| Analyte quantification is an area of great practical interest for biomedical and commercial applications. Measurement in highly scattering media such as tissue, blood, and other biological fluids is challenging using conventional spectroscopic methods. Likewise, measurement of compounds on-line or in opaque containers can be extremely challenging. This thesis presents general approaches for analyte quantification in liquids by the frequency analysis of ultrasound waves.;Initially, this research focused on the quantification of analytes using dispersive hydrogel sensors. These sub-micron hydrogels vibrate at characteristic resonance frequencies when exposed to ultrasound. The resonance frequencies of the sensors could be modulated by generating molecularly imprinted pockets that would recognize and bind to molecules. Changes in the measured ultrasound frequency spectrum upon binding could then be detected. Using this approach, quantification of theophylline between 10 muM and 6.1 mM was shown.;Improvements in analyte quantification using ultrasound were made by designing hydrogel sensors coupled to antibodies. The high affinity of antibodies for a specific antigen allowed the quantification of acetaminophen between 3.5 nM and 20.8 nM. The highly selective antibodies were also demonstrated to allow acetaminophen determination in a variety of biological media. Though the ultrasound frequency profiles change in the different media, frequencies used for multilinear determination of the analyte were clustered in certain regions of the spectrum. This indicated that the resonance frequencies were minimally affected by the media.;Antibody-linked sensors were also used to determine concentrations of tumor necrosis factor-alpha. The larger mass of this analyte compared to the acetaminophen was accompanied by an increase in sensitivity; concentrations between 115 pM and 592 pM could be determined. Based on this connection between analyte mass and the detection limit, dendrimer-based antibody sensors were then designed to further explore the mass dependence. Although the dendrimer sensor had a different vibration mechanism than the first antibody sensor, it was shown that the mass could be related to shifts in the resonance frequency components.;During propagation, ultrasound undergoes distortion processes that are characteristic of the chemicals present in the solution. This non-linear distortion of the waveform can be measured in the ultrasound frequency profile. In the second portion of this research, it was shown that characteristic changes in the ultrasound frequencies could be measured and correlated with variations in mixture composition. Volume fractions in three-component mixtures of water, methanol, and ethanol could be estimated simultaneously with errors between 2.9% and 3.8% using a hierarchical calibration approach.;The nonlinear distortion was likewise used to determine the composition of commercial beverages. The two major components, ethanol and carbohydrates, were determined simultaneously. Though a wide range of other constituents at various levels are present in beverages, the major components could be estimated with errors of 0.81% for ethanol and 17 mg/mL for carbohydrates.;Overall, the methods presented in this thesis show improvements over conventional approaches for analyte quantification. For point-of-care diagnostics, these methods are attractive due to the short measurement time and the minimal sample treatment required. Likewise, simple methodology and ability to examine optically opaque samples makes this technique highly applicable to on-line measurement in a variety of commercial manufacturing processes. Further applications and possible refinements are discussed in the conclusions chapter. |
