One of the goals of studies of the interstellar medium (ISM) is the determination of the physical properties of the individual clouds which constitute the neutral component of the ISM --- their temperatures, densities, relative abundances, electron fractions, internal turbulent motions, and structures --- and how those properties differ for clouds in different environments. Accurate estimates of these quantities, for a significant set of individual interstellar clouds, are required to test any fundamental, global model of the ISM. Determination of individual cloud properties requires both high spectral resolution and access to the satellite UV, where the resonance (ground state) lines of many important species are found.
Very high resolution optical spectra of Na I, Ca II, K I, Ti II, Ca I, CN, CH, and CH+ [obtained at McDonald (0.5 km/s), Kitt Peak (1.2 km/s), and the AAT (0.3 km/s), toward 20-60 bright stars with 0.01 < E(B-V) < 0.35] reveal much (though not all!) of the component structure present along the various lines of sight (see figure below), and provide constraints on individual cloud structure, abundances, temperature, and internal turbulent motions. The spectra, the statistics of individual cloud properties derived from fits to the observed line profiles, and comparisons among these species suggest that:
| High-resolution spectra of interstellar Na I and Ca II absorption toward epsilon Ori. The resolution, S/N ratio, and source of each spectrum is indicated. Tick marks above the spectra indicate components identified in fitting the profiles; 25 components (at least) are required for Ca II. Component-to-component differences in relative abundance and line width are evident. Note also the resolved hyperfine structure in several Na I components (e.g., at 3 km/s). You can view a similar plot of Na I, K I, Ca I, Ca II, and Ti II absorption, also toward epsilon Ori. |
| Distribution of line width b (= FWHM/1.665) for individual components in the Na I (upper) and Ca II (lower) surveys. Comparisons of the line widths of different species can provide information on the temperature, internal turbulent motions, and relative spatial distributions of those species. Even though the lines of sight sampled in the two surveys are very similar, the median b-values differ significantly: 0.73 km/s for Na I and 1.31 km/s for Ca II. For a subset of the 53 best-determined "corresponding" components seen in both Na I and Ca II, the median b-values are 0.65 km/s for Na I and 0.84 km/s for Ca II -- contrary to expectations based on the respective atomic weights. The Na I and Ca II thus cannot be identically distributed; Ca II probably occupies a somewhat larger volume, characterized by a larger temperature and/or larger turbulent velocities. |
|
Spectra available: If you are interested in high-resolution optical absorption-line data for a specific line of sight, check these catalogues of existing high-resolution and medium-resolution spectra. If I have not already observed your favorite line of sight, I may be able to in a future observing run (or I may know of someone else who has spectra).
Return to top of this page
The statistics of component line widths and adjacent component separations derived from the high-resolution (FWHM < 1.5 km/s) optical surveys imply that, for many lines of sight, much of the component structure will not be discernible, and many individual components will not be resolved, in UV spectra obtained with the Hubble Space Telescope GHRS or STIS (FWHM = 3.0-3.5 km/s). Furthermore, the optical spectra show that components with very different properties (e.g., overall column densities, line widths, Na I/Ca II ratios) are often separated by only several km/s along a given line of sight. Fortunately, however, the component structure deduced from the (few) high-resolution optical spectra can be used to model the lower resolution UV spectra, so that abundances for many other species can be derived for individual IS clouds from the transitions observed in the UV -- instead of the line of sight averages heretofore available. With these individual cloud abundances, one can then address questions concerning the detailed behavior of depletion in different environments, and can apply various diagnostics to understand the structure and physical properties (e.g., local density, electron fraction, ambient radiation field) of the individual clouds.
| The detailed component structure (noted by the tick marks) seen in the high-resolution K I and Ca II spectra of 23 Ori cannot be discerned in the lower resolution UV spectra of Mg I and Fe II obtained with the GHRS. The optical data can be used to model the UV lines, however, to obtain detailed abundances and physical conditions for individual clouds. For example, the two strongest components (near +22 km/s) appear to be due to relatively cold, predominantly neutral gas (T ~ 100 K), with n_e about 0.15 cm^-3, n_H about 10-15 cm^-3, and total thickness about 12-16 pc. The implied fractional ionization (0.01) is a factor 30 higher than expected from photoionization of heavy elements -- suggesting some slight ionization of H. The depletions of many typically depleted elements appear to be less severe than those found for the main component(s) toward zeta Oph, by factors of 2-3. Is there a connection between lower local densities and less severe depletions? |
I have also obtained both moderately high resolution (2.5-3.0 km/s) optical spectra and archival IUE spectra for a number of fainter, more heavily reddened stars (0.25 < E(B-V) < 1.15) whose far-UV extinction is either very steep or very shallow. Correlations noted between certain neutral species ratios, molecular abundances, and the far-UV extinction may shed light on the scattering properties of the grains in the far-UV, on the radiation environment, and on the structure of the clouds. The high resolution optical spectra are crucial for the proper interpretation of the lower resolution UV data -- the anomalous and/or preferential depletions reported for some of these lines of sight are due in part to inadequate understanding of the detailed component structure.
These detailed studies of the Galactic ISM provide an important background and context for studies both of the interstellar media in other galaxies and of the QSO absorption-line systems (QSOALS).
Detailed studies of other sightlines are in progress; see also papers by Fitzpatrick & Spitzer.
Return to top of this page
Translucent interstellar clouds, defined as having total visual extinctions A_V between 1 and 5 magnitudes, represent a transitional regime between the thinner diffuse clouds (A_V < 1 mag; traditionally studied via optical and UV absorption lines of various atomic and a few molecular species) and the thicker dense, molecular clouds (A_V > 5 mag; typically studied via mm-wave emission lines of various molecular species). With HST (GHRS and STIS) and FUSE, these translucent clouds now can be explored via the rich variety of atomic and molecular transitions in the UV --- including those of molecular hydrogen. We (Welty, Snow, Morton) have obtained high-resolution optical spectra of Na I, K I, Ca I, Ca II, CH, CH+, and CN for most of the 36 stars in the FUSE survey of translucent clouds. The detailed component structures derived from these spectra are being used to model the absorption from other atomic and molecular species (e.g., H_2, HD) in the lower resolution FUSE spectra. Our recent mini-survey of CO and C2 indicates that:
Return to top of this page
Interstellar absorption lines were first recognized as interstellar by virtue of their invariance amidst the shifting stellar lines in the delta Ori A spectroscopic binary system (Hartmann 1904). Recently, however, it has become evident that (1) some clearly interstellar lines do vary in strength and/or velocity, and (2) small-scale spatial structure is commonplace. While variability is to be expected in disturbed regions such as supernova remnants (e.g., toward HD 72127, behind the Vela SNR), it is also seen in more quiescent regions. For example, within the last 15 years, strong, narrow Na I absorption has appeared at v ~ -38 km/s toward the halo star HD 219188; that absorption continued to strengthen until mid-2000, then declined by a factor of 2 by the end of 2006. The line of sight appears to be moving into/through a relatively cold, quiescent intermediate velocity (IV) cloud, due to the 13 mas/yr proper motion of HD 219188; the variations in Na I probe length scales of 2--38 AU/yr. Comparisons among UV spectra from the HST GHRS (1994-1995) and the HST STIS (2001, 2003, 2004) suggest that (1) the total N(H) (~ 6 X 10^{17} cm^{-2}) and the (mild) depletions did not change significantly between 1994 and 2004, and (2) the local hydrogen density n_H and electron density n_e increased by factors of 2-4 between 1995 and 2001, then declined slightly through 2004. The relatively high fractional ionization, n_e/n_H ~ 0.1, implies that hydrogen must be partially ionized. The high ratio N(Na I)/N(H_tot) suggests that the variations in Na I do not imply large local pressures or densities in this case. The variations in Na I, C I, Ca II, and C II* appear to be due to differences in density and ionization [and not N(H)] over scales of tens of AU.
See also papers by Meyer, Lauroesch, Crawford, and collaborators.
Return to top of this page
In the ISM, elements heavier than H and He are generally found to be less abundant, relative to H, than they are in the Solar photosphere. These deficits are generally ascribed to "depletion" of those elements into interstellar dust grains, under the assumption of Solar total (gas+dust) abundances (though recent work suggests that this assumption may not be strictly true). The abundances of a number of elements detected for the first time in HST GHRS spectra have clarified the depletion pattern seen in moderately dense interstellar clouds. In addition, the heavy elements Ga, Ge, As, Se, Kr, Cd, Sn, Tl, and Pb (with atomic numbers ranging from 31 to 82) provide probes of nucleosynthetic processes other than those responsible for the lighter elements. Comparisons of the depletion properties of Si, Ge, Sn, and Pb, all in the same column of the periodic table (so that they should show similar chemical behavior), may provide information regarding various processes thought to occur on the surfaces of dust grains in the ISM.
| Variation of interstellar depletion [log(X/H)_observed - log(X/H)_Solar] with "condensation temperature", in an average "dense interstellar cloud". Circles show average depletions derived from observations of at least three stars; triangles, those from fewer than three stars. Squares identify depletions inferred indirectly from trace ionization stages. The correlation with condensation temperature is tighter than that with first ionization potential; deviations from the general trend may be explained by various chemical processes thought to occur either during grain formation or in the ISM. Less severe depletions are found for warmer, more diffuse clouds in the Galactic disk and halo, and also for high-velocity clouds (in which the grains may have been significantly disrupted by shocks). |
See also papers by Cardelli, Meyer, Sofia, and collaborators.
Return to top of this page
The diffuse interstellar bands (DIBs), first recognized in 1922 by Heger, refer to a large number of interstellar absorption features (several hundred now known) which are (1) broader than the identified atomic and molecular lines, (2) generally, but somewhat loosely correlated with the total amounts of gas and dust present, and (3) of unknown origin. We (Snow, York, Hobbs, Thorburn, Oka, McCall, Sonnentrucker, Welty) are compiling a database of high S/N, moderate resolution optical spectra (3700--10,000A), high resolution optical spectra of selected atomic and molecular lines, UV and far-UV spectra, and IR spectra of about 60 stars --- in order to compare the DIB strengths with the abundances of various atomic and molecular species and with various diagnostics of cloud physical conditions --- and thus to understand better the behavior and origin(s) of the DIBs.
Return to top of this page
Spectra obtained with the HST GHRS and with IMAPS have revealed high-velocity (HV) absorption features (at velocities typically about -100 km/s) toward the Galactic disk stars 23 Ori, zeta Ori, tau CMa, and eta Tau. The HV absorption is seen prominently in the strong lines of C II, C II*, Mg II, Al II, Al III, Si II, and Si III (and in some cases, in lines of S II, S III, Fe II, and perhaps Fe III) -- but not in the lines of O I and N I that are typically very strong in neutral (H I) gas. These HV clouds are thus essentially fully ionized. While the various ionization ratios would be broadly consistent with collisional ionization equilibrium at a temperature of about 25,000 K, the line widths of individual resolved components in the HV gas indicate that the temperatures are actually 8,000-10,000 K. These HV clouds thus appear not to be in equilibrium -- they have cooled more quickly than the higher ionization states could recombine, presumably after having been shocked. While the total elemental abundances in the HV gas are in some cases uncertain, due to not having data for all potentially significant ionization stages, it appears that Al, Fe, and perhaps Si may still be somewhat depleted, even in gas that has been strongly shocked. The HV gas observed toward eta Tau exhibits remarkably strong C II*, relative to C II -- perhaps due to optical pumping; the HV gas in this case is likely to be circumstellar material within or near this binary system.
Return to top of this page
The Solar System currently is apparently immersed in a warm, diffuse interstellar cloud -- whose properties define the boundary conditions of the Solar System. The properties of this "local" or "surrounding" interstellar cloud may be investigated by observing the optical and UV interstellar absorption lines present in the spectra of very nearby stars. In collaboration with P. Frisch, I have been obtaining high-resolution, high S/N ratio Ca II and Na I spectra toward a number of nearby stars, in order to investigate the morphology and properties of the nearest interstellar cloud(s).
See also the proceedings of the recent Local Bubble and Beyond conference (IAU Colloquium 166) and recent papers by Redfield & Linsky.
Return to top of this page
Return to my home page
Go to UC Astro home page
Last modified 04 Feb 2011