XMM-Newton Users Handbook

4.6.2 RGS Multipointing Mode (MPM)

A small proportion of defects comprising bad or “cool” columns or pixels is inevitable in CCD detectors. In a dispersive instrument like the RGS, these detector defects, usually one pixel wide, correspond to fixed narrow wavelength intervals of about 10 mÅ per pixel (the RGS line-spread function has a FWHM of about 7 pixels). These wavelength defects sometimes coincide with features in the source spectrum that are important for the scientific success of an observation (see Tables 13 and 14).

One consequence of the pointing stability of XMM-Newton is that point-source photons of a given energy always end up in the same pixel of the detector. If an important spectral feature happens to fall on a defect, up to 20-50% of line flux may be lost depending on the filtering scheme employed. In addition, small defects can be difficult to distinguish from true, weak absorption lines. Further, there are the larger 70-100 mÅ gaps between the RGS CCDs.

When these circumstances arise, it is possible to change the spacecraft pointing in order to move the defects away so that more important parts of the spectrum are not affected.

The RGS Multipointing mode (MPM) allows the user to make five different pointings, with offsets in the dispersion direction of 0, $\pm$15 and $\pm$30 arcsec, corresponding to 0, $\pm$35 and $\pm$69 mÅ. In this way, the maximum flux loss in a spectral line is reduced significantly, and it becomes easier to distinguish weak absorption lines from defects. The dispersion step size is limited by the need to keep X-ray sources safely within the EPIC small windows.

For RGS, it is not necessary to start a new observation when slewing to the offset position. The instrument can be left on and useful observation time gained. On the other hand, EPIC pn offset maps must be generated at each pointing leading to a significant loss of EPIC pn exposure time.

A number of considerations must be taken into account in order to decide if the MPM could be applied with profit for a given observation.

• The mode is only relevant if the RGS is the prime instrument.

• The mode is in particular relevant if wavelengths important for the scientific success of the observation are affected by chip defects. This is also important if line positions are unknown in advance, for instance with unknown redshifts of foreground absorbers. Also, the continuum improves, as any small remaining detector defects will be smoothed.

• The mode is designed for high-quality RGS spectra, which is simply a requirement on the number of counts expected. The signal-to-noise ratio should be at least 10 in significant RGS resolution elements of $\approx$ 70 mÅ, equivalent to 100 counts.

• The mode would not benefit sources larger than a few arcmin, such as supernova remnants or the cooling cores of clusters of galaxies, because smoothing of the spectrum would automatically be achieved through the spectral degradation caused by the spatial extent of the source. Only for extended sources with a diameter of 30 arcsec or less is the mode of potential interest.

• For bright sources and relatively small exposure times, a significant fraction of the EPIC pn exposure time may be lost due to the overhead associated with the acquisition of offset tables at every position.

• When time variability of spectral lines is the main scientific driver, the user should carefully consider possible changes of sensitivity during the observation due to the different location of detector defects, versus the time-averaged spectral quality improvement due to the MPM. This holds in particular when the equivalent width or shape of the line vary. For lines with constant equivalent width (e.g. interstellar or intergalactic absorption lines) the MPM can be used without problems.