2004ApJ...608..957A

This study addresses, primarily, positional uncertainties of sources imaged by the X-ray cameras on the Röntgensatellit (ROSAT). Monte Carlo simulations are exploited to estimate the precision of centroiding point-source ``photon clouds'' in ROSAT event lists. Measurements of bright coronal stars demonstrate that the design specification 6" accuracy of a ROSAT pointing center was achieved in practice. Furthermore, the accuracy often can be improved to 2" or better by matching serendipitous X-ray sources to an optical catalog like the US Naval Observatory A2. Even a low signal-to-noise detection usually can be localized well enough to obtain a unique identification with a cataloged object or to establish that no bright optical counterpart is present. Secondarily, and in a more general context, the study addresses detection significance and flux limits in the sparse counting regime. Building on previous work by Gehrels and Kraft, Burrows, & Nousek, scaling laws are developed to estimate detection thresholds and lower and upper bounds on possible source fluxes. Although the specific application of ROSAT was in mind, the flux limit approximations apply to other situations governed by Poisson statistics; including ultraviolet and X-ray spectra with low source counts and nonnegligible backgrounds.