Data reduction and calibration methods for multicomponent analysis in spectrophotometry; Simultaneous determination of aluminum, iron, and titanium with 8-quinolinol

Abstract

Data reduction and calibration methods in an overdetermined multicomponent analysis are discussed on the accuracy and precision of determination. In the data analysis, the matrix representation of the Lambert-Beer�s law (AKC method) and their reversed form, expressing the concentration as the function of absorbance (CPA method) are compared. The CPA method is superior to the AKC method in the accuracy and precision when the number of the analytical wavelengths is equal to, or a few more than the number of component. But there is a problem in the calculation of the inverse matrix in the highly over-determined case with regard to the analytical wavelengths. On the. other hand, the AKG method has several advantages in the course of date analysis. In the AKG method, the data collected from many wavelengths can be dealt with; therefore, we can observe not only the individual spectrum of each component by calculating the absorptivity matrix after the calibration, but also the difference spectrum by subtracting the raw data spectrum from the calculated one after the analysis. From these spectra, validity of the Lambert-Beer�s law, or the additivity of the spectrum of each component can be ascertained. In both methods, the accuracy and precision of determination are improved by adding the nonzero intercept term to the calibration model, and by the calibration with standard mixtures instead of a standard solution of single component. The present methods could be successfully applied to the simultaneous determination of aluminium, iron, and titanium in the standard reference materials: NBS 81a (glass sand) and BGS 314 (silica brick). © 1983, The Japan Society for Analytical Chemistry. All rights reserved.