Capillary electrophoresis in pure nonaqueous solvents

Abstract

The use of nonaqueous solvents in capillary electrophoresis (CE) was investigated. Nonaqueous solvents used included methanol, ethanol, acetonitrile, dimethylformamide, and molten salts. Both indirect, and direct UV detection as well as electrochemical detection were examined. The behavior of electroosmotic flow and Joule heating was examined as well as selectivity patterns for a variety of inorganic and organic ionic species. Analytes examined included inorganic anions, alkali and alkaline earth metal ions, alkanesulfonates, alkyl sulfates, linear alkylbenzensulfonates, derivatized and free amino acids, and fatty acids. Selectivity was dependent not only on the nature of the solvent, but also on the nature and concentration of the electrolyte. With many inorganic ions completely reversed separation order relative to aqueous systems was observed. Although the addition of metal ions caused changes in resolution, complete separation of alkanesulfonates (C2-C16) and alkyl sulfates (C8-C18) was possible by a change from protic (methanol) to aprotic conditions (addition of acetonitrile). The partial separation of positional isomers of linear alkyl benzensulfonate homologues (C10-C15) was achieved in methanol/acetonitrile mixtures. The complete separation of alkali and alkaline earth metal ions, which has not been reported to date, and separation of potassium and ammonium were observed in a methanol/imidazole electrolyte. Significantly different migration order was achieved for dansylated amino acids compared to aqueous systems. Changes in selectivity of free amino acids were possible in basic and acidic electrolytes. Complete separation of a wide range of fatty acids (C1-C20) was possible in less than 20 min. The unique selectivity patterns in nonaqueous solvents was related to changes in solvation and ion-interaction effects. The ability to easily adjust selectivity via choice of electrolyte and to inject aqueous samples directly into the nonaqueous electrolytes suggests that such systems may offer significant advantages for many analytical problems. The results illustrate that when selectivity, efficiency, quantitation, and detection limits are taken into consideration, nonaqueous CE approaches should, in many situations, offer a unique alternative to aqueous based CE analyses. Electroosmotic flow was appreciable, and was cathodal in basic solutions. Reversal of electroosmotic flow was observed in the presence of an excess of a strong acid (HCl or HClO4) in the separation electrolyte. The nature of electrolyte anion (chloride, perchlorate, and acetate) and solvent (methanol/acetonitrile mixtures) had significant effects on the electroosmotic flow in acidic conditions. Evidence was observed for ion adsorption (protons and anions) on to silica surfaces and ion-interaction in the electrolyte. Adsorption and ion-interaction could be used to control both the direction and magnitude of electroosmotic flow. Reproducibility of electroosmotic flow was good under appropriate experimental conditions (%RSD = 1.1).