X-raying the interstellar medium: ROSAT observations of dust scattering halos.

We have studied X-ray halos around 25 point sources and four supernova remnants using ROSAT observations. All sources were observed during the ROSAT all-sky survey, 8 point-like sources and 2 supernovae remnants have been studied additionally using ROSAT pointed observations. The shapes of the X-ray halos were fitted with commonly used dust models (e.g., the Mathis-Rumpl-Nordsiek grain size distribution), resulting in a determination of the fractional halo intensity, which in turn can be converted into a dust column density. From the simultaneously obtained X-ray spectra, the cold gas absorption and hence an equivalent hydrogen column was determined. A surprisingly good correlation exists between the simultaneously measured dust and hydrogen column densities, indicating that gas and dust must be to a large extent cospatial. For the X-ray sources with known optical counterparts, the visual extinction correlates well with the X-ray derived dust scattering optical depth τ_sca_: τ_sca_=0.087xA_v_(mag)xE(keV)^-2^ and N_H_[cm^-2^/A_v_]=1.79x10^21^. The strong correlation between both quantities indicates that, with only few exceptions, intrinsic absorption by the X-ray source itself does not affect the derived absorption column densities. This method offers also the interesting possibility of verifying the optical identification of an X-ray source by using an X-ray property itself. In particular we find that for Cas A the optical extinction and the derived X-ray scattering are on the regression line but are too low for the observed absorption. Since Cas A's supernova was not observed optically, this supernova event must have been obscured by a local dust cloud. Finally, we have found a high degree of azimuthal symmetry in all the dust scattering halos of our sample, indicating, that the majority of dust grains responsible for the halo formation cannot be highly clumped in clouds.