Spectroscopic analysis and detection of biological systems
Several applications of near- and mid-infrared spectroscopic methods to biological systems have been investigated.^ Near-infrared spectroscopy has been used for determining alcohol solutions. Spectra nonlinearities due to hydrogen bonding and temperature perturbations were observed. Four different calibration algorithms were tried on the Near-infrared alcohol spectra: (1) multilinear regression (MLR), (2) principal component regression (PCR), (3) partial least squares (PLS), and (4) artificial neural networks (ANN). Among the four methods, ANN gave best performance when dealing with nonlinear phenomenon; while PLS and PCR methods could give good approximations to solving nonlinearities by decomposing the raw spectra into orthogonal principal components; MLR showed the worst prediction ability due to the fact that MLR can only pick up wavelengths which showed minimum nonlinear effects from the spectra and thus lack of the ability to compensate both nonlinear effect simultaneously. Both multivariate statistical and ANN methods can be used for quantitative analysis of alcohol solutions in the laboratory or can be readily adapted for continuous monitoring under process control conditions.^ Near-infrared spectroscopic multicomponent analysis employing fiber optics to carry the light to and from the area of measurement have been demonstrated for on-line real time fermentation monitoring. The fiber optics link between the instrument and measurement area is intrinsically safe, free from electromagnetic interference. All of these factors make real-time, in situ monitoring much simpler than it has been in the past.^ By taking the advantage of the adsorption enhancement exhibited when molecules are adsorbed on very thin films of silver or gold particles, a phenomenon known as surface-enhanced infrared absorption (SEIRA). A novel enzyme and/or antibody immobilized biosensor has been constructed. A silicon wafer was used as a substrate, and a thin film of gold was vacuum evaporated onto the silicon surface. The metal films shown in SEIRA, are discontinuous and consist of small ($\ll$ wavelength of light) metal particles. The incident electric field is greatly enhanced around the metal surface through the excitation of localized plasmon resonance of the metal particles, resulting in enhanced absorption. In the preliminary study, we employed Fourier transform-infrared-external reflectance spectroscopy (FTIR-ERS) method. And various types of antibody/antigen systems have been investigated. SEIRA biosensors are easy to handle, fast analyzing time and specific for different antibody/antigen systems.^ A pattern recognition algorithm, Mix-Match search, has been modified for identification of solid and liquid mixtures. It was successfully used to identify Mid-Infrared spectra. This new method is nondestructive, time saving in comparison with traditional analyzing methods for mixture identification. ^
Chemistry, Analytical|Chemistry, Biochemistry|Engineering, Chemical
"Spectroscopic analysis and detection of biological systems"
Dissertations and Master's Theses (Campus Access).