1. Metal abundance evolution in distant galaxy clusters observed by XMM-Newton Alessandro Baldi Astronomy Dept. - University of Bologna INAF - OABO In collaboration with: S. Ettori (INAF-OABO), I. Balestra (MPE-Garghing), P. Tozzi (INAF-OATS), S. Molendi (INAF-IASF Milano), F. Gastaldello (INAF-IASF Milano)

2. Measures of metal content at high z Balestra et al. (2007) obtained single emission-weighted estimates of 56 clusters (at 0.3 < z < 1.3) from Chandra and XMM-Newton Measuring Fe abundance within (0.15-0.3) R vir they found a negative evolution of Z(Fe) with z : Z(Fe) ≈ 0.4 Z  at 0.3 ≤ z ≤ 0.5 Z(Fe) ≈ 0.25 Z  at z ≥ 0.5 This result has been confirmed by Maughan et al. (2008) on a sample of 116 Chandra clusters at 0.1 < z < 1.3, where Z drop by 50% between z=0.1 and z≈1 This evolution is not simply driven by the appearance or disappearance of the cool cores In the XMM-Newton sample by Anderson et al. (2009, 29 clusters at 0.3 < z < 1.3) the metal abundance evolution is not as evident Maughan et al. (2008)

3. XMM-Newton high redshift cluster sample We selected a sample of 39 galaxy clusters at 0.4 < z < 1.4 from the XMM-Newton archive, with sufficient S/N to perform a spatially resolved spectral analysis (2-3 bins). Taking advantage of EPIC XMM-Newton high throughput and effective area, we performed a spatially resolved spectral analysis of the clusters in the sample, not only excising the cool cores (as in Maughan et al.) but also analyzing the emission in the outer regions of the clusters. The aim of this work is to determine if the decrease of Z with redshift observed by Balestra et al. & Maughan et al. is due entirely to physical processes associated with the production and release of Fe into the ICM, or partially associated with a redistribution of metals connected to the evolution of cool cores.

4. XMM-Newton data reduction and analysis Observation Data Files (ODF) processed to produce calibrated event files using the XMM-Newton SAS v10.0.0 Intervals of very high background removed using a double filtering: a threshold of 0.20 cts/s is applied to a light curve in 100s bins extracted in the 10-12 keV energy band. A further 3  clipping algorithm is then applied to a 2-5 keV light curve extracted from the residual events MOS1 & MOS2 background treated using spectral modeling instead of direct spectral subtraction, following the recipe of Leccardi & Molendi (2008). We used only MOS observations in the analysis because of problems in the pn background modeling and inconsistencies in the kT and Z measures between MOS and pn.

5. MOS background modeling We used the method developed by Leccardi & Molendi (2008) on their low redshift sample of XMM-Newton clusters. They analyzed a large compilation of “blank field” MOS1 and MOS2 observations, to characterize the following components of the background: X-ray background from the Galaxy Halo (HALO) Cosmic X-ray background (CXB) Quiescent soft protons (QSP) Cosmic ray induced continuum (NXB) Fluorescence emission lines The background parameters were estimated fitting the background model in a 10’-12’ ring and then rescaled appropriately for the cluster spectra, apart from the fluorescence emission lines left free to vary in normalization.

6. Spectral analysis strategy We determined r 500 iteratively following the formula derived by Vikhlinin (2006): Spectra in the following spatial bins were extracted and fitted with XSPEC v12.5.1 using Cash statistics: 0-0.15 r500 0.15-0.4 r500 >0.4 r500 We used a 1-T thermal mekal model where kT, Z and normalizations were left free to vary, fixing Galactic absorption and redshift.

7. Spectral analysis results Each spatial bin does not shows a clear evolution of the metal abundance with the redshift. This is evident, fitting the z vs. Z distribution with a power-law in the form Z  z -  :  =+0.1  0.3 for r<0.15r 500  =-0.2  0.3 for r=0.15-0.4r 500  =-0.6  0.8 for r>0.4 r 500 Baldi et al. 2011 in prep.

8. Spectral analysis results We averaged the abundance in 3 different redshift bins: 0.4 < z < 0.5 0.5 < z < 0.7 0.7 < z < 1.4 Abundance evolution is at less than 1  in all three spatial bins, agreeing with the power-law fits However, the evolution in the 0.15-0.4r 500 bin is agreement with theoretical predictions (Ettori 2005) and with Maughan et al. (2008) Baldi et al. 2011 in prep. Ettori (2005)

9. Abundance evolution studies & the next generation observatories The next generation of X-ray observatories (especially Athena) could improve dramatically our knowledge of abundance evolution in galaxy clusters at high redshift To give an idea of how these future missions could shed new light into the ICM enrichment in the early stages of the Universe, we performed 50ksec Athena - XMS spectral simulation for all the galaxy clusters in our sample with XSPEC Each simulated cluster spectrum (with the same spatial bins as in XMM analysis) was fitted with an XSPEC mekal model

10. Individual elemental abundances The high S/N XMS spectra would allow to investigate the evolution in abundance of the individual elements out to z≈1 (at least in a single spatial bin) OOMgMg SiSiSS

11. Abundance ratios and SN yields The abundance ratios between  elements and Fe would allow the comparison with the metal abundance yields expected from different SN types and therefore to study the history of ICM enrichment through SNIa and SNII. [Mg/Fe] [Si/Fe] [S/Fe]

12. Summary We presented a sample of 39 galaxy clusters at 0.4<z<1.4 extracted from the XMM-Newton archive. A spatially resolved spectral analysis of the clusters in the sample revealed no clear evidence of an evolution in abundance with z at every radius, not confirming the results of Balestra et al. (2007) and Maughan et al. (2008) obtained from Chandra data. Although several factors could explain this discrepancy (e.g. different redshift range, different averaging methods, etc.), we cannot exclude the presence of abundance evolution because of low statistics (leading to large errors on Z) for most of the clusters in the sample Athena XMS spectral simulation of the clusters in the sample showed how the high count statistics expected could confirm (or deny) with a higher degree of confidence the presence of an evolution in abundance. Abundance of individual elements could also be measured with small statistical errors down to z≈1, allowing to trace the ICM enrichment history through SNIa and SNII.