Novel on-line methods are presented for the extraction, preconcentration and analysis of zwitterionic biomolecules using sol-gel-coated columns coupled to a conventional UV/visible detector. The presented approaches do not require any additional modification of the commercially available standard CE instrument. Extraction, stacking, and focusing techniques were used in the preconcentration procedures. The positively charged sol-gel coatings were created using N-octadecyldimethyl[3-(trimethoxysilyl) proply]ammonium chloride (C18-TMS) in the coating sol solutions. Due to the presence of a positively charged quaternary ammonium moiety in C18-TMS, the resulting sol-gel coating carried a positive charge. The negatively charged sol-gel coatings were due to the presence of sulfonate groups, which was formed from the oxidation of thiol groups in precursor mercaptopropyltrimethoxysilane (MPTMS) by hydrogen peroxide. Besides MPTMS, tetramethoxysilane (TMOS) and n-octadecyltriethoxysilane (C18-TEOS) were also used to prepare the sol solution for the creation of the negatively charged coatings. For extraction, the pH of the samples was properly adjusted to impart a net charge opposite to the sol-gel coatings. When a long plug of the sample was passed through the sol-gel-coated capillary, extraction was achieved via electrostatic interaction between the charged sol-gel coating and the charged sample molecules. The extracted analytes were then desorbed and focused via local pH change and stacking. The local pH change was accomplished by passing buffer solutions with proper pH values, while a dynamic pH junction between the sample solution and the background electrolyte was utilized to facilitate solute focusing. The developed methods showed excellent extraction and preconcentration effects on both positively and negatively charged sol-gel-coated columns. On-line preconcentration and analysis results obtained on the sol-gel coated columns were compared with those obtained on an uncoated fused silica capillary of identical dimensions using conventional sample injections. The described procedure provided a 150 000-fold enrichment effect for alanine on the positively charged sol-gel-coated column. On the negatively charged sol-gel-coated column, the presented sample preconcentration technique provided a sensitivity enhancement factor (SEF) on the order of 3 x 103 for myoglobin, and 7 x 103 for asparagines. The developed methods provided acceptable repeatability in terms of both peak height and migration time.

Based on the linear relationship between concentration of H2O2 and the decrease of electrochemiluminescence (ECL) intensity in a Ru(bpy)32+/TPA system, procedures for the indirect determination of glucose with a flow injection analysis were developed. By passing solutions of glucose through a FIA system containing a glucose oxidase (GOx) immobilized sol-gel column and an ECL system of Ru(bpy)32+ and TPA, glucose can be determined optimally with a detection limit of 1.0 £gM in a linear dynamic range of 1.0 ¡V 200.0 £gM. A repetitive injection of glucose (100 £gM) and human serum solutions gave satisfactory reproducibility with relative standard deviations of 1.3 (N=31) and 3.9 % (N=42) respectively. Interference due to the presence of ascorbic acid, uric acid or other reducible agents in solution can be corrected by passing sample solutions through another sol-gel column that contained no GOx. From the agreement between the contents of glucose in human serum and soft drink analyzed by the developed method and those obtained by the spectroscopy method based glucose assay kit and satisfactory recovery of glucose from interferent containing solutions, the feasibility of the developed method for real sample analysis was confirmed. One of the major purposes of this study was to develop new immobilization approaches and flow cell designs for the fabrication of regenerable ECL-based sensors with improved sensitivity, convenience and long-term stability. Silica particulates were used as immobilization support in ECL sensors for TPA and NAD(P)H and in biosensors for glucose and glucose-6-phosphate¡]G6P¡^. The first ECL flow cell was fabricated from a glass tube, and a platinum wire was used as working electrode held at +1.3 V. The volume of the flow cell was about 50 £gL. An Ag/AgCl electrode and a piece of Pt wire were used as the reference and counter electrode respectively and placed downstream of the working electrode. Ru(bpy)32+ immobilized silica particulates with 1/3 silica sol content showed the best performance for TPA determination, and the sensitivity of TPA determination was dependent upon the amount of Ru(bpy)32+ immobilized in silica particulates. The lowest level of analyte detected for TPA was 0.02£gM, and linear range was from 0.02£gM to 5£gM. Up to a certain concentration level, it was found that Ru(bpy)32+ was tightly held in silica particulates and did not leach out into aqueous solutions, even with continuous flow for up to ten hours. Ru(bpy)32+ immobilized silica particulates were characterized of well activity and high stability; that stored at 0¢J exhibited its original activity for up to one year. The second ECL flow cell was fabricated from a piece of epoxy block supported Pt electrode (1 ¡Ñ 2 cm) as counter electrode, a piece glass window and a polyethylene spacer with 78 £gL cell volume, two 2.0-cm length of 0.6-mm diameter platinum wires were used as working electrodes held at +1.1 V, and an Ag/AgCl electrode as reference electrode. All three electrodes were incorporated within the main body of the cell. One of the biosensor design packed Ru(bpy)32+ incorporated silica particulates in the ECL flow cell, and a glucose dehydrogenase (GDH) immobilized silica sol-gel column is placed between the sample injection valve and the flow cell. The ECL response to samples containing glucose and cofactor (NADP) results from the Ru(bpy)33+ ECL reaction with NADPH produced by glucose dehydrogenase. This ECL biosensor was shown applicable for both NAD+- and NADP+- dependent enzymes, where NADH detection ranged from 0.50£gM ¡V 5.0 mM NADH and NADPH detection ranged from 1.0£gM - 3.0 mM NADPH. Glucose can be determined in a linear dynamic range of 5.0 - 500 £gM. Another biosensor design immobilized glucose-6-phosphate dehydrogenase¡]G6PDH¡^onto the Ru(bpy)32+ -doped silica particulates through silica chemistry and then packed these particulates into the ECL flow cell. By passing samples containing G6P and cofactor (NAD) through the ECL flow cell, G6P can be determined in a linear dynamic range of 10.0 £gM-1.0 mM. The regenerable ECL biosensor was characterized of good reproducibility and well stability for flow injection analysis. A repetitive injection of NADH (100 £gM) and G6P¡]500£gM¡^gave satisfactory reproducibility with relative standard deviations of 2.8 %¡]N=105¡^and 2.8 % (N=40) respectively.