NATURE 



[June 24, 1922 



fore^ had to be kept not above 1425° C, which is near 

 the delta-gamma transformation point. Moreover, it 

 was found necessary to counteract the influence of 

 grain growth at these very high temperatures. For 

 this purpose the temperature was varied every 5 or 

 ID minutes so as to bring the iron into the range of 

 another modification, e.g. in taking the beta iron 

 photogram the temperature was raised from time to 

 time to about 1000° C, and when the gamma and delta 

 irons were investigated the heating current was 

 broken off now and then, so as to let the wire cool 

 down into the alpha range. Photograms of delta iron, 

 which always contained some gamma iron, show, 

 accordingly, lines characteristic both of the delta and 

 gamma varieties. 



Experiments carried out on the above lines show that 

 alpha, beta, and delta irons have a body-centred cubic 

 lattice, whereas gamma iron has a face-centred cubic 

 lattice. The former lattice contains an atom at each 

 corner of the cube and one in the centre, the latter 

 has an atom at each corner and also in the centre of 

 each face. It follows, therefore, that although there are 

 generally considered to be four modifications of iron, 

 three of them possess one and the same crystal form. 

 For alpha iron at the ordinary temperature a has been 

 found to be 2*87 A.U. ; at 800° C. it has increased to 

 2-90 A.U., and at 1425° C. to 2-93 A.U. This agrees 

 very well with the known dilatation coefficient of alpha 

 iron. The heat expansion of gamma iron also manifests 

 itself in the increase of alpha-gamma from 3-63 A.U. 

 at 1100° C. to 3'68 A.U. at 1425° C, which agrees well 

 with the fact that the coefficient of gamma iron is 

 greater than that of alpha iron. 



At first sight it appears rather startling that the 

 transformation which takes place at A3 (beta to gamma) 

 is reversed at the higher temperature A4 (gamma to 

 delta). However, the diagrams of Weiss and Foex 

 which the author reproduces, showing the change of 

 magnetic susceptibilities of iron, indicate that alpha, 

 beta and delta iron probably possess one and the 

 same structure. Evidence as to the existence of delta 

 iron has gradually been accumulating during recent 

 years, and this modification can now be regarded as 

 being well established. Henceforth, it must take its 

 place in the iron-carbon equilibrium diagram. 



Great interest attaches to the authors' experiments 

 on the influence of carbon on the space-lattice of iron 

 in hardened steels (Fig. 2). These show that the gamma- 



iron lattice of austenitic steels is enlarged by dissolved 

 carbon. A steel with 1-98 per cent, carbon has been 

 found to have a somewhat larger lattice when quenched 

 from 1100° C. than from 1000° C. The investigations 

 thus show that carbon has a distorting influence on 



the lattice, and a further interesting point concerning: 

 this effect may be noted. The carbon atoms may be 

 situated in the cavities of the iron lattice, but the 

 author concludes that they are distributed quite 

 irregularly in the austenite crystals. If their deform- 

 ing influence were of a local nature, the interference 

 radiation would be diffuse. The lines of the photo- 

 grams of the austenite containing high carbon are,, 

 however, very distinct and clear, thus proving that 

 the iron lattice is uniformly deformed. 



Photograms of martensite, the characteristic con- 

 stituent of hardened steel, show three very faint and 

 diffuse lines, but their cloudiness makes it difficult 

 exactly to determine the position of their intensity 

 maxima. The a-values of the alpha iron in martensite 

 are not, therefore, so trustworthy as those of gamma 

 iron in austenite, but the various modifications of iron 

 in all the four photograms published have given the 

 same value, namely, 2-90 A.U., which indicates that the 

 alpha iron lattice of martensite is likewise enlarged by 

 the carbon atoms present. On the basis of these results, 

 the authors discuss the deeply interesting question, 

 whether martensite is a two-phase system or a homo- 

 geneous solid atom disperse solution. If the iron 

 lattice is uniformly deformed, it seems probable that 

 martensite, like austenite, is a true solid solution of 

 carbon of iron. If, however, the photogram of 

 martensite is identical with that of pure alpha iron, it 

 would indicate that the hardened steel contains a mass 

 of alpha iron particles, free from carbon. The photo- 

 grams thus far obtained point in the direction of the 

 first of these possibilities, and it seems probable that 

 martensite is a real atom disperse solution. The diffuse- 

 ness of the line in the photograms gives very important 

 information as to the structure of this constituent. 

 As Scherrer shows, the lines of a Debye-Scherrer 

 photogram get broader and more diffuse in proportion 

 as the crystal powder is more finely divided. No 

 quantitative comparisons have yet been made for 

 martensite, but from a qualitative comparison, of a 

 film of steel (o-8o per cent of carbon), with the photo- 

 gram of an extremely fine-grained gold colloid, the 

 authors conclude that the steel is as highly disperse 

 as the colloid. The lines of the martensite seem to be 

 about as broad as those in the gold photogram. The 

 ranges of homogeneous lattice in the steel have accord- 

 ingly, on an average, an extension of about 20 A.U., 

 and each of them contains only a few hundred atoms. 

 In the concluding section of their 

 paper the authors publish photograms 

 of the iron carbide, cementite, FcgC. 

 This they have found to be identical 

 with the well-known crystal plates of 

 speigel iron. By means of a Laue 

 photogram and investigations of an 

 orientated rotating crystal of speigel, 

 it has been possible to deduce the 

 crystal data of cementite. The authors 

 conclude that it belongs to the ortho- 

 rhombic system. Its ratio of axes is 

 0-670 : 0-755 : I. The dimensions of its elementary 

 parallelopiped are 4-53, 5-11, and 6-77 A.U. The 

 base group consists of four molecules, FcgC, which 

 corresponds to a specific weight of 7-62 for cementite, 



H. C.H, C. 



NO. 2747, VOL. 109] 



