The Research Of Improving The Analysis Accuracy Of Complex Solution Components By Near-infrared Absorption Spectroscopy

To estimate and improve the accuracy of the quantitative analysis of thecomponents concentration in the complex solution by near infrared absorptionspectroscopy, the analytic model was built to systematically analyze the influence ofvarious error sources, including the inherent characteristics of analyte and the factorsintroduced from the measuring and analysis process. The methods of improving theaccuracy of quantitative analysis were proposed. The effectiveness and applicablescope of these methods were verified by quantitative analyzing the total protein inserum using near infrared absorption spectroscopy.The main research work involves:1. From the standpoint of the chemical content and optical properties of theanalyte, the concentration, the shape and overlap degree of extinction coefficientcurve, as well as the range of the absorption coefficient and scattering coefficientwere considered as influencing factors of quantitative analysis accuracy by NIRS. Theoverlapping coefficient was presented to describe the degree of the overlap betweenthe extinction coefficient curves of the analyzing component and the backgroundcomponents. The certain relationship between the accuracy of the concentrationanalysis and the overlapping coefficient was verified by the quantitative analysis ofaqueous alcohol solution.2. From the standpoint of the measurement conditions, the noise of thespectrometer device and the number of wavelengths for establishing the model areinvolved in the accuracy analysis. The concept of equivalent signal-to-noise ratio(ESNR) was presented. The influence of ESNR and the number of wavelength forestablishing model on the accuracy of analysis was discussed3. The concentration distribution of analyzing component and backgroundcomponent was considered as an effect factor of modeling and prediction accuracy.The influence under the conditions of absorption analyzing component, absorptionbackground component, scattering analyzing component, and scattering backgroundcomponent were discussed, respectively. The influence of the samples number andsample cell thickness on the accuracy of analysis was analyzed. 4. From the perspective of obtaining the scattering information and taking theadvantage of the scattering properties of analyte, three approaches for improving theaccuracy of concentration analysis were proposed: the squared term or higher orderterms spectrum, multi-optical path spectrum and multi-point distributed spectrummethod. This research was analyzed the influence of these three approaches on theaccuracy of concentration analysis with different absorption coefficients andscattering coefficients. The effectiveness of the high order terms spectrum and themulti-optical path spectrum methods were verified by the quantitative analyzing oftotal protein in serum.