SVRC–QSPR model for predicting saturated vapor pressures of pure fluids

Knowledge of thermo-physical properties of organic chemicals is essential to chemical and process design applications. Vapor pressure is one such property used directly in process calculations and as input to property-prediction models. Although experimental determination of vapor pressures remains an option, often it is not possible to measure vapor pressure data experimentally for toxic or yet to be synthesized molecules. Current vapor pressure models, which utilize traditional physical properties as inputs, are limited by their range of applicability and/or by poor suitability for generalization. Further, recent quantitative structure–property relations (QSPR) models for vapor pressure have been limited to single-temperature generalizations (e.g., 298 K); thus, the distinct advantages offered by advances in computational chemistry as they relate to structure–property model generalizations have not been fully realized. s study, we present an integrated approach for developing a generalized model which is capable of predicting accurately the vapor pressure of organic chemicals over the entire saturation range (the triple point to the critical point). The approach uses a theoretical framework to develop the fluid behavior model and QSPR to generalize the parameters of the model. Specifically, we first apply our scaled variable reduced coordinates (SVRC) model to a diverse dataset containing over 1221 molecules involving 73 classes of chemicals, and then we generalize the SVRC parameters using structure–property (SP) models. For this modeling effort, reliable experimental vapor pressure data were obtained from the DIPPR database. The results for 52,445 data points indicate that: (a) the SVRC model represents these saturated vapor pressure data with 0.35% average absolute deviation (AAD), and (b) the generalized SVRC–QSPR model predicts the saturated vapor pressures with 0.5% AAD.