Are the Sublimation Thermodynamics of Organic Molecules Predictable

摘要: We compare a range of computational methods for the prediction sublimation thermodynamics (enthalpy, entropy, and free energy sublimation). These include model from theoretical chemistry that utilizes crystal lattice minimization (with DMACRYS program) quantitative structure property relationship (QSPR) models generated by both machine learning (random forest support vector machines) regression (partial least squares) methods. Using these we investigate predictability enthalpy, entropy sublimation, with consideration whether such method may be able to improve solubility schemes. Previous work has suggested major source error in schemes involving thermodynamic cycle via solid state is modeling change away state. Yet contrary this conclusion other found inclusion terms as enthalpy QSPR ...