High-resolution X-ray plasma diagnostics of stellar coronae

Dissertation zur Erlangung des Doktorgrades des Fachbereichs Physik der Universität Hamburg

vorgelegt von Jan-Uwe Ness aus Oldenburg

Hamburg, 2002


With the discovery of sunspots in the 17-th century a new field of work in Astronomy was opened. The physical origin of these dark spots on the solar surface has been found to be tightly connected with the presence of strong magnetic fields. Another spectacular phenomenon is the solar corona, the outermost region of the Sun. The high temperature found in the corona in connection with a very low density has posed a lot of questions. Again, the magnetic activity of the Sun seems to play a fundamental role in controlling the processes observed in the solar corona. An introduction for a long tradition of observing the solar corona especially in the X-ray wavelength band is given in this work. Starting with earth-bound observations of the visible corona, and proceeding with spectral diagnostics, impressive pictures with high spatial resolution are presented. A detailed structure with the distribution of the coronal plasma, obviously confined to loop-like structures following closed magnetic field lines, is seen on these pictures.
As a far more difficult task, the observation of stellar coronae, which are not spatially resolvable, is introduced. From X-ray observations carried out with the satellites Einstein and ROSAT it is known that the formation of coronae must be a common phenomenon in all late-type stars that occurs at the transition from a cool star's photosphere to space. A classification of X-ray properties similar to a Hertzsprung-Russell diagram connecting X-ray emission with fundamental stellar parameters is desirable, but only the rotation of cool stars has so far been found to be connected with X-ray emission.
New approaches can be made using the recent X-ray missions Chandra and XMM, which allow high spectral resolution. The emission lines observable with these missions are used for deriving physical properties of the coronae of Capella, Algol, and Procyon. Four more cool stars are analyzed, but in less detail. A special software has been developed in the frame of this work, the algorithm being especially dedicated to Chandra LETGS spectra, but also useful for all kinds of spectral data, e.g., XMM-RGS.
Preliminary analysis of the spectra yields a distribution of plasma layers with different properties, e.g., temperatures. At least a high temperature and a low temperature component is found for the active stars Algol and Capella, but not for the inactive star Procyon. Optical depth effects seem to be negligible, such that the basic assumption of an optically thin plasma is justified at least within the scope of the current data. The continuum of Algol's coronal X-ray spectrum can be fitted by a bremsstrahlung continuum, and an upper temperature as well as an emission measure are derived from the analysis of the continuum.
The main focus of this work is the application of the method of He-like triplets for deriving plasma densities. The coronal plasmas for Capella, Procyon, Algol, Epsilon Eridani, Alpha Centauri A and B, and UX Ari are probed. I generally find low densities for most of these seven stars and higher densities only for Epsilon Eridani and Algol. For Algol this result depends, however, on the influence of the UV radiation field emitted by the B star companion. No convincing trend could be found allowing the general conclusion of higher densities in more active stars. It is noteworthy, however, that active stars with lower densities as e.g., Capella are all RS CVn systems, i.e., binaries comprising two cool stars, possibly each of them sustaining a hot corona. An interaction between these two coronae is well conceivable. Quite convincing though is the trend of the presence of hot plasma in the more active stars, while inactive stars only have a ``cool'' component around 2MK.
At low densities constraints on parameters connected with structural information as, e.g., the loop scale length or filling factors can be made. That information enables us to compare stellar coronae with the solar corona. The densities measured for the solar corona are comparable with densities measured in inactive stars. Loop scale lengths can therefore be assumed to be very similar to the Sun, such that the structural properties of these coronae are likely the same as observed for the Sun. A different picture must be drawn for the more active stars, which have to accommodate the higher emission measure either in structures with similar sizes but higher densities, or they are characterized by a plasma distribution with higher filling factors. An isolated active corona as, e.g., in the Algol system seems to prefer the first option, while RS CVn systems can well harbor a binary corona with a large volume and low densities. Therefore these two configurations must be treated differently.

Jan-Uwe Ness