4. A Little Background on X-ray Astronomy and XMM-Newton Data

Some familiarity with the basics of X-ray astronomy in general and XMM-Newton in particular are not required when preparing and analyzing the data, but users may find a little context to be insightful. Here, we present here a short introduction/refresher to two things that have a strong bearing on the SAS tasks that we'll run and the decisions we'll make when processing our data. People who are very new to this topic may wish to read “The Absolute Beginners Guide to XMM” and references therein first.

4.1 Events and Event Patterns

The CCDs that detect X-rays are photon-counters. When an X-ray photon or cosmic ray hits a CCD (i.e., there is an “event”), the instrument records 1) the incident energy or Pulse Height Amplitude (PHA; more on this below), 2) the time, 3) the position on the CCD, and 4) the distribution of pixels over which a charge cloud spreads, which is called a “pattern”. An event can be registered in one pixel or several, and the probability of an X-ray generating a certain pattern is dependent on its energy. (It is similar “event grade” for ASCA and Chandra data.) The patterns themselves are named for the number of pixels that are involved in a detection, so singles have one pixel above a certain threshold value, doubles have two pixels above the threshold, triples have three, and quadruples have four. Examples of these are shown in Figure 4.1, where the small numbers next to the diagrams refer to the pattern number. It can be seen that a single pattern event has pattern = 0, double pattern events have pattern numbers that range from 1 to 4, triples have pattern numbers that range from 5 to 8, and quadruples have pattern numbers that range from 9 to 12. These are the only patterns that are recognized as valid events, and the data need to be filtered by pattern (among other things). Exactly which patterns are kept or discarded is determined by the instrument, its mode, and the science that the user wants to investigate. However, as a general rule, the highest quality event detections are made with lowest pattern values.

Some older publications that use X-ray data may refer to the PHA; it was common in older missions to use this to make spectra. However, this is no longer recommended. Instead, a related quantity - the Pulse Invariant (PI) - is used. The PI is the gain-corrected PHA, and the gain varies across the detector. Information about the gain mapping is contained in spectrum response files, in addition to other important information. Instructions on how to make these are included in §7.11, §8.8, §9.6, and §10.5.

Figure 4.1: EPIC camera patterns from Figure 7 of Turner et al. 2001, A&A, 365, L27. Events that have pattern values between 0 and 12, seen on the left side, are considered valid. In each, the red pixel is above threshold and has the largest signal, the green pixels have signals above threshold, the white pixels have signal below the threshold, and the crosses indicate pixels that are ignored.

4.2 Soft Proton Contamination

XMM-Newton has very high throughput, and an orbit that is highly eccentric, with much of its operational time spent outside Earth's protective magnetic field. The observatory is therefore susceptible to soft proton flares during observations. As suggested by the name, these protons have low energies ($<$ 300 keV). They show variability that has been linked to the solar wind, and can be found in the outer magnetosphere and interplanetary space. They can interact with the telescope and detectors, sometimes getting scattered into the focal plane of the X-ray detectors. The flares are unpredictable, sudden, and can be quite large - hundreds of counts/second across the bandpass - and can last for anywhere between seconds to hours. Progress has been made on understanding them and accounting for their effects (e.g. Fioretti et al. 2024, A&A, 691, A229 and references therein) but at present, the recommended way to deal with them is to remove the parts of observations that show contamination. However, it should be noted that the amount of flaring that needs to be removed depends in part on the object that is observed; a faint, extended source will be more affected than a very bright point source.

4.3 The General Approach

As a result of the above, a general procedure is followed for reprocessing XMM-Newton X-ray data (i.e., data from either the EPIC or RGS instruments):

1)	Apply the latest calibrations by rerunning the pipeline.
2)	Apply standard filters, which incude instrument- and mode-dependent pattern filters.
3)	Make a light curve to look for soft proton contamination and remove it if needed.

After these steps have been taken, the data are ready for more advanced procedures, such as extracting spectra, performing source detection, or making radial profiles.