ray images of octahedral com­pared with cuboid growth material, and the diffraction contrast at growth-sector boundaries which arises from oherency strains of localized defects in the boundaries. The ratio of growth rates on {111} compared with cuboid surfaces may vary continuously and smoothly during a sub­ stantial part of the crystal’s growth history, or may have major or minor discontinuities. Particles ranging up to micrometre sizes are found dis­persed in cuboid growth sectors. They can be detected by the X-ray diffraction contrast they generate through straining the diamond matrix surrounding them. They may be resolved individually by ultramicroscopy (if the specimen shape is favourable), and by their individual diffraction-contrast images (provided that inter-particle distances exceed a few micro metres). In some parts of one specimen a dense sheet of particles coincides with the growth-sector boundaries at which adjacent cuboid growth sectors met, and a zone about 20 μm thick on each side of the boundary is depleted of particles. As to whether the bodies are precipitates or inclusions the X-ray topographs show no instances of ‘grown-in’ dislocations originating at the particles (such as are often observed at inclusions of micrometre sizes and greater which have been trapped at growing crystal surfaces), and the few dislocations observed which issue from regions dense in diffraction-contrast-producing bodies are themselves ‘decorated' by apparently similar bodies. These observations favour precipitation, but do not exclude the possibility that the bodies are entirely inclusions. They also admit the possibility that the present size of the bodies was attained through post-growth segregation upon smaller inclusions. Cuboid-growth material not distorted by the diffraction-contrast-producing bodies can exhibit a low integrated Bragg reflexion, indicating very good lattice perfection. Octahedral-growth material Bragg reflects more strongly (sometimes markedly so) indicating unresolved lattice imperfections finely distributed. Octahedral growth sectors show strong ‘spike’ reflexions in comparison with cuboid sectors; and where the diamond matrix integrated reflexion is lowest, ‘spike’ reflexion is weakest. The ‘spike’ topographs show that the density of {100}-orientation platelets (presumably largely nitrogen impurity) is less (sometimes much less) in cuboid than in octahedral-growth material of similar epoch. It is suggested that {111} facets of diamond act as sinks for nitrogen present in the surrounding growth medium.