MILLIMETER OBSERVATIONS OF A SAMPLE OF HIGH-REDSHIFT OBSCURED QUASARS

We present observations at 1.2 mm with Max-Planck Millimetre Bolometer Array (MAMBO-II) of a sample of z greater than or similar to 2 radio-intermediate obscured quasars, as well as CO observations of two sources with the Plateau de Bure Interferometer. The typical rms noise achieved by the MAMBO observations is 0.55 mJy beam(-1) and five out of 21 sources (24%) are detected at a significance of >= 3 sigma. Stacking all sources leads to a statistical detection of < S-1.2 mm > = 0.96 +/- 0.11 mJy and stacking only the non-detections also yields a statistical detection, with < S-1.2 mm > = 0.51 +/- 0.13 mJy. At the typical redshift of the sample, z = 2, 1 mJy corresponds to a far-infrared luminosity L-FIR similar to 4 x 10(12) L-circle dot. If the far-infrared luminosity is powered entirely by star formation, and not by active galactic nucleus heated dust, then the characteristic inferred star formation rate is similar to 700 M-circle dot yr(-1). This far-infrared luminosity implies a dust mass of M-d similar to 3 x 10(8) M-circle dot, which is expected to be distributed on similar to kpc scales. We estimate that such large dust masses on kpc scales can plausibly cause the obscuration of the quasars. Combining our observations at 1.2 mm with mid-and far-infrared data, and additional observations for two objects at 350 mu m using SHARC-II, we present dust spectral energy distributions (SEDs) for our sample and derive a mean SED for our sample. This mean SED is not well fitted by clumpy torus models, unless additional extinction and far-infrared re-emission due to cool dust are included. This additional extinction can be consistently achieved by the mass of cool dust responsible for the far-infrared emission, provided the bulk of the dust is within a radius similar to 2-3 kpc. Comparison of our sample to other samples of z similar to 2 quasars suggests that obscured quasars have, on average, higher far-infrared luminosities than unobscured quasars. There is a hint that the host galaxies of obscured quasars must have higher cool-dustmasses and are therefore often found at an earlier evolutionary phase than those of unobscured quasars. For one source at z = 2.767, we detect the CO(3-2) transition, with S-CO Delta nu = 630 +/- 50 mJy km s(-1), corresponding to LCO(3-2) = 3.2 x 10(7) L-circle dot, or a brightness-temperature luminosity of L'(CO(3-2)) = 2.4 x 10(10) K km s(-1) pc(2). For another source at z = 4.17, the lack of detection of the CO(4-3) line suggests the line to have a brightness-temperature luminosity L'(CO(4-3)) < 1 x 10(10) K km s(-1) pc(2). Under the assumption that in these objects the high-J transitions are thermalized, we can estimate the molecular gas contents to be M-II2 = 1.9 x 10(10) M-circle dot and < 8 x 10(9) M-circle dot, respectively. The estimated gas depletion timescales are tau(g) = 4 Myr and <16 Myr, and low gas-to-dust mass ratios of M-g/M-d = 19 and <20 are inferred. These values are at the low end but consistent with those of other high-redshift galaxies.