Inherited geochemical diversity of 3.4 Ga organic films from the Buck Reef Chert, South Africa

Archean rocks contain crucial information about the earliest life forms on Earth, but documenting these early stages of biological evolution remains challenging. The main issue lies in the geochemical transformations experienced by Archean organic matter through its multi-billion-year geological history. Here we present spatially resolved chemical investigations conducted on 3.4 Ga organic films from the Buck Reef Chert, South Africa which indicate that they possess significantly different chemical compositions. Since these organic films all underwent the same post-depositional geological history, this geochemical diversity is most likely inherited, reflecting original chemical differences which were not completely obliterated by subsequent burial-induced degradation processes. These results demonstrate that early Archean organic films carry chemical information directly related to their original molecular compositions. This paves the way for the reconstruction of the initial chemical nature of organic microfossils found in ancient rocks, provided that the geologically-induced chemical transformations they underwent are properly constrained. Information on the initial chemical nature of Archean organic materials can survive in rocks despite a complex geological history, according to spatially resolved chemical analyses of 3.4-billion-year old organic films.

T he ancient fossil record contains crucial information about the evolution of early life on Earth [1][2][3] . Yet, the biogeochemical signals that can be retrieved from ancient rocks are not pristine: multiple alteration processes inevitably modify the chemical composition of organic molecules over geological timescales, which makes it particularly challenging to reconstruct the original chemical nature of Paleoarchean organic-rich structures, and thence to distinguish authentic microbial remains from abiotic ones [4][5][6] .
Many studies have reported the occurrence of organic films in a number of Paleoarchean cherts 13,15,[28][29][30][31][32][33][34][35][36][37][38][39] . Yet, besides the work of Westall, et al. 32 and Alleon, et al. 13 , these studies did not investigate the molecular composition of these microstructures. Here, we report results of spatially resolved elemental and molecular investigations conducted using transmission electron microscopy (TEM) and scanning transmission X-ray microscopy (STXM) on 3.4 Ga organic films from the Buck Reef Chert (BRC, Kaapvaal Craton, South Africa). We interpret the geochemical diversity of the BRC organic films investigated as inherited, i.e., the chemical differences between the films result from initial chemical differences, even though these organic films underwent a significant degradation during their geological history.

Results
Scanning electron microscopy (SEM) observations reveal that the BRC organic films are draped over the microquartz crystals ( Fig. 1). Besides slight wrinkles, these films display an overall smooth surface texture, with clear imprints of microquartz crystals. Parts of some of the films are below the freshly exposed surface investigated, i.e., trapped within the microquartz matrix of the sample (see also Supplementary Fig. 1).
According to Raman data, the BRC organic films are rather homogeneous in terms of structural order of their aromatic skeleton. In fact, all these films display similar Raman signals typical of disordered carbonaceous materials, with a rather intense and narrow D1 band and a less intense composite G + D2 band ( Fig. 2 and Supplementary Fig. 2). Application of the RSCM (Raman spectroscopy of carbonaceous material) geothermometer 40 indicates that the organic films were exposed to peak temperature conditions around 300°C, which is consistent with previous studies of the geological history of the BRC 41,42 . Assimilated, SEM and Raman spectroscopy testify to the syngenicity of the BRC organic films, i.e., they were trapped within the chert at the time of its formation.
TEM investigations conducted on Focused ion beam (FIB) foils extracted from five BRC organic films (red segments on Films 1-5, Fig. 1) evidence the absence of gap/space between the organic films and the chert matrix, even at the nanometer scale (Fig. 3). Besides the microquartz crystals, no mineral phases can be observed within the FIB foils investigated: the organic films are exclusively composed of carbonaceous compounds (Fig. 3).
Spatially-resolved chemical investigation using STXM reveals that the BRC organic films possess variable N/C values (from 0.05 to 0.16 ± 0.02; Fig. 4a), indicating a significant chemical variability among the BRC organic films. The C-X-ray absorption near edge structure (XANES) spectra (Fig. 4) display multiple absorption features. The large and more or less asymmetrical features centered at about 285.2 eV signal the presence of olefinic/aromatic groups and heterocycles, while the more or less defined peaks/ shoulders centered at 286.7 eV reveal the presence of imine (C=N), nitrile (C≡N), carbonyl (C=O) and/or phenolic (Ar-OH) groups 27,43,44 . The peaks at 289.4 eV signal the presence of hydroxyl (OH) groups while the relatively intense contributions at 288 eV likely reflect the contribution of aliphatic carbons, although a contribution from amide ((R 1 , R 2 )N-C=O) groups at 288.2 eV cannot be excluded 27,43,44 . Of note, the relative intensities of these features do not vary within each film (Supplementary Fig. 3) but vary significantly between the films investigated ( Fig. 4 and Supplementary Fig. 4), confirming the significant chemical differences between the BRC organic films investigated. In fact, the organic films 1, 2, and 4 contain more heterocycles and O-rich and N-rich functional groups than the organic films 3 and 5, the organic film 3 being richer in aliphatics and hydroxyls than the organic film 5. Differences between the films are less obvious at the N K-edge. Still, the features centered at 401.2 eV, likely related to the presence of amide, imine and/or nitrile groups, are only detected in the organic films 3 and 5, and the contribution of N-rich heterocycles (the peak at 399.9 eV and the shoulder around 398.8 eV 27,43,45 ) is higher for the organic films 1, 2, and 4 than for the organic films 3 and 5. Altogether, despite a similar morphology, the BRC organic films investigated here are not chemically identical.

Discussion
Inherited geochemical diversity. The BRC organic films investigated have experienced only partial degradation despite a long geological history punctuated by episodes of metamorphism in the lower greenschist facies (Fig. 2). The molecular composition of the BRC organic films is not that different, at the first order, from that of organic films found in chert of the 3.4 Ga Strelley Pool Formation 13 . Such preservation, with significant quantities of hydrogen, nitrogen and oxygen retained in the BRC organic films, may be related to their early entombment within silica. In fact, experimental work demonstrated that such early entombment within silica limits the molecular degradation of microorganisms during advanced diagenesis 21,23 . Early entombment is believed to explain the high level of molecular preservation previously reported for organic microfossils in cherts of the 1.9 Ga Gunflint 12 and the 3.4 Ga Strelley Pool 13 formations. This study thus exemplifies once again that Archean cherts are pertinent targets for the search for (at least partially) chemically preserved remains of ancient life.
While the BRC organic films are morphologically and texturally similar, they differ significantly from each other in terms of elemental and molecular compositions (Fig. 4). These chemical variations cannot be due to post-depositional mineralogical effects. In fact, although distinct mineral phases will not impact similarly the thermal degradation of organic molecules 17,[21][22][23][24]26,46 , the BRC organic films investigated are only associated to microquartz (Figs. 1 and 3). The chemical variability observed here is also difficult to explain by postdepositional processes involving fluid circulation. In fact, because liquid water enhances their reactivity [47][48][49][50][51][52] , the circulation of fluids can only be detrimental to the molecular preservation of organic compounds. Yet, there is no organic carbon at microquartz crystal boundaries in the chert matrix which would testify to the infiltration of oil or to the fluid-deposition of organic carbon, in contrast to what was described in the 340 Ma Red Dog Zn-Pb deposit (Alaska) 53 or in the 3.5 Ga chert of the Mount Ada Basalt (Pilbara, Australia) 54 .
Rather than a result of post-depositional alteration processes, the chemical variability observed here can be seen as a legacy, i.e., this geochemical diversity is very likely inherited. In other words, the BRC organic films investigated here were initially chemically different and post-depositional degradation processes did not completely obliterate these differences. This is totally consistent with recent experimental results having suggested/predicted that organic molecular heterogeneities can withstand diagenesis and be recognized in the geological record 27 .  Implications for the origin/nature of the BRC organic films. As is the case of many microstructures found in Paleoarchean cherts 5,6 , determining whether the BRC organic films investigated here are biogenic or abiotic is not straightforward. Carbonaceous laminations and 100 µm-thick films interpreted as fossilized microbial mats were previously reported in the BRC 29,34,35,37 .
Because the organic films investigated here are much smaller, the direct comparison is difficult. Organic films interpreted as fossilized biofilms were previously described in a black chert sample from another unit of the Kromberg Formation 31 and from the Strelley Pool Formation 13,15 . In contrast to the BRC organic films investigated here, the Kromberg organic films were reported to be composed of multiple layers of parallel 0.25 µm-thick filaments that are thickly coated with a heterogeneously textured film that ranges from ropy, granular, smooth, to holey 31 , while the Strelley Pool organic films were reported to display surface textures that range from granular to smooth and slightly wrinkled 13,15 . In addition, contrasting with the BRC organic films investigated here, the Strelley Pool organic films were reported to have broadly similar elemental and molecular compositions, consistent with geologically altered biological compounds, whatever their fine scale texture 13 . It thus appears difficult to interpret the BRC organic films investigated here as being similar in nature to the Kromberg and Strelley Pool organic films. Even though this would not constitute a biosignature in itself, the absence of amide and carboxylic functional groups in some of the BRC organic films investigated (Fig. 4) makes it difficult to conclude on their biogenicity at this stage. The fact that the BRC organic films were initially chemically different (cf. above) appears consistent with a biogenic origin. In modern stromatolites, extracellular polymeric secretions can indeed display heterogeneous chemical degradation rates, and N-enriched organic residues are produced from the degradation of carbohydrates by heterotrophic microorganisms 55 . Yet, it can be conceived that the BRC organic films investigated here were produced by abiotic processes having occurred prior to the geological history of the BRC. Abiotic processes can indeed lead to the formation of nitrogen-rich organic compounds and chemically heterogeneous kerogen-like particles as those found in carbonaceous chondrites [56][57][58][59] . In addition, abiotic hydrothermal or atmospheric processes on the Archean Earth, may have produced condensed forms of organic matter, as already suggested for some carbonaceous materials of the BRC 35 . Worse, experimental studies of thermal degradation processes have shown that abiotic organic compounds may produce organic residues difficult to distinguish from remains of biogenic organic compounds 27,60 , especially given that the thermal degradation of biogenic organic compounds has been shown to be more abstruse than historically believed [16][17][18][19][20]22,[24][25][26]46,[61][62][63] . At the end, although they fulfill many (if not all) of the commonly used criteria of biogenicity, even the most conservative ones 64 , further quantitative constraints on the impact of a geological history on biogenic and abiotic organic compounds appear necessary to determine the exact origin of the BRC organic films investigated here.
Concluding remarks. Archean rocks have been exposed to complex geological processes that have caused the degradation of biomolecules and/or abiotic synthesis of organic molecules with chemical features similar to those produced by living organisms 5,6,65 . Here, we provide compelling evidence that inherited molecular differences can survive a 3.4-billion-year-long geological history punctuated by metamorphic events. Altogether, the present results highlight that ancient organic microstructures still carry information on their initial chemical nature. This initial nature could be reconstructed provided robust experimental constraints on the impact of a geological history on the chemical transformations of organic compounds. Undoubtedly, such reconstructions, associated with detailed chemical characterization such as exemplified in this work, would have major implications for our understanding of early life on Earth and potentially beyond.  69 . Silica precipitation occurred directly from seawater on a subsiding volcanic platform 35,69,70 or resulted from the hydrothermal silicification of volcanoclastic deposits in a shallow water platform 36,71,72 . For this study, BRC samples of the lower black and white chert were collected in 2017 (25°55′33.6″S, 30°54′28.8″E; top of the K1e Member). A black chert sample was selected for this study, because Paleoarchean organic structures interpreted as fossilized biofilms were previously reported in similar rocks of the same Kromberg Formation 29,31,34,35,37 . The chert sample investigated here was only mechanically fractured to expose fresh surfaces for further Raman, SEM, TEM, and STXM experiments, as previously done for similar Paleoarchean cherts 12,13,54 , thereby avoiding organic contamination issues related to resin embedding and polishing.
Raman spectroscopy. Raman spectroscopy was used to document the degree of structural organization of the aromatic skeleton of the carbonaceous material composing the films investigated. Raman data were obtained using a Renishaw INVIA spectrometer (IMPMC, Paris, France) in a confocal configuration equipped with a 514.5-nm wavelength 50-mW Modulaser Ar laser, on a surface of freshly fractured chert sample, at room temperature. The configuration used yields a horizontal resolution of ≈1 μm for a laser power delivered at the sample surface below 1 mW, thereby preventing irreversible thermal damage 73,74 . Extraction of spectral parameters from peak fitting procedures and estimation of peak temperature experienced by organic carbon were done following the methodology described by Beyssac, et al. 40 .
Scanning electron microscopy. SEM was used to document the texture of the organic films measured using Raman before their extraction by FIB milling. Chert fragments analyzed with Raman were mounted on aluminum stubs without any additional preparation, except Pt coating. SEM observations were performed with a SEM-field emission gun ultra 55 Zeiss (IMPMC, Paris, France) at 2 and 10-kV accelerating voltage for secondary and back-scattered electron analyses, respectively.
Focused ion beam. FIB foils were extracted for TEM and STXM investigations using a FEI Strata DB 235 (IEMN, Lille, France). Milling at low Ga-ion currents allowed minimizing artefacts such as gallium implantation, amorphization, composition changes or redeposition of the sputtered material 75,76 .
Transmission electron microscopy. TEM analyses were performed on FIB foils to document, down to the nanoscale, the texture and relationships between microquartz and the organic materials composing the films investigated. TEM observations were conducted with a JEOL 2100 field emission gun microscope (IMPMC, Paris, France) operating at 200 kV. Z-contrast STEM imaging was performed using the high-angle annular dark field mode.
STXM coupled with XANES spectroscopy. STXM-based XANES spectroscopy analyses were performed on FIB foils to document both the carbon and nitrogen speciation and the N/C values of the films investigated using the HERMES STXM beamline at the synchrotron SOLEIL (Saint-Aubin, France) 77,78 . Energy calibration was done using the well-resolved 3p Rydberg peak of gaseous CO 2 at 294.96 eV for the C K-edge, and using the 1 → π* photoabsorption resonance of gaseous N 2 at 400.8 eV for the N K-edge. XANES hypercube data (stacks) were collected with a spatial resolution of 35 nm at energy increments of 1 eV over the 250-450 eV region and at energy increments of 0.1 eV over the carbon (270-340 eV) and the nitrogen (390-450 eV) K-edges with a dwell time of less than 1 ms per pixel to prevent irradiation damage 79 . Stack alignment and extraction of XANES spectra were done using the aXis2000 software (ver2.1n). The XANES spectra shown here are the sum of hundreds of spectra collected over homogeneous organic-rich areas of several hundreds of square nanometers. N/C atomic ratio values and quantitative spectral parameters were estimated using both the methodology outlined by Alleon, et al. 80 and using the QUANTORXS freeware 44 .

Data availability
The data supporting the findings of this study are available on Zenodo (https://doi.org/ 10.5281/zenodo.4015138). The samples are available from the corresponding author (J.A.) upon request.