XMM-Newton Users Handbook

3.3.4 Intrinsic energy resolution of EPIC

The CCD pixels of the EPIC cameras are energy sensitive, enabling non-dispersive spectroscopy. The resolving power of these cameras is determined by the intrinsic energy resolution of the individual pixels. The spectral resolution for a point source located at the nominal pointing position of both the EPIC MOS and pn CCDs, as a function of energy, is displayed in Figs. 25 and 26. Also plotted in Fig. 25 are the measured in-flight $FWHM$ of the Al K$\alpha$ and Mn K$\alpha$ on-board calibration lines, after correction for charge transfer inefficiency (CTI) by the SAS, obtained in five different orbits. This shows a degradation of approximately 13% in the energy resolution of the MOS since the launch of XMM-Newton. The degradation is due to an increase in the CTI of the CCDs with time in orbit. Work is currently being carried out to improve the CTI correction technique and hence the energy resolution.

For the pn camera, no significant degradation of the $FWHM$ has been found.

Figure 25: Temporal evolution of the EPIC MOS energy resolution (FWHM) as a function of energy. The solid curve is a best fit $E^{0.5}$ function to ground calibration data between 0.1 and 12 keV; all events with pattern 0-12 were included in the analysis. Below around 0.6 keV (shown by the dotted region) surface charge loss effects distort the main photopeak signficantly from a Gaussian and hence the effective energy resolution. The data points represent MOS 1 in-flight measurements of the FWHM of the Al $K_\alpha$ (1.478 keV) and Mn $K_\alpha$ (5.893 keV) lines in five different epochs. It should be noted the rapid degradation of the resolution between the first and the second epoch, and the recovery and subsequent stability after the cooling of the MOS camera (performed between November and December 2002, i.e. between Rev.#530 and Rev.#560. In the main panel measurement error bars are smaller than the symbol size. In the insets a zoom of the spectral ranges around the nominal line positions is shown. Typical standard deviations in each epoch range between 3 and 7 eV, and 8 and 18 eV for the Al and Mn line, respectively.

Figure 26: Left panel: MOS 1 energy resolution as a function of energy for singles (blue) and singles+doubles (black) events. Right panel: pn energy resolution as a function of energy for: a) single events at the boresight (position "Y9" in the canned response matrices; black); b) double events at the boresight position (red); c) single events closest to the readout node (position "Y0", blue); d) double events closet to the readout node (green). Line widths are based on monocromatic line spectra, simulated with the SASv7.1 canned response matrices.