EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 512 (March-April 2010) A&A, 512 (2010) L8 Full HTML
Issue
A&A
Volume 512, March-April 2010
Article Number L8
Number of page(s) 6
Section Letters
DOI https://doi.org/10.1051/0004-6361/201014339
Published online 07 April 2010
A&A 512, L8 (2010)

LETTER TO THE EDITOR

Detection of CO in Triton's atmosphere and the nature of surface-atmosphere interactions[*]

E. Lellouch1 - C. de Bergh1 - B. Sicardy1,2,[*] - S. Ferron3 - H.-U. Käufl4

1 - LESIA, Observatoire de Paris, 5 place Jules Janssen, 92195 Meudon, France
2 - Université Pierre et Marie Curie, 4 place Jussieu, 75005 Paris, France
3 - ACRI-ST, 260 route du Pin Montard, BP 234, 06904 Sophia-Antipolis Cedex, France
4 - European Space Observatory, Karl-Schwarzschild-Strasse 2, 85748 Garching bei München, Germany

Received 1 March 2010 / Accepted 13 March 2010

Abstract
Context. Triton possesses a thin atmosphere, primarily composed of nitrogen, sustained by the sublimation of surface ices.
Aims. We aim at determining the composition of Triton's atmosphere to constrain the nature of surface-atmosphere interactions.
Methods. We perform high-resolution spectroscopic observations in the 2.32-2.37 $\mu $m range, using CRIRES at the VLT.
Results. From this first spectroscopic detection of Triton's atmosphere in the infrared, we report (i) the first observation of gaseous methane since its discovery in the ultraviolet by Voyager in 1989; and (ii) the first ever detection of gaseous CO in the satellite. The CO atmospheric abundance is remarkably similar to its surface abundance, and appears to be controlled by a thin, CO-enriched, surface veneer resulting from seasonal transport and/or atmospheric escape. The CH4 partial pressure is several times higher than inferred by Voyager. This confirms that Triton's atmosphere is seasonally variable and is best interpreted by the warming of CH4-rich icy grains as Triton passed southern summer solstice in 2000. The presence of CO in Triton's atmosphere also affects its temperature, photochemistry, and ionospheric composition. An improved upper limit on CO in Pluto's atmosphere is also reported.

Key words: planets and satellites: atmospheres - Kuiper belt: general

1 Introduction

Like Pluto, Neptune's satellite and probably former Kuiper-Belt object Triton possesses a tenuous, predominantly nitrogen atmosphere, in equilibrium with surface ices mostly composed of N2 and a variety of other species. The most volatile of these species, CH4 and CO, must be present in trace amounts in the atmosphere as well. However, depending on the precise mechanisms of surface-atmosphere interactions, the expected atmospheric abundances vary by orders of magnitude, and except for the detection of CH4 in the UV by Voyager in 1989, observations have been severely lacking. Progress in IR-detector technology has now made the remote study of thin and distant atmospheres possible. Following our observations of methane in Pluto's atmosphere (Lellouch et al. 2009), we report the first spectroscopic detection of Triton's atmosphere in the infrared.

2 VLT/CRIRES observations and CH4 and CO measurements

Spectroscopic observations of Triton were obtained on July 4, 2009, using the CRIRES infrared echelle spectrograph (Käufl et al. 2004) installed on ESO VLT (European Southern Observatory Very Large Telescope) UT1 (Antu) 8.2 m telescope. We focused on the regions of the (2-0) band of carbon monoxide and the $\nu_3$+$\nu_4$ band of methane, covering the 2318-2330, 2334-2345, 2349-2359, and 2363-2373 nm ranges. We used the instrument in adaptive optics mode with a slit of 0.4'', providing a mean spectral resolution of 60 000, and acquired spectra during $\sim$4 h. A large Doppler shift (-23 km s-1) ensured proper separation of the target lines from the telluric absorptions.

The resulting spectrum shows the detection of many lines due to methane in Triton's atmosphere, particularly at 2320-2330 nm (Fig. 1). This is the first observation of gaseous methane since its discovery by Voyager (Herbert & Sandel 1991). As for our study of Pluto's CH4, we constructed a direct line-by-line atmospheric model of Triton, integrated over angles and including solar lines reflected off Triton's surface as well as the telluric transmission (see details in Lellouch et al. 2009). The spectrum was first modelled by assuming a single-temperature layer, with Triton's atmospheric methane mean temperature (T) and column density (a) as free parameters. We inferred T = 50+20-15 K and a = $0.08~\pm$ 0.03 cm-am (Fig. 2 online). The same analysis for Pluto inferred T = 90+25-18 K and a = 0.75+0.55-0.30 cm-am. This confirms that Pluto's atmosphere is warmer than Triton's, as a result of its higher methane abundance.

\begin{figure} \par\includegraphics[width=5.5cm,angle=270,clip]{14339fg1.ps} \vspace*{-3mm} \end{figure} Figure 1:

Black: Triton spectrum at 2320-2330 nm observed by CRIRES/VLT. The spectral resolution is 60 000. Red and blue curves show synthetic spectra, including telluric and solar lines. Red: methane column density in Triton's atmosphere = 0.08 cm-am. Triton's thermal profile, based on Voyager measurements, is taken from Krasnopolsky (1993) and a Voyager-like vertical distribution is used for methane (Herbert & Sandel 1991, entrance profile). Blue: no methane. The continuum slope is caused by the red wing of the $\nu _3 + \nu _4$ band of solid methane. Here, as in Fig.3, the vertical unit approximately represents the geometric albedo (but uncorrected for telluric and solar lines). Models are shifted vertically by -0.07 for clarity.

Open with DEXTER

The error bars in the inferred mean methane temperature are too large to constrain the methane vertical distribution. Instead, we used the Voyager-determined thermal structure (temperature versus altitude, Krasnopolsky et al. 1993) and methane vertical profile (Herbert & Sandel 1991, ingress UV occultation profile). The latter shows a decrease in the CH4 mixing ratio with altitude with a scale height of $\sim$20 km, due to photolysis. We obtained the same column density as above, indicating a partial pressure of methane of $9.8 \pm 3.7$ nbar, i.e., a surface density of ( $1.9 \pm 0.7) \times 10^{12}$ cm-3. This appears to be 4 +5-2.5 times higher than inferred from Voyager in 1989, adopting the CH4 number densities of Herbert & Sandel (1991) and Strobel & Summers (1997) ( $4.7 \times 10 ^{11}$ cm-2 within a factor of 1.7, averaging ingress and egress). An even greater enhancement factor (5 +6-2) is found if the Krasnopolsky & Cruikshank (1995) reanalysis of the Voyager UV data, which gives CH $_4 = 3.1 \pm 0.8 \times 10^{11}$ cm-3 at the surface, is used. Results are independent of the surface pressure, as collisional broadening is negligible. They clearly demonstrate that the CH4 partial pressure has increased in the past 20 years.

The 2335-2365 nm part of the Triton spectrum (see excerpts in Fig. 3) shows the detection of 8 lines due to the CO(2-0) band (R2-R5, P2, P3, P5 and P8), providing the first detection of CO in its atmosphere. An accurate determination of the CO abundance is difficult, because at infinite spectral resolution, these features are very narrow, saturated Doppler-shaped lines. Nonetheless, assuming a vertically uniform CO distribution, and utilizing the whole set of CO lines (see Fig. 4 online), we determine a CO column of 0.30 cm-am, i.e., a CO partial pressure of 24 nbar, within a factor of 3. The column density CO/CH4 ratio is nominally $\sim$3.75 (surface partial pressure ratio CO/CH4 $\sim$2.5) with a factor of 4 uncertainty. Deriving the CO/N2 and CH4/N2 mixing ratio is complicated by the surface pressure in 2009 being unknown. Stellar occultation results (Olkin et al. 2007; Sicardy et al. 2003; Elliot et al. 1998, 2000a) indicate that the pressure has doubled in $\sim$10 years from the 14 $\mu $bar value determined by Voyager in 1989 (Gurrola 1998). A reasonable assumption for 2009 is 40 $\mu $bar, which provides CO/N $_2 \sim 6 \times 10^{-4}$ and CH4/N $_2 \sim 2.4 \times 10^{-4}$ at the surface, to within factors of 3 and 1.4, respectively. The CO abundance that we determine is many times lower than previous upper limits (Broadfoot et al. 1989; Young et al. 2001).

\begin{figure} \par\includegraphics[width=5.5cm,angle=270]{14339fg3.ps} \vspace*{-3mm} \end{figure} Figure 3:

Black: portions of the Triton spectrum at 2335-2340 nm and 2350-2355 nm. Five CO lines (R5, R4, R3, P2 and P3) are present in these spectral ranges. Red and blue curves show synthetic spectra, including telluric and solar lines, as well as Triton's methane. Red: CO column density in Triton's atmosphere = 0.30 cm-am. Blue: No CO. The mismatch in the ``continuum'' level over 2351-2354 nm is caused by absorption in the (2-0) band of CO ice (see e.g. Quirico et al. 1999; Grundy et al. 2010), not included in models.

Open with DEXTER

3 Discussion

Near-infrared observations have detected CO and CH4 on Triton's surface with mixing ratios of 0.05% and 0.1% relative to N2, and, at least for CH4, mostly in solid solution in N2 (Cruikshank et al. 1993; Quirico et al. 1999; Grundy et al. 2010). In this situation, the expected partial pressure of each species is the product of its solid mole fraction and its pure vapor pressure (Raoult's law for an ideal mixture). This scenario implies atmospheric abundances of CO and CH4 about 1 and 3 orders of magnitude lower than observed, respectively (Fig. 5). Although Henry's law may be more applicable to the N2-CH4 system than Raoult's, this does not reduce the discrepancy by more than a factor of 2-3. This problem has been studied in the case of Pluto's atmospheric methane, present at the 0.5% level (Young et al. 1997; Lellouch et al. 2009), i.e. also considerably enriched over its solid solution equilibrium value. The origin of this enhancement is thought to ultimately lie in the seasonal evolution of the N2-dominated solid solution. Preferential sublimation of N2 initially creates a thin surface layer enriched in the less volatile species. Further evolution of this layer may lead to either (i) the formation of chemically pure grains in vertically or geographically segregated deposits (``pure ice'' scenario); or (ii) the establishment of a homogeneous ``detailed balancing layer'' controlling the surface-atmosphere exchanges. In the ``pure ice'' case, the atmospheric mixing ratios are in simple proportion of the pure vapor pressures at the relevant ice temperatures (which may differ for different species) and, except for the main species controlling the pressure, of the fractional area covered by each ice. For CH4 on Pluto, Stansberry et al. (1996) demonstrated that pure CH4 lag deposits (whose existence has been proven by observations, see Douté et al. 1999) assume higher temperatures than N2 because of their reduced sublimation cooling and preferential formation in regions of high insolation. Even if covering only a few percent of Pluto's surface, these patches can explain the observed atmospheric abundance of methane. Alternatively, the ``detailed balanced'' model (Trafton 1990; Trafton et al. 1998) predicts that surface-atmosphere exchanges in the presence of atmospheric escape and seasonal transport lead to an atmospheric composition reflecting that of the accessible ice reservoir from which it is replenished. When no fractionation (e.g., diffusive) occurs during escape or transport, the atmospheric mixing ratios are identical to those in the volatile reservoir. This is accomplished by the thin surface veneer, enriched in the less volatile species, throttling off the N2 sublimation, and in permanent equilibrium with the atmosphere according to Raoult's law.

\begin{figure} \par\includegraphics[width=5.5cm,angle=270]{14339fg5.ps} \vspace*{-3mm} \end{figure} Figure 5:

Abundance measurements and equilibrium curves for Triton's volatiles. The black, pink, and light blue curves show the vapor pressure of pure N2, CO, and CH4 ices, calculated from Fray and Schmitt (2010). For CO and CH4, the thick curves (red and blue, respectively) show the partial pressures based on Raoult's law for an ideal solid solution with N2, with CO and CH4 respective abundances of $5 \times 10^{-4}$ and $1 \times 10^{-3}$ in the ice (Quirico et al. 1999). For CO and CH4, the thin red and blue lines show the partial pressures expected in the framework of the ``detailed balancing model'' (see text). The black square shows the 1989 Voyager pressure measurement (14 nbar), which corresponds to sublimation equilibrium of N2 ice at 37.25 K. The green symbol represents the Voyager-measured CH4 partial pressure (Herbert & Sandel 1991); it is plotted at $38 \pm 1$ K, the surface temperature measured by Voyager and the N2 ice temperature inferred by Tryka et al. (1994). The blue star and the red square show the CH4 and CO abundances determined in this work. They are plotted at a temperature of $39 \pm 1$ K. A temperature of 39 K corresponds to a plausible 40 $\mu $bar pressure for Triton's atmosphere in 2009; 40 K, which corresponds to a N2 vapor pressure of 65 $\mu $bar, is a reasonable upper limit, given the stellar occultation inferences that Triton's surface pressure has doubled in the 10 years following the Voyager epoch. The CO partial pressure we measure is consistent with expectations from the detailed balancing model, while CH4 is lower.

Open with DEXTER

\begin{figure} \par\includegraphics[width=6cm,angle=0]{14339fg6.eps} \vspace*{-3mm} \end{figure} Figure 6:

Atmospheric mixing ratios and composition of the ice boundary layer (``film'') in the detailed balancing model. ( Top): CO-N2 system. A surface pressure of 40 $\mu $bar is assumed, as estimated for 2009. The dark blue line is the CO/N2 mixing ratio expected for pure ices. The green curve shows the CO mole fraction in the ice surface film, and the black curve is the CO/N2 atmospheric mixing ratio derived from the composition of the ice film by applying Raoult's law. The range of CO/N2 atmospheric mixing ratios inferred in this work for this pressure ((2- $18) \times 10^{-4}$) is indicated by the blue-colored region. It implies a CO/N2 mixing ratio in the surface film of (1.4- $12) \times 10 ^{-3}$ (see text). The surface film is therefore largely dominated by N2, and the total pressure is defined by N2 equilibrium at 39.075 K. ( Bottom): CH4-N2 system. Calculations are here performed for a 14 $\mu $bar pressure, appropriate for the Voyager conditions. The Voyager-determined CH4/N2 atmospheric mixing ratio at the surface level ((1.1- $3) \times 10^{-4}$) is indicated by the yellow region. The colored lines have the same meaning as in the top panel, with CH4 replacing CO. The intersection of the black line with the colored area shows that explaining the observed CH4 mixing ratio and the total pressure requires a 38.3-39.6 K surface temperature and would imply a very high CH4 mole fraction (50-80%) in the surface film, well beyond the solubility limit of CH4 in N2. The same conclusion is reached if the CH4 amounts measured in 2009 are used. Note that these diagrams remain similar at other surface pressures, the only change being the required ice temperature to sustain the total pressure.

Open with DEXTER

Our observations provide discriminating keys between these scenarios. The case for CO is most straightforward because: (i) CO is not subject to diffusive separation from N2 upon escape; and (ii) as the ratio of its vapor pressure to that of N2 is largely insensitive to temperature in the relevant range (e.g., CO/N 2 = 0.112 at 36 K and 0.166 at 41 K), its atmospheric mixing ratio should be roughly seasonally constant. The observed CO abundance is over two orders of magnitude lower than the pure CO vapor pressure (Fig. 5), and one might envisage that the atmospheric CO results from the sublimation of pure CO patches covering $\sim$0.4% of the surface. However, we regard this scenario as unlikely. Although the pure vs. isolated form of CO on Triton's surface has not yet been proven by observations, the miscibility of CO and N2 in the solid phase in all proportions and the similarity of their vapor pressures argue for a co-condensation of the two species on Triton's surface. This is further supported by the remarkably similar longitudinal distribution of the N2 and CO ice bands at Triton (Grundy et al. 2010), strongly indicative of a spatially constant CO/N2 ice mixing ratio. In addition, even if pure CO patches occurred on Triton's surface, they would probably not be able to elevate the CO atmospheric abundance by means of the mechanism envisaged for Pluto methane. This is because CO is not buoyant in N2, restricting the dispersal of CO-rich gas in the background atmosphere and therefore the sublimation rates of the CO patches (Stansberry et al. 1996). The detailed balancing model instead provides a physically-expected interpretation of the atmospheric CO/N2 mixing ratio being consistent with its value in the ice phase (Fig. 5). Based on this scenario, the N2-CO composition of the surface boundary layer (``film'') can be established by simple application of Raoult's law, providing $q_{\rm CO}$(film) = $q_{\rm CO}$(atm)$\times$p $_{\rm sat}$(CO) / $p_{\rm sat}$(N2) (Trafton et al. 1998). Adopting again $p = 40~\mu$bar, our observed $q_{\rm CO}$(atm) = (2- $18) \times 10^{-4}$ indicates $q_{\rm CO}$(film) = (1.4- $12) \times 10 ^{-3}$. Therefore the surface veneer is still dominated by N2 and the presence of CO does not significantly modify the N2 atmospheric pressure, defined by equilibrium at 39.1 K (Fig. 6). Because it may be as thin as a few molecular layers, the surface film may not be visible in the near-IR spectra.

The case for CH4 is more complex. As previously realized (Cruikshank et al. 1993; Yelle et al. 1997; Strobel & Summers 1997; Strobel et al. 1996), the CH4 atmospheric mixing ratio at the surface measured by Voyager ($\sim$ $1.8 \times 10^{-4}$) is at least three orders of magnitude higher than expected for an ideal mixture. However, we note that it is also smaller, by a factor of $\sim$6, than the ice CH4/N2 mixing ratio, and as such does not agree with the detailed balancing model in its simplest form. Unlike CO, CH4 is subject to atmospheric photolysis and mass separation, and its vapor pressure is more temperature-dependent. This probably makes the surface/atmosphere abundance relationship for CH4 complex and seasonally variable. In any case, the phase diagram of N2-CH4 is not obviously consistent with the formation of a CH4-rich solid solution veneer (Stansberry et al. 1996). Explaining the range of observed CH4 atmospheric abundance would require a CH4 mole fraction in the surface film as high as 50-80 % (Fig. 6), well beyond the solubility limit of CH4 in N2 (Prokhvatilov & Yantsevich 1983). The formation of pure CH4 ice grains, decoupled from the mixture and not influencing its sublimation (Stansberry et al. 1996; Spencer et al.1997), further evolving into a lag deposit, may be a more plausible outcome. Using a Bond albedo of 0.85 (Triton's polar cap) and an emissivity of 0.7-1, a reasonable subsolar temperature for these pure methane patches is 45-48 K. Applying the Stansberry et al. (1996) Pluto model, we then find that methane patches covering 0.5-1% of Triton's surface are sufficient to maintain a $\sim$ $2 \times 10^{-4}$ atmospheric mixing ratio. Although there is no evidence of such patches in Triton's near-IR spectrum, the methane longitudinal distribution of CH4 ice differs from that of N2, and small areas of CH4-dominated ice, notably near 300$^{\circ}$ longitude, are not inconsistent with observations (Grundy et al. 2010). In contrast, the existence of widespead pure methane ice is ruled out; therefore the emphasized fact that the Voyager-measured methane partial pressure was consistent with vapor pressure equilibrium of pure CH4 ice at 38 K is probably coincidental.

After the Voyager encounter, a variety of seasonal N2 cycle models (see review in Yelle et al. 1997) were explored in attempting to explain Triton's visual appearance and the then measured surface pressure. These models, which essentially differed in terms of the assumed ice and substrate albedos and thermal inertia, had limited success, leaving unanswered the simple question of where the ice is on Triton. Yet, they made distinctive predictions about the short-term evolution of Triton's atmosphere. High thermal inertia models predicted a pressure increase as Triton approached and passed Southern summer solstice in 2000 (Triton subsolar latitude moved from 45.5 S in 1989 to a maximum 50 S in 2000 and 47 S in 2009). This is a consequence of increased insolation on, and attendant sublimation of, the Southern polar cap (Spencer & Moore 1992; Forget et al. 2003). In contrast, ``dark frost'' models (Hansen & Paige 1992) or low thermal inertia models predicted a pressure decrease from $\sim$1980 onwards, due to the exhaustion of the seasonal southern cap and re-condensation of N2 on the invisible winter pole. The discovery of the pressure increase in the 1990's, and the persisting signature of N2 and other ices in Triton reflectance spectrum with no obvious temporal evolution (Grundy et al. 2010), strongly argue that the bright deposits covering most of Triton southern hemisphere are relatively stable seasonal deposits. Our observation that the methane partial pressure has increased by a factor of $\sim$4 from 1989 to 2009 is qualitatively consistent with the reported pressure increase and the above interpretation. Since the CH4 vapor pressure varies more sharply with temperature than N2, we expect the atmospheric methane to be currently increasing more rapidly than pressure, but multi-volatile seasonal models will be needed to fully interpret our results. A direct measurement of Triton's current pressure is also highly desirable, and could be obtained by redeterminating the N2 ice temperature from its 2.15 $\mu $m band (Tryka et al. 1994).

The detection of CO also has implications for Triton's atmospheric thermal structure, photochemistry, and ionosphere. CO is an important cooling agent, by means of radiation in its rotational lines (Krasnopolsky et al. 1993; Strobel & Summmers 1997; Elliot et al. 2000b). It enriches atmospheric chemistry by introducing additional oxygen species (Krasnopolsky & Cruikshank 1995). Most importantly, it profoundly modifies ionospheric composition by providing a source of C atoms and C+ ions and by suppressing the N+ concentration at the benefit of CO+ and NO+ (see review in Strobel & Summers 1997). Although the error bar on the CO abundance is large, all previous considerations on the role of CO now have direct observational support.

During the same observing night, we also searched for CO in Pluto's atmosphere, covering the region of the (3-0) band near 1.57 $\mu $m. Only an upper limit (1 cm-am) was obtained. For a characteristic surface pressure of 15 $\mu $bar (Lellouch et al. 2009), this indicates CO/N2 < $5 \times 10^{-3}$. While an improvement over previous results (Bockelée-Morvan et al. 2001; Young et al. 2001), this upper limit does not provide a strong constraint when compared to the measured CO ice mole fraction ( $1 \times 10^{-3}$, Douté et al. 1999). Nonetheless, given the similarity of the two bodies, the above considerations about surface control of atmospheric CO at Triton should also apply to Pluto. In 2015, observations with the ALICE and Rex instruments on New Horizons will provide measurements of the surface pressure and CH4 and CO abundance in Pluto's atmosphere. We anticipate that CO will be measured at a uniform ratio of 0.001 and that the methane mixing ratio will show horizontal variability associated with local time and methane patch distribution.

Acknowledgements
This work is based on observations performed at the European Southern Observatory (ESO), proposal 383.C-0609. We thank F. Forget, B. Schmitt, and L. Trafton for illuminating discussions.

References

  1. Broadfoot, A. L., Atreya, S. K., Bertaux, J. L., et al. 1989, Science, 246, 1459 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  2. Bockelée-Morvan, D., Lellouch, E., Biver, N., et al. 2001, A&A, 377, 343 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  3. Cruikshank, D. P., Roush, T. L., Owen, T. C., et al. 1993, Science, 261, 742 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  4. Douté, S., Schmitt, B., Quirico, E., et al. 1999, Icarus, 142, 421 [NASA ADS] [CrossRef] [Google Scholar]
  5. Elliot, J. L., Hammel, H. B., Wasserman, L. H., et al. 1998, Nature, 393, 765 [NASA ADS] [CrossRef] [Google Scholar]
  6. Elliot, J. L., Person, M. J., McDonald, S. W., et al. 2000a, Icarus, 148, 347 [NASA ADS] [CrossRef] [Google Scholar]
  7. Elliot, J. L., Strobel, D. F., Zhu, X., et al. 2000b, Icarus, 143, 425 [NASA ADS] [CrossRef] [Google Scholar]
  8. Forget, F., et al. 2000, BAAS, 32, 45.01 [Google Scholar]
  9. Fray, N., & Schmitt, B. 2010, Planet. Space. Sci., in press [Google Scholar]
  10. Grundy, W. M., Young, L. A., Stansberry, J. A., et al. 2010, Icarus, 205, 594 [NASA ADS] [CrossRef] [Google Scholar]
  11. Gurrola, E. M. 1995, Ph.D. Thesis, Stanford University, Palo Alto [Google Scholar]
  12. Hansen, C. A., & Paige, D. A. 1992, Icarus, 99, 273 [NASA ADS] [CrossRef] [Google Scholar]
  13. Herbert, F., & Sandel, B. R. 1991, JGR, 96, 19241 [Google Scholar]
  14. Kaeufl, H.-U., Ballester, P., Biereichel, P., et al. 2004, SPIE, 5492, 1218 [Google Scholar]
  15. Krasnopolsky, V. A., & Cruikshank, D. P. 1995, JGR, 100, 21271 [Google Scholar]
  16. Krasnopolsky, V. A., Sandel, B. R., Herbert, F., et al., R.J. 1993, JGR, 98, 3065 [Google Scholar]
  17. Lellouch, E., Sicardy, B., de Bergh, C., et al. 2009, A&A, 495, L17 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  18. Olkin, C. B., Elliot, J. L., Hammel, H. B., et al. 1997, Icarus, 129, 178 [NASA ADS] [CrossRef] [Google Scholar]
  19. Prokhvatilov, A. I., & Yantsevich, L. D. 1983, Sov. J. Low Temp. Phys. 9, 94 [Google Scholar]
  20. Quirico, E., Doute, S., Schmitt, B., et al. 1999, Icarus, 139, 159 [NASA ADS] [CrossRef] [Google Scholar]
  21. Sicardy, B., et al. 1998, BAAS, 30, 49.02 [Google Scholar]
  22. Spencer, J. R., & Moore, J. M. 1992, Icarus, 99, 261 [NASA ADS] [CrossRef] [Google Scholar]
  23. Spencer, J. R., et al. 1997, in Pluto and Charon, ed. S. A. Stern, & D. J. Tholen (The University of Arizona Press), 435 [Google Scholar]
  24. Stansberry, J. A., Spencer, J. R., Schmitt, B., et al. 1996, Planet. Space Sci. 44, 1051 [NASA ADS] [CrossRef] [Google Scholar]
  25. Strobel, D. F., & Summers, M. E. 1995, in Neptune and Triton, ed. D.P. Cruikshank (The University of Arizona Press), 1107 [Google Scholar]
  26. Strobel, D. F., Zhu, X., & Summers, M. E. 1996, Icarus, 120, 266 [NASA ADS] [CrossRef] [Google Scholar]
  27. Trafton, L. M. 1990, ApJ, 359, 512 [NASA ADS] [CrossRef] [Google Scholar]
  28. Trafton, L. M., Matson, D. L., & Stansberry, J. A. 1998, in Solar System Ices, ed. B. Schmitt, C. de Bergh, & M. Festou (Kluwer Academic Publishers), 773 [Google Scholar]
  29. Tryka, K. A., Brown, R. H., Chruikshank, D. P., et al. 1994, Icarus, 112, 513 [NASA ADS] [CrossRef] [Google Scholar]
  30. Yelle, R. V., Lunine, J. I., Pollack, J. B., et al. 1995, in Neptune and Triton, ed. D.P. Cruikshank (The University of Arizona Press), 1031 [Google Scholar]
  31. Young, L. A., Elliot, J. L., Tokunaga, A., et al. 1997, Icarus, 127, 258 [NASA ADS] [CrossRef] [Google Scholar]
  32. Young, L. A., Cook, J. C., Yelle, R. V., et al. 2001, Icarus, 153, 148 [NASA ADS] [CrossRef] [Google Scholar]

Online Material

\begin{figure} \par\includegraphics[angle=270,width=16cm,clip]{14339fg2.ps} \end{figure} Figure 2:

Zoomed image of several CH4 lines showing sensitivity of the spectrum to Triton's methane. Some portions of the spectrum, especially over 2320-2326 nm, are relatively independent of temperature, while high-energy lines at 2353-2359 nm show increased temperature sensitivity. The top two panels show sensitivity to the methane abundance. Blue, red, and green synthetic spectra have 0.03, 0.08, and 0.20 cm-am of methane. Triton's thermal profile is taken from Krasnopolsky (1993) and a Voyager-like vertical distribution is used for methane (Herbert & Sandel 1991, entrance profile). Based on these models, the best fit methane column density is determined to be $0.08 \pm 0.03$ cm-am (i.e. $\pm $40%). The bottom two panels show sensitivity to methane temperature. The previous best-fit model using Voyager thermal profile is shown in pink. Other models assume an isothermal atmosphere with temperature of 30 K (dark blue), 50 K (red) and 80 K (green). These fits indicate a mean methane temperature of 50 +20-15 K.

Open with DEXTER

\begin{figure} \par\includegraphics[angle=270,width=16cm]{14339fg4.ps} \end{figure} Figure 4:

The eight detected CO lines from Triton's atmosphere. X-axis units are nm and Y-axis units are arbitrary. Lines are compared with models including 0, 0.03, 0.30, and 3 cm-am of CO. The slow change in absorption depth with change of column density is caused by heavy saturation of particularly narrow Doppler-shaped lines at $T \sim 50$ K. Based on these models, the best fit CO column density is determined to be 0.30 cm-am with a factor of 3 uncertainty.

Open with DEXTER

Footnotes

... interactions[*]
Figures 2 and 4 are only available in electronic form at http://www.aanda.org
...[*]
Senior member of the Institut Universitaire de France.

All Figures

  \begin{figure} \par\includegraphics[width=5.5cm,angle=270,clip]{14339fg1.ps} \vspace*{-3mm} \end{figure} Figure 1:

Black: Triton spectrum at 2320-2330 nm observed by CRIRES/VLT. The spectral resolution is 60 000. Red and blue curves show synthetic spectra, including telluric and solar lines. Red: methane column density in Triton's atmosphere = 0.08 cm-am. Triton's thermal profile, based on Voyager measurements, is taken from Krasnopolsky (1993) and a Voyager-like vertical distribution is used for methane (Herbert & Sandel 1991, entrance profile). Blue: no methane. The continuum slope is caused by the red wing of the $\nu _3 + \nu _4$ band of solid methane. Here, as in Fig.3, the vertical unit approximately represents the geometric albedo (but uncorrected for telluric and solar lines). Models are shifted vertically by -0.07 for clarity.

Open with DEXTER
In the text

  \begin{figure} \par\includegraphics[width=5.5cm,angle=270]{14339fg3.ps} \vspace*{-3mm} \end{figure} Figure 3:

Black: portions of the Triton spectrum at 2335-2340 nm and 2350-2355 nm. Five CO lines (R5, R4, R3, P2 and P3) are present in these spectral ranges. Red and blue curves show synthetic spectra, including telluric and solar lines, as well as Triton's methane. Red: CO column density in Triton's atmosphere = 0.30 cm-am. Blue: No CO. The mismatch in the ``continuum'' level over 2351-2354 nm is caused by absorption in the (2-0) band of CO ice (see e.g. Quirico et al. 1999; Grundy et al. 2010), not included in models.

Open with DEXTER
In the text

  \begin{figure} \par\includegraphics[width=5.5cm,angle=270]{14339fg5.ps} \vspace*{-3mm} \end{figure} Figure 5:

Abundance measurements and equilibrium curves for Triton's volatiles. The black, pink, and light blue curves show the vapor pressure of pure N2, CO, and CH4 ices, calculated from Fray and Schmitt (2010). For CO and CH4, the thick curves (red and blue, respectively) show the partial pressures based on Raoult's law for an ideal solid solution with N2, with CO and CH4 respective abundances of $5 \times 10^{-4}$ and $1 \times 10^{-3}$ in the ice (Quirico et al. 1999). For CO and CH4, the thin red and blue lines show the partial pressures expected in the framework of the ``detailed balancing model'' (see text). The black square shows the 1989 Voyager pressure measurement (14 nbar), which corresponds to sublimation equilibrium of N2 ice at 37.25 K. The green symbol represents the Voyager-measured CH4 partial pressure (Herbert & Sandel 1991); it is plotted at $38 \pm 1$ K, the surface temperature measured by Voyager and the N2 ice temperature inferred by Tryka et al. (1994). The blue star and the red square show the CH4 and CO abundances determined in this work. They are plotted at a temperature of $39 \pm 1$ K. A temperature of 39 K corresponds to a plausible 40 $\mu $bar pressure for Triton's atmosphere in 2009; 40 K, which corresponds to a N2 vapor pressure of 65 $\mu $bar, is a reasonable upper limit, given the stellar occultation inferences that Triton's surface pressure has doubled in the 10 years following the Voyager epoch. The CO partial pressure we measure is consistent with expectations from the detailed balancing model, while CH4 is lower.

Open with DEXTER
In the text

  \begin{figure} \par\includegraphics[width=6cm,angle=0]{14339fg6.eps} \vspace*{-3mm} \end{figure} Figure 6:

Atmospheric mixing ratios and composition of the ice boundary layer (``film'') in the detailed balancing model. ( Top): CO-N2 system. A surface pressure of 40 $\mu $bar is assumed, as estimated for 2009. The dark blue line is the CO/N2 mixing ratio expected for pure ices. The green curve shows the CO mole fraction in the ice surface film, and the black curve is the CO/N2 atmospheric mixing ratio derived from the composition of the ice film by applying Raoult's law. The range of CO/N2 atmospheric mixing ratios inferred in this work for this pressure ((2- $18) \times 10^{-4}$) is indicated by the blue-colored region. It implies a CO/N2 mixing ratio in the surface film of (1.4- $12) \times 10 ^{-3}$ (see text). The surface film is therefore largely dominated by N2, and the total pressure is defined by N2 equilibrium at 39.075 K. ( Bottom): CH4-N2 system. Calculations are here performed for a 14 $\mu $bar pressure, appropriate for the Voyager conditions. The Voyager-determined CH4/N2 atmospheric mixing ratio at the surface level ((1.1- $3) \times 10^{-4}$) is indicated by the yellow region. The colored lines have the same meaning as in the top panel, with CH4 replacing CO. The intersection of the black line with the colored area shows that explaining the observed CH4 mixing ratio and the total pressure requires a 38.3-39.6 K surface temperature and would imply a very high CH4 mole fraction (50-80%) in the surface film, well beyond the solubility limit of CH4 in N2. The same conclusion is reached if the CH4 amounts measured in 2009 are used. Note that these diagrams remain similar at other surface pressures, the only change being the required ice temperature to sustain the total pressure.

Open with DEXTER
In the text

 \begin{figure} \par\includegraphics[angle=270,width=16cm,clip]{14339fg2.ps} \end{figure} Figure 2:

Zoomed image of several CH4 lines showing sensitivity of the spectrum to Triton's methane. Some portions of the spectrum, especially over 2320-2326 nm, are relatively independent of temperature, while high-energy lines at 2353-2359 nm show increased temperature sensitivity. The top two panels show sensitivity to the methane abundance. Blue, red, and green synthetic spectra have 0.03, 0.08, and 0.20 cm-am of methane. Triton's thermal profile is taken from Krasnopolsky (1993) and a Voyager-like vertical distribution is used for methane (Herbert & Sandel 1991, entrance profile). Based on these models, the best fit methane column density is determined to be $0.08 \pm 0.03$ cm-am (i.e. $\pm $40%). The bottom two panels show sensitivity to methane temperature. The previous best-fit model using Voyager thermal profile is shown in pink. Other models assume an isothermal atmosphere with temperature of 30 K (dark blue), 50 K (red) and 80 K (green). These fits indicate a mean methane temperature of 50 +20-15 K.

Open with DEXTER
In the text

  \begin{figure} \par\includegraphics[angle=270,width=16cm]{14339fg4.ps} \end{figure} Figure 4:

The eight detected CO lines from Triton's atmosphere. X-axis units are nm and Y-axis units are arbitrary. Lines are compared with models including 0, 0.03, 0.30, and 3 cm-am of CO. The slow change in absorption depth with change of column density is caused by heavy saturation of particularly narrow Doppler-shaped lines at $T \sim 50$ K. Based on these models, the best fit CO column density is determined to be 0.30 cm-am with a factor of 3 uncertainty.

Open with DEXTER
In the text

Copyright ESO 2010

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.