[Previous] | [Session 80] | [Next]
R. I. Reid (Dominion Radio Astrophysical Observatory), P. P. Kronberg (Los Alamos National Laboratory)
This paper describes a generalization and improvement of Kronberg, Dyer, Burbidge, and Junkkarinen's (1991) method for detecting weak gravitational lenses by using polarized jets as backgrounds. Gravitational lenses produce coherent distortions of background sources, but do not rotate their polarization vectors. In the absence of lensing, the intrinsic polarization angle is often aligned with the morphology of the source, so discrepancies can reveal the presence of a foreground lens, without requiring the lens to be strong enough to produce multiple images. The original method compared the direction of synchrotron jets, as defined by a spline in the image plane, to their polarization structure. This new method uses the gradient of the entire image, which permits us to both add information from jet lobes, and to directly compare the lens model to the visibility measurements of our observations in the uv plane where the uncertainties are well understood.
The lens parameters (mass, location, and major and minor axes if warranted by the data) are found by minimizing \chi2 for the difference of the observed and model visibilities, and the lens is tested for significance. Typically the errors for the gradient directions are much less than the polarization angle errors, so the weights for \chi2 are just the thermal noise of the visibilities. Otherwise gradient direction uncertainties must also be accounted for with a Monte Carlo technique.
This work has been supported by NSERC grant #5713 to P. Kronberg, and a NSERC Visiting Fellowship for R. Reid at the Dominion Radio Astrophysical Observatory.
[Previous] | [Session 80] | [Next]
Bulletin of the American Astronomical Society, 34, #4
© 2002. The American Astronomical Soceity.