Volume 573, January 2015
|Number of page(s)||21|
|Section||Cosmology (including clusters of galaxies)|
|Published online||08 January 2015|
Left: comparison between the temperatures determined using the old version 1.3.1 and the new version 2.0.1 of AtomDB. Different colors represent groups with an abundance higher (red) and lower (blue) than 0.6 solar. Here, we plotted only the innermost temperature bin, and excluded the groups for which we determined only a global temperature: IC 1262, NGC 6338, RXC J1840. Right: the same as in the left panel, but comparing the abundances instead of the temperatures. Different colors represent groups with a temperature higher (red) and lower (blue) than 1.5 keV.
|Open with DEXTER|
Left: temperature profile of NGC 3402 derived by fitting the spectra with different plasma models and abundance tables. Right: normalization values obtained by fitting a spectrum with different plasma models and abundance tables.
|Open with DEXTER|
Version 2.0 of AtomDB is available since 2011. With respect to the older version 1.3, it includes significant improvements on the iron L-shell data. As we show below, this change strongly affects the temperature and abundance determination for low-mass systems. We used the temperature and the APEC normalization from the spectral fits to determine the gas and total masses. Thus, the use of different AtomDB versions has to be taken into account when comparing our results with the ones in literature. Here, we analyze the main effects.
To show how the temperature and abundance determination change, we fitted the innermost bin of the galaxy groups in our sample using the two AtomDB versions. The results are shown in Fig. A.1. While at temperatures higher than 1.5 keV the temperature difference is quite small, at very low temperatures (i.e., kT<1 keV) the temperatures obtained using the version 2.0 are up to 18% higher than the ones obtained using version 1.3. At the same time, the obtained abundance is 20–30% lower with a trend of a larger deviation for higher temperatures. More in detail, we note that when the group abundance is relatively low (A< 0.6 A⊙), the temperature deviation arises only for kT< 1 keV. In contrast, when the abundance is relatively high (A> 0.6 A⊙), small differences can be observed already at a temperature of ~2 keV.
To investigate how much this influences the total mass estimate we used NGC 3402 as a test case because of its low temperature and good quality of data. Since many authors use a MEKAL model instead of the APEC model, we also included this thermal plasma model in our analysis. We determined the temperature profile for the different models and for different abundance tables (i.e., from Anders & Grevesse 1989 and Asplund et al. 2009). As shown in Fig. A.2 (left panel), while the profile from MEKAL and the old AtomDB version agree quite well, the new AtomDB has a higher temperature at all radii. This translates into a total mass higher by ~10%. Although the effect is weak, a higher temperature is also obtained when the old abundance table from Anders & Grevesse (1989) instead of the most recent table from Asplund et al. (2009) is used.
In Fig. 7 we compared the gas mass at a given radius for different works and found that the difference for most of the objects is of ~10%. Since we used the APEC normalization to estimate the central electron densities to better understand whether the new AtomDB can explain part of the difference, we compared the normalization of the spectrum extracted from an annulus of ~7 arcmin (to maximize the S/N) and fitted with the different models. As shown in the right panel of Fig. A.2 (for display purposes we only show the MOS1 normalization, but the trend is similar for MOS2 and pn, although with different values), the normalization with the new AtomDB is ~10% lower than the older one with the MEKAL one lying in between. Depending on the combination of abundance tables and plasma models used in a particular paper, the difference can be up to ~15–20%. Since the central electron density scales with the square root of the normalization from XSPEC, using the new AtomDB can give a lower central density (and so a lower integrated gas mass) of up to 7–10%. The difference in gas mass for NGC 3402 is ~6% at R2500 and ~10% at R500, so the use of different AtomDB versions alone can explain the different gas mass shown in Fig. 7.
The use of the new AtomDB version results in a total mass higher by 10% than the mass derived using an older version. At the same time, the gas mass will be up to 10% lower than the value obtained using the old AtomDB versions. Given these results, the gas fraction for the less massive galaxy groups obtained with the most recent version of the AtomDB can be up to 20% lower than the mass derived with the old AtomDB version. Of course, this is an upper limit because we used NGC 3402 for the calculation, one of the coolest groups in our sample, which implies a larger difference between the different AtomDB versions, and not all the low temperature objects show such a large difference. Furthermore, in general, the temperature profiles obtained with the new AtomDB version cannot be simply scaled up because, as we showed, the difference in temperature depends both on the real temperature and on the associated metallicity. For example, in the outer regions where the temperature is lower, the metallicity is lower as well, which mitigates the real difference in the temperature estimation (see, e.g., Fig. A.1). This result highlights the importance of taking this problem into account for comparisons between different papers.
At first glance, A194 can be confused with a merging system because it shows three X-ray peaks: the main one in the NE, a second one in the SW, and a third in the center. Mahdavi et al. (2005) argued that the SW source is a background cluster of galaxies at z = 0.15. Sakelliou et al. (2008) confirmed that although it might be possible that the SW source is a background cluster, it is not possible with the XMM-Newton data to exclude that the source is at the same redshift as A194. We used an extraction area with a radius of 1 arcmin centered on the source and found that it is better fitted by a thermal plasma at redshift 0.15 than by a model with a redshift fixed at 0.018, in agreement with the finding of Mahdavi et al. (2005). In particular, we obtained a temperature of and metallicity of with χ2/ d.o.f. = 137/114 when z = 0.15 and a temperature of and metallicity with χ2/ d.o.f. = 173/114 if z = 0.018. Thus, since the peak is probably the BCG of a background cluster, we decided to exclude a region corresponding to R500 from the analysis of A194 to minimize the effect that it would have on the derived properties. By using the M-T relation derived only using the other objects in the sample, we then estimated for A194 a mass of M ~ 6 × 1013 and R500 ≈ 500 kpc which corresponds to about 3.5 arcmin at the A194 redshift. To be on the safe side, we excluded 4 arcmin around the SW peak. The flux in the 0.1–2.4 keV band from this 4 arcmin region is ~10-12 erg/s/cm2, so even excluding this region, the net flux of A194 is ~8.7 × 10-12 erg/s/cm2, well above the flux limit threshold.
By extracting a spectrum from a region with a radius of 15′′ around the NW source, we found that it is consistent with that of an AGN type 2 (an intrinsic absorption component was needed to fit the spectrum). The redshift of the source is 0.0182 consistent with the redshift of the cluster and a luminosity of 3 × 1041 erg/s, suggesting that it is accreting inefficiently. The estimated flux is 1.3 × 10-13 ergs/s/cm2.
Both regions were excluded from the analysis of the group properties.
A3390 shows two X-ray peaks that are centered on two bright galaxies at the cluster redshift. We extracted a spectrum from a region of 4′ around the two X-ray peaks to estimate the redshift of the two clumps with the X-ray data alone. We did not find any evidence that the two clumps have a different redshift. We estimated the temperature and surface brightness profiles of each component independently by excluding a region of 10 arcmin around the second subcluster.
IC 1633 appears as a relaxed group with no usual signs of a merger, such as a radio halo or a mismatch between the X-ray peak and the cD galaxy. Instead, from the exposure-corrected, background- and point-source-subtracted image we note that there is a strong elongation to the north of the emission peak (i.e., higher surface brightness). This feature together with the separation of more than 30 kpc between the EP and EWC suggests we are probably observing an unrelaxed system.
A3574 has two components separated by ~0.6 Mpc that are accepted as independent clumps (Böhringer et al. 2004). The main one is the eastern clump (A3574E), whose central galaxy is a Seyfert galaxy: IC 4329A. This galaxy carries 75% of the total flux of the clump (Böhringer et al. 2004), but the net flux of ~7.3 × 10-12 ergs/s/cm2 is still above the flux threshold of this paper.
This system is clearly in the process of merging, with a bright subclump just to the south of the main X-ray peak.The subclump corresponds to a small group of galaxies apparently falling into the gravitational potential of the main group.
NGC 4936 is the lowest redshift group analyzed in this sample. Its surface brightness profile has an unusual outer β value of 0.32 ± 0.01. The X-ray image shows that the X-ray peak, which is centered on the cD galaxy, is surrounded by a very faint extended emission.
Despite its overall regular and spherical X-ray emission, this galaxy group shows an anomalous temperature distribution with a central temperature peak surrounded by a relatively cool shell.
This remarkable feature has also been observed by different authors (e.g., Vikhlinin et al. 2006; O’Sullivan et al. 2007; Sun et al. 2009Eckmiller et al. 2011) and different instruments (e.g., XMM-Newton and Chandra). Combining XMM-Newton, Chandra, and VLA observations O’Sullivan et al. (2007) concluded that the most likely explanation for this feature is the interplay between the cool-core region and a previous period of AGN activity.
Surface brightness profiles for NGC 1132, RXCJ2315.7-0222, and RXCJ1840.6-7709.
|Open with DEXTER|
Fit parameters with a double β model.
© ESO, 2015
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.