Logo PTB

Compressed FTIR spectroscopy using low-rank matrix reconstruction

05.01.2021

Sample interferogram of the L. tarentolae film is shown for pixel (50, 50) in a) and the corresponding spatial image of interferometer position t / Δt = 1715 in b). The interferometer domain around the center burst is magnified in c) and the pixel domain of interest that is subsequently analyzed in the bioanalytical fingerprinting application is shown as a red rectangle in the pixel domain reconstruction in d).

Fourier transform infrared (FTIR) spectroscopy is a powerful technique in analytical chemistry. Typically, spatially distributed spectra of the substance of interest are registered simultaneously using FTIR spectrometers equipped with array detectors. Scanning methods such as near-field FTIR spectroscopy are a promising alternative providing higher spatial resolution. However, serial recording severely limits their application due to the long acquisition times involved and the resulting stability issues.

In order to speed up such measurements a novel mathematical technique of low-rank matrix reconstruction of a compressed FTIR measurement has been developed. With this method the measurement time of the scanning methods is significantly reduced. To demonstrate the method, data from a pilot study of biological samples (Leishmania strain protozoes) have been analyzed by randomly selecting only 5% of the interferometry data. The results obtained for bioanalytical fingerprinting using the proposed approach have been shown to be essentially the same as those obtained from the full set of data. This finding can significantly foster the practical applicability of high-resolution serial scanning techniques in analytical chemistry, and it is also expected to improve other applications of FTIR spectroscopy and spectromicroscopy. An open source Python library enables the method to be used easily.

The research was performed by the working groups Data Analysis and Measurement Uncertainty and IR Spectrometry, in collaboration with the Freie Universität Berlin (AG Prof. Rühl) within the DFG Project „Bayesian Compressed Sensing“ (EL 492/1-1, RU 420/13-1).

Publication:

M. Marschall, A. Hornemann, G. Wübbeler, A. Hoehl, E. Rühl, B. Kästner, and C. Elster, "Compressed FTIR spectroscopy using low-rank matrix reconstruction," Opt. Express 28, 38762 (2020).

Contact:

B. Kästner, 7.11, E-Mail: Opens local program for sending emailBernd.Kaestner(at)ptb.de
M. Marschall, 8.42, E-Mail: Opens local program for sending emailManuel.Marschall(at)ptb.de