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

---

GHRS Spectrally Resolved Profiles of the Lyman Emission Collisionally Excited in the Jovian Aurorae

R. Prangé, L. Pallier, and C. Emerich
Institut d'Astrophysique Spatiale, Orsay, France

L. Ben Jaffel
Institut d'Astrophysique de Paris, Paris, France

D. Rego, J.T. Clarke, G.E. Ballester
University of Michigan, Ann Arbor, MI 48104 USA

Institut d'Astrophysique de Paris, Paris, France on leave from Institut d'Astrophysique Spatiale, Orsay, France

Abstract:

We present here the first Lyman profiles taken at high spectral resolution (70mÅ) in the Jovian aurorae. As predicted by models developed beforehand, but for the first time on a planet, we detect a strong central reversal of the line reminiscent of solar spectra. This is attributed to radiative transfer effects on photons excited deep in the atmosphere by energetic charged particles from the magnetosphere. We show that the observed profiles are in very good agreement with the synthetic ones, and that the agreement is sensitive enough to the characteristics of the particles and of the atmosphere to be used as a diagnostic probe. Finally, we find that the reversal is not at the center of the line, suggesting the possibility of relative motions of a few between the source and the overlying atmosphere.

Keywords: Jupiter, aurorae, magnetosphere, UV

Introduction

The powerful magnetic field of Jupiter sustains a giant magnetosphere where plasma processes are still moderately understood. In-situ measurements, limited to five rapid fly-bys of deep space missions (Pioneers, Voyagers and Ulysses), have not provided a global view so far. A complementary approach to these processes consists of remote-sensing of the Jovian aurorae, excited by precipitation of energetic magnetospheric charged particles along high latitude field lines. Moreover, the power input in the auroral zone being so large, several 10 watts (Livengood et al. 1992), it presumably controls the dynamics and energetics of the atmosphere on a planetary scale (Sommeria et al. 1995, Emerich et al. 1996). Imaging of the aurorae may allow to identify the magnetically conjugate active regions in the magnetosphere (Gérard et al. 1994, Prangé et al. 1996), whereas spectroscopy gives access to the signature of the precipitating particles. Early studies have focussed on the far-ultraviolet H Werner and Lyman bands as the hydrocarbons lying above the auroral source layer exhibit a dramatic absorption cross-section increase shortward of 1400Å (Yung et al. 1981, Livengood et al. 1990), allowing direct determination of the overlying hydrocarbon optical depth, and given an atmospheric model, of the particle penetration depth, and eventually of their energy. The Lyman emission also bears the signature of the incident particles (energy and species), but it has been much less studied so far (Harris 1993) because it was much more difficult to model (i) the combined effect of atmospheric excitation rate and of multiple scattering on escaping photons, (ii) the `hot' wing excited by charge-exchanged fast atoms when protons are precipitating; and because limited spectral resolution did not allow to resolve the line profile. Such models, including the effect of radiative transfer have now been developed for incident electrons and protons (Rego 1994, Rego et al. 1994, 1996a). In the range of energies expected from magnetospheric particles (above 100 eV--1 keV), the model predicts a significant core reversal of the line, together with a line broadening which depend on the overlying H column density (i.e., penetration depth, energy). An energy-dependent red-shifted component may also be observed when protons are the primary particles.

Observations

Spectra of Jupiter were taken on June 25, 1994, with the Goddard High Resolution Spectrograph (GHRS) on HST, with the Echelle A grating and the 2 arcsec aperture. Four targets were selected in the northern polar region of Jupiter, on auroral structures for which previous high spatial resolution FOC images had suggested different types of magnetospheric particle precipitation. Pointing was obtained by an offset from Io, providing an accuracy of 0.1 arcsec. The wavelength range, 1214--1220Å covers the Lyman line and a few H Werner lines at a resolution of 70mÅ. In order to remove the geocoronal contribution efficiently, we have dedicated a full orbit to geocoronal spectra taken at the same solar zenith angles and with the same exposure time (nine minutes) as the Jovian spectra. Standard STScI procedures have been used to reduce the data. Figure 1 displays the co-added spectra obtained at the four locations after subtraction of the geocoronal emission (significantly Doppler-shifted by the relative Earth's motion, and located at the foot of the blue wing).

 
Figure: High resolution spectra in the Jovian aurora.

It is immediately clear that the Lyman profiles exhibit a deep core reversal, a feature common in solar and stellar emission lines, but never observed so far on a planetary spectrum. This reversal was predicted by Rego et al.'s model as the result of radiative transfer effects in an optically thick atmosphere. By contrast, it is not observed at lower latitude, in spectra of the solar reflected line (Emerich et al. 1996), also in agreement with model predictions (see the solar scattered contribution in Figure 2). Similar core reversals, although not as deep, have recently been observed also in a series of Ech-A spectra taken at the equatorward edge of the southern auroral region (Emerich et al. 1996).

The second feature, also characteristic of solar line profiles, is that the line shape is not symmetric, with two peaks of different intensity, as if the reversal were not at the center of the line. This effect may affect either the blue or the red wing, depending on the spectrum, with significant variations in the ratio of the peaks. It must be the signature of changes in the nature of the original source and/or in the state of the atmosphere. This suggests that the `rest frame' of the line may differ between the source region and the overlying atmosphere where multiple scattering occurs, with relative motions up to 4 along the line of sight to account for the strongest asymmetry observed.

Comparison with Model Spectra

The model Lyman profile includes the solar reflected and auroral contributions. The solar reflected contribution is computed with a code derived from Ben Jaffel et al. (1988), and using the doubling-adding method. This contribution depends only on the atmospheric model used and on the zenith angle of the line of sight (i.e., on the distance of the target from the center of the disc). The auroral contribution is self-consistently computed (intensity and line profile) from the altitude distribution of the excitation rate by precipitating particles and the subsequent effect of the overlying atmosphere on the escaping photons (multiple scattering by H and hydrocarbon absorption). The radiative transfer calculations at Lyman are performed for both angle averaged partial frequency redistribution and complete frequency redistribution (AAPR and CR). For electron energies above 100 eV (protons above 2 keV), the atmosphere is thick enough that cumulative effects of the redistribution of photons, and of their path opacity before escape produces a central reversal associated with a line broadening. This effect is increasing with the energy of the incident particles. The reversal is also broader and the wings more developed when the frequency redistribution is complete.

This model has been recently used to interpret Lyman profiles previously obtained in Cycle 3 with GHRS/G160M at lower resolution, 570mÅ (Clarke et al. 1994), using a set of photochemical atmospheric models possibly representative of the auroral atmosphere. The resolution did not allow to detect the central reversal of the line, but permitted a search for a best fit to the overall profile (especially the FWHM and the wings of the line). This was found using an H column density [H] = 10 cm, and an eddy diffusion coefficient K= 10 cm (Rego et al. 1996b). We make use of the same atmospheric model for our first attempt of comparison between a model spectrum and the observations. Figure 2 shows the spectra resulting from the model spectrum (solar contribution, plus auroral contribution scaled to the observed intensity, both of them at the location of the first spectrum in Figure 1) for three typical incident particle energy with AAPR, once folded with the Ech-A Line Spread Function (LSF), appropriate to extended sources.

 
Figure: Comparison of synthetic auroral spectra with the first spectrum in Figure 1. see text

The first conclusion is that the agreement between the synthetic spectra and the observed one is globally remarkable. Most of the differences in fact result from approximations in this very preliminary study. For instance, the two peaks are slightly too large because we have had to use the pre-COSTAR LSF (a new one is not available so far), and we cannot account for the difference in peak intensities on both sides of the reversal because we have not yet included any asymmetry effect in our excitation/radiative transfer coupling code. We have also plotted on the last panel, the synthetic profile for 1 keV electrons and CR, for which the fit is not as good (excessive width and wings). We see that the critical parameters, such that the separation between the peaks (reversal width), the depth of the reversal, and the shape of the line wings vary sufficiently from one model to the other that they can be used as a sensitive diagnostic of the incident particles in a future more thorough study. With the atmospheric model used here, we find that the separation is best reproduced with 1 keV incident electrons, AAPR, whereas the depth of the reversal is best fitted with 10 to 100 keV. This is probably an indication that the atmospheric model used is not appropriate to the real atmosphere. Therefore, we can anticipate that the spectra will also help constraining the auroral atmospheric models, very poorly constrained so far.

Acknowledgments:

We are grateful to the STScI team, and especially Alex Storrs and Steve Hulbert for their assistance in the execution of the observations and reduction of the data.

References:

Ben Jaffel, L., Magnan, C., & Vidal-Madjar, A. 1988, Astron. Astrophys., 204, 319

Clarke, J.T., Ben Jaffel, L., Vidal-Madjar, A., Gladstone, G.R., Waite, H., Jr., Prangé, R., Gérard, J.C., Ajello, J, & James, G. 1994, ApJ, 430, L73

Emerich, C., Ben Jaffel, L., Prangé, R., Clarke, J.T., Ballester, G.E., & Sommeria, J. 1996, this issue

Gérard, J.C., Dols, V., Prangé, R., & Paresce, F. 1994, Planet. Space Sci., 42, 905

Harris, W., M. 1993, PhD Dissertation, University of Michigan, Ann Arbor, MI, USA

Livengood, T.A., Strobel, D.F., & Moos, H.M. 1990, Geophys. Res. Lett., 95, 10375

Livengood, T.A., Moos, H.M., Ballester, G.E., & Prangé, R. 1992, Icarus, 97, 26

Prangé, R., Rego, D., Southwood, D., Zarka, P., Miller, S., & Ip, W.H. 1996, this issue

Rego, D. 1994, PhD. Dissertation, Université Paris-Sud, Orsay, France

Rego, D., Prangé, R., & Gérard, J.C. 1994, J. Geophys. Res., 99, 17075

Rego, D., Prangé, R., & Ben Jaffel, L. 1996a, J. Geophys. Res., submitted

Rego, D., Clarke, J.T., Ben Jaffel, L., Ballester, G.E., Prangé, R., & McConnell 1996b, J. Geophys. Res., submitted

Sommeria, J., Ben Jaffel, L., & Prangé 1995, Icarus, 119, 2

Yung, Y.L., Gladstone, G.R., Chang, K.M., & Ajello, J.M., 1981, ApJ, 254, L65