The X-ray variability of a large, serendipitous sample of spectroscopic quasars

We analyze the X-ray variability of 264 Sloan Digital Sky Survey spectroscopic quasars using the Chandra public archive. This data set consists of quasars with spectroscopic redshifts out to z 5 and covers rest-frame timescales up to Δt _sys 2000 days, with three or more X-ray observations available for 82 quasars. It therefore samples longer timescales and higher luminosities than previous large-scale analyses of active galactic nucleus (AGN) variability. We find significant (≳3σ) variation in 30% of the quasars overall; the fraction of sources with detected variability increases strongly with the number of available source counts up to 70% for sources with ≥1000counts per epoch. Assuming that the distribution of fractional variation is Gaussian, its standard deviation is 16% on ≳1 week timescales, which is not enough to explain the observed scatter in quasar X-ray-to-optical flux ratios as being due to variability alone. We find no evidence in our sample that quasars are more variable at higher redshifts (z > 2), as has been suggested in previous studies. Quasar X-ray spectra vary similarly to some local Seyfert AGNs in that they steepen as they brighten, with evidence for a constant, hard spectral component that is more prominent in fainter stages. We identify one highly variable Narrow Line Seyfert 1-type spectroscopic quasar in the Chandra Deep Field-North. We constrain the rate of kilosecond-timescale flares in the quasar population using 8 months of total exposure and also constrain the distribution of variation amplitudes between exposures; extreme changes (>100%) are quite rare, while variation at the 25% level occurs in <25% of observations. [O III] λ5007 emission may be stronger in sources with lower levels of X-ray variability; if confirmed, this would represent an additional link between small-scale (corona) and large-scale (narrow-line region) AGN properties.