Obviously, from an astrobiological point of view, organic compounds containing C, H, N, O, S, and P, are of primary astrobiological interest [10,11,12,13,14]. Soluble organic matter (SOM) can be extracted from powdered meteorite by common solvents (e.g., methanol or dichloromethane) and these extracts can be ionized by electrospray ionization in both the positive and negative ion modes. Nevertheless, inorganic content, such as the presence of mass deficient MgSO₄, Na salts, and minerals can also be identified in otherwise organic rich chondrites [10].

For illustration, we present here some of our recent results using the ultrahigh resolution LTQ-Orbitrap (Grenoble, France) to analyze the soluble organic matter from a Martian meteorite (Figure 1). Small meteor rocks were ground and a methanol:toluene (1:2) solvent mixture was used for extraction. After centrifugation, direct infusion ESI was used in the positive ion mode to generate ions. The spectrum in Figure 1A contains a large number of peaks showing hundreds of components. Note that the vertical (intensity) axis is logarithmic in Figure 1A so the intensity range shown is 1–50,000. A mass defect versus exact mass (MDvEM) plot (for appropriate references, see Materials and Methods below) of the same and intricate MS is shown in Figure 1B. MDvEM analysis reveals immediately the very high degree of organization in the datasets, with obvious series in CH₂, but also addition of minerals, due to the interaction within the complex matrix (Figure 1B). Intensity dimension is color coded while the vertical axis plots the difference of one exact mass to the nearest integer. Repetition of stoichiometric patterns generates lines in this diagram. Here, the most visible oblique lines are due to the multiple addition of CH₂.

The insoluble organic material (IOM) consists mostly of unsaturated polycyclic macromolecules that are difficult to ionize by ESI. Thus, the availability of dual MALDI/ESI sources (such as on the Bruker Apex Qh or SolarixR instruments) is incredibly useful. We detected organic molecules (ions) using simply laser desorption ionization (LDI) of Orgueil. In general, our results confirmed that ESI and LDI are complementary ionization techniques that probe different parts (organic content) of chondritic meteorites [10] (data not shown, but see other LDI results below).

Recent LTQ-Orbitrap studies also revealed that carbonaceous meteorites (such as Murchison and Lonewolf Nunatks 94,103 meteorites) contain a wide range extraterrestrial nucleobasis. [11] Tandem MS/MS fragmentation and comparison with reference standard provided unambiguous proof for the presence of purine, adenine, 2,6-diaminopurine and 6,8-diaminopurine that can be considered as components of an “extended genetic alphabet”.

Note here that although FT-MS measurements (Orbitrap or ICR) provide reliable chemical compositions, it is difficult to distinguish between structural isomers based solely on (ultrahigh resolution) tandem MS/MS fragmentation. Other techniques, such as nuclear magnetic resonance (NMR) are extremely useful and improve the information content of analysis. A world leader in these combinative studies is a group led by Schmitt–Kopplin in the Helmholz–Zentrum in Munich, Germany [12,13]. Their recent studies include the one on Soltmany ordinary chondrite [12] and Murchison meteorite [13].

For correctness, we also note that conventional low resolution MS coupled with chromatography is still justified to use and provides useful information on organic content. An interesting recent example is a study when NWA, Gold Basin, Orgueil, and Al Haggounia meteorite powders were used to catalyze synthesis of nucleosides and other prebiotic compounds from formamide under proton irradiation [14]. Under these hostile conditions, about 30 different organic compounds have been detected by GC-MS. Because most of these compounds are not volatile enough derivatization was necessary before the analysis [14].

Based on the literature and our own experience, we can conclude that ultrahigh resolution MS (and tandem MS/MS) experiments are important to reveal detailed chemical information on extraterrestrial organic matter delivered to the Earth by meteorites. Although ESI is definitely a powerful ionization method, for the analysis of certain insoluble organic material we strongly recommend the use of laser desorption ionization (LDI) as a complementary ionization method (for other examples, related to laboratory “tholins”, see below).

Atmospheric processes have long been known to generate organic molecules. As mentioned in the Introduction, Titan’s atmosphere is well known due to the successful Cassini–Huygens mission. Specifically, the INMS [4,5] mass spectrometer detected hydrocarbons and N-containing species up to m/z 100. Due to low (unit) resolution, however, the precise assignments are ambiguous. Heavier positively and negatively charged ions (with higher m/z up to several hundreds or even thousands) were detected by the CAPS-IBS [6] and CAPS-ELS [7,8] mass spectrometers, respectively. This is very important in situ information, but again, due to low resolution, it prevents reliable identification of the detected ions. Moreover, it is not clear how much and by what mechanism nitrogen is incorporated in Titan’s organic aerosols, i.e., the formation mechanisms of astrobiologically important organics are not fully understood.

To improve our understanding, in line with experiments, theoretical modeling of the kinetics of Titan’s atmospheric processes is of primary importance. These studies are out of the scope of the current paper but we refer here to the work of Vuitton, Lavvas, Yelle, and others [15,16,17,18,19].

Obviously, for generating ion–molecule (or even molecule–molecule) interactions an energy source is required. On Titan (and, in general, other Solar System bodies with atmosphere), one of these sources are high enough energy photons in the UV range that are capable of ionizing/dissociating N₂ (and/or CH₄). Using tunable synchrotron radiation at the Advanced Light Source, Imanaka and Smith recently demonstrated the first evidence of nitrogenated organic aerosol production by extreme ultraviolet–vacuum ultraviolet irradiation of a N₂/CH₄ gas mixture [20]. The products of these UV-induced reactions were then analyzed by a low resolution quadrupole mass analyzer and by ultrahigh-mass-resolution FT-ICR (Bruker 9.4 T Apex Qh, Bruker, Billarica, MA, USA) mass spectrometry with LDI. The N₂/CH₄ photolytic solid products at 60 and 82.5 nm irradiation wavelengths indicate the predominance of highly nitrogenated compounds. ¹⁵N labeling predicted that short lived nitrogen containing reactive species, such as the HC₂N radical might play a role in the formation of C_xH_yN_z species. We note, however, that isotope labeling of the gas mixtures with ¹³C clearly indicated the presence of unsaturated hydrocarbons at 60 nm irradiation [20].

FT-ICR cells can be used not only to detect large number of ions with high resolution and m/z precision but also to perform ion-molecule reactions in the cell. Maybe the simplest example is hydrogen/deuterium exchange (HDX) reactions of isomeric ions [21]. Different HDX rates indicate structural isomers of isomeric ions. In some cases, reaction products are also different (see, e.g., ND₃ addition to b_n peptide fragment ions, when n > 5 [22]).

A novel experiment relevant to modeling Titan’s atmospheric chemistry has been performed inside the cell of a FT-ICR mass spectrometer [23]. The ICR cell served as a “reaction chamber” into which gas mixtures representative of Titan’s atmospheric composition were injected. These gas mixtures were then irradiated with extreme ultraviolet radiation from the Elettra synchrotron. The influence of various gas mixtures and irradiation frequencies were studied on the ion evolution. This approach allowed the authors to compare the theoretical model predictions [15] to the laboratory experiments and to test for missing and/or poorly constrained pathways. Overall, the comparison between observed and modeled ion intensities validates the kinetic model. Unfortunately, the ICR cell is not infinite in size and its surface may initiate (catalyze) certain reactions implicating the involvement of heterogeneous chemistry on aerosols that might efficiently produce HCN and NH₃ in Titan’s upper atmosphere.

In parallel to the above-mentioned comprehensive studies, we also performed simple experiments in the FT-ICR cell of an IonSpec 4.7 T instrument (IonSpec, Irvine, CA, USA) where 40 eV electrons were used as the energy source (unpublished results). The electrons were generated using a regular filament mounted outside the FT-ICR cell. Again, N₂/CH₄ gas mixtures were leaked into the ICR cell and the electron flux was varied as well as the reaction time (i.e., ions evolving time). Typical spectra are shown in Figure 2a,b.

It is expected that 40 eV electrons ionize N₂ and CH₄. The irradiation time of 15 s of ionization with 40 eV electrons is sufficient time to ionize these molecules efficiently. A 2 s reaction time was allowed after the electron beam was turned off so other products are also formed (within the altogether 15 + 2 = 17 s time window). Not surprisingly, products that mimic chemical ionization (CI) when CH₄ is used as a CI gas (see, e.g., CH₅⁺ and C₂H₅⁺) have also been detected. Ultrahigh resolution allows us to take a “closer look” to identify HCNH⁺ (most likely protonated HCN), N₂⁺, and the ethylene molecular ion (CH₂=CH₂⁺) all with m/z 28 nominal mass. When the gas mixture was allowed to react further for 600 s after the applied 15 s electron flux, the ionized N₂ disappeared and the ratio of CH₂N/CH₂CH₂ increased (see inset in Figure 2b). Note that the intensity of C_xH_y ions has been gradually increased and shifted to higher m/z masses (up to about m/z 180) and they overwhelmingly dominate the 600 s reaction time spectrum (full m/z range not shown). Ions with C_xH_yN_z compositions were detected at very low intensity (or not detected at all).

Unfortunately, albeit not surprisingly, similar to the Elettra experiments mentioned above [23], we also faced the problem of the involvement of the surface, as well as residual H₂O after 10 s of reaction time, after which the H₃O⁺ ion becomes dominant. Evidence for surface chemistry was found in the negative ion mode by the detection of Cl⁻ and Br⁻ anions (spectrum not shown) that, most likely, are sputtered products from the mesh located between the filament and the ICR cell (that functions as an “on–off” switch to control the electron beam flux in and out of the ICR cell).

In spite of the shortcomings mentioned above, we believe that using the ICR cell as a gas phase “reactor” is justified and the ions produced are informative for the very early stage of ion formation generated by either high energy photons (in the UV range) or electrons (in the tens of eV energy range). In fact, surface chemistry, albeit a different one related, e.g., to microscopic dust in the atmosphere, may play a role in the formation of C_xH_yN_z species that can grow to colloidal droplets responsible for the haze (non-transparent atmosphere). These droplets can rain down with liquid methane to Titan’s surface and can be exposed to additional chemical reactions, most notably hydrolysis in ammonia/ice pounds (see discussion below).

The first detailed ultrahigh resolution MS study on non-volatile laboratory tholin components (molecular weight > 120 Da) was published back in 2003 [24]. A homemade glow discharge (high voltage (HV)) reactor was built in the laboratory of Mark Smith (then at the University of Arizona) to generate laboratory tholin samples. This reactor was later modified and produced more pure tholin samples [25]. If methane is constantly replenished in the atmosphere, non-volatile aerosols have had billions of years to accumulate on Titan’s surface. Thus, determining the chemical composition of its constituents is crucial to better understand and predict what might happen to them on the surface. Without ultrahigh resolution, significant chemical information would have been lost. Most notably, we realized that at almost every integer m/z there are several ions that are related to a 2CH₂/2N permutation that produce homolog series that grow and diminish with increasing m/z. (The mass difference between 2CH₂ and 2N is 28.031300 − 28.006148 = 0.025152 Da, which is easily detectable using FT-MS even at higher m/z values.) These series become more unsaturated with increasing molecular mass. The success of this study inspired us as well as others to perform additional analysis by using ultrahigh resolution mass spectrometry [24,25]. Similar trends were observed by Pernot et al. [26] in a more detailed study.

It is well known that most of the prebiotic molecules on Earth contain oxygen. Because there is almost no oxygen containing compounds in Titan’s atmosphere, the obvious question emerged: is there any way to incorporate oxygen into the C_xH_yN_z molecules that are initially formed in a reductive atmosphere?

One avenue is to add the N₂/CH₄ mixtures with small amount of CO, perform the laboratory simulations and analyze the products by ultrahigh resolution mass spectrometry. A detailed study by Hörst et al. was published in 2011 [27]. The laboratory simulations were performed in the PAMPRE (Production d’Aerosols en Microgravité par Plasma Réactifs) apparatus [28], producing aerosols by cold plasma activation of a gaseous N₂/CH₄/CO mixtures of 96.2%/2.0%/1.8% and 93.2%/5.0%/1.8%. Isotopically labeled C¹⁸O was used in these studies to further confirm the oxygen contribution from CO. The LTQ-Orbitrap mass spectrometer (Grenoble, France) was used to obtain the ultrahigh resolution mass spectra. Although the gas composition was significantly different from that of Titan’s upper atmosphere, the study is important and clearly shows that when simple oxygen containing compounds (such as CO) are present, complex molecules containing oxygen are definitely formed. Most interestingly, several prebiotic molecules have been detected and confirmed, including adenine, cytosine, uracil, thymine, guanine, glycine and alanine [27].

An alternative way for oxygen incorporation is hydrolysis reactions of Titan aerosols on the surface of Titan. This sounds bizarre at first because the temperature is so low (about 95–100 K) on the surface. Obviously, water is frozen but water mixed with some ammonia has a significantly lower freezing point (e.g., 29.4% ammonia/water is liquid at about 193 K). Ammonia/water ponds can, furthermore, be heated up by meteorite impacts and they can stay melted for several thousands years.

We performed some qualitative hydrolysis reactions in H₂O and ammonia/water [29]. It was found that laboratory tholin components incorporate significant amount of oxygen within an hour at 373 K or within a few hours at 293 K. Exposure to air for several weeks also led to oxygen incorporation but the process was, as expected, much slower than in the liquid phase. In the same paper, we also proposed some characteristic fragmentation patterns using low resolution ion trap and high resolution (FT-ICR) tandem MS/MS measurements. In the positive ion mode, characteristic losses include the loss of NH₃, HCN, C₂H₄, CH₃CN, CH₂=NH, etc. Another importance of this work was to demonstrate the ESI ionization with methanol:acetonitrile (1:1) solvent combination is an inert mode to generate ions from tholin components, i.e., when laboratory tholins are harvested carefully and anaerobically, oxygen contamination is negligible during the ESI process and the solutions are stable for at least 1–3 days allowing us more than enough time to gather ultrahigh resolution mass spectrometry data on the samples.

More comprehensive studies on laboratory hydrolysis have been carried out by Neish et al. in water and 14% ammonia/water solutions at different temperatures (−20, 0, 20, and 40 °C) [30,31,32]. Kinetic curves and rate constants have been determined at those temperatures for both decay processes (i.e., when C_xH_yN_z species decayed due to oxygen contribution) and for oxygen growth reactions (i.e., the growth of C_xH_yN_zO_n species). (Note that experiments with ¹⁸O labeled water clearly proved that the source of oxygen was water and not contamination from moisture in air.) Activation energies were then calculated from Arrhenius plots for several species. The activation energies were found to be in the range of 50 ± 10 kJ/mol (decay) and 60 ± 10 kJ/mol (growth) [30]. Extrapolating down to about 100 K, these values encourages us to say that hydrolysis reactions can, indeed, occur within 3000–10,000 years, which is a very short time on the planetary time scale. Again, we note for correctness, that no phase transitions (i.e., freezing of melted ponds) are assumed. But even then, meteorite impact can heat up local ammonia/water environment that might cool down within a couple of days or weeks but oxygen enrichment can occur within this relatively short time even though not at full scale

Similar to the work of Hörst et al. amino acids asparagine, glutamine, and histidine were detected among the hydrolysis products with 14% ammonia/water [32]. It is interesting to note here that narrow band experiments with resolution >200,000 clearly showed [31] that tholin components not only react with oxygen but also with NH₃ as was proved by using labeled ¹⁵NH₃. The power of using even higher resolution (up to almost 500,000) is demonstrated in Figure 3 and Table 1 that shows the unambiguous identification of additional species, e.g., the odd electron molecular ions for ions with the same nominal mass of m/z 192. The lower spectrum practically demonstrates an isotopic fine structure experiments that is widely used in metabolomics. Even though the bottom spectrum clearly shows ten different species, this spectrum is relatively simple compared to those obtained for species containing other mass deficient elements, such as S, P, and halogens.

Data in Table 1 show that components 5, 9, and 11 are the well-characterized species related to the above-mentioned 2CH₂/2N replacements. In addition, it is expected to detect the ¹³C isotope peaks for species with m/z 191 nominal mass (3, 6, 10). This sample is the same as in used in Reference [25] (denoted as UHVT_001) so it has been stored in a −80 °C freezer for years in a parafilmed bottle but has been opened a couple of times during these years so the appearance of small oxygen containing “contaminants” is not surprising (peaks 1, and 4). What is especially interesting and not reported before is the detection of odd electron molecular ions (peaks 3 and 7). Not shown in Table 1 but the largest error for accurate mass measurement that is possible to achieve with internal calibration and with a 15 T magnet is about 20 ppb. (First, the broadband spectrum is calibrated, ion compositions are identified and confirmed, then an ion in the narrow range is used to recalibrate the isotopic fine structure spectrum.) Obviously, the top spectrum in Figure 3 with a ≈R = 63,000 resolution is already informative but separation of peaks 3, 4 and 7, 8 is not adequate enough.

It is important to note here that the spectra shown in Figure 3 have been obtained on our Bruker SolarixR 15T FT-ICR instrument (Ohio State University, OH, USA) with an alternative ionization technique, laser desorption ionization (LDI). We first realized the effectiveness of LDI for laboratory tholin ionization back in 2005 [29] by generating LDI ions up to m/z 1000. That time LDI with FT-ICR capability was not available in our laboratory so only a time-of-flight (TOF) instrument (a Bruker Ultraflex III MALDI TOF-TOF, University of Arizona, Tucson, AZ, USA) was used. The resolution of TOF instruments (usually <50,000) is not always adequate to detect all “isobaric” ions, although they still can provide useful results just as demonstrated in [33] for tholin samples or de Marcellus et al. for aldehyde and carbohydrate molecules evolution in laboratory simulations by GC-TOF mass spectrometry [34].

Most of the studies mentioned above focused on positively charged ions (except with [33] that presented results for negatively charged ions). Recently, we performed a comparative study for positively and negatively charged ions by using both ESI and LDI ionization methods [25]. Characteristic differences have been observed: more saturated positively charged ions represent mostly amines and imines, meanwhile negatively charged ions are more unsaturated and represent compounds with nitrile (cyano) functionality (CN groups). In some cases, exotic C_xN_z ions are observed in the negative LDI spectra, such as the C₁₀N₅⁻ anion that is most presumably a pentacyano cyclopentadienyl anion. Tandem MS/MS fragmentation confirmed the cyano functionality by losing HCN as a neutral molecule from several ions. Tandem MS/MS fragmentation and quantum chemical calculations suggested the formation of ion-molecule complexes, with a common anion of C₂N₃⁻. (This ion at m/z 66 was first mentioned and studied in detail by Carrasco et al. [33].) Nevertheless, further experimental and theoretical studies are necessary to confirm the existence of stable ion-molecule complexes.

It is clear from the experimental and theoretical studies mentioned above that structural isomers do exist. Thus, even with the ultrahigh resolution power of FT-MS, isotope distinction is not always possible. Other than theoretical calculations, there are three different mass spectrometry related experiments that can be used to distinguish isomers: (i) ion mobility; (ii) gas-phase HDX; and (iii) variable wavelength infrared multiphoton dissociated spectroscopy (IRMPD). Although we have performed some ion mobility and HDX experiments in our laboratories, the result for laboratory tholins are not conclusive enough to be publishable. On the other hand, we successfully used IRMPD action spectroscopy in the CLIO laser facility [35] to obtain information on functional groups. The evaluation of the data and theoretical calculations are in progress and the results will be published elsewhere. For illustration, in Figure 4 we show one IRMPD action spectrum that was obtained for a C₄H₃N₄⁻ anion (that we assume is an isomer (or isomers) of “deprotonated tetracyanide”). To our best knowledge, this is the first time a nitrile (CN stretching) band was detected at around 2180 cm⁻¹ in the CLIO facility. This is an important result because it provides proof for the existence of –CN groups in highly unsaturated anions.