CORONAL GAS IN THE HALO .2. ORFEUS OBSERVATIONS OF GALACTIC HALO STARS

We present new intermediate-resolution spectra (lambda/Delta lambda = 3300) of fourteen early-type halo stars near the O VI resonance double t at 1031.93 and 1037.62 Angstrom These spectra were obtained with the Berkeley extreme and far-ultraviolet (EUV/FUV) spectrometer in the OR FEUS telescope aboard the space platform ASTRO-SPAS during the mission of space shuttle Discovery down in 1993 September. Most of the target s are at high \b\, at a mean distance of about 2.8 kpc, providing an o pportunity to study conditions in a region of the Galaxy comparatively unexplored by Copernicus (see work of Jenkins). The ORFEUS halo stars confirm the surprising paucity of O VI in the halo reported by Hurwit z et al. for two extragalactic sight lines through the southern Galact ic hemisphere. We find an O VI scale height significantly lower than t hose reported for other highly ionized species; h(0) is between about 80 pc and 600 pc if the midplane density n(0) is between 1.5 and 5 x 1 0(-8) cm(-3). Only upper limits on N V exist for our sight lines; the O VI/N V ratio is constrained to lie above similar to 3. The O VI/C IV ratio is about 1-3. There is an inconsistency between the mean abunda nce of coronal gas in the halo derived from our O VI data and the much greater abundance derived from N V (reported by Sembach & Savage) and the O VI/N V ratio of similar to 15 observed in the disk and predicte d by various models. We identify a foreground contribution from the ra dio loop I/soft X-ray enhancement along many of the high-b sight lines studied in N V as the most likely cause of the discrepancy. Unless th e O VI/N V ratio in the halo falls well below its expected value, the projected column of coronal gas varies by a factor of similar to 100 a mong high-b sight lines. Insofar as it is sensible to discuss ''typica l'' values under these circumstances, we arrive at projected columns o f 8 x 10(13) cm(-2) for O VI and 5 x 10(12) cm(-2) for N V. For compar ison with models or other observations that are inherently exclusive o f low \z\ coronal gas, smaller columns should be adopted. The Galactic fountain model of Benjamin & Shapiro provides a self-consistent frame work for interpretation of our observational results and various studi es of C IV. The data are consistent with a flow velocity of similar to 15 km s(-1), isochoric cooling, a mass flux of similar to 3 M(.) yr(- 1) per side, a density of similar to 10(-2) cm(-3), and a cooling leng th of a few hundred pc. There is weak evidence for an excess of O VI a long sight lines characterized by foreground absorption from intermedi ate negative velocity (INV) gas detected primarily in low-ionization s pecies. The tentative O VI excess is about 4-8 x 10(13) cm(-2) and may be produced in interfaces between the cool INV gas and a hot, low-den sity medium. We constrain the product of the midplane pressure and fil ling factor in the hot phase of the interstellar medium; at log T = 6. 0, P/k x f(h) is probably <4000-12,000 cm(-3) K assuming isothermal ga s, collisional ionization equilibrium, and solar abundances.