Study On The Continuous Wavelength UV-Spectrophotometry Of Multicomponent Analysis

Chemical information is science on computer, mathematics and statistics analysis technology. As chemical information develops and various software of spectrum chemical information appear, combination of chemical information and spectrum technique makes spectrum analysis come into a new era, and is a highly future method of modern spectrum analysis technique.A new method of continuous wavelength UV-spectrophotometry is brought forward to multicomponent analysis, which is combined by chemical information and UV-spectrophotometry. This method treats the absorptivity and concentration of compound as independent variables, establishes mathematics model, and is solved by generalized reduced gradient. Continuous wavelength UV-spectrophotometry regards the multicomponent analysis question as a constrained optimization. The study' s main results are as follows: 1. establishing mathematics model of multicomponent analysis byUV-spectrum: min f = generalized reduced gradient is selected to optimize object function on the basis of the mathematics model' s characteristic which is few degrees of freedom and many nonlinear constraints. (1)Study is carried through on different independent variables project. The same experiment data is analyzed by this method which treats concentration of compound and absorptivity as independent variables and that method which only treats the concentration of compound as independent variables. The result of this mothed is that the average relative error is 2. 1%, that method is less than 4.9%, which shows this method is better than that one obviously. (2) Research is carried through on algorithms of optimizing mathematics model. The analytical result is unaltered by three different algorithms, which indicates the light demands of algorithms. (3) Initial value of concentration of compound is studied. The analytical result is unaltered when different initial value of concentration of compound, which indicates that the model is self-contained, the algorithm is steady and finds the optimum. (4) Every decision-making factor' s force in target function is researched, which affects the analytical result. Every decision-making factor in target function is endowed with different weight by normalization. The result improves evidently for multicomponent of wide concentration, which shows every decision-making factor' s forcein target function is different. (5) Study is carried through on precision. Multicomponent analysis result of three kinds of compound system is that relative standard deviation is less than 3.362%, which shows that has supernal precision.2. Multicomponent analysis result of three kinds of compound system is that benzoic acid and potassium sorbate' s average relative error is 2. 1%; phenol, quinoline, pyridine' s average relative error is 1.7%; sunset yellow, tartrazine, amaranth, ponceau 4R, allurared' s average relative error is 2. 4%.3. Multicomponent qualitative and quantitative analysis is carried through primarily. Qualitative and quantitative analysis result is correct for analyzing 1-3 components in 10 standard database, which indicates that multicomponent qualitative and quantitative analysis is precise.