Unimolecular reactions possess a unique interest in that, as Perrin (‘Ann. Physique,’ vol. 11, p. 5, 1919) first pointed out, for the occurrence of such, some type of interaction between radiation and matter must take place. Although such reactions appear to be extremely rare, many physical processes such as evaporation, ionisation in gases at high temperatures and radio-active decay, proceed at rates conforming to a unimolecular law; true chemical reactions which are definitely unimolecular and not pseudo-unimolecular in character are, on the other hand, stated by many (
e. g
., Lowry, ‘Trans. Farad. Soc.,’ vol. 17, p. 596 (1922) ) to be non-existent. In order to substantiate this statement, it is clearly necessary to prove the more complex character of any reaction which satisfies the usual criteria of unimolecular change. The thermal decomposition of gaseous nitrogen pentoxide apparently fulfils these conditions, for Daniels and Johnston (‘J. Am. C. S.,’ vol. 43, p. 53 (1921)) showed that the reaction proceeded according to a unimolecular law over wide ranges of variation of pressure, and Lueck (
ibid
., vol. 44, p. 757 (1922)) obtained practically identical unimolecular constants for the decomposition in solution in carbon tetrachloride and chloroform. On the other hand, Daniels, Wulf and Karrer (
ibid
., vol. 44, p. 2402 (1922) ) suspected the reaction to be autocatalytic, owing to the apparent retardation of the reaction velocity in the presence of ozone, but the experiments of one of us (Hirst, ‘J. C. S.,’ vol. 127, p. 657 (1925), and of White and Tolman (‘J. Am. C. S.’ vol. 47, p. 1,240 (1925)) proved this to be erroneous. In addition, it has been shown that the reaction proceeds uniformly according to the unimolecular law even in the presence of extensive glass surfaces, or of gases which may be either indifferent, such as argon and nitrogen, or the products of reaction, such as nitrogen tetroxide or dioxide or oxygen. The rate of reaction may be expressed in the form -
d
C/
dt
= 4·98 × 10
13
e
-24.700/RT
. C. Attempts have been made to interpret the experimental results on the hypothesis that the reaction is in reality bimolecular, and only apparently unimolecular in character; but owing to the abnormally large value of the energy of activation, namely, 24,700 calories per gram. molecule, the number of molecules which could be activated per second by inelastic collision, calculated according to the kinetic theory, falls far short of the observed reaction rate, being, in fact, some 10
5
times smaller.
The thermal decomposition of nitrogen pentoxide at low pressures