Nuclear Magnetic Biosignatures in the Carbonaceous Matter of Ancient Cherts: Comparison with Carbonaceous Meteorites

Abstract

The search for organic biosignatures is motivated by the hope of understanding the conditions of emergence of life on Earth and the perspective of finding traces of extinct life in martian sediments. Paramagnetic radicals, which exist naturally in amorphous carbonaceous matter fossilized in Precambrian cherts, were used as local structural probes and studied by electron paramagnetic resonance (EPR) spectroscopy. The nuclear magnetic resonance transitions of elements inside and around these radicals were detected by monitoring the nuclear modulations of electron spin echo in pulsed EPR. We found that the carbonaceous matter of fossilized microorganisms with age up to 3.5 billion years gives specific nuclear magnetic signatures of hydrogen (¹H), carbon (¹³C), and phosphorus (³¹P) nuclei. We observed that these potential biosignatures of extinct life are found neither in the carbonaceous matter of carbonaceous meteorites (4.56 billion years), the most ancient objects of the Solar System, nor in any carbonaceous matter resulting from carbonization of organic and bioorganic precursors. These results indicate that these nuclear signatures are sensitive to thermal episodes and can be used for Archean cherts with metamorphism not higher than the greenschist facies. Key Words: Kerogen—Biosignatures—Origin of life—Archean—EPR spectroscopy. Astrobiology 13, 932–947.

1. Introduction

O ne of the primary goals in astrobiology is the identification of biosignatures of primitive life in ancient rocks. The most ancient organic traces of life on Earth, about 3.5 billion years old (Archean), are preserved in the form of carbonaceous microstructures in silicified sediments (cherts). The origin of these microstructures, biogenic or abiotic, has been vehemently debated in the past decades (Schopf, 1993; Mojzsis et al., 1996; Brasier et al., 2002; Garcia-Ruiz et al., 2003; Derenne et al., 2008; Schopf and Kudryavtsev, 2012), and undisputable molecular biosignatures are necessary to assess the biogenicity of the carbonaceous matter (CM) in Archean cherts. However, these materials have often been submitted to high temperature and pressure during diagenesis and metamorphic episodes, so that molecular biosignatures may have suffered from severe alteration. The loss of functional groups, heteroatoms, and hydrogen occurring during aromatization of the CM may ultimately lead to crystalline graphite (Buseck and Huang, 1985; Wopenka and Pasteris, 1993; Bustin et al., 1995; Beyssac et al., 2002, 2003; Bernard et al., 2010), with the total loss of memory of the biological origin of this carbon material. This typically occurs for metamorphic grade in the amphibolite facies or higher (van Zuilen et al., 2012), implying that molecular biosignatures can only be sought in formations with lower metamorphic grades.

Even in a chert containing biogenic CM, a part of this material may originate from various contamination processes such as hydrothermalism, endolithic microorganisms, weathering, and anthropogenic contamination (Hoering, 1967; Roedder, 1981; Campbell, 1982; Westall and Folk, 2003). Consequently, the detection method must be able to distinguish primitive biosignatures with the same age as the host rock from more recent biosignatures due to contamination.

In addition, the presence of carbonaceous microstructures in a sedimentary rock is not sufficient to assess that primitive life was present at the epoch of the sediment deposit, because the CM may also have an abiotic origin (Luque et al., 1998; Garcia-Ruiz et al., 2003). As far as very ancient rocks are concerned, a part of this contamination may be of extraterrestrial origin (meteorites, interplanetary dust particles, comets). For example, carbonaceous meteorites, among the most primitive objects of the Solar System (4.5 Gyr), contain a significant amount of CM with chemical composition similar to that of terrestrial kerogens (Hayatsu et al., 1977). Remarkably, this very ancient extraterrestrial CM contains organic molecules of biological importance at the part-per-million level (Cronin et al., 1988). During the early history of the Solar System, Earth and Mars were intensively bombarded by meteorites, micrometeorites, and comets, so that a considerable amount of CM has probably been deposited on these planets (Chyba and Sagan, 1992; Maurette et al., 2006), with the possibility that a fraction was fossilized in sedimentary deposits. This eventuality is of particular concern for the search for traces of organic carbon in the oldest sedimentary rocks on Mars.

The in situ detection of molecular biosignatures in a composite material composed of a silicified rock with embedded amorphous and extremely disordered CM is a real challenge for spectroscopic analysis. For example, the well-known Raman spectroscopy, which gives a typical signature of CM (see, e.g., Quirico et al., 2009; Bower et al., 2013), is not always sufficient to make a clear distinction between biogenic and abiotic CM (Pasteris and Wopenka, 2003; Marshall et al., 2010; Bower et al., 2013). More generally, spectroscopic methods based on electromagnetic radiation in the UV-visible range are limited to the sample surface or subsurface because of the strong optical absorption of most rocks. Spectroscopic methods that work at long wavelengths, such as nuclear magnetic resonance (NMR), are more adapted to a selective detection of molecular biosignatures. In addition to the high resolution of this technique, the weak interaction (dipolar magnetic) of the radiofrequency field with matter allows a deep penetration of the radiation in the sample. Unfortunately, NMR suffers from a low sensitivity and for this reason is not adapted to rocks containing small amounts of CM, sometimes at the part-per-million level. Moreover, NMR spectra are often perturbed by paramagnetic species (transition metal impurities, radicals, etc.) that are always present in natural materials.

Electron paramagnetic resonance (EPR) spectroscopy offers a good compromise between sensitivity (much better than NMR) and resolution (lower than NMR). Its frequency domain in the microwave range allows a deep penetration of the electromagnetic field probe into the sample, so that EPR spectra correspond to the bulk of the material and are much less sensitive to surface alteration and contamination (Binet et al., 2008). All carbonaceous materials, natural or synthetic, contain paramagnetic defects (hereafter referred to as radicals), which can be detected with high sensitivity by EPR spectroscopy. These defects can be described as molecular fragments with unpaired electrons responsible for the EPR response. This is the case for synthetic amorphous carbon (Robertson, 2002), coals (Uebersfeld et al., 1954; Retcofsky et al., 1975; Mrozowski, 1988a, 1988b; Whelan and Thompson-Rizer, 1993), carbonaceous shales (Rigali and Nagy, 1997), cherts (Skrzypczak-Bonduelle et al., 2008; Bourbin et al., 2012), and carbonaceous meteorites (Binet et al., 2002, 2004a; Gourier et al., 2008; Delpoux et al., 2011). However, compared to NMR, standard cw-EPR (continuous wave EPR) suffers from a rather poor resolution about 10³ times less than NMR, so that radicals in all kinds of carbonaceous materials always give the same type of EPR signal, mostly in the form of a more or less narrow and symmetrical line characterized by a g-factor (related to the magnetic field value of the line) close to the free electron spin value 2.0023 (Fig. 1) (Uebersfeld et al., 1954; Retcofsky et al., 1975; Mrozowski, 1988a; Whelan and Thompson-Rizer, 1993; Binet et al., 2002; Skrzypczak-Bonduelle et al., 2008). The width of this signal is due to the hyperfine (hf) interactions of the unpaired electron spin of the radical with all neighboring nuclei, giving hundreds of overlapping lines that merge into a single EPR line without apparent structure. This unresolved hf structure, which is expected to contain useful information, can be resolved with pulsed-EPR techniques, whereby nuclear magnetic resonances of various nuclei with nonzero nuclear spin (¹H, ¹³C, ¹⁴N, ³¹P, ²⁹Si,…) located in the vicinity of the unpaired electron can be indirectly detected by recording the nuclear modulation of the electron spin echo (Höfer et al., 1986; Schweiger and Jeschke, 2001). With these pulsed EPR methods, the unpaired electron spins of radicals play the role of internal probes, which transfer the microwave excitation to neighboring nuclear spins at the nanometer scale. The resulting nuclear magnetic excitations are transferred back electron spins and read in the form of modulations of the electron spin echo at nuclear frequencies.

FIG. 1.

cw-EPR spectra at room temperature of Gunflint (1.9 Gyr) and Warrawoona (3.5 Gyr) cherts, CM extracted from the Orgueil meteorite, and molecules encapsulated in silica and treated at 550°C.

In this work, we used the bi-dimensional version of this electron spin echo technique, known as hyperfine sublevel correlation (HYSCORE) spectroscopy (Höfer et al., 1986), to investigate the nuclear transitions of ¹H, ¹³C, and ³¹P in the vicinity of organic radicals in cherts with different degrees of evolution that contain fossil CM of biological origin. The magnetic resonance signals were compared with those of the abiotic CM in carbonaceous meteorites. The influence of metamorphism was experimentally studied by monitoring the effect of carbonization on biological molecules encapsulated in silica.

2. Materials and Methods

Three kinds of materials were studied in this work: lightly metamorphosed cherts containing microorganisms fossilized in the form of carbonaceous microstructures (referred to as F-type), CM extracted from two carbonaceous meteorites (referred to as M-type), and CM originating from carbonization of bioorganic molecules encapsulated in silica (referred to as C-type).

2.1. Cherts and meteorites

Three chert samples (silicified sediments) containing carbonaceous remains of microorganisms (F-type) were provided by the Museum National d'Histoire Naturelle (Paris, France). A 3.49-billion-year-old (Archean) chert (PPRG 006 from the Precambrian Paleobiology Research Group Collection, courtesy of W. Schopf) was collected in the Dresser Formation of the Warrawoona group, Pilbara Craton, Western Australia. It hosts among the oldest putative microfossils (Schopf, 1993). A 1.88-billion-year-old (Paleoproterozoic) chert, referenced as GF 74-1-11 from the Awramik collection, comes from the Schreiber Beach locality of the Gunflint Formation (Ontario, Canada) and contains well-preserved prokaryotic microorganisms (Barghoorn and Tyler, 1965). These two cherts are metamorphosed in the greenschist facies. A non-metamorphized 45-million-year-old chert (Eocene) referenced as PPRG 456 from the Schopf collection comes from the Clarno Formation, John Day basin in Oregon (USA), and contains well-preserved remains of vascular plants.

Carbonaceous matter of meteorites (M-type) was extracted from the mineral host of two meteorites by the usual HF/HCl treatment (Durand and Nicaise, 1980), which does not affect the structure and the concentration of radicals (Binet et al., 2002). Samples of Orgueil (France, fallen in 1864) and Tagish Lake (Canada, fallen in 2000) meteorites were provided by the Museum National d'Histoire Naturelle (Paris, France) and the NASA Johnson Space Center (Houston, USA) with the permission of J. Brook (the finder), respectively. The bulk elemental compositions of the insoluble residues of the HF/HCl treatments (the amorphous CM) were determined by elemental analysis at Wolf Laboratories (Rueil-Malmaison, France) for Orgueil and by Atlantic Microlab, Inc. (Norcross, Georgia, USA) for Tagish Lake. The resulting compositions of the CM are C₁₀₀H₇₂O₁₈N₃S₂ for Orgueil and C₁₀₀H₄₆O₁₅N₁₀ for Tagish Lake (Binet et al., 2002, 2004b). The samples still contain inorganic residues that amount to 8.9% and 7.2%, respectively.

2.2. Encapsulation of organic and bioorganic molecules

C-type CM (carbonization of organic molecules) was prepared to study the effect of thermal metamorphism on biological material. Several types of molecules and biomolecules were encapsulated in silica by a three-step sol-gel method (Livage et al., 2001) in order to reproduce the chemical environment in a chert: glucose, adenosine triphosphate (ATP), deoxyribonucleic acids (DNA), and a protein (bovine serum albumin, BSA). The first step of the synthesis is the preparation of aqueous solutions of about 10 mg·mL⁻¹ of biomolecules from Sigma-Aldrich. Two kinds of macrobiomolecules were used: DNA from salmon sperm and BSA in the form of lyophilized powders. For DNA, it was necessary to add a few milliliters of a 0.1 mol·L⁻¹ NaOH solution to the solution of DNA to obtain a neutral pH, which does not hinder the gel formation. The other molecules are anhydrous α-D-glucose (powder) and a solution of the disodium salt of adenosine-5′-triphosphate (ATP). To allow for the gel formation, a 100 mM aqueous solution was adjusted to pH 7 with a Trizma base buffer. The second step consisted of the preparation of the 0.75 mL·L⁻¹ sodium silicate solution with 3.3 g of Na₂Si₃O₇ (27% SiO₂,10% NaOH) in 20 mL of H₂O. Then 700 μL of HCl 3 mol·L⁻¹ were poured under magnetic stirring in 5 mL of solution of silicate of sodium to reach pH ∼7. Finally, 5 mL of solution containing the solvated biomolecule was poured at once. The gel appeared in a couple of minutes. The last step was the sample dehydration by freeze-drying. The gels were immersed into liquid nitrogen in order to freeze them very quickly and solidify the water contained in these samples. They were then subjected to a pressure of several millibar to sublimate the water, for approximately 50–60 h. This process allowed a progressive dehydration of the samples without altering their structure, giving a final product in the form of white powders.

2.3. EPR experiments

All the experiments were performed at X-band (9.5 GHz). Continuous wave–EPR spectra were recorded at room temperature with a Bruker ELEXSYS E500 spectrometer equipped with a high-sensitivity 4122SHQE/0111 microwave cavity. Pulsed-EPR experiments were performed at temperatures in the range from 4 to 30 K with a Bruker ELEXSYS E580 spectrometer equipped with a helium flow cryostat. The HYSCORE method was used for the detection of ¹H, ¹³C, ³¹P, and ²⁹Si nuclei. A spin echo is generated by the pulse sequence π/2-τ-π/2-t ₁-π-t ₂-π/2-τ-echo at the time τ after the last π/2 pulse; π/2 and π represent the flip angles of the electron magnetization. The echo intensity is measured by varying stepwise the times t ₁ and t ₂ at constant τ. The pulse lengths were 16 and 32 ns for π/2 and π pulses, respectively. For both t ₁ and t ₂ domains, 256×256 data points were measured. A second-order polynomial background subtraction was used to remove the unmodulated part of the echo. The magnitude spectrum was obtained after a 2-D Fourier transform of the spectra by using a Hamming apodization window function. The final 2-D spectrum represents the correlations between nuclear frequencies corresponding to the two spin states (up and down) of the electron and gives the nuclear frequencies and the hyperfine interactions of nuclei in the vicinity of the unpaired electron spin (Höfer et al., 1986). We always checked that the HYSCORE spectra correspond to the probed cw-EPR lines by recording three-pulse echo-detected EPR spectra (π/2-τ-π/2-T-π/2-τ-echo), which represent the echo intensity versus the magnetic field strength.

The samples were prepared by introducing several hundreds of milligrams of material in 4 mm quartz tubes. In the case of bioorganic molecules encapsulated in silica (C-type), the tubes were then maintained in a vacuum (10⁻² bar) for at least 4 h and sealed to avoid oxidation of powders during thermal treatments and the effect of the paramagnetism of molecular oxygen on the EPR signal of radicals. This encapsulated organic matter was carbonized by step heating treatments of the sealed tubes during short times (15 min) with temperature increasing by 50°C at each step. The cw-EPR spectrum was recorded at room temperature after each thermal step. The thermal step at 550°C gives an EPR line with almost the same lineshape (Lorentzian) and the same linewidth as the 1.9-billion-year-old Gunflint chert (Skrzypczak-Bonduelle et al., 2008).

3. Results

3.1. cw-EPR of carbonaceous matter

The three types of CM (F, M, and C) give a cw-EPR signal at 9.4 GHz consisting of a single narrow line representing the first derivative of the microwave absorption (Fig. 1). This signal is due to polyaromatic fragments containing unpaired electrons with spin S=1/2 delocalized in 2p_π orbitals of the sp² carbon backbone (Binet et al., 2002; Robertson, 2002). Such single EPR lines are ubiquitous in carbonaceous materials, including nanographene and nanographite (Joly et al., 2010; Barbon and Brustolon, 2012). The magnetic moments may originate from defects or from localized states in the edge of the polyaromatic fragments (Enoki and Kobayashi, 2005). The CM of coals, meteorites, and non-metamorphosed (or slightly metamorphosed) cherts can be viewed as a network of polyaromatic (sp² carbons) fragments linked together by more or less complex diamagnetic (S=0) aliphatic chains (sp³ carbons). Only a very small fraction of the polyaromatic units contain unpaired electron spins and are thus detected by EPR.

In addition to the narrow EPR line of CM, cherts and meteorite samples studied in this work contain other signals that have been reported in previous papers (Binet et al., 2002; Skrzypczak-Bonduelle et al., 2008). Meteoritic materials exhibit a broad ferromagnetic resonance signal due to magnetite and other residual magnetic minerals (Binet et al., 2002). However, this ferromagnetic resonance signal is so broad that it only contributes to a slight deviation of the baseline of the narrow CM line. The three chert samples exhibit other paramagnetic signals close to the CM line. In addition to a weak Fe³⁺ signal at g≈4.2, we always observed E′-center of quartz (an electron trapped at a dangling sp³ orbital of \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$${ \rm SiO}_3^{3 - }$$ \end{document} ), and the [AlO₄/h]⁰ center, which is an O⁻ ion adjacent to a substitutional Al³⁺ in Si site. In addition to these three paramagnetic defects, the 3.49-billion-year-old Warrawoona chert shows a very weak signal of vanadyl VO²⁺ ions, which is an inorganic biosignature (Gourier et al., 2010). Also, the 45-million-year-old Clarno chert exhibits four weak, very narrow lines that are unambiguously attributed to isolated CH₃° radicals (Skrzypczak-Bonduelle et al., 2008). All these defects are magnetically diluted in the mineral matrix and are slowly relaxing (except perhaps Fe³⁺), so that their presence does not affect the EPR line of CM. This was checked by studying different fragments of each chert sample, as the width and shape of the CM line did not depend on the intensity of Fe³⁺ and other defect signals, which are sample-dependent. The iron(III) content is not known and varies from fragment to fragment in the same sample.

The two EPR parameters of importance in single line spectra are the peak-to-peak linewidth ΔB _pp and the g-factor, defined as g=hν/βB _res, where h is Planck's constant, ν the microwave frequency, β the electron Bohr magneton, and B _res the magnetic field at the center of the resonance line. These parameters are gathered in Table 1 for the samples studied in this work. The parameters for the molecules encapsulated in silica are given in Table 1 for the thermal step 550°C, a temperature for which the EPR lines are very similar to that of Gunflint chert. In addition to ΔB _pp and g, the EPR lineshape of radicals is an interesting proxy that depends on the age of the CM and thus provides a test for syngenicity of the organic matter in Archean cherts (Bourbin et al., 2013). It may be argued that the slight distortion of some EPR lines in Fig. 1 could reflect a Dysonian profile (Barbon and Brustolon, 2012) resulting from electronic conductivity. A Dysonian line being homogeneous, however, it cannot give spin echoes in contrast with the experimental results presented below.

Table 1.

Origin and cw-EPR Parameters of the Samples

Carbonaceous matter	Origin	Sample	g-factor	Linewidth (mT)	Reference
Fossils (type F)	Cherts (microcrystalline SiO₂)	Warrawoona (3.5 Gyr)	2.0033±1.5×10⁻⁴	0.12±0.04
		Gunflint (1.9 Gyr)	2.0034±1.5×10⁻⁴	0.232±0.04	Skrzypczak-Bonduelle et al., 2008
		Clarno (45 Myr)	2.0040±4×10⁻⁴	0.72±0.04
Meteoritic (type M)	Carbonaceous matter extracted by the HF/HCl method	Orgueil	2.0031±3×10⁻⁴	0.38±0.02
		Tagish Lake	2.0032±1.5×10⁻⁴	0.56±0.02	Binet et al., 2004b
Carbonization of molecular precursors (type C)	Molecules encapsulated in sol-gel silica	Protein (BSA)	2.0034±1.5×10⁻⁴	0.30±0.02
		DNA	2.0038±1.5×10⁻⁴	0.35±0.02	This work
		ATP	2.0027±1.5×10⁻⁴	0.36±0.02
		Glucose	2.0033±1.5×10⁻⁴	0.26±0.02

The linewidth ΔB _pp reflects the broadening of the EPR line by unresolved hf interaction with hydrogen (¹H) nuclei and by the dipolar interaction between radicals (Skrzypczak-Bonduelle et al., 2008; Bourbin et al., 2012). This can be seen by artificial aging, in which the evolution of ΔB _pp is monitored during isochronal thermal treatments. Figure 2a shows the results for the different precursors (BSA, DNA, ATP, glucose) encapsulated in SiO₂ and the values measured for the 45-million-year-old Clarno chert treated under the same thermal conditions (interrupted line). The linewidth decrease is due to the loss of hydrogen atoms that occurred during the polymerization of the CM (Skrzypczak-Bonduelle et al., 2008). It appears that the carbonization of molecules encapsulated in SiO₂ follows the same evolution of ΔB _pp as a young chert containing non-mature CM (Clarno in Fig. 2a). The linewidths measured for Gunflint (0.23 mT, G in Fig. 2a) and Warrawoona (0.12 mT, W in Fig. 2a) are the same as those for Clarno and organic molecules heat treated at ∼550°C and ∼700°C, respectively.

FIG. 2.

Variation of EPR parameters with thermal treatment temperature for molecules encapsulated in silica. The discontinuous line represents the evolution for the Clarno chert (45 Myr). (a) Peak-to-peak linewidth ΔB _pp. The two arrows indicate the values for Warrawoona (W) and Gunflint (G) cherts. (b) g-factor.

Differences between thermally treated cherts and carbonized molecules encapsulated in SiO₂ are more pronounced in the evolution of g-factor (Fig. 2b). The deviation of g from the free-spin g-value g_e =2.0023 reflects the presence of heteroelements heavier than carbon (mainly oxygen) in the vicinity of the electron spin. The decrease of the g-factor during carbonization is due to oxygen loss in the form of H₂O and CO₂. It appears that the thermal evolution found for Clarno is close to that of BSA and ATP but very different from that of glucose and DNA, except at the end of the thermal treatment. This suggests that proteins may have dominated the biomolecular precursors at the origin of the CM fossilized in Clarno. Arrows in Fig. 2b indicate the g-factors for Gunflint (G) and Warrawoona (W) cherts. Interestingly, these values are close to those measured for the CM originating from glucose precursor, which may indicate that carbohydrates could be the main precursors at the origin of the CM of Precambrian cherts.

This analysis of the evolution of resonance lines by cw-EPR can be also extended to the lineshape, which evolves from a Lorentzian line toward a stretched Lorentzian line upon natural and artificial aging (Skrzypczak-Bonduelle et al., 2008; Bourbin et al., 2013). This joined evolution of the g-factor, ΔB _pp, and the lineshape upon natural and accelerated aging clearly indicates that the EPR of radicals in fossil CM would be strongly influenced by any thermal event that occurs during metamorphic episodes subsequent to the formation of the sediment. However, a detailed study of the EPR lineshape (Bourbin et al., 2013) shows that the evolution of radicals in fossil CM is controlled by metamorphism for cherts younger than 2 Gyr, while it is controlled by the age for cherts older than 2 Gyr with metamorphic grade not higher than the greenschist facies.

The following part of this paper focuses on the analysis of the unresolved hf structure of the EPR line, which contributes to the EPR linewidth. It is partly due to the interaction of the radical electron spin with various nuclei (¹H, ¹³C and ³¹P, ²⁹Si,…) inside or around these radicals. The hf interactions being controlled by the chemical bonding and the structure of radicals in the CM, these interactions should be much more sensitive to the origin of the molecular precursors than the g-factor, the linewidth, and the lineshape.

3.2. Pulsed EPR of cherts and meteorites

Nuclear modulation effects of spin echo amplitude in pulsed EPR spectroscopy give a direct measurement of the hf interactions between the unpaired electron spins of the aromatic radicals and various nuclear spins inside or in the vicinity of the radicals. Typical HYSCORE spectra are shown in Fig. 3 for Clarno (45 Myr), Gunflint (1.9 Gyr), and Orgueil and Tagish Lake meteorites, plotted as the intensity versus radio frequencies ν ₁ and ν ₂ (in MHz). These spectra are in the range of nuclear frequencies, as usually measured in NMR spectroscopy. For the sake of comparison, all the spectra of Fig. 3 are recorded with the same interval τ=136 ns between the two first π/2 pulses. This condition is a good compromise to detect the hyperfine spectrum of both ¹H and ¹³C in the same spectrum. In principle, the shape of the ¹H-hyperfine spectrum depends on τ by the so-called “blind spot” effect, and low τ are generally preferred to avoid lineshape distortion (Schweiger and Jeschke, 2001). However, we did not observe significant blind-spot effects by varying τ, and as shown below, the most important information was obtained by using high τ values. Despite the fact that the spectrum of the Tagish Lake meteorite suffers from a low signal-to-noise ratio, it is clearly of the same type as that of the Orgueil meteorite. It appears that all the HYSCORE spectra are similar to those found for coals (Ikoma et al., 1998, 2002); however, comparison of spectra in Fig. 3 shows that the biogenic CM of cherts (type F) is different from that of meteorites (type M), despite the fact that the latter has the same chemical composition as a terrestrial coal (Hayatsu et al., 1977; Sephton et al., 1999). The characteristics of HYSCORE spectra in the different kinds of materials are summarized in Table 2 and are detailed below.

FIG. 3.

HYSCORE spectra at 9 K and τ=136 ns of two cherts from the Gunflint (1.9 Gyr) and Clarno (45 Myr) formations (F-type carbon) and for the CM of Orgueil and Tagish Lake meteorites (M-type carbon). Spectra on the right represent the contour map of HYSCORE (projection in the frequency plane). Color images available online at www.liebertonline.com/ast

Table 2.

Nuclear Magnetic Signatures of the Carbonaceous Matter

Carbonaceous matter	Sample	¹ H	¹³ C	²⁹ Si	³¹ P double frequency
Fossils (type F)	Warrawoona (3.5 Gyr)	No signal^*	Weak	Narrow and weak	Present
	Gunflint (1.9 Gyr)	“Horn shape”	Weak	Narrow and intense	Present
	Clarno (45 Myr)	“Horn shape”	Weak	Narrow and intense	Present
Meteoritic (type M)	Orgueil	Linear ridge	Broad and intense	Broad and weak	Absent
	Tagish Lake	Linear ridge	Broad and intense	Broad and very weak	Absent
Carbonization of molecular precursors (type C)	Protein (BSA)	Linear ridge	Broad and intense	Broad and weak	Absent
	DNA	Linear ridge	Broad and intense	Broad and weak	Absent
	ATP	Linear ridge	Broad and intense	Broad and weak	Absent
	Glucose	Linear ridge	Broad and intense	Broad and weak	Absent

The potential biosignatures are represented in bold type.

The lack of ¹H signal in the Warrawoona chert is due to the total loss of hydrogen in this highly mature carbonaceous matter (Skrzypczak-Bonduelle et al., 2008).

3.2.1. ¹H hyperfine interactions

The shape of the ¹H-HYSCORE signal reflects the nature of the C-H bonds in radicals of the CM. These radicals are composed of a polyaromatic core terminated by hydrogen atoms or more or less branched aliphatic chains. Radicals in the different types of CM differ by the shape and the size of the aromatic part and by its branching level. Three types of hydrogen atoms can be considered as schematized in Fig. 4. The unpaired electron is delocalized on carbon 2p_π orbitals of the aromatic fragment and interacts with nuclear spins by two different hf interactions: the isotropic interaction A _iso, which is proportional to the electron spin density at the nucleus, and the dipolar interaction T, which depends on distance R between electron and nuclear spins as T α R ⁻³. The first type of hydrogen atoms to be considered belongs to the aliphatic chains. As these ¹H nuclei are the most distant from the electron spin (they are noted H_d in Fig. 4), they interact with the electron spin only by a weak dipolar interaction that gives the central peak along the diagonal ν ₁=ν ₂ at the nuclear frequency ν _n=14.5 MHz of ¹H (matrix line). The second type of hydrogen atoms is localized in benzylic positions (noted H_b in Fig. 4) and characterized by a hf interaction dominated by the isotropic coupling term A _iso, with only a weak dipolar contribution T. They appear in the HYSCORE spectrum as pairs of spots along an axis ν ₁=−ν ₂ crossing the diagonal ν ₁=ν ₂, with a spitting equal to A _iso. In the case of a distribution of hf couplings for H_b atoms, these spots overlap and give a spectrum in the form of a linear ridge along the antidiagonal ν ₁=−ν ₂ (Gourier et al., 2008). The third type of hydrogen atoms is localized on the edges of the aromatic rings (noted H_a). These aromatic hydrogen atoms are characterized by a significant dipolar interaction T in addition to the isotropic interaction A _iso. In a disordered CM, this gives peaks and ridges that are positively shifted from the antidiagonal and exhibit a “horn shape.” The splitting between the extrema of the ridge is equal to A _iso+2T, and the shift from the antidiagonal is 9T ²/32ν _n (Pöppl and Kevan, 1996; Schweiger and Jeschke, 2001) (Fig. 4).

FIG. 4.

Schematic representation of a hypothetical branched polyaromatic fragment in primitive CM. The electron is delocalized on the 2p_π orbitals of aromatic rings. Distant hydrogen (H_d), benzylic hydrogen (H_b), and aromatic hydrogen (H_a) represent the three types of hydrogen atoms that can be identified in HYSCORE spectra. Schematic HYSCORE spectra for H_d, H_b, and H_a hydrogen atoms are represented with the indication of the hyperfine parameters. A _iso and T are the isotropic and dipolar components of the hyperfine interaction, respectively. Color images available online at www.liebertonline.com/ast

It thus appears that the shape of the ¹H-HYSCORE spectrum should mainly depend on the branching level of aromatic radicals. A poorly branched radical is dominated by aromatic (H_a) hydrogen atoms, which gives the “horn shape” spectrum observed for the biogenic CM (F-type) fossilized in cherts (Fig. 3). Alternatively, H_a atoms are much less abundant in highly branched radicals corresponding to meteoritic CM (Gardinier et al., 2000; Cody and Alexander, 2005), which are dominated by benzylic (H_b) hydrogen atoms as shown by the linear ¹H ridge observed in meteoritic (type M) CM (Gourier et al., 2008).

More details on the hyperfine coupling can be deduced from the analysis of ¹H ridges, although the lack of well-resolved hyperfine peaks in the spectra indicates a distribution of hf parameters due to the strongly disordered character of the CM. Figure 5 shows the ¹H-HYSCORE spectra of Gunflint chert recorded at τ=128 ns and τ=348 ns. One type of hydrogen dominates the spectrum at τ=128 ns, while a more complex structure appears at τ=348 ns, revealing the presence of several types of hydrogen atoms. At both τ values, the dominant hydrogen atoms are characterized by A _iso≈−3 MHz and T≈+7 MHz. The significant contribution of the dipolar interaction and the negative value of the isotropic coupling indicate that these dominant hydrogen atoms are bonded to aromatic carbon atoms (H_a hydrogen in Fig. 4). The same features are observed for Clarno (45 Myr). The spectrum for Warrawoona (3.5 Gyr) exhibits no ¹H signal, which is in agreement with the high maturity of its CM. Indeed, the CM has lost its hydrogen as shown also by its very narrow cw-EPR linewidth (Fig. 1 and Table 1). On the contrary, the lack of vertical shift in the proton ridge of meteoritic CM (Fig. 3) indicates a distribution of A _iso values with very small dipolar contributions, pointing to dominant hydrogen atoms in benzylic position (H_b type). This shows that aromatic radicals in meteoritic CM are more branched (less aromatic hydrogen) than in cherts, as previously inferred from ¹³C-NMR spectroscopy (Gardinier et al., 2000; Cody and Alexander, 2005).

FIG. 5.

HYSCORE spectra at 4 K of Gunflint chert in the ¹H frequency range, recorded at two τ values. The vertical frequency shift is represented by an arrow. Color images available online at www.liebertonline.com/ast

3.2.2. ¹³C hyperfine interaction

¹³C-HYSCORE spectra of biogenic CM and meteoritic CM are also clearly different (Fig. 3 and Table 2). The carbon signal centered at the nuclear frequency ν _n=3.7 MHz of ¹³C is weak and narrow in cherts, while it is broad and intense in meteorites. Because of the low ¹³C abundance (1.1%), the intensity and shape of carbon signal are very sensitive to the geometry and environment of the radical. Ikoma et al. (1998, 2002) showed that the signal intensity ratio ¹³C/¹H increases with the number of equivalent carbon nuclei interacting with the unpaired electron of the radical. This ratio is clearly larger in meteorites than in cherts, which may be related to the high aliphatic branching of aromatic units in meteoritic CM as already deduced from the shape of the ¹H-HYSCORE spectra and from ¹³C NMR (Gardinier et al., 2000; Cody and Alexander, 2005). The large number of carbon atoms located at short distances from the polyaromatic unit increases the probability for the unpaired electron to interact with a ¹³C nucleus. This is not the case with the biogenic CM in cherts, which exhibits a weak ¹³C signal due to distant carbons and a ¹H signal typical of aromatic hydrogens (H_a). These differences in the ¹³C/¹H ratio may also reflect size differences of aromatic fragments (Ikoma et al., 1998, 2002). This would indicate that aromatic radicals of fossil CM in cherts are smaller than those of meteoritic CM. Nevertheless, HYSCORE spectroscopy shows that ¹H and ¹³C signatures of biogenic CM can be easily distinguished from those of meteorites, despite the fact that they exhibit the same cw-EPR signal (Fig. 1).

3.2.3. Evidence of ³¹P and ¹⁴N in biogenic carbonaceous material

For particular τ values, a narrow HYSCORE peak appears at 12.2 MHz in cherts, which is lacking in meteoritic CM, as shown in Fig. 6 for Clarno, Gunflint, and Warrawoona cherts recorded with τ=136, 156, and 136 ns, respectively. As this peak does not correspond to ¹³C or ¹H nuclear frequencies, it indicates the presence of a heteroelement localized inside or sufficiently close to carbonaceous radicals to interact with the electron spin via a dipolar interaction. There are no nuclei with such a high nuclear frequency; however, this value matches with a double frequency transition of ³¹P (I=1/2, ν _n=6.1 MHz, abundance 100%) that should appear at 2ν _n=12.2 MHz. Compared with the ¹H signal, the intensity of the ³¹P signal increases with the age of the CM and is the dominant signal for the oldest sample (Warrawoona, 3.5 Gyr). However, this attribution of the peak at 12.2 MHz to a double frequency transition of ³¹P may appear speculative in absence of a direct observation of a ³¹P spectrum around 6.1 MHz.

FIG. 6.

HYSCORE spectra at 9 K showing the double frequency ³¹P transition in three cherts: Clarno (τ=136 ns), Gunflint (τ=256 ns), and Warrawoona (τ=136 ns). Color images available online at www.liebertonline.com/ast

Convincing proof of the presence of ³¹P was obtained by studying the CM extracted from the Warrawoona chert by the HF/HCl extraction procedure (see Section 2.1). As the sample is an almost pure carbonaceous material after this demineralizing reaction, the HYSCORE intensity is much higher, and the spectrum reveals features that cannot be observed in a pristine chert, which contains only a small amount of CM. The HYSCORE spectrum at 4 K (Fig. 7) of the extracted CM shows two clear indications for the presence of ³¹P nuclei in addition to other nuclei and confirms the lack of hydrogen already observed in the chert (Fig. 6) and expected for such primitive CM (3.5 Gyr). In the positive quadrant (+,+) of Fig. 7, corresponding to hyperfine coupling regime A<2ν _n for nuclear spin I=1/2, three peaks can be distinguished that are centered at 5.8 MHz. They correspond to double frequency transitions of ²⁹Si (ν _n=2.9 MHz). At a slightly higher frequency, three peaks centered at the frequency of ³¹P (ν _n=6.1 MHz) correspond to hf interaction parameters A _iso=8 MHz and T=1 MHz with phosphorous. Double quantum (dq) transitions of ¹⁴N are also clearly observed at low frequency. The negative quadrant (+,−) corresponds to hyperfine coupling regime A>2ν _n, with peaks centered at frequency |A|/2, and split by 2ν _n. In addition to double quantum peaks of ¹⁴N, two pairs of peaks can be distinguished that correspond to ¹³C (|A|/2=7.8 MHz) and ³¹P (|A|/2=6.9 MHz), identified by their splitting 2ν _n=3.7 and 12.2 MHz, respectively. This result clearly shows that at least two types of ³¹P nuclei are bonded to the radical and characterized by their hyperfine interaction A≈8 MHz and A≈13.8 MHz, respectively. The strong ¹³C interaction A≈15.6 MHz is compatible with carbon atoms of the polyaromatic rings, which are directly in contact with the unpaired electron spin. The presence of several types of ¹⁴N nuclei in such mature CM indicates that these nitrogen atoms are incorporated in the aromatic rings, which protected them from elimination during diagenesis and metamorphism. In contrast, the ¹⁴N signal is very weak or lacking in the meteoritic CM.

FIG. 7.

HYSCORE spectrum at 4 K and τ=130 ns of the CM extracted from the Warrawoona chert, showing the hyperfine interactions with ³¹P, ¹³C, ¹⁴N, and ²⁹Si nuclei (dq indicating a double quantum transition). The (+,+) quadrant corresponds to ν ₁>0 and ν ₂>0, and the (+,−) quadrant corresponds to ν₁>0 and ν₂<0. Color images available online at www.liebertonline.com/ast

The clear difference between the HYSCORE spectrum of the pristine chert, which contains only a double frequency peak of ³¹P with no measurable hf interaction, and the HYSCORE spectrum of the extracted CM, with measurable ³¹P hf interactions of two distinct phosphorous atoms and no double frequency peaks, may be a direct consequence of the demineralization procedure. The presence of multifrequency ³¹P transitions in pristine cherts (four-frequency transitions at 24 MHz are also observed, not shown) indicates that P atoms belong to a phosphorous-rich component in contact with the CM, so that there are no direct C-P bonds. This separation of the CM component and P-rich component explains the presence of magnetically equivalent ³¹P nuclei responsible for the multifrequency transitions, which weakly interact by dipolar interaction with the electron spins of the CM, giving a single peak. This strongly suggests that these P-rich components are embedded in the carbonaceous microstructures, as already observed for other inorganic compounds in Precambrian rocks (van Zuilen et al., 2012) and in meteorites (El Amri et al., 2005). The extraction of the CM by the demineralization procedure could have destroyed the P-rich microprecipitates, with formation of direct C-P bonds in the carbonaceous network that are responsible for measurable hf couplings with both isotropic and dipolar contributions. Whatever the nature of these P-rich components (clusters, nanoparticles, etc.), their systematic presence in contact with biogenic carbonaceous particles in cherts is a strong indication of the biological origin of phosphorous. They may originate from the decomposition of nucleic acids and other P-rich biomolecules during diagenesis. The fact that only ³¹P, ¹³C, and ¹⁴N interactions are detected in the Warrawoona chert is due to the complete hydrogen loss in this very mature CM. In the younger Clarno chert, the less mature CM gives a strong ¹H signal and a weaker ³¹P signal, indicating less abundant and (or) less organized P-rich phases. The Proterozoic chert (Gunflint) is an intermediate situation. Also, the presence of various types of ¹⁴N in the biogenic CM extracted from cherts, which are lacking in the meteoritic CM, is a strong indication of the biologic origin of nitrogen, which may originate from the decomposition of proteins and nucleic acids.

3.2.4. Interaction with ²⁹Si nuclei

As the CM is embedded in a Si-rich mineral matrix in cherts and in meteorites, it is not surprising to observe the systematic presence of a ²⁹Si signal (4% natural abundance), with characteristics depending on the origin of the CM. This signal is weak and broad in the meteoritic CM, while it appears narrow and intense in cherts except for the oldest one (Warrawoona). However, a ²⁹Si signal is clearly observed when the CM is extracted from this chert (Fig. 7). The differences in the ²⁹Si signals may reflect different types of chemical bonding between the CM and the mineral matrix. We expect that direct C-Si bonds should give stronger hyperfine interactions (broader ²⁹Si signals) than C-O-Si bonds. If this interpretation is correct, it means that carbonaceous particles would be dominantly linked to the mineral by C-O-Si bonds in cherts and by C-Si bonds in meteorites.

3.3. Effect of thermal treatments on the hyperfine interactions

It was shown in Section 3.1 that thermal treatments of a young chert and of biomolecules encapsulated in a SiO₂ matrix give the same evolution of the cw-EPR line of CM. To evaluate the effect of thermal metamorphism on HYSCORE biosignatures, we compared in more detail the radicals produced by carbonization of biomolecules encapsulated in SiO₂ (C-type CM) and the radicals produced by natural aging in cherts (F-type CM).

For each type of encapsulated molecules (BSA, DNA, ATP, glucose), the HYSCORE spectra were recorded at three τ values (136, 156, and 200 ns) after each thermal step (400°C, 450°C, 500°C, 550°C, 600°C, and 650°C). Figure 8 displays examples of HYSCORE spectra recorded at 9 K and τ=136 ns after the thermal step at 550°C. This temperature was chosen because it gives a cw-EPR line with the same linewidth ΔB _pp (related to the H/C ratio of the CM) as Gunflint chert.

FIG. 8.

HYSCORE spectra at 9 K and τ=136 ns of molecular precursors encapsulated in sol-gel silica and treated at 550°C. Color images available online at www.liebertonline.com/ast

Several features of the C-type CM are similar to those observed in meteoritic (M-type) CM but differ from F-type CM (Figs. 8 and 3). First, the shape of the ¹H ridges is linear and very similar in C-type and M-type CM. This linear ridge is the signature of the distribution of highly branched aromatic radicals. This similitude of C- and M-type CM is confirmed by the ¹³C signal, which is broad and intense in both types of carbonaceous materials, while it is weak and narrow in F-type CM. It is also confirmed by the ²⁹Si signal, which is broad in both M- and C-type CM, while it is narrow and generally intense in F-type CM. The presence of similar ¹H, ¹³C, and ²⁹Si hf structures in C- and M-type CM that are different from those of F-type CM suggests that the organic precursors at the origin of the meteoritic CM were probably submitted to a thermal event before incorporation in the parent body. However, this thermal episode was not strong enough to eliminate hydrogen atoms. The cosmochemical implications of this result are beyond the scope of the present paper, and more work is needed to explore the thermal history of the meteoritic CM by EPR spectroscopy.

The second characteristic of C-type CM is the absence of ³¹P signal even after carbonization of P-rich precursors (DNA and ATP), while this signal is always present in F-type CM. This indicates that P-rich molecules produced during thermal treatment of DNA and ATP are not embedded in the CM in the form of P-rich phases as in F-type CM. Instead, they are localized at large distances from the electron spins. As dipolar interactions are proportional to R ⁻³, where R is the distance between the ³¹P nuclei and the unpaired electron spin of radicals, the amplitude of the nuclear modulations of the spin echo vanishes at large distances, which explains the lack of ³¹P signal for carbonized DNA and ATP.

This possible effect of temperature on the separation of phosphorous-rich precipitates from the CM also occurs for thermal treatment on a young chert. The 2-D projections of the HYSCORE spectra at τ=200 ns of Clarno treated at 300°C and 400°C compared with the pristine sample are shown in Fig. 9. No HYSCORE spectra could be detected for thermal treatments at higher temperature. The pristine sample shows a very weak ¹³C signal, the double frequency peak of ³¹P and a ¹H ridge with two peaks (arrows) corresponding to the anisotropic hyperfine interaction of H_a hydrogen atoms. These features are the main characteristic of F-type CM in cherts. Thermal treatments modify the ¹H spectrum, with the loss of H_a hydrogen atoms, as revealed by the linearization of the proton ridge and the lack of shifted peaks, indicating that the radicals are more branched than in pristine chert. The vanishing of the double frequency ³¹P peak can also be observed. This confirms that a thermal episode can destroy the P-rich phases and modifies the branching level of aromatic moieties. Consequently, ¹H, ¹³C, and ³¹P HYSCORE spectra of CM provide nuclear magnetic biosignatures only in cherts that have not been submitted to a metamorphic episode. However, the fact that these biosignatures are observed in Warrawoona and Gunflint cherts shows that they are not strongly affected by a metamorphic grade reaching the greenschist facies. This confirms the results obtained for the evolution of the cw-EPR lineshape (Bourbin et al., 2013).

FIG. 9.

HYSCORE spectra at 30 K of Clarno chert recorded with τ=200 ns. (a) pristine sample; (b) treated at 300°C; (c) treated at 400°C.

4. Discussion

Organic radicals are always present in primitive CM (Skrzypczak-Bonduelle et al., 2008; Binet et al., 2002) and can thus be used as internal probes to study their chemical environment at a subnanometer scale by the indirect detection of nuclear resonances of nuclei with nonzero nuclear spin by using pulsed EPR spectroscopy. The main results are summarized in Table 2, highlighting the features that can be considered as possible biosignatures of primitive life.

Concerning the cw-EPR line itself, it was previously shown that the progressive loss of hydrogen (¹H) during the Precambrian was responsible for the EPR line narrowing observed upon increasing age of the chert and of a specific modification of the EPR lineshape, which moved from a pure Lorentzian shape to a stretched Lorentzian shape (Skrzypczak-Bonduelle et al., 2008). The amount of stretching increases with the age of the CM, the most stretched line corresponding to the oldest CM (∼3.5 Gyr), which allowed us to propose a method for dating the CM with respect to its mineral matrix (Bourbin et al., 2013). On the opposite, the hydrogenated CM of carbonaceous meteorites (∼4.6 Gyr) always gives a purely Lorentzian EPR line (Binet et al., 2002, 2004b). Archean CM and meteoritic CM differ also by the temperature dependence of their EPR lines (proportional to the paramagnetic susceptibility), which obeys a Curie's law specific for S=1/2 radicals in Archean cherts and a more complex law indicative of S=1 diradicals in meteorites (Binet et al., 2004a, 2004b; Delpoux et al., 2011). However, such cw-EPR analyses, which give information on the age and the origin (terrestrial versus extraterrestrial) of the CM, give no indication of their biogenic/abiotic origin. In this work, the indirect detection of nuclear magnetic frequencies by pulsed EPR showed that the intensity and shape of the ¹H spectrum is sensitive to the biogenicity of the organic precursors. As anticipated, the intensity of the ¹H spectrum decreased in cherts of increasing age and vanished in the oldest Archean chert. This feature can be used to detect any biogenic contamination (endolithic bacteria, for example) in Archean cherts, as the CM originating from bacteria that contaminated Archean cherts long after the sediment formation (i.e., from the Proterozoic to present) would exhibit intense ¹H spectra. An important result is that the shape of the ¹H HYSCORE spectrum is also sensitive to the origin of the precursors and their maturation degree. At the molecular level, this feature is interpreted as the direct consequence of the degree of aliphatic branching of aromatic radicals, which controls the ratio C=H_a/(H_a+H_b+H_d), with C=0 for totally branched radicals and C=1 for radicals with no aliphatic branching. It appeared from pulsed EPR that C≈0 is characteristic of meteoritic CM and of carbonized molecular precursors, biogenic or not. On the contrary, CM from fossilized organisms is characterized by 0<< C<1. The sensitivity to thermal treatments of the nuclear magnetic spectra of CM indicates that the biosignatures are no longer valid for CM that experienced a metamorphism higher than the greenschist facies.

However, these ¹H biosignatures cannot be used for Archean cherts, as CM has lost most of its hydrogen in this case. After checking that there is no ¹H spectrum in an Archean sample, indicating a very low level of biological contamination, the next step is the study of other isotopes, the most interesting ones being ¹³C and ³¹P (and possibly ¹⁴N). The ¹³C spectrum is complementary to that of ¹H, as it is also sensitive to the branching level of aromatic radicals. The most branched radicals (meteoritic CM and CM from carbonized precursors) give strong and broad ¹³C ridges, while poorly branched CM (fossils) gives very weak and narrow ¹³C peaks, independently of its age. Phosphorous is also a promising signature of a biological origin, as we systematically found a ³¹P signal in cherts containing fossilized bacteria. This signal is lacking in meteorites and, more surprisingly, in CM formed by carbonization of phosphorous-rich biomolecular precursors (DNA, ATP). This indicates that only the natural fossilization process is able to maintain phosphorous atoms at close distance (subnanometer scale) from the radicals. The most likely origin of this phosphorous is the decomposition of nucleic acids (DNA, RNA) of fossilized organisms. We are presently studying Archean chert samples from South Africa (Josefsdal Formation), which fully confirm the presence of ³¹P close to aromatic radicals. The chemical form of this phosphorous is not yet determined and will be the object of future work.

It must be stressed that all these nuclear magnetic biosignatures are only those corresponding to CM responsible for the narrow EPR line. It might be possible that other forms of CM are also present but are not detected by cw-EPR and thus cannot be probed by HYSCORE spectroscopy. For example, more or less graphitized carbons give broad lines with various shapes, width, and g-factor anisotropy (Barbon and Brustolon, 2012). If such forms of carbon are present in our samples, their cw-EPR signals are too weak to be detected, and we checked by echo-detected EPR that all the HYSCORE spectra correspond to the narrow cw-EPR line of CM. Moreover, the presence of graphitized carbon appears very unlikely in the studied cherts, because they suffer from a metamorphism that is too low (greenschist facies or lower) to transform carbon into graphite. This was precisely the reason why we selected these well-documented samples.

It appears that cw-EPR and pulsed EPR are complementary to Raman spectroscopy for the analysis of primitive CM. We believe that combination of Raman parameters of the well-known G and D lines (width, intensity ratio, etc.), cw-EPR parameters (g-factor, width, lineshape, magnetism), and pulsed-EPR parameters (intensity and shape of ¹H, ¹³C, ³¹P resonances) would constitute fingerprints of the origin, age, and evolution of primitive CM. In addition to CM, it is worth noting that EPR can also be used to detect inorganic biosignatures, as for example biogenic vanadyl VO²⁺ ions (Gourier et al., 2010) and biogenic MnO₂ mineralizations (Kim et al., 2011).

5. Conclusion

It may be concluded that EPR spectroscopy in its different modalities constitutes a powerful tool to investigate primitive organic matter in relation with the origin of life on Earth. In this work, we showed that the intensity and the shape of EPR-detected nuclear resonances of ¹H, ¹³C, and ³¹P are markers of the precursors (biological, meteoritic), the age, and the thermal history of CM. The main advantage of EPR is a unique combination of sensitivity, resolution, and selectivity. In addition, this technique is non-invasive, as pristine samples can be studied without preparation, and CM can be detected at the part-per-million level in the bulk of small (centimeter- to millimeter-sized) pieces of rocks. In association with other methods (Bandfield et al., 2001; Marzo et al., 2007; Derenne et al., 2008; Orofino et al., 2010; Westall and Cavalazzi, 2011; Bower et al., 2013), EPR offers a great potential for the search for fossil traces of life on Mars.

Footnotes

Acknowledgments

The authors acknowledge the Centre National d'Etudes Spatiales (CNES) and the Institut National des Sciences de l'Univers (INSU) of CNRS for financial support. We also thank François Robert from Museum National d'Histoire Naturelle (Paris) for providing the chert samples. The carbonaceous matter of the Tagish Lake meteorite was provided by Luann Becker, from University of California, Santa Barbara (California, USA).

Abbreviations

ATP, adenosine triphosphate; BSA, bovine serum albumin; CM, carbonaceous matter; cw, continuous wave; EPR, electron paramagnetic resonance; hf, hyperfine; HYSCORE, hyperfine sublevel correlation; NMR, nuclear magnetic resonance.

References

Bandfield

J.F.

, Moreau

J.W.

, Chan

C.S.

, Welch

S.A.

, Little

2001. Mineralogical biosignatures and the search for life on Mars. Astrobiology, 1:447–466.

Barbon

, Brustolon

2012. An EPR study on nanographites. Appl Magn Reson, 42:197–210.

Barghoorn

E.S.

, Tyler

S.A.

1965. Microorganisms from the Gunflint chert. Science, 147:563–577.

Bernard

, Beyssac

, Benzerara

, Findling

, Tsvetkov

, Brown

G.E.

Jr.

2010. XANES, Raman and XRD study of anthracene-based cokes and saccharose-based chars submitted to high temperature pyrolysis. Carbon, 48:2506–2516.

Beyssac

, Goffé

, Chopin

, Rouzaud

J.N.

2002. Raman spectra of carbonaceous material in metasediments: a new geothermometer. Journal of Metamorphic Geology, 20:859–871.

Beyssac

, Goffé

, Petitet

J.B.

, Froigneux

, Moreau

, Rouzaud

J.N.

2003. On the characterization of disordered and heterogeneous carbonaceous materials by Raman spectroscopy. Spectrochim Acta A Mol Biomol Spectrosc, 59:2267–2276.

Binet

, Gourier

, Derenne

, Robert

2002. Heterogeneous distribution of paramagnetic radicals in insoluble organic matter from the Orgueil and Murchison meteorites. Geochim Cosmochim Acta, 66:4177–4186.

Binet

, Gourier

, Derenne

, Robert

, Ciofini

2004a. Occurrence of abundant diradicaloid moieties in the insoluble organic matter from the Orgueil and Murchison meteorites: a fingerprint of its extraterrestrial origin? Geochim Cosmochim Acta, 68:881–891.

Binet

, Gourier

, Derenne

, Pizzarello

, Becker

2004b. Diradicaloids in the insoluble organic matter from the Tagish Lake meteorite: comparison with the Orgueil and Murchison meteorites. Meteorit Planet Sci, 39:1649–1654.

10.

Binet

, Gourier

, Derenne

2008. Potential of EPR imaging to detect traces of primitive life in sedimentary rocks. Earth Planet Sci Lett, 273:359–366.

11.

Bourbin

, Derenne

, Westall

, Gourier

, Rouzaud

J.N.

, Gautret

2012. EPR study of Archean microbial mats analogues and comparison with Archean cherts: towards a possible marker of oxygenic photosynthesis? Orig Life Evol Biosph, 42:569–585.

12.

Bourbin

, Gourier

, Derenne

, Binet

, Le Du

, Westall

, Kremer

, Gautret

2013. Dating carbonaceous matter in Archean cherts by electron paramagnetic resonance. Astrobiology, 13:151–162.

13.

Bower

D.M.

, Steele

, Fries

M.D.

, Kater

2013. Micro Raman spectroscopy of carbonaceous material in microfossils and meteorites: improving a method for life detection. Astrobiology, 13:1–11.

14.

Brasier

M.D.

, Green

O.R.

, Jephcoat

A.P.

, Kleppe

A.T.

, van Kranendonk

M.J.

, Lindsay

J.F.

, Steele

, Grassineau

N.V.

2002. Questioning the evidence for Earth's oldest fossils. Nature, 416:76–81.

15.

Buseck

P.R.

, Huang

B.J.

1985. Conversion of carbonaceous material to graphite during metamorphism. Geochim Cosmochim Acta, 49:2003–2016.

16.

Bustin

R.M.

, Rouzaud

J.N.

, Ross

J.V.

1995. Natural graphitization of anthracite: experimental considerations. Carbon, 33:679–691.

17.

Campbell

S.E.

1982. Precambrian endoliths discovered. Nature, 299:429–431.

18.

Chyba

, Sagan

1992. Endogenous production, exogenous delivery and impact-shock synthesis of organic molecules: an inventory for the origin of life. Nature, 355:125–132.

19.

Cody

G.D.

, Alexander

C.M.O'D.

2005. NMR studies of chemical structural variation of insoluble organic matter from different carbonaceous chondrite groups. Geochim Cosmochim Acta, 69:1085–1097.

20.

Cronin

J.R.

, Pizarello

, Cruikshank

D.P.

1988. Organic matter in carbonaceous chondrites, planetary satellites, asteroids and comets. Meteorites and the Early Solar System. Kerridge

J.F.

, Mathews

M.S.

University of Arizona Press: Tucson, 819–957.

21.

Delpoux

, Gourier

, Vezin

, Binet

, Derenne

, Robert

2011. Biradical character of D-rich carriers in the insoluble organic matter of carbonaceous chondrites: a relic of the protoplanetary disk chemistry. Geochim Cosmochim Acta, 75:326–336.

22.

Derenne

, Robert

, Skrzypczak-Bonduelle

, Gourier

, Binet

, Rouzaud

J.N.

2008. Molecular evidence for life in the 3.5 billion year old Warrawoona chert. Earth Planet Sci Lett, 272:476–480.

23.

Durand

, Nicaise

1980. Procedures for kerogen isolations. Kerogen. Durand

Technip: Paris, 35–53.

24.

El Amri

, Maurel

M.C.

, Sagon

, Baron

M.H.

2005. The micro-distribution of carbonaceous matter in the Murchison meteorite as investigated by Raman imaging. Spectrochim Acta A Mol Biomol Spectrosc, 61:2049–2056.

25.

Enoki

, Kobayashi

2005. Magnetic nanographite: an approach to molecular magnetism. J Mater Chem, 15:3999–4002.

26.

Garcia-Ruiz

J.M.

, Hyde

S.T.

, Carnerup

A.M.

, Christy

A.G.

, van Kranendonk

M.J.

, Welham

N.J.

2003. Self-assembled silica-carbonate structures and detection of ancient microffossils. Science, 302:1194–1197.

27.

Gardinier

, Derenne

, Robert

, Behar

, Largeau

, Maquet

2000. Solid state CP/MAS 13C NMR of the insoluble organic matter of the Orgueil and Murchison meteorites: quantitative study. Earth Planet Sci Lett, 184:9–21.

28.

Gourier

, Robert

, Delpoux

, Binet

, Vezin

, Moissette

, Derenne

2008. Extreme deuterium enrichment of organic radicals in the Orgueil meteorite: revisiting the interstellar interpretation? Geochim Cosmochim Acta, 72:1914–1923.

29.

Gourier

, Delpoux

, Skrzypczak-Bonduelle

, Binet

, Ciofini

, Vezin

2010. EPR, ENDOR and HYSCORE study of the structure and the stability of vanadyl-porphyrin complexes encapsulated in silica: potential paramagnetic biomarkers for the origin of life. J Phys Chem B, 114:3714–3725.

30.

Hayatsu

, Matsuoka

, Scott

R.G.

, Studier

M.H.

, Anders

1977. Origin of the organic matter in the early Solar System, VII: the organic polymers in carbonaceous chondrites. Geochim Cosmochim Acta, 41:1325–1339.

31.

Hoering

T.C.

1967. The organic geochemistry of Precambrian rocks. Researches in Geochemistry, 2

Abelson

John Wiley: New York, 87–111.

32.

Höfer

, Grupp

, Nebenfürh

, Mehring

1986. Hyperfine sublevel correlation (HYSCORE) spectroscopy: a 2D ESR investigation of the squaric acid radical. Chem Phys Lett, 132:279–282.

33.

Ikoma

, Ito

, Tero-Kubota

, Akiyama

1998. HYSCORE study on coal radicals. Energy Fuels, 12:1363–1368.

34.

Ikoma

, Ito

, Tero-Kubota

2002. Exploring radicals in carbonaceous solids by means of pulsed EPR spectroscopy. Energy Fuels, 16:40–47.

35.

Joly

V.L.J.

, Kigushi

, Hao

S.J.

, Takai

, Enoki

, Sumii

, Amemiya

, Muramatsu

, Hayashi

, Kim

Y.A.

, Endo

, Campos-Delgado

, Lopez-Urias

, Botello-Mendez

, Terrones

, Dresselhaus

M.S.

2010. Observation of magnetic edge state in graphene nanoribbons. Phys Rev B Condens Matter Mater Phys, 81. 10.1103/PhysRevB.81.245428.

36.

Livage

, Coradin

, Roux

2001. Encapsulation of biomolecules in silica gels. J Phys Condens Matter, 13:R673–R691.

37.

Luque

F.J.

, Pasteris

J.D.

, Wopenka

, Rodas

, Barrenechea

J.F.

1998. Natural fluid-deposited graphite: mineralogical characteristics and mechanisms of formation. Am J Sci, 298:471–498.

38.

Kim

S.S.

, Bargar

J.R.

, Nealson

K.H.

, Flood

B.E.

, Kirschvink

J.L.

, Raub

T.D.

, Tebo

B.M

, Villalobos

2011. Searching for biosignatures using electron paramagnetic resonance (EPR) analysis of manganese oxides. Astrobiology, 11:775–786.

39.

Marshall

C.P.

, Edwards

H.G.M.

, Jehlicka

2010. Understanding the application of Raman spectroscopy to the detection of traces of life. Astrobiology, 10:229–243.

40.

Marzo

G.A.

, Politi

, Blanco

, Brunetto

, Fonti

, Orofino

2007. On potential spectroscopic detection of microfossils on Mars. Earth Moon Planets, 101:127–140.

41.

Maurette

, Duprat

, Engrand

, Kurat

2006. From Earth to Mars with micrometeorite volatiles. Adv Space Res, 38:701–708.

42.

Mojzsis

S.J.

, Arrhenius

, McKeegan

K.D.

, Harrison

T.M.

, Nutman

A.P.

, Friend

C.R.I.

1996. Evidence for life on Earth before 3,800 million years ago. Nature, 384:55–59.

43.

Mrozowski

1988a. ESR study of carbonization and coalification process part I: carbonaceous compounds. Carbon, 26:521–529.

44.

Mrozowski

1988b. ESR study of carbonization and coalification process part II: biological materials. Carbon, 26:531–541.

45.

Orofino

, Blanco

, D'Elia

, Lichelli

, Fonti

, Marzo

G.A.

2010. Study of terrestrial fossils in phyllosilicate rich soils: implication in the search for biosignatures on Mars. Icarus, 208:202–206.

46.

Pasteris

J.D.

, Wopenka

2003. Necessary, but not sufficient: Raman identification of disordered carbon as a signature of ancient life. Astrobiology, 3:727–738.

47.

Pöppl

, Kevan

1996. A practical strategy for determination of proton hyperfine interaction parameters in paramagnetic transition metal ion complexes by two-dimensional HYSCORE electron spin resonance spectroscopy in disordered systems. J Phys Chem, 100:3387–3394.

48.

Quirico

, Montagnac

, Rouzaud

J.-N.

, Bonal

, Bourot-Denise

, Duber

, Reynard

2009. Precursors and metamorphic condition effects on Raman spectra of poorly ordered carbonaceous matter in chondrites and coals. Earth Planet Sci Lett, 287:185–193.

49.

Retcofsky

H.L.

, Thompson

G.P.

, Raymond

, Friedel

R.A.

1975. Studies of ESR linewidths in coals and related materials. Fuel, 54:126–128.

50.

Rigali

J.M.

, Nagy

1997. Organic free radicals and micropores in solid graphitic carbonaceous matter at the Oklo natural fission reactors, Gabon. Geochim Cosmochim Acta, 61:357–368.

51.

Robertson

2002. Diamond-like amorphous carbon. Mater Sci Eng R Rep, R37:129–281.

52.

Roedder

1981. Are the 3800-Myr-old Isua objects microfossils, limonite-stained fluid inclusions, or neither? Nature, 293:159–162.

53.

Schopf

J.W.

1993. Microfossils in the Early Archean Apex chert—new evidence of the antiquity of life. Science, 260:640–646.

54.

Schopf

J.W.

, Kudryavtsev

A.B.

2012. Biogenicity of Earth's earliest fossils: a resolution of the controversy. Gondwana Research, 22:761–771.

55.

Schweiger

, Jeschke

2001. Principle of Pulsed Electron Paramagnetic Resonance. Oxford University Press: Oxford, UK.

56.

Sephton

M.A.

, Pillinger

C.T.

, Gilmour

1999. Small scale hydrous pyrolysis of macromolecular material in meteorites. Planet Space Sci, 47:181–187.

57.

Skrzypczak-Bonduelle

, Binet

, Delpoux

, Vezin

, Derenne

, Robert

, Gourier

2008. EPR of radicals in primitive organic matter: a tool for the search of biosignatures of the most ancient traces of life. Appl Magn Reson, 133:371–397.

58.

Uebersfeld

, Etienne

, Combrisson

1954. Paramagnetic resonance, a new property of coal-like materials. Nature, 174:614.

59.

van Zuilen

M.A.

, Fliegel

, Wirth

, Lepland

, Qu

, Schreiber

, Romashkin

A.E.

, Philippot

2012. Mineral-templated growth of natural graphite films. Geochim Cosmochim Acta, 83:252–262.

60.

Westall

, Cavalazzi

2011. Biosignatures in rocks. Encyclopedia of Geobiology. Thiel

, Reitner

Spinger: Berlin, 189–201.

61.

Westall

, Folk

R.L.

2003. Exogenous carbonaceous microstructures in Early Archean cherts and BIFs from the Isua Greenstone Belt: implication for the search for life in ancient rocks. Precambrian Res, 126:313–330.

62.

Whelan

J.K.

, Thompson-Rizer

1993. Chemical methods for assessing kerogen and proto-kerogen types and maturity. Organic Geochemistry. Engel

, Macko

Plenum Press: New York, 289–353.

63.

Wopenka

, Pasteris

J.D.

1993. Structural characterization of kerogens to granulite-facies graphite: applicability of Raman microprobe spectroscopy. Am Mineral, 78:533–557.

Nuclear Magnetic Biosignatures in the Carbonaceous Matter of Ancient Cherts: Comparison with Carbonaceous Meteorites

Abstract

1. Introduction

2. Materials and Methods

2.1. Cherts and meteorites

2.2. Encapsulation of organic and bioorganic molecules

2.3. EPR experiments

3. Results

3.1. cw-EPR of carbonaceous matter

3.2. Pulsed EPR of cherts and meteorites

3.2.1. 1H hyperfine interactions

3.2.2. 13C hyperfine interaction

3.2.3. Evidence of 31P and 14N in biogenic carbonaceous material

3.2.4. Interaction with 29Si nuclei

3.3. Effect of thermal treatments on the hyperfine interactions

4. Discussion

5. Conclusion

Footnotes

Acknowledgments

Abbreviations

References

3.2.1. ¹H hyperfine interactions

3.2.2. ¹³C hyperfine interaction

3.2.3. Evidence of ³¹P and ¹⁴N in biogenic carbonaceous material

3.2.4. Interaction with ²⁹Si nuclei