Quantitative Structure Property Relationship Models for the Prediction of Gas Heat Capacity of Benzene Derivatives Using Topological Indices

摘要

Graph-theoretical approach represents simple and efficient means to quantitative structure property relationship (QSPR) studies. Graph-theoretical topological indices are high potential descriptors for modeling and predicting physicochemical properties of chemical compounds. A QSPR study was performed for prediction of gas heat capacity (C-v) of 69 benzene derivatives using Wiener (W), Szeged (Sz), first order molecular connectivity ((1)x), Balaban (J), hyper-Wiener(WW), Wiener polarity (W-P) and Harary (H) topological indices. The calculation was performed by the ab-initio method at HF/6-31G(d) level of theory. The relationship analysis between heat capacity (C-v) and topological indices was done by using multiple linear regression (MLR) method, with heat capacity (C-v) as dependent variable and seven independent variables to generate the equation that relates the structural features to the heat capacity (C-v) properties. The results show good models with three-seven parameters linear equations. The best model in this study is contains three descriptors ((1)chi,W, Sz) are included, with values of the correlation coefficient (r=0.943), the standard error (s=9.667Jmol(-1)K(-1)), the Fisher-ratio (F=172.475), the adjusted coefficient of determination (r(adj)(2) = 0.883) and Durbin-Watson value (D=1.728), which indicate that these descriptors, play an important role in effect on heat capacity (C-v) of benzene derivatives.