The density of coronal plasma in active stellar coronae

We have analyzed high-resolution X-ray spectra of a sample of 22 active stars observed with the High Energy Transmission Grating Spectrometer on Chandra in order to investigate their coronal plasma density. Densities were investigated using the lines of the He-like ions O VII, Mg XI, and Si XIII. Si XIII lines in all stars of the sample are compatible with the low-density limit (i.e., ne1013 cm-3. Mg XI lines betray the presence of high plasma densities up to a few times 1012 cm-3 for most of the sources with higher X-ray luminosity (>~1030 ergs s-1) stars with higher LX and LX/Lbol tend to have higher densities at high temperatures. Ratios of O VII lines yield much lower densities of a few times 1010 cm-3, indicating that the ``hot'' and ``cool'' plasma resides in physically different structures. In the cases of EV Lac, HD 223460, Canopus, \u3bc Vel, TY Pyx, and IM Peg, our results represent the first spectroscopic estimates of coronal density. No trends in density-sensitive line ratios with stellar parameters effective temperature and surface gravity were found, indicating that plasma densities are remarkably similar for stars with pressure scale heights differing by up to 3 orders of magnitude. Our findings imply remarkably compact coronal structures, especially for the hotter (~7 MK) plasma emitting the Mg XI lines characterized by the coronal surface filling factor, fMgXI, ranging from 10-4 to 10-1, while we find fOVII values from a few times 10-3 up to ~1 for the cooler (~2 MK) plasma emitting the O VII lines. We find that fOVII approaches unity at the same stellar surface X-ray flux level as characterizes solar active regions, suggesting that these stars become completely covered by active regions. At the same surface flux level, fMgXI is seen to increase more sharply with increasing surface flux. These results appear to support earlier suggestions that hot 107 K plasma in active coronae arises from flaring activity and that this flaring activity increases markedly once the stellar surface becomes covered with active regions. Comparison of our measured line fluxes with theoretical models suggests that significant residual model inaccuracies might be present and, in particular, that cascade contributions to forbidden and intercombination lines resulting from dielectronic recombination might be to blame