Wide uses of azo dyes produce a great risk of high residuals of carcinogenic aromatic amines, and hence it is important to rapidly analyze these carcinogenic compounds in the textile products to guarantee product safety. In the present work, a surface enhanced Raman spectroscopic (SERS) method was developed for rapid detection of carcinogenic aromatic amines in textiles. In this method, target aromatic amines are extracted from textiles, and then gold nanoparticles are added to the organic extractant, which assemble into closely packed Au array at liquid interface in situ. Finally, fingerprint SERS signals of the target aromatic amines are detected on the generated Au array on the basis of strong chemical interaction between the aromatic amines and the Au surface. The proposed method provided good reproducibility with a relative standard deviation of 3.5% for ten parallel tests of benzidine. It was applied to analyze 70 textile products. To strengthen the spectroscopic data processing, a cluster analysis model was established with 50 samples to automatically identify the spectra based on the good signal reproducibility. The other 20 samples were used as test sets to validate this model. It was found that all the positive samples were successfully identified with false positive rate of 20%. With the addition of the Artificial Intelligence step, the reliability of the discriminant results can be ensured.
Keywords: Aromatic amines; Clustering analysis; Interface self-assembly; Rapid screening; Surface-enhanced Raman scattering; Textiles.
Copyright © 2021 Elsevier B.V. All rights reserved.