Systematic errors on black hole spin

Determining black hole spin from broad iron lines is potentially the best approach in active galactic nuclei. However this measurement is rather subtle and although a number of spin values have been published in recent years, the affect of systematic errors remains largely unquantified. We explore systematic effects on the measured spin parameter, illustrating these effects using all available Suzaku and XMM-Newton data on the Seyfert 1 galaxy Fairall 9. We explore effects resulting from different data analysis techniques, and imperfect spectral models. We find that systematic errors dominate over statistical errors and uncertainties in the underlying accretion disk Physics. We identify a particularly important degeneracy that exists between narrow emission features in the iron K band and the blue wing of the broad line - CCD spectra do not have sufficient spectral resolution to separate these two features. Moreover, we find that the spin parameter can be sensitive to modeling of the soft part of the spectrum, i.e. the modeling of the soft excess. We investigate the extent to which time-variability of the spectrum (i.e. multi-epoch fitting) can be used to resolve these issues. Including time variability studies, we can conclude that the soft excess in Fairall 9 does not originate from only blurred reflection. Instead an additional component is required to permit consistent modeling of the entire spectrum over time.