906 STEEL 



fire or reverberatory rurnace carbon, silicon, manganese, and some iron being burnt 

 in either case, with the production of carbonic oxide, silicates of protoxide of iron 

 and manganese, and malleable iron we have, in the blowing of a charge weighing five 

 tons an amount of work done in about two and a half to three days in its performance 

 in the puddling furnace. It has been pointed out by Jordan, that the principal part 

 of the heat developed in the process is due to the combustion of silicon, which when 

 oxidised to silicic acid, combines with protoxide of iron, and other bases, and remains 

 in the bath in the form of slag ; while in the case of carbon, a considerable portion of 

 the heat is expended in volatilising the carbonic oxide produced, which escapes at the 

 temperature of the melted metal, and burns to waste at the mouth of the converter. 

 If the calorific power of silicon be assumed to be the same as that of carbon, the amount 

 of heat produced by the combustion of one kilogramme of silicon to silicic acid will be 

 8,000 units, 1 when burnt in pure oxygen, or 6,382 in air ; the difference between the two 

 quantities corresponding to the amount required to heat up the inert nitrogen. Under 

 the latter condition, one kilogramme of carbon will produce only 475 effective units, 

 beingthe difference between 2,473 units theoretically developed and 1,998 units carried 

 off by the gaseous products carbonic oxide and nitrogen, supposing them to escape at a 

 temperature of 1400. The use of steam instead of air as an oxidising agent, is, in 

 the case of the combustion of iron or carbon, always disadvantageous on account of the 

 great amount of heat required to free the oxygen from its combination with hydrogen, 

 which is not reproduced to the same extent in the subsequent formation of carbonic 

 oxide or protoxide of iron. With silicon, however, the conditions are somewhat 

 different, as there is a small sensible gain. This will explain the reason why the use 

 of steam in the refinery is only recommended for a few minutes at the commencement 

 of the operation, that is, as long as free silicon remains in the pig-iron under 

 treatment. 



By applying the quantities given above to the calculation of the amount of heat 

 developed in the blowing in one ton of Bessemer pig-iron of the ordinary quality pro- 

 duced in the south of France (which has the following composition per ton of 1,000 

 kilogrammes: carbon, 42'50 ; silicon, 20 - 00 ; iron and manganese, 937'50 = 1000'00), 

 Jordan arrives at the following results : 



Kilogrammes Units of heat 



The combustion of 20 of silicon produces . . . 127,648 



42 '5 of carbon produces . . . 20,176 



87'5 of iron and manganese produces 66,237 



Or a total of . 214,061 



If we take the specific heat of molten malleable iron at 0'16, the amount of heat 

 developed will be sufficient to raise the temperature of the metal, which is supposed 

 to be completely decarbonised, about 1350 above that of the cast iron when run into 

 the converter. 



The great heating power of silicon is, therefore, to be regarded as the reason for the 

 use of dark-grey iron in the Bessemer process ; under ordinary circumstances, about 

 '2 or 2'5 per cent, silicon being considered as essential. Jordan states that in the 

 steel works in the south of France the process could only be carried out by running 

 the cast iron directly from the blast-furnace into the converter. The amount of silicon 

 as a heat-producer in the Bessemer process may be, to some extent, taken by mangan- 

 ese; as is the case in Styria, where the cast iron used is smelted from the spathic ores. 

 It is, however, less advantageous, because the deficiency in silica, which is required to 

 flux the protoxide of manganese formed, can only be supplied by the destruction of the 

 siliceous lining of the converter. The corrosive action of manganese on the hearths 

 of blast-furnaces where spathic ores are smelted has already been noticed. 



Although silicon is an essential component of good Bessemer pig-iron, it is of im- 

 portance that the amount per cent, should be somewhere about the same as, or not 

 very much more than that of the carbon. An excess of the former element, works 

 prejudicially in two ways: first, it gives rise to an increased waste of iron in the slag ; 

 and secondly, it cannot be completely removed before the whole of the carbon is burnt 

 away, so that it may happen in the blowing of such metal, that, although the process 

 is apparently complete, as determined by the usual indication of the cessation of 

 the flame from the converter, sufficient silicon is retained in the decarbonised metal to 

 render the finished steel brittle and useless. Snelus gives the following analyses in 

 illustration of this point : 



1 TMs is in excess of the real ninount, which has recently been determined to be 7,000. Jordan's 

 original figures are however preserved, as the quantities are only given as approximations, for the 

 purpose of illustrating the theory of the process, and not as absolute numerical determinations 



