Analysis of the solvent strength parameter (linear solvent strength model) for isocratic separations in...

The solvent strength parameter (slope) and intercept (log k_w) of the linear solvent strength model are determined for methanol-water (17 columns), acetonitrile-water (15 columns), acetone-water (7 columns), tetrahydrofuran-water (6 columns), and 2-propanol-water (4 columns) for varied compounds. It is shown that the linear region of the plots of the retention factor (log k) against the volume fraction of organic solvent (ϕ) is largely a system property independent of solute type for compounds with accessible retention factors. Solvent strength parameters for methanol (3.12 ± 0.12), acetonitrile (2.78 ± 0.18), acetone (2.71 ± 0.11), and tetrahydrofuran (2.95 ± 0.24) are assigned for siloxane-bonded silica and porous graphitic carbon columns as an average of the compound-specific experimental values. Statistical analysis indicates that the compound-specific solvent strength parameters are column dependent and limited as a general solvent property. Evaluation of the column-specific solvent strength parameters provides further evidence that the average solvent strength parameter is column dependent. The column-specific solvent strength values are compared with the hypothetical water-organic solvent distribution systems using the system constants of the solvation parameter models as variables confirming the column dependence of the solvent strength parameter. The column-specific solvent strength parameters can be predicted by the solvation parameter model with a typical deviation of about 0.12 over a range of 1.69 to 6.33 for the experimentally determined values. The intercept of the linear solvent strength model is shown to be both column and organic solvent dependent. Compound column-specific values are adequately described by the solvation parameter model. Log k_w cannot be recommended as a descriptor of solute properties since it has no clear connection to a real distribution system.