(466c) Molecular Dynamics Simulation And Thermodynamic Modeling Of The Self-Assembly Of The Triterpenoids Asiatic Acid And Madecassic Acid In Aqueous Solution Conference: AIChE Annual MeetingYear: 2007Proceeding: 2007 AIChE Annual MeetingGroup: Engineering Sciences and FundamentalsSession: Computational Studies of Self Assembly II Time: Wednesday, November 7, 2007 - 5:29pm-5:46pm Authors: Goldsipe, A. C., Massachusetts Institute of Technology Stephenson, B. C., Massachusetts Institute of Technology Blankschtein, D., Massachusetts Institute of Technology Asiatic acid (AA) and madecassic acid (MA) are bioactive phytochemicals with pentacyclic structures similar to bile salts. Although their surface activity was previously observed, little is known about the potential of these amphiphiles for self-assembly. To complement recent experimental investigations of their micellization behavior, AA and MA were studied by atomistic molecular dynamics computer simulations and molecular-thermodynamic modeling. In the simulation studies, AA and MA were initially randomly distributed in aqueous solution and simulated for 75 nanoseconds. Visual inspection of the simulations revealed spontaneous micelle formation within a few nanoseconds. Consequently, the simulation data was quantitatively analyzed to determine micelle properties, including aggregation numbers, principal moments of the radius of gyration, internal order parameters, and solvent accessible surface areas. These simulations clearly demonstrate the self-assembly of two non-traditional surfactants into dynamic micellar aggregates. Although the simulations provide significant information about micelle structure, it is difficult and computationally expensive to estimate accurately the critical micelle concentration (CMC) from simulation data. Inspired by the recent success of a hybrid model using both computer simulations (CS) and a molecular-thermodynamic (MT) model, a modified CS-MT model was developed to quantify the free-energy change associated with micelle formation in order to predict the CMCs of AA and MA. Although the modified CS-MT model overestimates the CMCs (by a factor of 3.5 for AA and by 50% for MA), it correctly predicts that AA has a lower CMC than MA. In contrast, the monomer concentrations in the simulation cells overestimate the experimental CMCs by a factor of 140 for AA and 180 for MA. These results demonstrate that the hybrid model of computer simulations and molecular-thermodynamic theory can provide detailed and quantitative information about micelles of novel surfactants with a fraction of the computational expense of computer simulations alone.