A

iCSnieoptc©

K

1

stbmoaa

1d

Available online at www.sciencedirect.com

Spectrochimica Acta Part A 69 (2008) 1301–1310

CW- and pulsed-EPR of carbonaceous matter in primitivemeteorites: Solving a lineshape paradox

Olivier Delpoux a, Didier Gourier a,∗, Laurent Binet a, Herve Vezin b,Sylvie Derenne c, Francois Robert d

a CNRS, Ecole Nationale Superieure de Chimie de Paris (ENSCP, ParisTech), Laboratoire de Chimie de la MatiereCondensee de Paris, UMR-CNRS 7574, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France

b CNRS, Universite des Sciences et Technologies de Lille, Laboratoire de Chimie Organique et Macromoleculaire,UMR-CNRS 8009, 59655 Villeneuve d’Ascq, France

c CNRS, Ecole Nationale Superieure de Chimie de Paris (ENSCP, ParisTech), Laboratoire de Chimie Bioorganiqueet Organique Physique, UMR-CNRS 7573, 11 rue Pierre et Marie Curie, 75231 Paris Cedex 05, France

d CNRS, Museum National d’Histoire Naturelle, Laboratoire d’Etude de la Matiere Extraterrestre UMS-CNRS 62,61 rue Buffon, 75005 Paris, France

Received 2 August 2007; accepted 18 September 2007

bstract

Insoluble organic matter (IOM) of Orgueil and Tagish Lake meteorites are studied by CW-EPR and pulsed-EPR spectroscopies. The EPR lines due to polycyclic paramagnetic moieties concentrated in defect-rich regions of the IOM, with concentrations of the order of 4 × 1019 spin/g.W-EPR reveals two types of paramagnetic defects: centres with S = 1/2, and centres with S = 0 ground state and thermally accessible triple state= 1. In spite of the Lorentzian shape of the EPR and its narrowing upon increasing the spin concentration, the EPR line is not in the exchangearrowing regime as previously deduced from multi-frequency CW-EPR [L. Binet, D. Gourier, Appl. Magn. Reson. 30 (2006) 207–231]. It isnhomogeneously broadened as demonstrated by the presence of nuclear modulations in the spin-echo decay. The line narrowing, similar to anxchange narrowing effect, is the result of an increasing contribution of the narrow line of the triplet state centres in addition to the broader linef doublet states. Hyperfine sublevel correlation spectroscopy (HYSCORE) of hydrogen and 13C nuclei indicates that IOM• centres are small

olycyclic moieties that are moderately branched with aliphatic chains, as shown by the presence of aromatic hydrogen atoms. On the contraryhe lack of such aromatic hydrogen in triplet states suggests that these radicals are most probably highly branched. Paramagnetic centres areonsiderably enriched in deuterium, with D/H ≈ 1.5 ± 0.5 × 10−2 of the order of values existing in interstellar medium. 2007 Elsevier B.V. All rights reserved.

l corr

wbpahss

eywords: Primitive carbonaceous matter; Meteorites; EPR; Hyperfine subleve

. Introduction

Carbonaceous meteorites, the most primitive objects of theolar system, contain a substantial amount of carbon with upo 90% made of an insoluble, acid resistant amorphous car-onaceous material, generally referred to as insoluble organicatter (IOM) [1,2]. This IOM preserves the record of the first

rganic synthesis in Early Solar System 4.56 billion years (Byr)go. IOM contains soluble organic molecules such as aminocids and sugars, and could have supplied Early Earth and Mars

∗ Corresponding author. Tel.: +33 1 4427 6706; fax: +33 1 4634 7489.E-mail address: Didier-gourier@enscp.fr (D. Gourier).

rtaNi[h

386-1425/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.saa.2007.09.045

elation spectroscopy; Spin-echo correlation spectroscopy

ith organic molecules during the periods of intense meteoriticombardments, which ended up ≈ 3.9 Byr ago [3–5]. The amor-hous and strongly disordered character of the IOM preventsny detailed structural analysis, however, a general descriptionave been proposed from combination of various techniques,uch as NMR spectroscopy [6,7], pyrolysis coupled to masspectroscopy [8], transmission electron microscopy [9] anduthenium tetroxide oxidation [10]. They converge on a descrip-ion of IOM as an extended network of aromatic (sp2 carbon) andliphatic (sp3 carbon) moieties with other elements such as H, O,

and S. Despite the fact that meteoritic IOM has a bulk chem-cal composition similar to that of terrestrial type III kerogens8], its structure is different and is made of small aromatic unitsighly substituted and cross-linked by short aliphatic chains [7].

1302 O. Delpoux et al. / Spectrochimica Act

FAr

AFcoaIb

acpiiot

FoG

scEPintiotpnbbdlt(t

obaWrts((aeoe

ig. 1. Schematic representation of the insoluble organic matter of meteorites.polycyclic cluster with an odd number of carbon atoms is paramagnetic and

epresents a IOM• centre.

very schematic representation of meteoritic IOM is shown inig. 1. On the contrary, IOM of oldest terrestrial rocks (micro-rystalline siliceous rocks referred to as cherts) mainly consistf extended polyaromatic structures, which can be interpreteds a disordered graphitic material [11]. This primitive terrestrialOM preserves the oldest life record on Earth, dated about 3.5illion years [12].

Amorphous carbonaceous materials, synthetic or natural,lways contain high concentration of paramagnetic defects thatan be easily detected by EPR spectroscopy [13–17]. Exam-les of continuous wave (CW) EPR signal of IOM are shownn Fig. 2 for two carbonaceous meteorites (Orgueil and Tag-

sh Lake) and a Precambrian chert (Gunflint formation, 1.9-Byrld). Unfortunately, CW-EPR is somewhat less informative inhis case as it invariably gives a unique line with a Lorentzian

ig. 2. X-band (9.5 GHz) CW-EPR spectra at room temperature of insolublerganic matter of Orgueil and Tagish Lake meteorites (4.56 Byr) and of theunflint chert (1.9 Byr). The dotted lines represent Lorentzian shape functions.

wI

2

Nl[CaemcoJJdbmfUft

a Part A 69 (2008) 1301–1310

hape at g = 2, which is identical in meteoritic IOM and in ancientherts [18]. For this reason it is not surprising that pioneeringPR investigations on meteorites in the 1960s [19–21] and onrecambrian cherts in the 1980s [15] could not give relevant

nformation on cosmochemistry and primitive life, and wereot followed by deeper investigations. This drawback is dueo the fact that structural information contained in the hyperfinenteraction is lost in the lineshape (inhomogeneous broadening)r, more problematical, is averaged out by exchange interac-ion if the Lorentzian shape results from an exchange narrowingrocess (homogeneous line). When the EPR line is inhomoge-eously broadened, information from hyperfine interactions cane partially recovered by detecting nuclear frequencies eithery electron–nuclear double resonance (ENDOR) [22] or byetecting nuclear modulations of the electron spin-echo enve-ope (ESEEM) in pulsed-EPR [23]. The hf interaction is lost ifhe EPR is exchanged narrowed and only a free induction decayFID) can be obtained by pulsed-EPR if the electronic relaxationime T2 is sufficiently long.

We show in this work that EPR of paramagnetic defectsf IOM in carbonaceous meteorites exhibits a paradoxicalehaviour, with characteristics of both exchange narrowingnd dipolar broadening processes. Indeed CW-EPR at X- and-bands clearly show the manifestation of an exchange nar-

owing process, while pulsed-EPR indisputably shows thathe Lorentzian line is inhomogeneously broadened by unre-olved hyperfine interaction. Spin-echo correlation spectroscopySECSY) and hyperfine sublevel correlation spectroscopyHYSCORE) demonstrate that exchange narrowing is onlypparent. The most striking result is the demonstration of anxtreme deuterium enrichment of paramagnetic centres in IOMf Orgueil and Tagish Lake meteorites. This result is of consid-rable interest from the cosmochemical point of view in relationith the recent discovery of deuterium hot-spots in meteoritic

OM by NanoSims [24].

. Experimental part

A piece of the Orgueil meteorite was provided by Museumational d’Histoire Naturelle of Paris. The IOM was iso-

ated from the mineral part by the standard HF/HCl treatment7,25]. The ground meteorite was extracted with water andH2Cl2/MeOH, saponified with KOH and hydrolyzed with HClnd HF/HCl. We previously checked that this treatment had noffect on the EPR signal as the same treatment on coals did notodify the lineshape, the linewidth, the g-factor and the spin

oncentration [26]. The IOM sample from the Tagish Lake mete-rite, isolated by the same method, was provided by the NASAohnson Space Centre (Houston, USA) with the permission of. Brook (the finder). The two samples consist in a black pow-er containing a few percent of mineral residues [26,27]. Theulk elemental compositions of the IOM were determined by ele-ental analysis at Wolf Laboratories (Rueil Malmaison, France)

or Orgueil and by Atlantic Microlab, Inc. (Norcross, Georgia,SA) for Tagish Lake. Their compositions are C100H72O18N3S2

or Orgueil and C100H46O15N10 for Tagish Lake. A small quan-ity (2 mg) of IOM was directly introduced in the EPR quartz tube

ica A

wrcwtwnsbfOmmtt

optOs(sE(lttlssagoccAfG

3

3

nattttomdaf0

asnsnithethbω

dCtqN

p(found for Tagish Lake meteorite (not shown). The temperaturedependence of the normalized spin concentration in meteoriteshas been attributed to the presence of two types of IOM• cen-

Fig. 3. Temperature variation of the CW-EPR line of paramagnetic centres inthe carbonaceous matter of Orgueil meteorite. (a) Peak-to-peak width of the

O. Delpoux et al. / Spectrochim

ithout special precaution because the EPR line was found to beemarkably stable over more than 5 years without any detectablehange of the EPR parameters (g-factor, lineshape, intensity). Itas not possible to increase the quantity of IOM introduced in

he tube because the resonant cavity of the EPR spectrometeras perturbed by the presence of small particles of ferromag-etic mineral residues in the IOM [26]. The Precambrian cherthown in Fig. 2 for comparison with meteorites was providedy Museum National d’Histoire Naturelle of Paris. It comesrom the Shreiber Beach locality of the Gunflint formation inntario (Canada). This 1.9 Byr-old rock contained carbonaceousicrostructures, which are uncontested fossilized prokaryoticicroorganisms [28]. As the IOM content of this chert is rela-

ively high, its spin concentration ≈ 1017 spin/g [18] is sufficiento avoid IOM extraction by the demineralising treatment.

CW-EPR and pulsed-EPR measurements were carriedut at X-band (9.64 GHz) between 4 K and room tem-erature with Bruker ELEXSYS E500 and E580 spec-rometers, respectively. Several pulse sequences were used.ne-dimensional spectroscopy was carried out with the

tandard two-pulse electron spin-echo envelope modulation2P-ESEEM) sequence (π/2–τ–π–τ–echo), the 3P-ESEEMequence (π/2–τ–π/2–T–π/2–τ–echo) and the matched 3P-SEEM sequence (π/2–τ–tp–T–tp–τ–echo). The matched pulse

tp) is set to tp = 72 ns for proton enhancement nuclear modu-ation. Two-dimensional techniques were also carried out: thewo-pulse spin-echo correlation spectroscopy (SECSY) fromhe π/2–t1–π/2–t1–t2 sequence [29] and the hyperfine sub-evel correlation spectroscopy (HYSCORE) from the four-pulseequence π/2–τ–π/2–t1–π–t2–π/2–τ–echo [30]. In all theseequences τ, T, t1 and t2 represent time delays between orfter microwave pulses. Hyperfine interactions were investi-ated with HYSCORE, which appears more suited than usualne-dimensional ESEEM experiments in the case of disorderarbonaceous matter such as coals [31] because it disentanglesomplex 1D-ESEEM spectra by splitting overlapping peaks.dditional CW-EPR spectra at 95 GHz (W-band) were per-

ormed at the Grenoble High Magnetic Field Laboratory (CNRS,renoble, France).

. Results and discussion

.1. CW-EPR: evidence of exchange narrowing

IOM of meteorites is made of small aromatic clusters con-ected by short aliphatic chains [7], most of them possessingn even number of carbon atoms for stability reasons and arehus diamagnetic. If a small proportion of these clusters con-ain an odd number of carbon atoms, and thus of π electrons,hey are paramagnetic and give an EPR spectrum. Fig. 2 showshe CW-EPR spectra recorded at X-band and room temperaturef paramagnetic defects in IOM of Orgueil and Tagish Lakeeteorites, compared with that of the 1.88-Byr old chert. These

efects will be hereafter referred to as IOM• centres. The spectrare identical and exhibit the same Lorentzian shape, the same g-actor ≈ 2.003, and similar linewidths, which amount 0.38 and.55 mT for Orgueil and Tagish Lake meteorites, respectively,

EWvca

cta Part A 69 (2008) 1301–1310 1303

nd 0.23 mT for the chert. In the case of meteorites, the line-hape is Lorentzian in the whole temperature range 4–300 K, andarrows upon increasing the temperature. A narrow Lorentzianignal in disordered solids may originate either from an exchangearrowing mechanism or from a dipolar broadening mechanismn a magnetically diluted spin system [32]. In principle thesewo mechanisms cannot coexist if the paramagnetic defects areomogeneously distributed because in these two situations thexchange frequency ωe exhibits different relations with respecto the distribution of local fields �B induced by unresolvedyperfine interaction, dipolar broadening and g-factor distri-ution. Exchange narrowing is determined by the conditione > γ�B while dipolar broadening imposes the inverse con-ition ωe < γ�B. As the exchange energy J = �ωe is due tooulombic repulsion between electron spins, it increases with

he concentration N of electron spins. The well-known conse-uence of strong exchange is a line narrowing upon increasing[32].Fig. 3 shows the temperature dependence of the peak-to-

eak linewidth (Fig. 3a) and the concentration of IOM• centresFig. 3b) in Orgueil meteorite. The same variations were also

PR at X-and W-bands (9.5 GHz and 95 GHz, respectively). The solid line for-band corresponds to the theoretical variation calculated with Eq. (1) from the

ariation at X-band. (b) Variation of the spin concentration with temperature. Theontribution of S = 1/2 radicals (IOM• centres) and S = 0 centres with thermallyccessible triple state (S = 1, TATS) are indicated.

1 a Act

tcibTfaiohTcwtbmp(Toaae

�

wg

ω

wcoEIospcgcalafc

N

Attlfo

ai

tNlcFatsOs[4pcoois

3

tInaiarrfirtnialdeabseEi

ifta

304 O. Delpoux et al. / Spectrochimic

res [33]. One type is a spin doublet (S = 1/2) responsible for theonstant part of the spin concentration in Fig. 3b. The other types a ground state spin singlet (S = 0) with a thermally accessi-le triplet state (TATS, S = 1) at 0.1 eV from the singlet state.he effect of TATS is an increase of the spin concentration by a

actor ≈ 170% between 130 K and room temperature in Orgueilnd Tagish Lake [27,33]. The peak-to-peak linewidth decreasesn the same temperature range, which is the usual manifestationf an exchange narrowing mechanism (Fig. 3a). This behaviouras never been found in terrestrial kerogens, which indicates thatATS are specific to meteoritic IOM [33]. The line narrowing atonstant spin concentration between 4 K and about 50 K (Fig. 3)as also observed in terrestrial kerogens (coals). It was attributed

o a motional narrowing effect due to thermally activated tum-ling of the macromolecular network [34]. The hypothesis ofotional and exchange narrowing has been confirmed by com-

aring the temperature dependence of the linewidth at X-band9.4 GHz) and at W-band (95 GHz), as shown in Fig. 3a [34].he full lines in this figure represents the theoretical variationf the linewidth at X- and W-bands under the hypothesis ofmotional narrowing below 60 K and an exchange narrowing

bove 130 K. These curves were obtained with the followingxpression calculated in the Redfield limit ωe � γ�B:

Bpp(T ) = af (T )2 + c

ωc(T )(1)

here ωc(T) is the correlation frequency of the fluctuating fieldiven by

c(T ) = ωe(T ) + ωm0 exp

(−�E

kT

)(2)

ithωm0 = 2.0 × 109 rad/s being the tumbling contribution to theorrelation frequency, and �E/k = 22 K is the activation energyf the low temperature tumbling motion. The first term ωe(T) inq. (2) is the exchange contribution to the correlation frequency.

t depends on temperature owing to the temperature dependencef the spin concentration. However, for T < 130 K, where thepin concentration is constant, ωe(T) does not depend on tem-erature, and has the constant value ωe0 = 3.2 × 109 rad/s. Theonstant c in Eq. (1) represents the contribution of g-anisotropy,-distribution and hyperfine interaction to the linewidth, with= 2.25 × 109 mT rad/s at X-band and c = 3.45 × 109 mT rad/st W-band. The term af(T)2 in Eq. (1) is the contribution of dipo-ar interactions between IOM• centres to the local fields, with= 3.4 × 108 mT rad/s and the temperature-dependent function

(T) is obtained from the temperature variation of the spin con-entration of Fig. 3b [33]:

∝ f (T ) = 1 + 1.75

1 + 0.002 exp(1210/T )(3)

ll the parameters a, c, ωe, ωm0 and �E have been adjusted fromhe temperature variation of �Bpp at X-band, and the model was

ested on its ability to predict the variation at W-band (continuousine in Fig. 3a). Despite the number of adjustable parameters, theair agreement between Eq. (1) and the experimental variationf linewidth between 4 K and 300 K at W-band is a convincing

ctLt

a Part A 69 (2008) 1301–1310

rgument in favour of an averaging of hyperfine and spin–spinnteractions by exchange narrowing processes.

It is worth noticing that Eq. (1) contains one addi-ional adjustable parameter, which is the spin concentration≈ 5 × 1020 spin/g at 300 K for Orgueil [34]. This value is

arger by two orders of magnitude than the bulk average spinoncentration 7.1018 spin/g measured from EPR intensity (seeig. 3b). This apparent disagreement disappears only if wessume that IOM• centres accumulate in restricted regions ofhe IOM, in the form of “defect hot-spots”. The presence ofuch local accumulations of defects has already been shown forrgueil and Murchison meteorites on the basis of microwave

aturation measurements and ENDOR enhancements factors26]. For example in the case of Orgueil, the two methods gave.1 ± 0.6 × 1019 spin/g and 3.1 ± 0.6 × 1019 spin/g at room tem-erature, respectively, which is also much larger than the averageoncentration. Despite some disagreements as to the exact valuef local concentrations, all these methods point to the occurrencef “defect hot-spots” in meteoritic IOM, in which exchangenteractions appear the most likely to explain the Lorentzianhape and the line narrowing at high spin concentration.

.2. Pulsed-EPR: evidence of dipolar broadening

The exchange-narrowing regime of the CW-EPR line implieshat the condition ωe > γ�B holds for Orgueil and Tagish LakeOM, and that the Lorentzian line is homogeneous. We expecto spin-echo in pulsed-EPR with such a pure homogeneous line,s echoes are the consequence of inhomogeneous broadening,.e if the condition ωe > γ�B holds. Surprisingly both Orgueilnd Tagish Lake IOM give spin-echoes in all the temperatureange 4–300 K. The case of Orgueil is shown in Fig. 4, whichepresents the integrated CW-EPR line and the two-pulse echoeld sweep EPR (echo-detected EPR). The echo-detected EPRepresents the variation of the echo intensity as a function ofhe external magnetic field. Thus it represents the inhomoge-eous component of the EPR spectrum. It appears clearly thatts shape perfectly matches the integrated CW-EPR lineshapet room temperature. It is slightly narrower than the CW-EPRine at 10 K, however the similitude of the two types of spectraemonstrates that the EPR line is inhomogeneously broadenedven at room temperature, and is definitely not homogeneouss predicted for an exchange narrowing regim. This paradoxicalehaviour indicates that the origin of the line narrowing at highpin concentration is more complex than expected for a simplexchange mechanism. Also, the inhomogeneous character of thePR line opens the possibility to study the unresolved hyperfine

nteraction.In principle three mechanisms can be at the origin of such

nhomogeneous broadening. The first one is a distribution of g-actors (equivalent to g-strain in solids) and(or) the anisotropy ofhe g-factor. This effect is frequency dependent and we expect

splitting of the Lorentzian line when the frequency is suffi-

iently high to give a g-anisotropy or a g-distribution larger thanhe linewidth. However, the lineshape of meteoritic IOM remainsorentzian at frequency as high as 285 GHz without any distor-

ion or splitting, in contrast to the lineshape of coals, which split

O. Delpoux et al. / Spectrochimica A

Fsa

att

tIab

�

wttN(i

�

piCisi

sp

r[sssebSEitbta 2ing all nuclear frequencies that can be detected along ν1 axis. TheSECSY contour maps and the projections along ν1 and ν2 areshown in Fig. 5 for Orgueil. Along ν1, fundamental and com-

ig. 4. Comparison of the first integral of the CW-EPR line, the two-pulse field-weep EPR (echo detected EPR) and the SECSY-EPR at room temperature (top)nd 10 K (bottom) of the IOM of Orgueil meteorite.

bove 95 GHz by the g-anisotropy [34]. We may thus concludehat a g-factor distribution (or anisotropy) is not responsible forhe inhomogeneous line in meteoritic IOM.

The second possible source of inhomogeneous broadening ishe occurrence of dipolar interactions between electron spins ofOM• centres. In a magnetically diluted medium, dipolar inter-ctions give a Lorentzian line with a peak-to-peak width giveny [35, p. 125]:

Bdippp = 4π2

9gβN ≈ 8 × 10−21 N (4)

here N is expressed in spin/cm3 and �Bdippp in mT. Since

he density of our samples is about 1 g/cm3, the concentra-ion can also be expressed in spin/g. From the concentration≈ 4 × 1019 spin/g found for Orgueil at room temperature, Eq.

4) gives �Bdippp = 0.32 mT, which is comparable with the exper-

mental value �Bpp = 0.38 mT. However, Eq. (4) implies that

Bdippp should decrease (N decreases, see Fig. 3b) at lower tem-

erature, which is not the case as the experimental linewidthncreases from 0.38 mT at 300 K to 0.80 mT at 4 K (see Fig. 3a).onsequently the strong discrepancy between Eq. (4) and exper-

mental results, show that dipolar interaction between electronpins cannot account for the observed inhomogeneous broaden-

ng.

The third possible mechanism is the broadening by unre-olved hyperfine interaction, which can be demonstrated inulsed-EPR by observing the modulation of the echo decay

Fopaf

cta Part A 69 (2008) 1301–1310 1305

esulting from the dipolar part of the hf interaction (ESEEM)23]. The relation between the ESEEM and the EPR line-hape can be studied with the SECSY experiment based on theequence π/2-t1-π/2-t1-t2. As the maximum of the echo inten-ity occurs at time t1 after the second pulse, the shape of thecho is recorded during time t2 for each t1 step, and the ensem-le of data is Fourier transformed to give a two-dimensionalECSY spectrum. Such experiments can be compared to 2P-SEEM echo decay versus field sweep. The main difference

n the SECSY sequence is a flip angle β = π/2 instead of π inhe 2P-ESEEM sequence. This takes advantage of increasingandwitdth excitation. Moreover SECSY sequence is often usedo measure variation of the electronic phase memory time Tmcross CW-EPR spectra along ν axis, the second pulse refocus-

ig. 5. Two-dimensional maps of spin-echo correlation spectroscopy (SECSY)f IOM in Orgueil meteorite at room temperature (top) and 9 K (bottom). Therojections along ν2 (EPR dimension) and along ν1 (ESEEM dimension) arelso represented. Modulation frequencies at the proton frequency νH and 13Crequency νC and their combinations at 2νH, 3νH and 2νC are indicated.

1 a Acta Part A 69 (2008) 1301–1310

bfaν

pl(

arttisst(atfiOwwatpthsttEtoectacaoim

3

apirameν

it

Fig. 6. Matched 3P-ESEEM (a) and 3P-ESEEM (b) spectra at 10 K of the IOMoπ

a

ts1oEddpcwaEτ

tostodc3lw

fciHi

306 O. Delpoux et al. / Spectrochimic

ination cross peaks are clearly observed at 9 K at hydrogenrequency and its harmonics νH = 14.6 MHz, 2νH = 29.2 MHznd 3νH = 43.8 MHz, and 13C cross peaks can be suspected atC = 3.7 MHz. These cross peaks are broadened along ν1 by thehase memory time Tm. At room temperature, nuclear modu-ations can be seen at νH, 2νH for hydrogen and 2νC for 13CFig. 5).

Of particular interest is the projection of the SECSY plotlong ν2, which represents the shape of the EPR spectrumesponsible for the nuclear modulations. The SECSY-EPR spec-ra are shown in Fig. 4 (in MHz) and Fig. 5 (in mT). At 9 K,he appearance of a Pake-doublet shape for the SECSY-EPRs different from the single line of the echo-detected EPR, ashown in Fig. 4b. This indicates that SECSY experiment haselected one component, hereafter referred to as IOM1

• cen-re, characterized by a large hf interaction with one hydrogenA = 23 MHz or 0.82 mT). As SECSY experiment probes Tmcross CW-EPR spectrum, the apparent doublet structure at lowemperature can be attributed to a species with a long Tm and suf-cient bandwidth excitation to be separated from other species.n the contrary, the echo-detected EPR, which integrates thehole echo, is much less resolved because of Tm and π band-idth limitations. Thus the EPR spectrum at low temperature

ppears to be the sum of a component with no apparent hf split-ing, referred to as IOM2

• centres, responsible for the majorart of the EPR line, and a minor component due to IOM1

• cen-res characterized by an hf interaction of 23 MHz with a singleydrogen atom (Fig. 4). Simulation of the echo-detected EPRpectrum shows that IOM1

• centres contribute by less than 10%o the total EPR line at low temperature. Upon increasing theemperature above 130 K, the CW-EPR and the echo-detectedPR become narrower as the result of the increasing contribu-

ion of TATS, characterized by a narrow line. This can be seenn the SECSY-EPR at room temperature (Figs. 4 and 5a), whichxhibits a narrow component (TATS) superimposed to a broaderomponent (IOM• centres). As TATS represent about 50% ofhe paramagnetic centres at room temperature (see Fig. 3b), thepparent line narrowing upon increasing the temperature is theonsequence of the growing contribution of TATS (narrow line)t the expense of IOM• centres (broader line), and not the effectf an increase of the Heisenberg exchange resulting from thencrease of TATS concentration, as previously deduced from

ulti-frequency CW-EPR spectroscopy [34].

.3. Structure of IOM• centres

The modulation of the primary echo generated in 2P-SECSYrising from hyperfine interaction refocused by the second π/2ulse, it clearly shows the presence of H and 13C hf interactionsn both Orgueil (Fig. 5) and Tagish Lake (not shown), whicheflect their relatively high H/C ratio, 0.72 and 0.46, respectively,nd confirm the attribution of IOM• centres to organic radicaloieties. The relatively poor resolution of two-pulse ESEEM

xperiments is due to the fact that the peak broadening along1 is determined by the phase memory time Tm. The resolutions better in 3P-ESEEM experiment as the width of the correla-ion peaks is determined by the much longer nuclear relaxation

alfa

f Orgueil meteorite. Time delay τ = 136 ns, pulse length of 16 ns was used for/2 pulses and matched pulse length tp = 72 ns. The modulations in time domainre shown in the inserts.

ime T2n (or T1 if T1 < T2n) [23]. Fig. 6 shows the 3P-ESEEMpectra of Orgueil at 10 K. In addition to the hydrogen peak at4.6 MHz, the 13C peak (I = 1/2, abundance 1.11%) is clearlybserved at its nuclear frequency 3.7 MHz in the standard 3P-SEEM spectrum (Fig. 6b). The weak peak at 2.2 MHz is likelyue to deuterium (I = 1, natural abundance on Earth 0.015%) asemonstrated below. Of particular interest is the fact that thiseak could not be detected at such low abundance, which indi-ates that IOM• centres are considerably enriched in deuterium,ith D/H ratio of at least 1% to be detected. The role of hydrogen

s the principal component of the unresolved hf interaction of thePR line is shown with the matched 3P-ESEEM sequence (π/2--tp-T-tp-τ-echo), whereby the high turning angle tp is matchedo proton nuclear frequency (72 ns). It enhances the detectionf protons by defocusing electron coherence on allowed tran-itions and refocusing on forbidden transitions during the firstransfer. During the time evolution, a double forbidden transferccurs with the detection of allowed transition [36]. The resultsisplays in Fig. 6a clearly show a pure proton modulation that isonsiderably enhanced compared to those detected in standardP-ESEEM sequence. This clearly confirms that the broadeningine effect observed is due to unresolved hyperfine interactionith hydrogen.More precise information on IOM• centres was obtained

rom measurement of hf interactions by hyperfine sublevelorrelation spectroscopy (HYSCORE). This 2D spectroscopys powerful to disentangle congested ESEEM spectra [23,30].YSCORE contour maps for Orgueil and Tagish Lake are shown

n Figs. 7 and 8, respectively. The contribution of different nuclei

re clearly identified by ridges extending along independentines ν1 = 2νn − ν2 crossing the diagonal ν1 = ν2 at the nuclearrequency νn. The temperature dependence of the proton hf inter-ction, which dominates the EPR lineshape and is responsible

O. Delpoux et al. / Spectrochimica Acta Part A 69 (2008) 1301–1310 1307

Fig. 7. HYSCORE spectra at 9 K (a and b) and room temperature (c) of theIOM of Orgueil meteorite. Pulse lengths of 28 ns and 56 ns were used for π/2and π pulses, respectively. (b) Extended view of the low frequency part of thespectrum at 9 K, showing the respective contributions of 13C, D, 15N and 14Nhighlighted by dotted lines.

Fig. 8. HYSCORE spectra at 9 K of the IOM of Tagish Lake meteorite. Pulselengths of 28 ns and 56 ns were used for π/2 and π pulses, respectively. (a)Goo

fttraν

wpapoonsmtt

eneral view of the spectrum (τ = 136 ns) and (b) extended view at τ = 184 nsf the low frequency part of the spectrum, showing the respective contributionsf 13C and D.

or the line narrowing at high temperature, can be observed onhe proton spectra obtained from slice of the HYSCORE alonghe proton ridge. Fig. 9a shows such slices for Orgueil at 10 K andoom temperature, obtained from the contour maps of Fig. 7and c. At 10 K the spectrum is the sum of a narrow peak atH = 14.6 MHz due to distant hydrogen atoms weakly interactingith the electron spin, and a broad superposition of correlationeaks due to anisotropic hf interactions with several hydrogentoms of IOM• centres. At room temperature only the centraleak of distant H can be clearly observed. As the contributionf TATS to the EPR spectrum is important at room temperature,ne can deduce that contrary to IOM• centres, TATS centres doot possess hydrogen atoms in their structure, and the electron

pin interacts only with distant H atoms. Owing to the polyaro-atic nature of paramagnetic centres in IOM, it is thus possible

o conclude that IOM• centres possess aromatic hydrogen, i.ehey are moderately branched with aliphatic chains (see Fig. 1),

1308 O. Delpoux et al. / Spectrochimica Act

FIF

wa

wetfosfaasrinvl

oaposogl

|rpai

A

wQaaitEisditTc[wL≈1

3

aFcomgpigotsbaeHotctt

ig. 9. Hydrogen (a) and 13C (b) spectra at 9 K and room temperature of theOM of Orgueil meteorite, obtained from slices along the HYSCORE ridges ofig. 7.

hile TATS are highly branched so that all boundary sp2 carbonsre linked to aliphatic chains.

This interpretation is confirmed by 13C-HYSCORE spectra,hich appear clearly on the contour maps (Figs. 7 and 8). The

volution of the 13C spectrum is better evidenced on slices alonghe carbon ridge (Fig. 9b), which show that the spectra are dif-erent at 10 K and at room temperature. Fig. 9b shows the casef Orgueil. Only IOM• centres contribute at 10 K, while thepectrum at room temperature is the sum of the contributionsrom both IOM• centres (60%) and TATS (40%). The satellitest ±3 MHz from 13C frequency, which appear at room temper-ture are due to TATS while the central part of the HYSCOREpectrum is due to IOM• centres. The very good signal-to-noiseatio of 13C-HYSCORE at room temperature is a convincingndication that the lack of proton ridge at room temperature isot only the result of a loss of sensitivity arising from the 1/Tariation of EPR intensity, but is also the consequence of theack of hydrogen atoms in the structure of TATS.

The distribution of correlation peaks along the hydrogen ridgef IOM• centres is a consequence of the highly disordered char-cter of IOM, with a distribution of size and environment of theolyaromatic radicals. However, structural information can bebtained from the tips of the ridge, where HYSCORE inten-

ity falls to zero. They give the hf interaction at a canonicalrientation of the magnetic field for the most coupled hydro-en atom. The edges of the proton ridge correspond to theargest hf component |HAmax| ≈ 12.3 ± 0.2 MHz for Orgueil and

s

|

a Part A 69 (2008) 1301–1310

HAmax| ≈ 10.2 ± 0.2 MHz for Tagish Lake. In a polyaromaticadical with a hydrogen atom on the ith carbon, the three com-onents k = x, y, z of the hf interaction with the ith hydrogenre the sum of the isotropic interaction ρπ,iQCH and the dipolarnteraction ρπ,iA

CHk :

ik = ρπ,i(Q

CH + ACHk ) (5)

here ρπ,i is the spin density in the ith pz carbon orbital,CH = −67.3 MHz to −84.1 MHz is the Mc Connell factor forneutral •CH fragment [37], and ACH

k = −2.8 MHz, −36 MHznd +38 MHz for the k = x, y and z components of the dipolarnteraction, respectively [38]. The directions x and z correspondo the axis of the pz orbital and the C H bond, respectively. Fromq. (5), the largest hf component |HAmax| of an aromatic H lies

n the range 103ρπ,i ≤ |HAmax| ≤ 120ρπ,i (in MHz), and corre-ponds to the orientation y, with B0 perpendicular to the C Hirection and the pz orbital. The experimental value of |HAmax|ndicates that the spin densities on carbon atoms of IOM• cen-res are characterized by ρπ ≤ 0.12 for Orgueil and ρπ ≤ 0.1 foragish Lake, as deduced from Eq. (5). These values are typi-al of small aromatic clusters such as perylene and coronene31]. This interpretation is reinforced by the 13C hf interaction,hich is clearly observed in Figs. 7 and 8 for Orgueil and Tagishake, respectively. It amounts ≈6.1 ± 0.1 MHz for Orgueil and3.3 ± 0.1 MHz for Tagish Lake. These values also agree with

3C hf values for small polyaromatic radicals [31].

.4. Evidence of deuterium enrichment

It is important to note that deuterium correlation peaksre easily observed in HYSCORE spectra, as shown inigs. 7b and 8b. This indicates that the D/H ratio in IOM•entres is larger than about 0.5% in Orgueil and Tagish Lake,therwise deuterium could not be observed. A more accurateeasurement using a reference sample with known D/H ratio

ives D/H = 1.5 ± 0.5 × 10−2 [39]. This value is very high com-ared to that of protosolar hydrogen (D/H = 25 ± 5 × 10−6), ands close to the mean value D/H ≈ 2 × 10−2 of the interstellaras phase [40,41]. Such extreme deuterium enrichment of mete-ritic IOM could indicate that IOM• centres are a heritage ofhe interstellar medium pre-dating the formation of the solarystem. However, this interpretation implies a statistical distri-ution of deuterium atoms on H sites of IOM• centres becausell possible H position would be partially deuterated. However,xamination of hydrogen and deuterium HYSCORE shows that

are distributed on several sites as shown by the weak maximan the proton ridge at 10 K (Fig. 9a), while deuterium exhibitwo well identified maxima that can be directly observed on theontour plot (see Fig. 8b for Tagish Lake) or on slice throughhe deuterium ridge (case of Orgueil, not shown). This indicateshat deuterium occupies predominantly one specific site.

Deuterium substitution of a hydrogen atom in IOM• centres

hould give hf interaction related to the proton hf interaction by

DA| = |HA|gn(D)

gn(H)= 0.153|HA| (6)

ica A

wgtO|wliwqidrhdopntrt[

4

gtm

(

(

(

(

(

A

v

R

[

[

[

[[

[[

[

[

[[

[

[

[[

[

[

[

[[

O. Delpoux et al. / Spectrochim

here gn(D) = 0.85744 and gn(H) = 5.58569 are the nuclear-factors of deuterium and hydrogen, respectively. Fromhe experimental value for the most coupled hydrogen ofrgueil, |HAmax| = 12.3 MHz, we expect a hf interaction

DAmax| = 1.9 MHz for deuterium. This value fairly agreesith the experimental value |DAmax| = 2.0 ± 0.3 MHz. Simu-

ation of the HYSCORE spectrum (not shown) of Orgueilndicates that this hyperfine interaction is mainly isotropic,ith Aiso ≈ 1.9 MHz, Adip ≈ 0 MHz and P = 0.1 MHz for theuadrupolar interaction. The same result is obtained for Tag-sh Lake, with an experimental hf interaction 1.9 ± 0.3 MHz foreuterium. Thus we may conclude that deuterium atoms haveeplaced H atoms only at the position that exhibits the largestf interaction. The cosmochemical consequences of the extremeeuterium enrichment of IOM• centres and of the localizationf deuterium at a specific C H bond are discussed in anotheraper [39,42]. The fact that these centres are very inhomoge-eously distributed and concentrated in defect-rich regions ofhe IOM [26] implies that they constitute also deuterium-richegions. This could provide a simple explanation to the observa-ion of deuterium “hot-spots” in meteoritic IOM by NanoSims24].

. Conclusion

Combination of CW-EPR and pulsed-EPR spectroscopiesives a more precise description of paramagnetic centres inhe insoluble organic matter (IOM) of primitive meteorites. The

ain conclusions of this study are the following:

1) Meteoritic IOM contains at least two types of π-type, poly-cyclic paramagnetic centres: IOM• centres with S = 1/2, andTATS centres with S = 0 ground state and thermally acces-sible triple state S = 1 [33]. All these paramagnetic centresare concentrated in defect-rich regions of the IOM, withconcentrations of the order of 4 × 1019 spin/g [26].

2) In spite of the Lorentzian shape and the line narrowing uponincreasing the spin concentration, the EPR line is not inthe exchange-narrowing regime as previously deduced frommulti-frequency CW-EPR [34]. The presence of a spin-echowith nuclear modulation of its envelope demonstrates thatthe EPR line is inhomogeneously broadened by unresolvedhyperfine interactions with hydrogen nuclei.

3) Comparison of echo-detected EPR and SECSY-EPR showsthat the EPR line of TATS is narrower than that of IOM•centres. The consequence is a progressive line narrowingsimilar to an exchange narrowing when the proportion ofTATS increases upon raising the temperature.

4) Hyperfine sublevel correlation spectroscopy (HYSCORE)of hydrogen and 13C nuclei indicates that IOM• centresare small polycyclic moieties that are moderately branchedwith aliphatic chains, as shown by the presence of aromatichydrogen atoms. On the contrary the lack of such aromatic

hydrogen in TATS suggests that these radicals are mostprobably highly branched.

5) IOM• centres are considerably enriched in deuterium, withD/H ≈ 1.5 ± 0.5 × 10−2 of the order of D/H existing in inter-

[

[

cta Part A 69 (2008) 1301–1310 1309

stellar medium. The fact that Deuterium enriched IOM•centres are concentrated in defect-rich regions of the IOMimplies the existence of deuterium “hot-spots”, which havebeen observed by NanoSims [24].

cknowledgment

The authors gratefully acknowledge Luann Becker for pro-iding us the IOM from Tagish Lake meteorite.

eferences

[1] J.M. Hayes, Geochim. Cosmochim. Acta 31 (1967) 1395–1440.[2] M.A. Sephton, Nat. Prod. Rep. 19 (2002) 292–311.[3] C.F. Chyba, C. Sagan, Nature 355 (1992) 125–132.[4] J.R. Cronin, S. Pizarello, D.P. Cruikshank, in: J.F. Kerridge, M.S. Matthews

(Eds.), Meteorites and the Early Solar System, University of Arizona Press,Tucson, 1988, pp. 819–857.

[5] M. Maurette, Orig. Life Evol. Biosph. 28 (1998) 385–412.[6] G.D. Cody, C.M.O.D. Alexander, Geochim. Cosmochim. Acta 69 (2005)

1085–1097.[7] A. Gardinier, S. Derenne, F. Robert, F. Behar, C. Largeau, J. Maquet, Earth

Planet. Sci. Lett. 184 (2000) 9–21.[8] M.A. Sephton, C.T. Pillinger, I. Gilmour, Planet. Space Sci. 47 (1999)

181–187.[9] S. Derenne, J.N. Rouzaud, C. Clinard, F. Robert, Geochim. Cosmochim.

Acta 69 (2005) 3911–3918.10] L. Remusat, S. Derenne, F. Robert, Geochim. Cosmochim. Acta 69 (2005)

4377–4386.11] A. Marchand, J. Conard, in: B. Durand (Ed.), Kerogen-Insoluble Organic

Matter from Sedimentary Rocks, Technip, Paris, 1980, pp. 224–270.12] J.W. Schopf, A.B. Kudryavtsev, D.G. Agresti, T.J. Wdowiak, A.D. Graja,

Nature 416 (2002) 73–76.13] J. Uebersfeld, A. Etienne, J. Combrisson, Nature 174 (1954), 614-614.14] T.A. Dickneider, S. Scull, J.K. Whelan, N.V. Blough, Org. Geochem. 26

(1997) 341–352.15] P.I. Premovic, Naturwissenschaften 69 (1982) 479–482.16] H.J. von Bardeleben, J.L. Cantin, K. Zellama, A. Zeinert, Diam. Relat.

Mater. 12 (2003) 124–129.17] D.J. Keeble, K.M. Robb, G.M. Smith, H. El Mkami, S.E. Rodil, J. Robert-

son, J. Phys. Condens. Matter 15 (2003) 7463–7468.18] D. Gourier, L. Binet, A. Skrzypczak, S. Derenne, F. Robert, Spectrochim.

Acta Part A 60 (2004) 1349–1357.19] J. Duchesne, J. Depireux, C. Litt, Geochem. Int. 1 (1964) 1022–1024.20] A.P. Vinogradov, G.P. Vdovykin, I.N. Marov, Geochem. Int. 1 (1964)

395–398.21] K.F. Schulz, R.M. Elofson, Geochim. Cosmochim. Acta 29 (1965)

157–160.22] L. Kevan, L.D. Kispert, Electron Spin Double Resonance Spectroscopy,

John Wiley, New York, 1976.23] A. Schweiger, Angew. Chem. 30 (1991) 265–292.24] H. Busemann, A.F. Young, M.C.O’D. Alexander, P. Hoppe, S. Mukhopad-

hyay, L.R. Nittler, Science 312 (2006) 727–730.25] B. Durand, G. Nicaise, in: B. Durand (Ed.), Kerogen, Technip, Paris, 1980,

p. p. 35.26] L. Binet, D. Gourier, S. Derenne, F. Robert, Geochim. Cosmochim. Acta

66 (2002) 4177–4186.27] L. Binet, D. Gourier, S. Derenne, S. Pizarello, L. Becker, Meteorit. Planet.

Sci. 39 (2004) 1649–1654.28] E.S. Barghoorn, S.A. Tyler, Science 147 (1965) 563–577.29] B.R. Patyal, R.H. Crepeau, D. Gambil, J. Freed, Chem. Phys. Lett. 175

(1990) 445–452.30] P. Hofer, A. Grupp, H. Nebenfurh, M. Mehring, Chem. Phys. Lett. 132

(1986) 279–282.31] T. Ikoma, O. Ito, S. Tero-Kubota, K. Akiyama, Energ. Fuels 12 (1998)

1363–1368.

1 a Act

[[

[[

[

[

[

[

[40] J. Geiss, G. Gloecker, Space Sci. Rev. 84 (1998) 239–250.

310 O. Delpoux et al. / Spectrochimic

32] C.P. Slichter, Principles of Magnetic Resonance, Springer, Berlin,1992.33] L. Binet, D. Gourier, S. Derenne, F. Robert, I. Ciofini, Geochim. Cos-

mochim. Acta 68 (2004) 881–891.34] L. Binet, D. Gourier, Appl. Magn. Reson. 30 (2006) 207–231.35] A. Abragam, Principles of Nuclear Magnetism, Clarendon Press, Oxford,

1961.36] G. Jeschke, R. Rakhmatullin, A. Schweiger, J. Magn. Reson. 131 (1998)

261–271.37] F. Gerson, W. Huber, Electron Spin Resonance Spectroscopy of Organic

Radicals, Wiley-VCH, Weinheim, 2003.

[

[

a Part A 69 (2008) 1301–1310

38] W. Gordy, Theory and Application of Electron Spin Resonance, John Wiley,New York, 1980.

39] D. Gourier, F. Robert, O. Delpoux, L. Binet, H. Vezin, A. Moissette, S.Derenne, submitted for publication.

41] P.R. Mahaffy, T.M. Donahue, S.K. Atreya, T.C. Owen, H.B. Niemann,Space Sci. Rev. 84 (1998) 251–263.

42] L. Remusat, F. Robert, S. Mostefaou, A. Meibom, O. Delpoux, L. Binet,D. Gourier, S. Derenne, Geochim. Cosmochim. Acta 75 (2007) A832.