IMAGING THE HOST GALAXIES OF HIGH-REDSHIFT RADIO-QUIET QSOS

We present new deep K-band and optical images of four radio-quiet QSOs at z approximate to 1 and six radio-quiet QSOs at z approximate to 2.5, as well as optical images only of six more at z approximate to 2.5. We have examined the images carefully for evidence of extended ''fuzz'' from any putative QSO host galaxy. None of the z approximate to 2.5 QSOs shows any extended emission, and only two of the z approximate to 1 QSOs show marginal evidence for extended emission. Our 3 sigma detection limits in the K images, m(K) approximate to 21 for an isolated source, would correspond approximately to an unevolved L* elliptical galaxy at z = 2.5 or to 2-3 mag fainter than an L* elliptical at z = 1, although our limits on host galaxy light are weaker than this due to the difficulty of separating galaxy light from QSO light. We simulate simple models of disk and elliptical host galaxies, and find that the marginal emission around the two z approximate to 1 QSOs can be explained by disks or bulges that are similar to 1-2 mag brighter than an unevolved L* galaxy in one case and similar to 1.5-2.5 mag brighter than L* in the other. For two other z approximate to 1 QSOs, we have only upper limits (L approximate to L*). The hosts of the high-redshift sample must be no brighter than about 3 mag above an unevolved L* galaxy, and are at least 1 magnitude fainter than the hosts of radio-loud QSOs at the same redshift. If the easily detected K-band light surrounding a previous sample of otherwise similar but radio-loud QSOs is starlight, then it must evolve on timescales of greater than or similar to 10(8) yr (e.g., Chambers and Chariot 1990); therefore our nondetection of host galaxy fuzz around radio-quiet QSOs supports the view that high-redshift radio-quiet and radio-loud QSOs inhabit different host objects, rather than being single types of objects that turn their radio emission on and off over short timescales. This is consistent with the general trend at low redshifts that radio-loud QSOs are found in giant elliptical galaxies while radio-quiet QSOs are found in less luminous disk galaxies. It also suggests that the processes responsible for the spectacular properties of radio-loud AGNs at high redshifts might not be generally relevant to the (far more numerous) radio-quiet population.