;o6 



NATURE 



[SliJTEMIilir }^, "1910 



The microscope does not show in what constituent the 

 siUcon crystallises. It is known that in grey irons it is 

 associated with the ferrite and pearlite ; but grey iron is 

 the final result of the decoinposition of carbide of iron 

 and possibly silico-carbides, which primarily form during 

 solidification, .and although the silicon in the decomposed 

 product may be entirely associated with the iron, it is no 

 proof that initially some of it may not have crystallised 

 with the carbides. 



In the white Cleveland iron, previously referred to, it 

 is probable that the several constituents are present in the 

 following proportions : — 



Silico-pearlite, the residue of the original 



austenite octahedral crystallites 

 Iron carbide in plates 

 Iron, phospho-carbide eutectic 

 Manganese sulphide 

 Iron sulphide 



Per cent. 



42-50 

 33-66 

 23-10 

 0-38 

 0-36 



100-00 

 When fractionally dissolving the powdered metal in 

 acid, it was the iron and associated silicon of the pearlite 

 which passed into solution, and the carbide and phosphide 

 which remained insoluble, and as these contained only 

 0-I2 per cent, silicon, or about o-o6 per cent, on 100 parts 



KlG. 5.— .Same as Fig. 4. Section cut parallel to the 

 surface of a massive carbide plate. 

 The ground mass is carbide of iron, 

 'the white dots are sulphide of iron. 



of the original metal, it is evident that the pearlite must 

 have contained 1-89 — 0-06=1-83 per cent, of the silicon, or 



on 100 parts of it =4-3 per cent., and that about 



42-5 

 97 per cent, of the total silicon had crystallised with the 

 austenite. 



A little reflection will lead to the conclusion that if the 

 carbon in the Cleveland white iron were to be gradually 

 increased, the proportion of primary austenite crystallites 

 would decrease ; there would be less and less of them to 

 carry the silicon, and this element would be concentrated 

 in the diminishing solid austenite. It also follows that if 

 the carbon \Yere to be so increased that no primary 

 austenite would form, the silicon would have to crystallise 

 in some other constituent. 



In the e.xample, referred to above, of the chilled cast- 

 ing, the carbides contained only 0028 per cent, silicon, 

 or o-oi6 per cent, on the original metal. In this case, 

 therefore, about 98 per cent, had crystallised with the 

 primary austenite. 



The question as to what amouiit of silicon will crystal- 

 lise with the austenite so as to saturate it is probably 

 NO. 2132, VOL. 84] 



variable with other variables. To determine this by 

 chemical analysis would involve an exceedingly tedious 

 research. 



It is probable that as it increases, and as the austenite 

 approaches more and more nearly to the saturation point, 

 a gradually increasing proportion of the silicon will 

 crystallise with the carbides. 



It is well known that molten low silicon grey irons, in 

 the absence of any appreciable quantity of sulphur, gives 

 a white fracture when slightly chilled. Irons with above 

 5 per cent, silicon, when similarly treated, are supposed 

 not to behave in the same manner, and this is quite true 

 when any ordinary method of chilling is adopted. For 

 instance, when the liquid silicious glazed metal No. i 

 was run into water, the chilled iron contained graphite ; 

 but when a large drop was suddenly pressed into a sheet 

 as thin as paper between cold plates of iron, the chilled 

 metal was quite white, and no graphite could be detected 

 on dissolving it in nitric acid. The metal so chilled was 

 difficult to dissolve in acid, and the silica produced, instead 

 of forming a gelatinous bulky residue, remained in a 

 close, dense condition — indeed, the thin chilled sheet, 

 after all soluble matter had been removed, remained a 

 ri'gid sheet of dense coherent silica, whereas the same 

 metal allowed to cool slowly from the liquid state in a 

 sand mould yielded to acid gelatinous silica. 



Fig. 6. — Glazed Cleveland Iron after melting with 



little Sulphide of Iron. 



White crystals = primary carbo-silicide of iron. 



Dark = the second cementile. 



Comple.v structure = iron-carbon-phosphoi 



The different behaviour to acid treatment of the chilled 

 as contrasted with that of the slowly cooled metal indicates 

 that the condition of the silicon in rapidly chilled metal 

 is different from its condition in the same metal slowly 

 cooled. 



In 1895 Mr. T. W. Hogg, of Newburn Steel Works, 

 published an account of a very interesting observation, in 

 which he showed the difference in the silicon solubility in 

 different parts of the same pig iron, a portion of which 

 was white and a portion grey. The iron referred to 

 contained : — 



White part 



Combined carbon 3'88'\ 



Graphitic carbon o'4Sj' 



Silicon o'65 



Manganese 1'63 



He determined the solubility in dilute acid of the sil.icon 

 in each portion, and found that the silicon soluble in 

 hydrochloric acid was, in the grey part = about Si per 

 cent, and in the white part = about 48 per cent. 



He found also that the silica left on treating the two 



4-33 



Per cent- 



o-98-\ 

 3-68/ 

 085 

 I '60 



4 -£6 



