Development of a biofluid chemical measurement system using liquid core optical fiber Raman spectroscopy

Near Infrared (NIR) Raman spectroscopy can provide compositional information about chemicals dissolved in biological fluids. The Raman intensity is proportional to the amount of chemicals. It has been developed for years as a tool to measure biofluid chemical concentrations by illuminating sample and collecting Raman intensity holding the sample in a cuvette geometry. It has been found that the Raman intensity can be enhanced by increasing the excitation and collection sample volume in a liquid core optical fiber (LCOF) geometry. In this thesis, we present a biofluid chemical concentration measurement system using LCOF Raman spectroscopy.; A home-built LCOF Raman spectroscopy system designed for this purpose using 830 nm illumination is described in the thesis. The system is switchable between LCOF and traditional cuvette geometry. The system was characterized using aqueous solutions.; The Raman intensities of aqueous solutions from the two geometries were compared in both theory and experiment. The results agreed well with each other. As high as 15 fold Raman enhancement was observed.; The Raman spectra of biological chemicals in aqueous solution and spiked urine samples were acquired from LCOF and cuvette geometries. The concentrations were predicted using partial least squares (PLS) leave one out cross validation. The results from the two geometries were compared. Concentrations of creatinine were measured in both setups. The LCOF geometry had an advantage at shorter integration times because of Raman enhancement while the cuvette geometry gave better results at longer integration times due to a better system reproducibility.; The LCOF Raman intensity varies from sample to sample with sample absorption coefficient as well as the chemical concentration. This effect can add uncertainty to the concentration measurement. Biofluid samples from multiple patients vary a lot in absorption coefficient, which could cause as much as 20% uncertainty in concentration measurement. To remedy this, the sample absorption coefficient was measured at the same time using the LCOF as a sample holder for a single beam spectrophotometer. Raman spectra were normalized using the absorption coefficient information to remove the enhancement variation. The result of 20% ethanol aqueous solutions with different amounts of India ink showed the process can reduce the Raman peak variance from 60% to less than 1%.; 13 chemical concentrations were measured in blood serum and urine samples from over 70 patients. The concentration measurement was based on integration time of 1--2.5 min. Compared with reference concentration values, 9 of the 13 chemicals were measured to within the reference error level. Shortening the integration time to less than 20 s did not increase the prediction error significantly (less than twice the reference error level at full integration time). LCOF Raman spectroscopy has the potential to become a feasible tool for reagentless, multi-chemical biofluid chemical concentration measurement.