Cyanogen, cyanoacetylene, and acetonitrile in comet 67P and their relationship to the cyano radical
N. Hänni, K. Altwegg, H. Balsiger, M. Combi, S. A. Fuselier, J. De Keyser, B. Pestoni, M. Rubin, S. F. Wampfler
AAstronomy & Astrophysics manuscript no. 39580final © ESO 2021February 9, 2021
Cyanogen, cyanoacetylene, and acetonitrile in comet 67P and theirrelation to the cyano radical
N. Hänni , K. Altwegg , H. Balsiger , M. Combi , S. A. Fuselier , , J. De Keyser , B. Pestoni , M. Rubin , and S. F.Wampfler Physics Institute, Space Research & Planetary Sciences, University of Bern, Sidlerstrasse 5, CH-3012 Bern, Switzerlande-mail: [email protected] Department of Climate and Space Sciences and Engineering, University of Michigan, Ann Arbor, MI, USA. Space Science Directorate, Southwest Research Institute, San Antonio, TX, USA. Department of Physics and Astronomy, The University of Texas at San Antonio, San Antonio, TX, USA. Royal Belgian Institute for Space Aeronomy, BIRA-IASB, Brussels, Belgium. Center for Space and Habitability, University of Bern, Gesellschaftsstrasse 6, CH-3012 Bern, SwitzerlandReceived XXX / Accepted YYY
ABSTRACT
The cyano radical (CN) is one of the most frequently remotely observed species in space, and is also often observed in comets. Datafor the inner coma of comet 67P / Churyumov-Gerasimenko collected by the high-resolution Double Focusing Mass Spectrometer(DFMS) on board the Rosetta orbiter revealed an unexpected chemical complexity, and, recently, also more CN than expected fromphotodissociation of its most likely parent, hydrogen cyanide (HCN). Here, we derive abundances relative to HCN of three cometarynitriles (including structural isomers) from DFMS data. Mass spectrometry of complex mixtures does not always allow isolation ofstructural isomers, and therefore in our analysis we assume the most stable and abundant (in similar environments) structure, that isHCN for CHN, CH CN for C H N, HC N for C HN, and NCCN for C N . For cyanoacetylene (HC N) and acetonitrile (CH CN),the complete mission time-line was evaluated, while cyanogen (NCCN) was often below the detection limit. By carefully selectingperiods where cyanogen was above the detection limit, we were able to follow the abundance ratio between NCCN and HCN from3.16 au inbound to 3.42 au outbound. These are the first measurements of NCCN in a comet. We find that neither NCCN nor either ofthe other two nitriles is su ffi ciently abundant to be a relevant alternative parent to CN. Key words. comets:general – comets: individual: 67P / Churyumov-Gerasimenko – instrumentation: detectors – methods: data anal-ysis
1. Introduction
The CN radical is thought to be a key-species in the evolutionof prebiotic molecules and its abundance in various stellar andinterstellar environments is well established from remote ob-servations (Herbst & van Dishoeck 2009; van Dishoeck 2014).Hänni et al. (2020) reported the first in situ detection of the CNradical in comet 67P / Churyumov-Gerasimenko (67P hereafter).These latter authors showed that the abundance of CN cannotbe explained by photodissociation of HCN alone. These resultsare consistent with various previous remote observation studiesof other comets, for example of 6P / d’Arrest observed by DelloRusso et al. (2009). The first comet showing this discrepancywas C / Half ofFray’s sample of eight comets showed inconsistent production The yield of a daughter species (D) from a parent species (P) is deter-mined by two factors: the photodissociation rate of P (i.e., the numberof dissociations per second) and the quantum yield of P → D (i.e., the rates and the full sample showed shorter CN formation scalelengths than the HCN destruction scale length, especially be-low 3 au. Possible additional or alternative parents of CN shouldtherefore have higher photodissociation rates than that of HCN,which was estimated by Crovisier (1994) to be 1 . × − s − and by Huebner et al. (1992) to be 1 . × − s − at quiet Sunconditions at 1 au. The situation appears slightly di ff erent if pro-duction rates obtained from infrared (IR) spectroscopy are con-sidered, such as those reviewed by Dello Russo et al. (2016), be-cause the IR-based HCN production rate is significantly higherfor some comets than the radio-based one. The origin of the ob-served discrepancies is the topic of ongoing debate. However,Dello Russo et al. (2016) also reported incompatible CN andHCN production rates for 7 out of the 21 comets investigated.These latter authors suggest grains as the possible origin of theunexplained portion of CN; cf. also Hänni et al. (2020). Alterna-tively, many volatile candidate molecules have been suggested.However, the origin of CN in comets where HCN cannot beits sole parent remains to be found. Cyanogen, already men-tioned by Swings & Haser (1956), ranks high among the can-didates as it is a simple molecule that photodissociates into twoCN radicals. For this reason, and because of the lack of experi- portion of dissociations that occur according to this specific dissociationchannel). Article number, page 1 of 10 a r X i v : . [ a s t r o - ph . E P ] F e b & A proofs: manuscript no. 39580final mental values, the quantum yield of cyanogen photodissociationinto CN is often assumed to be equal to two (e.g., Bockelée-Morvan & Crovisier 1985). The reported photodissociation ratesvary. Bockelée-Morvan & Crovisier (1985) determined a pho-todissociation rate of 3 . × − s − . Other possible candidatemolecules that may produce CN upon photodissociation are thetwo nitrile species CH CN and HC N. Fray et al. (2005) reviewthe photochemical properties of these two molecules. The CNproduction rate from CH CN photodissociation is likely to be100 times lower than from HCN photodissociation and henceon the order of 1 × − s − . This results from a photodissocia-tion rate of 6 . × − s − (Bockelée-Morvan & Crovisier 1985)when the quantum yield is taken into account. For the vacuumultraviolet region, Kanda et al. (1999) reported that CN produc-tion is a minor dissociation channel and that the correspondingquantum yield may be lower than 0.02. HC N seems to have ahigher photodissociation rate. Values of between 2 . × − s − and 7 . × − s − were mentioned by Fray et al. (2005); however,the authors note that the quantum yield may be as low as 0.05(cf. Halpern et al. 1988). Haser model calculations indicate thatNCCN / HCN production rate ratios of between 0.15 and 0.85 andHC N / HCN production rate ratios of between 0.2 and 0.7 (underthe assumption of a quantum yield equal to 1) could explain thediscrepancies in the observed scale lengths. When the Rosetta spacecraft reached the inner coma and ap-proached the cometary nucleus of 67P for the first time, a newera of comet studies began. During the two years of the Rosettascience mission phase, from August 2015 through September2016, the DFMS, a high-resolution sector-field mass spectrom-eter that was part of the Rosetta Orbiter Spectrometer for Ionand Neutral Analysis (ROSINA) sensor package (Balsiger et al.2007), analyzed the neutral composition of 67P’s inner comafrom up-close. Comet 67P is a Jupiter-family comet (JFC) witha period of 6.45 years, a perihelion distance of 1.24 au, and anaphelion distance of 5.68 au. The tilted rotation axis of the cometleads to an inbound equinox in May 2015 at 1.7 au and an out-bound equinox in March 2016 at 2.7 au. 67P acquired its currentorbit after a close encounter with Jupiter in 1959. Although noreliable orbital parameters exist before 1923 (the year of an ear-lier encounter with Jupiter), 67P is believed to have resided in theinner Solar System for several thousand years (Maquet 2015).Rosetta was as close as 10 km from the cometary nucleus for thefirst time in October 2014. At that time, the northern hemispherewas experiencing a long but cold summer. Illumination condi-tions gradually changed on the inbound trajectory, giving wayto a short and intense summer on the southern hemisphere witha peak in outgassing shortly after 67P’s closest approach to theSun.Bulk abundances of CH CN and HC N relative to water werederived by Rubin et al. (2019) from data collected in May2015 (at approximately 1.6 au) over the southern hemisphere:(0.0059 ± ± CH CN and HNC were not considered for Haser modeling by Frayet al. (2005) because these molecules do not possess photodissociationrates higher than that of HCN. relative to hydrogen cyanide for the first time, namely fromin situ data collected in the inner coma of comet 67P by theROSINA / DFMS. The strategy pursued regarding data analysisis described in Sect. 2. Subsequently, the results are presentedin Sect. 3 together with the full mission data for cyanoacetyleneand acetonitrile and in Sect. 4 we discuss these in the context ofvalues published for other astrophysical environments.
2. Instrumentation and method
The DFMS and its fully identical laboratory twin were built inthe Mattauch-Herzog geometry (Mattauch & Herzog 1934). De-tails and specifications of the DFMS were given in Balsiger et al.(2007). The instrument was designed to analyze cometary neu-trals after electron-impact ionization (EI) with a 45 eV elec-tron beam. Because of the ionization process, the moleculesoften undergo fragmentation, which has to be taken into ac-count, for instance when neutral density data are to be derived.Although mass spectrometry does in principle allow to distin-guish structural isomers based on their fragmentation patterns,which may be di ff erent due to the di ff erent structural entities ofthe molecules, this distinction is not possible for nitrile species,primarily because of the lack of reference data. Nitrile specieshave not been calibrated in the DFMS laboratory twin modelat the University of Bern for reasons of toxicity. Consequently,the fragmentation analysis we performed relies on mass spec-tra from the National Institute of Standards and Technology(NIST) Standard Reference Database Number 69 (Steins 2018)and hence on data acquired with a standard 70 eV ionizing elec-tron beam. It should be noted that lower electron energies (atequal charge density) lead to lower fragment yields. For C HN,only data for cyanoacetylene are available. For C H N, spec-tra of two isomers are available, namely acetonitrile (CH CN)and methylisocyanide (CH NC). However, these are very similarand can barely be distinguished, especially as the di ff erent ion-ization voltage would expectedly impose (slightly) di ff erent sig-nal intensities. The analysis presented here is therefore based onthe NIST fragmentation patterns of acetonitrile and cyanoacety-lene, which also are the most common isomers for comets (cf.e.g. Mumma & Charnley 2011). Also, for the full mission datafor CHN, the fragmentation of the more stable isomer hydrogencyanide (HCN) is used. Notably, no fragmentation data are avail-able for its isomer hydrogen isocyanide (HNC). For C N , onlythe mass spectrum of NCCN was measured by Stevenson (1950).NCCN produces a main (molecular ion) peak on m / z =
52 u / e andonly small signals on other masses, which are due to fragmentsand / or isotopologs (roughly 10% combined). Among the fourpossible structural isomers of C N , the most stable is NCCN.Botschwina & Sebald (1990) reported CNCN and CNNC to beless stable than NCCN by 102 and 302 kJ · mol − , respectively.We therefore consider it reasonable to attribute the C N signalto NCCN. However, we chose not to correct for fragmentatione ff ects and directly compare the NCCN signal on m / z =
52 u / e(corresponding to 90% of the total intensity due to NCCN) to theHCN signal on m / z =
27 u / e (corresponding to 84% of the totalintensity due to HCN).The data evaluation regarding NCCN had to be performed man-ually spectrum by spectrum, based on a least-squares fitting rou-tine. This is because the signal intensity was low, often even The Hill notation for chemical formulas is used in this text to de-note a molecule with several structural isomers if no specific isomer isaddressed, for instance because distinction based on the fragmentationpatterns is not possible.Article number, page 2 of 10. Hänni et al.: NCCN, CH CN, and HC N in comet 67P and their relation to CN ClC N C H D F M S i n t en s i t y [ i on s / s ] m/z [u/e] datag1+g2fit Fig. 1.
Example mass spectrum collected on 15 March 2016 (10:59 am)of mass 52 u / e –the mass of the molecular ion of NCCN– in the DFMSneutral gas mode. The sum-curve of the two Gaussians (g1 + g2) is plot-ted with a dashed black line for each peak while the overall fit is shownwith a solid black line. Error bars indicate statistical error margins. below detection limit. Moreover, the peak position is overlap-ping with the CH Cl signal to the left and with the C H N andalso the relatively large C H signals on the right. Each spectrumhas been background-subtracted and mass-scaled prior to fitting.The characteristic peak shape of DFMS is best reproduced with adouble-Gaussian peak profile when the second Gaussian (g2) isabout 5-10% of the height of the main Gaussian (g1) and about300-400% of its width (Le Roy et al. 2015; De Keyser et al.2019). Width and height ratios of the two Gaussians remain con-stant within one mass spectrum. A typical mass spectrum withan NCCN signal on mass 52 u / e is shown in Fig. 1. For reasonsdiscussed in Sect. 3, the abundance of NCCN was deduced rel-ative to HCN as a neutral density ratio n NCCN / n HCN . This ratiocan be obtained from the ratio of the observed ions per spectrum c NCCN / c HCN as follows: c NCCN c HCN · (cid:32) m HCN m NCCN (cid:33) − . · σ HCN σ NCCN = n NCCN n HCN , (1)where σ denotes the electron-impact ionization cross-sectionand m denotes the mass of the respective species. The fac-tor ( m HCN / m NCCN ) − . is an empirical correction —establishedbased on noble gas calibration measurements with the DFMSlaboratory twin instrument—that is meant to account for themass-dependent instrument sensitivity and detector yield (Cal-monte 2015; Wurz et al. 2015; Rubin et al. 2019). The error re-lated to this correction is estimated to be 15-20%. Because of thetoxicity of the substances under consideration, EI cross-sectionswere taken from Pandya et al. (2012) for HCN (approximately3.4 Å at 45 eV) and NCCN (5.1 Å at 45 eV). These values yield σ HCN /σ NCCN = .
66, a term used in Eq. 1. This ratio is expectedto have low errors as the cross-sections were calculated based onthe same method. The respective Binary Encounter Bethe (BEB)values are 4.5 Å for NCCN and 3.0 Å for HCN, also with a ra-tio of σ HCN /σ NCCN = . n NCCN / n HCN is governed by the relative statistical error of the NCCN sig-nal, while the HCN signal, being orders of magnitude larger,has a negligible contribution. For very small signals, a 5% error All statistical errors are given as 1 σ errors. must be added to the statistical error to account for background-subtraction errors and another 10% error to account for fittingerrors. Uncertainties related to fragmentation e ff ects and ioniza-tion cross-sections are systematic and the according error is es-timated to not exceed 20% in total. HCN is normally measuredat a lower gain step than the minor species, which adds a 10%systematic error to the ratios (Schroeder et al. 2019). Systematicerrors are taken into account in Table 1, where the total errormargins are given, but are neglected elsewhere.The data evaluation procedure applied to derive HC N andCH CN neutral densities was described in great detail by Ru-bin et al. (2019) and in the literature cited therein. Unlike forNCCN, the number of ions per spectrum, which was used as astarting point for the analysis presented in this work, was derivedwithout applying a mass scale. Fragmentation e ff ects have alsobeen considered. The statistical error of the neutral density of amoderately abundant species like HCN is approximately 10%.In order to obtain the total error, the systematic errors indicatedabove have to be included.
3. Results
In several spectra collected between 13 and 16 March 2016, justwhen 67P reached its outbound equinox, the mass spectromet-ric signature of NCCN was clearly observed. Notably, the abun-dance of species may be highly variable on short timescales, forexample due to the spacecraft trajectory or due to variability ofthe coma itself. This is obvious from Fig. 2, which shows theuncorrected number of ions collected on mass 52 u / e during theindicated time period. On 13 March, the signature of NCCN isnot detectable and also the C H signal is often rather weak. Thesituation completely changes within a day and the signature ofNCCN becomes clearly apparent. Based on these data we in-vestigated correlations with other coma species. The resultingPearson correlation coe ffi cients are listed in Table A.1 in theAppendix A. Relatively good correlation (R = + HNC (thereaction CN + HCN is known to possess an energy barrier). Asobserved for other comets (e.g., Mumma & Charnley 2011), itis likely that also for 67P a (presumably small) portion of theHCN observed by DFMS is actually HNC. For reasons detailedin Sect. 2, structural isomers with identical exact masses cannotalways be distinguished and we therefore assumed the observedmolecules to have the structure of the most stable isomer.Thanks to the good correlation of NCCN and HCN, we wereable to successfully locate periods where the NCCN signaturewas detectable, namely by looking at periods where the HCNabundance peaks. This is illustrated in Fig. 3 (top), which showsthe HCN signal c HCN during the full Rosetta mission time. Dayswhere the signature of NCCN was found are indicated by blackbars. Data from these days were evaluated as described in Sect. 2in order to extract the neutral density ratio n NCCN / n HCN as shownin Fig. 3 (bottom). The full mission data of the neutral abundanceratios n CH CN / n HCN and n HC N / n HCN are included. Although thecorrelation of NCCN and HCN was good in March 2016, the
Article number, page 3 of 10 & A proofs: manuscript no. 39580final - ,
00h 03 - ,
12h 03 - ,
00h 03 - ,
12h 03 - ,
00h 03 - ,
12h 03 - ,
00h 03 - ,
12h 03 - , -100-50 0 50 100 c X [ i on s / s ] l a t. [ deg ] lat. c HCN c CH Cl c NCCN c C H count 10 - ,
00h 03 - ,
12h 03 - ,
00h 03 - ,
12h 03 - ,
00h 03 - ,
12h 03 - ,
00h 03 - ,
12h 03 - , -100-50 0 50 100 Fig. 2.
Number of ions per spectrum c X , where X denotes the species observed on mass 52 u / e (i.e., CH Cl, NCCN, and C H ). The underlyingdata were collected in the four days between 13 and 16 March 2016, including the day of the outbound equinox on 16 March 2016. Countingsymbols on the x-axis mark times of analyzed spectra. HCN data has been evaluated in an automated data evaluation pipeline and time-interpolatedas described in Rubin et al. (2019). Error bars indicate statistical errors. dependence on heliocentric distance is obviously di ff erent forthe two species, which is a phenomenon widely known for var-ious cometary species. However, based on mass spectrometricdata we cannot fully exclude that, first, the signal assigned toHCN is only due to hydrogen cyanide and does not containcontributions from hydrogen isocyanide and, second, a possi-ble contribution of the isocyanide would be time-independent.A time-variable contribution of hydrogen isocyanide (if it frag-ments di ff erently from its isomer) in principle could be respon-sible for such variations of the abundances relative to HCN toa certain extent (cf. e.g., Mumma & Charnley 2011). The sameis true for possible contributions of isomers of the other nitrilespecies. Also, the CH CN relative abundance, which generallydisplays less variation because of a better correlation with HCN,clearly varies around the average value of the n CH CN / n HCN ra-tio of 0.083 with peak values around 67P’s perihelion passage.The same trend is observed for both of the other nitrile species,HC N and NCCN, which vary more pronouncedly. While ra-tios around perihelion passage are clearly above average (aver-age n HC N / n HCN = . n NCCN / n HCN = . ff erencebetween the high and low values is more than an order of magni-tude. The large scatter of the n HC N / n HCN values for 67P, indicat-ing the comparably poor correlation with HCN, leads to identi-cal averages for the full mission as compared to the time aroundperihelion. The average relative abundances of the nitrile speciespresented in this work are summarized in Table 1, including bulkabundances from Rubin et al. (2019) where available, as wellas values reported for other astrophysical environments (to bediscussed in Sect. 4). The bulk values from Rubin et al. (2019)represent the outgassing as observed between the end of Mayand beginning of June 2015. The values for n HC N / n HCN foundby Rubin et al. (2019) are consistent within error limits withthe averages presented in this work, while their n CH CN / n HCN isclearly falling below. Still far from the Sun (at approximately3.1 au in November 2014), Le Roy et al. (2015) derived a ratioof CH CN / HCN of 0.066 for the northern hemisphere and 0.025for the southern hemisphere, which bracket the bulk value fromRubin et al. (2019). At that time, only an upper limit for HC Ncould be derived. However, momentary outgassing, as captured for instance by Rubin et al. (2019) or Le Roy et al. (2015), maywell di ff er from a full mission average and also from the in-tegral over the full mission time, as presented by Läuter et al.(2020). The latter authors derived abundances of 14 major comaspecies relative to water based on an inverse coma model andROSINA / DFMS data. Their Fig. 5 compares their findings to therelative abundances from Rubin et al. (2019) (amongst others),demonstrating reasonable consistency.Based on the data presented in Fig. 3 (bottom), the heliocentricdistance dependence of the species targeted in this work can beobtained. This is usually achieved by fitting a power law to theslope of the production rate of a given species as a function of theheliocentric distance r h . The abundance of a species X relativeto HCN, as derived in this work, is proportional to its productionrate relative to HCN. Therefore, the slope of the production rateof species X relative to HCN corresponds to the di ff erence be-tween the two individual slopes. Using the same time periods I a , I b , and I c as Läuter et al. (2020), we obtained the exponents x a , x b , and x c from fitting the following power law to our data: n X n HCN ∝ r xh . (2)Adding to x the appropriate power-law exponent for HCN re-ported in Table 4 of Läuter et al. (2020) yields the desired power-law exponent of the heliocentric distance dependence of ourspecies X, that is of CH CN and HC N (albeit with less con-fidence). The same procedure was applied to the CN data fromHänni et al. (2020). The NCCN data are unfortunately too lim-ited for such a fitting. However, as n NCCN / n HCN seems to fol-low n HC N / n HCN quite well, a similar heliocentric distance de-pendence can be expected. Table 2 summarizes the accordinglyderived power-law exponents for the distinct time periods I , in-cluding those for HCN, H O, and CO found by Läuter et al.(2020). Notably, the correlation coe ffi cients for period I c arerather low (R ≈ CN and CN, less for HC N). Nev-ertheless, they are even lower for the other periods, especiallyfor period I a , which reflects the large scatter of the data. Läuteret al. (2020) used period I c to group molecules into two groups,one that follows CO and one that rather follows H O. Using the
Article number, page 4 of 10. Hänni et al.: NCCN, CH CN, and HC N in comet 67P and their relation to CN
200 400 600 c HCN [ i on s / s ] r [ k m ] r or r h c HCN n CN / n HCN n CH CN / n HCN n HC N / n HCN n NCCN / n HCN
200 400 600 10 -5 -4 -3 -2 -1 -
11 2015 -
01 2015 -
03 2015 -
05 2015 -
07 2015 -
09 2015 -
11 2016 -
01 2016 -
03 2016 -
05 2016 -
07 2016 - n X / n HCN r h [ au ] -5 -4 -3 -2 -1 -
11 2015 -
01 2015 -
03 2015 -
05 2015 -
07 2015 -
09 2015 -
11 2016 -
01 2016 -
03 2016 -
05 2016 -
07 2016 - Fig. 3. (top) Number of ions per spectrum of HCN ( c HCN ; with 30% error bars and fragmentation not taken into account) during the full Rosettascience mission as derived by Rubin et al. (2019). The cometocentric distance ( r ) is over-plotted (solid blue line) together with the days whenNCCN was found (black vertical bars). (bottom) Neutral density ratio of species X with respect to HCN ( n X / n HCN ), where X is CH CN, HC N,NCCN, and CN (from Hänni et al. (2020) for comparison). The respective total mission average values are included as solid horizontal lines forthe nitriles investigated in this work. The inbound equinox (5 May 2015), perihelion (13 August 2015), and the outbound equinox (16 March2016) are indicated (orange vertical bars) together with the heliocentric distance ( r h ; solid blue line). The period from May to September 2015used to derive the close-to-perihelion values in Table 1 is indicated in pale red. The data shown here for cyanogen are also listed in Table B.1 inAppendix B. Statistical error bars are omitted for visual clarity (except for NCCN). An impression can be obtained for the case of n CN / n HCN fromFig. 4 (top) in Hänni et al. (2020). same criteria, CH CN and HC N fall into the H O group ratherthan into the CO group. This seems to suggest that they are em-bedded in a water-ice matrix. CN, in contrast, follows a positiveexponent in time period I c , making an origin of CN from one ofthe considered nitrile molecules highly unlikely. However, thepositive exponent of CN could be the result of the interdepen-dence between cometocentric and heliocentric distance. For adirect comparison, Fig. 3 also includes the lower limit neutral Hänni et al. (2020) do not correct for the portion CN lost due tofragmentation inside the instrument because this portion has not beendetermined for the DFMS. density ratio n CN / n HCN , as derived and reported by Hänni et al.(2020) and used for fitting the power-law exponents listed in Ta-ble 2.Returning to the initial question regarding the search for alterna-tive or additional parent species of the CN radical, the followingcan be stated on the basis of our analysis of relative abundances:in the first few months of the Rosetta science mission, HC N andNCCN abundances on the order of 0.1% or less relative to HCNfall far below the minimum relative abundance of 10-20%, as re-quired by the Haser modeling of Fray et al. (2005). Rememberthat CH CN can be ruled out as parent species due to unfavor-able photochemical properties, despite its higher abundance of
Article number, page 5 of 10 & A proofs: manuscript no. 39580final
Table 1.
Abundances of the targeted nitrile species relative to HCN as resulting from our analysis of ROSINA / DFMS data from 67P ( a ) and otherastrophysical environments. Environment NCCN / HCN HC N / HCN CH CN / HCN67P (full mission) 0.00078 ± ± ± ± ± ± ( b ) – 0.0029 ± ± ( c ) – 0.079 0.082C / ( d ) – 0.022 ± ± ( e ) – 0.057 ± ± ( f ) – 0.0036 ± ± Notes. ( a ) Total error margins include statistical errors for the averages calculated as standard deviation divided by the square root of the numberof actual measurements in the averaged time period (excluding time-interpolated data points). ( b ) Values from Rubin et al. (2019). ( c ) Values fromBockelée-Morvan et al. (2000). ( d ) Values from Biver et al. (2015). ( e ) Values from Bergner et al. (2018); 40% error margins estimated from theerror bars in their Fig. 12. The average over the sample of five disks was calculated considering rotational temperatures of 30 K (CH CN / HCN)and 70 K (HC N / HCN), respectively, not taking into account CH CN relative abundance upper limits given for two of the five disks. ( f ) Valuesfrom Drozdovskaya et al. (2019).
Table 2.
Power-law exponents for CH CN, HC N, and CN for three time intervals I ( a ) . I a I b I c Species 3.1–2.3 au 1.7–2.2 au 2.4–3.6 au11 / / / / / / CN + N + + + Notes. ( a ) Further information on the employed time intervals I a , I b , and I c can be found in Table 3 of Läuter et al. (2020). up to about 10% relative to HCN. Also, abundances of HC Nand NCCN relative to HCN are at least two orders of magnitudelower than the abundance of CN relative to HCN. Judging fromthe photorates of the targeted species, the assumed gas velocityof a little less than 1 km / s, and the fact that ROSINA / DFMS datawere collected at a few tens of kilometers above the cometarysurface, parent species cannot be less abundant than daughterspecies. An exemplary estimation of HCN photodissociation atcometocentric distances relevant for Rosetta and at heliocentricdistances relevant for 67P may be found in Hänni et al. (2020).Also, the approximate factor of two between some radio and IRHCN production rates —compare for example Fray et al. (2005)to Dello Russo et al. (2016) — does not substantially changethe picture. On the contrary, the fact that the relative CN abun-dance from Hänni et al. (2020) is considered a lower limit ratherincreases the discrepancy. The fact that the situation graduallychanges towards the outbound equinox, where n CN / n HCN reachesa minimum, is due to the di ff erent heliocentric distance depen-dence of these species relative to HCN (cf. discussion aboveand Table 2). While the relative abundance of CN decreases,reaching a minimum around the time of the outbound equinoxin March 2016, the relative abundances of the nitriles studied inthis work peak around perihelion.
4. Discussion
Having derived a relative abundance of cyanogen for the firsttime and having followed this and the relative abundances oftwo other nitrile species during the full Rosetta mission phase,this section compares our results to other astrophysical environ-ments where the targeted species were observed. A selection ofvalues can be found in Table 1. For the following discussion, itis important to remember that absolute abundances of molecules with multiple isomers derived from ROSINA / DFMS data shouldbe considered as upper limits. This is due to the fact that otherisomers, besides the most likely one considered in the data eval-uation processes for this work, may contribute to the observedsignal. However, because this is the case for all of the investi-gated nitrile species (incl. HCN), the e ff ect is probably compen-sated at least partially when abundance ratios are considered. Unlike the smaller and less active JFC 67P, the large and activeOort cloud comet (OCC) Hale-Bopp shows very similar rela-tive abundances of HC N and CH CN. HC N is less abundantin 67P than in Hale-Bopp, while CH CN is more abundant closeto the Sun. Nevertheless, the full mission average of this latterspecies matches the value from Hale-Bopp exactly. Hale-Boppwas observed between 0.9 and 1.2 au and hence the solar fluxmust have been considerably higher than for 67P between 1.24and 1.74 au. This, together with the heliocentric distance depen-dence of n HC N / n HCN , could (at least partially) explain the higherHC N / HCN ratio for Hale-Bopp (at about 1 au) than for 67P (atabout 1.5 au). HCN itself is less abundant relative to water in67P than in Hale-Bopp by almost a factor of two, comparingRubin et al. (2019) and Bockelée-Morvan et al. (2000). Notably,the HCN abundance measured with radio spectroscopy is lowerthan the one measured with IR spectroscopy by a factor 1.4 forHale-Bopp, meaning that CH CN / HCN and HC N / HCN may beoverestimated (assuming that other nitriles are not similarly bi-ased).Another comet for which the relevant abundance ratios havebeen determined is the OCC C / Article number, page 6 of 10. Hänni et al.: NCCN, CH CN, and HC N in comet 67P and their relation to CN (IRAM), Biver et al. (2015) observed this comet at a heliocen-tric distance of 1.3 au, very similar to the perihelion distance of67P. The authors identified 21 molecules and retrieved produc-tion rates , including those of HCN, CH CN, and HC N. TheCH CN / HCN ratio in comet C / N / HCN ratio found bythese latter authors is roughly double that found in this work.Bockelée-Morvan & Biver (2017) review abundances of the tar-geted nitriles relative to water by giving a range for a sample ofmore than ten comets. Interestingly, the range of relative abun-dances reported for HC N / H O includes and exceeds the rangeof relative abundances reported for CH CN / H O. This overallpicture does not change when the comparably narrow range ofvalues of HCN abundances relative to water is taken into ac-count: There is a (smaller) portion of comets where HC N ismore abundant than CH CN and a (larger) portion where it issimilarly or less abundant. 67P and C / N in comets is under debate.
A similar scenario was deduced from observations of a sam-ple of protoplanetary disks by Bergner et al. (2018): while theabundances of CH CN relative to HCN cover a narrow range ofroughly 2 to 6 per cent, those of HC N relative to HCN cover abroad range of a few to more than 120 per cent (cf. their Fig. 12).For CH CN / HCN, the assumption of lower rotational tempera-tures leads to values more similar to what is observed for comets,while for HC N / HCN the opposite is the case. This leads to diskaverage abundances of CH CN relative to HCN and for a ro-tational temperature of 30 K that are comparable to the Rosettafull mission average reported here. The average from the comet’ssouthern hemisphere (that is from around the perihelion of 67P)is somewhat higher. However, for HC N / HCN, even the valuesfor the highest rotational temperature of 70 K are still an order ofmagnitude higher than the values we report for 67P. This deple-tion of HC N relative to HCN, when comparing 67P to Bergner’sdisk sample, remains unexplained, although it is observed forother comets, such as for instance for comet C / N / HCN in one of the disks is 50 K,making it reasonable that a rotational temperature of 30 K leadsto overestimation of the ratio. However, comparison of the ni-trile abundances in 67P to those in the vicinity of the low-massprotostar IRAS 16293-2422 B (Drozdovskaya et al. 2019) leadsus to the conclusion that, while HC N / HCN is similarly abun-dant, CH CN / HCN is overabundant in the protostar. Generally,comparison to disk observations is limited by the fact that oftenthe midplane, where comets are most likely formed, cannot beobserved due to optical thickness.Unfortunately, direct remote spectroscopic observations ofNCCN are hampered by its lack of allowed rotational transi-tions and very low vibrational band strength (Crovisier 1987).Therefore, Agúndez et al. (2015, 2018) proposed the use ofNCCNH + (protonated cyanogen) or CNCN (isocyanogen) as The model used by Biver et al. (2015) to retrieve production ratesincludes neutral and electron collisions. proxies. Based on the detection of CNCN, NCCNH + , C N,CH CN, C H CN, and H CN in the prestellar core L1544, Vas-tel et al. (2019) developed a chemical network to explain theobserved abundances. The main cyanogen production pathwaysconsidered in the network are the CN + HNC and N + C Nreactions in the gas phase. Their calculations yield an NCCNgas phase abundance a factor 100 larger than that of its isomerCNCN. Although the modeled column density of NCCN in thegas phase is comparable to that of HCN, its modeled grain sur-face abundance seems to be relatively low. Judging from whatwe learned about comet 67P, a relatively low grain surface abun-dance of NCCN with respect to HCN seems reasonable.
The detection of cyanogen in Titan’s atmosphere by the IR spec-trometer IRIS onboard Voyager 1 (Kunde et al. 1981) was laterconfirmed based on data from the Cassini mission (e.g., Teanbyet al. 2006, 2009; Magee et al. 2009; Cui et al. 2009). Abun-dances of cyanogen, and of HCN, HC N, and CH CN, which arealso observable with Atacama Large Millimeter / submillimeterArray (ALMA; e.g., Thelen et al. 2019; Cordiner et al. 2019;Lellouch et al. 2019), were consistently observed to underlaystrong latitudinal and vertical variations. In Titan’s atmosphere,a specific nitrogen(N )- and methane(CH )-dominated chem-istry is at play. Ultraviolet photons from the Sun, photoelec-trons, and magnetospheric electrons dissociate molecular nitro-gen and methane in the mesosphere and thermosphere into rad-icals, which may recombine to form nitriles (Lara et al. 1996;Wilson & Atreya 2004). Loison et al. (2015) proposed the for-mation of NCCN from HCCN + N, in addition to the previous re-actions. A photochemical network of the neutral reactions yield-ing nitrile and imine species is shown in their Fig. 8. Thanks toROSINA / DFMS, onboard the Rosetta spacecraft, comet 67P hasnow become the second environment in our Solar System wherecyanogen has been firmly detected. Obviously, we cannot expectto observe neutral–neutral or neutral–ion coma chemistry at 67P.This is because (i) the densities at cometocentric distances of afew tens of kilometers are too low (and hence the mean free pathtoo large) for a relevant amount of collisions to occur, and (ii) thesublimating species have little time to be photoionized on theiroutbound trajectory up to such distances. But also at the surface,chemistry seems unlikely (Fuselier et al. 2015, 2016).
5. Summary and Conclusions
The most relevant findings of this work can be summarized asfollows:1. A time-variable signal of C N was identified inROSINA / DFMS data. Assuming the most stable iso-mer to be responsible for this signal, it may be assigned tocometary cyanogen NCCN.2. NCCN correlates well with CHN (presumably HCN) inMarch 2016. Thanks to this correlation, several other peri-ods with NCCN signals above the detection limit were iden-tified and abundances of NCCN relative to HCN were de-rived. Within 1.24-1.74 au, the average ratio n NCCN / n HCN isequal to 0.0018 ± N andCH CN, were derived with respect to HCN as detailed inRubin et al. (2019). The average ratio n HC N / n HCN is equal
Article number, page 7 of 10 & A proofs: manuscript no. 39580final to 0.0046 ± n CH CN / n HCN is equal to 0.12 ± ff erent heliocentric distance de-pendencies. Furthermore, e ff ects of possible contributions ofother structural isomers (which may fragment di ff erently)cannot be fully ruled out.5. Neither NCCN nor HC N are abundant enough (relative toHCN) to explain the CN radical density (relative to HCN)observed with ROSINA / DFMS and reported by Hänni et al.(2020). The search for CN parent species has to be contin-ued.Presenting evidence against several nitriles as possible par-ents of the cyano radical in comet 67P, this work may help futureinvestigations of the origin of the cometary CN to pursue morefruitful leads.
Acknowledgements.
We gratefully acknowledge the work of the many engineers,technicians and scientists involved in the Rosetta mission and in the ROSINAinstrument in particular. Without their contributions, ROSINA would not haveproduced such outstanding results. Rosetta is an ESA mission with contributionsfrom its member states and NASA. Work at the University of Bern was funded bythe Canton of Bern and the Swiss National Science Foundation (200020 182418).S.F.W. acknowledges the financial support of the SNSF Eccellenza ProfessorialFellowship PCEFP2_181150. Especially, we thank Dr. N. Fray and Dr. M. N.Drozdovskaya for engaging in fruitful discussions on the topic.
References
Agúndez, M., Cernicharo, J., de Vicente, P., et al. 2015, A&A, 579, L10Agúndez, M., Marcelino, N., & Cernicharo, J. 2018, ApJ, 861, L22A’Hearn, M., Millis, R., Festou, M., et al. 1983, IAU Circ., 3802, 1Altwegg, K., Balsiger, H., & Fuselier, S. A. 2019, ARA&A, 57, 113Balsiger, H., Altwegg, K., Bochsler, P., et al. 2007, Space Sci. Rev., 128, 745Bergner, J. B., Guzmán, V. G., Öberg, K. I., Loomis, R. A., & Pegues, J. 2018,ApJ, 857, 69Biver, N., Bockelée-Morvan, D., Moreno, R., et al. 2015, Science Advances, 1,1500863Bockelée-Morvan, D., Crovisier, J., Baudry, A., et al. 1984, A&A, 141, 411Bockelée-Morvan, D. & Biver, N. 2017, Philosophical Transactions of the RoyalSociety of London Series A, 375, 20160252Bockelée-Morvan, D. & Crovisier, J. 1985, A&A, 151, 90Bockelée-Morvan, D., Lis, D. C., Wink, J. E., et al. 2000, A&A, 353, 1101Botschwina, P. & Sebald, P. 1990, Chemical Physics, 141, 311Calmonte, U. M. 2015, Sulfur isotopic ratios at 67P / Churyumov-Gerasimenkoand characterization of ROSINA-DFMS FM and FSCordiner, M. A., Teanby, N. A., Nixon, C. A., et al. 2019, AJ, 158, 76Crovisier, J. 1987, A&AS, 68, 223Crovisier, J. 1994, J. Geophys. Res., 99, 3777Cui, J., Yelle, R. V., Vuitton, V., et al. 2009, Icarus, 200, 581De Keyser, J., Altwegg, K., Gibbons, A., et al. 2019, International Journal ofMass Spectrometry, 446, 116232Dello Russo, N., Kawakita, H., Vervack, R. J., & Weaver, H. A. 2016, Icarus,278, 301Dello Russo, N., Vervack, R. J., J., Weaver, H. A., et al. 2009, ApJ, 703, 187Drozdovskaya, M. N., van Dishoeck, E. F., Rubin, M., Jørgensen, J. K., & Al-twegg, K. 2019, MNRAS, 490, 50Fray, N., Bénilan, Y., Cottin, H., Gazeau, M. C., & Crovisier, J. 2005,Planet. Space Sci., 53, 1243Fuselier, S. A., Altwegg, K., Balsiger, H., et al. 2016, MNRAS, 462Fuselier, S. A., Altwegg, K., Balsiger, H., et al. 2015, A&A, 583, A2Halpern, J. B., Miller, G. E., Okabe, H., & Nottingham, W. 1988, J. Photochem.Photobiol. A, 42, 63Haser, L. 1957, Bulletin de la Societe Royale des Sciences de Liege, 43, 740Herbst, E. & van Dishoeck, E. F. 2009, ARA&A, 47, 427Hänni, N., Altwegg, K., Pestoni, B., et al. 2020, Monthly Notices of the RoyalAstronomical Society, 498, 2239Huebner, W. F., Keady, J. J., & Lyon, S. P. 1992, Ap&SS, 195, 1Kanda, K., Nagata, T., & Ibuki, T. 1999, Chemical Physics, 243, 89Kunde, V. G., Aikin, A. C., Hanel, R. A., et al. 1981, Nature, 292, 686Lara, L. M., Lellouch, E., López-Moreno, J. J., & Rodrigo, R. 1996, J. Geo-phys. Res., 101, 23261 Läuter, M., Kramer, T., Rubin, M., & Altwegg, K. 2020, MNRAS, 498, 3995Le Roy, L., Altwegg, K., Balsiger, H., et al. 2015, A&A, 583, A1Lellouch, E., Gurwell, M. A., Moreno, R., et al. 2019, Nature Astronomy, 3, 614Loison, J., Hébrard, E., Dobrijevic, M., et al. 2015, Icarus, 247, 218Magee, B. A., Waite, J. H., Mandt, K. E., et al. 2009, Planet. Space Sci., 57, 1895Maquet, L. 2015, A&A, 579, A78Mattauch, J. & Herzog, R. 1934, Zeitschrift fur Physik, 89, 786Mumma, M. J. & Charnley, S. B. 2011, ARA&A, 49, 471Pandya, S. H., Shelat, F. A., Joshipura, K. N., & Vaishnav, B. G. 2012, Interna-tional Journal of Mass Spectrometry, 323-324, 28Rubin, M., Altwegg, K., Balsiger, H., et al. 2019, MNRAS, 489, 594Schroeder, I. R. H. G., Altwegg, K., Balsiger, H., et al. 2019, A&A, 630, A29Steins, S. E. 2018, Chemistry WebBook, NIST Standard Reference DatabaseNumber 69, National Institute of Standards and TechnologyStevenson, D. P. 1950, J. Chem. Phys., 18, 1347Swings, P. & Haser, L. 1956, Atlas of representative cometary spectraTeanby, N. A., Irwin, P. G. J., de Kok, R., et al. 2009, Icarus, 202, 620Teanby, N. A., Irwin, P. G. J., de Kok, R., et al. 2006, Icarus, 181, 243Thelen, A. E., Nixon, C. A., Chanover, N. J., et al. 2019, Icarus, 319, 417van Dishoeck, E. F. 2014, Faraday Discussions, 168, 9Vastel, C., Loison, J. C., Wakelam, V., & Lefloch, B. 2019, A&A, 625, A91Wilson, E. H. & Atreya, S. K. 2004, Journal of Geophysical Research (Planets),109, E06002Wurz, P., Rubin, M., Altwegg, K., et al. 2015, A&A, 583, A22
Article number, page 8 of 10. Hänni et al.: NCCN, CH CN, and HC N in comet 67P and their relation to CN
Appendix A: Results of the correlation analysis
The C N signal observed in March 2016, as shown in Fig. 2,was correlated to the signals of the other species analyzed byRubin et al. (2019). The results are listed in the following Ta-ble A.1: Table A.1.
Results of the correlation analysis performed based on datafrom March 2016.
Species R( c C N ) ( a ) c H O c CO c CO ( b ) c O c H S c NH c CH c CH O c C H c CH OH c CHN c SO c CH CHO c OCS c HNCO c C H c CH CN c CS c CH NO c C H O c C H c HC N Notes. ( a ) The resulting Pearson correlation coe ffi cients (R) are arrangedaccording to decreasing bulk abundance relative to water of the isomerconsidered by Rubin et al. (2019) in their Table 2. ( b ) As c CO sometimesis dominated by the CO fragment of CO , this correlation coe ffi cientmay not be representative. Appendix B: Neutral density ratio data for cyanogen
For the convenience of the reader, we list the results of our analy-sis regarding cometary cyanogen separately, namely the neutraldensity ratio n NCCN / n HCN as a function of the Rosetta missiontime. These data were shown in Fig. 3 (bottom).
Article number, page 9 of 10 & A proofs: manuscript no. 39580final
Table B.1.
Neutral density ratio of cyanogen and hydrogen cyanide n NCCN / n HCN as a function of the Rosetta mission time. r h is the heliocentricdistance. Date r h [au] n NCCN / n HCN
Stat. error [%] No. of spectra averaged ( a ) Detector row ( b ) . × −
40 1 A2014-10-24 3.12 2 . × −
29 1 A2015-05-06 1.70 2 . × −
33 1 A2015-05-07 1.69 1 . × −
41 1 A2015-07-30 1.25 1 . × −
29 11 A + B2015-11-22 1.72 5 . × −
44 2 A2016-03-14 2.57 1 . × −
22 15 B2016-03-15 2.58 1 . × −
19 15 B2016-03-16 2.59 1 . × −
20 12 B2016-07-18 3.42 6 . × −
25 2 A
Notes. ( a ) We include this value to indicate the statistical basis of the individual data points. ( b ) Of the two redundant micro-channel plate / linearelectron detector array (MCP //