Predictive methods have been employed to characterize chemical separation mediums including solvents and absorbents. These studies included creating Abraham solvation parameter models for room-temperature ionic liquids (RTILs) utilizing novel ion-specific and group contribution methodologies, polydimethyl siloxane (PDMS) utilizing standard methodology, and the micelles cetyltrimethylammonium bromide (CTAB) and sodium dodecylsulfate (SDS) utilizing a combined experimental setup methodology with indicator variables. These predictive models allows for the characterization of both standard and new chemicals for use in chemical separations including gas chromatography (GC), solid phase microextraction (SPME), and micellar electrokinetic chromatography (MEKC). Gas-to-RTIL and water-to-RTIL predictive models were created with a standard deviation of 0.112 and 0.139 log units, respectively, for the ion-specific model and with a standard deviation of 0.155 and 0.177 log units, respectively, for the group contribution fragment method. Enthalpy of solvation for solutes dissolved into ionic liquids predictive models were created with ion-specific coefficients to within standard deviations of 1.7 kJ/mol. These models allow for the characterization of studied ionic liquids as well as prediction of solute-solvent properties of previously unstudied ionic liquids. Predictive models were created for the logarithm of solute's gas-to-fiber sorption and water-to-fiber sorption coefficient for polydimethyl siloxane for wet and dry conditions. These models were created to standard deviations of 0.198 and 0.122 logunits for gas-to-PDMS wet and dry, respectively, as well as 0.164 and 0.134 log units for water-to-PDMS wet and dry, respectively. These models are particularly useful in solid phase microextraction separations. Micelles were studied to create predictive models of the measured micelle-water partition coefficient as well as models of measured MEKC chromatographic retention factors for CTAB and SDS. The resultant predictive models were created with standard deviations of 0.190 log units for the logarithm of the mole fraction concentration of water-to-CTAB, 0.171 log units for the combined logarithms of both the mole fraction concentration of water-to-CTAB and measured MEKC chromatographic retention factors for CTAB, and 0.153 log units for the combined logarithms of both the mole fraction concentration of water-to-SDS and measured MEKC chromatographic retention factors for SDS.

The Abraham Solvation Parameter Model (ASPM) is a linear, free-energy relationship that can be used to predict various solute properties based on solute-solvent interactions. The ASPM has been used to predict log (K or Cs,organic/Cs,gas) values, as well as log (P or Cs,organic/Cs,water) values for solute transfer into the following organic solvents: 2-methoxyethanol, 2-ethoxyethanol, 2-propoxyethanol, 2-isopropoxyethanol and 2-butoxyethanol. The derived log (K or Cs,organic/Cs,gas) correlations describe the experimental data to within 0.14 log units (or less). The derived log (P or Cs,organic/Cs,water) correlations describe the experimental data to within 0.16 log units (or less). The ASPM has also been used to predict the enthalpies of solvation of organic solutes dissolved in the following solvents: acetic acid, dimethyl carbonate, diethyl carbonate, 1-butanol, 1-pentanol, 1-hexanol. The derived enthalpy of solvation correlations, using the L solute descriptor, describe the experimental data to within 2.50 log units (or less). The derived enthalpy of solvation correlations, using the V solute descriptor, describe the experimental data to within 3.10 log units (or less). Validation analyses have been performed on several of the correlations; and, as long as the solute descriptors fall within the given ranges as reported, the original correlations show good predictive ability for determining 1) solute transfer into, and 2) enthalpy of solvation for the aforementioned solvents.

Experimental data were obtained for the computation of mole fraction solubilities of three dichloronitrobenzenes in organic solvents at 25oC, and solubility ratios were obtained from this data. Abraham model equations were developed for solutes in tributyl phosphate that describe experimental values to within 0.15 log units, and correlations were made to describe solute partitioning in systems that contain either "wet" or "dry" tributyl phosphate. Abraham model correlations have also been developed for solute transfer into anhydrous diisopropyl ether, and these correlations fit in well with those for other ethers. Abraham correlations for the solvation of enthalpy have been derived from experimental and literature data for mesitylene, p-xylene, chlorobenzene, and 1,2-dichlorobenzene at 298.15 K. In addition, the enthalpy contribution of hydrogen bonding between these solutes and acidic solvents were predicted by these correlations and were in agreement with an established method. Residual plots corresponding to Abraham models developed in all of these studies were analyzed for trends in error between experimental and calculated values.