Searching For Breaks In The Hard X-ray Power Spectral Density Functions Of AGN



The physics of accretion disks in active galactic nuclei (AGN) have been the subject of intensive recent work. While their huge energy output over a broad range of wavelengths (i.e. radio to X-ray) allows for detailed spectral characterization, their extremely small sizes are not resolved by even the best observatories. The Event Horizon Telescope is the current best hope for directly imaging the accretion disk of black holes however it will only be able to observe the two nearest black holes Sgr A\(^{*}\) and M87. A large resolved survey of extragalactic AGN is not possible currently or even in the near future. Time variability analysis in different spectral bands is one of the best ways to study the structure of accretion disks as well as the mechanisms controlling the mass accretion rate crucial to understanding the physics of AGN.

We propose to combine short timescale NuSTAR observations with the latest Swift/Burst Alert Telescope (BAT) long timescale light curves (Shimizu et al., 2013) to study the time variability of a sample of AGN designed to span a wide range of black hole mass, AGN luminosity and Eddington ratio in the previously unexplored hard X-ray band. We will use a new maximum likelihood technique (Zoghbi et al., 2013) to construct the power spectral density functions (PSD) over a large range of timescales (years to hundreds of seconds) as well as measure the important break frequency in the PSD for the first time at high energies. The PSDs will allow us to test relationships between the observed properties of the AGN (e.g. \(M_{BH}\) and \(L/L_{Edd}\)) and time variability (Kelly et al., 2013) and to provide important constraints on models of accretion.

Background and Motivation

Accretion onto super-massive black holes (SMBH) is a complex process that produces an extremely large amount of radiation spanning nearly the entire electromagnetic spectrum. As material spirals into the SMBH, the loss of gravitational potential energy is converted into thermal energy and radiated away producing the characteristic optical-UV AGN SED (i.e. “Big Blue Bump”). Due to the high mass of the SMBH though, the accretion disk itself does not heat to high enough temperatures to produce X-ray emission. Rather, theoretical models must invoke a hot corona to explain the large amounts of X-ray emission seen in AGN. In particular X-rays are thought to be produced through thermal Comptonization of seed photons from the accretion disk by the hot plasma sitting above the disk (Haardt et al., 1993). Further a so-called “Compton hump” that peaks around 30 keV can be produced through reflection of the original X-rays by the relatively cold accretion disk (George et al., 1991; Nandra et al., 1994; Magdziarz et al., 1995; Rivers et al., 2011).

Apart from their spectral characteristics, AGN also exhibit large amounts of variability at all wavelengths. This variability is highly aperiodic over a broad range of timescales from years down to hours. The variability power as measured by the power spectral density function (PSD) is seen to decrease as a function of increasing temporal frequency with the shape normally consistent with a power law, \(P\propto\nu^{-\alpha}\) (Lawrence et al., 1987; Lawrence et al., 1993; Green et al., 1993). The value of \(\alpha\) depends on the timescale with longer timescales displaying \(\alpha\sim1\) and shorter timescales displaying \(\alpha\sim2\) (Edelson 1999, Uttley 2002).

Variability provides a window into the small scale physics and structure of the emitting components near the SMBH that current and future telescopes can only dream to image. By monitoring and analyzing the variability of AGN, we can probe deep into the heart of the strongest gravitational potentials in the universe. Ultra-hard X-rays (\(\gtrsim\)15 keV) in particular are thought to originate closest to the SMBH given their strong short-timescale variability and high energy. Due to the high energy, ultra-hard X-rays are also unaffected by absorption from intervening material as long as the hydrogen column density (\(N_{\rm H}\)) is less than \(\sim10^{24}\) cm\(^{-2}\) (i.e. “Compton-thin”).

Given the lack of sensitive monitoring at ultra-hard X-rays in the past, most X-ray variability studies have focused on the 2–10 keV energy range using the Rossi X-ray Timing Explorer (RXTE) and XMM-Newton. Through dedicated monitoring campaigns spanning years to hours significant progress has been made in understanding the variability properties of AGN. Specifically, breaks in the PSD of AGN were measured for a moderately large sample spanning a range of black hole masses (\(M_{\rm BH}\)) and accretion rates (e.g Uttley et al., 2002; Uttley et al., 2005; McHardy et al., 2004; McHardy et al., 2005; Markowitz et al., 2003). A relationship was found between the break timescale, \(M_{\rm BH}\), and \(L_{\rm Bol}\) (bolometric luminosity) that spanned not only a large range of AGN properties but also scaled all the way down to galactic black holes (McHardy et al., 2006); a strong indication that the same accretion process powers the emission in both sets of objects.

With the launch of Swift/BAT, long term monitoring of AGN at ultra-hard X-rays became automatic. Swift/BAT continuously searches the sky for gamma-ray bursts in the 14–195 keV energy range. Due to its wide field of view (\(\sim60\times100\,\rm{deg^{2}}\)), Swift/BAT covers roughly 75–85% of the sky every day. This has allowed for nearly uniform coverage of the entire ultra-hard X-ray sky and the creation of the largest ultra-hard X-ray source catalog (Baumgartner 2013) containing over 1000 sources, with more than 700 identified as AGN.

In Shimizu et al. (2013), we took advantage of the \(>5\) years of continuous monitoring of the ultra-hard X-ray sky to analyze the long-term light curves for 30 of the brightest AGN in the Swift/BAT sample by constructing the PSD. We attempted to measure for the first time the break in the PSD at these high energies to determine if the same McHardy et al. (2006) relationship exists. Further any evolution of the break timescale with energy would give insight into the geometry of the corona assuming the timescale is related to a radial distance away from the SMBH. Unfortunately, due to the relatively low sensitivity of Swift/BAT, the time resolution of our light curves was limited to \(\sim5\) days. The PSD was therefore limited to timescales \(>10\) days given Nyquist sampling. We were successful in measuring the slopes of the low frequency PSD finding nearly all of the AGN were consistent with a slope of -1, very similar to the PSD in the 2–10 keV range.

NuSTAR, launched in 2012, however provides a window to the shorter timescales necessary to measure the PSD breaks. Its high sensitivity (1000 times more sensitive than Swift/BAT) and remarkable timing capabilities allow one to create high quality light curves that can cover timescales from several hundred to several hundred thousand seconds. NuSTAR also has overlapping energy coverage with Swift/BAT between 14–75 keV. The PSD constructed from the NuSTAR light curves therefore can be combined with the Swift/BAT light curves and extend the PSD over nearly six orders of magnitude in temporal frequency.

In this proposal we will outline our plan for selecting Swift/BAT AGN from the NuSTAR archive, our method for constructing and analyzing the PSD using a new maximum likelihood technique, and the expected scientific goals and implications.