Polarimetry of compact radio sources
Magnetic fields are probably involved in feeding gas into AGN, but their strengths and orientations are mostly not known and estimated by equipartition arguments, i.e., assuming that an equal amount of energy is comtained in particles and in magnetic fields. This is not very satisfactory, since one does not know whether this assumption is justified or not.
One can actually measure magnetic fields in the following way. Some radio sources have good evidence for a foreground gas which produces a narrow region of decreased radio emission along the jet (actually, this region casts a radio shadow onto us). The mechanism can be identified as free-free absorption, i.e., elastic scattering of radiation by free electrons. If this absorber is interspersed with a magnetic field, then the electric vecor position angle, or EVPA, of polarized emission is rotated in the magneto-ionized plasma. This is called Faraday rotation, and since radio sources emit highly polarized synchrotron emission, one should be able to measure such an effect.
Measuring the free-free absorption and the rotation of the EVPA can yield a direct measurement of the magnetic field strength, because free-free absorption is proportional to the square of the particle density in the foreground medium, and Faraday rotation is directly proportional to particle density. Thus, the two effects can be disentangled and the magnetic field strength can be measured.
We have searched for such an effect in the five nearby radio sources NGC1052, NGC4261, Hydra A, Centaurus A and Cygnus A, using the VLBA at 15 GHz. Al sources turned out to entirely unpolarized (at levels of 0.5% to 1%) in the core and along the jets, with the exception of Cygnus A, which showed a degree of polarization of 0.4%. This is surprising because the emission is certainly polarized synchrotron emission, and so the polarization must have been removed somewhere.
Considering many possible causes of depolarization, the least exotic explanation is a spherical accretion flow onto the AGN (Bondi accretion). If it is turbulent enough so that many "Faraday cells" with different EVPAs are averaged together in the synthesized beam, then the net polarization observed by us is severly decreased.
As Faraday rotation is inversely proportional to the wavelength squared, the effect is much lower at shorter wavelength, and polarization should be observable at, say, 43 GHz.
Collaborators: Alan Roy, Uwe Bach, Denise Gabuzda, Thomas Krichbaum
Back to main page 
15 GHz polarimetric image of Hydra A. The contours indicate flux densities of 0.87 mJy/beam x 2N. Colors indicate fractional polarization. The small patches of up to almost 80% polarized intensity 2 mas south-west of the core are imaging artefacts and not significant; the source is entirely unpolarized.
Last modified: Thu Apr 1 20:01:51 CEST 2004