2017MNRAS.469.4675M

Quasar redshifts require the best possible precision and accuracy for a number of applications, such as setting the velocity scale for outflows as well as measuring small-scale quasar-quasar clustering. The most reliable redshift standard in luminous quasars is arguably the narrow [O III] λλ4959, 5007 emission line doublet in the rest-frame optical. We use previously published [O III] redshifts obtained using near-infrared spectra in a sample of 45 high-redshift (z > 2.2) quasars to evaluate redshift measurement techniques based on rest-frame ultraviolet spectra. At redshifts above z = 2.2, the Mg II λ2798 emission line is not available in observed-frame optical spectra and the most prominent unblended and unabsorbed spectral feature available is usually C IV λ1549. Peak and centroid measurements of the C IV profile are often blueshifted relative to the rest-frame of the quasar, which can significantly bias redshift determinations. We show that redshift determinations for these high-redshift quasars are significantly correlated with the emission-line properties of C IV (i.e. the equivalent width, or EW, and the full width at half-maximum, or FWHM) as well as the luminosity, which we take from the Sloan Digital Sky Survey Data Release 7. We demonstrate that empirical corrections based on multiple regression analyses yield significant improvements in both the precision and accuracy of the redshifts of the most distant quasars and are required to establish consistency with redshifts determined in more local quasars.