Extraction of spectral information of additives from activated manganese dioxide products using adaptive kernel independent component analysis

• Published in: Guang pu xue yu guang pu fen xi Vol. 31; no. 5; p. 1340
• Main Authors: Wang, Guo-qing, Peng, Yang, Liu, Shao-wen, Sun, Xiao-li, Zhao, Jian-bo, Sun, Yu-an, Liu, Ying-fan
• Format: Journal Article
• Language: Chinese
• Published: China 01.05.2011
• ISSN: 1000-0593
• DOI: 10.3964/j.issn.1000-0593(2011)05-1340-04

The additives were abstracted from the manganese dioxide products with four kinds of organic solvents, ether, acetone, chloroform and toluene. The extracts were then baked and their attenuated total reflectance (ATR) FTIR spectra were measured using liquid membrane method. The number of chemical components of the additives was determined by median absolute deviation (MAD), and the spectral information of the pure component was extracted by kernel independent component analysis (KICA). The extracted spectral information of the additives is accordant to that of the practically used compounds. An adaptive kernel independent component analysis (AKICA) was proposed for directive extraction of spectral information from chemical mixtures. The results demonstrated that the AKICA method provides an alternative approach to extracting spectral information from the chemical mixtures without previously chemical or physical preseparation for direct extracting spectral information of pure components in the mixed system.