Polarimetry is a powerful tool for the study of magnetic fields on interstellar scales and larger scales (e.g., Table 2).
A review has been carried out of the polarimetric observations to date, and their deductions for the magnetic fields on large scales (1 pc and up), in the Milky Way galaxy and beyond. In this review, I have attempted to tie together the many observational pieces of the puzzle of the galactic magnetic fields in the Milky Way (e.g., Table 1). If at places it appears incomplete, it should be viewed as an incentive to continue to piece the puzzle together to the end, Future trends over the subject areas covered in this review have been indicated.
The pace of progress in observational polarimetry is hampered by the relative lack of polarimetric instruments, by the larger amount of time needed as compared to other observational searches, and possibly by the necessity to do a better job of educating fellow astronomers on Telescope Time Allocation Committees about the usefulness of polarimetry.
Narrow magnetized features are found in the interstellar medium of the Milky Way disk (filaments, outflow cavities, edges of molecular clouds), and in the Galactic Center.
A few magnetic field maps of dusty molecular clouds observed at Extreme-Infrared (800 µm) wavelengths have now been published in the 1990s, e.g., W75N-IRS1, M17-SW (e.g., Figure 3 here), Sagittarius B2, and MonR2 core. Preliminary resuls indicate that the magnetic field lines in MonR2 core and in M17-SW do bend, suggesting that there is some evolution of the magnetic field inside a molecular cloud, but without excessive tangling of the magnetic field lines. Future trends: Extreme Infrared polarimetric observations of the magnetic field distribution in molecular clouds are in their infancy. Air-borne Far-Infrared (e.g., ISO, SOFIA) and Ground-based Extreme-Infrared (e.g., JCMT, CSO) polarimetry of dust aligned by magnetic fields in molecular clouds is likely to better reveal the complete structure of molecular clouds, and their dynamical interactions with the interstellar medium gas and with other clouds. Classification systems of magnetic fields in molecular clouds have been discussed (e.g., Fig. 2 here).
Numerous interstellar superbubbles near the sun have been found to be magnetized, causing local deviations of the larger scale regular galactic magnetic field in the Milky Way (e.g., Fig. 6 here). The magnetic field in the shells of superbubbles follows the relation B ~ nk with k 1.0, suggestive of shocked expanding gas (e.g., Figure 5 here). A similar exponent B ~ n1.0 is expected in the shells of supernovae remnants. A similar B ~ n1.0 is expected in the compressed edges of molecular clouds near HII regions. Future trends: as there are still less than a dozen shells with a measured magnetic field strength, more observations are needed for better statistical analyses, and to look for other predicted trends. The theories of magnetic fields in superbubbles are likewise evolving rapidly.
Outside of these shocked expanding shells, there is a B ~ nk relation for gas and magnetic field, and the data show k = 0.5 for objects from n > 100 cm-3 up to n > 1011 cm-3, suggestive of an equilibrium between various energies (e.g., Figure 1 here). Future trends: observations of even more dense objects, nearer n 1020 cm-3, are needed to test theories of star formation, and the predicted flattening of the B ~ nk curve.
Data also show a B ~ nk relation, but with k 0.2 for objects with size > 100 pc and n < 100 cm-3, suggestive of relatively free gas motions bunching up along magnetic field lines (e.g., Figure 13 here).
In the Galactic center, magnetic field strengths of . 1000 µGauss over small regions of dense gas have been measured. Future trends: the presence or absence of strong magnetic field over large scales in the galactic center region is still problematic, and in need of more observations. Much theoretical work and observational studies need to be done before the overall magnetic field structure within the central 200 parsecs of our Galaxy is understood. Certain elements of a dynamo may be present there, with gas motions changing vertical, poloidal magnetic fields into toroidal, azimuthal magnetic fields.
There is an observed relationship between the galactic magnetic field B and the star formation rate SFR, or star formation efficiency SFE, of the form B ~ SFRj and B ~ SFEj, with j 0.13 (e.g., Figure 12 here). In combination with the B ~ nk law with k 0.2 over large scales, one finds that SF ~ n1.3 for nearby spiral galaxies. Future trends: more observations are needed to ascertain the SFE and SFR over sub-galactic regions, such as inside a specific spiral arm within the Milky Way Galaxy or within another spiral galaxy.
The most likely seed for galactic magnetism could be recent local ejecta from stellar winds and supernovae. The most likely amplification and orientation mechanism could be the galactic dynamo. Future trends: theories of seed magnetic fields are evolving rapidly, and tests of their predictions should be amenable to observations soon.
Over a galactic scale, the observations indicate that the Milky Way probably has 4 main spiral arms (e.g., Figure 7 here), 2 magnetic field reversals (e.g., Figure 8 here), and a large scale azimuthal ASS magnetic field shape (e.g., Figure 9 here).
Of the three main methods used to derive galactic magnetic fields, the Faraday rotation method seems to be the most accurate for giving the galactic magnetic field strength, which is about the same as that obtained from the equipartition method (e.g., Figure 11 here). The Faraday rotation method is the only one capable of giving the galactic magnetic field direction. Future trends: There are hundreds of QSO and galaxies with a detected RM (e.g., Figure 4 here), but there are thousands of QSO and distant galaxies that lack a known RM, due to their faintness. Better sensitivity in Faraday observations should help extend the areas of coverage of this technique.
In the 1990s, most large scale magnetic fields in spiral galaxies seem amenable to interpretation with the currently favored dynamo theories. Future trends: while the dynamo theories can readily explain many features of the galactic magnetic fields, the question of the viability of the dynamos or its replacement by some galactic fountain theories has not been settled entirely.
A majority of spiral galaxies have the axisymmetric magnetic field shape mazim = 0. A number of spiral galaxies have the bisymmetric magnetic field shape mazim = 1 (e.g., Figure 10 here). In M31, the azimuthal ASS magnetic field shape prevails, without magnetic field reversal (e.g., Figure 14 here). The mazim = 1 shape may be due to past tidal interactions with nearby companions (e.g., Figure 15 here), as evidenced by a distorted HI tail outside of the galaxy (MD = 4). Future trends: Statistics suggest that a simple B shape corresponds to a simple HI shape, and that a complex B shape corresponds to a complex HI shape (e.g., Figure 16). Less than 2 dozen galaxies have had their global magnetic field structure assessed, and more observations are needed to test more predictions of recent theories in these areas.
The intracluster gas inside clusters of galaxies contains an intergalactic intracluster magnetic field near 1 µGauss, as detected by its contribution to the rotation measure. The best seed for this intracluster intergalactic magnetic field appears to be the ejecta from the interstellar medium of galaxies in clusters, via stellar winds, supernovae, galactic fountains, tidal strippings, etc. Future trends: as less than a dozen clusters have had their magnetic field strengths measured, more observations are needed to better test the fitting of the recent theories of magnetic fields in clusters of galaxies.
I thank Ms. Lyne Séguin (NRCC-Ottawa) for help in drawing some of the figures (2, 3, 6, 9, 15); the PGPLOT software was employed for the others. I am grateful to many people who spent much of their scientific lives on researching mid-scale and large-scale magnetic fields.