The X-ray variability of NGC 6814 - Power spectrum

NGC 6814 is unique among Seyfert galaxies in showing strong evidence for periodic X-ray behavior, first detected from an EXOSAT "long-look observation. " Here simulation techniques are used to obtain the X-ray variability power spectrum of the unevenly sampled Ginga light curve of the object. A simple power law is not an adequate description of the power spectrum, with the residuals showing excess power on time scales consistent with the periodicity of 12, 100 +/- 100 s seen in the EXOSAT data. Over 75% of the total source variability is due to the periodic component (rms amplitude of 36%). The residual variability can be described as the more usual "flicker noise" f^-1. 1^ power law. This continues unbroken to time scales shorter than 300 s, so that the region from which the variability arises must he smaller than 10¹3^ cm. Folding techniques, combined with the simulations to obtain the error, give the period as 12, 132 +/- 3 s. While the period is consistent with that seen in the EXOSAT data, the shape of the folded light curve is very different. There are three main peaks, two of which are separated by an extremely sharp dip, instead of the single peak and small harmonic structure observed by EXOSAT. Using the dip as a fiducial mark, a second Ginga observation of the source taken 1 year later is found to be consistent with being completely periodic and phase-coherent with the first Ginga observation. Thus the period is consistent with being constant over a time span of 6 years, but phase coherence is only maintained on time scales of <= 1 year. Subtle differences in the shape of the folded light curve with energy, and the very large amount of power in the periodic component, suggest occultation as its origin, although amplification of variability from an X-ray emitting "hot spot" through gravitational lensing is also possible. The former suffers from the nature of the periodic time scale, while the latter is unattractive because it is difficult to trivially produce the different shape of the folded light curve seen in the Ginga and EXOSAT data. That both these models probably fail to provide an adequate explanation may he due to the added complexity of anisotropic X-ray emission, suggested by the discrepancy between the lack of soft photons implied by the flat spectrum and the copious source of soft photons available from reprocessing in the iron-line-producing material.