NATURE 



[March i8, 1909 



speedily decompose. The explosive combustion of ethylene 

 may, therefore, be represented by the following scheme : — 



H„C:CH„ 



HO.CH:CH„ 



HO.CH:CH.OH 



i C0H0+H..O "» 2CH„0 = 2C0 + 2Hj 



\2C+ft2+H.p/ 



In a sufficient supply of oxygen, the transition from the 

 original hydrocarbon to the diliydroxy state is probably 

 so rapid that no breaking down of the ethylenic structure 

 occurs in passing through the initial monohydroxy stage. 

 Indeed, it is conceivable that under the extreme conditions 

 of detonation the passage from o to 2 may be effected in a 

 single molecular impact. The dihydroxy derivative would 

 at once break down into carbon monoxide and hydrogen, 

 via formaldehyde. 



But when the oxygen supply is reduced below the equi- 

 molecular proportion, it is evident that the initial mono- 

 hydroxy derivative cannot all be oxidised to the dihydroxy 

 stage ; some of it would, therefore, decompose partly into 

 acetylene and steam and partly also into carbon, hydrogen, 

 and .steam, together with some methane. 



In a similar manner the combustion of ethane would 

 involve the rapid passage through ethyl alcohol to acelalde- 

 hyde, and steam, with subsequent decomposition of the 

 aldehyde into carbon, hydrogen, methane, and carbonic 

 oxide, with the proviso that a reduction of the oxygen 

 supply below the equimolecular proportion would bring 

 about in some measure the decomposition of the alcohol 

 into ethylene and steam, &c., at stage i. 



H3C.CH3 — > CH3.CH2OH 



CH3.CH(OH)2 

 CHj.CHO + HoO 



/ C„H4+H„0 \ 



|.2C + 2Hj+H50/ 



r CHj + CO ~» 



But the cases of ethane and ethylene are typical of all 

 other hydrocarbons, so that it may be said that, in general, 

 the mechanism of explosive combustion involves (i) the 

 initial formation and subsequent decomposition of hydroxyl- 

 ated (or " oxygenated ") molecules ; (2) in a sufficient 

 supply of oxygen, the independent oxidation of the decom- 

 position products; (3) in an insufficient oxygen supply, the 

 subsequent breaking down of unsaturated hydrocarbons, 

 interactions between carbon and steam, or between oxides 

 of carbon, hydrogen, and steam, the final system depend- 

 ing on the amount of available oxygen, the temperature of 

 the flame, and the rate of cooling. 



Experiment V. — The influence of different rates of cool- 

 ing of the flame on the final system may be illustrated by 

 firing an equimolecular mixture of ethane and oxygen in 

 two glass vessels having approximately the same volume 

 but widely different surface areas. For this purpose I 

 have selected (i) a tube about i metre long and 2 cm. 

 internal diameter, and (2) a globe of 8-5 cm. internal 

 diameter. Both these vessels have the same volume (about 

 300 c.c), but the surface area of the tube is very nearly 

 three times that of the globe. It is therefore to be ex- 

 pected that, in consequence of the more rapid cooling of 

 the flame, there will be a greater accumulation of the 

 primary combustion products in the case of the tube 

 experiment. On comparing the results of the two ex- 

 plosions, it is at once evident that more water and less 

 carbon have been produced in the case of the tube ; more- 

 over, the pressure ratio />,//>, is 1-45, as compared with 

 about 1-75 in the globe experiment, and an examination 

 of the products would show that the lower ratio is 

 accounted for by the much greater survival of acetylene, 

 ethylene, and aldehydic products in the tube experiment. 

 These facts, which are set forth in the following table, are 

 in complete harmony with the hydroxvlation theory. 



Experiment VI. — The experiments I have so far shown 

 you refer more particularly to the initial period of " m- 

 flammation " in explosive combustion, that is to say, to 

 the conditions ordinarily prevailing in hydrocarbon flames. 

 The question may be asked whether or not the views I 

 have advanced are applicable to the extreme conditions of 

 " detonation " or of explosions under high initial pressures. 

 This question can best be answered by a consideration of 

 NO. 2055, VOL. 80] 



the behaviour of an equimolecular mixture of ethane and 

 o.xygen under these extreme conditions. 



Inflammation 0/ an Equimolecular Mixture of Ethane 

 and Oxygen. 



It is .difficult to set up detonation in this mixture ; the 

 gases must be fired at an initial pressure of about ij atmo- 

 spheres in a stout leaden coil of about i-inch internal 

 diameter. Even then it is necessary to start the explosion 

 wave in a special firing piece containing electrolytic gas 

 under pressure. I therefore regret that, owing to the 

 special arrangements requisite for success, it is not possible 

 to make the experiment to-night. I will, however, carry 

 out an experiment on the explosion of the gases at an 

 initial pressure of 15 atmospheres. 



The cylindrical steel bomb on the table is part of an 

 apparatus recently installed in the fuel and metallurgical 

 laboratories of the University of Leeds for investigations 

 on gaseous explosions under high pressures. The bomb is 

 about a foot long with an external diameter of 4 inches, 

 and the central cylindrical explosion chamber is 8 inches 

 long by I inch in diameter. It has been tested by hydraulic 

 pressure up to 1000 atmospheres, and has been repeatedly 

 used for experiments with mixtures of hydrocarbons and 

 oxygen at initial pressures of as much as 40 atmospheres. 

 The bomb is now connected, through a valve at the top, 

 with a standard Bourdon gauge, and contains an equi- 

 molecular mixture of ethane and oxygen at a pressure of 

 15-8 atmospheres. The valve will now be closed, and the 

 mi.xture fired by means of an electrical arrangement in 

 the special firing piece. 



All that is audible of the explosion is a sharp click, and 

 on opening the valve connecting with the gauge again 

 the final pressure of the cold products pf explosion is re- 

 corded. After applying the necessary correction for the 

 "dead space" in the gauge connections, the final 

 " corrected " pressure is as nearly as possible 30-8 atmo-, 

 spheres, corresponding to a ratio />,//>, = 1 '93. I would 

 now direct your attention to the tabulated results of a 

 similar bomb experiment carried out a few weeks ago at 

 Leeds at an initial pressure of 25 atmospheres, and also at 

 the same time to those of another experiment in which the 

 gases were detonated in a lead coil at an initial pressure of 

 li atmospheres. 



In both these experiments carbon was deposited, and it 

 is evident also that steam was formed. The ratio p^lp. 

 Avas as nearly as possible 2-0 instead of the 2.5 required 

 by the theory of the preferential combustion of carbon. 

 Moreover, a notable feature of the results is the presence 

 of as much as 7 per cent, of methane among the products 

 of the experiment at 25 atmospheres ; the fact that so 

 much methane survived when all other hydrocarbons were 

 battered to pieces during the explosion (no traces of either 

 acetylene or ethylene being found in the products) is a 

 remarkable testimony to its relatively great stability at the 



