基于紫外光谱法和有监督模式识别技术对氯化钠、氢氧化钠和β-苯乙胺种类的鉴别

doi:10.3964/j.issn.1000-0593(2021)02-0448-06

Abstract
References
Related Citation (15)

Download: PDF (2389 KB)
Export: BibTeX | EndNote (RIS)

Abstract β-phenylethylamine(PEA) is a very important chemical that intermediates raw materials. During the synthetic process of PEA, the final product usually contains NaCl, NaOH and PEA. Therefore, it is of great significance for the synthesis and qualitative measurement of PEA to identify the species of NaCl, NaOH, PEA and the mixture. A rapid method for the identification of NaCl, NaOH, PEA and mixtures was established by UV spectroscopy. Therefore, the absorption spectra of NaCl, NaOH, PEA and the mixture were measured by UV spectroscopy in the interval 190～400 nm. Firstly, the PLS method was used to extract the principal components of the UV spectrum, and a few principal components were used to replace the original variables to reduce the complexity of the model. PLS was used to extract the distribution of score vector values of NaCl, NaOH and PEA in the space of the first three principal components. The accumulative contribution rates of the first three principal components of NaCl, NaOH and PEA were 96.64%, 99.44% and 99.95%, respectively. Therefore, the first three principal components of NaCl, NaOH and PEA contain most of the spectral information. Secondly, three principal components were used as input variables to identify the species of NaCl, NaOH and PEA by using LDA, Sigmoid SVM, RBF-SVM, RBF-ANN, BP-ANN and Artificial Bee Colony(ABC) combined with BP-ANN(ABC-BP-ANN), and the overall sensitivity were 95.6%, 95.6%, 95.9%, 95.8%, 96.9% and 99.6%, respectively. Since the characteristic absorption peaks of NaCl and NaOH are very similar, the score vectors of the principal component would overlap each other, which led to misjudgment for the species identification of NaCl and NaOH. By comparing the results of six classification methods, it is known that ABC-BP-ANN is the best, BP-ANN is the second, RBF-SVM and RBF-ANN are similar, but slightly worse than BP-ANN, LDA and Sigmoid-SVM are the worst. Finally, the mixtures of seven different mole fractions were prepared which ranged from 0% m·m^-1 to 60% m·m^-1(The molar fraction of the mixture is the percentage of PEA in the mixture), then RBF-SVM, BP-ANN and ABC-BP-ANN are used to identify the species of the mixture. From the results of sensitivity and specificity, it can be concluded that the classification result of ABC-BP-ANN is the best, BP -ANN is the second, RBF-SVM is the worst, and the results of the mixtures are consistent with those of the single component. The results indicated that UV spectroscopy combined with ABC-BP-ANN pattern recognition technology could successfully identify the species of NaCl, NaOH, PEA and the mixture. This method can be used as a simple, rapid and reliable method for the species identification of NaCl, NaOH, PEA and the mixture, and it can also provide theoretical basis and technical support for the synthesis and quality control of PEA.

Key words：β-phenylethylamine; UV spectroscopy; Artificial Bee Colony combined with BP-ANN; Sensitivity; Species identification of the mixture

Received: 2019-12-09 Accepted: 2020-04-04

ZTFLH:

TS227

Corresponding Authors: TANG Xiao-jun E-mail: xiaojun_tang@xjtu.edu.cn

Cite this article:

URL:

https://www.gpxygpfx.com/EN/10.3964/j.issn.1000-0593(2021)02-0448-06 OR https://www.gpxygpfx.com/EN/Y2021/V41/I02/448