Scrutinizing the unresolved x-ray background in the CDFS field via transdimensional sampling
Abstract
We analyse the Chandra Deep Field South (CDFS) data using a novel inference framework called probabilistic cataloging, where fair samples are taken from the posterior distribution of the catalog space of point sources in the field consistent with the merged CDFS exposure. This requires constructing a Markov chain of samples via transdimensional proposals involving births and deaths as well as within-model proposals. Our point source model is also hierarchical, allowing the shape of the flux distribution of the point sources to be inferred simultaneously. By marginalizing over point sources below the detection threshold, robust uncertainty estimates can be placed on the flux and color distribution of Active Galactic Nuclei (AGNs) as well as the normalization and spectrum of the truly isotropic background. We validate our method using simulated deep Chandra exposures and show that the isotropic background emission can be constrained at the 10% level, which is a result that takes into account its covariance with unresolved AGNs and the particle background of Chandra. We compare our method to results derived from fluctuation analyses, i.e., study of the 1-point function of photon counts across the image, and discuss how our results relate to the AGN population synthesis models.
- Publication:
-
AAS/High Energy Astrophysics Division #16
- Pub Date:
- August 2017
- Bibcode:
- 2017HEAD...1611304D