QSPR models for enthalpy and entropy of organic compounds based on a set of norm indices

Abstract

In this study, quantitative structure–property relationship (QSPR) models are developed to predict the standard vapourisation enthalpy (ΔvapH°), enthalpy of fusion (ΔfusH), entropy of fusion (ΔfusS), and standard sublimation enthalpy (ΔsubH°) of organic compounds. Molecular size and shape parameters describing the structural information are introduced to improve the accuracy and robustness of the predictive effect. The correlation coefficients (R2) for ΔvapH°, ΔfusH, ΔfusS, and ΔsubH° are 0.981, 0.851, 0.888, and 0.844, respectively, and these results indicate the high accuracy of models. Because some experimental values of the ΔfusH and ΔfusS are less than average absolute error (AAE) of QSPR models, which leads to negative values by linear models, nonlinear models are proposed to mitigate this problem. Additionally, a comparison with previously published results further demonstrates that these QSPR models can achieve high precision for data set comprising numerous chemical substances. From a holistic perspective, the novel QSPR models can be widely used in multiple domains, such as chemical reaction and predesign prediction.