Calixarenes are supramolecules consisting of phenol units linked via methylene bridges. Covalently bonded to a silica gel, they act as a reversed-phase. Their application as a stationary phase in HPLC has been recently reviewed.
1Additionally to the well known interactions of alkyl-bonded silicas, they can give rise to retention by forming complexes [2][3][4][5][6][7][8][9][10] and by providing p-p-interactions 3,7,11,12 or p-electron-transfer.
7-12For that reason, calixarene-bonded phases exhibit special selectivities. It is obvious that differences of selectivities can be distinct between different kinds of calixarene-bonded phases and especially between calixarene-bonded phases and common C18-phases. In order to select suitable columns for new separation problems, it would be beneficial to numerically describe the retention characteristics and the selectivity of each column. Numerous works have been done in that field and mathematical models relating the retention factor to properties of the column, the solute and the mobile phase were developed. [13][14][15][16][17][18][19][20] The retention has been related to physicochemical parameters or, in more empiric ways, to parameters derived from retention data.The retention and selectivity characteristics of novel calixarene-bonded HPLC-phases are described here. Therefore a multiple term equation has been used. The mechanisms of retention are estimated with its single parameters.
Theory
Mathematical modelsA linear equation describing solute retention was introduced by Snyder:Here the retention of a solute ln k¢ is defined by its retention in pure water, a solute specific value S and the volume fraction of organic modifier f. Yet the equations applicability is limited to a range of 0.2 -0.8f. Equations valid over the whole range of mobile phase composition were described by Schoenmakers et al. 14,15 and by Bosch, Bou and Rosés. 16 Schoenmakers quadratic equation uses factors based on Hildebrands solubility parameter. 21 The factors can be calculated from physical parameters.The model of Rosés et al. is based on the Dimroth Reichardt polarity parameter ET(30). 22 This parameter had been related to log k¢ by Johnson et al. 17 and has been found to give better linear relations to log k¢ than the volume fraction of methanol. Under the assumption that stationary phase polarity is independent of mobile phase composition, Rosés model predicts retention factors by use of two variables. This leads to a simple equation which can be handled easily. Pappa-Louisi et al.23 studied various two-and three-parameter-models concerning their applicability to predict ln k¢. The best predictions could be given by a three-parameter-model, while the two-parametermodels gave comparatively poor results. They pointed out that generally the largest errors occur for predicting retention times of late eluting substances at low modifier concentrations and that predictions deteriorate if the observed modifier range expands.However, although such "compact" systems include only a small number of p...