IRON 311 Precisely similar results are calculable from the analyses of Tiinner, Ebelmen, and others who have examined the composition of the gases at different levels of the blast furnace ; the variations in the amounts of carbon and oxygen relatively to the nitrogen at the lower levels in all cases are of such a nature as to indicate that the amount of decomposition of iron cyanide with evolution of nitrogen is very considerable, i.e., that the reduction of iron oxide by alkaline cyanides takes place to an extent constituting a very considerable fraction indeed of the total amount of reduction. The amount of alkaline cyanides disseminated through the gases of a furnace at different levels varies inversely with the height above the tuyere ; thus, in the course of Lowthian Bell s experiments, the following analyses were made by the present writer of the sub stances dissolved by water through which known large volumes of the gases were aspirated, being drawn from the diil erent levels into a large gasometer, the weights being given in grammes per cubic metre of gas (at and 760 mm. ), and the experiments being all made within a short time of one another (all on the same day) : Height above tuyere in) feet j 8 24 CO 76 Exit pipe after leaving furnace. 73-47 14-15 918 16 05 3-47 Sodium 39-23 17-84 1(1-69 7-99 1-72 Cyanogen 49-06 15-76 7-C7 5"94 4-73 Other substances Cl-31 10-10 9-85 19-38 il-40 , Total constituents of the) fume soluble in water j 223-07 G285 43-39 49-36 21-32 The amounts of alkaline cyanides were found to be considerably variable from day to day when the gases from any given perforation were examined ; thus, for example, the following amounts of com bined cyanogen were obtained in two other series of observations with the first and last of these perforations : 1st Day. 2,1 Day. 6th Day. 9th Day. 11-34 3-57 13th Day. 15th Day. Eight ft. above tuyere 19-00 12-03 4-00 17-32 6 -CO 20-61 2-91 9-16 1-79 In the furnace examined the quantity of gases at a few feet above the tuyere level per unit weight of pig iron made would be about 6 parts by weight, so that per 100 grammes of pig the gases would weigh about 600 grammes, occupying about 45 cubic metre. When the amount of cyanogen combined as cyanides disseminated through the gases was 20 grammes (equivalent to 50 grammes of potassium cyanide) per cubic metre (a quantity often exceeded), the potassium cyanide per 100 grammes of pig would consequently be about 22 5 grammes, or about -g- of the weight of the pig iron, and consequently about t x |- = | roughly of the oxygen in the form of ferric oxide in the ore originally used ; hence evidently the influence exerted by the combined cyanogen upon the removal of the last portions of oxygen must have been very considerable indeed, especially as the cyanides that escape in the gases from the hearth probably represent con siderably less than the total amount generated there, a considerable proportion being used up in deoxidizing the iron oxide pan passu with its formation. That this is so has indeed been urged long ago by Bunsen and Playfair, who found that the gases drawn from a per foration 2 feet 9 inches above the tuyere of the Alfreton furnace contained cyanogen compounds equivalent to from 8 to 10 grammes of potassium cyanide per cubic metre of gas, much smaller amounts than those above mentioned, but greater than those found on some other occasions when the alkaline substances contained in the fume consisted chiefly of carbonates. The chief source of the alkalies which form the cyanides is the coke used as fuel, but the ore and flux also usually contain small quantities ; when a furnace is newly blown in, the amount of cyanides is necessarily very small ; but a very few weeks use suffices to cause an accumulation of a quantity sufficient to exert a marked influence on the chemical actions taking place, whilst a somewhat longer period brings the accumulation up to the final working average attained when the alkaline compounds mechanically carried off in the fume, and escaping altogether from the furnace through not being inter cepted and filtered out by the substances in the upper part, together with those in the cinder, just equal the alkalies brought in by the fuel and burden jointly. Jt is highly probable, although not abso lutely demonstrated, that when charcoal is used as fuel the forma tion of alkaline cyanides is promoted, owing to the increased quan tity of potassium carbonate in the ash of the charcoal as compared with coke ; and that this is one of the reasons why the consumption of carbon in the form of charcoal in the Styrian, American, and Swedish furnaces is often less per ton of iron made than that of coke in even the best of the large English furnaces, the greater ease with which the ores are reduced as compared with English ones being, at any rate in certain cases, another circumstance diminishing the quantity of fuel requisite. A large number of direct observations as to the progressive changes undergone by the minerals in descending through the fur nace have been made, more especially by Ebelmen, Tiinner, and Lowthian Bell, with the general result of showing that the changes as a whole are substantially those above described ; as the iron ore sinks, it becomes deoxidized at a.rate which at first gradually in creases, the_temperature rising ; but by and by the reduction ceases to increase in rate, and would probably almost stop were the inner portions of the lumps as much reduced as the outer portions. Neither direct experiments on the ores in the furnace, nor laboratory experiments, nor the results deduciblc from the examination of the composition of the gases at different levels indicate that under the conditions of the blast furnace interior complete deoxidation of the ore _ensues until the level of the hearth is reached and the iron begins to fuse, the agents completing the dooxidation being partly the carbonaceous matter of the solid fuel, but to a much greater extent the finely divided carbon precipitated from the carbon oxide in the upper part of the furnace, and the alkaline cyanides. 20. Development and Ajyyropriation of Heat in the Blast Furnace. The sources of heat in the blast furnace are two in number, viz., the heat brought in by the hot blast, and that generated by the combustion of the fuel. The former of course varies considerably with the nature of the heating arrangement and with the actual weight of blast employed per unit weight of iron smelted ; thus, if the weight of air used be 5 5 times that of the pig iron made (110 cwts. of blast per ton of pig), if its temperature be 500 C., and the average specific heat of its components 23, the heat brought in per unit weight of pig made will be 5 5 x 500 x 0-23 = 632-5 heat units, the weight of the pig iron being the unit of weight ; and similarly in other cases. The heat generated by the combustion of the fuel, again, depends, first, on the amount of fuel burnt and the pro portion of inert matters (ash) in it and other circumstances modifying its heat of combustion, and, secondly, on the relative amounts of carbon oxide and dioxide formed, In transforming 1 part of amorphous carbon into carbon dioxide, the heat evolution (the materials and products being all at the ordinary temperature) is close to 8000, the following values having been found by different observers : Favre and Silbermann..... ......................... 8080 Wood charcoa Despretz ................................................ 7912 Do. Andrews ................................................ 7900 Do. Fuvre nnd SObennann .............................. 8047 Gas carbon. Do. do ............................... <797 Graphite. If, again, carbon oxide be burnt to dioxide, the amount of heat is near to 2400 per unit weight of carbon oxide. Favre and Silbermann ......................................................... 2403 Andrews ........................................................................... 2431 Hence the heat given out in burning one part by weight of carbon to carbon oxide must be 8000 - x 2400 = 2400, since 3 parts of carbon yield 7 of carbon oxide. If then a given quantity of coke containing 95 per cent, of carbon be burnt, two-thirds to carbon oxide and one- third to carbon dioxide, the heat produced will be that is, the heat developed by this combustion of one part by weight of fuel would suffice to raise the temperature of 4053 parts by weight of water through 1" C. ; or generally, if - - of the carbon be Jl m + n burnt to carbon oxide and to carbon dioxide, p being the per- m + n ccntage of carbon in the coke (the trifling amount of hydrogen being neglected), the heat development per 100 parts by weight of coke is p( m x 2400 + - " - x 8000 ) One part by weight of hydrogen m+n m+n ) furnishes about 34,000 heat units when burnt to liquid water, between 28,000 and 29,000 if burnt to vaporous steam ; so that, if q be the percentage of hydrogen, the total heat development per one part of fuel is close to p( m . x 24 + x 80 ) + q x 285 ; if m+n m+n J q is loss than 5 (as is usually the case) the error caused by neglecting the term involving q altogether is not greater than that due to the uncertainty about the precise values of the heat evolved in burning carbon to carbon oxide and to carbon dioxide (taken above approximately as 2400 and 8000 respectively). Knowing the quantity of fuel (coke) burnt and the average com position of the waste gases, together with the amount of flux (lime stone) employed, the quantity of carbon dioxide and oxide formed by the combustion of the coke and the amount of blast employed to burn it can be readily calculated ; for example, in one out of many series of observations made by Lowthian Bell with the present writer s co operation, it was found that the average composition by weight of the issuing gases was
