Cross-calibration of the XMM-Newton EPIC pn and MOS on-axis effective areas using 2XMM sources

Aims. We aim to examine the relative cross-calibration accuracy of the on-axis effective areas of the XMM-Newton EPIC pn and MOS instruments. Methods. Spectra from a sample of 46 bright, high-count, non-piled-up, isolated on-axis point sources are stacked together, and model residuals are examined to characterize the EPIC MOS-to-pn inter-calibration. Results. The MOS1-to-pn and MOS2-to-pn results are broadly very similar. The cameras show the closest agreement below 1 keV, with MOS excesses over pn of 0-2% (MOS1/pn) and 0-3% (MOS2/pn). Above 3 keV, the MOS/pn ratio is consistent with energy-independent (or only mildly increasing) excesses of 7-8% (MOS1/pn) and 5-8% (MOS2/pn). In addition, between 1-2 keV there is a silicon bump - an enhancement at a level of 2-4% (MOS1/pn) and 3-5% (MOS2/pn). Tests suggest that the methods employed here are stable and robust. Conclusions. The results presented here provide the most accurate cross-calibration of the effective areas of the XMM-Newton EPIC pn and MOS instruments to date. They suggest areas of further research where causes of the MOS-to-pn differences might be found, and allow the potential for corrections and possible rectification of the EPIC cameras to be made in the future.