(382r) Improve Understanding of the Temperature-Dependence of Liquid Viscosity of Organic Compounds through Quantitative Structure-Property Relationship and Correlation

Liquid viscosity is an important parameter in process design. Ideally, experimental data for any given compound would be available to aid in the design. However, experimental data for liquid viscosity can be hard to come by. The DIPPR database, which has data for 47 different thermophysical properties for more than 2300 compounds, only has experimental liquid viscosity values for just under half the compounds. That leaves a majority of the compounds in the database have predicted liquid viscosity values. Prediction methods, which are largely group contribution methods, do a reasonably good job below the normal boiling point for compounds that contain the same groups used to develop the method. For a wide range of compounds, prediction methods have proven to be inadequate. Recent work in the field of Quantitative Structure-Property Relationship (QSPR) has shown promise in improving liquid viscosity prediction. These QSPR models have been developed for predicting liquid viscosity at a given temperature (usually 298.15 K). This work seeks to go a step further by taking a QSPR model at a reference liquid viscosity and adding temperature dependence to better predict liquid viscosity over a range of temperatures.