Improving Infrared Spectroscopy Characterization of Soil Organic Matter with Spectral Subtractions

Podsumowanie

SOM underlies many soil functions and processes, but its characterization by FTIR spectroscopy is often challenged by mineral interferences. The described method can increase the utility of SOM analysis by FTIR spectroscopy by subtracting mineral interferences in soil spectra using empirically obtained mineral reference spectra.

Streszczenie

Soil organic matter (SOM) underlies numerous soil processes and functions. Fourier transform infrared (FTIR) spectroscopy detects infrared-active organic bonds that constitute the organic component of soils. However, the relatively low organic matter content of soils (commonly < 5% by mass) and absorbance overlap of mineral and organic functional groups in the mid-infrared (MIR) region (4,000-400 cm^-1) engenders substantial interference by dominant mineral absorbances, challenging or even preventing interpretation of spectra for SOM characterization. Spectral subtractions, a post-hoc mathematical treatment of spectra, can reduce mineral interference and enhance resolution of spectral regions corresponding to organic functional groups by mathematically removing mineral absorbances. This requires a mineral-enriched reference spectrum, which can be empirically obtained for a given soil sample by removing SOM. The mineral-enriched reference spectrum is subtracted from the original (untreated) spectrum of the soil sample to produce a spectrum representing SOM absorbances. Common SOM removal methods include high-temperature combustion ('ashing') and chemical oxidation. Selection of the SOM removal method carries two considerations: (1) the amount of SOM removed, and (2) absorbance artifacts in the mineral reference spectrum and thus the resulting subtraction spectrum. These potential issues can, and should, be identified and quantified in order to avoid fallacious or biased interpretations of spectra for organic functional group composition of SOM. Following SOM removal, the resulting mineral-enriched sample is used to collect a mineral reference spectrum. Several strategies exist to perform subtractions depending on the experimental goals and sample characteristics, most notably the determination of the subtraction factor. The resulting subtraction spectrum requires careful interpretation based on the aforementioned methodology. For many soil and other environmental samples containing substantial mineral components, subtractions have strong potential to improve FTIR spectroscopic characterization of organic matter composition.

Wprowadzenie

Soil organic matter (SOM) is a minor constituent by mass in most soil samples but is implicated in multiple properties and processes underlying soil functions, such as nutrient cycling and carbon sequestration¹. Characterizing the composition of SOM is one of several approaches to link SOM formation and turnover with its role(s) in soil functions^2,3. One method of characterizing SOM composition is Fourier transform infrared (FTIR) spectroscopy, which offers detection of functional groups that constitute organic matter in soils and other environmental samples (e.g., carboxyl C-O, aliphatic C-H)⁴. However, the utility of FTIR spectroscopy for revealing SOM functional group composition is challenged by the dominant mineral component for the majority of soils (typically > 95% mass) due to strong inorganic absorbances that challenge or severely limit detection and interpretation of organic absorbances.

Spectral subtractions offer a way to improve FTIR spectroscopic characterization of organic matter in soil samples. Subtracting mineral absorbances from the soil spectrum can be used to enhance absorbances of organic functional groups of interest in the analysis of SOM composition

(Figure 1).

Advantages of spectral subtractions over standard FTIR spectroscopy (i.e., soil spectra) include:

(i) Improved resolution and interpretation of organic absorbance bands compared to normal soil spectra. Though interpretation of organic bands in soil spectra can be performed by assuming that the relative differences in absorbance are due to differences in organic functional groups, this limits comparisons to samples with the same mineralogy and relatively high SOM content, and may be less sensitive to changes in organic bands, even those considered to be relatively mineral-free (e.g. aliphatic C-H stretch)⁵

(ii) Analysis of soils beyond high SOM samples or organic matter-enriched extracts or fractions

(iii) Highlighting changes induced by experimental treatments from mesocosm to field scales⁶

Additional applications of spectral subtractions in FTIR analysis of SOM include complementing structural and molecular characterizations (e.g., NMR spectroscopy, mass spectrometry)^5,7, identifying the composition of SOM removed by an extraction or destructive fractionation⁸, and fingerprinting SOM composition for forensic purposes⁹. This method is applicable to a wide variety of mineral-organic mixtures beyond soils, including sediment¹⁰, peat¹¹, and coal^12,13.

The potential of spectral subtractions to improve FTIR spectroscopic characterization of SOM is demonstrated using examples of organic matter removal to obtain mineral reference spectra, and then, using these mineral reference spectra, performing and evaluating ideal and non-ideal spectral subtractions. This demonstration focuses on diffuse reflectance infrared Fourier transform (DRIFT) spectra collected in the mid-infrared region (MIR, 4,000 - 400 cm^-1), as this is a widespread approach for the analysis of soil samples⁴.

The two example methods of SOM removal for obtaining a mineral-enriched reference spectrum are (i) high-temperature combustion ('ashing') and (ii) chemical oxidation, using dilute sodium hypochlorite (NaOCl). It should be noted that these are examples of commonly employed SOM removal methods, rather than prescriptive recommendations. Other methods of SOM removal may offer reduced mineral artifacts and/or enhanced removal rates (e.g., low-temperature ashing)¹⁴. High-temperature ashing was one of the first methods used to obtain mineral-enriched reference spectra for performing subtractions, initially for OM-enriched samples derived from soils (e.g., dissolved organic matter, litter)^15,16 followed by its application to bulk soil samples^17,18. The example chemical oxidation used to remove SOM is based on the method of NaOCl oxidation described by Anderson¹⁹. This was originally developed as a pretreatment for removing organic matter in soil samples prior to X-ray diffraction (XRD) analysis, and has been investigated as a potential chemical fractionation sensitive to SOM stabilization^20,21. Both high-temperature removal and chemical oxidation using NaOCl can entail soil-specific artifacts and have limitations on spectral interpretation that should be considered when selecting a method of SOM removal^14,22.

Protokół

1. Prepare Soil for Non-treated DRIFT Spectroscopy and SOM Removal

1. Sieve the soil to < 2 mm using a stainless-steel mesh (the 'fine-earth fraction').
   NOTE: This demonstration employs two soils of similar texture but a nearly 3-fold difference in total SOM content (Table 1).

2. SOM Removal by Chemical Oxidation: Example of NaOCl

1. Adjust the pH of 6% w/v NaOCl to pH 9.5 by adding 1 M HCl dropwise to the solution while mixing and measuring with a pH meter.
   NOTE: Most commercial bleaches (e.g., Clorox) are suitable in quality and concentration (typically 3-7% NaOCl v/v) but will have pH > 12. As NaOCl oxidation of organic matter is pH-dependent, and pH 9.5 is recommended for its use with soil samples^19,23, it is necessary to adjust the pH of most commercially available bleaches.
2. Add 25 mL NaOCl (6% w/v, pH 9.5) to 4 g soil (sieved, air dried) in a 50-mL conical tube and mix by sonication (600 s, output frequency 20 kHz, power 200 W).
3. Incubate the mixture in a hot-water bath (15 min, 80 °C) to increase oxidation rate.
4. Centrifuge to obtain a clear supernatant (e.g., 15 min at 4,000 × g for coarser textured soils; room temperature). Manually decant the supernatant into a waste container.
   NOTE: The concentration of NaOCl in the supernatant (conservatively assuming no oxidation and thus no consumption of NaOCl) is the same as commercially available bleach for household use. Finer textured soils may require longer centrifugation time (e.g., up to an additional 15-30 min) at a given centrifuge speed (e.g., 4,000 × g) to obtain a clear supernatant.
5. Repeat steps 2.3 and 2.4 twice for a total of three oxidation steps.
6. After the last oxidation step, add 20 mL deionized H₂O (dH₂O) to the soil and mix for 5 min using a horizontal shaker (120 rpm). Centrifuge for 15 min at 4,000 × g and room temperature. Repeat for a total of three treatments.
7. Using a spatula and dH₂O from a squirt bottle as needed, extract and wash out the soil pellet from the bottom of the centrifuge tube into a plastic weigh boat (or another container with high surface area). Oven-dry (60 °C maximum, 48 h) to an air-dried state.
8. Once the soil sample is dried, quantify total organic carbon content by combustion-gas chromatography using a C/N analyzer²⁴. Calculate SOM removal as the difference in organic carbon concentration before and after oxidation treatment.
   NOTE: Due to loss of organic matter and soil structure, soil will be prone to crusting, in particular for soils with low sand content. It may be necessary to apply gentle pressure and/or hand grinding to re-homogenize the crusted soil. Soils with inorganic carbon (i.e., carbonates) require additional steps for quantifying organic carbon by combustion-gas chromatography^25,26.

3. SOM Removal by High-Temperature Combustion

1. Measure ~1-2 g of soil (sieved, air dried) into a porcelain crucible using a spatula.
2. Heat at 550 °C for 3 h using a muffle furnace.
   NOTE: This is an example method of SOM removal using combustion at a relatively high temperature. Refer to Discussion on alternative procedures (e.g., temperature).

4. DRIFT Spectroscopy

NOTE: For this example, the FTIR spectrometer software listed in the Table of Materials will be used.

1. Acquire spectra of untreated soil and mineral enriched reference sample (treated to remove SOM).
   1. Prepare the soil samples.
      1. Dilute the samples (optional).
         1. Use analytical grade KBr (or other halide salt) dried at 105 °C and stored in a desiccator to remove residual moisture. For soil samples, effective KBr dilutions can be achieved at a range of 1-33%, in contrast to < 1% for pure compounds.
         2. Mix soil and KBr for a final sample size of 100-400 mg. For example, for a 3% dilution, gently grind 12 mg of dry sample with 60 mg of KBr for 60 s with an agate mortar and pestle. Then, 'fold in' 328 mg of KBr to fully homogenize the sample.
         3. Use serial dilutions with KBr to obtain a high final dilution rate ( < 1%). Perform replicate dilutions to ensure reproducibility, especially since diluted samples use 10¹-10² less soil than neat samples.
      2. Grind untreated and treated soil samples to similar consistency by hand grinding and sieving (e.g., 250 µm using a 60# sieve).
         NOTE: Compared to hand grinding, greater consistency is facilitated by automation, in particular by ball milling. However, the relatively small amount of soil used in SOM removal (e.g., 1-3 g for ashing due to crucible volume) means that hand grinding may be more practical.
   2. Collect the background spectrum.
      1. Load a sample of KBr (ground in the same manner as soil samples (see 4.1.1.2) to mimic soil matrix effects) into a sample cup or plate well.
         NOTE: The "background spectrum" is different from the mineral-enriched reference spectrum (see 4.1.3) used for performing subtractions. The background spectrum will be used by the software to remove atmospheric and other ambient absorbances during collection of spectra on soil samples. All software descriptions are specific to the chosen software and will need to be adapted to other software.
      2. Purge the spectrometer chamber with CO₂- and H₂O-scrubbed air (via a purge gas generator) or with N₂ gas for greater consistency in collection conditions. For example, collection of spectra under ambient atmosphere may entail small fluctuations in humidity and CO₂ that can cause changes in absorbance spectra.
         NOTE: Newer spectrometers may have mirrors (e.g., gold, SiC) that can potentially reduce humidity effects.
      3. Collect a background spectrum using the same detector and acquisition parameter settings, including scan number, wavenumber range, and resolution, that will be used to collect spectra of samples.
         1. Open the drop-down menu for Experiment and select the desired experimental collection method (e.g., acquisition mode).
            NOTE: In this example using chosen spectrometer (see the Table of Materials), the selected method is iS50 Main Compartment.
         2. Click the Experimental Setup icon to select spectral acquisition parameters.
         3. Under the Collect tab, check that the number of scans and resolution are appropriate for experimental objectives; for example, a common setting for DRIFT spectra of neat soils is 128 scans at 4 cm^-1 resolution. Click Ok to save changes.
         4. Click the Collect Background icon to collect a background spectrum. Save the background spectrum for use in the collection of spectra of soils (treated and untreated).
   3. Acquire spectra of soil samples.
      NOTE: Use the same acquisition parameters to collect background and sample (untreated soil, mineral-enriched soil) spectra. Differences among detectors in acquisition time and resolution pose trade-offs that impact collection time and spectral quality. Typical scan numbers for soil spectra range from 128-512 scans. Scan number can be decreased and replicates averaged to obtain a total target scan number. For example, two analytical replicates - the same sample loaded in two separate wells - can be collected using 64 scans each and averaged for a total of 128 scans.
      1. Load the soil sample. To ensure consistent loading and minimize surface roughness, pour samples into the sample cup (or well) to the point of slightly overfilling above the lip or edge of the cup. Then, surface-smooth the soil in the cup using a flat edge (e.g., razor) such that the height of soil sample in the cup is flush with the lip of the cup.
         NOTE: Due to the interaction of infrared light with a matrix such as soil in diffuse reflectance mode, sample loading can influence DRIFT spectra. Samples should not be tamped or subjected to pressure because packing density can affect absorbance. Finer particle size of samples ensures greater ease of surface smoothing (see 4.1.2.1). Depending on the spectrometer model and the sample density, the mass of sample needed to fill a sample cup will range from 300 to 600 mg. In the case of plate wells, this also depends on the well size. Plates with a greater number of wells will have smaller wells and will therefore require less sample. For example, 96-well plates commonly have a well volume of 360 µL whereas 24-well plates have a well volume of 3.4 mL.
      2. Collect spectra of untreated and treated soil samples. First check that the background spectrum collected previously (see 4.1.2.3.4) is used. Click Experimental Setup. Under the Collect tab, select Use specified background file and load the background spectrum file. Click Ok to save changes. To commence spectral collection on the soil, click Collect Sample.
         NOTE: Re-load the same sample in a different well or sample cup to collect the replicate spectra to account for scattering artifacts produced by surface roughness and by variability in matrix density.
2. Perform spectral subtractions.
   NOTE: The subtraction factor (SF) weighs the degree to which absorbances in the mineral reference spectrum are subtracted from absorbances at the corresponding wavenumber in the spectrum of the untreated soil. For subtractions focused on improving resolution of organic absorbances to characterize SOM, it is recommended to utilize the entirety of the MIR afforded by most spectrometers (e.g., 4,000 to 650 or 400 cm^-1, depending on the detector). The next steps describe an empirical method for determining the SF. All software descriptions are specific to the chosen software and will need to be adapted to other software.
   1. Zero out peaks by using the subtraction option of the software program to change the subtraction factor (SF) to minimize or reduce a target mineral peaks and/or mineral peaks, and/or to maximize a linear baseline¹⁴.
   2. Simultaneously select the untreated and treated soil spectra and click the Subtract icon (top center of screen);the first spectrum selected (untreated soil) will be the spectrum from which the second spectrum (treated soil) will be subtracted.
   3. Use the vertical toggle bar or arrows to increase or decrease the SF (left-hand of screen). Observe the changes in the previewed subtraction spectrum.
      1. Use this iterative feature to determine an appropriate SF as described in Representative Results. The numerical SF value appears in the middle of the toggle bar. To adjust the range of SF values, use the Finer and Coarser buttons.
   4. Click Add (upper right-hand of screen) to load the calculated subtraction spectrum into a window.
      NOTE: Since the majority of mineral absorbances are not linear with concentration in most (if not all) soil samples, it is usually not feasible to remove all mineral peaks. It is recommended that mineral peaks considered less prone to inversion (e.g., quartz-like Si-O at 2,100-1,780 cm^-1)¹⁴ be used as the target peak to zero-out by adjusting the SF.
   5. Record and report methodological details on how the subtraction was performed with sufficient detail to allow independent calculation of the same subtraction spectrum from the untreated soil spectrum, including: (1) the wavenumber region used for subtraction, (2) the SF or range of SFs used, and (3) the (mineral) peak or region targeted for zero-ing out.
      NOTE: A good test of the reliability of a subtraction is to have it performed anew by the same user and/or independently by a different user using the reported subtraction parameters.
3. Interpret the spectra.
   1. Perform spectral interpretation using various resources available to analyze and interpret the resulting subtraction spectra, in particular assignments of absorbances to organic functional groups⁴.
      NOTE: Other uses of subtraction spectra include multivariate analysis (e.g., principal component analysis), chemometric prediction of soil analytes²⁷, and even forensic fingerprinting⁹.

Wyniki

The method of SOM removal has practical as well as theoretical implications for the interpretation of subtraction spectra. For example, mineral alterations from high temperature ashing can manifest as losses or appearances of peaks and/or as shifted or broadened peaks in the mineral reference spectrum. These spectral artifacts are prone to occur in regions of overlap with organic bands at 1,600-900 cm^-1,²² compromising interpretation of organic bands. Co...

Dyskusje

The method of removing SOM carries two considerations: 1) the amount of SOM removed, and 2) absorbance artifacts in the resulting mineral reference spectrum. It is fortunately possible— and arguably necessary— to identify and quantity these issues in order to avoid biased interpretations of SOM composition from the resulting subtraction spectrum. Ideally, spectral subtractions would employ a mineral-only reference spectrum to yield a spectrum of 'pure' SOM. In reality, the resulting subtraction spectr...

Materiały

Name	Company	Catalog Number	Comments
Nicolet iS50 spectrometer	Thermo Fisher Scientific	912A0760	infrared spectrometer used to collect spectra
EasiDiff	Pike Technologies	042-1040	high throughput sample holder
OMNIC	Thermo Fisher Scientific	INQSOF018	software used to perform subtractions
6% v/v sodium hypochlorite	Clorox	n/a	generic store-bought bleach for oxidative removal of soil organic matter
Type 47900 Furnace	VWR International	30609-748	muffle furnace for ashing soils to removal soil organic matter
VWR Gooch Crucibles, Porcelain	VWR International	89038-038	crucibles for ashing
VWR Tube 50 mL Sterile CS500	VWR International	89004-364	for sodium hypochlorite
Forced air oven	VWR International	89511-414	for drying soils after oxidation and water washes
VersaStar pH meter	Fisher Scientific	13 645 573	for measuring pH of oxidation solution