Using principal component analysis to understand the variability of PDS 456

05 Jun 2018

We present a spectral-variability analysis of the low-redshift quasar PDS 456 using principal component analysis. In the XMM-Newton data, we find a strong peak in the first principal component at the energy of the Fe absorption line from the highly blueshifted outflow. This indicates that the absorption feature is more variable than the continuum, and that it is responding to the continuum. We find qualitatively different behaviour in the Suzaku data, which is dominated by changes in the column density of neutral absorption. In this case, we find no evidence of the absorption produced by the highly ionized gas being correlated with this variability. Additionally, we perform simulations of the source variability, and demonstrate that PCA can trivially distinguish between outflow variability correlated, anti-correlated, and un-correlated with the continuum flux. Here, the observed anti-correlation between the absorption line equivalent width and the continuum flux may be due to the ionization of the wind responding to the continuum. Finally, we compare our results with those found in the narrow-line Seyfert 1 IRAS 13224–3809. We find that the Fe K UFO feature is sharper and more prominent in PDS 456, but that it lacks the lower energy features from lighter elements found in IRAS 13224–3809, presumably due to differences in ionization.