... values

Our background subtraction procedure assumes that the CRXB does not vary within the field of view. The low value of the local CRXB (as compared to blank field value) may partly be due to a higher galactic absorption (Sect. 5.3). We thus cannot exclude that the CRXB actually varies within the FOV, in view of the observed  ${N_{\rm H}}$ variations (see Fig. 7). However, this further test shows that the double subtraction is really needed only for the cluster outskirts ( $5\hbox{$^\prime$ }{-}12\hbox{$^\prime$ }$), closest to the region chosen to compute the local background (r>16'). This should minimise any artifact due to our neglect of possible background radial variation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.