July 14, 1923] 



NA TURE 



59 



remove work-hardness and render the crystals equiaxed, 

 were subjected to degrees of tensile strain varying from 

 two to ten per cent, extension on three inches of the 

 parallel portion of the testpiece. After this they were all 

 heated to 550° C. and kept thus for six hours. Finally, 

 they were etched in a ten per cent, solution of sodium 

 hydrate. It will be seen that the crystals in the 

 testpiece extended two per cent, are very coarse, and 

 that as the strain is increased the crystal size diminishes 

 until at ten per cent, it has become quite small. But 

 there is a further point to be noted, namely, that in 

 all the eight cases shown, large crystals have also 

 formed in the broad heads of the testpieces, where 

 the strain must have been less. 



The problem which we set ourselves was to convert 

 the crystals, numbering about 1,687,000, in the parallel 

 portion of a testpiece 4 in. x i in. x 0-125 i"-j! i^to a 

 single crystal. Three treatments, two thermal and 

 one mechanical, are necessary. The testpiece in the 

 original condition was cold-rolled, and as a result the 

 crystals were very much elongated and worked into 

 one another. It had first to be heated so that it might 

 be completely softened and new equiaxed crystals of 

 approximately uniform size produced. The most 

 suitable temperature was found to be 550° C. and the 

 time six hours. It had next to be strained to the 

 required amount, which was equivalent to a tensile 

 strain of 2-4 tons per square inch. Finally, it had 

 to be heated so that the potentiality of growth con- 

 ferred by strain could be brought fully into operation. 

 This was a somewhat lengthy operation, and involved 

 a heat treatment beginning at 450° C. and finishing 

 at 600° C. over a period of about 100 hours. After 

 these treatments, on an average about one testpiece 

 in four is converted into a single crystal over the 

 parallel portion. Sometimes this space is occupied by 

 two, three, or even four crystals, but never by more 

 than that. 



The production of these very large crystals has 

 enabled us to determine the tensile properties of 

 single crystals and compare them with those of the 

 aggregates of minute crystals of which such bars are 

 usually composed. In the latter case very uniform 

 results are obtained, the ultimate stress varying from 

 4-5 to 4*7 tons per square inch, and the percentage 

 extension on three inches being from 36 to 38. The 

 values obtained in tests of specimens consisting of 

 single crystals varied, however, from 2-80 to 4-08 tons 

 per square inch, while the extension varied from 34 

 to 86 per cent, measured on three inches. These 

 variations in properties were accompanied by differ- 

 ences in the method of stretching and the types of frac- 

 ture which hav^e provided a means of classifying them. 



Speaking broadly, five types may be distinguished. 

 In certain cases the testpieces narrowed in breadth 

 gradually from the shoulders towards the fracture, 

 and the metal necked sometimes almost to a point. 

 In other cases the testpiece remained broad, losing 

 sometimes only one per cent, in breadth, but became 

 very thin. In the third case the testpiece both 

 narrowed and thinned uniformly, and a noticeable 

 feature of this type is the sloping of the sides, so that 

 the section after pulling is no longer a right-angled 

 parallelogram but one with acute and obtuse angles. 

 Slip bands were usually well marked, and were inclined 



NO. 2802, VOL. I I 2] 



to the axis at different angles. In the fourth type 

 the testpieces not only narrowed and thinned but in 

 addition necked at the fracture, and in all cases a 

 sideways slip was evident. In the fifth type may be 

 included all the testpieces which produced twin crystals 

 on being pulled. No signs of these were visible before 

 stress was applied. In some cases only a few twins 

 resulted, while in others the testpiece was twinned 

 all over. In every case the testpiece buckled and 

 crumpled to a certain extent, owing to the shifting 

 of portions of the sheet into a twinning position. 

 These differences in the method of distortion and 

 fracture are due to differences in the original orientation 

 of the crystal in the testpiece. 



Monocrystalline testpieces were also prepared in 

 round bars of diameter 0*564 and 0*798 of an inch 



Fig. 2. — Fractured testpieces of single crystals in round bars, showing how 

 in each case the bar draws down in one dimension and produces 

 a wedge-shaped (double-grooved) fracture. By permission of the 

 Institute of Metals. 



respectively. The deformation of these testpieces 

 under tensile stress was very remarkable, and deserves 

 special mention. On one hand, a bar consisting 

 of the usual aggregate of small crystals drew down 

 with a roughening of the surface, the maintenance of 

 a circular cross section, and a cup-and-cone fracture. 

 On the other hand, the single crystals flattened very 

 much in one dimension, whereas the other dimension 

 differed but little from the original diameter of the 

 bar, and the end result was not a cup-and-cone fracture 

 but a double groove. The bar when subjected to 

 tensile stress slipped principally on one plane, which 

 subsequent investigations by Mr. G. I. Taylor and 

 Miss C. F. Elam have shown to be an octahedral 

 plane. When it began to break it drew down sharply 

 in the same direction in which it had thinned, and 

 a lens-shaped area was formed. As the bar pulled 

 apart this became flatter and flatter ; it parted first 

 at each side and then in the middle. The final result 

 was a curious double-grooved fracture with flow lines. 

 Fig. 2 shows the fractured testpieces of five single 



