When the homogeneous thermal decomposition of nitrous oxide was first studied in connection with the theory of gaseous reactions, the principal problem was to decide whether the activation of the molecules occurred independently of collisions, as would have been required by the radiation theory of activation. The influence of pressure on the rate of reaction showed definitely that the activation depended on a collisional process, in which sense the reaction proved to be bimolecular. The characteristic of an ideal bimolecular reaction is that the time of half change should be inversely proportional to the initial pressure. It was in fact found that the reciprocal of the half change period when plotted against initial pressure gave a straight line, which, however, did not pass through the origin. This meant that at low pressures a reaction of the first order was occurring, as well as the bimolecular change. This first order reaction was not further investigated, as it seemed quite possible that it was a surface reaction, the intrusion of which became relatively more serious as the pressure fell. It was observed, furthermore, that the complete course of a decomposition at a given initial pressure was not represented very well by the usual bimolecular equation; this, however, was capable of explanation in terms of an autocatalytic effect of the by-products of the reaction, since small amounts of the higher oxides of nitrogen were known to be formed in addition to the oxygen and nitrogen constituting the main products. More recently two new observations have been made, rendering desirable a fuller investigation of some of the details about the reaction, which have hitherto been regarded as of less importance than its general interpretation in terms of the collisional mechanism.The first of these is the observation of Volmer and Kummerow that, at low partial pressures of nitrous oxide, inert gases exert an accelerating influence on the decomposition. This suggests that the low pressure unimolecular part of the decomposition is perhaps really homogeneous, and also of the “quas-unimolecular type” which is subject to the influence of foreign gases. The second of the observations referred to is that of Voliner and Nagasako, who state that, between 1 and 10 atmospheres, the whole decomposition becomes of the first order. Thus the second order reaction observed in the earlier experiments, which were not carried out at pressures greater than an atmosphere, would be the low pressure part of a quasi-unimolecular reaction, The difference in mechanism between a true bimolecular reaction and the quasi-unimolecular reaction would be simply that in the former the nitrous oxide reacts at the moment of collision, while in the latter it survives the activating collision for a definite period and then splits up spontaneously into N
2
and an oxygen atom, unless in the meantime it has been deactivated.
Prediction of the Progressive Loss of Winding Clamping Pressure in Power Transformers