The emission-line properties of the QSO population at z approximately 2

We compare the ultraviolet emission-line properties of the three subclasses of the QSO population (radio quiet, flat radio spectrum radio loud, steep radio spectrum radio loud) during the 'QSO epoch' at z approximately 2. Our total sample consists of 79 objects that are well matched in luminosity and redshift, which should allow unbiased comparison of their spectral properties. We confirm recent results that radio-loud objects have, on average, narrower and stronger C IV lambda 1549 and C III) lambda 1909 emission lines than radio-quiet objects. This is likely the effect of larger average intermediate-line and/or narrow-line regions in radio loud objects, including a relative excess of emission on very extended (approximately 100 kpc) scales. We additionally find that radio-loud objects show a significantly wider range in the asymmetry of the C IV lambda 1549 profile than radio-quiet objects, with cases of both redward and blueward asymmetry. By contrast, radio-quiet objects almost exclusively show blueward asymmetry in this line, and cases of strong redward asymmetry appear to be limited to steep-spectrum objects. Statistically, there is a closer correspondence between the emission-line properties of radio-quiet and flat-spectrum objects than between either of these subclasses and steep-spectrum objects, whose mean emission-line properties differ strongly from those of the radio-quiet sample. This result is similar to that found for the properties of the optical emission lines in low-redshift objects and is not easily understood under the proposed unified models of the flat- and steep-spectrum sources. We also find that 20%-30% of radio-loud objects show narrow (approximately 1000 km/s) emission in He II lambda 1640. No cases of such emission are found among radio-quiet objects. There is evidence that this spectral feature arises in extended ionized halos, which may thus be associated with the objects' radio ejecta, cluster cooling flows, or a combination of both.