## Abstract

By new experimental studies on hydrocarbon oxidation, taken in conjunction with results already available, an attempt is made to form a systematic picture of the following phenomena. (a) The complex dependence of rate upon hydrocarbon and oxygen pressures: this varies in a complicated way from substance to substance and with the experimental conditions. In the region of relatively low temperature, if the maximum rate is set approximately proportional to [RH]$^x$ [O$_2$]$^y$, x varies from 3 to 1 and y from + 1 to -1. An inverse correlation between x and y is shown over a range of combustible substances. (b) The highly characteristic form of the temperature dependence: there are often regions where the rate falls as the temperature increases. (c) The way in which the reaction orders change with temperature: the x and y of the above equation may rise above the values given when the temperature is increased. After fairly sharp maxima they fall agam as the temperature is further increased. (d) The reactant concentration dependences of $\pi$, the ratio of reaction rate to peroxide concentration: this varies with rather less than the first power of [RH] and with a small negative power of [O$_2$]. (e) The reactant concentration dependences of t*, the time to the attainment of the maximum rate: this increases with [RH] and decreases with [O$_2$]. These phenomena can be correlated in terms of a reaction mechanism based on a combination of simple and well-documented steps: $(1) \mathrm{RH} + \mathrm{O}_2 \rightarrow \mathrm{R}^. + \mathrm{HO} ^._2, {2}R^. + \mathrm{O}_2 \rightarrow \mathrm{RO}^._2$, (3) $\mathrm{RO}^._2 + O_2 \rightarrow$ end of chain, (3) $\mathrm{RO}^._2 \rightarrow$ end of chain, $\mathrm{RO}^._2 + \mathrm{RH} \rightarrow \mathrm{ROOH} + \mathrm{R}^., (5) \mathrm{ROOH} \rightarrow \alpha\mathrm{R}^., (6)\mathrm{ROOH} \rightarrow$ end of chain, (7)$\mathrm{R}^. + O_2 \rightarrow$ end of chain, (7') $\mathrm R \rightarrow$ end of chain. With the assumption of the attainment of stationary states and with the introduction of certain approximations, the scheme leads to the following general expression for the maximum rate of reaction: $\rho_{\max.} = \frac{C_1[RH]^3 [O_2]}{w(w-C_2[RH])},$ where $w = a+b[O_2]+c[\mathrm{RH}]+a'/[\mathrm{O}_2]+b'+c'[\mathrm{RH}] [\mathrm{O}_2]$ and [RH] and [O$_2$] refer to reactant concentrations. a, a', b, b',c, c', C$_1$ and C$_2$ are composite rate constants. According to the relative values of the constants the most diverse types of behaviour appear as special cases covered by the general expression. Thus the apparently contrasting results of much published work can be understood. At lower temperatures (e.g. 280 $^\circ$C) the phenomena can be mostly explained by the retention in the denominator of only those terms including a, b and c. With increasing temperature the relative influence of the terms in C$_2$ and a' must be assumed to increase in order to account for the increase in reaction orders referred to under (c). With still further increase in temperature the fall-off in reaction orders and rate must be attributed to the eventual emergence in the denominator of a concentration-independent term as the dominant one. This is probably associated with a predominant importance of process 3'. When this spontaneous destruction of peroxy radicals goes far enough, the 'low temperature' reaction is snuffed out and the field is left clear for the 'high temperature' reaction.