Special Issue Reports

Organic compounds on comet 67P/Churyumov-Gerasimenko revealed by COSAC mass spectrometry

See allHide authors and affiliations

Science  31 Jul 2015:
Vol. 349, Issue 6247, aab0689
DOI: 10.1126/science.aab0689


Comets harbor the most pristine material in our solar system in the form of ice, dust, silicates, and refractory organic material with some interstellar heritage. The evolved gas analyzer Cometary Sampling and Composition (COSAC) experiment aboard Rosetta’s Philae lander was designed for in situ analysis of organic molecules on comet 67P/Churyumov-Gerasimenko. Twenty-five minutes after Philae’s initial comet touchdown, the COSAC mass spectrometer took a spectrum in sniffing mode, which displayed a suite of 16 organic compounds, including many nitrogen-bearing species but no sulfur-bearing species, and four compounds—methyl isocyanate, acetone, propionaldehyde, and acetamide—that had not previously been reported in comets.

The study of the chemical composition of comets provides key information about the raw materials present in the early solar system (1, 2). Ground and space-based observations have identified over 20 organic molecules in comet comae (3, 4), a subset of which are of prebiotic interest (5, 6).

The Cometary Sampling and Composition (COSAC) experiment on Rosetta’s lander Philae was designed to detect and identify organic molecules in the material of comet 67P (7). It consists of a gas chromatograph (GC) and a time-of-flight mass spectrometer (TOF-MS) to analyze samples delivered by the sample drilling and distribution system (SD2). COSAC can also operate in sniffing mode, in which the MS accumulates data without active sampling by SD2. Molecules that have passively entered the instrument are ionized, accelerated, and finally registered by COSAC. MS sniffings were made several times between launch and arrival at the comet, including during a fly-by of Lutetia (8). MS sniffings were made on arrival at 67P from 10 km above the surface, after initial touchdown, and at the final resting site (Fig. 1).

Fig. 1 Mass spectra taken by COSAC in “sniffing mode.”

Top (green): spectrum taken 25 min after first touchdown; the m/z 18 peak reached a height of 330 counts, but the spectrum is truncated to show smaller peaks more clearly; middle (red): final spectrum, taken 2 days later at the current Philae position; bottom (blue): first spectrum, obtained in orbit 27 days before landing, from a distance of 10 km.

The Philae lander first touched down on 67P on 12 November 2014 at 15:34:04 UTC and then bounced. The impact excavated about 0.4 m3 of solid material (9), some of which would have entered COSAC’s two exhaust pipes, which are on the bottom of the lander (10), and then stuck to the inside of the 2-cm-wide pipes. The temperature in these pipes was 12° to 15°C (10), midway between the cold cometary exterior and the heated interior of the lander, allowing volatile organics to sublime and be detected by the MS in a measurement that began at 16:00:30 UTC and ended at 16:02:50 UTC, when the lander was about 150 m above the surface on its first bounce. We focus here on this spectrum (green in Fig. 1), which differs fundamentally in the number and intensity of its peaks from the undisturbed spectra taken before and after, and represents excavated cometary material (10). Our approach was to find the best fit to this spectrum of a superposition of standard National Institute of Standards and Technology (NIST) mass spectra (11) of candidate cometary molecules. The spectral deconvolution methodology used is similar to that used in other space missions [such as the Ion and Neutral Mass Spectrometer (INMS) measurements by Cassini and the GCMS on the Huygens probe] (1214).

Because COSAC has a mass resolution of only 300, single mass peaks cannot be resolved into different molecular species [e.g., CO, N2, and C2H4, all at a mass/charge ratio (m/z) of 28, are indistinguishable]. Analysis (10) was limited to compounds below m/z 62 because signals beyond this value are too faint to be distinguished reliably from noise. The peak at m/z 78, for example, is not real: Several ions coincidentally ended up in a single channel, leaving neighboring ones empty (10). All conceivable molecules were first listed and their fragmentation patterns evaluated (table S1). Elimination of molecules with incompatible fragmentation patterns (for reasons described in table S3) led to a short-list of candidate molecules (table S2). We further reduced the short list by making the fit in order of decreasing mass, starting from m/z 59 (10), and eliminating unstable and unsaturated species. This yielded a good fit to all peaks (except m/z 15 and a fraction of m/z 29, Fig. 2) with 16 species from several families of molecules—alcohols, carbonyls, amines, nitriles, amides, and isocyanates—in a consistent combination (Table 1). Peaks for m/z < 10 were not included in the fit because they are not listed in the standard NIST mass spectra (11). The molecular abundances of these compounds relative to that of water (Table 1) were corrected for electron cross section (table S4). The absence of ions at m/z 32 indicates a lack of sulfur-bearing species (Fig. 2 and Table 1). Amino acids were not included in the fit because the molecular ion peaks of glycine (m/z 75) and alanine (m/z 89) are negative after background subtraction, thereby suggesting that they are noise. Although fragment peaks assigned to glycine and alanine in the NIST standard spectra (11) are present in the COSAC spectrum (in the m/z 30s to 40s range), any contribution to these fragment peaks from amino acids is difficult to disentangle from the contributions of other species.

Fig. 2 The fit to the observed spectrum.

Comparison of the COSAC original mass spectrum (black bars for each integer mass) and the spectrum reconstructed from the best fit (orange bars to right of original signal). The peak heights are normalized to 100 for the m/z 18 peak (which has been truncated).

Table 1 The 16 molecules used to fit the COSAC mass spectrum.

View this table:

The main source of error is the low signal intensity, averaging about 100 counts (table S1). A statistical square-root of n approach yields a standard deviation of 10%. In addition, the NIST standard spectra (11) have a 15% error. Considering formal error propagation and the uncertainties in our peak-fitting algorithm, we estimate that the abundances given in Table 1 are accurate to about a factor of 2. The fit of a mass spectrum whose peaks result from the superposition of different molecular species is intrinsically degenerate, with several possible solutions (10).

The absence of large quantities of NH3, HCHO, and CO2 in our best fit may seem surprising because they were expected to be present as components of cometary ice. NH3 (m/z 17) was not needed for the fit, but the presence of small quantities seems likely. However, this is hard to quantify because the large H2O peak at m/z 18 implies a substantial contribution to the m/z 17 peak from the OH fragment peak of H2O, which is difficult to distinguish from any NH3 contribution. HCHO (m/z 30) and CO2 (m/z 44) are not included because m/z 30 is mainly accounted for by fragment peaks of other molecules, rather than by the molecular ion of HCHO, and m/z 44 is mainly accounted for by fragment contributions from acetamide, formamide, and acetaldehyde, rather than by CO2.

We initially tried a fit that started with the assumption that m/z 44 came from CO2, but no acceptable fit could be achieved to the remaining peaks. If all of m/z 44 were ascribed to CO2, our sample would only contain 3% of CO2 relative to water. Using the procedure described above, we found that a more sensible fit for all mass peaks, especially m/z 57, 58, and 59, could only be achieved by assuming a CO2 concentration of less than 0.1%. The low abundance of CO2, NH3, and HCHO could indicate that the excavated COSAC sample came from an area depleted in volatile ice components. Observations by the Visible, Infrared and Thermal Imaging Spectrometer (VIRTIS) from the Rosetta orbiter (15) do suggest a dark surface depleted in volatiles, consisting mainly of refractory organic macromolecular materials, with very little ice on the surface. Studies (16) using the Rosetta Orbiter Spectrometer for Ion and Neutral Analysis (ROSINA) indicate that volatile ices sublime diurnally and seasonally, with CO2 ranging from 3% relative to water in local summer (the present case) to 80% in local winter.

The COSAC findings differ from those of Ptolemy (17) because COSAC sampled particles excavated by the impact (10) that entered the warm exhaust tubes located on the bottom of the lander, where they pointed toward the surface, whereas Ptolemy sampled ambient coma gases entering exhaust tubes located on top of the lander, where they pointed toward the sky (possibly with the addition of some dust that made its way around the lander). That COSAC detected far more nitrogen-bearing compounds than Ptolemy agrees with earlier observations that nitrogen was more abundant in the dust than in the gas of comet Halley (18). The Ptolemy team interpret their mass spectrum as fragments of polyoxymethylene polymer, with a strong CO2 peak of intensity 20% relative to water. COSAC did not detect ambient coma gases (which were dominated in Ptolemy data by CO2 with a few polymer fragments). The COSAC MS maintains a constant pressure; thus, subliming gases from our ground sample pushed the ambient coma gases outside the COSAC MS. Before sublimation, the total pressure inside the COSAC MS was dominated by CO2, in line with Ptolemy data and prelanding COSAC spectra. After sublimation, the total pressure inside the COSAC MS was due to the sum of the partial pressures of all the sublimed ground materials. This can explain the missing CO2 in the post-touchdown spectrum. The displacement of ambient coma gases by subliming ground materials and the temperature of 12° to 15°C in the COSAC exhaust tubes, which is too low to break down any refractory polymers in the ground materials, combine to explain why COSAC did not detect any polymer fragments.

The COSAC molecules form a consistent set related by plausible formation pathways (Fig. 3). A nitrogen source such as NH3 must originally have been abundant to form the many N-bearing species, but could since have mostly evaporated or been used up in reactions. All the COSAC organics can be formed by UV irradiation and/or radiolysis of ices due to the incidence of galactic and solar cosmic rays: alcohols and carbonyls derived from CO and H2O ices (19), and amines and nitriles from CH4 and NH3 ices (20). Hydrolysis of nitriles produces amides, which are linked to isocyanates by isomerization.

Fig. 3 Possible formation pathways of COSAC compounds.

Species in red are not confidently identified; species in green are reported for the first time in comets by COSAC.

Several of the COSAC compounds, such as HCN, CH3CN, and HNCO, are present in the comae of most comets (1). Others, such as CH3CHO, HCONH2, CH2(OH)CH2(OH), CH3NH2, and C2H5NH2, have only been found in a few comets. Four molecules reported by COSAC—CH3NCO, CH3COCH3, C2H5CHO, and CH3CONH2—have not been previously reported in a cometary environment, and CH2OHCHO has only been reported as an upper limit. These cometary molecules are all predicted by our generalized formation scheme (Fig. 3). CH2OHCHO is an efficient initiator in the prebiotic formation of sugars (21). HCN is a key molecule in the prebiotic synthesis of amino acids (21, 22) and nucleobases (21) and even offers an elegant pathway to sugars (23). HCONH2 provides a prebiotic route to nucleobases (24). HCONH2 (24) and CH3CONH2 (21) catalyze phosphorylation of nucleosides to nucleotides, in which amines also play a role (21). Isocyanates play a major role in the prebiotic synthesis of peptides, through the so-called isocyanate route (22). The complexity of cometary nucleus chemistry and the importance of N-containing organics imply that early solar system chemistry fosters the formation of prebiotic material in noticeable concentrations.

Supplementary Materials


Materials and Methods

Figs. S1 and S2

Tables S1 to S4

References (2532)

References and Notes

  1. Full details and methods are in the supplementary materials on Science Online.
  2. Acknowledgments: This work was supported by Deutsches Zentrum für Luft- und Raumfahrt (DLR) under contract no. 50 QP 1302, and by Centre National d’Etudes Spatiales (CNES) grants at LATMOS and LISA. U.J.M. was funded by L’Agence Nationale de la Recherche (ANR-12-IS07-0006). Support from the NASA Astrobiology Institute (P.E.), the NASA Postdoctoral Program at Goddard Space Flight Center (T.G.), Programme de Développement d’Expériences scientifiques through Enterprise Ireland (S.McK.L.), and the Welch Foundation (A.J.MacD.) is acknowledged. Rosetta is a European Space Agency (ESA) mission with contributions from member states and NASA. The Philae lander was provided by a consortium led by DLR, Max-Planck-Institut für Sonnensystemforschung, CNES, and Agenzia Spatiale Italiana. The data used in this study are available through ESA’s Planetary Science Archive (www.rssd.esa.int/index.php?project=PSA&page=rosetta).
View Abstract

Stay Connected to Science

Navigate This Article