A multi-analytical approach to studying the chemical composition of typical carbon sink samples

Peatlands in southern South America (Tierra del Fuego region, TdF) play a key role in the ecological dynamics of Patagonia. It is, therefore, necessary to increase our knowledge and awareness of their scientific and ecological value to ensure their conservation. This study aimed to assess the differences in the distribution and accumulation of elements in peat deposits and Sphagnum moss from the TdF. Chemical and morphological characterization of the samples was carried out using various analytical techniques, and total levels of 53 elements were determined. Furthermore, a chemometric differentiation based on the elemental content of peat and moss samples was performed. Some elements (Cs, Hf, K, Li, Mn, Na, Pb, Rb, Si, Sn, Ti and Zn) showed significantly higher contents in moss samples than in peat samples. In contrast, only Mo, S and Zr were significantly higher in peat samples than in moss samples. The results obtained highlight the ability of moss to accumulate elements and to act as a means to facilitate the entry of elements into peat samples. The valuable data obtained in this multi-methodological baseline survey can be used for more effective conservation of biodiversity and preservation of the ecosystem services of the TdF.

Sampling. Eight ombrotrophic and mostly undisturbed peat bogs and Sphagnum moss samples were collected in 2018. Peatland patches were sampled in duplicate and selected according to the peatland map of Roig et al. 29 . A map of the location of eight sampling sites (A, B, C, D, E, F, G, and H) and the geographical coordinates of the selected points are shown in Fig. 1. Except for site B, which is located in Ushuaia, the most populated city of the archipelago, and near the International Airport, the other sites are in low anthropogenic impacted areas. Peat samples were collected using a field drill, and cores were extracted to a depth of 20 cm, divided into 10 cm segments, and filled in plastic bags. The peat and Sphagnum moss samples were dried at 40 °C for two days, ground (using a ceramic mortar and pestle), and homogenized in the laboratory. All samples were stored at 4 °C until analysis.
Analysis of powder samples. Carbon 30 . The analytical procedure consists of the combustion of the sample (1000 °C) followed by catalytic oxidation and reduction processes, separation of the gases produced in the gas-chromatographic column, and analysis of the same with a thermal conductivity detector (set at 290 °C); signal processing and determination of the percentage of elements present in the sample with the EAGER 200 program.
High resolution morphology study and semi-quantitative analysis. Scanning electron microscope (SEM) analysis was performed with a high-resolution field emission SEM (HR-FESEM; model Auriga 405; Carl Zeiss Microscopy GmbH, Jena, Germany) equipped with an energy dispersive spectrometer for X-ray microanalysis (XEDS, model Quantax; Bruker, Berlin, Germany). Before analysis, a small portion of the powder sample was fixed to specific support and coated with a thin layer of C by a sputtering machine (Q150T Turbo-Pumped Sputter Coater/Carbon Coater; Quorum Technologies Ltd., East Sussex, United Kingdom) to make its surface conductive. HR-FESEM XEDS acquisitions were performed under a high vacuum (10 −6 hPa) at 20 keV accelerating voltage. Micrographs were acquired by secondary electron detector (SED) at a working distance (WD), magnification, spot size, and tilt angle conditions properly adjusted to optimize image resolution. The microanalysis was performed at WD and magnification ranging from 3.0 to 7.8 mm and 100× to 45,000×, respectively.
Mercury analysis. Mercury was determined using an advanced mercury analyzer (AMA-254, Altec Ltd., Prague, Czech Republic) according to EPA method 7473 31 . Each sample (~ 20 mg) was analyzed in duplicate, and a certified reference material (BCR 482) was used for the analytical procedure's quality assurance and control. Trueness bias percentage and precision were 2.3 and 3.2%, respectively. The limit of determination and quantification (LOD and LOQ, respectively), defined as the Hg concentration corresponding to three and ten times the standard deviation of the blanks (n = 10), was 0.0001 and 0.004 mg kg −1 , respectively. A Hg reference standard solution (1,002 ± 7 mg L −1 in 10% HNO 3 ; SCP Science, Baie D'Urfé, Quebec, Canada) in 1% v/v HNO 3 (67% suprapure, Carlo Erba Reagents, Milan, Italy) was used to prepare calibration solutions (in the range of 0-50 μg L −1 ). All aqueous solutions were prepared using deionized water with a resistivity of 18.  25 and 300 mg, respectively, were irradiated at the RA-3 reactor (thermal flux 3.1013 cm −2 s −1 , 8 Mw) for 4.5 h. Two measurements with 7 and 30-day-decay counted from the end of irradiation, were done using GeHP detectors (30% efficiency, 1.8 keV resolution for the 1332.5 keV 60Co peak). The Gamma Vision software was employed to acquire the gamma spectra, and elemental mass fractions were calculated using software developed at the NAA laboratory. For quality control purposes, Andesite ACH-1 and Grass (Poeaceae) from WEPAL 2011-4 were used as control samples for peat and Sphagnum moss analysis, respectively 33 (Table S1). The INAA LODs are shown in Tables 1 and S2. Table 1. Results of the total elemental content (mg/kg d.w.) in peat and Sphagnum moss by AMA, ICP-MS, ICP-OES, and INAA (n = 16 for each matrix). a Wavelength selected for the elemental analysis by ICP-OES. b Isotope selected for the elemental analysis by ICP-MS. c Non-parametric Mann Whitney test was applied: "-" = not determined; "ns" = not significant at p > 0.05; "*" = p < 0.05; "**" = p < 0.01; "***" = p < 0.001. Numbers in bold in the same row indicate significant differences. Total digestion of the powder samples was performed in a close microwave oven system (Ethos1 Touch Control; Milestone, Sorisole, Bergamo, Italy) using an acid mixture of HCl-HF-HNO 3 , according to Bettinelli et al. 35 , Astolfi et al. 36 , and Gaeta et al. 37 . Briefly, weighed amounts (~ 100 mg) of the samples were transferred into polytetrafluoroethylene vessels; to these, 1 mL H 2 O 2 (30% superpure, Merck, Darmstadt, Germany), 1 mL HCl (30% suprapure, Carlo Erba Reagents, Milan, Italy) and 3 mL HNO 3 (method M1) or 1 mL HF (40% superpure, Sigma-Aldrich Chemie GmbH, Steinheim, Germany), 1 mL HCl and 3 mL HNO 3 (method M2) were added. The solutions obtained were heated with microwave energy for 40 min using a program with temperature ramps up to 180 °C. The obtained digests were diluted to 20 mL with deionized water and filtered (0.45 μm cellulose nitrate membrane; GVS Filter Technology, Indianapolis, IN, USA). The samples by method M1 were further diluted 1:10 with deionized water. After these procedures, the digests obtained with M1 or M2 method were analyzed by ICP-MS or ICP-OES, respectively. Method blanks were periodically analyzed alongside the samples to check for any losses or cross-contamination. The LODs and LOQs are shown in Table 1. A standard reference material (NIST 1515, Gaithersburg, MD, USA) was analyzed for quality control purposes. For all the certified elements (Al, B, Ba, Ca, Cd, Ce, Co, Cr, Cu, Fe, K, La, Mg, Mn, Mo, Na, Ni, P, Pb. Rb, S, Sb, Sr, U, V, W, and Zn), the trueness bias percentage was between -9.7 (La) and 17% (Ca) of the expected value, and the precision as repeatability was between 0.5 (K)-25% (Sb) 36 . For the other elements (As, Be, Bi, Cs, Ga, Li, Nb, Se, Si, Sn, Te, Ti, Tl, and Zr) there are indicative levels in the peat and Sphagnum moss samples. For the elements in common between the different analytical techniques (As, Ce, Co, Cr, Cs, Fe, La, Rb, Sb, and Zn), a quality control of the data was carried out through intertechnical comparisons.
Statistical analysis. Statistical analyses were conducted using IBM® SPSS® Statistics 27 software (IBM Corp., Armonk, NY, USA). For each element, values below the LOD were replaced with a value equal to half the LOD 38,39 . When the percentage of values < LOD exceeded 30%, the element was excluded from the statistical dataset.
The differences in the sample concentration were tested by Kruskal-Wallis and pairwise post-hoc tests and Mann-Whitney test. Probability values from multiple pairwise comparisons were adjusted using Bonferroni corrections 40 .
Principal component analysis and variable selection by stepwise approach were performed with JMP 16 Pro (SAS Institute) to highlight possible sample grouping. Autoscaling pretreatment was applied on data matrices before the chemometric analyses.  2 and S1-S4, and Tables S3 and S4). In all the samples the following elements Al, Ca, Cl, Mg, Na, O, P, S, and Si are present in variable percentages, while Cr and Cu are found only in peat and Fe, K, and N in Sphagnum moss (Tables S3 and S4). Oxigen (from 25.8 to 45.6% in peat and from 29.7 to 44.5% in Sphagnum moss) is the most abundant element. The surface of the peat samples appears layered and porous, while the Sphagnum moss samples show globular and fibrous formations. Backscattered electron images (Fig. 2) provide a differentiation between the organic matter (weak brightness) and the inorganic fraction (strong brightness) in the samples. In both matrices, there are metallic spherical particles (Fig. 2). The inorganic spheroidal particles present in the upper layers of peat are of predominantly anthropogenic origin such as industrial activities and coal combustion, since cosmic spherules (generally produced by combustion processes) are extremely rare in peat 41 Table 2). Peat is made up of organic material that contains about 50% C (Table 2), so C concentrations reflect changes in peat humification and related mass losses in peatlands 16 . The decomposition of organic material in peat leads to the formation of phenolic structures derived from lignin 43 . These structures are more difficult to degrade than proteins and sugars 43,44 . The O/C and H/C ratios can give information on the carbohydrate content and the percentage saturation of C within the organic molecule. Therefore, a higher aromaticity in the samples will be given by lower H/C ratios 43 ; while a lower O/C ratio indicates a lower carbohydrate level and/or higher organic content in the peat sample 45 . The C/N ratios are also related to the decomposition processes of organic matter since the microbial consumption of organic substances determines a decrease in the abundance of C compared to N 16,46 . However, decomposition by microorganisms is limited due to the lack of nutrients in the peat and the low pH 16 . Therefore, the C/N ratio is used as an indicator of the degree of decomposition of peat and the mass loss of peat 16,46 . The C/N ratios in peat are an indicator of the degree of humification, where low C/N ratios indicate high humification of the peat 46 . Previously reported ratios of C/N values for peat and bog vegetation mainly ranged from 40 to 100 16,46 . Overall, the results obtained in this study (Table 2) are in agreement with those of the literature 16,43,46 . The samples with the highest C/N ratio are those in site H while those with the highest degree of humification are present in sites E for peat and G for Sphagnum moss. Generally, changes in the humification index are related to changes in environmental conditions, i.e. different climates and resulting differences in peat accumulation and decomposition 46 . The high C/N ratio, between 60 and 153 (sites A, D, G, and H for peat and A, B, C, E, F, and H for Sphagnum moss) suggests rapid peat growth under conditions of humidity while lower values may indicate drier marshes and, therefore, exposure of peat to aerobic decay for more extended periods 46,47 . Differences in vegetation can be explained by lower rainfall and higher nutrient input through marine sprays 48,49 . www.nature.com/scientificreports/ The S content of peat consists of organic S, sulphate and sulphides 50 . Not much information is found in the literature on S in peat and Sphagnum moss. Organic S in Scottish peat is predominant with about 64% of total S being C-bonded sulphur and 27% ester sulphate S 51 . Some authors did not find a clear correlation between S intake and its concentration in peat 52 . The S levels measured in the Sphagnum mosses sampled in the Tierra del Fuego region range between 0.01 and 0.26%, which is similar to those reported for non-polluted sites 53 . Figures S5 and S6 show the infrared spectra of the different samples, while Table S5 presents a summary of the main bands observed in peat and the assignments of the related functional groups. The spectra present wide bands typical of natural organic matter and are due to the superposition of single absorption bands 43 . The main features of the spectra are: a broad band around 3400 cm −1 due to O-H stretching of various groups like alcohol and phenols; two peaks at 2920 and 2850 cm -1 due to C-H stretching and characteristics of aliphatics (fats, wax, lipids); a region between 1720 and 1420 cm −1 assigned to C=O stretch of carbonyl and carboxyl groups (carboxylic acids and aromatic esters), aromatic C=C and asymmetric COO− group vibrations (lignin and other aromatics and aromatic or aliphatic carboxylates), and OH deformations and C=O stretch of phenols or C-H deformation (phenolic and aliphatic structures); and, finally, absorption bands in the 1100-1000 cm −1 region were allocated to combination of C-O stretching and O-H deformation of polysaccharides 46,54-56 .

Chemical composition of peat and Sphagnum moss samples. For all elements analyzed in peat and
Sphagnum moss samples, higher total levels of each element were found in Sphagnum moss than in peat samples except Mo, S and Zr ( Table 1). The most abundant elements in both matrices were Al, Ca, Fe, S, Si, and Zr, ranging from 440 (Zr) to 12,000 (Si) mg kg -1 . Cesium, Hf, K, Li, Mn, Na, Pb, Rb, Si, Sn, Ti, and Zn levels in Sphagnum moss were about 2 (Na) to 18 (Zn) times significantly higher than that in peat. In contrast, significantly higher levels of elements were found in the peat samples than in the Sphagnum moss samples for Mo, S, and Zr. Considering the element levels for each matrix and site (Tables S6 and S7), no significant differences were highlighted, showing the uniform distribution of elements in each matrix in the area under study.
Some elements, such as Pb and Hg, tend to bind strongly to organic material 50 . Lead is considered immobile and well-preserved in peat profiles 57,58 . Even most of the elements that represent the deposition of dust, such as Si and Zr, or others, such as Al, Ti, Sc and rare earth elements (REE), which are mainly identified as lithogenic tracers, should be immobile and stable 50 . However, some elements, such as Pb, tend to form organic complexes at low pH values and for this reason can move along peat profiles 50 .
Some studies showed that the increase in Hg concentration was linked to a greater decomposition of the peat (low C/N ratio) and not necessarily to a greater contribution from atmospheric depositions 59 . Our data, in agreement with other studies 60,61 , showed no correlation between Hg and peat decomposition (p < 0.05). This highlights a homogeneity in the composition of the sampled peat.
Our results match those of Wang et al. 25 , which show that moss accumulates most elements by deposition, including toxic elements, and can be a means of facilitating the accumulation of metal(loid)s in the soil. Furthermore, wet or dry atmospheric deposition of particles on the moss surface can subsequently be solubilized or washed away with precipitation [62][63][64] . Thus, peat can be enriched with metals and significantly contribute to elemental levels in moss 25 .
Each country has defined risk levels associated with different metal(loid) concentrations [65][66][67] . For example, Table S8 shows the threshold levels established by Legislative Decree 152/2006 and subsequent amendments and additions 68 and the standards established in the Finnish legislation for contaminated soil 69 . Concerning Finnish legislation, the highest concentration levels are defined by the main land uses, i.e., industrial or transport sites and other land uses. The second level of concentration is the so-called "guide value". If this is exceeded, the area presents a level of contamination with ecological risk (e) or health risks (t). Different guide values are set for industry and transport areas (highest guide value), and for all other land uses (lowest guide value). For the assessment of agricultural land, the threshold of lower guide values for the sample applies. The elemental levels in peat were all below regulatory standards, except As in site E (13.0 ± 1.0 mg/kg), which exceeds the Finnish threshold value of 5 mg kg −1 . However, even if anthropogenic arsenic pollution is widespread, As in soil is generally considered mainly of geological origin, with a higher background concentration in clayey soils 66 . Chemometric analysis. The above presented results showed several differences between peat samples from moss samples. Chemometric analyses have been applied to confirm this differentiation and to visualize sample differences.
Usually, a chemometric treatment is needed when the sample size is large and exceed the variables' number. Nevertheless, principal component analysis should be applied to reduced observations to perform an explorative analysis and to provide a graphical representation of a natural grouping of samples.
In this regard, t, a stepwise variables selection was carried out for all the analyses performed aiming to reduce the number of variables and select the most informative ones. The minimum Bayesian information criterion (BIC) was applied to point out which variables exhibited the best separation among peat and moss samples.
The variables' stepwise results for C, H, N, O and S contents were reported in Table S9; the only variables excluded were %N and %O. These findings match the analytical results differences. Principal component analysis was performed with the selected variables and reported in Fig. 3. The first two principal components accounted for 78.3% of the total variability. Peat samples (red) are all located in the right part of the plot, while the moss ones are in the left one, except the moss of the G site. The grouping of peat and moss samples is more evident on PC1, which was mostly influenced by %C and O/C variables. So, these variables seem to better characterize the samples' grouping.  (Table S10). It is noteworthy that these wavelengths fall in the typical range of lignin and cellulose compounds, as reported in Table S5. Therefore, the peat and moss samples' differences in the organic matter should be attributed to these compounds. Principal component analysis plots for FTIR vibration bands were reported in Fig. 4. The first two PCs accounted for 96.1% of the total variance. Variables showed a similar contribution on PC1 and PC2. Moss samples (blue), except for the G site, are located in the lower part of the graph, while peat ones (red) are in the higher one.
In the multi-element analysis, the variable stepwise included twenty-nine of forty-nine elements determined, i.e. S, Ca, Hg, Mo, P, Zr, As, Hf, Be, La, U, Al, Ce, Ba, Mn, Ni, Sm, Th, Pb, Cr, Si, Lu, Ti, Sn, Cu, Rb, Ga, Zn and Bi (Table S11). Notably, the variables included to differentiate samples belong to different classes such as macroelements, trace-elements, and REEs. This finding confirms the use of different element classes to characterize environmental samples. Principal component analysis has been performed with the selected elements (Fig. 5). The first two PCs explained 74.3% of the total variability with a partial grouping of peat and moss samples. In fact, moss samples from A, C, E and G sites are clearly located in the right part of the score plot, while the samples from the other sites appear partially overlapped with peat ones. The major differences between peat and moss are mainly due to the different Zn and As contents, while the similarities are attributable to the S and Zr contents.

Conclusions
The elemental distribution and physicochemical characteristics in peat deposits and Sphagnum moss of the Tierra del Fuego region (south Patagonia) were estimated. Our study demonstrated the need to monitor the content of elements in the peatlands of Southern Argentina to assess the risk of anthropogenic pollution and moss's role in the metals' biogeochemical cycle. The anthropic impact could, in fact, endanger the ecosystems of peat bogs, sites of great ecological importance since they can be a refuge for rare flora and fauna species and, for this reason, play a fundamental role in the conservation of biodiversity and safeguard of ecosystem services. Therefore, it is important to design peat bog research and management strategies to increase the knowledge and awareness of scientific communities and local populations on the importance of these ecosystems. This study provides baseline information useful for the evaluation of the resilience capability of the Patagonian ecosystem in a mid-long term.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.