next up previous contents index
Next: HST Future Science Up: Planetary Science and Previous: Supersonic Motions in

Science with the Hubble Space Telescope -- II
Book Editors: P. Benvenuti, F. D. Macchetto, and E. J. Schreier
Electronic Editor: H. Payne

The CO Abundance on Neptune from its UV Reflection Spectrum

Régis Courtin, Daniel Gautier
Département de Recherche Spatiale, CNRS, Observatoire de Paris-Meudon, 92195 Meudon, France

Darrell Strobel
Department of Earth and Planetary Sciences, Johns Hopkins University, Baltimore, MD 21218 USA



Observations of Neptune obtained with the Faint Object Spectrograph were analyzed in search of CO absorption between 1900 and 2300Å. We report on the marginal detection of two Cameron bands--the (0,0) band at 2064Å and the (1,0) band at 1992Å--and on the determination of the CO abundance using a comprehensive radiative transfer model. Rayleigh-Raman scattering is found to have a dominant influence on the UV albedo of Neptune in that spectral range, whereas haze particles contribute only slightly to the overall opacity. We derive a CO mole fraction of in the 30--800 mbar pressure range, in agreement with previous determinations from microwave transitions. It confirms that the source of CO on Neptune is located deep in the troposphere of the planet.


The unexpected detection of carbon monoxide (and hydrogen cyanide) in Neptune's stratosphere (Marten et al. 1993) has led to significant progress in the understanding of the physico-chemistry of that planet. Two possible origins have been identified for CO: a) reactions between products of the methane photodissociation and infalling oxygen atoms that come from Neptune's satellite Triton (Moses 1992); b) uplifting of carbon monoxide produced in the deep atmosphere through thermochemical reactions between methane and water vapor (Lodders & Fegley 1994). The vertical distribution of CO resulting from these two scenarios is drastically different. Several observations made after the initial detection (see Table 1) have already shown the vertical distribution of CO to be inconsistent with an external origin (Marten et al. 1993).

Table 1: Determinations of Neptune's CO abundance

This paper presents the detection of two CO absorption bands in Neptune's UV reflection spectrum, and the determination of the CO mole fraction from these measurements using a comprehensive radiative transfer model.


The data that we analyzed were retrieved from the HST archive. They were obtained in August 1992 under the GTO observing program 1200-GTO/OS-86G (Caldwell, Principal Investigator). The observations used the FOS/BL -- G190H configuration with a 1 arcsec aperture. The spectral dispersion is 1.47Å/diode, but the spectral resolution resulting from the extended aperture is close to 4Å. The processing of the raw data is described elsewhere (Courtin et al. 1996). Figure 1 shows the calibrated FOS measurements after correction for instrument straylight. That correction represents between 10 and 20% of the measured fluxes in the spectral range of interest. Because of the difficulties encountered in accurately removing solar spectral features from planetary reflection spectra, we chose to reference the Neptune measurements to those obtained on Uranus two months before with the same instrument configuration.

Figure: Calibrated FOS spectra of Uranus and Neptune, corrected for instrument straylight.

In addition, we need to determine the geometric albedo of each planet in order to validate the radiative transfer models. We used the solar irradiance data of Anderson & Hall (1989) for that purpose. Figure 2 shows the geometric albedo of Neptune derived from the FOS data of Fig. 1 after binning in 4.4Å wide intervals. It is compared to the 50Å-bandwidth results of Wagener et al. (1986) from IUE measurements. Figure 3 represents the Neptune-to-Uranus ratio of the geometric albedos obtained with three values of the bin width (4.4, 5.9 and 7.3Å/bin). Three depressions appear systematically, two of which--around 1992 and 2064Å--coincide with the first two bands of the Cameron system of the CO molecule. The third depression around 2049Å cannot be explained at the moment.

Figure: Geometric albedo of Neptune in binning intervals of 4.4Å. The IUE determinations of Wagener et al. (1986) (filled circles) were used as a check of the FOS calibration. The solid line corresponds to the result of the radiative transfer model described in Section 3.

Figure: Neptune-to-Uranus geometric albedo ratio derived from the data of Fig. 1 with various binning intervals. The positions of the first three Cameron bands of CO are indicated.

Modeling the Reflection Spectrum

The geometric albedo of each planet was modeled with a two-stream radiative transfer code (Toon et al. 1989, McKay et al. 1989) including Rayleigh-Raman scattering from the dominant molecular constituents (84.7% H, 15% He, 0.03% N), as well as Mie scattering from haze particles. The scheme designed by Pollack et al. (1986) was used to approximate the details of Raman scattering. For the haze, the results of the Voyager investigation at Uranus (Rages et al. 1991), and Neptune (Pryor et al. 1992) were adopted with some modification of the mean particle sizes and the use of an ad hoc extrapolation of the index of refraction to fit both the IUE- and FOS-derived albedos. The details of the haze models are given in Courtin et al. (1996).

The result of the Neptune modeling is shown in Fig. 2 at the nominal resolution of the FOS data (4Å). As far as molecular absorption is concerned, other than CO, only CH needs to be taken into account. However, its cross-section because negligible longwards of 2000Å. wavelengths longer than 2000Å. Actually, all the features present in the simulated spectrum are the result of Raman scattering which introduces a shift of solar lines by about 15Å (S rotational), 26Å (S rotational), or 184Å (Q(1) vibrational), depending on the H transition which is excited by the solar photons. A similar fit was obtained in the case of Uranus.

Results and Implications

Neptune-to-Uranus ratio spectra were calculated for various values of the Neptune CO mole fraction. The upper limit of the CO abundance on Uranus from the microwave measurements is about 100 times smaller than the Neptune value. Therefore, the use of the Uranus spectrum as reference does not introduce any bias in the search for CO absorption on Neptune. Results are shown in Fig. 4

Figure: Neptune-to-Uranus albedo ratio spectra computed with f=1, 3, and 5, compared to the FOS data points affected with their formal error bar.

for f=1, 3, and 5. The ratio spectrum still contains a few features associated with Raman scattering. This is due to the fact that Neptune's stratosphere is less hazy than that of Uranus; therefore, the influence of Raman scattering on its reflection spectrum is stronger. The CO (0,0) and (1,0) Cameron bands in the simulated spectra match the observed features. Their contrast, however, is of the same magnitude as the error bars. Thus, we can only claim a marginal detection of these transitions. The (2,0) band appears even fainter in the synthesized spectra and situated on the flank of a strong Raman feature. Therefore, its detection would be quite difficult.

A goodness-of-fit parameter was defined in the 1950--2100Å interval, which led to a best fit value f. The main source of uncertainty is the random noise affecting the observations. A small contribution stems from uncertainties concerning the correction of instrument straylight, as well as the extrapolation of the CO absorption cross-section to low temperatures (50--60 K). Figure 5 presents a comparison of our determination

Figure: Comparison of the present determination of the CO mole fraction with those obtained from the microwave transitions.

of f with those given in Table 1. Most of the CO absorption in our model takes place between 30 and 800 mbar, even though Rayleigh-Raman optical depth unity is reached near 40 mbar. The present result, although less precise than those obtained from the microwave transitions, confirms that the vertical distribution of CO in Neptune's atmosphere is constant.


The observations that we analyzed represent a great improvement over the data gathered so far on the UV reflection spectrum of Neptune. Thanks to the increased spectral resolution and sensitivity afforded by the Faint Object Spectrograph, we were able to detect CO absorption features around 2000Å. Our interpretation is in good agreement with results obtained previously in the microwave range, and confirms the notion that the source of CO on Neptune is located deep in the troposphere of the planet.


This work was supported by the French ``Programme National de Planétologie'' (CNRS-INSU) and by NASA through Grant No. AR-4910.01-92A from the Space Telescope Science Institute.


Anderson, G. P. & Hall, L. A. 1989, J. Geophys. Res., 94, 6435

Courtin, R., Gautier, D., & Strobel, D. 1996, Icarus, in press

Guilloteau, S., Dutrey, A., Marten, A., & Gautier, D. 1993, A&A, 279, 661

Lodders, K. & Fegley, B. Jr. 1994, Icarus, 112, 368

Marten, A., Gautier, D., Owen, T., Sanders, D., Matthews, H. E., Atreya, S. K., Tilanus, R. J. P., & Deane, J. 1993, ApJ, 406, 285

McKay, C. P., Pollack, J. B., & Courtin, R. 1989, Icarus, 80, 23

Moses, J. I. 1992, Icarus, 99, 368

Naylor, D. A., Davis, G. R., Griffin, M. J., Clark, T. A., Gautier, D., & Marten, A. 1994, A&A, 291, L51

Pollack, J. B., Rages, K., Baines, K. H., Bergstralh, J. T., Wenkert, D., & Danielson, G. E. 1986, Icarus, 65, 442

Pryor, W. R., West, R. A., Simmons, K. E., & Delitsky, M. 1992, Icarus, 99, 302

Rages, K., Pollack, J. B., Tomasko, M. G., & Doose, L. R. 1991, Icarus, 89, 359

Rosenqvist, J., Lellouch, E., Romani, P.N., Paubert, G., & Encrenaz, T. 1992, ApJ, 392, L99

Toon, O. B, McKay, C. P., Ackerman, T. P. & Santhanam, K. 1989, J. Geophys. Res., 94, 16287

Wagener, R., Caldwell, J., & Fricke, K.-H. 1986, Icarus, 67, 281

next up previous contents index
Next: HST Future Science Up: Planetary Science and Previous: Supersonic Motions in