Full Resolution Deconvolution of Complex Faraday Spectra

ABSTRACT Polarized synchrotron emission from multiple Faraday depths can be separated by calculating the complex Fourier transform of Stokes’ parameters as a function wavelength squared, known Synthesis. As commonly implemented, introduces an additional term $\lambda _0^2$, which broadens real and imaginary spectra but not amplitude spectrum. We use idealized tests to investigate whether information recovered with clean process restoring beam set narrower width peak in ‘full’ resolution spectrum _0^2=0$. find that _0^2$ choice makes no difference, except for smaller beam. With this beam, accuracy phase stability are unchanged single components. However, using components, we (a) better discrimination (b) significant reductions blending structures tomography images, (c) reduction spurious features maps. also discuss limited on scales comparable note clean-bias, reducing amplitudes. present examples MeerKAT L-band data. revisit maximum depth, surveys sensitive, introduce variable Wmax, power drops factor 2. most cannot resolve continuous distributions unless full is used.