Jan-Uwe Ness

First detection of X-ray emission from Saturn

First indication for X-rays from Jupiter was found 1983 with the Einstein-Telescope. Unusually high X-ray emission was detected from Jupiter and it was difficult to associate a plausible production mechanism. For the Earth Bremsstrahlung from charged solar wind particles cause the emission of X-rays, but this mechanism cannot explain the high X-ray emission found for Jupiter. The most plausible scenario was thought to be connected with the volcanically active moon Io. Heavy ions are produced and precipitate on the atmosphere of Jupiter and cause X-ray emission that should show signatures from Sulfur and Oxygen, atoms in Jupiters atmosphere that are excited into higher energetic states by these ions.

In my project I analysed planetary data obtained with the ROSAT-Telescope. I analysed observations of Jupiter, Saturn, Uranus and Neptun, but found only X-rays from Jupiter and Saturn (Ness&Schmitt 2000 A&A, 355, 394). The detection of X-rays from Saturn was only very marginal and I describe the analysis procedure briefly.
For the analysis the proper motion of the planet had to be corrected, because the Satellite does keep the planet in it's focus. This was done by assigning each individual photon (for which positions on the detector and arrival times are recorded in a so-called event list) a new position in a transformed image depending on the arrival time. This transformation was chosen in a way that all planetary photons will be located at a designated position while photons from eventual fixed sources in the images will be smeared out according to an inverse path of the planet. The planetary photons are extracted in an extraction box collecting as many source photons while avoiding to collect too many background photons. An additional box is placed adjacent to the source in a source free region and the number of photons in this box is assumed to be the background level that needs to be subtracted.
Fig. 1


Fig. 1: Jupiter in the soft X-ray band [0.2-0.64 keV]
The image is transformed, such that the proper motion of the planet is accounted for. The photons next to Jupiter are due to pointing problems of the telescope during the exposure.

Fig. 2


Fig. 2: Exposure of Saturn in the same energyrange and after analogous processing.
The green box indicates the expected position of the object. 203 counts have been counted within this area for Jupiter and 14 counts for Saturn (background subtracted).


For the cases of Jupiter and Saturn I show in Fig 1 and Fig 2 the transformed ROSAT-PSPC images. While Jupiter is clearly detected the detection of Saturn is very questionable. The extraction box placed on the expected position of Saturn contains 22 photons, while from the background only 7.6 photons are expected. In order to rule out that this is and event of pure chance I place altogether 2000 boxes of the same size adjacent to other and covering the complete field. This is shown in Fig. 3; the center of the image is the expected position of Saturn. In Fig. 4 I show the result from this counting exercise. The horizontal axis represents the numbers of photons counted in the different boxes while the vertical axis is the number of boxes where the individual photon number were counted. It can be seen that the box containing Saturn is (by chance?) the box with the highest encountered number of photons, while the count number in the complete field follows a Poisson distribution. The solid line is a best fit Poisson curve with an expected mean of 7.54 counts. Under these circumstances the probability to find 22 photons in box placed at a designated position by pure chance is 1.7 10-5.
Fig. 3


Fig. 3:Complete image with Saturn expected in the center and N=2000 boxes covering the complete field for counting photons.

Fig. 4


Fig. 4:Number of boxes encountered with given photon counts


Clearly a confirmation of this detection is desirable and I proposed an observation to be carried out by the Chandra telescope. This observation was granted and Saturn was observed for 70 ksec in April 2003 and the analysis is presented in Ness, J.-U., Schmitt, J.H.M.M., Wolk, S.J., Dennerl, K., and Burwitz, V. A&A, 418, 337 (2004).



last modified: Friday, 14-Nov-2008 21:43:21 CET