590 REPORT — 1891. 



the contact of tbe ore and the reducing gas was not sufficiently prolonged, led to a 

 great increase in the height of blast-furnaces, but without, as Griiner showed, 

 diminishing the proportion of carbonic oxide escaping from the throat. The re- 

 duction of an iron ore by carbonic oxide only takes place within certain well- 

 defined limits, and a knowledge of the laws of chemical equilibrium would have 

 saved thousands and thousands of pounds which have been wasted in building 

 unduly high furnaces. I would add that large sums have also been sacrificed in 

 the vain attempt to smelt oxide of zinc in the blast-furnace, for which operation 

 patents have frequently been sought, in ignorance or defiance of the readiness with 

 which the inverse action occurs, so that the reducing action of carbon on oxide of 

 zinc may be balanced by the re-oxidation of the reduced zinc by carbonic anhydride, 

 which is the product of the reduction. A further instance may be borrowed from 

 an electro-chemical process which has been adopted for obtaining alloys of 

 aluminium. As is well known, all attempts to effect the direct reduction of 

 alumina by carbon have failed, because the reaction 



2(Al203) + 3C = 4Al + .'3C02 



requires 783 '2 calories, while only 291 calories would result from the conversion 

 of carbon into carbonic anhydride, therefore the reaction cannot be effected ; but in 

 Cowle's process aluminium is nevertheless liberated when alumina is mixed with 

 charcoal and strongly heated by the passage of an electric current. This result is 

 due, not to a simple reduction of alumina, but to its dissociation at the high tem- 

 perature produced by the passage of a current of 1,600 amperes between carbon 

 poles, the liberated aluminium being at once removed from the system by metallic 

 copper which is simidtaneously present and may not be without action itself. 



An instance of the importance of these considerations is presented in the 

 manufacture of steel by the basic process. Much care is devoted to obtain- 

 ing conditions which will ensure not only the elimination, but the order of the 

 disappearance of the impurities from the molten pig iron. In the basic process as 

 conducted in the closed converter, the phosphorus does not disappear until the 

 carbon has left the fluid bath, whilst, when the open-hearth furnace is used, the 

 elimination of the phosphorus may be effected before that of the carbon, and it is 

 asserted that if the carbon goes before the phosphorus is got rid of, a further 

 addition of carbon is necessary. A curious and subtle case of chemical equilibrium 

 is here presented. In the open-hearth furnace and Bessemer converter respectively, 

 the temperatures and pressures are different', and the conditions as to the presenta- 

 tion of oxygen to the fluid bath are not the same. The result is that the relative 

 rates of oxidation of the phosphorus and carbon are different in the two cases, 

 although in either case, with a given method of worldng, there must be a ratio 

 between the phosphorus and carbon in which they disappear simultaneously. 

 The industrial bearing of the question is very remarkable. In the basic Bessemer 

 process the tendency of the phosphorus to linger in the bath renders an ' after- 

 blow ' necessary, it may be only of a few seconds' duration, but much iron is never- 

 theless burnt and wasted, and Mr. Gilchrist tells me that if this after-blow could 

 be avoided, a saving of some six per cent, of the yield of steel would be effected 

 annuallj', the value of which, at the present rate of output and price of steel, is no 

 less than a quarter of a million sterling. 



The larger loss of sulphur by the steel in the converter than that which occurs 

 in the open-hearth furnace, and the increase in the percentage of manganese, which 

 leaves the slag and returns to the bath of metal in the converter at the end of the 

 ' blow,' will probably be traced to the disturbance of equilibrium which attends 

 very slight variations in the conditions, especially as regards temperature and 

 pressm-e, under which the operations are conducted. 



In the blast-furnace the reducing action must be greatly dependent on the rate 

 at which alkaline cyanides are formed, and Hempel has recently shown, by the 

 aid of well-devised experiments, that the quantity of cyanides which may be 

 obtained at a high temperature from carbon, nitrogen, and alkaline oxides, increases 

 as the pressure becomes greater. 



Metallurgical chemistry is, in fact, a special branch of chemical science which 



