Abstract

We report on a NuSTAR and XMM-Newton program that has observed a sample of three extremely luminous, heavily obscured WISE-selected active galactic nuclei (AGNs) at z similar to 2 across a broad X-ray band (0.1 -79 keV). The parent sample, selected to be faint or undetected in the WISE 3.4 mu m (W1) and 4.6 mu m (W2) bands but bright at 12 mu m (W3) and 22 mu m (W4), are extremely rare, with only similar to 1000 so-called W1W2-dropouts across the extragalactic sky. Optical spectroscopy reveals typical redshifts of z similar to 2 for this population, implying rest-frame mid-IR luminosities of nu L-nu (6 mu m) similar to 6 x 10(46) erg s(-1) and bolometric luminosities that can exceed L-bol similar to 10(14) L-circle dot. The corresponding intrinsic, unobscured hard X-ray luminosities are L(2-10 keV) similar to 4 x 10(45) erg s-(1) for typical quasar templates. These are among the most AGNs known, though the optical spectra rarely show evidence of a broad-line region and the selection criteria imply heavy obscuration even at rest-frame 1.5mm. We designed our X-ray observations to obtain robust detections for gas column densities N-H <= 10(24) cm(-2). In fact, the sources prove to be fainter than these predictions. Two of the sources were observed by both NuSTAR and XMM-Newton, with neither being detected by NuSTAR (f(3-24 keV) less than or similar to 10(-13) erg cm(-2) s(-1)), and one being faintly detected by XMM-Newton (f(0.5-10 keV) similar to 5 x 10(-15) erg cm(-2) s(-1)). A third source was observed only with XMM-Newton, yielding a faint detection (f(0.5-10 keV) similar to 7 x 10(-15) erg cm(-2) s(-1)). The X-ray data imply these sources are either X-ray weak, or are heavily obscured by column densities N-H greater than or similar to 10(24) cm(-2). The combined X-ray and mid-IR analysis seems to favor this second possibility, implying the sources are extremely obscured, consistent with Compton-thick, luminous quasars. The discovery of a significant population of heavily obscured, extremely luminous AGNs would not conform to the standard paradigm of a receding torus, in which more luminous quasars are less likely to be obscured, and instead suggests that an additional source of obscuration is present in these extreme sources.