X - ray spectroscopy of the active dM stars : AD Leo and EV Lac

We present results of an analysis of new observations of the active M dwarfs AD Leo (GJ 388) and EV Lac (GJ 873), performed with the X-ray satellite SAX, and compare them with both published and new analyses of ROSAT PSPC observations of these stars. The PSPC spectra can be fitted with one(EV Lac) or two-component (AD Leo) isothermal mekal models, and very low metallicity (∼ 0.1 solar); we have found clear evidence of spectral variations in correspondence of a large flare observed during one of the PSPC observations of EV Lac, requiring the addition of a second component with log T ∼ 7.5 to fit the flare spectrum. The SAX light-curves of AD Leo and EV Lac also show the occurrence of several flares. A two-component model does not provide an adeguate fit of the SAX spectra, regardless of the value of coronal metallicity. These spectra require at least three thermal mekal components and best-fit coronal plasma metallicity below solar for AD Leo and only marginally below solar for EV Lac, with 90% confidence ranges on Z/Z = 0.220.35 and 0.36-1.11, respectively. We have also fitted the SAX spectra of AD Leo and EV Lac with model spectra from constant cross-section static coronal loops. One-loop models fail to fit the observed spectra. A second loop component, that accounts for most of the plasma emission at high energy, is required to obtain a fit of statistical quality just slightly worse than the 3-T fits. In the case of EV Lac both the 2-loop and the 3-T fits are unable to reproduce the observed emission below 0.5 keV. The available evidence points toward the existence of various (at least two) main classes of coronal emitting structures: the dominant one is composed of hundreds of compact loops, with relatively low maximum temperature and length smaller than 0.1 the stellar radius, covering no more than 1% of stellar surface; the second class, responsible for the high energy emission, is composed at least of tens of quite elongated loops, covering a very small fraction of stellar surface. We find no-evidence of loops with length comparable to the stellar radius.