A model that yields chemical accuracy for a broad range of organic molecules is presented. The range of applicability of such an accurate model is very broad: it can be used by chemists to predict equilibria while fostering new chemistries and allow process engineers to make more reliable designs. The model which is group-contribution (GC) based, estimates gas phase standard enthalpy of formations (ΔfH°gas) of organic compounds. To achieve the chemical accuracy, a systematic property-data-model analysis, which allows efficient use of knowledge of the experimental data of ΔfH°gas and the molecular structural information is employed. Based on the findings of property-data-model analysis, new structural parameters are defined and included in the GC-model to provide additional structural information for compounds having large correlation errors and to thereby improve accuracy of ΔfH°gas predictions through better correlation of data. For parameter estimation, a data-set containing 861 experimentally measured values of a wide variety of organic compounds (hydrocarbons, oxygenated compounds, nitrogenated compounds, multi-functional compounds, etc.) is used. The developed property model for ΔfH°gas is fully predictive and is based exclusively on the molecular structure of the organic compound. Compared to other currently used property prediction methods, the developed GC-model for ΔfH°gas provides significant improvement in accuracy with an average absolute error of 1.75kJ/mol and standard deviation of 2.61kJ/mol.
Fluid Phase Equilibria, 2013, Vol 348, p. 23-32
Enthalpy of formation; Group-contribution method; Quantum mechanical method; Chemical accuracy; Benson method