Introduction

The earliest traces of Life occur in rocks dating from 3.5 to 3.2 Gyr ago (Westall et al. 2006a, b; Ueno et al. 2008; Derenne et al. 2008; Shen et al. 2009; Javaux et al. 2010; Westall 2011). In the Archean Eon (3.8 − 2.5 Gyr ago), the early atmosphere contained a very low concentration of O2 with estimates ranging from 0.001 % (Kump 2008) to 2 % (Towe 1996) of its present value. This period was followed by a drastic increase in O2 in the atmosphere between about 2.3 to 2.4 Gyr ago (Holland 1984; 1999; Barley et al. 2005).

The origins of this so-called Great Oxidation Event (GOE) are debated: photolysis of water vapour with subsequent loss of hydrogen to space, changes in the redox state and the influence of volcanic gas output (Kasting et al. 1993; Holland 2002; Gaillard et al. 2011), or short to long-term consequences of the appearance of oxygenic photosynthesis (Kopp et al. 2005; Catling and Claire 2005)? Studies of rocks 3.2 to 3.5 Gyr old document the probable existence of anoxygenic photosynthesis (Perry et al. 1971; Tice and Lowe 2004; Kappler et al. 2005; Konhauser et al. 2005; Olson 2006; Westall et al. 2006a; Bosak et al. 2007; Croal et al. 2009; Westall 2011; Westall et al. 2011), but the exact timing of the appearance of oxygenic photosynthesis is highly debated. There are three main schools of thought (Buick 2008):

  1. (i)

    Oxygenic photosynthesis appeared very early on Earth and the Archean atmosphere always showed high O2 rates (Ohmoto 1997; Ohmoto et al. 2006). Note that Ohmoto (Ohmoto et al. 2008) now suggests the presence of oxygen oases, rather than a totally oxidized atmosphere.

  2. (ii)

    Oxygenic photosynthesis appeared long before the GOE, but it took hundreds of millions of years to fill all near-surface oxygen sinks (Catling and Claire 2005).

  3. (iii)

    Oxygenic photosynthesis arose shortly before the GOE, causing major environmental and atmospheric changes (Kopp et al. 2005).

Hypothesis (i) is highly controversial and the evidence for early oxygenic photosynthesis is conflicting. On the one hand, the presence of detrital pyrite in paleosoils and mass independent sulfur isotopes indicate an anoxic Archean atmosphere prior to the GOE (Rye and Holland 1998; Rasmussen and Buick 1999; Farquhar et al. 2000; Pavlov and Kasting 2002), whereas, on the other hand, Schopf et al. (1983), Hofmann et al. (1999) and Schopf (2011) hypothesise the existence of oxygenic photosynthesis on the basis of microfossil morphologies similar to those of oxygenic cyanobacteria, stromatolite-like constructions resembling those produced by oxygenic photosynthetic microbial mats, and carbon isotope compositions consistent with fractionation by oxygenic photosynthesisers. These interpretations are contested, however: the microfossils are probably hydrothermal precipitations (Brasier et al. 2002), the stromatolite-like structures are most likely to have been produced by anoxygenic photosynthesisers (Allwood et al. 2006), and the isotopic fractionation range of oxygenic photosynthesisers overlaps with that of other microorganism groups, such as anoxygenic photosynthesisers (Reysenbach and Cady 2001). Nevertheless, another argument for oxygenic conditions based on possible evidence for transported uranium (uranium can only be transported in the oxidized state) in rocks > 3.7 Ga-old was put forward by Rosing and Frei (2004). Data from younger 2.5 to 2.9 Gyr old sediments based on (a) redox-sensitive metals (Anbar et al. 2007), (b) sedimentology and micropaleontology (Buick 1992; Kazmierczak and Altermann 2002) and (c) carbon isotopic measurements (Nisbet et al. 2007; Thomazo et al. 2011) also suggest the existence of oxygen and oxygenic photosynthesis, therefore strongly supporting model (ii). Finally, a recent modeling study, based on protein folding, proposes a molecular clock according to which oxygenic photosynthesis appeared 2.9 Gyr ago (Wang et al. 2011). The appearance of this metabolic process is a key step in the evolution of life towards large and complex organisms (Catling et al. 2005).

This study aims at investigating a new potential marker of oxygenic photosynthesis, by using a comparative approach between modern microbial mats and natural chert samples. Photosynthetic microbial mats are considered to be modern analogues of ancient stromatolites (Schopf et al. 1983). Samples were collected in the lake “La Salada de Chiprana” (Spain), the only permanent hypersaline ecosystem in Western Europe (Vidondo et al. 1993). Grown in a shallow-water environment, the complex microbial mats present in “La Salada de Chiprana” contain two types of bacteria using oxygenic and anoxygenic photosynthesis (Jonkers et al. 2003), respectively Microcoleus-like bacteria (MLB) (cyanobacteria) and Chloroflexus-like bacteria (CLB) (chloroflexacea). Comparison of mat samples enriched in each of these two bacteria was performed using Electron Paramagnetic Resonance (EPR) spectroscopy. This spectroscopy detects the electron spin of organic radicals in carbonaceous matter with high sensitivity (Uebersfeld and Erb 1956) and has therefore already been used to characterize the maturation of polyaromatic organic matter in coals (Retcofsky et al. 1968; Villey et al. 1981; Oberlin et al. 1978; Mrozowski 1988a, b; Rouzaud et al. 1991) in Precambrian cherts (Skrzypczak-Bonduelle et al. 2008) and in meteorites (Binet et al. 2002, 2004; Gourier et al. 2008). The radicals detected by EPR consist of more or less complex aromatic moieties (with sp2 carbons), where the unpaired electron spin occupies a π-type molecular orbital. Several parameters can be deduced from the EPR line based on the amplitude, the width and the resonance field of the signal. Those parameters were determined during analysis of thermally treated biomass mainly comprising oxygenic and anoxygenic photosynthetic bacteria and were compared with various Precambrian chert samples.

Material and Methods

Material

“La Salada de Chiprana”, located in the Ebro River basin, northern Spain (41°14′30″N, 0°10′50″W), is a hypersaline lake (78 ‰) in which the most abundant ions are Mg2+ and SO3 2− (Vidondo et al. 1993; Jonkers et al. 2003). With a total surface of 31 ha and a maximum depth of 5.6 m, the lake lies on the Upper Oligocene-Miocene Caspe Formation, which is mainly composed of sand and silt stones. The lake hosts extensive microbial mats, exhibiting a multilayer structure and containing MLB and CLB.

MLB is a mat-building cyanobacterium, common in marine and hypersaline environments. In laboratory experimental conditions, it has been shown capable of oxygenic photosynthetic activity at sulphide concentrations as high as 500 μmol.L−1 H2S, but it is also able to carry out anoxygenic photosynthesis using sulphide as an electron donor (Jørgensen et al. 1986). During the day, MLB has a high rate of CO2 fixation (Guerrero and De Wit 1992). In situ measurements of oxygen dynamics during daily cycles have shown that oxygen concentration increases directly after sunrise and rapidly reaches over five times air saturation (Jonkers et al. 2003; Ludwig et al. 2005).

CLB are multicellular anoxygenic, filamentous, green non-sulphur bacteria that can use anoxygenic photosynthesis to produce chemical energy in the form of the adenosine triphosphate (ATP) molecule (Pierson and Castenholz 1995). They are characterized by the production of bacteriochlorophyll a (BChla), and in some species, by the additional production of BChlc or BChld. Chloroflexus aurantiacus, one of the most studied species of the phylum, is a thermophilic species (Pierson and Castenholz 1974). However, the presence of a diverse number of Chloroflexus-like organisms is also reported in intertidal marine and submerged hypersaline microbial mats (Pierson et al. 1994). In the “La Salada de Chiprana” lake, 16 phylotypes of the phylum Chloroflexus, were identified (Bachar et al. 2007). These anoxygenic, facultative anaerobic organisms may have played an important role in the evolution of photosynthesis, given their characteristics close to both green sulphur bacteria and purple bacteria, as well as their ability to perform cellular respiration in aerobic environments (Bruce et al. 1982).

20 cm*10 cm*10 cm portions of the microbial mats were sampled in July and October 2010 and cultured in laboratory, in aerated natural lake water, at ambient temperature, and under Philips-HPI T Plus 400 W lamps. Routine microsensor profiling of oxygen concentrations carried out in microbial mats of the same sampling sites showed no significant change in oxygen metabolism during the maintenance period in these conditions (Ludwig et al. 2006). The microbial mats were dissected using magnifying glasses in order to separate layers dominated by MLB from those dominated by CLB. The microbial composition in the two subsamples was checked using optical microscopy (Fig. 1). Jonkers et al. (2003) reported the presence of smaller organisms, such as diatoms, in the mat that occur in both samples. The bacteria were rinsed twice with distilled water, frozen and then lyophilized. Note that no significant change occurred in the microbial composition of the mat between July and October 2010 and that the same dissection protocol was used.

Fig. 1
figure 1

Optical microscopy of an aliquot of (a) the first bacteria sample, revealing MLB colonies (b) the second bacteria sample, revealing the presence of CLB (small filaments), together with Diatoms (small ovals) and few MLB (large filaments)

Full size image

Nine natural cherts of various ages, metamorphic facies, and geographical origins were also studied for comparison. Data resulting from a previous study of 12 other natural chert samples were included in this study (Skrzypczak-Bonduelle et al. 2008). The sample characteristics are given in Table 1. All selected cherts had a metamorphic grade lower than the greenschist facies. These cherts are silicified sediments containing the carbonaceous remains of microorganisms (and macroorganisms in the most recent samples). The silica phase in the cherts is microcrystalline quartz.

Table 1 Description of the studied chert samples
Full size table

Methods

Kerogen isolation was performed using an HF/HCl acidic treatment (Durand and Nicaise 1980).

Optical microscopy of modern bacteria was performed using a Leica DLMB microscope in transmitted mode, equipped with a 205 Leica C Plan 506076 objective. 3 to 4 mg of lyophilized CLB and 5 to 10 mg of lyophilized MLB were placed in EPR tubes in quadruplicates. A 1:1 mixture of MLB and CLB was obtained by crushing 10.3 mg of lyophilized CLB and 10.3 mg of lyophilized MLB in a glass mortar. 4 to 6 mg of the mixture was then introduced into EPR tubes in duplicates. About 100 mg of intact chert rock were sampled and analyzed by EPR without further preparation.

All EPR tubes were sealed under vacuum in order to eliminate O2. The accelerated ageing consisted in thermal treatment in gradual steps from 20 °C to 720 °C. Each step consists in a 50 °C rise of temperature for 15mn (Skrzypczak-Bonduelle et al. 2008).

EPR analyses were performed at ambient temperature and at X-band (9.5 GHz) on a Bruker ELEXSYS E500 spectrometer equipped with a high sensitivity Bruker 4122SHQE/0111 microwave cavity. A 2 mW to 20 mW microwave power was used. The spectrometer was calibrated using the diphenylpicrylhydrazyl (DPPH) standard with a known g factor (g = 2,0037).

High-resolution transmission electron microscopy (HRTEM) is a relevant tool for obtaining information on the organization of carbonaceous matter (structure, nanostructure, texture) by imaging directly the profile of the polyaromatic layers (Oberlin 1989). Observations were carried out using a Jeol 2011 microscope operating at 200 keV. Image analysis was conducted after skeletonization as described by Rouzaud and Clinard (2002).

Raman spectra were obtained using a Renishaw InVIA Reflex microspectrometer equipped with a 514-nm Spectra Physics argon laser at 20 mW. The laser was focused on the sample using a DMLM Leica microscope with a 100 X objective. The laser power at the sample surface was set at around 1 mW. The signal was detected by a Peltier cooled RENCAM CCD detector. The spectrometer was calibrated using a silicon standard before each session. Kerogens isolated from cherts and carbonaceous matter from 720 °C thermally treated MLB and CLB were crushed in an agate mortar before deposition on a glass slide for analysis.

Results and Discussion

Optical Microscopy of Bacterial Samples

Observation of the separated bacteria using optical microscopy revealed that the first batch, which was sampled in the upper layer of the microbial mat, is chiefly composed of MLB colonies (Fig. 1a), whereas the second batch, sampled right under the first layer, is mainly composed of CLB, together with diatoms and few MLB filaments (Fig. 1b).

It is obvious that perfect separation of the two bacteria is not possible. However, the optical images show that each dissected layer was dominated by one of the bacteria species. The other organisms reported by Jonkers et al. (2003) contribute to each dissected layer and are therefore taken as part of the background signal.

EPR Study of the Thermally Treated Bacteria

Three main parameters (described on Fig. 2a) can be derived from EPR lines of radicals in the disordered carbonaceous matter within each sample. The peak-to-peak amplitude App, for a given peak-to-peak linewidth ΔBpp, reflects the number of spins in the sample, which is directly related to the total amount of radicals. When ΔBpp varies, the line intensity is calculated as I = App.ΔBpp 2. To avoid consideration of the mass of organic matter introduced in the tube, each thermal treatment series on a sample was normalized so that the maximum intensity attained was equal to 1. The g-factor, calculated from the resonance field of the signal, reflects the chemical composition of the radicals. Its deviation from the free spin value (g e  = 2.0023) increases with the O/C ratio (Retcofsky et al. 1968; Rouzaud et al. 1991). The peak-to-peak linewidth ΔBpp reflects the diversity of organic radicals (distribution of radicals), their local concentration (dipolar interaction),and the hydrogen content (unresolved hyperfine interaction) (Skrzypczak-Bonduelle et al. 2008).

Fig. 2
figure 2

a Definition of parameters measured on the EPR line of radicals in carbonaceous matter (b) EPR spectra of the thermally treated bacteria

Full size image

The evolution of the EPR spectra and their associated parameters during thermal treatment of the two types of bacteria is shown in Figs. 2b and 3 respectively. Only the 320 °C to 720 °C temperature range is represented for the thermal treatment since there is no significant EPR signal for samples treated at lower temperatures. Moreover, the EPR line shape evolution during thermal treatments is the same for both samples, from Gaussian to Lorentzian and then to a stretched-Lorentzian shape, and is therefore not reported here. Three domains (II, III and IV) of evolution of the organic matter in chert samples in the 320 °C to 720 °C temperature range were previously distinguished (Skrzypczak-Bonduelle et al. 2008). Note that Domain I is not mentioned in this paper, since it describes organic matter that underwent thermal treatment lower than 320 °C, for which, as mentioned above, no significant EPR signal was observed. Both types of bacteria show the same EPR behavior with increasing temperature, as described below.

Fig. 3
figure 3

Evolution of (a) the normalized intensity (b) the g-factor and (c) the peak-to-peak linewidth of the EPR signal of thermally treated MLB (mixed black line) and CLB (dashed blue line). Error bars reflect the dispersion of the values for the quadruplicate

Full size image

Domain II ranges from about 320 °C to about 450 °C, the temperature range in which there is a loss of carboxyl and hydroxyl groups. In this particular study, EPR spectra in Domain II show a slight increase in their intensity, an almost steady g-factor, and an almost constant linewidth indicating that no C-H hydrogen loss occurs in this temperature range.

Domain III is defined for thermal treatment from about 450 °C to about 620 °C, a temperature range in which hydrocarbon loss occurs. In this study, there is a drastic increase in the intensity, and a slight decrease in the EPR linewidth in Domain III (see insert in Fig. 3c), revealing an increase in radical concentration. A slight decrease in the g-factor is observable (Fig. 3b), pointing to the loss of some heteroelements, mainly oxygen. The linewidth decrease in this domain reveals a hydrogen loss (Skrzypczak-Bonduelle et al. 2008).

Domain IV, for thermal treatment higher than about 620 °C, is characterized by a sharp decrease in intensity, indicating that the recombination rate of radicals is faster than their creation rate, the concentration of which reached a maximum (Fritsch et al. 1974). The g-factor remains constant, pointing to no significant change in the chemical nature of the radicals. Surprisingly, there is significant broadening in linewidth, especially for the MLB sample. The linewidth at 720 °C for MLB and CLB is 1.9 ± 0.6 mT and 0.7 ± 0.1 mT respectively. Four hypotheses can be put forward to account for this selective broadening:

  1. (i)

    Unresolved hyperfine interaction between the radicals and the hydrogen nuclei (Retcofsky et al. 1975). However, as highlighted in the insert in Fig. 3c, the linewidth decreases at the beginning of Domain III. This is related to hydrogen loss and thus rules out any major hyperfine interaction between the radicals and hydrogen nuclei.

  2. (ii)

    The presence of molecular dioxygen O2, which is paramagnetic, and interacts with organic radicals, resulting in an EPR line broadening by dipolar interactions (Bates et al. 1995). This hypothesis appears unlikely owing to the fact that all EPR experiments were conducted under vacuum.

  3. (iii)

    The creation of a new type of radical due to specific organic precursors in the organic matter of MLB. This hypothesis appears unlikely, given that organic functions are removed from the carbonaceous matter during the first stages of carbonization. On the contrary, the EPR linewidth difference occurs only for the higher stages of carbonization,in which radicals recombine and are annihilated.

  4. (iv)

    Selective modification of radicals due to high temperature reactions. This reaction may be catalyzed by an element (e.g. a transition metal, …) associated with the organic matter of MLB only. This hypothesis appears the most likely and is developed below.

In the absence of hydrogen, which is the case during the highest stage of carbonization (domain IV), the EPR linewidth is driven by the distance R between unpaired electron spins, i.e. between radicals: the larger the distance, the smaller the linewidth, which varies as R −3 due to dipolar interactions between electron spins. Statistical theory predicts that the linewidth of a Lorentzian line (which is the case in the present study) in a magnetically diluted spin system is related to the spin concentration N by (Abragam 1961):

$$ \varDelta {B_{pp }}=\frac{{4{\pi^2}}}{9}g\beta N\approx 8.12\times {10^{-21 }}N $$
(1)

with N in cm−3 and ΔB pp in mT. In the following, we distinguish between the mean spin concentration N in the sample, which is directly measured by the EPR intensity, and the local spin concentration N loc experienced by each radical, which is measured by the linewidth. Figure 3a shows that the EPR intensity and thus the mean concentration N decreases above 600 °C. Assuming an homogeneous distribution of radicals in the 620 °C carbonized material and using Eq. 1, we may calculate the spin concentration at 620 °C N ≈ N loc  ≈ 0.5 × 1020 cm−3. At 720 °C, the EPR intensity decreases by a factor ≈ 6 for MLB, which means that N decreases by the same factor. Thus the mean spin concentration N in MLB at 720 °C is N ≈ 0.8 × 1019 cm−3. However, due to the linewidth broadening, when using Eq. 1, we find that N loc  ≈ 2.5 × 1020 in MLB.

This implies that the global spin concentration N decreases in MLB between 620 °C and 720 °C, while the local spin concentration N loc increases in the same temperature range. This points to the formation of shallow radical aggregates characterized by their local concentration N loc  > N (Eq. 1). Such a formation of radical aggregates is consistent with formation at a local scale, possibly related to a catalytic effect. To have a better understanding of the origin of these shallow aggregates of radicals, Raman spectroscopy and HRTEM analyses of the carbonized bacteria were run.

Raman Spectroscopy and HRTEM of the Thermally Treated Bacteria

Raman spectroscopy and HRTEM were performed on the 720 °C thermally treated bacteria for which the EPR linewidth difference is the largest. The Raman spectra of the heated microorganisms are very similar, as shown in Fig. 4, with two main bands at 1,344 and 1,570 cm−1. The spectra were deconvoluted with a four-band fitting (G, D1, D2 and D3) using a Voigt profile, as described in Beyssac et al. (2002). This allowed determination of the position, intensity, area and width for each band. Raman parameters were calculated (Table 2), comprising the D1 full-width at half maximum (cm−1), along with the R1 and R2 ratios, defined as: \( R1=\frac{Intensity(D1) }{Intensity(G) } \) and \( R2=\frac{Area(D1) }{{Area\left( {G+D1+D2+D3} \right)}} \) (Beyssac et al. 2003). The D1 full-width at half maximum is known to decrease, with increasing degree of organization in the carbonaceous matter, whereas the R1 and R2 ratios increase (Beyssac et al. 2003; Bernard et al. 2010). In the case of the 720 °C thermally treated bacteria, Raman parameters are similar (Table 2), thus indicating that the organic matter in the thermally treated MLB and CLB has the same structure.

Fig. 4
figure 4

Raman spectra of 720 °C thermally treated (a) CLB and (b) MLB bacteria

Full size image
Table 2 Raman parameters for 720 °C thermally treated Chloroflexus-like bacteria (CLB) and Microcoleus-like bacteria (MLB)
Full size table

HRTEM observations were run on the 720 °C thermally treated MLB and CLB (Fig. 5). Associated HRTEM parameters (Table 3) were calculated from skeletonization of the HRTEM images of the two samples (as illustrated in Fig. 5a, b), following the method of Rouzaud and Clinard (2002). These parameters are again very similar, confirming the similarity of the organic matter structure inferred from Raman observations. It can therefore be concluded that the degrees of carbonization of MLB and CLB are similar.

Fig. 5
figure 5

High Resolution TEM images and image skeletonization of 720 °C thermally treated (a) CLB and (b) MLB bacteria

Full size image
Table 3 HRTEM parameters for 720 °C thermally treated Chloroflexus-like bacteria (CLB) and Microcoleus-like bacteria (MLB)
Full size table

It is important to note that radicals detected by EPR constitute only a very small part (much less than 0.1 %) of the carbonaceous matter. On the contrary, Raman spectroscopy and HRTEM reflect the totality of the material and not the close environment of the radicals influenced by the specific element contained in MLB.

In conclusion, the structure of the organic matter does not drive the EPR linewidth of the thermally treated bacteria. This confirms that the specific broadening of the EPR line is only due to the aggregation of a small quantity of radicals, which are undetectable by Raman and HRTEM. This aggregation, which occurs only in MLB, must be due to the presence of some catalytic element present in MLB and not in CLB, and which is incorporated or adsorbed on carbonaceous particles. It must be stressed that the slight linewidth increase for the CLB sample (Fig. 3c) is inferred to be due to the presence of small amounts of MLB in the CLB sample, due to limitations in the separation method (Fig. 1b).

EPR Study of Natural Cherts

The variation of the EPR linewidth (ΔBpp) of 15 chert samples with ages ranging from 0.4 to 3.5 Gyr is reported in Fig. 6. A decrease of ΔBpp is observed from samples through the Phanerozoic and the Proterozoic. A similar trend was previously reported for coals (Retcofsky et al. 1975) and cherts (Skrzypczak-Bonduelle et al. 2008). As indicated above, this EPR line narrowing is due to progressive hydrogen loss upon ageing. For Archean samples, the ΔBpp encompasses a large range of values, from 0.05 mT to 0.35 mT. A similar dispersion was already reported for a smaller set of Archean cherts (Skrzypczak-Bonduelle et al. 2008). This is unlikely to result from the maturity of the sample, which is the result of age, temperature and pressure. This maturity was evaluated using the literature (Table 1) and Raman spectra of the cherts (Fig. 7, Table 4). For instance, the Buck Reef (6) and Middle Marker (7) Archean cherts have more or less the same age (3.42 and 3.72 Ga, respectively) and similar Raman spectra (Fig. 7), revealing similar metamorphic grades. However, those two samples show different linewidths (Fig. 6). Similarly, the nature of the sediment and its environment of formation do not seem to be the main cause of linewidth dispersion in Archean samples. For instance, the Middle Marker chert (7) (Lanier and Lowe 1982) and the Zwartkoppie chert (J) (Lowe and Knauth 1977) both formed in a shallow-water environment of deposition but nevertheless show different linewidths.

Fig. 6
figure 6

Peak-to-peak width of the EPR line versus age of the studied cherts. Results from this work (numbers 1 to 8, empty squares) and from Skrzypczak-Bonduelle et al. (2008) (letters A to L, plain diamonds). Numbers and letters refer to samples as identified in Table 1. Dotted line is a guide for the eye and does not represent any correlation

Full size image
Fig. 7
figure 7

Raman spectra of the carbonaceous matter in the studied cherts identified in Table 1. Only the Rhynie sample from Skrzypczak-Bonduelle et al. (2008) was available for analysis

Full size image
Table 4 Raman parameters for carbonaceous matter embedded in cherts
Full size table

If not chiefly related to the metamorphic grade, the age of the sample, or the geological setting, the EPR linewidth dispersion has to be related to the nature of the carbonaceous precursor. The g factor evolution shows that heteroelements are lost during the first stage of carbonization. The influence of organic functions is thus unlikely and the EPR linewidth dispersion reflects the presence of some aggregates of radicals similar to those observed for MLB. Based on the aforementioned results on artificially aged microbial samples, one might assume that a specific metal element plays a catalytic role in the radical aggregation. This element could be, for example, a transition metal ion originating from a specific metalloenzyme of the metabolism of the precursor bacteria. Similar linewidth broadenings were reported by Mrozowski (1988b) during EPR study of thermally treated biological materials (higher plants), which are complex oxygenic photosynthesizers, the metabolism of which is based on metalloenzymatic activity. In a similar fashion, the EPR linewidth broadening in the cherts might be a marker of the metabolism of the micro-organisms, the remains of which are embedded in the chert.

It is important to note that if linewidth broadening in cherts may be a marker of metabolism, the absence of linewidth broadening does not imply the absence of this metabolism. Namely, transition metals are known to migrate in the mineral matrix during time and thermal events, as documented for vanadium (Gourier et al. 2010). However, the higher the organic matter concentration, the slower the migration of the transition metal (Gourier et al. 2010). Those observations imply that the aforementioned proposed marker is only applicable to chert samples having zones of concentrated organic matter. This preliminary study highlights the possibility of EPR as a potential tool to access information on the metabolism of past micro-organisms, for example oxygenic photosynthesis.

Conclusion

The EPR study of thermally treated oxygenic and anoxygenic photosynthetic bacteria, MLB and CLB respectively, showed selective broadening of the EPR signal of the MLB bacteria, for thermal treatments at temperatures higher than 600 °C. This difference was attributed to the presence of catalytic activity due to an element associated with MLB, creating radicals aggregates. However, despite Raman spectroscopy analysis and HRTEM observations, both showing the formation of similar carbonaceous materials, the exact nature of this reaction could not be determined. Similar broadening of EPR linewidth was also found in some natural carbonaceous Archean cherts, without any relationship to sample age, location or metamorphic grade, therefore suggesting that it might be a metabolic signature. Further analysis is needed in order to understand the mechanisms controlling this increasing linewidth in the EPR signal.