The Simultaneous Multicomponent Determination In Complicated System By Chemomtrics Methods

Chemometrics is an interdisciplinary subject. It includes statistics, mathematics, computer science and chemistry. In recent years, with the development of computer science, more and more people attach importance to chemometrics. Multi-components simultaneous determinations in the complicated analytical system, without chemical separation, are the main research objects of analytical chemists, which is also a fast developing research field in chemometrics.Flow injection-chemiluminescence (FI-CL) method is an available trace-analysis technique. However, the fatal disadvantage of the chemiluminescence method is its poor selectivity. The geological sample with many interference elements is the complicated analytical system. The general method is separating them before determination. However, it is not fit for the fast determination of FI-CL. To avoid the troublesome separation process among elements, chemometrics method is introduced to the research, to solve the difficulty of signal separation of similar elements and realized the simultaneous determination them.It has been proved that Near-infrared (NIR) spectroscopy is a powerful analytical tool in the pharmaceutical industries, especially in the quantitative analysis of the pharmaceutical tests during the last two decades. NIR spectrum needs being analyzed by the algorithm of chemometrics, because NIR spectrum can not be distinguished directly. Artificial neural network (ANN) and partial least squares (PLS) methods are more and more widely applied in parsing NIR spectrum.Chinese herbal medicines is the complicated analytical systems. NIR spectrum of Chinese herbal medicine is serious superpose than chemical medicine, which is also a vacant research field of analytical chemistry. Cofrel is an important anti-cough medicine in the world. The contents of benproperine phosphate, which is effective ingredient in Cofrel medicine, is 19.49~20.10mg in one table of 120mg. The prediction results of benproperine phosphate by conventional ANN usually can not satisfy the precision requirement of qualitative control during Cofrel medicine manufacture.In this paper, we attempted to use Kalman Filtering, PLS and a double artificial neural network (DANN) to analyze respectively the commixed CL kinetics of multi-components in the complicated analytical system, and realized the simultaneous determination Nb and Ta Pt and Pd Au and Pt in the geological sample by FI-CL method. The DANN algorithm be used to parse the NIR spectrum of Cofrel medicine and add dummy components to eliminate interferential components, and to use a improved PLS algorithm were used to parse NIR spectrum of Cofrel medicines and Chinese herbal medicine-cortex fraxinis. The results show that there are not significant difference between this method and HPLC. The main research work is described as follows.Simultaneous determination of the trance amounts of Nb and Ta by Kalman Filter analysis with FI-CLA fast and highly sensitive Kalman Filter (KF) analysis with FI-CL method for simultaneously determining the trace amounts of niobium and tantalum without separating them were developed in this chapter. Luminol-K3Fe(CN)6-H2O2 was used as luminescent reagents. After depositing-separation of most interference ions, 1% EDTA as the mask reagent was used to dispel the effect of residual interference elements. The mixed chemiluminescence kinetic curve of Nb and Ta was analyzed by KF in this method to realize the simultaneous determination of Nb and Ta in the national standard samples. The detection limit of the method was 2.1 X 10'9g/g and 4.0x10~9g/g, the relative standard deviation (RSD) was 4.9% and 3.3% for Nb and Ta, respectively.Simultaneous determination of the trance amounts of Pt and Pd by KF-FI-CLIn this chapter, a fast and highly sensitive KF analysis with FI-CL method for simultaneously determining the trace amounts of Pt and Pd without separating them was reported. Luminol-HbOa-OP was used as luminescent reagents. Cationic exchange resin was used to dispel the effect of interference elements. The mixed chemiluminescence kinetic curve of Pt and Pd was analyzed by KF in this method to realize the simultaneous determination of Pt and Pd in the national standard samples. The detection limit of Pt and Pd was 2.20x10"8g/mL and 2.14xl0"8g/mL, the relative standard deviation (RSD) was 0.19% and 3.00% for Pt and Pd, respectively. The relative error (RE%) of detecting the national standard samples was less than 14%.Simultaneous Determination of Au and Pt by PLS-FI-CLIn this chapter, a partial least squares (PLS) analysis combined with on-line and sensitive FI-CL was developed to determine the trace amounts of Au and Pt simultaneously , without the boring process of separating them. Luminol-H202 was used as luminescent reagents. After separation of the most metal ions by cationic exchange resin on-line, EDTA was used as the mask to dispel the effect of residual interference elements. The positions of the catalyzing peak of the luminescence kinetic curve of Au and Pt were different, which was the basis of the simultaneous determination by parsing of PLS. The detection limits of Au and Pt were 1.58xlO"9mol/L and 2.00x1 O^mol/L, respectively, and the relative standard deviation (RSD) 1.25% and 1.59%. The relative error(RE%)of detecting mixed samples was less than 10%.Application of PLS method-NIR transmission spectroscopy in multi-component determination of the cortex fraxinisIn this chapter, PLS method was used to parsed NIR transmission spectrum of aesculin and aesculetin in Chinese herbal medicine-cortex fraxinis, without fussy separating. The effect of the wavelength range, the numbers of principal components and derivative NIR on the results was discussed. The contents of aesculin and aesculetin, which were extracted from 14 cortex fraxinis, have been predicted. Compared the results with those of HPLC, the relative errors (RE%) of aesculin were less than 5.32%, and less than 25% while the contents of aesculetin were more than 0.30%. The results testify glancingly that PLS method-NIR transmission spectroscopy can be used to determine the multi-component for Chinese herbal medicine.Quantitative Analysis of Cofrel Medicines by DANN(PLS)-NIR SpectroscopyIn this chapter, a double artificial neural network (DANN) and improved PLS algorithm were used to parse NIR reflectance spectrum of Cofrel medicines, respectively. The contents of benproperine phosphate, which was the effective ingredient in Cofrel medicines, were accurately nondestructive quantitatively predicted. The effect of the wavelength range, conventional spectra, first-derivative spectra and second-derivative spectra on the results were discussed respectively. Compared the results with those of HPLC, the relative errors (RE%) of benproperine phosphate were less than 0.18% and 0.42%, respectively. The analytical results could be applied to qualitative control of Cofrel medicines.