Sp. Maran et al., OBSERVING STELLAR CORONAE WITH THE GODDARD HIGH-RESOLUTION SPECTROGRAPH .1. THE DME STAR AU MICROSCOPII, The Astrophysical journal, 421(2), 1994, pp. 800-808
We report on an observation of AU Mic taken with the Goddard High Reso
lution Spectrograph aboard the Hubble Space Telescope. The data consis
t of a rapid sequence of spectra covering the wavelength range 1345-13
75 Angstrom with a spectral resolution of 10,000. The observations wer
e originally intended to search for spectral variations during flares.
No flares were detected during the 3.5 hr of monitoring. A method of
reducing the noise while combining the individual spectra in the time
series is described which resulted in the elimination of half of the n
oise while rejecting only a small fraction of the stellar signal. The
resultant spectrum was of sufficient quality to allow the detection of
emission lines with an integrated flux of 10(-15) ergs cm(-2) s(-1) g
reater. Lines of C I, O I, O V, Cl I, and Fe XXI were detected. This i
s the first indisputable detection of the 1354 Angstrom Fe XXI line, f
ormed at T similar to 10(7) K, on a star other than the Sun. The line
was well resolved and displayed no significant bulk motions or profile
asymmetry. From the upper limit on the observed line width, we derive
an upper limit of 38 km s(-1) for the turbulent velocity in the 10(7)
K plasma. An upper limit is derived for the flux of the 1349 Angstrom
Fe XII line, formed at T similar to 1.3 x 10(6) K. These data are com
bined with contemporaneous GHRS and IUE data to derive the volume emis
sion measure distribution of AU Mic over the temperature range 10(4)-1
0(7) K. Models of coronal loops in hydrostatic equilibrium are consist
ent with the observed volume emission measures of the coronal lines. T
he fraction of the stellar surface covered by the footpoints of the lo
ops depends upon the loop length and is less than 14% for lengths smal
ler than the stellar radius. From the upper limit to the estimated wid
th of the Fe XXI line profile we find that we cannot rule out Alfven w
ave dissipation as a possible contributor to the required quiescent lo
op heating rate.