2005ApJ...619..297V

The rotation measure RM and location L of over 350 pulsars in the Milky Way are used to constrain the structure of the large-scale magnetic field in the Milky Way disk. Various magnetic field directions in various concentric circles are employed in the mapping to fit the RM and L values, allowing the testing of various models. This approach can build up signal-to-noise ratios by averaging across large rings (of size 1 kpc); at worst, this approach may miss a (double) field reversal occurring within a ring. My current work here begins with the assumption of concentric rings. In the best pulsar-based fit, an overall clockwise-going magnetic field (as seen from the north Galactic pole) extends radially at least from 1 to 12 kpc from the Galactic center, except for a 2 kpc wide anticlockwise magnetic field located in an annulus between 4 and 6 kpc from the Galactic center. The best fit has a very special feature in the form of a string of H II regions located in the anticlockwise annulus. For the mean pulsar RM as averaged over radial rings, the system of a clockwise-going Galactic magnetic field with only one anticlockwise-going field in an annulus produces the lowest reduced χ². All other fits with a different number of field-reversed annuli have a 4 times larger reduced χ² or worse. Thus a completely clockwise fit (no reversed ring) has a 5 times larger reduced χ². Fits do not support more than one magnetically reversed annulus in the Galaxy, coming up with an 8 times larger reduced χ² for two opposite rings, 4 times larger for four opposite rings, 4 times larger for six opposite rings, and 9 times larger for 12 alternating zones. Splitting the pulsars above from those below the Galactic plane, the best fit with a single reversed ring holds true separately for pulsars above the plane and for pulsars below the plane. The origin of the anticlockwise annulus could be due to a number of factors (internal or external) or could be primordial (regular or chaotic).