Fourier transform infrared spectroscopy

Our estimations of carbohydrate and aromatic contents are based on a newly developed analysis technique for FTIR spectra. FTIR is a common spectroscopic method for analyzing the composition of solid-phase organic matter. When used with attenuated total reflectance (ATR), this method is relatively fast and inexpensive, but is not fully quantitative. FTIR data are typically analyzed either qualitatively by changes in the shape of the spectra, or semi-quantitatively with ratios of peak heights, most commonly humification indices (i.e., ratios of aromatic:carbohydrate or aliphatic:carbohydrate peaks)^{19,21,49,76–79}. A disadvantage of humification indices, as with other peak ratios, is that it is difficult to discern whether humification is driven by changes in carbohydrates vs. aromatics and aliphatics.

Other studies have overcome this problem by correlating FTIR data with wet chemistry-based measurements of carbohydrates, lignin, lipids, proteins, and other compounds, allowing FTIR to be used more quantitatively. These studies include simple calibrations with ratios of peak heights^76,77,79, as well as more complex multivariate techniques such as partial least squares^80–82. However, neither of these techniques provide a basis for estimating relative abundances of individual compounds, apart from ratios, that are not directly calibrated. Some studies have isolated individual peaks and correlated them to wet chemistry⁸³, but this technique is relatively uncommon.

In this study, we introduce a new FTIR data processing method that allows for more thorough quantification of compound classes. First, instead of normalizing peak heights relative to other peaks via humification indices, we better isolate each compound by instead normalizing peaks to the integrated area of the entire spectrum. Next, we use a set of calibration standards to compare these normalized peak heights with wet chemistry analyses, specifically % cellulose + hemicellulose (carbohydrate peak, ~1030 cm⁻¹) and % Klason lignin (aromatic peaks, ~1510 and ~1630 cm⁻¹)²², and show that these measures are linearly correlated. Thus, this study not only provides a method for estimating concentrations of carbohydrates and aromatics, it also suggests that other area-normalized peak heights may be interpretable as relative abundances for cross-sample comparison of individual compound classes.

In preparation for FTIR analysis, peat and plant samples were freeze-dried and then ground to a fine powder for 2 min using a SPEX SamplePrep 5100 Mixer/Mill ball grinder. Calibration standards (for description, see Calibration of FTIR data) were dried at 50 °C, ground in a Wiley mill to pass through a 60-mesh screen, and re-dried to constant weight at 50 °C²². FTIR spectra were collected with a PerkinElmer Spectrum 100 FTIR spectrometer fitted with a CsI beam splitter and a deuterated triglycine sulfate detector. Transmission-like spectra were obtained with a Universal ATR accessory with a single-reflectance system and made from a zinc selenide/diamond composite. Samples were placed directly on the ATR crystal, and force was applied so that the sample came into good contact with the crystal. Spectra were acquired in % transmittance mode between 4000 and 650 cm⁻¹ (wavenumber) at a resolution of 4 cm⁻¹, and four scans were averaged for each spectrum. The spectra were ATR-corrected to account for differences in depth of beam penetration at different wavelengths, and then baseline-corrected, with the instrument software. Spectra were then converted to absorbance mode for subsequent data analysis.