A micromachined absorbance and fluorescence detection cell for application to capillary electrophoresis within planar glass substrates (chips) is described. A microfabricated U-cell for absorbance provides a longitudinal path 120-140 µm long parallel to the flow direction and gives at least a 10-fold increase in absorbance compared to an absorbance path transverse to the flow direction. Absorbance detection limits of 0.003 AU gave ∼6 µM detection limits for hydrolyzed fluorescein isothiocyanate dye. The same device can be used for longitudinal fluorescence excitation with a 20-fold improvement in signal-tobackground levels due to reduced scattering, utilizing a form of sheath flow. Fluorescence detection limits of ∼20 000 molecules and 3 nM were obtained for fluorescein.

Methods to fabricate planar capillary electrophoresis devices integrated with a postcolumn reactor in fused silica (quartz) and Pyrex glass are presented. Quartz is etched at ∼1 μm/min with a 2.1:1 width-to-depth ratio using a Cr/Au/Cr metal mask and concentrated HF/HNO(3). On-chip postcolumn reaction of o-phthaldialdehyde (OPA) and amino acids gave theoretical plate numbers up to 83 000 and ∼90 ms peak widths, corresponding to 14 plates/V and a 0.5 μm theoretical plate height. The reactor geometry caused only a 10% degradation in efficiency.

A glass microchip is described in which reagents and serum samples for competitive immunoassay of serum theophylline can be mixed, reacted, separated, and analyzed. The device functions as an automated microfluidic immunoassay system, creating a lab-on-a-chip. Electroosmotic pumping was used to control first the mixing of 50-fold-diluted serum sample with labeled theophylline tracer in a 1:1 ratio, followed by 1:1 mixing and reaction with anti-theophylline antibody. The 51-nL on-chip mixer gave the same concentration as dilution performed off-chip, within 3%. A 100-pL plug of the reacted solution was then injected into an electrophoresis separation channel integrated within the same chip. Measurements of free and bound tracer by fluorescence detection gave linear calibration curves of signal vs log[theophylline] between 0 and 40 mg/L, with a slope of 0.52 ± 0.03 and an intercept of −0.04 ± 0.04 after a 90-s reaction time. A detection limit of 0.26 mg/L in serum (expressed before the dilution step, actual concentration of 1.3 μg/L at the detector) was obtained. Recovery values were 107% ± 8% for 15 mg/L serum samples.

Clinical interest in the use of capillary electrophoresis (CE) has recently been extended to the microchip environment. Clinical analyses demand careful handling of complex samples that are often limited in quantity and in concentration. The integrated sample handling and analysis capabilities of microchip substrates thus seem ideally suited to clinical applications. This review surveys the development of sample handling (injection, mixing, and reaction) and separation elements on-chip. The integration of these elements to create a variety of clinical analyzers has been demonstrated. The application of microchip CE systems to human serum protein analysis, immunoassay, and DNA studies is reviewed, along with various other clinical applications. In addition, the clinical potential of the lab-on-a-chip concept is discussed.

The affinity constant of a monoclonal antibody to fluorescently labeled bovine serum albumin (BSA) was measured in diluted mouse ascites fluid using a microfluidic chip to perform affinity capillary electrophoresis. Borofloat glass-based devices could be used repeatedly with samples for many months. On-chip separations were performed in less than 60 s, and 30-60 s was required for manual sample exchange. The change in peak height for BSA with increasing BSA/anti-BSA concentration ratio was used to determine concentration changes in bound and free BSA. A Scatchard plot analysis gave an affinity constant (more exactly the intrinsic association constant) of 3.5+/-0.6 x 10(7) M(-1) for a 1:1 stoichiometric ratio. Two affinity complexes were separated. One complex was identified by the Scatchard method as having a 1:1 stoichiometric ratio. The other complex is proposed to have a stoichiometry with an excess of anti-BSA to BSA, most likely (anti-BSA)2-BSA, on the basis of a faster migration time than the 1:1 complex, a decrease in the amount of this complex with increasing [BSA], and predictions of theoretical models for multi-valent antigens. Potential applications of microchip-based devices in affinity measurements are discussed.