470 



NATURE 



[September 5, 1907 



7///1 K.M'I.OSION OF GASES. 



THK eailust woiU on ihc explosion of gasc-s was 

 that of lliiniphry Davy, who in 1817 publislicd 

 tliose celebrated experimenls on " the propagation of 

 llanie through small tubes and orifices " which led 

 him to the construction of the miner's safety-lamp. 



More than half a century later Bunsen devised the 

 in>n-luminous gas burner, observing that unless the 

 llow of the mixture of coal-gas and air exceeded a 

 certain rate the flame became unsteady and passed 

 down the tube. Hunsen believed that this rate repre- 

 sented the velocity with which an explosion would 

 travel in the combustible gases in a closed tube, and 

 he obtained definite values for a number of mixtures 

 by leading the gases through an orifice at the end of 

 a tube, igniting the jet, and determining the minimum 

 speed at which the g.-ises must be forced through the 

 tube to prevent the flame passing back through the 

 opening. The rates of explosion measured in this 

 way were comparativclv slow, the fastest observed 

 being about thirty- 

 '' '■ '' hcven yards a second. 



Hut' in i88i Ber- 

 thelot and Vieille 

 discovered that when 

 an explosive mixture 

 is ignited at the end 

 of a long pipe, the 

 \elocity of the explo- 

 sion rapidly increases 

 from its point of 

 origin until it reaches 

 ,•1 maximum velocity, 

 which remains con- 

 stant however long 

 the column of gas 

 may be, and which 

 greatly exceeds the 

 speeds of combus- 

 lion measured by 

 Bunsen ; this dis- 

 coverv was confirmed 

 bv the independent 

 investigations o f 

 Mallard and L e 

 Chatelier, published 

 at the same time. 

 Berthelot gave the 

 name " ronde explo- 

 < n.ime -^ sive" (detonation- 



Fic, I. wave) to the flame 



travelling with its 

 maximum velocity, thus distinguishing it from 

 the variable progressive combustion which precedes 

 its development. The velocity of the explosion- 

 wave constitutes a physical constant which has .'i 

 specific value for each inflammable mixture; measure- 

 ments by Berthelot and H. B. Dixon have .shown 

 that it is approximately equal to the velocity of sound 

 in the burning gases at the temperature of the explo- 

 sion. For a mixture of hydrogen and oxygen in 

 equivalent proportions the velocity is about 3000 yards 

 a second. 



Mallard and Le Chatelier succeeded in recording 

 the slow movements of the flame of progressive com- 

 bustion by photographing the fl.ish on a piece of 

 sensitised p.iper fixed on a revolving cylinder. They 

 found that when the gases are ignited at the open 

 end of a long tube, the flame travels for some distance 

 with a uniform slow velocity of the order measured 

 by Bunsen ; the flame next begins to vibrate, swing- 

 ing backwards and forwards with oscillations of 

 increasing .-miplitude ; then it either dies down or 

 sometimes the gas detonates. If the gas is fired near 

 NO. 1975. VOL. 76] 



the closed end of the tube, the movement of the flame 

 is uniformly accelerated until the detonation is set 

 up. Le Chalelier's apparatus was not fast enough 

 to analyse the wave of detonation itself. 



The apparatus used by Prof. H. B. Dixon consists 

 of a drum carrying .a narrow strip of Eastman film, 

 which can be rotated at the rate of 100 metres a 

 second. The explosion tube is fixed horizontally, and 

 the image of the fl.inie is focussed on to the vertically 

 moving film. The photographs show an inclined line 

 of light compoinided of the two motions. Fig. i is 

 a photograph of the explosion of cyanogen with 

 oxygen. The mixture was fired near the middle of 

 a tube by an electric spark s. The flame moves 

 .slowly in both directions; to the left it passes out of 

 the field of view, to the right its speed increases until 

 at X the detonalion-wave is set up. The detonation- 

 wave, moving with constant velocity, is represented 

 bv a straight lino {x, d), while the slower movement 

 of the progressive combus- 

 tion preceding it is shown 

 as a curve (s, x), the steep- 

 ness of which diminishes 

 as the motion of the flame- 

 accelerates — the speed ol 

 the drum being uniform. 

 The duration of the flash 

 was less than i/ioo second. 

 The period before the de- 

 tonation is dislinguishcii 

 not only by the slow move- 

 ment of the flame, but also 

 by slow and incomplete 

 combustion and feeble 

 luminosity. 



The initiation of the de- 

 tonation-wave is marked by 

 certain characteristics — (i) 

 a sudden increase in intens- 

 ity of the flame, accom- 

 panied by an instantaneous 

 rise in pressure; it is found 

 that glass tubes are most 

 often fractured at the point 

 where the detonation ori- 

 ginates; (2) rapid and 

 complete combustion ; {_),) 

 the setting up of a strongly 

 luminous backward wave 

 (.r, r) — the so-called " re- 

 ton.'Ltion-wave " — vv h i c h 

 under certain conditions 

 travels ;is rapidly as the 

 detonation-wave itself. The Fn.. 2 



sudden rise in pressure is 



due to the increase of chemical action, and this pres- 

 sure not only produces the forward detonation, but also 

 sends a backward wave of compression into the slowly- 

 burning gases behind it; this compression-wave raises 

 the temperature of the combining gases and increases 

 the luminosity. It should be observed that the light 

 produced by the explosion is chiefly due to particles 

 knocked from the glass and raised to incandescence ; 

 the small particles suspended in the burning gases 

 glow by the heat imparted to them bv the hotter but 

 invisible gaseous products of combustion. 



The detonation-wave is set up only after the flame 

 has run sotiic distance, which depends on the nature 

 of the mixture .and on the size of the spark. 



Fig. 2 shows the explosion of hydrogen and oxygen 

 in a closed glass tube too short to allow of the de- 

 tonation being set up. The gas is fired in the middle 

 of the tube, and the flame spreads right and left with 

 faint luminosity. The fl.'ime is preceded by an in- 

 visible compression-wave which travels with the 



