The European Large-Area Infrared Space Observatory Survey V: a BeppoSax hard X-ray survey of the S1 region

We present BeppoSAX observations of the southern S1 region in the European Large-Area Infrared Space Observatory (ISO) Survey (ELAIS). These observati ons cover an area of similar to1.7 deg(2) and reach an on-axis (similar to0 .7 deg(2)) 2-10 keV (hard X-ray, or HX) sensitivity of similar to 10(-13) e rgs s(-1) cm(-2). This is the first HX analysis of an ISOCAM survey. We det ect nine sources with a signal-to-noise ratio SNRHX > 3, four additional so urces with a 1.3-10 keV (total X-ray, or T) SNRT > 3, and two additional so urces that seem to be associated with QSOs having SNRT > 2.9. The number de nsities of the SNRHX > 3 sources are consistent with the ASCA and BeppoSAX log N-log S functions. Six BeppoSAX sources have reliable ISOCAM 15 mum cou nterparts within similar to 60". All these ISOCAM sources have optical coun terparts of R < 20 mag. Five of these sources have been previously opticall y classified, four as QSOs and one as a broad absorption line (BAL) QSO at z = 2.2. The remaining unclassified source has X-ray and photometric proper ties consistent with those of a nearby Seyfert galaxy. One further HX sourc e has a 2.6 <sigma> ISOCAM counterpart associated with a galaxy at z = 0.32 5. If this ISOCAM source is real, the HX/MIR properties suggest either an u nusual QSO or a cD cluster galaxy. We have constructed MIR and HX spectral energy distributions to compute the expected HX/MIR ratios for these classe s of objects up to z = 3.2 and assess the HX/MIR survey depth. The BAL QSO has an observed X-ray softness ratio and HX/MIR flux ratio simi lar to those of QSOs but different from those found for low-redshift BAL QS Os. This difference can be explained in terms of absorption, and it suggest s that high-redshift BAL QSOs should be comparatively easy to detect in the HX band, allowing their true fraction in the high-redshift QSO population to be determined. The QSOs cover a wide redshift range (0.4 <z <2.6) and ha ve HX/MIR flux ratios consistent with those found for nearby IRAS and optic ally selected Palomar-Green QSOs. This suggests that MIR-selected QSOs of R <20 mag come from the same population as optically selected QSOs. We confir m this with a comparison of the B/MIR flux ratios of MIR and blue-band-sele cted QSOs.