Date of Award


Degree Type


Degree Name

Master of Science in Agricultural and Biological Engineering


Agricultural and Biological Engineering

First Advisor

Abigail S. Engelberth

Committee Chair

Abigail S. Engelberth

Committee Member 1

Mario G. Ferruzzi

Committee Member 2

Nathan S. Mosier


The market for bio-based products from plant sources is on the rise. There is a global challenge to implement environmentally clean practices for the production of fuels and pharmaceuticals from sustainable resources. A significant hurdle for discovery of comparable plant-derived products is the extensive volume of trial-and-error experimentation required. To alleviate the experimental burden, a quantum mechanics based molecular modeling approach known as the COnductor-like Screening Model for Real Solvents (COSMO-RS) was used to predict the best biphasic solvent system to purify silymarins from an aqueous mixture. Silymarins are a class of flavonolignans present in milk thistle ( Silybum marianum L.), which has been used in traditional eastern medicine to treat liver disease. More recently, silymarins have been studied as a cancer treatment therapy due to their antioxidant properties, but effective large-scale separation methods need to be developed. Previous research has shown that these compounds can be fractionated using centrifugal partition chromatography (CPC), but not to an acceptable level of purity. Due to previous incomplete fractionation, the silymarins are ideal compounds to assess the use of a molecular modeling approach to predicting partitioning in a CPC separation. The COSMO-RS method was implemented using the software programs HyperChem, TmoleX, and COSMOthermX in order to calculate partition coefficients for the six silymarin compounds in various solvent systems. The partition coefficient for each silymarin in each solvent system was verified by experimentation using the shake flask method and compared to the results of the model.