Estimation of the flash points of saturated and unsaturated hydrocarbons

This work introduces a novel, reliable and simple correlation for predicting the flash point of various types of saturated and unsaturated hydrocarbons containing cyclic and acyclic paraffins, olefins, alkynes and aromatic hydrocarbons. It is shown that the new method is based on the number of carbon and hydrogen atoms and some specific molecular moieties, which can easily be used for any types of hydrocarbons. The largest available experimental data consisting of 441 diverse hydrocarbons are used here in order to derive and test the general correlation. For a data set containing 423 of these 441 hydrocarbons, the root mean square (rms) and average absolute deviations are 7.7 and 5.7 K, respectively. The estimated flash points for 18 further hydrocarbons containing complex molecular structures have been compared with one of the best available new group additivity methods, which give much lower values of the rms and average absolute deviations.