Effect of alcohol aggregation on the retention factors of chiral solutes with an amylose-based sorbent: modeling and implications for the adsorption mechanism.

Various displacement models in the literature have been widely used for understanding the adsorption mechanisms of solutes in various chromatography systems. The models were used for describing the often-observed linear plots of the logarithms of the retention factor versus the logarithms of the polar modifier concentration CI(0). The slopes of such a plot was inferred to be equal to the number of the displaced modifier molecules upon adsorption of one solute molecule, and were generally found to be greater than 1. In this study, the retention factors of four structurally related chiral solutes, ethyl lactate (EL), methyl mandelate (MM), benzoin (B), and pantolactone (PL), were measured for the amylose tris[(S)-α-methylbenzylcarbamate] sorbent, or AS, as a function of the concentration of isopropanol (IPA) in n-hexane. With increasing IPA concentration CI(0), the slopes increase from less than 1, at a concentration range from 0.13 to 1.3M, to slightly more than 1 at higher concentrations. Such slopes cannot be explained by the conventional retention models. It was found previously for monovalent solutes that such slopes can only be explained when the aggregation of the mobile phase modifier, isopropyl alcohol, was accounted for. A new retention model is presented here, accounting for alcohol aggregation, multivalent solute adsorption, multivalent solute-alcohol complexation, alcohol adsorption, and solute intra hydrogen-bonding, which occur in these four solutes. The slope is found to be controlled by three key dimensionless groups, the fraction of the sorbent binding sites covered by IPA, the fraction of the solute molecules in complex form, and the fraction of the IPA molecules in aggregate form. The limiting slope at a very high IPA concentration is equal to the value of (x+y)/n, where x is the number of the solute-sorbent binding sites and y is the number of the alcohol molecules in the solute-alcohol complex, and n is the alcohol aggregation number. The model was tested with the HPLC data of two sets of chiral solutes, one set of new data presented here and of one set of literature data by Gyimesi-Forrás et al. (2009), for which there is no known intramolecular H-bonding. For the first set of solutes, the values of the equilibrium constants for intramolecular hydrogen bonding were calculated from our previous IR data. The value of the parameter y was fixed on the basis of the number of the solute functional groups, IR data, and the results of DFT and MD simulations. The retention factors in pure hexane (k0) were found experimentally for EL, MM, and B; for PL they were estimated from the data. Then the values of x and the complexation equilibrium constants were estimated. The model fits fairly well our new data, and less well the more-limited literature data, for which the k0 values were unavailable, and the retention factors were obtained over a narrow range of IPA concentrations. For EL and PL, results of infrared spectroscopy, density functional theory, and molecular dynamics simulations indicated strong solute-IPA complexation, and multiple solute-sorbent binding sites, which are consistent with the fitting results. Hence, the new model has been shown to be more reliable than the previous models for estimating the numbers of the potential binding sites of multivalent solutes.