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

2. Talk Outline
Introduction to galaxy clusters and metal abundance studies
Data reduction and analysis of the XMM-Newton cluster sample
Spectral fitting strategy & global X-ray properties of the sample
Results on abundance evolution & comparison with previous works
Further work in progress and a glimpse to the future…

3. Why clusters of galaxies?
Clusters of galaxies are the largest virialized structures in the Universe, arising from gravitational collapse of high peaks of primordial density perturbations.
They represents unique signposts in the Universe, where the physical properties of the cosmic diffuse baryons can be used to trace the past history of cosmic structure formation (e.g. Peebles 1993; Rosati et al. 2002; Voit 2005).
The hot thin gas permeating the cluster gravitational potential tipically reaches T≈ K and emits in the X-rays via thermal bremsstrahlung.
Jenkins et al. (1998)
A 1689
Peng et al. (2008)

4. Why clusters of galaxies?
X-ray studies of galaxy clusters provide:
an efficient way of mapping the overall structure and evolution of the Universe;
an invaluable means of understanding their internal structure and the overall history of cosmic baryons.
The hot gas is chemically enriched by SN explosions, whose signatures are visible in the X-ray spectra as emission lines (e.g. Werner+ 2006, 2007)
The distribution of metals in the ICM, as well as their evolution as a function of z are the ‘footprint’ of cosmic star formation history and crucial to trace the effect of SN feedback on the ICM in different cosmic epochs (e.g. Ettori 2005; Borgani+ 2008) 2A
Sersic
Werner+ 2006, 2007

5. Metal distribution in clusters: ROSAT, ASCA & BeppoSAX
Several studies have been presented in the literature on the radial distribution of Z in galaxy clusters
ROSAT & ASCA: Finoguenov, David & Ponman (2000) derived the radial distribution of Fe, Si, Ne and S on 11 relaxed CC clusters, where they found that:
the total Fe abundance decreases with radius
the Si, Ne and S abundances are either flat or decrease less rapidly
BeppoSAX: De Grandi & Molendi (2001) on 17 nearby clusters (9 CC & 8 NCC at z < 0.1) found:
a strong enhancement in the abundance in the central regions of the CC clusters
a flatter Z profile in the NCC clusters
the metallicity in CC clusters is higher than in NCC clusters at every R
similar results where also found by Irwin & Bregman (2001) on 12 clusters at 0.03 ≤ z ≤ 0.2 CC
De Grandi & Molendi (2001)
NCC

6. Metal distribution in clusters: XMM-Newton & Chandra
Leccardi & Molendi (2008), using XMM-Newton data on 50 clusters, found that <Z>≈0.45 Z in the center and decreases out to ≈0.2 r180; beyond 0.2r180 the profile is consistent with being flat at Z≈0.25 Z.
Chandra:
Vikhlinin et al. (2005) on a sample of 11 low redshift CC clusters observed a negative gradient in the distribution of Z, peaked toward the center.
Baldi et al. (2007) on a sample of 7 CC and 5 NCC clusters observed a negative gradient in both CC and NCC clusters, steeper in CC and flatter in NCC.
Leccardi & Molendi 2008

7. 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 ( ) Rvir 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
Maughan et al. (2008)

8. 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.
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.

9. XMM-Newton high redshift cluster sample
Baldi et al (A&A submitted)

10. XMM-Newton data reduction and analysis
Observation Data Files (ODF) processed to produce calibrated event files using the XMM-Newton SAS v9.0.0
Intervals of very high background removed using a double subtraction method (fixed ct-rate threshold + 3clipping algorithm) and visually inspecting the light-curves.
MOS1 & MOS2 background treated using spectral modeling instead of direct spectral subtraction, following the recipe of Leccardi & Molendi (2008).
PN data used only in the imaging analysis (i.e. to subtract point sources) because of problems in background characterization and inconsistencies with MOS temperature and abundance measure

11. MOS background modeling
Leccardi & Molendi (2008) 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.

12. Spectral analysis strategy
We determined r500 iteratively applying the formula derived by Vikhlinin (2006) to the r500 annulus:
Spectra in the following spatial bins were extracted and fitted with XSPEC v using Cash statistics:
r r500
r 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.

13. X-ray global properties
Baldi et al. 2011
(A&A submitted)
The <T> distribution ranges from 1.5 to 11 keV, with a peak around 4 keV
The abundance <Z> measured in the r500 range is always < 0.5 Z

14. Integrated abundance evolution
Baldi et al (A&A submitted)
No clear evolution with redshift is observed considering the emission from the whole cluster (0-0.6r500) and excising the core ( r500)
Fitting the data points with a power-law Z  (1+z)- we obtain:
0-0.6 r500 :  = 0.80.5
r500 :  = 0.60.7
The most significant deviation from = 0 is observed in the r500 sample, where  = 0.750.74 at a 90% c.l.

15. Spatially resolved abundance evolution
Baldi et al (A&A submitted)
No evidence of correlation between redshift and metal abundance is observed also spatially resolving the cluster emission.
Fitting the z vs. Z distribution with a power-law we have:
0-0.15r500 : = 0.20.7
r500 : = 0.40.7
>0.4 r500 : = 1.71.9
The lack of evolution is confirmed also by a Spearman’s rank  analysis which shows a high probability of no correlation (p>0.5)

16. Temperature-abundance correlation
Baldi et al (A&A submitted)
A correlation between the temperature and the abundance measured in galaxy clusters has been previously reported (e.g. Baumgartner et al. 2005, Balestra et al. 2007)
Considering the integrated emission from the whole cluster, we fit a powerlaw Z  kT- obtaining:
0-0.6 r500 :  = -0.10.1
r500 :  = -0.10.2
Although the abundance shows a higher spread in its values in lower temperature clusters (kT < 5 keV), no obvious trend between kT and Z is observed

17. Comparison with previous works
Considering the emission from the clusters as a whole we could detect only a mild evolution (<2)
A comparison with the previous results obtained on the abundance evolution in the literature is possible (Balestra et al. 2007, Maughan et al. 2008, Anderson et al. 2009) can be performed by selecting the clusters at z>0.4 and fitting the z-Z distribution with a power-law
Reference
Extr. Radius
N(obj)
Redshift 
Baldi 11
0-0.6 r500 39
0.750.47
0.280.01
r500
0.610.69
0.240.02
Balestra 07
rvir 46
0.680.50
Maughan 08
0-1r500 50
2.440.73
0.300.02
0.15-1r500
3.870.99
0.220.02
Anderson 09
S/N optimized 23
0.090.82

18. Comparison with previous works
All samples are not showing significant evolution (>2) except Maughan’s samples, which shows values of  significantly higher than all the others
The high  values in Maughan’s are probably driven by the stringent upper limits on Z measured in a high-z cluster (CLJ at z=1.03 with Z<0.04 Z)
Baldi et al (A&A submitted)
Removing the contribution of this cluster from Maughan et al. samples is lowering the values of  (=2.180.74 in the 0-1r500 sample; =-0.151.31 in the r500 sample)
This is clearly a caveat about being careful in interpreting results based on small samples
Reference
Extr. Radius
N(obj)
Redshift 
Baldi 11
0-0.6 r500 39
0.750.47
0.280.01
r500
0.610.69
0.240.02
Balestra 07
rvir 46
0.680.50
Maughan 08
0-1r500 50
2.440.73
0.300.02
0.15-1r500
3.870.99
0.220.02
Anderson 09
S/N optimized 23
0.090.82

19. Spatially resolved weighted abundance mean
We computed an error weighted abundance mean for each of the spatially resolved radial regions in three redshift bins:
0.4  z  0.5
0.5  z  0.7
0.7  z  1.4
Baldi et al (A&A submitted)
The z-Z relation is consistent with no evolution at every radius
Although no significant evolution is present, the trend in the r500 radial bin:
complements nicely the measures of Maughan et al. (2008)
broadly agrees with the predictions of Ettori (2005) model (the slopes are different at 1.5 level)

20. Radial dependence of the abundance
This is the first time that ICM abundance is parametrized as a function
of both radius and redshift Direct comparison with models of metal
diffusion (e.g. Ettori 2005) and simulations of chemical enrichment (e.g.
Fabjan et al. 2010) in the ICM
From the error weighted abundance measures, a radial dependence of the abundance on the radius is evident
Baldi et al (A&A submitted)
The dependence of Z on both r and z could be tested by fitting the relation:
The best fit values of the free parameters are:
Z0 = 0.36  0.03
a = 0.32  0.07
= 0.07  0.21
A significant negative trend of Z with the radius (and no redshift evolution) is present

21. More interesting science is on the way…
We are using the large dataset built from the XMM-Newton archive galaxy cluster sample to achieve more interesting science results, e.g.:
Temperature profiles of high redshift clusters
Study of the evolution of elements other than Fe
And more is on the way…
SLIDES

22. Temperature profiles at high z
We selected a sub-sample made of the 12 brightest clusters (MOS net counts > 3000) to extract a radial temperature profile
Baldi et al. 2011a
in prep.

23. Temperature profiles at high z
Solid line: Vikhlinin et al. (2006) relation renormalized to our data
Dotted line: Best-fit to our data, leaving key parameters in the Vikhlinin et al. relation free to vary
Dashed line: Best fit to our data at r > 0.15 r500
Baldi et al. 2011a
in prep.
Preliminary results are showing a different behavior of the temperature decrease at r > 0.15r500 with respect to the relation of Vikhlinin et al. (2006), derived in local clusters observed by Chandra

24. Elemental abundances at high z
We fitted simultaneously all the spectra divided in 4 different redshift bins to study the evolution with redshift of Mg, Si, S, Ni and Fe
Mg, Si, S and Ni Fe
Baldi et al. 2011b in prep.
Preliminary analysis shows a clear evolution with z only in Fe abundance, although Si is showing hints of evolution at z > 0.7

25. 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 structure formation in 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

26. 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) O Mg Si S

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

28. Summary
We presented a sample of 39 galaxy clusters at 0.4<z<1.4 extracted from the XMM-Newton archive.
We do not observe a statistically significant abundance evolution with z in the integrated emission from the whole cluster. The most significant deviation from no evolution is found in the 0-0.6r500 sample (90% c.l.)
Dividing the emission in 3 radial bins, no significant evidence of abundance evolution (al-ways below 1σ) could be observed.
We compared our results with previous works in the literature. Our results are fully consistent with Balestra et al. (2007), Maughan et al. (2008) and Anderson et al. (2009) samples at z > 0.4, as indicated by a power-law fit of the z-Z distributions (differences in γ and normalization always <2σ).
We computed error-weighted means of the spatially resolved abundances in three redshift bins. The abundance weighted mean is consistent to be constant with redshift at every radius. The trend in the r500 radial bin complements nicely the measures of Maughan et al., and agrees with the predictions of Ettori (2005) model.
We parametrized the dependence of the metal abundance on both the radius r and the redshift z with a function of two variables Z (r,z). The relation obtained show a significant negative trend of Z with the radius and, although no significant evolution with the redshift is detected, it represents the first time that the ICM abundance has been parametrized as a function of both r and z that could be directly compared with the models of diffusion of metals and the simulations of chemical enrichment in the ICM.