832 



NATURE 



[December 8, 192 



Solid Solutions and Inter-Metallic Compounds. 

 By Dr. Walter Rosenrain, F.R.S. 



MKTALLURC.ICAL research during the past 

 twenty years has l)ecn largely devoted to the 

 stud)- of alloys, and as one result we now possess a 

 series of more or less complicated equilibrium diagrams 

 representing the constitution of most of the binary 

 and of some of the ternary systems. While, on 

 one hand, increasing accuracy of methods has rendered 

 these diagrams far more complex than was at first 

 supposed, a careful examination of those which are 

 most thoroughly established suggests that, widely as 

 they vary among themselves, there are certain regulari- 

 ties which point to some common fundamental principle 

 which, if once grasped, would exhibit these varied 

 diagrams as parts of an intelligible whole. Fortun- 

 ately, at the time when this great mass of disconnected 

 knowledge lies awaiting synthetic treatment, the 

 results of X-ray analysis applied to the study of the 

 inner structure of crystals have become available. 

 As the result of an endeavour to apply these results 

 to the explanation of the behaviour of alloys systems, 

 the writer has arrived at a theory which, on a simple 

 basis, promises to afford an easy explanation of many, 

 if not of all, of the properties of alloys, and to afford a 

 much deeper insight into the nature of solid solutions 

 and of inter-metallic compounds, and through them 

 to throw new light on the nature of inter-atomic 

 relationships. 



The theory in question has been fully stated in two 

 recent papers, and need only be briefly summarised 

 here.^ A metallic solid solution is an aggregate of 

 crystals which, when in equilibrium, are homogeneous 

 in composition, so that both the solvent metal and 

 the solute metal are present in the same proportions 

 in all the crystals. The present theory of the con- 

 stitution of such cr)'stals is based on three funda- 

 mental principles, the first of which has now received 

 considerable experimental verification, while the other 

 two appear to follow almost unavoidably. The first 

 is that a solid-solution crystal is built up of the two 

 kinds of atoms, those of the solvent and of the solute, 

 upon a single space-lattice which is, substantially, that 

 of the solvent, so that the atoms of the solute may be 

 regarded as being simply substituted for an equal 

 number of atoms of the solvent on the " parent " 

 lattice. Measurements of the lattice-constants of 

 certain groups of solid-solution alloys and comparison 

 of the results with the measured densities of the alloys 

 have strongly confirmed this view. The evidence 

 already obtained indicates that this is the inner 

 structure of practically all inter-metallic solid solu- 

 tions, but some room for doubt may still exist in 

 regard to certain metalloids, such as carbon or 

 phosphorus. 



Next, in a cr)stal built up in this manner of two 

 kinds of atoms upon a single, simple space-lattice, the 

 inference can scarcely be avoided that a certain degree 

 of distortion of the lattice must result. The nature 

 of this distortion must depend upon the character of 



' "Solid Solutions," Second Annual Lecture of the Inst, of MetaU 

 Di\*ision, .American Inst. Mining Enfrineers, New York. Feb. 1923; and 

 "The Inner Structure of .\llov-s," Thirteenth Mav Lecture to the Inst, of 

 Metals, London, May, 1923. Joum. Inst. Metals. 1923, ii. 



NO. 2823, VOL. I 12] 



!it 



the two kinds of atoms concerned ; there ma\ 

 either expansion or contraction of the p.ir* m l.i' 

 and this may l)e either mainly local or n 



The degree and nature of this dLstortioi , 



upon the extent to which the .solute atom differs Ir 

 the solvent, and also upon the general character 

 the solvent lattice, but these are details which iv 

 not be considered here. We may pass on to the il ird 

 fundamental conception — that the extent to wl h 

 any given space-lattice can be distorted, and par' 

 larly expanded, is strictly limited — that thf-n- \ 

 fact, for each pair of atoms a limitir 

 which the bond between them — wh.t 

 ceases to act. This rule of a limiting maximum hi- 

 constant or parameter leads to a series of intert - 

 inferences. Thus, a uniform undistorted lattic< 

 a pure substance will be uniformly expanded by i : 

 until the limiting parameter is attained ; at this i ■ t 

 the atoms throughout the lattice will lo.se their p< • r 

 of cohesion and the crystal melts. In a solid sohi- w 

 crystal, the lattice may be locally expanded b\ ? ; » 

 presence of solute atoms; under thermal expar. w 

 those expanded regions of the lattice will reach : ; < 

 limiting parameter at a temperature where the !• > 

 expanded portions of the lattice are still well Ix 

 the limiting value ; the result will be comm.encen 

 of fusion in those regions of the crj-stals richest :i 

 solute and the formation of a liquid richer in sol -t 

 than the remaining solid. This consideration explai > 

 why, in solid solutions, we generally find a melt:- j 

 range instead of a single melting point. Where •\x 

 solute atoms cause expansion of the lattice the mi-h;: - 

 temperatures will be depressed by successive additi i- 

 of solute. On the other hand, where the present ( i 

 the solute atoms causes a contraction of the soKini 

 lattice, there will be a rise of melting point and 

 first liquid to be formed on fusion is richer in solu 

 than the residual solid. These latter inferences havi 

 been strikingly verified in such cases as those of sol 

 solutions formed by the addition of palladium 

 silver or of nickel to copper. 



A considerable number of further inferences can 

 drawn from the three fundamental principles of 

 present theory' of the inner structure of solid soluti( 

 — for example, the striking inverse relationship w 

 is found to hold between the solubility of one mi 

 in another and its hardening effect upon it. and 

 relationship between the hardness, high melting [x>iiii 

 and high elastic modulus of a metal on one side 

 its power of forming solid solutions on the ot 

 The theor}' has even made it possible to suggest 

 explanation of the properties of metals and alli 

 in regard to electrical conductivity. Whatever 

 true mechanism of electric conduction, there can 

 no doubt that it is associated with the movement 

 electrons through the metal ; it is now suggested 

 where the atoms lie on perfectly straight lines on 

 space-lattice the movement of electrons is en 

 unhindered and the metal in that state should exhil 

 super-conductivity. This can only be fully r 

 ver>- near the absolute zero, since at higher tem 



