EDP Sciences
Free Access
Issue
A&A
Volume 536, December 2011
Article Number L12
Number of page(s) 4
Section Letters
DOI https://doi.org/10.1051/0004-6361/201118189
Published online 19 December 2011

© ESO, 2011

1. Introduction

Titan is the only satellite of the Solar System to possess a dense atmosphere (1.5 bar), mainly composed of N2 with a few percent of CH4. The detection of hydrogen cyanide (HCN) and of more complex nitriles (HC3N and C2N2) obtained by IR spectrometers onboard spacecrafts (starting with Voyager in 1980) was of considerable interest, as these molecules are key intermediates in the synthesis of organic molecules. Their presence, along with that of several hydrocarbons, implies a complex photochemistry of methane, which is, on the one hand, coupled with that of N2, and on the other hand, enriched by ion-molecule reactions taking place in the upper atmosphere (>800 km). As a matter of fact, and as one of the major discoveries of the Cassini mission at Titan, numerous heavy organic molecules have been detected in Titan’s upper atmosphere by Cassini Ion Neutral Mass Spectrometer (INMS), (e.g. Waite et al. 2005; Vuitton et al. 2007). Despite these impressive results, from which a completely new view of Titan’s ion-neutral chemistry emerged, INMS cannot distinguish between species of identical mass, in particular between isomers. Heterodyne spectroscopy of strong rotational lines provides high sensitivity and frequency discrimination for the detection of minor species with permanent dipole moment – such as HCN, HC3N and CH3CN (Marten et al. 2002; Gurwell 2004), and CO (Gurwell & Muhleman 1995; Hidayat et al. 1998; Rengel et al. 2011) – and allows for full resolution of line profiles, as well as unique absolute wind measurements in Titan’s stratosphere and mesosphere from mapping of the Doppler shift in spatially-resolved data (Moreno et al. 2005).

In this paper, we report the first detection of a new species, hydrogen isocyanide (HNC), in Titan’s atmosphere, using the Heterodyne Instrument for the Far-Infrared (HIFI, de Graauw et al. 2010) onboard the ESA Herschel Space Observatory (Pilbratt et al. 2010).

2. Observations

Observations of Titan were performed on June 14 and December 31, 2010, as part of the Herschel guaranteed time key program “Water and related chemistry in the Solar System” (HssO, see Hartogh et al. 2009). We used the HIFI heterodyne receiver in band 1a (covering the 480–560 GHz spectral region), with two orthogonal polarizations (H and V) simultaneously.

The original goal of these observations was to study water vapor in Titan’s atmosphere. The receiver was therefore tuned to the H2O(110–101) rotational transition at 556.936 GHz, in the upper sideband (USB). In June 2010, the local oscillator (LO) frequency of the instrument was tuned to 550.390 GHz. Because HIFI is a double sideband (DSB) receiver, its lower sideband (LSB) simultaneously covered the 542.4–546.4 GHz range. For spectral analysis, we used two different spectrometers at a spectral resolution of 1.1 MHz (Wide Band Spectrometer, WBS) and 0.25 MHz (High Resolution Spectrometer, HRS). The telescope half power beam width (HPBW) at 556 GHz is 39′′. To minimize the contribution from Saturn, Titan was observed at eastern elongation (173′′ east of Saturn) using a position switch (PSw) observing mode. The reference sky position was chosen to be separated from Titan by +2′ in declination (i.e. North of Titan) in order to avoid any contamination from Saturn. Data reduction was carried out using the Herschel data reduction software (HIPE, Ott 2010), which calibrates the data flux and corrects for spacecraft and planet velocity. A polynomial baseline removal was performed to eliminate the standing waves.

In addition to the water line (which will be reported separately), we unexpectedly detected in the LSB a narrow emission line at 543.897 GHz (Fig. 1), which we identified as the HNC(6–5) transition, representing the first detection of HNC in Titan’s atmosphere. To confirm the detection, additional observations of Titan were obtained on December 31, 2010, this time at western elongation. To increase the signal-to-noise ratio (S/N), instead of position switch we used the standard frequency switch (FSw) observing mode, with a frequency throw of 94.5 MHz. To confirm that the line was indeed from HNC(6–5), we used a slightly different tuning, 1 changing the LO frequency to 550.325 GHz. The line was again unambiguously detected, separately on each of the two spectrometers (Fig. 1) at the frequency of HNC(6–5). For both observational epochs, the HNC emission is narrow (FWHM ~ 0.8 km s-1, or 1.5 MHz). The apparent diameter of Titan on December 31 was  ~ 6% smaller than on June 14 (Table 1). After scaling the HNC line intensity to the June 14 conditions (Fig. 1), the line intensity ratio between the two epochs (June/Dec.) is 1.02  ±  0.10 implying no significant intensity variation over this time period or between the leading and trailing sides.

thumbnail Fig. 1

Observations of HNC(6–5) at 543.897 GHz on Titan measured with Herschel/HIFI. Both H and V polarizations are averaged. 1a) Upper left panel: measurement with WBS (Δν = 1.1 MHz) on June 14, 2010. 1b) Upper right panel: measurement with HRS (Δν = 0.25 MHz) on June 14, 2010. 2a) Lower left panel: measurement with WBS on Dec. 31, 2010. 2b) Lower right panel: measurement with HRS on Dec. 31, 2010. Due to varying distance of Titan, observations on Dec. 31 are scaled in intensity to the conditions of June 14.

Open with DEXTER

Table 1

Summary of HIFI observations of Titan.

3. Modeling

The HNC spectra were analyzed with a line-by-line radiative transfer code accounting for the spherical geometry of Titan’s atmosphere (Marten et al. 2002, and references therein). This model considers HNC molecular opacities from the JPL catalog (Pickett et al. 1998), and also includes collision-induced opacities of N2–N2, N2–CH4, and CH4–CH4 (Borysow & Frommhold 1986, 1987; Borysow & Tang 1993). The pressure broadening coefficient (γ0) of HNC(6–5) was taken to be equal to that of HCN(6–5), γ0 = 0.129 cm-1 bar-1 at a reference temperature of 300 K and with a temperature dependency exponent n = 0.69 (Yang et al. 2008). The thermal profile used for our computation is a combination of (i) the temperatures measured by Huygens/HASI (Fulchignoni et al. 2005) in the troposphere at altitudes between 0–140 km; (ii) the disk-averaged Cassini/CIRS stratospheric temperatures (Vinatier et al. 2010) at altitudes between 140–500 km; (iii) the Cassini/INMS retrieved temperatures (i.e. 155 K in average, De La Haye et al. 2007) at altitudes between 1000–1500 km; (iv) a decreasing temperature from 165 K to 155 K at altitudes between 500 and 1000 km. Our modeling results, presented below, indicate that the HNC(6–5) line is optically thin (opacity τ ~ 0.12), therefore results are not too sensitive to the precise temperature profile.

The model computes spectra in units of fluxes or Rayleigh-Jeans temperatures (Trj). To compare Trj with the measured antenna temperature (Ta), we use the usual relationship: Ta = Trj/Fd × Beff/Feff, with Fd the geometrical dilution factor, Beff the telescope beam efficiency (i.e. 0.75), and Feff the telescope forward efficiency (i.e. 0.96). Because the HRS spectra are much noisier (due to their higher spectral resolution) than the WBS data, we focussed on the WBS spectra, averaging the two epochs to further improve the S/N.

3.1. Uniform mixing ratio profiles

The HNC narrow linewidth indicates that the line is mostly Doppler-broadened, with little or no contribution from Lorentz linewings. In a first series of models, we considered uniform vertical distributions of HNC, i.e. with a constant mixing ratio q0 above a given altitude z0. In this approach, one can discard HNC profiles with z0 smaller than 200 km, because the associated simulated lines are broader than the observed line (i.e. pressure-broadened, Fig. 2). While a value of z0 equal to 300 km is still acceptable, we favor uniform profiles in which z0 is greater than 400 km. This, in particular, is also consistent with HNC being restricted to the upper thermosphere (e.g. z0 = 1000 km). For such uniform distributions characterized by z0, the probed altitudes mainly lie between z0 and z0 + 100 km. Best–fit results of the coupled parameters z0 and q0 are given in Table 2. The retrieved column density depends slightly on z0, and lies in the range (0.6–1.5)  ×  1013 cm-2 for z0 varying between 400 and 1000 km, with higher values of z0 associated with smaller columns.

3.2. A thermospheric case

As illustrated above, our observations do not establish the vertical distribution of HNC, except for the loose constraint that it cannot be present in large amounts below 300 km. Nevertheless, as we discuss below, formation scenarios argue for a large fraction of HNC to be restricted to the thermosphere, with its distribution roughly following that of the electronic density in the lower ionosphere. To mimick this situation, we tested the case in which HNC is restricted to a layer from 900 to 1200 km, with a constant number density (i.e. a mixing ratio (q) increasing with altitude as q = q0 × (p0/p), with p0 the reference pressure corresponding to z0 = 900 km). For altitudes lower than z0, the mixing ratio is taken as zero. Taking into account the high molecular diffusion above the homopause at  ~1200 km (D = 0.5 −  − 1 × 1010 cm2 s-1, De La Haye et al. 2007; Yelle et al. 2008), we also assumed a constant mixing ratio above the homopause zh = 1200 km. With this model, the best fit is obtained for q0(900 km) = (2.0. It corresponds to an HNC column density of  cm-2, close to the one retrieved for uniform A or B cases (Table 2). Extending this approach to other values of z0, we also computed models with increasing mixing ratio with altitude as q = q0 × (p0/p)n. Many possible fits were found for z0 between 300–1000 km, further demonstrating that we cannot currently constrain the vertical distribution of HNC. The retrieved column density remains unchanged at (0.6–1.5)  ×  1013 cm-2.

thumbnail Fig. 2

Models of the HNC(6–5) line, assuming an uniform mixing ratio distribution of HNC above altitudes of 1000, 400, 300, and 200 km, corresponding to cases A, G, H, and I, respectively, in Table 2. These models are compared with the averaged WBS spectrum from June/Dec.

Open with DEXTER

4. Discussion

The possible presence of HNC in Titan’s atmosphere was first considered by Petrie (2001). Based on pre-Cassini photochemical models (e.g. Ip 1990; Galand et al. 1999; Banaszkiewicz et al. 2000) which predicted that HCNH+ is one important or even the dominant ion at the ionospheric peak, Petrie (2001) postulates the dissociative recombination of HCNH+ (HCNH+ + e  →  HNC + H) as an important source of HNC. This reaction, which produces HCN and HNC in approximately equal yields, was suggested as the origin of the high abundance of HNC (HNC  ~  HCN) in dense molecular clouds as early as 35 years ago (Watson 1976). Petrie (2001) further considered various loss processes for HNC, including (i) photon-induced isomerization (HNC + hν  →  HCN); (ii) photolysis (HNC + hν  →  H + CN); (iii) H-catalyzed isomerization (HNC + H  →  HCN + H); (iv) protonation with ionospheric cations (XH+ + HNC  →  X + HCNH+); and (v) reactions with neutral radicals (X + HNC  →  XCN + H or XH + CN). He dismissed (i) and (ii) as being insignificant compared to protonation and retained the other three loss mechanisms, focussing on process (v) with X = CH3. Using this limited chemistry, HNC would therefore result from the following sets of reactions: The detection of HCNH+ by INMS with a peak abundance of  ~103 cm-3 near 1100–1150 km (e.g. Waite et al. 2005; Vuitton et al. 2007; Cui et al. 2009) supports the above production scheme and allows one to evaluate it quantitatively, although key reaction rates (ki) are still uncertain. The dissociative recombination rate of HCNH+ depends on the electron temperature. The Cassini Radio and Plasma Wave Science (RPWS/LP) measurements have reported an electronic temperature of 500 K (Ågren et al. 2009; Galand et al. 2010) over 950–1100 km. Such a value is surprisingly high, because electronic temperatures closer to the  ~150 K gas temperature are expected below  ~1050 km (Richards et al. 2011). We therefore considered both Te = 150 K and Te = 500 K, giving k1a = k1b = 1.5 × 10-7 cm3 s-1 and 6.90 × 10-8 cm3 s-1, respectively (Semaniak et al. 2001).

Table 2

Retrieved HNC mixing ratio and column density assuming uniform profiles above altitude z0.

As discussed by Petrie (2001), protonation is likely to involve virtually all hydrogen-bearing ions, so it is reasonable to adopt a total XH+ concentration equal to the electron density. In particular, the HCNH+ + HNC reaction has been investigated theoretically, and suggested to be a significant loss channel for HNC in molecular clouds (Pichierri 2002). To our knowledge however, no measurements of the protonation rates for HNC are available. Petrie uses k2 = 3.5 × 10-9 cm3 s-1. The Su-Chesnavich formulae (see Woon & Herbst 2009) would give k2 = 5.2 × 10-9 cm3 s-1 at 150 K. As this value seems high compared to typical protonation rates (Vigren, priv. comm.), we here considered k2 = 3.5 × 10-9 cm3 s-1 and k2 = 1 × 10-9 cm3 s-1. With a peak electron density of  ~2000 cm-3 (Wahlund et al. 2009), this gives a loss rate of (2–7)  × 10-6 s-1.

Following Talbi et al. (1996) we adopt a k3 = 1 × 10-14 cm3 s-1 reaction rate (at 150 K) for the H-catalyzed isomerization reaction (3). Based on various models (e.g. Lara et al. 1996; De La Haye et al. 2008), the [H] mixing ratio near 1100 km is in the range (4–10)  × 10-4. Adopting 4 × 10-4 from De La Haye et al. (2008), which gives [H]  ~  1.3 × 106 cm-3 at 1100 km, implies a loss rate of  ~1.3 × 10-8 s-1 through reaction (3), much lower than the above losses from ion-neutral reactions. Regarding the CH3 + HNC (4) reaction rate, Petrie (2001) considered two extreme values (k4 = 10-14 and k4 = 5 × 10-11 cm3 s-1), but in a more recent study, Petrie & Osamura (2004) indicate a much lower rate (9.2 × 10-20 cm3 s-1 at 200 K). With this value and a CH3 concentration of  ~2 × 106 cm-3 at 1100 km (e.g. Hörst et al. 2008), the associated loss is entirely insignificant, and the HNC abundance is thus controlled by reactions (1a) and (2).

With the above scenario, the chemical lifetime of HNC is (1.4–5)  ×  105 s, i.e. 3 to 10 times shorter than Titan’s day and comparable to the day-to-night transport time (about 2 × 105 s for  ~ 50 m/s thermospheric winds, Müller-Wodarg et al. 2008). HCNH+ shows strong diurnal variations with maximum (dayside) concentrations of  ~1000 cm-3 at 1000–1200 km (Cui et al. 2009), and about five times less on the nightside, implying that (unobservable so far) diurnal variations of HNC can be expected. The HNC chemical lifetime is comparable to the transport timescale (1.5 × 105 s at 1000 km, for a  ~3 × 108 cm2 s-1 diffusion coefficient, Yelle et al. 2008) As a result, HNC in the ionosphere is affected by both chemical and transport losses, but its distribution may be reasonably close to photochemical equilibrium. In this framework, and considering only the protonation loss with [XH+] = [e], the HNC concentration can be simply written as [HNC] = k1a/k2 ×  [HCNH+], where the equality holds either for concentrations or column densities. With an HCNH+ dayside (relevant for our purpose) column density of 3.5 × 1010 cm-2, and using each of these two values considered for k1a and for k2, we obtain an HNC column of 7 × 1011–5.2 × 1012 cm-2. The upper range is consistent with the measurements, although somewhat marginally.

The above calculations only consider HNC production from HCNH+, while any other ion that includes the -CNH+ group may lead to HNC upon recombination. HCNH+ is the dominant ion measured by INMS, followed by C2H5+, HC3NH+, c-C3H3+, C3H5+ (Cui et al. 2009), but RPWS and the Cassini Plasma Spectrometer (CAPS) measurements indicate that heavy ions beyond the mass range of INMS ( > 100 amu) contribute significantly, constituting  ~5% of the ionosphere near 1100 km and becoming even dominant (50-70 %) over 950–1000 km altitude (Crary et al. 2009; Wahlund et al. 2009). Nitrile ions heavier than HCNH+ have higher recombination rates, e.g. by factors 3 to 5 for HC3NH+ and other species measured by Vigren et al. (2011). Remarkably, “effective” recombination rates for Titan’s ionosphere, based on CAPS and RPWS/LP measurements (Galand et al. 2010), increase below 1200 km, reaching  ~5 × 10-6 cm3 s-1 at 1000 km, a behavior attributed to the change of composition below the ionospheric peak and the progressive onset of heavy ions. Since heavy nitrile ions in this altitude range could provide significant additional sources of HNC, we conclude that a purely ionospheric source may be quantitatively viable for HNC, provided the protonation rates are not too high.

Even if a primarily ionospheric production is assumed, HNC will not necessarily be restricted to the ionosphere, as it must be transported downward to some extent by eddy mixing. This will further increase its chemical lifetime, in relation to the decline of ion-molecule reactions. Petrie & Osamura (2004) find that in the neutral atmosphere, H-catalyzed isomerization (3) may be the main loss channel for HNC, with additional contributions from reactions with CN and C3N. Photolysis and photo-isomerization of HNC, albeit negligible at ionospheric levels, must also become significant there, especially the latter, which has a very low energy threshold (8510 Å, see Petrie 2001). Additional formation routes to HNC in the neutral atmosphere should also be investigated, in particular its production through N(4S) + 3CH2. This pathway produces equal amounts of HCN and HNC and has been invoked in the regulation of the HNC/HCN abundance in molecular clouds (Herbst et al. 2000). We leave these issues for future photochemical modeling, with the goal of matching HNC and HCN simultaneously. We note that, in the model where HNC is restricted to altitudes above 1000 km, its mole fraction is  ~6 × 10-5. Comparing with the HCN abundance from INMS (2.0 × 10-4, Vuitton et al. 2007) indicates HNC/HCN  ~ 0.3; thus, a significant contribution of HNC to the mass 27 signal is likely.

In the Solar System, HNC has been observed in many comets since its first discovery in comet Hyakutake, with HNC/HCN abundance ratios varying from  ≤ 2% to 20% (Irvine et al. 1996; Lis et al. 2008). The inverse correlation of this ratio with heliocentric distance argues for the production of HNC from the thermal degradation of organic grains heated by the Sun (Lis et al. 2008). A production of HNC from solid (haze) material in Titan’s atmosphere might therefore not be excluded, although this remains largely speculative. In this respect, and more generally to constrain photochemical models, determining the vertical profile of HNC and its abundance in the region of the main haze (300–500 km) would be important. Although observing other (and stronger) additional lines of HNC might help slightly, this will probably require limb observations from a Titan orbiter equipped with a submillimeter instrument (Lellouch et al. 2010).


Herschel is an ESA space observatory with science instruments provided by European-led Principal Investigator consortia and with important participation from NASA.

Acknowledgments

HIFI has been designed and built by a consortium of institutes and university departments from across Europe, Canada, and the United States under the leadership of SRON Netherlands Institute for Space Research, Groningen, The Netherlands, and with major contributions from Germany, France, and the US. We acknowledge very useful discussions with M. Galand, E. Vigren, E. Herbst, S. Petrie and O. Dutuit. L.M. Lara’s work has been supported by the Ministry of Innovation and Science through the project AyA 2009-08011.

References

All Tables

Table 1

Summary of HIFI observations of Titan.

Table 2

Retrieved HNC mixing ratio and column density assuming uniform profiles above altitude z0.

All Figures

thumbnail Fig. 1

Observations of HNC(6–5) at 543.897 GHz on Titan measured with Herschel/HIFI. Both H and V polarizations are averaged. 1a) Upper left panel: measurement with WBS (Δν = 1.1 MHz) on June 14, 2010. 1b) Upper right panel: measurement with HRS (Δν = 0.25 MHz) on June 14, 2010. 2a) Lower left panel: measurement with WBS on Dec. 31, 2010. 2b) Lower right panel: measurement with HRS on Dec. 31, 2010. Due to varying distance of Titan, observations on Dec. 31 are scaled in intensity to the conditions of June 14.

Open with DEXTER
In the text
thumbnail Fig. 2

Models of the HNC(6–5) line, assuming an uniform mixing ratio distribution of HNC above altitudes of 1000, 400, 300, and 200 km, corresponding to cases A, G, H, and I, respectively, in Table 2. These models are compared with the averaged WBS spectrum from June/Dec.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.