Significant evidence has shown that soot can be formed from polycyclic aromatic hydrocarbon (PAH) in combustion environments, but the transition of high molecular PAH from the gas phase to soot in liquid or solid state remains unclear. In this study, relationships between the boiling points of various planar PAH and their thermodynamic properties are systematically investigated, to find a satisfactory marker for the phase transition event. Temperature-dependent thermodynamic properties, including entropy, specific heat capacity, enthalpy and Gibbs free energy are simultaneously calculated for PAHs, using density functional theory and three composite compound methods. Comparing results indicate that the individual G3 method, plus an atomization reaction approach, produces the most accurate thermochemistry parameters. Compared to entropy, enthalpy, and Gibbs free energy, the specific heat capacity at 298 K is found to be a better marker for the boiling point of PAH, due to the observed linear correlation, predictable characteristics, and fidelity of accuracy as a function of temperature. The correlation equation: Y=10.996X+122.111 is proposed (where Y is the boiling temperature (K) and X is Cp at 298 K (cal/K/mol)). The standard deviation is as low as 16.7 K when comparing the calculated boiling points and experimentally determined values for 25 different aromatic species ranging from benzene to ovalene (C32H14). The effects of carbon number, structural arrangement, and partial pressure on the boiling point of large planar PAH are discussed. The results reveal that the carbon number in large planar PAH are the dominant factor determining its boiling points. It is shown that PAH containing about 60-65 carbon atoms are likely to exist as liquids in flames, although the partial pressure of such species is very low.
Bibliographical noteKAUST Repository Item: Exported on 2020-10-01
Acknowledgements: Research at King Abdullah University of Science and Technology (KAUST) was supported by the KAUST Clean Fuels Consortium (KCFC) and its member companies. The authors gratefully acknowledge the KAUST Supercomputing Laboratory (KSL) for providing the computing resources and technical support.