BAT X-ray survey. I. Methodology and X-ray identification

We applied the maximum likelihood (ML) method, as an image reconstruction algorithm, to the BAT X- Ray Survey (BXS). This method was specifically designed to preserve the full statistical information in the data and to avoid mosaicking of many exposures with different pointing directions, thus reducing systematic errors when co- adding images. We reconstructed, in the 14-170 keV energy band, the image of a 90x90 deg(2) sky region, centered on (R.A., decl.) (105 degrees,-25 degrees), which BAT surveyed with an exposure time of similar to 1 Ms ( in 2005 November). The best sensitivity in our image is similar to 0.85 mcrab or 2: 0; 10(-11) ergs cm(-2). We detect 49 hard X- ray sources above the 4.5 sigma level; of these, only 12 were previously known as hard X- ray sources (> 15 keV). Swift XRT observations allowed us to firmly identify the counterparts for 15 objects, while 2 objects have Einstein IPC counterparts ( Harris et al. 1990); in addition to those, we found a likely counterpart for 13 objects by correlating our sample with the ROSAT All- Sky Survey Bright Source Catalog ( Voges et al. 1999). Seven objects remain unidentified. Analysis of the noise properties of our image shows that similar to 75% of the area is surveyed to a flux limit of similar to 1 mcrab. This study shows that the coupling of the ML method to the most sensitive, all- sky surveying, hard X- ray instrument, BAT, is able to probe for the first time the hard X- ray sky to the millicrab flux level. The successful application of this method to BAT demonstrates that it could also be applied with advantage to similar instruments such as INTEGRAL IBIS.