Potentially Massive and Global Non-Pyrogenic Production of Condensed “Black” Carbon through Biomass Oxidation

With the increased occurrences of wildfires worldwide, there has been an increase in scientific interest surrounding the chemistry of fire-derived “black” carbon (BC). Traditionally, wildfire research has assumed that condensed aromatic carbon (ConAC) is exclusively produced via combustion, and thus, ConAC is equated to BC. However, the lack of correlations between ConAC in soils or rivers and wildfire history suggests that ConAC may be produced non-pyrogenically. Here, we show quantitative evidence that this occurs during the oxidation of biomass with environmentally ubiquitous hydroxyl radicals. Pine wood boards exposed to iron nails and natural weather conditions for 12 years yielded a charcoal-like ConAC-rich material. ConAC was also produced during laboratory oxidations of pine, maple, and brown-rotted oak woods, as well as algae, corn root, and tree bark. Back-of-the-envelope calculations suggest that biomass oxidation could be producing massive non-pyrogenic ConAC fluxes to terrestrial and aquatic environments. These estimates (e.g., 163–182 Tg-ConAC/year to soils) are much higher than the estimated pyrogenic “BC” fluxes (e.g., 128 Tg-ConAC/year to soils) implying that environmental ConAC is primarily non-pyrogenic. This novel perspective suggests that wildfire research trajectories should shift to assessing non-pyrogenic ConAC sources and fluxes, developing new methods for quantifying true BC, and establishing a new view of ConAC as an intermediate species in the biogeochemical processing of biomass during soil humification, aquatic photochemistry, microbial degradation, or mineral–organic matter interactions. We also advise against using BC or pyrogenic carbon (pyC) terminologies for ConAC measured in environmental matrices, unless a pyrogenic source can be confidently assigned.


Section 2. Materials and methods for supplementary analyses 2.1. Quantification of organic carbon and hydrogen
Elemental analysis for carbon (C%) and hydrogen (H%) was performed using a Thermo Finnigan FlashEA 1112 elemental analyzer fitted with a CHN column.Samples were analyzed in triplicate and calibrated to a five-point external calibration curve of nicotinamide (CE Elantech, Inc.).Empty tin capsules were analyzed as blanks and to evaluate for any sample carryover.An aspartic acid standard (CE Elantech, Inc.) was also analyzed as a control sample to confirm the accuracy of the measurements.C% measurements were with relative standard deviations below 5%.

Quantification of metals
Metal (Cu% and Fe%) analysis was performed after ashing 20-30 mg solid samples at 600 O C in a temperature-controlled oven for 24 hours.The ash residue was then digested with 5 mL aqua regia (HNO3:HCl = 1:3 molar ratio) for 12 hours, and the acidic mixture was evaporated in a sand bath at 70 O C.Then, the residue was resuspended in 5 mL 65% HNO3 and the suspension was let to acid-digest for 6 hours, after which the acid was evaporated in a sand bath at 70 O C. Residue was resuspended in 5 mL 65% HNO3 once more and HNO3 was evaporated after the 6-hour digestion.Then, 5 mL 2% HNO3 and ~1 mg of La(NO3)3 were added, and the nitrates formed during the acid digestions were dissolved with the assistance of a 10-minute sonication.The solutions were filtered using 0.22 uM 0.2 µm Teflon (PTFE) filters.Dissolved Fe and Cu were quantified using a Shimadzu AA-7000 atomic absorption spectrometer with a flame atomizer.Copper nitrate (1000 ppm, Acros Organics) and iron nitrate (10000 ppm, ASSURANCE) certified standards were used for external calibration.Instrument and procedural blanks were used to evaluate for contamination and sample carryover.
As it was known that the Pine 2 sample had been pressure-treated, its control and Fe-exposed samples were additionally assessed using inductively coupled plasma -mass spectrometry (ICP-MS).Samples were prepared as described above without the addition of La(NO3)3.Samples were analyzed on a Thermo Scientific Element XR double-focusing magnetic sector field ICP-MS instrument at the College of Sciences Major Instrumentation Cluster (COSMIC) facility at Old Dominion University (Norfolk, VA).The instrument was equipped with a perfluoroalkoxy alkane (PFA) microflow nebulizer and a quartz cyclonic spray chamber.Samples were taken up via an Elemental Scientific SC µ-DX autosampler with PFA probe, PFA sample tubing, and PFA sample vials.All ICP-MS front-end and sample input components were nitric-acid cleaned and sample probe was rinsed with 2% HNO3 (matrix-matched) between samples.The Pine 2 Control and Fe-oxidized samples, along with a procedural blank, were screened for a variety of elements that are common in pressure-treatment agents (Al, As, B, Cd, Co, Cr, Cu, Fe, Mn, Ni, Pb, Sb, Si, Ti, V, Zn). 2 Concentrations were calculated by the Element XR software based on standard calibration curves.Computed calculations were manually double-checked afterwards.In addition to the procedural blank, a variety of other blanks were analyzed throughout the analytical sequence to verify for no background contamination.Metal concentrations were blank-corrected using the procedural blank and converted to weight percentages relative to sample weight (wt.%).

Solid-state nuclear magnetic resonance (NMR) analysis
Solid-state NMR analysis was performed on dried powdered samples, which were packed in a 4 mm Zirconia (ZrO2) rotor with a polychlorotrifluoroethylene (Kel-F) cap.Analysis was done on a 400 MHz (9.4 Tesla) Bruker BioSpin AVANCE II spectrometer fitted with a 4 mm magic angle spinning (MAS) probe at the COSMIC facility.One-dimensional quantitative 13 C spectra were acquired using the quantitative multi-pulse cross-polarization MAS (MultiCPMAS) pulse program. 3Samples were spun at the magic angle at 14 kHz and analyzed using a relaxation delay of 1 s, 5000 scans, 5 cross-polarization segments, and a total contact time of 3.30 ms.The obtained spectra were phased, calibrated to an external adamantane standard 4 , and multiplied by an exponential window function (EM) of 50 Hz.Spectra were then baseline-corrected and integrated in the following ranges: Methyl: 0-20 ppm, Methylene: 20-45 ppm, O-Alkyl: 45-90 ppm, di-O-Alkyl: 90-110 ppm, Aryl: 110-146 ppm, Aryl-O: 146-165 ppm, Carboxyl/Ester (COO): 165-184, Carbonyl (CO): 184-220 ppm.All data were processed using the Bruker TopSpin 4.0.7.software.

X-ray fluorescence imaging and Fe X-ray absorption near edge structure (XANES)
Micro X-ray fluorescence (μ-XRF) images and micro X-ray absorption near edge structure (μ-XANES) spectra at the Fe Kedge were collected on the X-ray Fluorescence Microprobe (4-BM) beamline at the National Synchrotron Light Source II (NSLS-II, USA).The μ-XRF images were collected with an incidence X-ray energy of 13,500 keV at a step size of 10/40 μm.Sample fluorescence was detected using a 7-element Vortex detector and was integrated in the pre-determined regions of interest.Fe Kedge XANES spectra were measured with a step size of 0.3 eV at the absorption edge and were aligned using a reference spectrum of Fe foil (first inflection point of the edge set at 7,112.0 eV).A focused beam (~ 10 × 10 μm 2 ) was used throughout the experiment.
Image processing and data analysis were performed in Larch and R. Energy alignment, and XANES spectral normalization were conducted using Athena.Energies of pre-edge features were determined using 2 nd derivative spectra in the GRAMS software.Pre-edge centroids were calculated as intensity-weighted means of pre-edge peak energies and compared with literature values. 5ll data have been corrected by applying a 0.9 eV energy shift because of a difference in calibrating Fe foil energy.

Extraction of organic matter
Base-extraction was selected as the method for obtaining a representative liquid extract of the solid control and Fe-oxidized wood samples.The extraction was performed using 0.1 M sodium hydroxide (Fisher, ACS Certified grade) at a ratio of 0.5 g sample/100 mL extractant.Suspensions were vigorously stirred on a shaker table for 24 hours.Then, the supernatant was removed and substituted with new 100 mL of extractant.The extraction was done three times over 3 x 24 hours to result in a total of 300 mL base-extract of each sample.Base-extracts were then filtered through pre-combusted 0.7 µm glass-fiber filters (GF/F, Whatman, 47 mm diameter) and cation-exchanged using a Dowex 50Wx8 resin (Acros Organics).A procedural blank of sodium hydroxide was processed the same way.All extractions were performed under inert (N2) atmosphere.The procedure followed the International Humic Substances Society (IHSS) guidelines and is evaluated and described in greater detail elsewhere. 6

Ultrahigh resolution mass spectrometry (FT-ICR-MS)
Cation-exchanged base-extracts were diluted to 50 mg/L carbon-equivalents (50 mgC•L -1 ) and then further diluted with methanol (CH3OH, Fisher Scientific, Optima LC-MS grade) to give 1:1 CH3OH:H2O mixtures.Samples were analyzed on a Bruker Daltonics 12-Tesla Apex Qe FT-ICR-MS housed in the COSMIC facility.The instrument was calibrated daily with a polyethylene glycol standard and instrument blanks were analyzed in-between samples to assure for no sample carryover.Samples were infused into the Apollo II electrospray ionization (ESI) source at flow rate of 120 µL/h and molecules were ionized in negative mode.Ionization voltages were optimized on a per-sample basis to assure for uniform spray currents across the dataset.The ionized molecules were collected in a hexapole, filtered by a quadrupole for a mass range of 200-1200 m/z, pre-concentrated in a second hexapole, and transferred into the ICR cell where 300 transients were collected.They were co-added, and the resultant free induction decay was zero-filled and sine-bell apodized.After a fast Fourier transformation, spectra were calibrated to naturally abundant fatty acids, dicarboxylic acids, and compounds belonging to the CH2-homologous series. 7Peaks with signal-to-noise above 3 were exported to MATLAB where salt, blank, and 13 C isotopologue peaks were removed from each spectrum.Molecular formulas were assigned to each mass list using the Molecular Formula Calculator from the National High Magnetic Field Laboratory (Tallahassee, FL).][10] No ambiguous assignments were left in the final formula lists (i.e., for each mass spectral peak there was only one molecular formula).Only formulas containing carbon, hydrogen, and oxygen elements are used hereafter (i.e., CHO formulas).
Molecular formulas are further classified based on their modified aromaticity index (AIMOD), a measurement of the doublebond density in a molecule. 11,12 ompounds with AIMOD = 0 are classified as "aliphatic".Molecules with 0 < AIMOD < 0.5 have either an aromatic moiety that is highly functionalized with aliphatic groups or have olefinic/alicyclic bonds.Molecules with 0.5 ≤ AIMOD < 0.67 are classified as aromatic.][13] The calculation for this index is shown below.
Data analysis was performed using codes of the MATLAB-based Toolbox for Environmental Research (TEnvR). 14

Section 3. Bulk elemental characteristics
Bulk elemental characterization revealed an enrichment in carbon and loss of hydrogen (Table S1), which resulted in the decrease of H/C ratios for the three specimens.This is indicative of forming aromatic carbon following exposure to Fenton chemistry (Figure 2).Solid-state 13 C NMR also confirmed the formation of aromatic carbon by showing an increase in aryl and phenolic (Aryl-O) functional groups (49 -98%, Figure S2).The increase in aromatic content is partially responsible for the darkening of the wood adjacent to each nail.In addition to phenolic groups, one-to two-fold increases in ketone (CO) and carboxyl (COO) groups are found (Figure S2), which are indicative of oxidation and are expected for these Fe-wood systems.Mass spectrometric analysis of alkaline extracts of the two pine samples showed a shift of the molecular composition towards a higher O/C ratio (Figure S4), which is a clear indication that the charcoalified materials had experienced oxidation relative to their control biomass.The observed changes in the organic matter composition show great similarity to what one would observe from plant litter being degraded in the environment 15,16 .We further confirmed the involvement of Fenton chemistry by examining the Fe oxidation state using synchrotron X-ray absorption near-edge fine structure (XANES).The presence of abundant oxidized Fe III in the wood samples adjacent to the Fe nails is consistent with the action of Fenton oxidation (Figures S5-S6).*Annual ConAC production is calculated over 12 years of Fe-exposure for the two pine boards and over 1 year of exposure of the maple board.It is expressed as percentage relative to the organic carbon (OC).ConAC in the controls is not included in the calculation, because fresh wood is ConAC-free. 17The average biomass-to-ConAC conversion (0.195 %/year) was computed by averaging the rates derived from the two pine boards (0.203, 0.186 %/year) .

Section 4. Structural characterization using one-dimensional NMR spectroscopy
Bulk structural characteristics of the two wood samples before and after exposure to Fenton chemistry were determined using one-dimensional nuclear magnetic resonance (NMR) spectroscopy (Figure S2).Solid-state 13 C NMR analysis is a classical approach for evaluating solid environmental matrices 18,19 , and utilizing the multi-pulse cross-polarization magic angle spinning 13 C NMR technique (MultiCPMAS 13 C NMR) allows for quantitative reporting of the total content of the various 13 C functionalities in the studied samples.The obtained MultiCPMAS 13 C NMR spectra for the control samples were highly characteristic for woody samples. 20Lignin and carbohydrate signatures are easily identified and seem to predominate in the spectra (left panels of Figure S2), as these biopolymers are of the highest abundance in woody biomass, carbohydrates being the most abundant and lignin being the second most abundant biopolymer, respectively. 21Lignin's presence is identified by the peaks associated with its methoxy (CH3O-, 56 ppm) and phenolic (CAryl-O, 147 ppm) functionalities.Cellulose's glycosidic units present seven peaks in the area between 60 and 110 ppm (C1-C6,).There are several peaks in the region of 0-45 ppm, where methyl (CH3-) and methylene (-CH2-) functionalities usually resonate.These are generally associated with acetylated glucose units in hemicelluloses and resinous substances.Pine 2 is a pressure-treated wood, thus its aliphatic resonances could also originate from alkyl groups in the quaternary ammonium ligands of commonly used pressure-treatment reagents. 2The peak at 172 ppm is associated with carboxyl groups (-COOH) or derivatives (-COOR) such as amides or esters, and this peak likely corresponds to acetyl esters of hemicellulose, cellulose lactones, and carboxyl groups of gluconic and glucuronic acids of oxidized cellulose.
Nearly all peaks found in the Fe-oxidized samples are present in the control samples (left panels of Figure S2), which is expected as both samples are from the same parent source.The distribution (peak intensity) of carbon moieties has changed after the decade-long Fenton oxidation -carbonyl, carboxyl, phenolic, and aromatic groups increase (right panels of Figure S2) showing that the exposure to Fe in the nail over decade-long exposure period has enriched the two Fe-oxidized samples in aromatic structures.This appears to be at the expense of cellulosic materials, as is evident by the diminution of di-O-alkyl and O-alkyl groups.Carbohydrates are labile towards oxidative processes and their degradation pathways via the Fenton reaction have been previously studied 22,23 , which explains the presented data here.It must be noted that the abundance of Fe in the Fe-oxidized samples (and copper in the Pine 2 samples) may have quenched some of the aromatic signals 24,25 , thus the abundance of aromatic carbons in the Fe-oxidized samples is likely much higher.

Section 5. Molecular characterization using ultrahigh resolution mass spectrometry (FT-ICR-MS)
To investigate the 12-year exposure to Fenton chemistry on the molecular level, ultrahigh resolution mass spectrometry was employed.Given that the utilized FT-ICR-MS instrument did not have an ionization source suitable for analyzing solid samples, the four samples of this study were analyzed after an alkali extraction, a classical approach from the soil sciences.Though a large fraction of the extracted sample remains as a solid residue (known as humin), the base-extract is usually representative of the structural composition of the evaluated sample. 6To confirm this for our samples, we employed solid-state NMR analysis on the freeze-dried base-extracts and compared them with the original solid samples (Figure S3).The base-extracts contained all functional groups native to the whole samples concluding that our base-extracts were representative.A large portion of the cellulose was not extracted (O-alkyl and di-O-alkyl peaks at 62-105 ppm), however, carbohydrates do not ionize efficiently in the ESI source, and are therefore barely covered by the analytical window of the ESI-FT-ICR-MS.Thus, their lower amount in the base-extracts was not of serious concern.

Figure S3
. Solid-state 13 C NMR spectra of whole samples (in blue or red) relative to freeze-dried base-extracts (in black).
7][28][29] While this instrument only measures mass-to-charge values for all ionized molecules in a sample, its ultrahigh precision allows for a unique molecular formula to be assigned to each mass peak.While not quantitative, it is another common tool for detecting the presence of ConAC in various samples. 30,31 imilar to the assumption employed for the BPCA method (i.e., quantified ConAC is labeled as pyrogenic BC), formulas with modified aromaticity index 11,12 AIMOD ≥ 0.67 corresponding to ConAC are often labeled as BC 10,32,33 , and if they contain a N or S atom, they are even referred to as black nitrogen 34 and black sulfur 35 , respectively.The assigned molecular formulas to Pine 1 and 2 controls and Fe-oxidized samples are classified using a presence/absence approach 36 in three categories: Fenton-labile (formulas present in the control samples); Fenton-resistant (formulas present in both samples); and Fenton-produced (formulas present in the Fe-oxidized sample).This approach allows one to observe what molecular features were produced upon the Fenton oxidation of wood (Figure S4).

Figure S4. Van Krevelen diagrams (H/C vs. O/C
) of FT-ICR-MS formulas identified only in the control (Fenton-Labile), only in the Fe-oxidized (Fenton-Produced), or in both samples (Fenton-Resistant) using a presence/absence approach. 36The number of formulas of each class (and corresponding percentages) are given in parentheses in the legends.The van Krevelen space is separated based on modified aromaticity index (AIMOD) thresholds. 11,12 eprinted with permission from Goranov. 1 Copyright ©, 2020, by Aleksandar Ivaylov Goranov, All Rights Reserved.
After exposure to Fenton chemistry there are clear shifts in the molecular composition of both woods.For both samples, molecular formulas with higher O/C ratio and lower H/C ratio evolve, which is also accompanied by loss of numerous aliphatic/olefinic compounds (blue markers).These compounds are likely carboxyl-containing aliphatic molecules recently found and proposed to be important in the formation process of aromatic compounds, including ConAC, in soils. 37,38 he pools of Fenton-produced molecules contain lignin-like and carbohydrate-like formulas, as well as some ConAC formulas falling under the AIMOD ≥ 0.67 line (Figure S4).If these data are interpreted using the traditional approaches in the wildfire literature 10,[32][33][34][35] , it can be concluded that ConAC is produced after Fenton exposure.This data complements the trends presented using BPCA analysis and are in agreement with the previously published non-pyrogenic pathways for formation of ConAC from ligninaceous molecules. 39,40 he observed changes are also consistent with humification reactions in soils 15,16 , which are known to produce aromatic and condensed aromatic structures. 41,42 Tis observation also gives more merit to the previously made hypothesis that soil humics are derived from lignin after processing via radical electrocyclization reactions. 38However, a more in-depth analysis of these samples using humic extractions must be done to directly test this proposition.A recent study, which determined that microbe-induced humification produces condensed moieties 43 , is also in agreement with what is presented here in which Fenton exposure leads to humification and formation of ConAC.The molecular results here also parallel with the quantitative structural data from solid-state 13 C NMR data (Figure S2), further validating our proposition of oxidative chemistry driving the chemical composition and formation of non-pyrogenic ConAC.
A change in the molecular composition of carbohydrates is also evident.Wood contains mainly cellulosic carbohydrates 44 , which are known to be labile towards the Fenton reaction. 22,23 t is likely that the Fenton-produced formulas (blue markers) are of gluconic and glucuronic acids, products of the oxidative degradation of cellulose.][47]

Section 6. Distribution of Fe and Cu in wood and Fe oxidation state in relation to the Fenton chemistry
The distribution of Fe and Cu in two wood cross-sections (73 × 1 mm 2 , 20 × 2 mm 2 ), starting from the Fe-exposed area to the adjacent unaffected "control" area, were evaluated using micron-sized synchrotron X-ray beams (Figure S5a).The abundance of Cu follows the dark and light growth bands of wood with alternating Cu-rich and Cu-poor regions (Figure S5b) suggesting that these bands originated from the pressure-treatment.In contrast, Fe is present in patches of high-abundance hotspots in addition to the normal patterning that follows tree growth bands (Figure S5b) suggesting that the non-hotspot Fe distribution is intrinsic to the initial wood material.The Cu abundance in the Fe-exposed area is less when compared to the unaffected area (Figure S5c) indicating leaching of material over the time of exposure. 48,49 he abundance of Fe is greater in the nail-exposed area, with a twofold increase in Fe fluorescence intensity (0.014 to 0.031, Figure S5c).This increase in Fe abundance around the Fe-nail is likely from oxidative leaching by Fe from the nail, and diffusion and accumulation in the wood. 49Our interpretation for leaching and diffusion is also supported by the observation of distinct random patches throughout the "control" areas of the wood suggesting that Fe(II) and Fe(III) species diffused throughout time and possibly precipitated as distinct Fe minerals (oxides, hydroxides, and/or oxyhydroxides) or Fe-ConAC complexes. 50o evaluate changes in the Fe oxidation state and examine the role of Fe in altering the organic matter, Fe K-edge XANES spectra were evaluated at randomly selected locations in both Fe-oxidized and control areas of the wood.The pre-edge peak maximum, indicative of Fe-oxidation state 5 , appears between 7,113 -7,115 eV (Figure S5d).These peak maxima correspond to two types of Fe species that are found across the Fe exposure continuum: one being a mixture of Fe(II) and Fe(III), and the other being close to pure Fe(III).While the Fe abundance varies significantly among the Fe-oxidized and control regions, pre-edge energies of Fe in the two regions are not statistically different (p = 0.98).Additionally, a great variation in Fe oxidation state is observed within a few millimeters around the Fe-oxidized region when compared to the background "control" area of the wood.This large variation in Fe oxidation states has likely resulted from the Fenton chemistry, in which Fe is actively oxidized and reduced, leaving both Fe(II) and Fe(III) species in the wood.Such constant cycling of redox Fe speciation has been shown to be a key driver in the cycling of organic matter in soils. 51  The baseline value estimating biomass conversion to refractory ConAC (0.063 ± 0.008 %) was calculated by averaging the last three time points (0.072, 0.058, 0.059) and computing the associated standard deviation.

Section 1 .Figure S1 .
Figure S1.Charcoalification of pine (left) and maple (right) wood boards exposed to Fe nails.The blue arrow/circle indicate zones with no visible charcoalification, which were sampled to serve as representative controls.The red circles indicate zones with clear charcoalification.Condensed aromatic carbon (ConAC) quantities relative to organic carbon as well as lengths of exposure to nails are shown in the legends.The greenish tint of Pine 2 (left) is due to pressure treatment for retarding microbial degradation.Adapted with permission from Goranov. 1 Copyright ©, 2020, by Aleksandar Ivaylov Goranov, All Rights Reserved.

Figure S2 .
Figure S2.One-dimensional solid-state 13 C NMR analysis of whole wood samples (Control colored in blue and Fe-oxidized in red).Left panels show the NMR spectra, with labeled peaks (top) and chemical functionalities (bottom panel).Carbohydrate and Lignin resonances, along with the corresponding carbon numbers, are labeled in green (C and L, respectively).Right panels show the integrated chemical shift regions, with percent change in functionality abundance shown under each label in the x-axis.Adapted with permission from Goranov. 1 Copyright ©, 2020, by Aleksandar Ivaylov Goranov, All Rights Reserved.

Figure S5 .
Figure S5.Cu and Fe distribution and Fe speciation in Pine 2. a) A photograph of the imaged sample is shown on the top with the darkened Fe-oxidized area in the right.X-ray imaged cross-sections are labeled with white bars.Images and data for the short section can be seen in Figure S6; b) XRF elemental maps of Cu and Fe; c) Plot of integrated Fe (red) and Cu (blue) Kα intensities across the long section.The mean and standard deviation of for Kα intensity for each of the shaded 30-mm zones are shown as well; d) Plot of centroid position of pre-edge peaks of Fe K-edge XANES spectra across the imaged sample, with the means and standard deviations for each 30-mm zone.Reference lines of Fe(II) and Fe(III) preedge centroids are from previously published measurements.5Note that the images in a) and b) are compressed and elongated in the vertical direction, respectively, for clarity.

Figure S6 .
Figure S6.Cu and Fe elemental distribution in the short section from Pine 2 (see Figure S5): a) Cu and Fe Kα images with the color bar showing the relative intensity; b) Integrated Fe and Cu Kα intensities on this section.

Figure S7 .
Figure S7.Solid-state13 C NMR spectra of biomass materials used in laboratory oxidations.Boxed is the region of aromatics where aryl (110 -146 ppm) and phenolic (146 -165 ppm) functional groups resonate.These correspond to phenols in lignin and tannins, monoaromatic groups in suberin, as well as aromatic rings in proteins.

Table S1 .
Elemental analysis and benzenepolycarboxylic acids (BPCA) quantification of ConAC in the two pine wood samples.LOD = limit of detection

Table S2 .
Screening of Pine 2 using inductively couple plasma -mass spectrometry for the presence of elements typically contained in pressure-treatment chemicals.All results are reported as parts per million (ppm) of metal relative to sample weight.

Table S4 .
Quantities of carbon and ConAC at different points for the harsh oxidation experiment of maple wood (Figure4) *corrected for extraneous carbon inputs from the H2O+HCl+FeSO4+H2O2 reagents **there were no extraneous ConAC inputs (procedural blanks were ConAC-free)