The kinetics of reactions in solution. Part I.—A comparison of the decomposition of chlorine monoxide in the gaseous state and in carbon tetrachloride solution

Abstract

In general the rate of a chemical reaction is profoundly influenced by the solvent in which the reaction takes place. In Menschutkin’s experiments on the combination of tertiary amines and alkyl halides, for example, the reaction velocity in benzyl alcohol was several hundred times greater than that in hexane. While it is clear that the influence of the solvent, which cannot be correlated definitely with any physical properties, belongs to the category of specific chemical effects, it is not easy to see in exactly what it consists. Christiansen and Norrish and Smith have found that the rate of a bimolecular reaction in solution appears usually to be several powers of 10 smaller than that of a hypothetical reaction occurring in the gaseous phase with the same energy of activation. Norrish and Smith therefore suggest that, since each encounter between molecules of the reacting substances takes place with a solvent molecule in very close proximity, and is thus virtually a ternary collision, the influence of the solvent is primarily a deactivating one. On the other hand, the decomposition of nitrogen pentoxide—a unimolecular reaction—takes place at the same rate and possesses the same heat of activation in chloroform, in carbon tetrachloride and a number of similar solvents as it does in the gaseous phase. In certain other solvents, quite different rates and heats of activation are found. Daniels regards as “normal” those solvents in which the reaction occurs at the same rate as in the gaseous phase, and attributes the deviations found with other, “abnormal,” solvents to a definite formation of complex molecules whose stability differ from that of free nitrogen pentoxide. The isomerisation of pinene has also been found to take place at the same rate in the gaseous and in the liquid states. As far as this not very abundant evidence goes, then, it would suggest that bimolecular reactions are usually much slower in solution than in the gaseous phase, while unimolecular reactions tend usually to be uninfluenced by the presence of the solvent. This simple contrast, however, is by no means correct. In the first place, the variation in the rates of unimolecular reactions with change of solvent appears on closer inspection to be not less marked than that of bimolecular reactions. This aspect of the matter is dealt with more fully in the following paper. In the second place, there have hitherto been no data available for comparing the rate of a bimolecular reaction in the gas phase with its rate in solution in one of the relatively inert solvents which had been used for nitrogen pentoxide. It is shown in the present paper, however, that the decomposition of chlorine monoxide, which is essentially bimolecular, although somewhat complex in mechanism, occurs at almost exactly the same rate in carbon tetrachloride as in the gaseous state, and that the heat of activation is substantially the same. Thus it appears that, contrary to first impressions, the kinetic type of the reaction has little to do with the influence of solvents. The chemical nature of the solvent is the determining factor in all cases. Both for bimolecular and unimolecular reactions there appear to exist “ideal” solvents, of which carbon tetrachloride is a good example. This is being confirmed by Mr. Bowen and the writers by measurements on the rate of decomposition of ozone in carbon tetrachloride solution. For reactions that cannot be made to take place in the gas phase at all, and this includes the majority of measurable reactions, the rate in carbon tetrachloride or an analogous solvent would then appear to be a standard or ideal rate equal to that which the gas reaction would possess. Comparison of the rate and the energy of activation of a given reaction in any solvent with the corresponding values for the reaction in carbon tetrachloride thus provides a means of ascertaining whether the solvent has increased or decreased the normal heat of activation, and whether its effect is to be regarded as deactivating or positively catalytic. The similarity of behaviour of chlorine monoxide in the gaseous phase and in carbon tetrachloride solution suggests that, for bimolecular as well as for unimolecular reactions, solvents must be divided into two classes, and that the deactivating influences postulated by Norrish and Smith are as specific as the influence of “abnormal” solvents on unimolecular reactions.