XMM-ESAS is intended to produce both images and spectra of the background for regions that cover a significant fraction of the FOV[*]. The production of images depends critically on the production of spectra, and the order of analysis can sometimes become confusing. This overview is intended to blunt some of that confusion.

Imaging: The purpose of XMM-ESAS, for imaging, is to produce images of all of the known backgrounds and foregrounds so that one can isolate the true cosmic X-ray emission. The background/foreground components that XMM-ESAS handles are:

In each case one must know that distribution of that emission across the FOV and the absolute strength of that emission.

For the QPB, the distribution across the FOV is given by the filter-wheel closed (FWC) data, while the normalization is provided by the “corner” data: the parts of the detectors that are shielded from the X-rays focussed by the mirrors[*] as can be seen in Figure 1.

For the SPF, the distribution across the FOV was laboriously determined by Kuntz & Snowden (2008)[*] and the normalization must be obtained by fitting the spectrum from the entire FOV. Thus, imaging relies heavily on spectroscopic analysis. Although the bulk of the SPF emission can be removed by excising the time periods containing the flares, there is nearly always some residual emission from the remainder of the observation. The amount of that residual can be determined from the spectrum of the entire FOV as the SPF spectrum is (usually) distinctly different from that of any other likely emission component.

For the SWCX, the distribution across the FOV is given by the standard exposure map and the normalization must be obtained by fitting the spectrum of the entire FOV. Thus, one can see once again the need to do spectroscopic analysis in order to obtain a background/foreground subtracted image.

Spectroscopy: The purpose of XMM-ESAS, for spectroscopy, is to produce the spectrum of the quiescent particle background for the observation. All other spectral components must be characterized by fitting. In order to get the best spectrum of the diffuse emission, however, one will want to remove the point sources, which will require at least some imaging.

Order of Operations: Whether one is interested in imaging or spectroscopy, the initial steps are the same. We assume the existence of freshly recalibrated event files. We first determine which chips (if any) are in anomalous states and need to be removed. We then determine which time segments should be removed due to soft proton flares. We then identify point sources, build region files describing the location of the point sources, and build masks to exclude those regions. Those masks are required both to exclude the sources in making images, as well as to exclude them from the spectrum of the diffuse emission.

The next step can be rather time consuming and requires some skill. We extract the spectrum from the region of interest excluding, of course, the point sources. We build a QPB spectrum. Then, if we are only interested in the spectrum, the analysis may begin.

If we are interested in imaging, we fit the spectrum from the FOV (or the region of interest) in order to determine the strength of the remaining soft proton flare contamination; that fitting need not be done precisely, nor in a completely astrophysically motivated manner. Similarly, the spectral fitting may be required in order to determine the amount of solar wind charge-exchange (SWCX) emission. Then, we build the QPB image, the soft proton image, and the SWCX image so that we may subtract them from the count image and divide by the exposure image to get a properly calibrated image of the flux from the source of interest.