2003ApJ...583..963A

We describe an extensive search with the Space Telescope Imaging Spectrograph (STIS) for ultraviolet coronal (T>10⁶ K) forbidden lines in a sample of 29 F-M dwarfs, giants, and supergiants. Measuring coronal lines in the 1150-1700 Å band with STIS has important advantages of superior velocity resolution and an absolute wavelength calibration compared with using the Chandra or XMM-Newton grating spectrometers to observe permitted transitions of the same ion stages in the kilovolt X-ray region. Fe XII λλ1242, 1349 (T∼2x10⁶ K) and Fe XXI λ1354 (10⁷ K) are well known from solar studies and have been reported in previous stellar work. A search for other coronal forbidden lines in the 1200-1600 Å region was largely negative. The few candidate identifications (e.g., Ar XIII λ1330 and Ca XV λ1375) are too faint to be diagnostically useful. We add new dwarfs to the list of Fe XII detections, including the nearby solar twin α Cen A (G2 V). Clear detections of Fe XXI were obtained in dMe stars, active giants, a short-period RS CVn binary, and possibly in active solar-type dwarfs. We developed a semiempirical method for removing the C I blend that partially affects the Fe XXI λ1354 profile. As discussed recently by Johnson et al., Capella (α Aur; G8 III+G1 III) displays clear Fe XXI variability between Goddard High-Resolution Spectograph (GHRS) and STIS observations 4 yr apart, which is apparently due to a substantial decline in the contribution from the G8 primary. We present an alternative model of the GHRS and STIS era profiles using information in the two sets of line shapes jointly, as well as knowledge of the behavior of Fe XXI profiles of other late-G ``clump'' giants similar to Capella G8. The full survey sample also provides a context for the apparent variability: the Fe XXI flux of the G8 star in the GHRS spectrum is nearly identical (in L<sub>FeXXI</sub>/L<sub>bol</sub>) to other clump giants of similar L<sub>X</sub>/L<sub>bol</sub>, but it had dropped at least a factor of 6 in the STIS measurement. The He II λ1640 Bα feature–which is thought to be responsive to coronal irradiation–also showed significant changes between the GHRS and STIS epochs, but the decrease in the G8 star was much smaller than Fe XXI. The Fe XII flux displays a correlation with the ROSAT 0.2-2 keV X-ray flux that can be described by an α=0.5 power law. Fe XXI exhibits a steeper, perhaps linear (α=1), correlation with the ROSAT flux down to an activity level of L<sub>X</sub>/L<sub>bol</sub>∼10<sup>–5</sup>, below which detections of the coronal forbidden line are rare. There is no evidence of large, systematic Doppler shifts in either Fe XII λ1242 or Fe XXI λ1354. This suggests that the emissions arise dominantly in confined structures, analogous to magnetic loops on the Sun, rather than, say, in a hot wind. The Fe XII and Fe XXI line widths generally are close to thermal (FWHM∼40-90 km.s<sup>–1</sup> at T∼10<sup>6.2</sup>-10<sup>7.0</sup> K), except for the Hertzsprung-gap giants 31 Comae (G0 III) and Capella G1 and the K1 subgiant primary of HR 1099, all of which show evidence for excess broadening in Fe XXI (Fe XII is obscured in these objects by broad N V λ1242 features). If the excess broadening is rotational, it implies that the hot coronae of ``X-ray-deficient'' 31 Com and Capella G1 are highly extended, contrary to the compact structures suggested by recent density estimates in a number of active coronal sources.