A Comprehensive View of Large-scale Magnetic Fields, with Emphasis on the Galactic Magnetic Field near the Sun

We examine the observations of large-scale magnetic elds in the Universe. We begin at the largest scale with clusters of galaxies and work our way down through galaxies and nally to the Milky Way, on which we concentrate in detail. We examine the observations of the Galactic magnetic eld, and their interpretation, under the philosophy that the Galactic magnetic eld is like that in other spiral galaxies. We use pulsar data, diiuse Galactic synchrotron emission, and starlight polarization data to discuss the Galaxy's global magnetic connguration and the uniform (B u), random (B r), and total (B t) components of the eld strength. We nd disagreement among conclusions derived from the various data sets and argue that the pulsar data are not the best indicator for large-scale Galactic eld. Near the Solar circle, we nd that the azimuthal average of B t is 4.2 G and we adopt B u 2:2 and B r 3:6 G. B t is higher in spiral arms, reaching 5:9 G. B t is higher for smaller R Gal , reaching 8:0 G for R Gal = 4:0 kpc. The pattern of eld lines is not concentric circles but spirals. The inclination of the magnetic spiral may be smaller than that of the Galaxy's spiral arms if our sample, which refers primarily to the interarm region near the Sun, is representative. However, it is not inconceivable that the local eld lines follow the Galaxy's spiral pattern, as is observed in external galaxies. I. MAGNETIC FIELDS ON THE LARGEST SCALES Detections of synchrotron radiation and Faraday rotation from clusters of galaxies provide secure evidence for intercluster magnetic elds. Mean Faraday Rotation Measures (RMs) of extragalactic radio sources located behind clusters of galaxies exceed those of a control sample (Kim, Tribble, and Kronberg 1991). Although there is a tight correlation of RM with mass ux in clusters with cooling ows (Taylor, Barton, and Gee 1994), the excess RM s are not restricted to clusters that have cooling ows(Goldschmidt and Rephaeli (1993). The excess RM , typically 30 rad m ?2 , combined with typical electron densities derived from the thermal X-ray emission, requires B u 0:02 G. Near the centers of the clusters the scale lengths for eld variability become very much smaller and the eld strengths larger; for example, for the Coma cluster the scale length becomes 1 kpc and the eld strength B u 8 …