Identification and quantitative analysis of chemical compounds based on multiscale linear fitting of terahertz spectra

lingbo qiao,yingxin wang,ziran zhao,zhiqiang chen
DOI: https://doi.org/10.1117/1.OE.53.7.074102
IF: 1.3
2014-01-01
Optical Engineering
Abstract:Terahertz (THz) time-domain spectroscopy is considered as an attractive tool for the analysis of chemical composition. The traditional methods for identification and quantitative analysis of chemical compounds by THz spectroscopy are all based on full-spectrum data. However, intrinsic features of the THz spectrum only lie in absorption peaks due to existence of disturbances, such as unexpected components, scattering effects, and barrier materials. We propose a strategy that utilizes Lorentzian parameters of THz absorption peaks, extracted by a multiscale linear fitting method, for both identification of pure chemicals and quantitative analysis of mixtures. The multiscale linear fitting method can automatically remove background content and accurately determine Lorentzian parameters of the absorption peaks. The high recognition rate for 16 pure chemical compounds and the accurate predicted concentrations for theophylline-lactose mixtures demonstrate the practicability of our approach. (c) 2014 Society of Photo-Optical Instrumentation Engineers (SPIE)
What problem does this paper attempt to address?