Electron microscopy, especially High-Resolution Transmission Electron Microscopy (HRTEM), allows the observation of aromatic layers in the IOM. Indeed when aromatic planes are quasi parallel to the incident electron beam (i.e. under the Bragg angle) and separated by a least 0.14 nm (the resolution of this microscope), fringes appear due to interferences between the transmission electron beam and the diffracted beams and are representative of the profile of the aromatic planes [48,49]. Then aromatic layers appear as fringes in HRTEM images. Image analysis allowed the determination of the properties of aromatic units. In Orgueil and Murchison, aromatic units primarily consist of 2 to 3 aromatic rings in diameter, i.e. 2 to 7 rings in each unit [17].

This observation is consistent with ¹³C solid-state NMR [14,20]. NMR is useful to determine the chemical environment of atoms (Fig. 1). Spectra deconvolution allows the determination and quantification of different types of carbon. These data reveal that IOM is highly aromatic (70% of carbon atoms are aromatic), with rather small aromatic units. These aromatic units are linked by short and branched aliphatic chains (see next section for more details). Furthermore, aromatic compounds are the main aromatic products released by pyrolysis [41,55,57], reinforcing previous assumptions on the aromatic behaviour of IOM.

Pyrolysis releases mainly aromatic compounds, ranging from one to seven rings [31,41,56], with methyl and ethyl substitutions. These products are characterized by a great diversity: all possible isomers are detected and no isomeric preference is observed. This property is shared with soluble compounds. This is likely related to an organosynthesis by carbon addition and aromatization of aliphatic chains to form aromatic units. It must be noted that no aliphatic compound is detected by pyrolysis, revealing that aliphatic chains are shorter than 8 carbon atoms.

The occurrence of moieties with thermally accessible triplet states was reported in the IOM of Orgueil and Murchison [5]. These moieties manifest themselves in Electron Paramagnetic Resonance (EPR) by an increase in spin concentration when the temperature is increased above 150 K (Fig. 2). This feature, not observed for terrestrial organic matter, was attributed to the occurrence of diradicaloids, i.e. aromatic moieties with two unpaired electrons radicals that can be stabilized by delocalisation of electrons over the aromatic rings. By theoretical calculation, Binet et al. [5] proposed that these species could be related to aromatic units of 10 to 15 rings, making units of 3 to 4 rings in diameter, consistent with the previous estimation.

A chemical degradation technique, ruthenium tetroxide oxidation, was chosen to investigate the aliphatic chains in the IOM of Orgueil and Murchison. Chemical degradations were performed on IOM from meteorites as far back as 1966 with the ozonolysis of Orgueil [7]. Moreover, following up this pioneering work, a number of other reagents were used in chemical degradations of IOM. They were trifluoroacetic acid, nitric acid, dichromate, copper oxide [25,26]. However, as far as we are aware of, no chemical degradation has been performed on IOM since that time. Moreover, except depolymerisation with trifluoroacetic acid, which was reported to yield C2 to C6 alkenes from Murchison IOM, no information on the aliphatic linkages could be derived from these chemical degradations.

Ruthenium tetroxide oxidation is a chemical degradation that has been largely used for characterizing the aliphatic and alicyclic portion of coals [30,58], kerogens [9,32] and humic substances [24] during the last 20 years. Ruthenium tetroxide is a mild and selective oxidising agent that preferentially destroys aromatic rings, converting them into CO₂. It was also shown that it converts, oxidatively, alcohols into ketones or aldehydes, aldehydes into acids, ethers into esters or lactones [28,34]. The aliphatic and acyclic structures liberated appear as carboxylic acids the carboxylic functional group marking points of attachment in the kerogen and positions of labile functional groups, such as carbon-carbon double bonds and ether links [58].

Ruthenium tetroxide oxidation of Orgueil and Murchison IOM [40] produces a series of α,ω-dicarboxylic acids ranging from C2 to C9, along with methyl and dimethyl derivatives for the shortest compounds, without any isomeric selectivity. These compounds are related to the aliphatic linkages between the polyaromatic units in IOM. It also produces tricarboxylic acids, which are related to branched linkages between three aromatic units. No monocarboxylic acid is detected, pointing to a small size of the side chains into the IOM. Recently, Huang et al. have used a modified protocol of ruthenium tetroxide oxidation on Murchison IOM and they have recovered monocarboxylic acids from 2 to 9 carbons atoms [27], but the isotopic ratios of the longer ones should indicate terrestrial contamination. It seems likely that only the shorter ones (shorter than C5), recovered along with their methyl and dimethyl derivatives, can be considered as indigeneous products from Murchison IOM, consistently with previous data [40].

From these products, it is possible to characterize the aliphatic chains in chondritic IOM. They are quite short: less than four carbons for side chains, from two to seven carbons for aliphatic bridges between aromatic units. Oxygen is included in these aliphatic chains within ester or ether functional groups. Aliphatic linking chains are substituted by methyl and ethyl groups. These substitutions are randomized as there is an isomeric diversity. Moreover, the occurrence of aliphatic chains linking several aromatic units is consistent with a high degree of cross-linking evidence by ¹³C solid-state NMR.

EPR reveals that radicals in Orgueil, Murchison and Tagish Lake IOM are heterogeneously distributed [4,6]. This feature is not observed in terrestrial organic samples and is specific to chondritic IOM. As shown on Fig. 3, radical concentration and the relaxation time of these radicals do not follow the trend of terrestrial organic matter. This means that radicals relax as if the concentration were higher. The only way to explain this observation is to propose that radicals are organized as clusters with high local concentrations compared to the bulk IOM.

All the previous results are summarized in Fig. 4. It represents the main information we have on the chemical structure of IOM of primitive chondrites. Nevertheless, for building a reliable molecular model, numerical molecular parameters are needed. Statistical parameters, obtained from several studies, are reported in table 1. Considering these parameters and the chemical functions evidenced in the IOM, it should be possible to draw a molecular model. This should be the goal of a future work.

These molecular data are nevertheless useful to understand the processes involved in the synthesis and evolution of the IOM. The aromatic units that constitute the chemical structure are related to large PAHs observed in the interstellar medium. Nevertheless, the aromatic units in IOM are smaller than interstellar PAHs (Fig. 5). This could be the result of a selective preservation of smaller PAHs in IOM prior to their destruction by UV-irradiation, inducing a shift in the distribution of the size of aromatic units towards the high values. This process is shown schematically in Fig. 5.

The oxygen-containing functions are likely related to the aqueous alteration experienced by primitive carbonaceous chondrites. Indeed, warm water circulation could have resulted in an oxidation of IOM leading to the formation of ether or ester functional groups. Thanks to solid-state ¹⁵N NMR, it was shown that nitrogen is contained in aromatic moieties as pyrroles functions [41]. The sequestration of nitrogen into aromatic units should prevent IOM from being the source of amino acids. The only way to produce amino acids from IOM is to bring another reservoir of nitrogen, for instance HCN, and to involve catalytic processes [33].

Pulsed Electron Paramagnetic Resonance has been used to analyze the isotopic composition of radicals in IOM of the Orgueil meteorite [16]. It is shown that radicals, which are heterogeneously distributed in the IOM, contain a very high deuterium concentration (D/H = 15000 ± 5000 × 10⁻⁶), much larger than the average bulk value D/H = 350 × 10⁻⁶ for this meteorite. This D-enrichment is comparable to the maximum enrichments observed in organics in the ISM (45,000 to 60,000 × 10⁻⁶ [36]).

The recent use of NanoSIMS ion probe provides new information on the distribution of the isotopic ratios in chondritic IOMs [12,47] with a spatial resolution of ca 100 nm. We have recently studied Orgueil IOM by NanoSIMS imaging of H, D, ¹²C, ¹⁴N and ¹⁶O. Maps of D/H, C/H, N/C and O/C were established to determine possible correlations between D/H isotopic ratio and any chemical or molecular feature. As previously observed for Murchison IOM [12,47], the D/H ratio is heterogeneous in Orgueil IOM, with small areas very enriched in D, with D/H ranging from 412 ± 30 to 723 ± 26 × 10⁻⁶ (see Fig. 6). In contrast, the bulk value is about 220 ± 22  × 10⁻⁶. These D-rich hot spots represent ≈ 2.8% of the IOM area. These hot spots are distributed all over the sample, without any specific pattern. No clear trend is observed between the D/H in these hot spots and the classical molecular parameters (for instance C/H, see Fig. 7) moreover, no association with inorganic acid-resistant minerals (mapped thanks to the O image) is observed.

We have also acquired images of Kainsaz (CO3) IOM and a 600 °C pyrolysis residue of Murchison IOM. Kainsaz IOM is known to be depleted in radicals (between 2 and 3 orders of magnitude less compared to Orgueil and Murchison IOM), likely because it has experienced thermal stress on the parent body [43]. The pyrolysis residue of Murchison IOM should represent a thermally processed sample of Murchison IOM. As no hot spots can be detected in those two samples, it clearly shows that D-rich hot spots are thermally sensitive, as one-hour heating of Murchison IOM at 600 °C made them vanish. In the case of Kainsaz IOM, two hypotheses could be proposed: (1) the organic precursor of the Kainsaz parent body was the same as that of Orgueil and Murchison ones, so the thermal metamorphism that has experienced the Kainsaz parent body [8] should have destroyed the carrier of D-rich anomalies, or (2) there is no common precursor and Kainsaz IOM never contained this carrier. The first hypothesis seems more likely and implies that thermal effects are responsible for the loss of the carrier of D-excess in chondritic IOM. It must be noted that organic radicals are known to be destroyed by heating stress.

Two different hypotheses could be proposed to account for the D-rich hot spots: (1) these hot spots are remnants grains of interstellar organic material or (2) they are related to organic radicals occurring in IOM, as they are known to be highly enriched in D. By combining data from NanoSIMS and pulsed EPR, it is possible to prove that all the D-excess is borne by organic radicals. This will be described in more details in an upcoming paper. As these radicals are concentrated in small areas, they define small regions in the IOM which could constitute D-rich hot spots. Moreover, these radicals have no specific molecular properties compared to the rest of the IOM [5]. So it is consistent to observe that C/H atomic ratios of D-rich hot spots are quite close to bulk values [12] as shown in Fig. 7.

GC-irMS (Gas Chromatography-isotope ratio Mass Spectrometry) permits the determination of the D/H ratio of aliphatic and aromatic moieties in Orgueil IOM [42]. Although aromatic moieties in chondritic IOM are known to be easily released by pyrolysis [41], the aliphatic linkages are recovered by ruthenium tetroxide oxidation [40]. Compound-specific D/H isotope measurements can then be performed on both the pyrolysis products and on the acids after derivatisation into their trimethylsilyl (TMS) esters.

Three types of organic groups (Fig. 8) hold most of the H in IOM. Type 1 is benzylic H, i.e. H bound to the first C atom linked to an aromatic ring (C atom in α position). Type 2 is non-benzylic aliphatic H. Type 3 is aromatic H. In order to understand the D/H distribution among the different products recovered during pyrolysis and chemical degradation (Table 2), three types of H were thus assumed to exhibit three different isotope compositions. For each compound the proportion of hydrogens of type 1, 2 and 3 was calculated. For this, it must be assumed that the recovered products bear the isotope labeling of the starting macromolecule before any laboratory degradation (Fig. 8); in other words, we must consider the location of H into the starting macromolecule and not in the recovered products. The C–H bond dissociation energy differs for the three types: the higher this energy, the stronger the bond. Any exchange between the hydrogen in these molecules and an external reservoir is affected by the binding energy. A stronger bond will undergo a lesser exchange and C–H bond-breaking is a prerequisite to D incorporation. The higher reactivity of the benzylic bond is well known and is reflected by a lower C–H bond dissociation energy than for the other aliphatic C–H bonds (87 kcal mol⁻¹ and 98 kcal mol⁻¹ for benzylic and aliphatic C–H bonds respectively). The C–H aromatic bond dissociation energy is even higher (111 kcal mol⁻¹) than for the aliphatic C–H bonds, independently of the size of the aromatic structure. So Type 1 H is supposed to be more exchangeable than Type 2, and Type 2 more exchangeable than Type 3.

D/H isotopic ratio of each analyzed compound can be calculated by considering the number of each type of hydrogen in each compound. This gives a mass balance equation for each analyzed compound with only three free parameters: the D/H ratio of Types 1, 2 and 3 hydrogen. By comparing iteratively between calculated and measured D/H values, D/H = 350 × 10⁻⁶, 240 × 10⁻⁶, 180 × 10⁻⁶ for hydrogen of Type 1, 2 and 3, respectively [42]. The good fit between measured and calculated D/H values is still observed if these Type 1, 2 and 3 D/H values are allowed to vary within ±10 × 10⁻⁶. This numerical result does not preclude the presence of other minor types of organic H, having D/H ratio different from those of Types 1, 2 or 3, but it demonstrates that, if present, these other types of H cannot be detected within the limit of the error bars on the D/H ratio. Such minor types of organic hydrogen could be the products of hydrothermal metamorphism on the parent body [11] and might be present as functional groups not detected in our experiments (as for example –OH groups).

The molecular determination of the D carrier in Orgueil IOM can be summarized in Fig. 9. Pulsed EPR revealed that, in the D-rich radicals, D is preferentially linked to the carbon atom in α-position relative to an aromatic unit, i.e. it is a benzylic hydrogen. This helps to determine the C–H bond energy for these radicals. We postulate that the IOM was a PSN product. In this pristine IOM, the organic sites had D/H ratios with values lying somewhere between the PSN and the D/H ratio of water (given by the –OH bearing minerals present in Orgueil [45]). After its synthesis, the IOM was transported into regions of the PSN where the physical conditions were comparable to the ISM, i.e. a cold ionized medium where ion–molecule reactions [50] could take place and where IOM acquired its deuterium enrichment. In this respect, ionized molecules (and more specifically the most abundant ion in such an environment, i.e. H₃⁺) rapidly become enriched in D even at low T [36]. Under these conditions, Type 1 hydrogens experienced a higher deuterium exchange rate than Type 2, and Type 2 higher exchange rate than Type 3, resulting in the observed relationship (Fig. 9). The radicals, exchanging even faster than the rest of the IOM should get more D-enriched than the bulk IOM, and as they are heterogeneously distributed in the IOM, could form D-rich small areas constituting the D-rich hot spots. Because the water D/H ratio also lies on the correlation defined by the three types of organic hydrogen (see Fig. 9), water might have acquired its deuterium enrichment under similar conditions.

Our model can be illustrated by the following reactions:

Taking the classical value for the exothermicity (ΔE/k = 260 K) of the reaction (1), the enrichment factor α in D can be approximated as: α = exp(260/T). No isotopic fractionation is assumed for the reactions (2) and (3); these reactions simply transfer deuterium. In an isotopic exchange between IOM and D-enriched H₃⁺, the maximum measured D/H ratio in IOM (i.e. 15,000 × 10⁻⁶) will represent the minimum possible D/H ratio of H₃⁺. This value gives α = (15,000 × 10⁻⁶/25 × 10⁻⁶) = 600, corresponding to T ≈ 40 K. These temperatures very likely occurred at the outer edges of the PSN as in the ISM. Recently, Ceccarelli and Dominik [13] have reported theoretical models that predict the occurrence of highly deuterated H₃⁺ in protoplanetary disks, which could exchange its D with IOM in the protosolar nebula.

Thus, these data support the theoretical idea according to which the deuterium enrichment of solar system molecular species could have taken place in the outer part of the turbulent protosolar disk ionized by UV light from the young Sun [1,23,44]. If, as we suggest, these organic molecules are contemporaneous with our solar system, one can easily imagine that they are products of protostellar disks in general. Indeed, because the D/H ratio of meteoritic water is consistent with our model, these processes could be at the origin of the deuterium enrichment in the inner solar system.

It must be noted that in the present interpretation, the isotopic species need not to reach equilibrium: the correlation is linked only to a property of isotopic exchange rates. Due to very low temperature, organic matter should rapidly condense into grains. This process should likely stop the D exchange as organics trapped into the ices cannot be reached by the ionized gas. This implies that the duration of the exchange reaction is limited, leading to a competition between each type of hydrogen. The dynamical implications of these ideas need to be explored further in future theoretical work.

This process is attractive because it explains both the occurrence of D-rich hot spots and D/H heterogeneity observed at molecular level, revealed by GC-irMS. Then the major issue that remains should be the origin of organic radicals. Indeed the occurrence of stable radicals in IOM is not yet explained. For instance it is not known if they were formed in or before the protosolar nebula in other words if they constitute presolar organic matter or if they were formed by a protosolar process.