Estimating and Correcting Mie Scattering in Synchrotron-Based Microscopic Fourier Transform Infrared Spectra by Extended Multiplicative Signal Correction

Abstract

We present an approach for estimating and correcting Mie scattering occurring in infrared spectra of single cells, at diffraction limited probe size, as in synchrotron based microscopy. The Mie scattering is modeled by extended multiplicative signal correction (EMSC) and subtracted from the vibrational absorption. Because the Mie scattering depends non-linearly on alpha, the product of the radius and the refractive index of the medium/sphere causing it, a new method was developed for estimating the Mie scattering by EMSC for unknown radius and refractive index of the Mie scatterer. The theoretically expected Mie contributions for a range of different alpha values were computed according to the formulae developed by Van de Hulst (1957). The many simulated spectra were then summarized by a six-dimensional subspace model by principal component analysis (PCA). This subspace model was used in EMSC to estimate and correct for Mie scattering, as well as other additive and multiplicative interference effects. The approach was applied to a set of Fourier transform infrared (FT-IR) absorbance spectra measured for individual lung cancer cells in order to remove unwanted interferences and to estimate ranges of important alpha values for each spectrum. The results indicate that several cell components may contribute to the Mie scattering.

Reflection contributions to the dispersion artefact in FTIR spectra of single biological cells

“…The background was recorded from a bare CaF 2 disk and thus the 30 spectrum indicates the changes in cellular reflection (specular and/or diffuse) relative to the bare CaF 2 . For all positions, the spectra are relatively noisy as the reflection from the cell/substrate interface is expected to be weak as the refractive index of the CaF 2 substrate (1.387 at 1696 cm -1 ) [21] is only 35 slightly larger than that of the cellular material normally considered to be about 1.3 [22]. At position (b), the spot is large enough that it samples some of the cytoplasm and the transflection spectrum shows typical cellular features associated with proteins (Amide I ~1650 cm -1 , Amide II 1565 cm -1 ).…”

Spectroscopic and chemometric approaches to radiobiological analyses

“…The resonant behaviour of this phenomenon (where scattering and absorption by structures within the cell) has recently been shown to produce these effects simultaneously [66]. Corrections for this effect may be provided with the extended multiplicative scatter correction (EMSC) [67,68]. Other contaminating effects in FTIRM spectra occur due to aborptions of water vapour and CO 2 , which may also be removed by existing embedded machine-dependent software corrections or machine-dependent in-house correction algorithms [69], while simple filtering and normalization procedures may correct for spectral noise and remove point-to-point variations in sample concentration.…”

Minimising contributions from scattering in infrared spectra by means of an integrating sphere

“…Based on these findings Kohler et al suggested a computational method to correct for the undulating baseline patterns. 21 These models could, however, not explain the origin of band distortions like derivative-like spectral line shapes, band splitting or inverted IR bands sometimes mistakenly called the "dispersion artefact". 22 In 2009, Bassan et al demonstrated then that such band distortions are a result of both, scattering and the fact that the real part n r and the imaginary part of the refractive index are linked via the Kramers-Kronig relationship.…”