Active galactic nuclei in the CNOC2 field galaxy redshift survey

We present a sample of 47 confirmed and 14 candidate active galactic nuclei (AGNs) discovered in the Canadian Network for Observational Cosmology field galaxy redshift survey (CNOC2). The sample consists of 38 objects identified from broad emission lines, eight from narrow [Ne v] emission, and 15 candidates from Fe II or Mg II absorption lines, one of which has been confirmed as a broad-line AGN via infrared spectroscopy. Redshifts of these AGNs range from z = 0.27 to z = 4.67, and the average absolute magnitude is M(B) similar or equal to -22.25, below the quasar/Seyfert division at M(B) = -23. Only two of the AGNs are detected at radio wavelengths. We find that only 0.3%+/-0.1% of galaxies brighter than similar toM* + 1 at 0.281 < z < 0.685 contain broad-line or [Ne v] AGNs. We find a total surface density of 278-400 AGNs deg(-2) to R = 22.09, comparable to previously published estimates. About 20% of these AGNs are classified as resolved or probably resolved in CFHT seeing and might be missed in surveys that target unresolved objects only. The sample includes several unusual objects: one with a very strong double-peaked Mg Ir emission line, several with unusual emission-line properties, one with an O III lambda 3133 broad absorption line, at least one with an optical absorption-line spectrum but broad Ha emission in the near-IR. No color selection criteria were involved in selecting this spectroscopically discovered sample. The sample is also unbiased against objects with luminous host galaxies, since the spectroscopy preferentially targeted extended objects. Simple color-color diagram selection criteria can recover similar to 81% +/- 6% of the CNOC2 AGNs, but several of the most unusual objects would be missing from such a color-selected sample. In the subsample of broad emission line-selected AGNs, the average equivalent widths for Mg II and C mi agree with the predictions of previous studies of the Baldwin effect. However, the average equivalent widths for C IV and Ly alpha are smaller than predicted by previous studies of the Baldwin effect at lower redshift. This may imply that the slopes of the C rv and Ly alpha Baldwin effects evolve with redshift, steepening with cosmic time. The broad emission line subsample also shows a higher incidence of associated Mg II lambda 2798 absorption than in most previous surveys and an incidence of associated C rv lambda 1549 absorption that may be more similar to that of radio-selected quasar samples than optically selected ones. This may arise from strong absorption being anticorrelated with optical luminosity or becoming less frequent with cosmic time or possibly because our selection method is not biased against objects with resolved spatial structure or reddened by dust associated with the absorbing gas.