From theoretical energy barriers to decomposition temperatures of organic peroxides

Stimulated by the needs arising from new regulatory frameworks, a number of equations have been proposed in recent years in view of predicting onset decomposition temperatures () of chemical compounds on the basis of quantitative structure-property relationships (QSPRs). The present study investigates an alternative route based on theoretical activation energies. For a set of 38 peroxides, this strategy outperforms state-of-the-art QSPR models, thus providing an attractive alternative to predict the stability of new compounds. Furthermore, it yields valuable insight into reaction mechanisms, pointing to a concerted decomposition pathway less energetically costly than the genuine O-O homolytic bond scission for dialkylperoxides with an ester group in position with respect to the peroxide group. In addition, a model similar in spirit to group contribution methods is put forward. The interest of the present approach is further illustrated by a comparison between present predictions and self-accelerating decomposition temperatures. This work demonstrates some limitations of popular QSPR methods for reactive hazards and should encourage less empirical approaches based on the systematic exploration of decomposition mechanisms. Indeed, according to present results, the key to improved models for predicting the thermal stability of organic compounds clearly lies in properly accounting for the most relevant paths to decomposition.