ON THE INFLUENCE OP CARBON, ETC., ON STEEL CORROSION. 87 



The influence of the condition of the carbide upon the relative electro- 

 chemical positions of the steels is, however, more definitely shown. 

 In the case of a saturated steel it is found that after three weeks' 

 immersion in sea-water the emulsified variety of pearlite is electro- 

 positive to the diffused and laminated varieties, whilst the conversion 

 of the pearlite into hardenite renders it electronegative in this form to 

 all the varieties of pearlite. 



Determination of Solubility in Acid Solutions. 



The solubilities of the various steels employed have been determined 

 in 1 per cent, solutions of H 2 S0 4 and HOI, in order to determine the 

 influence of carbon on the solubility of iron-carbon alloys, and also 

 to obtain an indication of the loss in weight per cent, sustained on im- 

 mersion in acid solutions, as contrasted with corrosion in such solutions 

 as sea-water. 



This was considered of importance in view of corrosion taking place 

 under ordinary conditions, especially in large towns, where atmospheric 

 acids contained in the rainwater play an important part. This section 

 is also of interest in the case of corrosion taking place in such solutions 

 as acid pit waters. 



The solubility tests were carried out on polished cylindrical bars 

 of the steels (1J inch long x § inch dia.), which were separately 

 immersed for a period of forty-eight hours in 100 c.c. of the acid 

 employed. The results obtained on immersion in 1 per cent. H 2 SO\i 

 are remarkably similar to those obtained in 1 per cent. HC1, both as 

 regards the actual values obtained with given specimens and in the type 

 of influence exerted upon the solubility by the carbon per cent, and 

 the treatment. The two sets are consequently dealt with together, 

 and the curves obtained with 1 per cent. H 2 S0 4 only are shown in 

 fig. 2. It may be noted that the practice of drawing smooth curves 

 through the experimentally determined points has been adopted 

 throughout in preference to the direct joining-up of these points by 

 straight lines, as is sometimes carried out. The method adopted is 

 found, in this case, to give much more satisfactory indications of the 

 real positions of the critical points in the curve than the alternative 

 one, whilst the possible degree of error introduced thereby is sufficiently 

 small to be negligible. 



The influence of carbon on the solubility is found to vary consider- 

 ably according to the treatment previously undergone by the steel. 



In the normalised, rolled, and annealed specimens the solubility 

 rises very rapidly from 010 per cent, up to 0'22 per cent., carbon 

 approx., after which it falls abruptly, reaching a minimum at about 

 0'30 per cent, to 0'40 per cent, carbon. Further rise of carbon from 

 0'40 per cent, to 0'96 per cent, carbon produces a general, but some- 

 what irregular, rise in solubility in the normalised and rolled speci- 

 mens, whilst in the annealed specimens a maximum solubility is reached 

 at 0'60 per cent, carbon, after which the values gradually decrease 

 again to 0'96 per cent, carbon. A comparison of these curves with 

 those given by Heyn and Bauer * as indicative of the probable influence 



4 Journ. Iron and Steel Inst., 1909, vol. 1, p. 109. * 



