Searching for the trigger of the AGN QPO: 8 years of RE J1034+396

TL;DR

This study analyzes eight years of XMM-Newton observations of the AGN RE J1034+396, revealing the non-stationary nature of its QPO, exploring its spectral and temporal evolution, and investigating potential triggers for the QPO phenomenon.

Contribution

The paper provides a detailed multi-epoch analysis of RE J1034+396, identifying the non-stationary behavior of its QPO and proposing a link between QPO-induced ionization changes and soft X-ray components.

Findings

QPO was present in some observations but absent in others.

The soft component lags behind hard X-rays by a light travel time of about 30 R_g.

Absorption alone cannot explain the QPO; ionization changes are involved.

Abstract

RE J1034+396 is one of the most extreme Narrow-line Seyfert 1s detected thus far, showing the only quasi-periodic oscillation (QPO) reliably detected in an Active Galactic Nucleus (AGN). Comparison with similar spectral and timing properties observed in the black hole X-ray binary (BHB) GRS 1915+105 suggests that RE J1034+396 is a super-Eddington accretor. A more complete understanding of the behaviour of RE J1034+396 can therefore lead to a unification of the accretion physics between such extreme AGN and super-Eddington BHBs. Here we report on our latest XMM-Newton observations of RE J1034+396, which no longer show the QPO, indicating that this source shows a non-stationary power-spectrum. We use spectral and temporal analysis across all five XMM-Newton observations of the source to probe the evolution of the object. The combination of the shape of the fractional variability with energy and the inferred velocity of absorbing material in the line-of-sight rules out an absorption-only method of creating the QPO. Instead the periodically changing absorption may be produced by the QPO causing a change in ionization state. We extend our analysis by including the covariance spectra which give much better signal to noise than an rms spectrum. These reveal a new aspect of the QPO, which is that there is also a small contribution from a soft component which is hotter than the soft excess seen in the mean spectrum. Folding the lightcurve on the QPO period shows that this component lags behind the hard X-rays. If this is due to re-processing then the lag corresponding to a light travel time across 30 R_g. Some of the remaining observations have similar energy spectra and covariance spectra, but none of them show a significant QPO, so we conclude that none of these features are the trigger for the appearance of the QPO in this object.