5.1. Spectral parameters
In Fig. 2 we plot the fitted parameters and for all stars of our sample. We have also included the Sun, assuming typical values of , , and . With the exception of the "old sun proxy" Hyi, all stars show significantly higher maximum temperatures than the Sun. On the other hand, no obvious difference is found in the spectral parameters for stars of different age. The TTS show spectral parameters similar to those of some much older stars in the Pleiades and Hyades. All TTS show , indicating a relatively high fraction of hot plasma.
It should be noted that the values for found in our fits in many cases considerably exceed the high temperature component from 2T fits for the same spectra (e.g. see Fig. 1). We interpret this as a hint that there is in fact a continuous temperature distribution and the 2T model underestimates the maximal temperature. Strong evidence for this assumption is found in the cases of EK Dra and Cet, where ASCA spectra are also available (see Güdel et al. 1996a): for EK Dra we find , while the 2T fit to the ROSAT spectrum gives a considerably cooler high temperature component of . However, the fit to the ASCA spectrum requires a strong third temperature component at (Güdel et al. 1996a), which clearly demonstrates the presence of significantly hotter plasma than found in the 2T ROSAT fit and is fully consistent with our maximum temperature. Furthermore, we can even explain the difference between and : in our spectra fitting simulations (see Appendix), we show that the most probable value to be found in a 2T fit to a CED spectrum with is , fully consistent with our fitting result for EK Dra. Qualitatively similar results are found for Cet.
In the following analysis we will investigate the relation of the spectral parameters to basic stellar properties. Therefore we exclude all known or suspected binaries from our sample, since one cannot know from which star in an unresolved binary system the X-rays originate.
5.2. Coronal temperature and spectral type
The first aspect to investigate is the dependence of the coronal temperature on the spectral type. In several studies it has been found that K and M stars show higher coronal temperatures compared to F and G stars (e.g. Schmitt et al. 1990; Stern et al. 1994). We cannot confirm this trend and find no significant dependence of the temperatures on the spectral type in our sample. Only if we focus on the Hyades stars a correlation seems to be present. However, no such correlation exists for the Pleiades stars (see Fig. 3) or for the other stars. This can be explained by the rotational velocities of these stars, as will be shown later. In any way, the spectral type seems not to determine the coronal temperatures.
We have also investigated possible relations of the coronal temperatures to the stellar mass, radius and surface gravity, but could not find any correlation.
5.3. Amplitude of X-ray emission and stellar rotation
It is well known that the X-ray luminosity of late type stars scales with rotation (Pallavicini et al. 1981). This is generally interpreted as a link between the coronal heating and the production of magnetic flux by a solar type dynamo. In Fig. 4 we plot the X-ray luminosity against the rotational velocity and the X-ray surface flux against the rotational period. The stars in our sample show the same dependencies as found in other studies: For slowly rotating stars the amplitude of the X-ray emission increases with rotation, whereas for fast rotators a saturation effect occurs. Very similar relations have been found for main-sequence stars (e.g. Dorren et al. 1995; Stauffer et al. 1994) as well as for TTS (e.g. Bouvier 1990; Gauvin & Strom 1992). This indicates that the amplitude of the X-ray emission is mainly determined by the rotation and related to the stellar age by the rotational evolution of the young stars (cf. Bouvier 1994 and Stauffer 1994). Only the Orion TTS P1659 and P1784 show a significantly higher X-ray surface flux than to be expected from their rotational periods, what might perhaps be explained by differential rotation (see Smith 1994).
5.4. Coronal temperature and X-ray activity
Further information can be gained by looking for a relation between the coronal temperature and the amplitude of the X-ray emission. Our data show a clear correlation between and , although the scatter is rather large (upper plot in Fig. 5). Similar correlations between the coronal temperature and the X-ray luminosity have already been reported in other studies. For example, Schmitt et al. (1990) found a correlation from 1T fits to the EINSTEIN spectra of late type stars and Güdel et al. (1996a, b) found from 2T fits to the ROSAT spectra of their sample of solar type field stars also used in our study.
If we use instead of (lower plot in Fig. 5), we find a much better correlation, extending over nearly two orders of magnitude in and four orders of magnitude in . TTS, young main sequence stars and field stars follow the same relation between and . Only the Hyades member L 86 clearly deviates from the correlation. This is a G2 star and its location in the plot might be explained by the possibility that a significant part the X-rays originates from an (up to now unknown) late type companion. In that case, we would underestimate the surface flux considerably.
We have performed a linear regression analysis (excluding L 86) and found the best fit relation
shown as the dotted line in Fig. 5. As noted by Basri (1987), the X-ray surface flux is a very good measure of stellar activity. This explains, why the correlation between and is much better than between and and shows that the coronal temperatures are determined by stellar activity.
Now we are in a position to explain why the Hyades seem to exhibit a correlation between and the spectral type, while the other stars in our sample do not (see section 5.2). The Hyades stars in our sample have very similar rotational velocities, ranging only from 10 km/sec to 15 km/sec. Since the rotational velocity scales with the stellar radius, the later type stars have shorter rotational periods than the earlier type stars. Finally, the dependence of activity on the rotational period and that of the coronal temperature on activity explains the apparent correlation between temperature and spectral type. No such apparent correlation is found for the Pleiades stars in our sample, because they exhibit widely different rotational velocities.
So we conclude that the coronal properties of young stars and even the TTS are determined by stellar rotation. The high X-ray luminosities and coronal temperatures can be explained by the fast rotation of the young stars. The rotational slow down during the early phases of stellar main sequence evolution is accompanied by a continuous decrease of X-ray activity and coronal temperature (for a recent review on the rotational evolution of TTS and young stars we refer to Bouvier (1994) and Stauffer (1994)).
© European Southern Observatory (ESO) 1997
Online publication: June 30, 1998