312 



NATURE 



[November 4, 1920 



Crystal Growth and Recrystallisation in Metals. 



'■"PHIi effects of heat on certain cast and plastic- 

 A ally deformed metals have been studied in 

 considerable detail by Prof. H. C. H. Carpenter 

 and Miss C. F. Elam in an investigation which 

 was published and discussed at the autumn meet- 

 ing of the Institute of Metals held at Barrow-in- 

 Furness under the presidency of Vice-Admiral Sir 

 George Goodwin. 



At the outset the authors state that the terms 

 "recrystallisation" and "crystal growth," which 

 signify quite different phenomena, have previously 

 been used for the most part indiscriminately and 

 interchangeably, with the result that the discussion 

 of this subject has necessarily been confused and 

 unsatisfactory. By "recrystallisation " is meant 

 the complete reorientation of a crystal or group 

 of crystals. The nevi' arrangement starts from 

 new centres, and is quite independent of the old 

 system of orientation. It is, in fact, the birth of 

 new, differently oriented crystals in a crystal 

 aggregate, and the gradual change of the old to 

 conform with the new. This always gives, in the 

 first instance, a refined structure. When all trace 

 of the old arrangement has disappeared it is con- 

 sidered that recrystallisation is complete. By 

 crystal growth is meant the rearrangement of 

 certain crystals in a crystal aggregate to conform 

 with the orientation of certain other crystals, 

 during which process the latter increase in size 

 by the addition of reoriented material at the same 

 time as the former decrease in size by the same 

 amount. This process necessarily leads to an in- 

 crease in crystal size. 



Many of the authors' experiments have been 

 carried out with an alloy of tin and antimony con- 

 taining about 15 per cent, of the latter metal. 

 This alloy possesses a very peculiar property, 

 which makes it of special value for investigating 

 the above phenomena, and for the first time has 

 enabled the stages of crystal growth in a metallic 

 complex to be studied experimentally. It pre- 

 sented, however, some special difficulties in polish- 

 ing and etching. Cutting, filing, and even grinding 

 on the finer emery papers bring about spon- 

 taneous recrystallisation of the surface layer. This 

 entirely obscures the genuine structure of the 

 specimen, and can be removed only by alternate 

 polishing and etching. Ammonium sulphide solu- 

 tion is the best reagent for developing the struc- 

 ture, and by alternately immersing the specimen 

 in the solution and rubbing it on selvyt or chamois 

 leather with magnesia moistened with ammonium 

 sulphide a very beautiful surface is obtained. This 

 reagent attacks different crystals to very different 

 degrees, so that some appear white, and others 

 black, under the microscope. This proved of con- 

 siderable assistance in the experiments, as it made 

 it easy to identify a particular area under exam- 

 ination. If a polished and etched specimen was 

 heated at i5o°-20o° C, and growth of any of the 

 crystals took place, the position of the new 



NO. 2662, VOL. 106] . 



boundary was marked by a line which is really 

 a difference of level almost as if the specimen had 

 been etched ; but there was no visible change in 

 the surface of the reoriented area. 



That these lines represent the position of the 

 new boundaries after heating could be shown by 

 (i) taking a photograph of an area which included 

 some of them ; (2) repolishing and re-etching the 

 specimen and rephotographing the same area. On 

 comparing the two photographs the boundaries of 

 the crystals as shown by the etching in {2) corre- 

 spond exactly with the new lines in (i). It was 

 always a simple matter to distinguish between a 

 growing crystal and one that was being grown 

 into. The alloy tarnishes on heating, being first 

 yellow, then orange, red, purple, blue, and green 

 in turn. Just as the crystals etch differently, so 

 they tarnish differently ; but they tarnish first in 

 accordance with the original etch. There is 

 always a colour contrast between adjacent crystals, 

 and the colour fixes the original boundaries. If 

 new boundaries appear iniiJe this colour boundary, 

 it shows that the adjacent crystal is growing into 

 it ; if outside, it shows that the crystal is growing 

 into its neighbour. 



A very remarkable feature about these boundary 

 markings is that they are formed only when the 

 specimen is removed from the furnace and cooled. 

 If a specimen is heated and cooled three times, 

 there will be three new boundaries round some 

 of the crystals ; if four times, there will be four, 

 and so on. But if another specimen be heated 

 along with the first, and taken out only when 

 the former was removed for the fourth time, one 

 new boundary marking alone would be visible. It 

 is, therefore, an arrest in the progress of growth 

 which makes these markings on the surface, and 

 further heating does not obliterate them. There 

 is, however, no sign of them after polishing and 

 etching, only the final and genuine crystal boundary 

 being then visible. 



The authors publish numerous photographs illus- 

 trating these phenomena, and the accompanying 

 series revealing the gradual decrease in size and 

 final disappearance of a crystal in four stages 

 is taken from a plate in their paper. The 

 crystals concerned are represented by letters 

 in Fig. I. Fig. 2 shows the effect of heating once. 

 The crystal A has been invaded by crystals C, D, 

 and E. Crystal C continues to grow on further 

 heating, and also crystal B to a small extent 

 (Figs. "3 and 4). The state of things after heat- 

 ing four times is shown in Fig. 5. All four 

 crystals have invaded A, arid C and E have met. 

 Fig. 6 is of the same area repolished and re- 

 etched, and shows that A no longer exists. The 

 orientations of crystals B, C, and D as revealed 

 by the etching are very different from that of A. 

 It is possible that the orientations of A and E 

 do not differ greatly, yet each crystal has takfen 

 a considerable share of A. 



