ON THE CHEMICAL COMPOUNDS CONTAINED IN ALLOYS. 141 



need at present in intermetallic chemistry. The difficulty of this subject, 

 whether we use the freezing-point curve or the microscope, is increased 

 by the uncertainty as to the maintenance of perfect equilibrium at each 

 stage of the cooling. 



Microscopic Examination. 



Osmond, Charpy, and Stead have shown us how much light the micro- 

 scope throws on our subject. 



The microscopic examination of the pattern shown by the polished 

 surface of an alloy that has, if necessary, been etched or heated to 

 produce oxidation colours seems to bring us nearer to the phenomena 

 than other methods of experiment. It is often quite easy to determine 

 which crystals formed first in the freezing (the primary crystallisation), 

 but there are certain types of pattern that are veiy puzzling. Among the 

 points calling for an answer are the following : 



1 . Does the existence of coated crystals, such as one finds in gunmetal 

 and also bronzes containing more tin than Cu3Sn, indicate the existence 

 of a second compound ? The answer is. Yes, in ,some cases — for example, 

 in the AuAl curve and, as Mr. Stead thinks, in the bronzes rich in tin, 

 where the Cu^Sn crystals are coated with CuSn. M. Le Chatelier has 

 lately pointed out that in all such cases the solid alloy is not in equili- 

 brium and that the eff'ects of annealing will generally be great. 



2. Can the existence of series of mixed crystals be detected by the 

 microscope ? 



M. Charpy and Mr. Stead both describe a similar structure, and are 

 disposed to attribute it to this cause. 



3. How far will the microscope supplement the very meagre indica- 

 tions that the curve sometimes gives of a compound ? 



As an answer one can take the portion of our copper-tin freezing-point 

 curve (fig. 10) between Cu4Sn and Cu^Sn. The curve is almost straight, 

 the swelling (one cannot call it a summit) corresponding to CugSn being 

 very slight. But the microscope shows Cu^Sn as a homogeneous body, 

 while alloys with a little more tin show new crystals embedded in this 

 and sharply separated. These new crystals increase as we add more tin, 

 until at CuaSn they fill the whole alloy ; thus the microscope is here much 

 more decisive in its indications than the curve. 



Rontgen-ray Photography. 



Skiagraphs of thin sections of alloy which contain one transparent 

 metal, such as sodium or aluminium, and one metal moi'e opaque, some- 

 times give fine views of the crystals in the alloy. This method has the 

 advantage of showing the structure of the alloy as it is before any etching 

 or other reagent has modified it. The two photographs shown were taken 

 some years ago by Mr. Heycock and myself. The first is aluminium 

 alloyed with ten per cent, of antimony. One sees that a heavy compound 

 has crystallised out first. This is in harmony with M. Gautier's curve 

 which presents no branch along which primary crystals of aluminium 

 could form. If a series of such photographs had been taken with increas- 

 ing percentages of Sb, we might have been able to locate the percentage 

 at which the compound was pure. 



