The Research Of The Qualitative Identification Of The Complex Multi-component System

In this thesis, we discuss the two chemometrics methods of band-target entropy minimization (BTEM) and target factor analysis (TFA) on the qualitative analysis of the open path Fourier transform infrared spectroscopy (OP/FT-IR) spectra. In order to reconstruct spectral characteristics of target molecules, we chose the complex atmosphere as major object. Though analyzing the similar degree between the target molecule reconstruction spectra by BTEM and TFA methods and the pure components reference spectra, the presence of the specific target molecules is determined. The primary contents of the thesis are summarized as follows:1. By analyzing the five times consecutive open path Fourier transform spectroscopy around the ranch, a systematic investigation was carried out on the qualitative analysis of the OP/FT-IR spectroscopy gas samples. The spectra were arranged in order of measuring time, which yielded five matrices for data analysis, the mixture spectra in the five data sets were1177,600,811,332,1184, and the spectrum wave-number range was750～1250cm-1. The spectral library data of target molecules, combined with TFA method, was used to qualitatively analyze the gas samples. The results shows that, based on the qualitative analysis of a single reference spectrum in the presence of a minor component of the interference with low signal intensity can overcome the interference, the spectral characteristics of the target molecule is extracted, and get a high degree of similarity between the reconstruction spectrum and the corresponding reference spectrum. The method is practical and facile.2. Using the band-target entropy minimization method for qualitative analysis of the gas samples, the results showed that, compared to the target factor analysis method, the band-target entropy minimization method does not require a reference spectrum of target molecules, the reconstruction result depends on the choice of the objective function, the optimal solution of the objective function corresponding to the spectra is the pure component spectrum. In addition, the method has been optimized, such as the choice of the target band, the penalty function and the Shannon entropy function in the objective function.3. Based on the measurements data obtained by OP/FT-IR, in the range of2850～3200cm-1, the two methods were used to detect the reconstruction effects of a trace of methane in the presence of strong interference NH3and water. The results showed that the reconstruction spectral of methane by the two chemometrics methods and the pure component spectra of methane are highly similar, which could determine the presence of methane component, indicating that the two methods has a perfect effect on a high degree of spectral overlap mixture.4. Compared the results of BTEM with TFA, The results show that although the reconstruction results by target factor analysis method is superior to the band-target entropy minimization, the band-target entropy minimization method can also qualitatively analyze the multi-component complex components in the samples. Meanwhile, the target factor analysis depend on the pure component spectral to reconstruct while the band-target entropy minimization method have not any reference information, So the applications based on the target spectrum entropy minimization method are wider than that of the target factor analysis method, which expand a more broader space for the qualitative analysis of the OP/FT-IR spectrum.