A preliminary study on constructing a high-dimensional asynchronous spectrum to analyze bilinear data

Ran Guo, Xin Zhang, Fei Zhang, Zhuo yong Zhang, Zhen qiang Yu, Yi zhuang Xu, Isao Noda, Yukihiro Ozaki

Research output: Contribution to journalArticlepeer-review

15 Scopus citations

Abstract

A novel approach to constructing high-dimensional asynchronous spectra (nD-Asyn) is proposed. Three theorems relevant to 1D slices of nD-Asyn are revealed. nD-Asyn is used to analyze bilinear data from mixtures containing multiple components obtained via hyphenated techniques. The spectral contribution of different components can be removed in a stepwise manner by increasing the dimensions of asynchronous spectra. As a result, the spectra of different components can be faithfully recovered even if the time-related profiles of different components severely overlap. Moreover, correct results can still be obtained via the nD-Asyn even if a considerable level of noise and baseline drift are present. The nD-Asyn approach is compared with MCR-ALS using different constraints in analyzing the data for a simulated and also for a real system. The nD-Asyn produced correct spectrum of every component. Only when complete constraints obtained from nD-Asyn method is utilized in the MCR-ALS calculation, correct spectra of all the components can be obtained. Thus, nD-Asyn can be used alone or in conjunction with MCR-ALS to analyze bilinear data containing contributions of multiple components.

Original languageEnglish (US)
Pages (from-to)76-84
Number of pages9
JournalSpectrochimica Acta - Part A: Molecular and Biomolecular Spectroscopy
Volume216
DOIs
StatePublished - Jun 5 2019
Externally publishedYes

Keywords

  • Bilinear data
  • SACPs
  • n-Dimensional asynchronous spectrum

ASJC Scopus subject areas

  • Analytical Chemistry
  • Atomic and Molecular Physics, and Optics
  • Instrumentation
  • Spectroscopy

Fingerprint

Dive into the research topics of 'A preliminary study on constructing a high-dimensional asynchronous spectrum to analyze bilinear data'. Together they form a unique fingerprint.

Cite this