Knowledge of the physical properties of organic compounds is necessary for the design and optimization of various standard chemical processes. Further, the normal boiling point can be used to ascertain relative volatilities of compounds for separation purposes and to determine the state of a compound. Since the normal boiling point for some substances is not easily measured by experiment, which may be due to the hazardous nature, expense, or physical characteristics of the chemical, a predictive method for the normal boiling point of complex or heavy hydrocarbon compounds is warranted. Utilizing 1280 experimental normal boiling point (NBP) measurements for compounds found in the DIPPR database, a quantitative structure-property relationship (QSPR) NBP prediction model was developed. This nonlinear model provides an effective means for determining NBP of organic chemicals including complex and higher boiling substances.

The hypothesis for this work was the utilization of an approach employing cause-and-effect to determine the extent of a given descriptor in accounting for variations in molecular volume, area, shape, polarity, association (VASPA), etc. of a molecule, rather than attempting to model the properties using QSPR directly. The quality of the predictions obtained for this diverse group of molecules demonstrates the validity of an integrated approach and provides credible evidence to support the above hypothesis. A 12 descriptor model resulted in absolute average deviation (%AAD) of 2.13 with a root mean square error of 16.7 K. After examination for outlying results and their subsequent elimination, the model had a %AAD of 1.7 and a RMSE of 9.8 K.