With the estimated PSDs from both Swift/BAT and NuSTAR for the first time we will have PSDs covering nearly 5 orders of magnitude in the frequency domain. From the PSDs we want to measure the shape and determine where the break occurs. To do this, as in \citet{Shimizu_2013}, we will use the Monte-Carlo PSRESP method from \citet{Uttley_2002}. This involves generating many realizations of a PSD from a grid of model parameters. The best fit model parameters are then chosen as the set that essentially minimizes \(\chi^2\).
Each AGN will be fit using two different models, an unbroken power law and a singly broken power law given by the following equations. \[\label{eq:unbroken} {\rm Unbroken:}\,P_{\nu} = A\left(\frac{\nu}{10^{-6}}\right)^{-\alpha}\] \[\label{eq:singly_broken} {\rm Singly\,Broken:}\,P(\nu) = \left\{ \begin{array}{rl} A\left(\frac{\nu}{\nu_b}\right)^{-\alpha} & \nu \leq \nu_b \\ A\left(\frac{\nu}{\nu_b}\right)^{-\beta} & \nu \geq \nu_b. \end{array} \right.\] \(\alpha\) and \(\beta\) are spectral indices, \(A\) is just a normalization factor, and \(\nu_b\) is the break frequency in the singly broken model. The goodness of fit from each model for each AGN will be compared to determine which model is correct. For all of the AGN where the singly broken model is the best fit we will have, for the first time, characteristic timescales in the 14–79 keV energy range.
To test the viability of our method, we have downloaded the NuSTAR observation for 3C 273 as well as the 70 month Swift/BAT light curve. The NuSTAR observation was reduced using nupipeline and nuproducts contained in HEASOFT. After binning the Swift/BAT and NuSTAR light curves to 5 day and 500 s bins respectively, the PSDs for each were measured using the innovative new technique presented in \citet{Zoghbi_2013}. Figure \ref{fig:3C273}, shows the PSD power for 3C 273 derived from the Swift/BAT and NuSTAR light curves.