The behaviour of a single crystal of antimony subjected to alternating torsional stresses

Previous experiments on the failure by fatigue of single crystals of aluminium, iron and zinc, representing the face-centred cubic, the body-centred cubic, and the close-packed hexagonal lattices, respectively, have shown that failure of metallic single crystals tends to occur by slip on the plane of greatest atomic density in the direction of greatest (linear) atomic density. The results obtained with iron seemed to indicate that of the two factors, the linear density is the more important. In all three lattices, however, the line of greatest density lay in the plane of greatest density, so that slip in the direction of the line of greatest density could always occur on the plane of greatest density and definite differentiation between the two factors was not possible. The structure of antimony (and also of bismuth), however, is such that the planes of maximum density do not contain any of the lines of maximum density, so that if the type of the slip plane were determined, definite evidence of the relative impor­tance of the two factors would be obtained. The present experiment was designed to yield this evidence; but in so far as the results are inconclusive, it is hoped to obtain further evidence by a similar experiment on a single crystal of bismuth. Lattice Structure.—The lattice structure of antimony as determined by A. Ogg (‘Phil. Mag.,’ vol. 42, p. 163 (1921)) and by James and Tunstall (‘Phil. Mag.,’ vol. 40, p. 233 (1920)) is a lattice of trigonal symmetry composed of two similar face-centred rhombohedral lattices, similarly orientated, displaced relative to each other along the longest diagonal of the rhombohedron (the axis of trigonal symmetry). The angle between any pair of edges of the rhombo­hedron is 86° 58' and the atoms are spaced along these edges at points 6·18 Å. apart. The ratio of the lengths into which the lattice points of either con­stituent lattice divide the long diagonals of the other lattice is given as 0·412 : 0·588 by Ogg and as 0·389 : 0·611 by James and Tunstall. For the purpose of the present report, the exact value of this ratio is of little importance; but where some value has to be inserted (e. g., in fig. 1) the value 0·4 : 0·6 has for convenience been assumed.