August 2;, 1891] 



NA TURE 



401 



at a certain critical temperature and pressure the slightest varia- 

 tion of either will destroy the equilibrium of the system and 

 induce chemical change. 



The aim of Boyle's chemical writings was to show that no 

 barrier exists between physics and chemistry, and to "serve the 

 commonwealth of learning by begetting a good understanding 

 betwixt the chemists and the mechanical philosophers," who 

 had, as he said, " been too great strangers to each other's dis- 

 coveries." In view of the dominant lines of research which 

 occupy chemists at the present time, such, for instance, as the 

 investigations of "osmotic pressure " and of the application of 

 Boyle's own law to salts in solution, he would feel that his hope 

 had been realized, and that, though he lived a century too soon 

 to take part in BerthoUet's discussion with Proust, he never- 

 theless shares BerthoUet's triumph in the long-delayed but 

 now rapid development of chemistry as a branch of applied 

 mechanics. 



We need, however, no longer look at these questions from 

 the point of view of Boyle, for our own interest in the application 

 of chemical mechanics to metallurgy is sufficiently vivid, as 

 instances to be given subsequently will show. 



Hitherto I have mainly dwelt on questions relating to oxida- 

 tion, but not less interesting is the history of the steps by which 

 an accurate knowledge was acquired of the other great process 

 practised by the metallurgist, the one to which Paracelsus was 

 the first to apply the name of "reduction." Its explanation 

 followed naturally from the elucidation of the phenomena of 

 combustion by Lavoisier, who in continuation of Macquer's 

 experiments of 177 1 proved, in conjunction with other workers, 

 that carbonic anhydride is produced when the diamond is burnt 

 in air or oxygen. Carbon has been known for ages as the most 

 important of the reducing agents, but when, in 177^, Lavoisier 

 heated oxide of lead and carbon together, he did not at first 

 recognize that carbonic anhydride had been produced, simply 

 because the volume of the gas set free was the same as if oxygen 

 merely had been liberated. He soon, however, saw that neither 

 the carbon alone, nor the oxide of lead alone, gave rise to 

 the evolution of carbonic anhydride, which resulted from the 

 mutual action of carbon and a constituent of the litharge. " This 

 last observation leads us insensibly," he adds, " to very import- 

 ant reflections on the use of carbon in the reduction of metals." 

 It most certainly did, and by 1815 an accurate, if incomplete, 

 view of reduction had passed into the encyclopaedias. It was seen 

 that the removal of oxygen from burnt metals, by carbon, "gives 

 the metals," as Fourcroy and Vauquelin put it, "a ne»v exist- 

 ence." Some ten years later Le Play attempted to show that 

 reduction is always effected by the intervention of carbonic 

 oxide, which elicited the classical rejoinder from Gay-Lussac, 

 who pointed out that "carbon alone, and at very moderate 

 temperatures, will reduce certain metallic oxides without the 

 intervention of carbonic oxide or of any other elastic fluid." 

 I mention these facts because metallurgists are slow to recognize 

 their indebtedness to investigators, and too often ignore the 

 extreme pains with which an accurate knowledge has been 

 acquired of the principles upon which their processes have been 

 based. 



The importance of a coherent explanation of reduction in 

 smelting pig-iron is enormous. The largest blast-furnaces in 

 18 1 5 hardly exceeded those in use in the previous century, and 

 were at most only 40 feet high, with a capacity of 5000 cubic 

 feet. At the present day their gigantic successors are sometimes 

 90 feet high, with a capacity of 25,000 cubic feet. This develop- 

 ment of the blast-furnace is due to the researches of a number 

 of investigators, among whom von Tunner, Lowthian Bell, and 

 Grimer deserve special mention. We are, however, forcibly 

 reminded of the present incompleteness of our knowledge of the 

 mechanism of reduction, when we remember that the experiments 

 of H. P. Baker have led us to believe that pure carbon cannot 

 be burnt in perfectly dry and pure oxygen, and therefore that 

 the reducing agent, carbonic oxide, cannot be produced at all 

 unless moisture be present. 



Ludwig Mond, Langer, and Quincke, teach us not only that 

 nickel can separate carbon from carbonic oxide, but the wholly 

 unexpected fact that dry carbonic oxide can at a temperature of 

 ICO take up nickel, which it again deposi'.s if heated to 150 . 

 Mond and Quincke, and, independently, Berthelot, have since 

 proved the existence of the corresponding compound of iron and 

 carbonic oxide, and it may safely be concluded that in the blast- 

 furnace smelting iron this peculiar action of carbonic oxide plays 

 an important part, and it doubtless aids the carburization of iron 



NO- I I 39, VOL. 44] 



by cementation. It is truly remarkable that the past year should 

 have brought us so great an increase in our knowledge of what 

 takes place in the reduction of an oxide of iron, and in the car- 

 burization of the liberated metal. My own experiments have, I 

 trust, made it clear that iron can, at an elevated temperature, be 

 carburized by the diamond in vacuo ; that is, in the absence of 

 anything more than "a trace " of an elastic fluid or of any third 

 element. Osmond has further shown within the last few months 

 that the action between iron and carbon is a mutual one, for 

 though carbon in the pure diamond form carburizes iron, the 

 metal in its turn, at a temperature of 1050°, attacks the diamond, 

 invests it with a black layer, and truly unites with it. 



The question of the direct carburization of iron (Darby's 

 process) by filtering the molten metal through carbon, promises 

 to be of much importance, for at present, as is well known, two 

 millions of tons of steel which are made in the Bessemer con- 

 verter in this country alone, are re-carbarized after " the blow" 

 by the addition of spiegeleisen. 



Carbonic oxide, moreover, would appear to be more chemically 

 active than had been supposed ; for during the present year 

 Berthelot has shown that the perfectly pure gas heated to 500° 

 or 550 produces carbonic anhydride with the deposition of 

 carbon at red heat, not by ordinary dissociation, but by decom- 

 position preceded by polymerization. He further shows that 

 carbonic oxide will decompose ammoniacal nitrate of silver, and 

 thus brings it into close connection with the aldehydes. 



(2) In turning to the modern aspects of metallurgical practice, 

 we shall see that the whole range of the metallurgist's field of 

 study is changing. It is no longer possible for him to devise a 

 series of operations on the evidence afforded by a set of equations 

 which indicate the completion of an operation ; ha has, as I 

 have already suggested, to consider the complicated problems 

 which have been introduced into chemistry from the sciences of 

 physics and mechanics. He has, in fact, no longer to deal 

 merely with atoms and molecules, but with the influence of mass. 

 As Ostwald points out, we are reminded that many chemical 

 processes are reciprocating so that the original products may be 

 obtained from the product of the reaction. The result of such 

 opposed processes is a state of chemical equilibrium, in 

 which both the original and the newly-formed substances are 

 present in definite quantities that remain the same so long as the 

 conditions, more especially temperature and pressure, do not 

 undergo further change. Again, in very many metallurgical 

 processes, reactions are rendered incomplee by the limitations 

 imposed by the presence of bodies which cannot be speedily 

 eliminated from the system, and the result may be to greatly 

 retard the completion of an operation. The time has come when 

 the principles of dynamic chemistry must be applied to the study of 

 metallurgical problems if they are to be correctly understood, and 

 it is, moreover, necessary to remember the part played by the 

 surface separating the different aggregates in contact with one 

 another. When, for instance, a reaction has to take place accom- 

 panied by the evolution of gas, there must be space into which 

 the gas can pass. The rate, therefore, at which change takes 

 place will obviously depend on the state of division of the mass. 



One of the most remarkable points in the whole range of 

 chemistry is the action engendered between two elements capable 

 of reacting by the presence of a third body. It may be, and 

 this is the most wonderful fact of all, that merely a trace of a 

 third body is necessary to induce reaction, or to profoundly 

 modify the structure of a metal. H. Le Chatelier and Mouret 

 have pointed out that in certain cases it is inaccurate to say that 

 the third body causes the reaction to take place, because, after 

 it has destroyed the inter-molecular resistances which prevented 

 the reaction taking place, the third body ceases to intervene. 

 This is apparently the case when platinum sponge effects the 

 union of oxygen and hydrogen, or, conversely, when very hot 

 platinum splits up water vapour into its constituent gases. 

 Future investigation will, it is to be hoped, show whether the 

 platinum does not exert some direct action in both cases. We 

 can no longer neglect the study of such questions from the point 

 of view of their practical application. The manufacture of red 

 lead presents a case in point. In " drossing" molten lead, the 

 oxidation of the lead is greatly promoted by the presence of a 

 trace of antimony ; and conversely, in the separation of silver 

 from molten lead, by the aid of zinc, H. Roessler and 

 Endelmann have recently shown that aluminium has a remark- 

 able effect in protecting the zinc from loss by oxidation, and 

 further, the presence of one-thousandth part of aluminium in 



