Pathlength Selection Method for Quantitative Analysis With Near-Infrared Spectroscopy

摘要

Recently, the near infrared (NIR) spectrum quantitative analysis has been widely used in measuring the concentrations of biological analytes in blood, tissue and other subtrates. In the region of NIR, absorptions are generally broad and therefore strongly overlapped. Accordingly more careful instrument configuration is required to enhance the prediction accuracy. The pathlength is one of the key parameters in the NIR instrument configuration. Optimal pathlength and absorbance for uniwavelength quantitative analysis in the presence of photometric errors has been described in the past. Since the absorptivity of the analytes changes greatly with the wavelength, the optimal pathlength for each waveband differs a lot. Multi-wavelength, instead of a single wavelength, is utilized in the NIR spectrum quantitative analysis of the absorption features. There is no single pathlength that fits for every measurement waveband. Therefore, it is more difficult to select pathlength for multi-wavelength analysis. In this paper, we discuss a new pathlength selection method called Combined Optimal-Pathlengths Method (COP Method), in which several pathlengths are applied in order that the COP spectrum in the quantitative analysis at every wavelength reaches the maximum sensitivity. The PLS (Partial Least Square) analysis results of COP spectra are compared with those using a single pathlength in simulations and experiments. They verify that COP method can enhance prediction accuracy of multi-wavelength quantitative analysis efficiently.