Ultra-high-resolution observations of interstellar Na I and Ca II K toward the high galactic latitude star HD 28491

We present very high resolution (0.32 km s(-1)) spectra of interstellar Na I D1, D2, and Ca II K absorption toward HD 28497 obtained with the Ultra-High-Resolution Facility at the 3.9 m Anglo-Australian Telescope. The star is located in projection in a highly disturbed interstellar region close to a number of identified features including the high galactic latitude molecular cloud MBM 20, the large Orion-Eridanus shell, seen in H alpha and H I 21 cm maps, and a filamentary loop structure between upsilon(LSR) = -12 and -4 km s(-1) in the Berkeley H I 21 cm survey and visible on the IRAS 100 mu m map. Toward HD 28497 we detect 13 absorption components in the Na I spectra, to a column density limit of 2 x 10(10) cm(-2), and 10 in Ca II K over a velocity range of similar to 70 km s(-1). Four absorption components in the Na I spectra show s-resolved hyperfine structure with b-values from 0.31 to 0.40 km s(-1) and column densities from 4.0 to 14 x 10(10) cm(-2). If we assume the clouds represented by these components have no internal turbulent velocities, their temperatures would range between 134 and 227 K. One of these hyperfine split (hfs) components, at upsilon(LSR) = -11.1 km s(-1), shows significant temporal variation in equivalent width compared to earlier (1977) observations, making this the first interstellar sight line outside the Vela supernova remnant to show a time-varying component. The feature may be associated with the filamentary loop structure seen in this region. There is poor correspondence between the Na I and Ca II profiles: we do not detect narrow Ca II profiles to the four hfs Na I components, and only three of the well-resolved components have the same Ca II and Na I radial velocities and consistent b-values. One of these components, at upsilon(LSR) = -30.0 km s(-1), has a low Na I/Ca II ratio and arises in a region where turbulent motions dominate-properties consistent with the hypothesis that the cloud lies close to HD 28497. In general, however, the Na I and Ca II occupy different gaseous phases in the ISM. We have compared our data with 21 cm emission profiles obtained from the recent Leiden/Dwingeloo H I survey. Based on agreement in the velocities, the Na I/Ca II ratio, and the kinetic temperatures, we conclude that the component at upsilon(LSR), = -7.5 km s(-1) is associated with the front side of the large, expanding Orion-Eridanus shell. Unexpectedly, the molecular cloud MBM 20 is not detected either in our absorption spectra or in the H I profiles, indicating that HD 28497 lies away from the core of MBM 20. Apart from the two features at -11 and -7.5 km s(-1) there is almost no agreement between the H I profiles and the optical spectra. Although we cannot rule out the possibility that most of the H I lies behind the star, this explanation seems unlikely because many of the H I features have previously been attributed to foreground phenomena. The beam sizes of the H I and the optical studies are quite different and this suggests a different explanation, namely that the physical sizes of the interstellar structures we detect in Na I and Ca II are not extensive enough to be detected in H I. If so, this raises questions about the usefulness in general of combining results obtained from H I 21 cm studies with results obtained from optical (or ultraviolet) studies of the interstellar gas.