366 



NA TURE 



[September 8, 1923 j 



the flame, a circumstance which seems to be of some 

 significance in relation to the mechanism of carbon 

 monoxide combustion. In conjunction with Prof. A. 

 Fowler, of the Imperial College, South Kensington, we 

 are now investigating it more closely with the view of 

 finding out its meaning. But the facts known warrant 

 us in concluding that the addition of a comparatively 

 small proportion of hydrogen has a peculiar influence 

 upon the combustion of carbon monoxide, whether at 

 high pressures (as in our bomb experiment) or in flame 

 combustion at ordinary pressures. 



The Mechanism of the Combustion of Carbon 

 Monoxide. 



To explain the peculiar influence of hydrogen or 

 steam upon the combustion of carbon monoxide, I 

 think it must be supposed that oxygen and carbon 

 monoxide molecules are mutually inert in flames, and 

 that before the carbon monoxide can be oxidised the 

 O2 molecules must be resolved either into atoms or 

 into " activated " steam. This precedent condition 

 can be brought about by the presence of hydrogen 

 (or maybe steam) in the mixture undergoing com- 

 bustion. For, according to my present view, an 

 undissociated O2 molecule on being heated in the 

 flame has its " residual affinities " sufficiently stimu- 

 lated to enable it to seize upon two hydrogen molecules, 

 forming initially an unstable vibratory complex H^Og. 

 Such a complex, being in an intensely vibratory con- 

 dition, would instantly break down (i) -partly into two 

 molecules of steam, also in a vibratory (and therefore 

 " activated ") condition, and (2) partly also into two 

 : atoms and two Hg molecules, thus : 



SH \ would give rise to 



j (I) « 



x2 : O 



/ 



H 



\H 



N/h; (simultaneously) | /,) (100 -«) (2O : +2H,). 



The ratio n/(ioo - n) would obviously depend upon 

 both temperature and environment. The higher the 

 temperature and the less hydrogen in the environment 

 the less the magnitude of n. But in all conditions the 

 hydrogen in a combustible mixture containing also 

 carbon monoxide functions as a resolver of Og mole- 

 cules simultaneously into (i) " activated " steam and 

 (2) : O atoms. Thus it is suggested that the primary 

 function of hydrogen as a promoter of the combustion 

 of carbon monoxide is to resolve the O2 molecules 

 (inert towards carbon monoxide) into : atoms and 

 " activated " OHg (reactive towards carbon monoxide), 

 itself being continuously regenerated in the process, as 

 is shown in Fig. 2. 



H 



^0 



^-H 





CO^=^CO, 



Fig. 2. 

 NO. 2810, VOL. 112] 



It may be observed that this view is similar to tl' 

 one advanced forty years ago by Prof. H. B. Dixo: 

 to explain his discovery of the mutual inertness of dr 

 carbon monoxide and oxygen in flames, but —"'<'■' 

 in one particular so as to make it more 

 to the further facts now known. He supjM)-(u lu 

 carbon monoxide is oxidised by OH, (but not by O.. 

 molecules in flames, the resulting hydrogen 

 mediately burnt to steam, which was thus c</. 

 regenerated, as follows : 



/(a) CO + OH,=CO, + H,) 

 \{h) 2H,+ 0,=2H,0 /• 



If, however, only such interactions (and no other- 

 occur, it is difficult to understand why the colour ani 

 spectrum of a flame of pure (moist) carbon monoxid 

 are so unlike those of hydrogen burning in air. Th- 

 characteristic spectrum of a carbon monoxide flam< 

 which extends far into the ultra-violet, would sr.-i i , 

 seem to be due to the formation in it of some < < »., 

 molecules in a more highly vibratory state than wou! 

 be likely to arise merely by interactions of CO an 

 OHg molecules. The difficulty in question is obviatec 

 and also other facts would be better explained, b 

 supposing (as I do) that an unstable vibratory H/ ' 

 complex, primarily formed by the interaction of (' 

 and Hg molecules, decomposes in each of two way 

 yielding : OHg and : O atoms, both of which are capable 

 of oxidising carbon monoxide. 



i 



The Energy- absorbing Function and Activatic'^ 

 of Nitrogen in the Combustion of Carbon 

 Monoxide. 



It next occurred to us to tr>'the effects of progressively 

 replacing the nitrogen of a normal carbon monoxicii 

 (2CO + O2 + 4N2) mixture by molecular equivalents ot 

 other gases, e.g. oxygen, carbon monoxide, or argon. 

 The first two of these gases are diatomic, and woul 

 have much the same densities and heat capacities ii 

 the nitrogen which they replaced ; and although the 

 might be expected to exert some " chemical mass 

 influence upon the combustion, yet in all other respect 

 they would act as " diluents." In argon we had an 

 absolutely inert monatomic gas of higher density, but 

 smaller volumetric heat capacity, than nitrogen, and 

 incapable of any internal vibrational energ)'. It would 

 therefore presumably be incapable of exerting any 

 effect upon the explosion other than that of mereh 

 sharing, by molecular collisions, in the increased kineti 

 energy acquired by the system as the result of tht 

 combustion. 



It may be observed that while the said replacement 

 of the nitrogen by the other gases would not affect i: 

 any way the total energy liberated on explosion, y*. i 

 the experiments showed that it affected somewhat the 

 proportion of the energy recorded by the gauge as 

 pressure (temperature) at the instant of maximum 

 pressure, and still more so the rapidity with which 

 the said pressure energy was developed. The most 

 important experimental results from this point of xiew 

 are summarised in the following table, and illustrated 

 by the set of pressure-time curves reproduced in Fig. 3. 

 Here it may be pointed out that the most essential 



