Calculation Of Solvation Free Energy By Multi-scale Modeling

Solvation free energy is the change of free energy during solvation process.It is used to describe the possibility of solute molecules be found in liquid phase in equilibrium.A number of thermodynamics properties are related to solvation,including partition coefficient,solubility,and association/dissociation constant et al.Calculation of gas-water phase exchange is of great importance for research on persistent organic pollutants(POPs).In addition,the partition coefficient of a molecule between water and 1-octanol(Poct-wat)is frequently used to study pharmaceutically related systems,evaluate adsorption,transfer,and excretion of pollutants and druglike compounds from a living organism.Meanwhile,precise calculation of solvation free energy is fundamental to simulations of many biological and chemical processes.Multiscale models including quantitative structure property relations(QSPR),Poisson-Boltzmann(PB),generalized Born model(GB),solvation model based on density(SMD),generalized amber force field(GAFF)and the conductor-like screening models for realistic solvation(COSMO-RS)were considered in this thesis in order to calculate solvation free energy of small neutral organic molecules in a variety of solvents.The predict power of different models was evaluated as well.In the third chapter,solvation free energy of small neutral organic molecules in organic solvents with a total number of 228 systems was calculated.The solvation free energy calculated with GAFF using the thermodynamics integration(TI)method was compared with experimental data,and with QSPR and COSMO-RS.All three models give a RMSD from experimental data lower than 4 kJ/mol.COSMO-RS yields significantly better performance than the other models after applying a ring-term correction,the physical meaning of which is unknown.For TI calculation,deviations from experiment occur when compounds containing nitro or ester groups are solvated into other liquids.This finding is consistent with results of properties calculated with pure liquids.In the fourth chapter,we compared the predict power of PB,GB(Still,HCT,OBC-I,OBC-II),SMD(at four different levels of theory)with explicit solvent model in GAFF on calculation of solvation free energy of organic molecules in organic solvents.PB and GB give poor agreement with explicit solvent model,and even worse with experimental data(RMSD is 14-15 kJ/mol).The main problem seems to be the prediction of the apolar contribution,that solvent entropy should be included.The quantum based SMD model gives significantly better agreement with experimental data than PB or GB,but not as good as explicit ones.The dielectric constant ￡ of the solvent is found to be a powerful predictor for the polar contribution,however,the Onsager relation may not hold for realistic solvent as suggested by explicit solvent and SMD calculations.From the comparison,we also find that with an optimization of the apolar contribution,the PB model gives slightly better agreement with experiments than the SMD model,whereas the correlation between the optimized GB models and experiments remains poor.Further optimization of the apolar contribution is needed for GB models to be able to treat solvents other than water.In the fifth chapter,we compared BAR(Bennett acceptance ratio),PB with TI for solvation free energy calculations of 95 compounds that include 5 main classes of pollutants.Neglectable difference was found between TI and BAR calculations with our simulation approach.The RMSD of PB from TI calculation is approximately 15 kJ/mol.A further analysis of each class of pollutants indicates that for most compounds(dioxins and PCBs),PB gives reasonable results,while large deviations were found for estrogen analogues,phthalates and veterinary medicine.A decomposition of solvation free energy into polar and apolar contributions indicates that in polar contribution,large deviation is associated with stronger solute-solvent interaction,and apolar contribution is the main factor to improve the accuracy of PB.