April 25, 1901] 



NATURE 



613 



I 



PRACTICAL PROBLEMS IN THE METALLO- 

 GRAPHY OF STEEL. 



CINCE Sorby in 1864 established the all-important 

 •^ fact that steel must be regarded as a crystallised 

 igneous rock, his work has been greatly expanded by the 

 international labours of many able microscopists. Much 

 of the work done, however, has been of academical 

 rather than of practical interest, and busy steel works' 

 metallurgists, appalled by the rapid growth of constituents 

 ending in " ile," of " eutectics" and of solid solutions of 

 carbon or carbides in unisolated allotropic modifications 

 of iron, are already beginning to ask themselves the 

 question. Is micrographic analysis going to be of any 

 real use to us, and, if so, in what direction ? The present 

 article is an attempt to very briefly answer the above 

 questions. 



The theory prevalent a quarter of a century ago that 

 steels of identical chemical composition would necessarily 

 have the same mechanical properties has long since 

 been discarded. But perhaps steel metallurgists have 

 not yet fully realised the disconcerting fact that steel of 

 excellent chemical composition, giving highly satisfactory 

 mechanical tests, may nevertheless utterly fail in use, 

 possibly with disastrous results. In other words, a 

 ductile steel which bends double cold without any sigi; 

 of flaw or failure may, under the influence of vibration, 

 snap like a piece of glass, though only subjected to 

 mechanical stresses well below its elastic limit. 



In connection with the materials of construction used 

 for high-speed engines, both land and marine, it is at the 

 present time a problem of paramount importance for the 

 scientific steel metallurgist to determine the cause of the 

 sudden infidelity of steel (or wrought iron) under 

 vibration. 



Data in the writer's possession prove beyond all doubt 

 that steel giving splendid chemical and mechanical tests 

 may rupture under vibration possibly in a few hours or 

 perhaps only after the lapse of twenty years. 



There is little doubt that in many cases the microscope 

 is capable of giving warning of the dangerous character 

 of a steel, chemically, and apparently mechanically, safe. 

 To intelligibly describe the structures of safe and 

 dangerous steels it is necessary to consider : 



(a) The micrographic constituents of structural steel. 



{b) The molecular migrations of these constituents 

 when at a red heat the metallic mass is in a semi-plastic 

 state. 



To put the case concretely, the chemical constituents 

 of a typical rail may (in addition to iron) be approxi- 

 mately taken as carbon 0*40, silicon 0*05, manganese o'go, 

 sulphur o 06, phosphorus o"o6 per cent., together with 

 small percentages of arsenic and copper. The micro- 

 graphic constituents of such a steel are : 



(i) The pale, simple constituent ferrite (in this case 

 somewhat impure iron). 



(2) The dark etching compound constituent pearlite, 

 consisting of mixed granules of iron and of a double 

 carbide of iron and manganese. 



(3) The dove-grey simple constituent sulphide of 

 manganese, MnS. 



It is important to remember that in manganiferous 

 steels the foregoing constituents are only completely 

 differentiated visually on slow cooling from a full red 

 heat, a fact which at once introduces the vital question of 

 the migration of constituents. Speaking broadly, it may 

 be said that sulphide of manganese is not, under working 

 conditions, capable of migration to any appreciable 

 extent. Thus it remains to consider only the migrations 

 of the ferrite and pearlite. 



The movements of these constituents on heating may 

 be termed " diffusion," and their movements on cooling 

 "segregation." 



On heating the typical steel specified to about 700° C. 



NO. 1643 VOL. 63] 



the compound constituent pearlite is converted, with 

 absorption of heat, into the simple constituent, marten- 

 site, at Osmond's point A^i. Then, passing through 

 Osmond's points A^,2-3, the constituents ferrite and 

 martensite diffuse one into the other till, at about 800° C, 

 molecular equilibrium is eventually established. 



If, however, the steel be cooled very slowly, the 

 molecules of martensite and of ferrite will perfectly 

 segregate in the respective proportions of about 45 and 

 55 per cent. Then at Ar. i, about 640'' C, the martensite 

 will decompose into the compound constituent pearlite, 

 which, owing to the presence of manganese, will be 

 granular and not laminated. On the other hand, if the 

 steel is somewhat quickly cooled in air, the segregation 

 of the constituents will be imperfect and the apparent 

 proportion of pearlite relatively large, because, owing to 

 the influence of the manganese present, the phenomenon 

 of constitutional segregation is retarded. 



As a matter of fact, the apparently large area of dark 

 pearlite is really an extremely intimate mixture of true 

 pearlite and unsegregated ferrite. 



The writer is aware that these statements may provoke 

 theoretical opposition, but they nevertheless describe the 



Fig. I. — Size of orieinal drawing, six inches ; magnification, 445 diameters. 

 Tae magnification here represented is about 166 diameters. 



observed facts, and by these, and not by theories, the 

 practical metallurgist must be guided. 



The micrograpli Fig. i shows the transverse section of 

 a rail web re-heated to 900" C. and allowed to cool in 

 air ; and this web exhibits the same structure in all the 

 three planes of section presently to be referred to. 



The micrograph Fig. 2 shows the structure when the 

 rail was slowly cooled in the re-heating furnace during a 

 period of 50 hours. 



In Fig. I the pale ferrite has imperfectly segregated 

 in the form of ragged and broken cell wails imperfectly 

 environing cells of pearlite mixed with unsegregated 

 ferrite. 



In Fig. 2 the pale ferrite and dark granular pearlite 

 have perfectly segregated mainly in the form of thick, 

 alternating laminae. The structure last named must be 

 regarded as highly dangerous, because under vibration 

 the adhesion between the constituents is liable to 

 gradually loosen and finally to be destroyed. Never- 

 theless, mechanical tests would initially reveal little 

 difference in the ductility of the two pieces of rail. 



