2.1. The Extreme Ultraviolet Explorer (EUVE) satellite
The EUVE satellite was launched in 1992 and operated till 2001, covering the 50-250 eV energy range (Bowyer & Malina 1991). The rectangular shape of the Lex/B (65-248 eV) filter resulted in images where the length of one side of the image greatly exceeds the other side: ~ 40' in width and more than 2° in length (see e.g. Fig. 1 in Durret et al. 2002).
The spatial scale was 13 pixels/arcmin. Due to the limited sensitivity of EUVE, exposure times on clusters were typically between several tens of ks and 1 Ms.
The Position Sensitive Proportional Counter PSPC instrument on board the ROSAT satellite first detected the soft excess component of clusters of galaxies at X-ray wavelengths. The PSPC had an effective area of about 200 cm2 at 0.28 keV, enabling soft X-ray studies. It had a large field of view (hereafter FOV) of 50' radius, covering the virial radius in most clusters and enabling in most cases the estimation of the local background. The low and stable internal background of the PSPC enabled reliable X-ray measurements at large radii where the background is important. A major problem for soft excess studies with the PSPC is that it did not cover energies above 2 keV. Thus, it could not be used to determine reliably the hot gas properties, which had to be measured elsewhere. A further limitation was the low energy resolution which did not allow detection of possible emission line blends emanating from the soft component. The angular resolution of the ROSAT PSPC was ~ 15". Details on the ROSAT PSPC instrument can be found in Briel et al. (1996).
An important change in the study of the soft excess came with the XMM-Newton satellite. The XMM-Newton European Photon Imaging Camera (EPIC) instruments PN and MOS extend the energy band coverage to 10 keV, thus enabling simultaneous determination of the hot gas and soft excess component properties.
The large collecting area of the EPIC telescopes (~ 1000 cm2 at 0.5 keV for the PN) provides the high statistical quality data necessary to examine the few 10% soft excess effect on top of the hot gas emission. The spectral resolution of the PN at 0.5 keV is 60 eV (FWHM), rendering it possible to resolve the emission lines emanating from the soft excess component. However, the relatively small FOV (15' radius) of EPIC prevents the study of the cluster outskirts for the nearest clusters with single pointings. The usage of offset pointings introduce the complex stray light problem which complicates the analysis of weak signals such as the soft excess. The more distant clusters, which would be covered out to the virial radius with a single pointing, are fainter, which reduces the quality of the signal. Thus, the XMM-Newton soft excess analysis is mostly limited to the central regions of nearby clusters. Details on the characteristics of XMM-Newton can be found in Turner et al. (2001) and Strüder et al. (2001), or in Ehle et al. (2006).
A further complication is the strong and flaring particle-induced detector background. A local background estimate is vital when analysing weak signals such as the cluster soft excess. This is usually not available for nearby clusters since they fill the FOV, and one has to resort to blank-sky based background estimates. This introduces uncertainties in the analysis, and further limits the cluster analysis to central regions where the background is not important. Together with the FOV limitations, the background problem limits the usefulness of XMM-Newton for measuring the soft excess in a large cluster sample.
The on-going calibration work on the PN and MOS instruments resulted in changes in the derived soft excess properties for a few clusters (Nevalainen et al. 2007). Thus, the EPIC results have some degree of systematic uncertainty based on calibration inaccuracies, and definitive results on the soft excess properties are not yet available (also see Sect. 4.1.2).
Chandra, launched in the same year as XMM-Newton (1999), has a similar eccentric orbit as the latter satellite and therefore suffers from comparable enhanced background problems. Its angular resolution is much higher (0.8" Half Energy Width) than that of XMM-Newton (14"). The advantage is that for extended sources like clusters of galaxies subtraction of contaminating background point sources can be done more accurately, and more importantly, blurring effects by the point spread function of the telescope can usually be ignored. These blurring effects were a serious problem for the BeppoSAX LECS data (Sect. 3.2.2). However, the effective area at low energies (E < 0.5 keV) of Chandra is an order of magnitude smaller than that of XMM-Newton, and time-dependent contamination that affects in particular the lower energies is a complicating factor in the analysis of Chandra data.
Suzaku is the fifth Japanese X-ray astronomy satellite, launched in July 2005 (Mitsuda et al. 2007). Unfortunately, the main observing instrument XRS, consisting of X-ray microcalorimeters used in space for the first time, did not last until the first space observations, due to the loss of liquid He (Kelley et al. 2007). On the other hand, the XIS instrument, which employs X-ray CCDs with improved performance for X-ray spectroscopy, is functioning well (Koyama et al. 2007). Regarding soft X-ray spectral studies of diffuse sources, XIS offers the best capability so far achieved with X-ray satellites.
The XIS system consists of 4 units of mirror and detector combinations. The X-ray mirrors are identical and have a focal length of 4.5 m and an effective area of about 500 cm2 at 2 keV (Serlemitsos et al. 2007). The angular resolution is about 2', limited by the light-weight design of the thin foil mirror.
The 4 CCDs in the focal plane operate jointly during the observations. The chips are square with an area of 10 mm × 10 mm and the number of pixels is 106. The field of view is 17' × 17'. One chip is back-illuminated, which gives a superior soft X-ray sensitivity with somewhat poorer energy resolution and higher background above 7 keV. The other 3 chips are of standard front illumination type. The typical energy resolution is 150 eV FWHM at 5.9 keV at the time of launch. The resolution degraded significantly with time (200 eV after 1 year), and the XIS team performed a charge injection operation after October 2006 to maintain the resolution around 170 eV. In November 2006, one of the 3 FI (front illumination) chips developed excess noise, and it has been switched off since then. This leaves a total of 3 chips, one BI (back illumination) and 2 FIs, in operation. Its background is lower than that of XMM-Newton.