I IRNACE REACTION^ [ \DER HKJH CASEOUS PRESSURES. 447 



MOIHHAN,* taking an intermediate position, asserts that the two materials only react 

 when in the form of vapour. 



The question, which for many reasons is of considerable importance, has never 

 received the detailed investigation which it deserves. 



Our experiments at atmospheric pressure, as we shall see, pointed to the fact that 

 a well-marked thermal reaction does take place, but not until the fusing point of 

 alumina is reached. 



HKROULT,! while admitting that reduction occurs, attributes it to electrolytic 

 action. Having carried out some experiments in which the reacting sulxstances were 

 In -a ted by radiation alone and in which good yields of aluminium bronze were 

 obtained, we contend that the assumption of electrolysis is by no means necessary. 



No information was available as to the temperature of vaporisation of metallic 

 aluminium, but various observations led us to believe that a large proportion of the 

 reduced metal was lost by volatilisation and subsequent combustion where the 

 furnace gases come in contact with the air. 



The high-pressure furnace seemed to us therefore particularly suitable for studying 

 this question, the advantages to be gained consisting firstly in the complete protec- 

 tion of the products from oxidation, and secondly in the decreased volatilisation which 

 might be expected under the high gaseous pressures. 



Some of the experiments tried under pressure to study this problem are given in 

 Table VIII., details of one experiment being reproduced in Table IX. 



By a cursory inspection of Table VIII. the two following facts may at once be 

 deduced : 



(1) That in the resistance furnace neither aluminium nor its carbide is produced. 



(2) That on the other hand all arc furnaces give a more or less marked reduction ; 

 although it will lie noticed (in section B) that the product chiefly occurs as carbide of 

 aluminium. 



In several cases small malleable lumps of the metal were condensed in the powdered 

 material surrounding the fused product. 



From this it would appear that the required conditions for which we are searching 

 had for some short period t>een accidentally fulfilled these conditions being the 

 rapid removal of the metal vapour from the reduction zone and its condensation 

 under circumstances which precluded carburisation. 



The idea that, by reducing the partial pressure of the carbon monoxide by 

 a circulation of hydrogen or coal gas, more favourable results would be attained led 

 to the experiments quoted in Table VIII., and D. From these we infer that the 

 reaction is considerably favoured by a dilution of the carbon monoxide. It is further 

 noticeable that this precaution results in an increase in the relative quantity of 



* H. MOISSAN, < The Electric Furnace,' London Ed., ARNOLD, p. 184. 



t P. L. T. HEROUI.T, 'Eng. Pat. 16853,' 18*7; alqg 'Congrf-s intern, des Mines et de la Me"tallurgie ' 

 (Paris), 1900. 



