| Nitrogen heterocyclic compounds (NHC) have relatively high molecular mass, complex spatial configurations, rich in hydrogen bonds and lone electron pairs of nitrogen. With rich chemical and biological activity, these compounds are often used as a dye, pesticide, lubricating oil additive and anti corrosion inhibitor. Meanwhile, as nitrogen heterocyclic ring containing compounds, they can be important raw materials of synthesizing the fourth generation of high energy density materials (HEDM). Therefore, clarify the reaction mechanism not only has the profound significance to the study of nitrogen heterocyclic compounds, simultaneously also has the very big help in the optimization of the reaction scheme, furthermore improve the yields.Through the analysis of compounds, infrared spectrum can acquire a lot of characteristic peak information; this information can be used to reflect the molecular structures, furthermore, analyzing the molecular structures from the mixture, so in this paper we choose to use infrared spectrum detection method to obtain synthetic mechanisms.The multivariate calibration method and the combination of online infrared technology, firstly to monitor the synthesis of 3-amino-4-amino oximidoflirazan (AAOF) as an example to determine a set of methodology which can successfully deduce the mechanism,, then the study of DAT synthetic mechanism validated the reliability of this set of methodology, specific content is as follows:One:the history and current situation of c he mo metrics and that of blind source separation are discussed; principles and algorithms of chemometric methods which involved in this thesis were discussed;Two:take monitoring the synthesis of AAOF as an example, through a variety of analytical algorithm for analysis obtained infrared spectra matrix, and established a set of methods with NMF as the core which can be used to derive reaction mechanism;(1) Subspace comparison method (SCM method) was proposed to determine the reaction components, and compared with the traditional PCA method; the results indicate that the SCM method can be more clear and intuitive to give component number. Then two MCR-ALS methods, namely the basic MCR-ALS method without the limiting conditions (BMCR-ALS method) and non-negative constraints of concentration added MCR-ALS method (NMCR-ALS method) were employed for the resolution of the mixed infrared spectra matrix. The experimental results show that these two kinds of MCR-ALS are able to achieve the target of unmixing data, but both of them required time-consuming evolving factor analysis (EFA) with over an hour to obtain the initial concentration, so we propose to use FastICA to take calculation. The separation results of each component hold legible infrared spectrum peaks, but the concentration curve is messy, so we need to find other effective signal separation method.(2) KICA and Jade were used on the mixing matrix infrared spectra unmixing. We found that the concentration outcomes of ICA methods have significant turnover, herein we introduce another kind of analytical method called NMF method to solve the spectrum. The results show that NMF method can accurately and quickly analyize the IR spectra matrix, and can greatly reduce the random rotational outputs.Three:the NMF method is successfully applied to the study of DAT synthetic mechanism, and the Gaussian software was employed to simulate the IR spectra of each component in aqueous solution, we also used the ATR element to measure the actual IR spectra of reactants and the product in aqueous solution, then matching NMF outputs with these corresponding spectra.The results showed that each spectra of component which calculated by NMF are very similar with the actual ATR-IR spectra and Gaussian fitting results, the derived synthetic mechanism was also consistent with the literature. Therefore NMF method is prove to be an effective and reliable tools in deduce the reaction mechanism from mixed mid-infrared data matrix. |
PDF Full Text Request