Astron. Nachr. / AN 000, No. 00, 1 – 6 (0000) / DOI please set DOI!

The XMM Cluster Outskirts Project (X-COP)

D. Eckert1,?, S. Ettori2,3, E. Pointecouteau4,5, S. Molendi6, S. Paltani1, C. Tchernin7, and The X-COP collaboration

1 Department of Astronomy, University of Geneva, ch. d’Ecogia 16, 1290 Versoix, Switzerland 2 INAF - Osservatorio Astronomico di Bologna, Via Ranzani 1, 40127 Bologna, Italy 3 INFN, Sezione di Bologna, viale Berti Pichat 6/2, 40127 Bologna, Italy 4 CNRS; IRAP; 9 Av. colonel Roche, BP 44346, F-31028 Toulouse cedex 4, France 5 Universite´ de Toulouse; UPS-OMP; IRAP; Toulouse, France 6 INAF - IASF-Milano, Via E. Bassini 15, 20133 Milano, Italy 7 Center for Astronomy, Institute for Theoretical Astrophysics, Heidelberg University, Philosophenweg 12, 69120 Hei- delberg, Germany

The dates of receipt and acceptance should be inserted later

Key words X-rays: galaxies: clusters - Galaxies: clusters: general - Galaxies: clusters: intracluster medium - cosmology: large-scale structure Galaxy clusters are thought to grow hierarchically through the continuous merging and accretion of smaller structures across cosmic time. In the Local Universe, these phenomena are still active in the outer regions of massive clusters (R > R500), where the matter distribution is expected to become clumpy and asymmetric because of the presence of accreting structures. We present the XMM-Newton Cluster Outskirts Project (X-COP), which targets the outer regions of a sample 14 of 13 massive clusters (M500 > 3 × 10 M ) in the redshift range 0.04-0.1 at uniform depth. The sample was selected based on the signal-to-noise ratio in the Planck Sunyaev-Zeldovich (SZ) survey with the aim of combining high-quality X-ray and SZ constraints throughout the entire cluster volume. Our observing strategy allows us to reach a sensitivity of 3 × 10−16 ergs cm−2 s−1 arcmin−2 in the [0.5-2.0] keV range thanks to a good control of systematic uncertainties. The combination of depth and field of view achieved in X-COP will allow us to pursue the following main goals: i) measure the distribution of entropy and thermal energy to an unprecedented level of precision; ii) assess the presence of non-thermal pressure support in cluster outskirts; iii) study the occurrence and mass distribution of infalling gas clumps. We illustrate the capabilities of the program with a pilot study on the cluster Abell 2142.

c 0000 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

1 Introduction Sembolini et al. 2016; Voit et al. 2005). However, non- gravitational processes induce an additional injection of en- In the hierarchical structure formation paradigm, galaxy tropy and can therefore be traced through the departures clusters are expected to form through the continuous merg- from the theoretical predictions (Chaudhuri et al. 2012). ing and accretion of smaller structures (see Kravtsov & Bor- Such departures have been observed for a long time in clus- gani 2012, for a review). In the local Universe, such pro- ter cores, where gas cooling and feedback from supernovae cesses should be observable in the outer regions of massive and active galactic nuclei are important (e.g. David et al. clusters, where galaxies and galaxy groups are infalling for 1996; Ponman et al. 1999; Pratt et al. 2010). More recently, the first time and smooth material is continuously accreted several works reported a deficit of entropy in massive clus- ters around the virial radius (see Reiprich et al. 2013, for a

arXiv:1611.05051v1 [astro-ph.CO] 15 Nov 2016 from the surrounding cosmic web. review), which has been interpreted as a lack of thermaliza- The hot plasma in galaxy clusters is expected to be 7 8 tion of the ICM induced, e.g., by an incomplete virializa- heated to high temperatures (10 − 10 K) through shocks tion of the gas (e.g. Bonamente et al. 2013; Ichikawa et al. and adiabatic compression at the boundary between the 2013; Kawaharada et al. 2010), non-equilibration between free-falling gas and the virialized intra-cluster medium electrons and ions (Hoshino et al. 2010), non-equilibrium (ICM, Tozzi et al. 2000). The thermodynamical properties ionization (e.g. Fujita et al. 2008), or weakening of the ac- of the gas retain information on the processes leading to cretion shocks (Lapi et al. 2010). However, these models the thermalization of the gas in the cluster’s potential well, −2/3 have received little support from cosmological simulations which is encoded in the gas entropy K = kT ne . Gravi- so far (e.g. Avestruz et al. 2015; Lau et al. 2015; Vazza et al. tational collapse models predict that the entropy of stratified 2010). cluster atmospheres increases steadily with radius, follow- ing a power law with index ∼ 1.1 (Borgani et al. 2005; The gas content of infalling dark-matter halos interacts with the ICM and is stripped from its parent halo through ? Corresponding author: e-mail: [email protected] the influence of the ram pressure applied by the ICM of the

c 0000 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 2 D. Eckert et al.: Instructions for authors main cluster. This process is expected to be the main mech- 2 Sample selection anism through which the infalling gas is heated up and viri- alized into the main dark-matter halo (Gunn & Gott 1972; Heinz et al. 2003; Roediger et al. 2015; Vollmer et al. 2001) To implement the strategy presented above, we selected a and it is believed to be key to the evolution of the cluster list of the most suitable targets to conduct our study. The galaxy population by quenching rapidly the star formation criteria used for the selection are the following: activity in clusters (Bahe´ & McCarthy 2015; Roediger & Bruggen¨ 2008). Recent observational evidence suggest that 1. SNR > 12 in the PSZ1 catalog (Planck Collabora- thermal conduction in the ICM is strongly inhibited (e.g. tion XXIX 2014): This condition is necessary to target Gaspari & Churazov 2013; Sanders et al. 2013). The long the most significant Planck detections and ensure that conduction timescale therefore delays the virialization of the SZ effect from all clusters be detected beyond R ; the stripped, low-entropy gas inside the potential well of the 500 2. Apparent size θ500 > 10 arcmin: Given the limited main cluster (De Grandi et al. 2016; Eckert et al. 2014), angular resolution of our reconstructed SZ maps (∼ 7 which causes the ICM in the outer regions of massive clus- arcmin), this condition ensures that all the clusters are ters to be clumpy (Mathiesen et al. 1999; Nagai & Lau 2011; well-resolved, such that the contamination of SZ flux Vazza et al. 2013). Since the X-ray emissivity depends on from the core is low beyond R ; the squared gas density, inhomogeneities in the gas distri- 500 3. Redshift in the range 0.04 < z < 0.1: This criterion bution lead to an overestimation of the mean gas density allows us to cover most of the azimuth out to R with (Eckert et al. 2015b; Nagai & Lau 2011; Simionescu et al. 200 5 XMM-Newton pointings (one central and four offset) 2011), which biases the measured entropy low. This effect whilst remaining resolved by Planck; needs to be taken into account when measuring the entropy 21 −2 4. Galactic NH < 10 cm : Since we are aiming at associated with the bulk of the ICM. In addition, large-scale maximizing the sensitivity in the soft band, this con- accretion patterns in the direction of the filaments of the dition makes sure that the soft X-ray signal is weakly cosmic web induce asymmetries in the gas distribution (e.g. absorbed. Eckert et al. 2012; Roncarelli et al. 2013; Vazza et al. 2011). Such filaments are expected to host the densest and hottest phase of the warm-hot intergalactic medium (e.g. Cen & This selection yields a set of the 15 most suitable targets Ostriker 1999; Dave´ et al. 2001; Eckert et al. 2015a), which for our goals. We excluded three clusters (A2256, A754, and are expected to account for most of the missing baryons in A3667) because of very complicated morphologies induced the local Universe. by violent merging events, which might hamper the analy- sis of the Planck data given the broad Planck beam. The re- maining 12 clusters selected for our study are listed in Table In this paper, we present the XMM-Newton cluster out- 1, together with their main properties. A uniform 25 ks map- skirts project (X-COP), a very large programme on XMM- ping with XMM-Newton was performed for 10 of these sys- Newton that aims at advancing significantly our knowledge tems in the framework of the X-COP very large programme of the physical conditions in the outer regions of galaxy (Proposal ID 074441, PI: Eckert), which was approved dur- 1 clusters (R > R500 ). X-COP targets a sample of 13 mas- ing XMM-Newton AO-13 for a total observing time of 1.2 sive, nearby clusters selected on the basis of their high Ms. The remaining 2 systems (A3266 and A2142) were signal-to-noise ratio (SNR) in the Planck all-sky survey mapped by XMM-Newton previously. Although the avail- of Sunyaev-Zeldovich (SZ, Sunyaev & Zeldovich 1972) able observations of A3266 do not extend all the way out to sources (Planck Collaboration XXIX 2014; Planck Col- R200, they are still sufficient for some of our objectives and laboration XXVII 2015). In the recent years, the progress we include them in the present sample. Finally, we add Hy- achieved in the sensitivity of SZ instruments allowed to dra A/A780 to the final sample. While the SZ signal from extend the measurements of the pressure profile of galaxy this less massive cluster is not strong enough to be detected clusters out to the virial radius and beyond (Planck Collab- beyond R500, a deep, uniform XMM-Newton mapping exists oration V 2013; Sayers et al. 2013). The high SNR in the for this system (see De Grandi et al. 2016, for more details). Planck survey ensures a detection of the SZ effect from our targets well beyond R500. X-COP provides a uniform 25 Our final sample therefore comprises 13 clusters in the 14 15 ks mapping of these clusters out to R200 and beyond, with mass range 2 × 10 < M500 < 10 M and X-ray tem- the aim of combining high-quality X-ray and SZ imaging perature 3 < kT < 10 keV. In Table 1 we also provide the throughout the entire volume of these systems. values of the central entropy K0 from the ACCEPT catalog (Cavagnolo et al. 2009), which is an excellent indicator of a cluster’s dynamical state (Hudson et al. 2010). According to this indicator, five of our clusters are classified as relaxed, 2 cool-core systems (K0 < 30 keV cm ), while the remaining 1 For a given overdensity ∆, R∆ is the radius for which eight systems are dynamically active, non-cool-core clus- 3 M∆/(4/3πR∆) = ∆ρc ters.

c 0000 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.an-journal.org Astron. Nachr. / AN (0000) 3

Table 1 Master table presenting the basic properties of the X-COP sample.

Name z SNR LX,500 kTvir Y500 M500 R500 θ500 K0 Ref 44 −1 −3 2 14 2 Planck [10 ergs s ] [keV] [10 arcmin ][10 M ] [kpc] [arcmin] [keV cm ] +0.30 A2319 0.0557 49.0 5.66 ± 0.02 9.60−0.30 43.17 10.56 1525 23.49 270.23 ± 4.83 2 ? +0.35 A3266 0.0589 40.0 3.35 ± 0.01 9.45−0.36 23.52 10.30 1510 22.09 72.45 ± 49.71 1 ? +1.01 A2142 0.090 28.4 8.09 ± 0.02 8.40−0.76 18.54 8.51 1403 13.92 68.06 ± 2.48 1 +0.19 A2255 0.0809 26.5 2.08 ± 0.02 5.81−0.20 11.17 4.94 1172 12.80 529.10 ± 28.19 1 +0.09 A2029 0.0766 23.2 6.94 ± 0.02 8.26−0.09 12.66 8.36 1399 16.08 10.50 ± 0.67 1 +0.11 A85 0.0555 22.8 3.74 ± 0.01 6.00−0.11 16.97 5.24 1205 18.64 12.50 ± 0.53 1 +0.07 A3158 0.059 19.8 2.01 ± 0.01 4.99−0.07 10.62 3.98 1097 16.03 166.01 ± 11.74 1 +0.07 A1795 0.0622 19.3 4.43 ± 0.01 6.08−0.07 6.43 5.33 1209 16.82 18.99 ± 1.05 1 +0.10 A644 0.0704 17.3 3.40 ± 0.01 7.70−0.10 7.22 7.55 1356 16.82 132.36 ± 9.15 3 +0.09 A1644 0.0473 16.1 1.39 ± 0.01 5.09−0.09 13.96 4.12 1115 20.02 19.03 ± 1.16 1 +0.04 RXC J1825 0.065 15.2 1.38 ± 0.01 5.13−0.04 8.39 4.13 1109 14.81 217.93 ± 6.33 4 † +0.32 ZwCl 1215 0.0766 12.8 2.11 ± 0.01 6.27−0.29 - 5.54 1220 14.01 163.23 ± 35.62 1 ? ‡ +0.08 A780 0.0538 - 2.25 ± 0.01 3.45−0.09 - 2.75 872 13.87 13.31 ± 0.66 1 Column description: 1. Cluster name. The clusters identified with an asterisk were mapped prior to X-COP. Abbreviated names: RXC J1825.3+3026, ZwCl 1215.1+0400, A780/Hydra A ; 2. Redshift (from NED); 3. Signal-to-noise ratio (SNR) in the Planck PSZ2 catalog (Planck Collaboration XXVII 2015). †In PSZ1 (Planck Collaboration XXIX 2014), but not in PSZ2 as it falls into the PSZ2 point source mask (see Table E.4 in Planck Collaboration XXVII (2015)). The SNR expected in PSZ2 from Eq. 6 and Table 3 in Planck Collaboration XXVII (2015) is about 16. ‡Below both PSZ1 and PSZ2 detection threshold; 4. Luminosity in the [0.5-2] keV band (rest frame); 5. Virial temperature; 6. Integrated Y parameter from the PSZ2 catalog; 7. Mass within an overdensity of 500, estimated using the M − T relation of Arnaud et al. (2005); 8. Corresponding value of R500 (in kpc); 9. Apparent size of R500 in arcmin; 10. Central entropy K0, from Cavagnolo et al. (2009); 11. Reference for the cluster temperature. 1: Hudson et al. (2010); 2: Molendi et al. (1999); 3: Cavagnolo et al. (2009) ; 4: This work (in prep.)

3 Abell 2142: a pilot study 2003) and estimated the median surface brightness from the resulting distribution. The median of the surface brightness Abell 2142 (z = 0.09, Owers et al. 2011) is the first clus- distribution was found to be a robust estimator of the mean ter for which the X-COP observing strategy was applied. In gas density (Zhuravleva et al. 2013), unlike the mean of the Tchernin et al. (2016) we presented our analysis of this sys- distribution, which is biased high by the presence of accret- tem out to the virial radius, highlighting the capabilities of ing clumps. The ratio between the mean and the median can X-COP. The results of this program are summarized here. In thus be used as an estimator of the clumping factor, Fig. 1 we show an adaptively smoothed, background sub- hρ2i tracted XMM-Newton mosaic image of Abell 2142 in the C = 2 , (1) [0.7-1.2] keV range, with Planck contours overlayed. hρi where h·i denotes the mean over radial shells (see Eckert 3.1 XMM-Newton surface-brightness profile et al. 2015b, for a validation of this technique using nu- merical simulations). This technique allows us to excise all We developed a new technique to model the XMM-Newton clumps down to the size of the Voronoi bins, which in the background by calculating two-dimensional models for all case of Abell 2142 corresponds to 20 kpc. the relevant background components: the non X-ray back- In Fig. 2 (reproduced from Tchernin et al. 2016) we ground (NXB), the quiescent soft protons (QSP), and the show the mean and median surface-brightness profiles of cosmic components. To validate our background subtrac- Abell 2142. The median profile is clearly below the mean tion technique, we analyzed a set of 21 blank fields totaling at large radii, which highlights the importance of clumping 1.3 Ms of data. The analysis of this large dataset yields a in cluster outskirts. A significant X-ray signal is measured flat surface-brightness profile, with a scatter of 5% around out to 3 Mpc from the cluster core (∼ 2R500), beyond which the mean value. This analysis allows us to conclude that the the systematics dominate. Note the significant improvement background level can be recovered with a precision of 5% in over previous XMM-Newton studies, which were typically the [0.7-1.2] keV band (see Appendix A and B of Tchernin limited to the region inside R500 (e.g. Leccardi & Molendi et al. 2016). 2008; Pratt et al. 2007). To measure the average surface-brightness profile free of the clumping effect, we applied the technique developed 3.2 Planck SZ pressure profile in Eckert et al. (2015b). Namely, in each annulus we com- puted the distribution of surface-brightness values by apply- Abell 2142 is one of the strongest detections in the Planck ing a Voronoi tessellation technique (Cappellari & Copin survey, with an overall signal-to-noise ratio of 28.4 using the

www.an-journal.org c 0000 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 4 C. Tchernin et al.: Joint X-ray/SZ analysis of Abell 2142D. Eckert et al.: Instructions for authors

10−4

27.70 10-2

27.60

−5 ]

10 3 27.50

10-3 27.40

−6 Y parameter 27.3010

-4 SZ, deprojection method 1 27.20 10 pressure [keV/cm

Declination SZ, deprojection method 2 27.10 10−7 X-ray

2.5 102 103 27.00 2 10-5 1.5 103 1 26.900.5 R [kpc] χ 0 −0.5 26.80−1 Fig. 10. Electron pressure profile. The grey solid line and shaded −1.5 Fig. 3 Electron pressure profile obtained using Planck −2 A2142 area show the best-fit pressure profile obtained by deprojecting 2 3 and XMM-Newton. The blue data points and gray shaded 26.70 10 10 R [kpc] the SZ data using the multiscale method (deprojection method 240.10 240.00 239.90 239.80 239.70 239.60 239.50 239.40 239.30 239.20 239.10 area show the deprojected SZ electron pressure obtained Right ascension 1). The red points show the spectroscopic X-ray data deprojected usingwith the a non-parametric method of Vikhlinin and et a al.parametric (2006) and method, the blue respec- triangles Fig. 9. Top panel: Compton y parameter profile from Planck data tively.show the The result red of data the points deprojection denote of the the pressurey-parameter profile data ob- using Fig. 1 Adaptively smoothed, background subtracted (red points). The grey solid line and shaded area show the best-fit tainedthe same from methodology X-ray spectral as Planck analysis. Collaboration (2013) (depro- profileXMM-Newton convolvedimage with the of instrument Abell 2142 PSF. in The the [0.7-1.2]data points keV are jection method 2). The blue data points are correlated and the correlatedband. The and corresponding the associated errorsPlanck correspondCompton-parameter to the square con- root associated errors correspond to the square root of the diagonal el- oftours the diagonal are shown elements in white. of the The covariance contour matrix. levels correspond Bottom panel: to fileements measured of the covariance from a spectral matrix. X-ray The dashedanalysis. and An dashed-dotted excellent residual of the fit to the Planck data. The dashed and dashed- 1, 3, 5, 7, 10, 15, 20, 30, 40, and 50σ. The red circle indi- agreementvertical lines between represent X-rayR500 andand SZR200 pressure, respectively. profiles is found C. Tchernin et al.: Joint X-raydotted/SZ analysis vertical of lines Abell represent 2142 R500 and R200, respectively. cates the estimated value of R500 ∼ 1, 400 kpc. over the range of overlap. This confirms that X-ray and SZ Appendix B). We refer to this surface brightness profile as ”pho- techniques return a consistent picture of the gas properties -12 in galaxy clusters and further validates the method that is tometric” in the following, to be distinguished from the spatially 10 put forward in X-COP. limited ”spectroscopic” surface brightness profile. the square)] root of the diagonal of the covariance matrix. This 2 To obtain the photometric surface brightness profile we ex-likely bias a direct visual comparison. The observed discrepancy around10-13R may therefore not be a physical e↵ect. We will come tracted photon images in the energy band [0.7-1.2] keV and 200 3.3density Joint profiles X-ray/SZ obtained analysis in Sect. 3.2 (Fig.6) with the SZ pres- created exposure maps for each instrument using the SAS taskback to the excess in the purely X-ray pressure profile compared s arcmin sure profile derived in Sect. 3.4 (Fig. 10). Using the equation to the2 SZ profile in the discussion section. eexpmap and the PROFFIT v1.2 software (Eckert et al. 2011). 10-14 Oncek T = theP gas/n , density we derived and pressure the 3D temperature profiles are profile determined, for both Due to the moderate resolution of the Planck satellite (trans- B e e The choice of the [0.7-1.2] keV band is motivated by the fact that thethe radial azimuthal profiles mean of and temperature the azimuthalkT = medianPe/ne densityand entropy profiles. this particular band maximizes the signal-to-noise ratio (Ettorilating into 7 arcmin on our y-map), we can not recover con- We remind−5 that/3 while the density profile obtained using the az- -15 K = P ne can be inferred. The gravitating mass profile et al. 2010; Ettori & Molendi 2011) and avoids the bright andstraints10 on the SZ pressure profile from the y parameter measure- e canimuthal also medianbe recovered is corrected by solving for the the presence hydrostatic of clumps, equilibrium the one variable Al K↵ and Si K↵ fluorescence lines, without a↵ectingments close to the cluster center. Therefore, we will consider in obtained using the azimuthal mean is not. the following only the radial range beyond 400 kpc ( 4 arcmin), equation, too much the statistics. Surface-brightness profiles were accu- -16 10 ⇠ / mulated in concentric annuli starting from the surface-brightnessradius beyond which the constraints on the pressure profile from The resulting joint X-ray SZ 3D temperature profiles are showndP in Fig.GM 11. The(< r uncertainties) in the temperature profile the SZEnergy flux [erg/(cm data are less impacted by the PSF blurring and therefore peak (R.A.=239.58, Dec=27.23), taking vignetting e↵ects into = −ρ 2 . (2) -17 weredr estimated byr combining the MCMC runs for the pressure account. NXB profiles were accumulated in the same regionsmore10 reliable. 500 1000 1500 2000 2500 3000 and density. At each radius, the temperature and its uncertainty from the NXB maps taking both the contribution of the qui- R [kpc] By comparing the gravitating mass with the gas mass escent particle background and the soft protons into account. obtainedwere drawn by integrating from the distribution the gas density of output profile, temperature the profile values. of In this figure we also show the deprojected spectroscopic tem- To model the contamination from residual soft protons, we ex-4.Fig.Fig. Joint 2 5. Surface-brightness X-ray/SZMean (blue) analysis and profiles median of (green) A2142 obtainedXMM-Newton with dif-intracluster gas fraction can also be recovered. ferent methods. The data points show the spectroscopic mea-perature profile obtained with the method of Ettori et al. (2010). tracted the spectra of the entire observations and fitted the high-Thesurface-brightness combination of profiles the X-ray of Abell and SZ 2142 signal in the can [0.7-1.2] be used In Fig. 4 we show the radial entropy profile of Abell surements (red), the azimuthally averaged profile (blue), and the energy part of the spectra (7.5-12 keV) using a broken power-lawtokeV recover band. the The thermodynamical red data points show quantities the results that obtainedcharacterize us- As expected, we observe di↵erent behaviors of the tempera- azimuthal median (green). 2142 obtained by combining X-ray and SZ data. Once model (see Leccardi & Molendi 2008). A 2D model for the con-theing ICM. a spectral In this analysis. section, The we horizontal combineThe green the dashed and three-dimensional blue and solid thin lines dashedture profile depending on if gas clumping is taken into account or lines show the corresponding best-fits obtained with the mul-again, the data are compared with the results of the spectro- tamination of residual soft protons was created using the ESASSZshow pressure the sky profile background with the level X-ray and gas the density level of profile systematics, to re- not. Indeed, the increase in the clumping factor towards the out- task proton following Kuntz & Snowden (2008). The details of tiscale deprojection method. The solid and dashed horizon-scopicskirts (see X-ray Fig. analysis, 7) causes which the temperature highlights profilethe much obtained broader from coverrespectively. the radial distribution of temperature, entropy, hydro- the soft-proton modeling technique are provided in Appendix A,statictal linesmass, correspond and gas fraction. to the Moreover,total background we can recoverlevel, and these to theradialthe azimuthal range accessible mean to steepen by the jointwith cluster-centric X-ray/SZ technique. distance The com- and a careful validation using blank-sky pointings is presentedquantitiesuncertainty largely on correctedthe background, for the erespectively.↵ect of the clumpedThe dashed gas. andobservedpared to the entropy profile profiles estimated are usingcompared the azimuthal with the prediction median tech- R R in Appendix B together with an assessment of systematic uncer-Thisdashed-dotted is obtained vertical by comparing lines represent the X-ray500 surfaceand 200 brightness, respectively.ofnique. numerical We also simulations note that the using spectroscopic gravitational X-ray collapse profile only closely tainties. measureddata from using the full thePlanck meanmission of the (seeazimuthal Table 1. photon A significant counts (Voitfollows et al. the 2005). X-ray/SZ profile obtained using the azimuthal me- In addition, we also derived the azimuthal median surfacedistributionSZ signal is with observed and the as wellone estimated out to 3 Mpc with from the the median cluster of dianInterestingly, (except for the when very combining last data point, the SZ but data this with is an the artefact me- of brightness profile following the method described in Eckert etthecore, distribution and it can (as be detailed readily in transformed Eckert et al. into 2015a). a high-quality the deprojection method). We will come back to this point in the where Da 349.8 Mpc is the angular distance to the cluster indian (clumping corrected) gas density profile, the recovered al. (2015a). Namely, Voronoi tessellation was applied on the pressure profile.⇠ For the details of the Planck analysis pro- 3 discussion section. cm, z = 0.09 is its redshift and np is the proton density (in cm ),entropy profile is consistent with the theoretical expectation count image to create an adaptively binned surface-brightness cedure we refer to Planck Collaboration V (2013). The e↵ect of the overestimate the density has a clear signa- 4.1.which Temperature is related profileto the electron density by ne = 1.21np, assumingwithin 1σ, whereas if the mean (biased) density profile is map with a minimum of 20 counts per bin. The median surface ture in the temperature profile. Similar e↵ects are also expected thatIn the Fig. plasma 3 we is show fully the ionized.Planck pressure profile obtained used, at large radii the entropy falls significantly below the brightness was then estimated in each annulus by weighting theAssuming that the ICM is an ideal gas, the joint X-ray/SZ 3D in the other thermodynamic quantities derived from the X-ray usingAfter two having different converted deprojection the surface-brightness methods (see Tchernin profile intoexpectations. In the latter case, the behavior of the entropy surface brightness of each bin by its respective surface. To es-temperatureemission measure, profile can we be deprojected recovered theby combining resulting profiles the X-ray to es-analysis. 4 et al. 2016). The results are compared with the pressure pro- profile is very similar to a number of recent Suzaku studies, timate the uncertainty in the median, we performed 10 boot- timate the 3D electron density profile. For the deprojection, we strap resampling of the surface-brightness distributions binned compared the output of two di↵erent methods: the multiscale 9 uniformly using Voronoi tessellation. The standard deviation of method described in Eckert et al. (2015b) and an onion-peeling the bootstrap realizations was then adopted as the error on the cmethod0000 WILEY-VCH (Ettori Verlag et al. GmbH 2010). & Co. KGaA, Both Weinheim methods assume spherical sym- www.an-journal.org median. In the [0.7-1.2] keV energy band, the systematic uncer- metry. The di↵erences between the output of the two procedures tainty in the subtraction of the background amounts to 5% of the thus gives us a handle of the uncertainties associated with the sky background component, as shown by an analysis of a set of deprojection. 22 blank-sky pointings (see Appendix B). This uncertainty was In the multiscale deprojection method, the projected profile added in quadrature to the surface-brightness profiles. To convert is decomposed into a sum of multiscale basis functions. Each the resulting surface brightness profiles into emission measure, component can then be easily deprojected to reconstruct the 3D APEC XMM-Newton we folded the model through the response profile. Following Eckert et al. (2015b), we decompose the pro- and computed the conversion between count rate and emission jected profile into a sum of King profiles, with s the projected measure. We note that the conversion factor is roughly indepen- radius, related to the line-of-sight distance and the 3D radius r, dent of the temperature in the energy band [0.7-1.2] keV, pro- by: r2 = s2 + `2 vided that the temperature does not fall below 1.5 keV. ⇠ In Fig. 5 we show the comparison between the spectroscopic N 2 3i/2 s and the photometric surface-brightness profiles. An excellent EM(s) = Ni 1 + , (2) r , agreement is found between the profiles obtained with the two i=1 2 c i ! 3 methods. We can see that XMM-Newton detects a significant X 6 7 th 6 7 emission out to almost 3 Mpc from the cluster core, which cor- where i represents the i basis46 function57 and s the projected responds to the virial radius R100 2R500. radius. The parameters of this fit are the normalization (Ni), the ⇠ core radii (rc,i) and the slopes (i). The number of components and the core radii used for the fit of the projected profile are de- 3.2. XMM-Newton deprojected electron density profile termined adaptively from the total number of data points, with 5 the condition that one basis function is used for each block of The normalization of the APEC model, in units of cm , is re- 4 data points. The relation between the projected and 3D pro- lated to the electron density (ne) by files can then be computed analytically (see Appendix A of Eckert et al. 2015b, for details). This method provides an ade- 14 10 quate representation of the observed profile and of the underly- Norm = n n d3r, (1) ((1 + z) D )2 e p ing density 3D profile, although the derived parameters have no · a Z

6 Astron. Nachr. / AN (0000) C. Tchernin et al.: Joint X-ray/SZ analysis of Abell 2142 5 C. Tchernin et al.: Joint X-ray/SZ analysis of Abell 2142 profile. As above, we investigated the e↵ect of clumping on the hydrostatic mass by comparing the results obtained with the az- 10 imuthal mean andX-ray/SZ, median photometric density median profiles. X-ray/SZ, photometric mean / 9 In Fig. 13 we show the combined X-ray SZ hydrostatic mass profiles obtainedX-ray spectroscopic for data the di↵erent input density profiles. 8 The mass profilePurely obtained gravitational collapse using model the azimuthal median increases 1015

7 steadily,1 while the one obtained with the azimuthal mean den-

500 sity profile shows an unphysical turnover at R > R200. Such 6

turnovers have been reported in the literature and interpreted as /M K/K tot 5 evidence for a significant non-thermal pressure contribution to M kT [keV] sustain gravity (e.g. Kawaharada et al. 2010; Bonamente et al. 4 X-ray/SZ, photometric median 2013; Fusco-Femiano & Lapi 2014). X-ray/SZ, photometric mean X-ray, spectra-photometric data 3 We also derived the hydrostatic mass profile using X-ray- X-ray/SZ, photometric median Weak Lensing, Umetsu et al. (2009) only information with the method described in Ettori et al. 14 10 X-ray, Piffaretti et al. (2011) 2 X-ray/SZ, photometric mean (2010). This method assumes a Navarro-Frenk-White (NFW) X-ray, Akamatsu et al. (2011) X-ray spectroscopic data 1 R/R 1.2 500 Galaxy kinematics, Munari et al. (2014) 1 form for the underlying mass profile and uses the deprojected gas 500 1000 1500 2000 2500 3000 th 1 density profile to reproduce the observed temperature profile es- 103 R [kpc] /K 0.8 timated with the spectral analysis by inversion of the hydrostatic R [kpc] obs 0.6 K Fig. 11. Deprojected temperature profile. Red points: equilibrium0.4 equation applied on a spherically symmetric ob- Fig. 13. Mass profile of A2142. Green: X-ray/SZ combined pro- ject.0.2 The best-fit on the 2 parameters describing the NFW mass Fig. 5 Mass profiles of Abell 2142 obtained by combin- Spectroscopic data (from Fig. 3) deprojected using the method / model (i.e. the concentration and R200 in the present1 analysis) is file using the azimuthal median density profile. Blue: X-ray SZ of Ettori et al. (2010); green and blue: combined X-ray/SZ 2 R/R ing Planck and XMM-Newton data. The green curve and then obtained using a minimization500 technique. Applying this combined profile using the azimuthal mean density profile. Red: profile using the azimuthal median and azimuthal mean density shadedNFW fit area to the show spectroscopic the mass profile X-ray data inferred using from the method the median of profiles, respectively. The dashed and dashed-dotted vertical method to the photometric median density profile, we measure Fig.Fig. 12.a 4 concentrationDeprojectedEntropy profiles entropyc = 3.00 of profile. Abell0.06 Topand 2142 panel:R200 obtained= the2249 red by dots16 com-kpc. (clumpingEttori et al.(2010). corrected) For comparison, gas density we profile, also plot while the mass for the mea- blue lines represent R500 and R200, respectively. ± ± showbiningHereafter, thePlanck X-ray alland spectroscopic referencesXMM-Newton to thedata, method whiledata. ofThe the Ettori green green et al. and curve (2010) blue and will curvesurements the reported mean (biased) in the literature. density Brown profile triangle: was used.M500 Thefrom red L M relation (Pi↵aretti et al. 2011); pink reversed trian- curvesshadedbe represent applied area show to the the the X-ray photometric entropy/SZ profiles profilemedian obtained densityinferred profile. using from the the az- me- curveX indicates the mass profile inferred from spectroscopic In Fig. 13 we compare the resulting mass profile with that gle: M200 from Subaru weak lensing (Umetsu et al. 2009); dark imuthaldian (clumping median and corrected) azimuthal gas mean density density profile, profiles, while respec- for the X-ray measurements (Ettori et al. 2010). The symbols indi- 4.2. Entropy profile tively.obtained The black using line the shows X/SZ the method. expectation We can from see that purely the two grav- meth- green square: M200 from Suzaku X-ray (Akamatsu et al. 2011); blueods curve lead the to consistent mean (biased) results. density Good agreement profile was is found used. in The par- cateblack the empty values triangle: of MM200200calculatedfrom galaxy from kinematics weak gravitational (Munari et Assuming that entropy is only generated by spherical virialisa- itational collapse (Eq. (5), Voit et al. 2005). The dashed and red pointsticular show between the the spectroscopic X-ray-only and X-ray the measurements. median X-ray/SZ The pro- al. 2014). tion shocks driven by hierarchical structure formation, we expect dashed-dotted vertical lines represent R500 and R200, respectively. lensing (pink triangle, Umetsu et al. 2009), galaxy kinemat- files. We also show the comparison of several mass measure- an entropy profile that follows the gravitational collapse model. Bottomblack curve panel: represents ratio of the the X-ray expectation/SZ entropy of profile pure gravitational(Kobs) over ics (black triangle, Munari et al. 2014) and Suzaku X-ray thecollapse entropyments (Voitfrom profile et theal. expected literature: 2005). from Akamatsu the purely et al. gravitational (2011) (Suzaku col-, as- In such a case, the gas with low entropy sinks into the cluster suming hydrostatic equilibrium); Umetsu et al. (2009) (Subaru, (green square, Akamatsu et al. 2011) center, while the high-entropy gas expands to the cluster out- lapse model (Kth): for the azimuthal median (in blue) and az- imuthalweak mean gravitational (in green) lensing); density profiles. Munari etThe al. horizontal (2014) (optical dashed spec- skirts (Voit et al. 2005; Pratt et al. 2010) and the resulting entropy troscopy, galaxy dynamics) and Pi↵aretti et al. (2011) (ROSAT, the azimuthal median density profile lies slightly below the pro- profile has a power-law shape given by, line represents the expectation for Kobs = Kth. file obtained from the azimuthal mean. At R , the di↵erence whichLX foundM arelation). deficit of These entropy measurements beyond R500 are. summarized Our analysis in 200 non-thermalbetween the azimuthal energy budget median in and galaxy the azimuthal clusters; meaniii) profilesdetect in- R 1.1 thusTable highlights 4. All our the mass importance measurements of gas are clumping consistent when within in- the K(R) = K 1.42 keV cm2. (5) Onerror the bars contrary, with the we mass can see measurements that the X/ madeSZ profile in Akamatsu obtained et al. fallingis of the gas order clumps of 6%. to study the virialization of infalling ha- 500 · R terpreting the results of Suzaku observations of cluster out- 500 ! using(2011), the azimuthal Umetsu median et al. (2009), method Munari rises et steadily al. (2014) with and radius Pi↵aretti los withinWe derived the potentialthe gas fraction well ofas athe function main structure.of radius by com- skirts. bining the gas mass profiles with the corresponding hydrostatic The quantity K500 is defined as (Pratt et al. 2010), out toet the al. maximum (2011). radius accessible in this study (3000 kpc ⇡ massWe profiles presented (see Sect. a pilot 4.3). study In Fig.15 on we the compare galaxy the cluster resulting Abell 2/3 2/3 R100).In Therefore, Fig. 5 we if showthe presence the gravitating of clumps mass is taken profile into account obtained M500 1 2/3 2 in the X-ray data the deviation observed with the blue curve 2142gas fraction (Tchernin profiles et with al. 2016), the expected demonstrating Universal baryon the full fraction poten- = , by solving the hydrostatic equilibrium equation (Eq. 2). In K500 106 14 h(z) keV cm (6) 4.4. Gas fraction profile from Planck, corrected for the baryon fraction in the form of · 10 M fb almost completely disappears and at R200 the entropy falls within tial of X-COP for the study of cluster outskirts. The cluster ! ! this case, we find that the mass profiles calculated with stars (Gonzalez et al. 2007). Interestingly, we can see that while 14 just 1-Because of the of self-similar their deep gravitationalexpectation. well, massive clusters are ex- is detected out to 2 × R500 ∼ R100 both in X-ray and SZ. where M500=8.66 10 M is the cluster mass at R500 = 1408 kpc theWe meanpected stress andto that retain median the the SZ matter profilese↵ect collected is nearly are consistent. since insensitive their formation. We to clump- measure Thus, the gas fraction profile derived from the combination of the az- · = ⌦ /⌦ = . 15 The two techniques provide a remarkably consistent picture (values derived from Munari et al. 2014), fb b m 0 15 ingM200 (e.g.the= relativeBattaglia (1.41 amount et± al.0. of 2015),03) baryonic× the10 di and↵Merence dark , in matter between good should agreement the be two close imuthal median density profile with the SZ pressure profile is is the cosmic baryon fraction, with ⌦ the baryon density, ⌦ ofalmost the gasconstant properties, and close and to the they expected can be value combined ( 13-14%), to recover the b m profileswithto the theis caused values Universal only calculated value. by our Recent treatment withPlanck weak ofobservations gravitationalgas clumping of thein lensing the power ⇠ 3 gas fraction profile derived using the azimuthal mean density the matter density, and h(z) = ⌦m(1 + z) + ⌦⇤ the ratio of the X-rayspectrum data.+0.18 This of CMBshows15 anisotropies the importance indicate of takinga Universal into baryon account frac- the thermodynamic properties of the gas and the gravitat- (1.24−0.16 ×10 M , Umetsu et al. 2009) and galaxy kine- profile and the SZ pressure profile increases with radius and ex- Hubble constant at redshift z to its present value. the e↵tionects⌦ dueb/⌦ tom the= presence0.153 0 of.003 clumps (Planck in the Collaboration derivation of 2015c). the ing mass profile out to the cluster’s boundary. Our results p +0.26 15± ceeds the cosmic baryon fraction. We also note that the gas frac- We derived the combined SZ and X-ray 3D entropy profile thermodynamicsmaticsCorrected (1.31− for0 quantities..23 the× stellar10 M fraction, , Munari which et accounts al. 2014). for Thus, 10-20% our of 5/3 highlighttion profile the obtained importance from purely of taking X-ray the information effect of is gas slightly clump- by using the equation K = P /n with the X-ray density datathe do total not show amount any of baryonssign of hydrostaticin galaxy clusters bias (e.g.,even Gonzalez when ex- et e e ingabove into the account expected when value.We measuring will discuss the these properties points further of the in gas profiles obtained in Sect. 3.2 (Fig. 6) and the SZ pressure profile al. 2007), we expect a gas fraction of 13-14%. tending our measurements out to R200. Sect. 5.3.2. derived in Sect. 3.4 (Fig. 10). In Fig. 12 we show the entropy 4.3. Hydrostatic mass at large radii, where accretion from smaller structures is im- profiles obtained using the azimuthal mean and median density AssumingAssuming that the spherical ICM is in symmetry, hydrostatic the equilibrium total gas mass within is given the by portant. In the near future, X-COP will bring results of sim- profiles. For comparison, we also show the entropy profile ob- gravitationalthe integral potential of the of gas the density cluster, over the the total cluster enclosed volume, mass at ilar quality for a sizable sample of a dozen clusters, allow- 2/3 4 Conclusion 5. Discussion tained with the spectroscopic X-ray information (K = kBT/ne ) the distance r from the cluster center canr be estimated as ing us to determine universal profiles of the thermodynamic from our deprojected temperature and gas density spectroscopic 2 dPgM(rgas) (< r) = 4GM⇡ tot⇢(

www.an-journal.org c 0000 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 6 D. Eckert et al.: Instructions for authors

References arXiv:1303.3286 Roediger, E. & Bruggen,¨ M. 2008, MNRAS, 388, 465 Akamatsu, H., Hoshino, A., Ishisaki, Y., et al. 2011, PASJ, 63, Roediger, E., Kraft, R. P., Nulsen, P. E. J., et al. 2015, ApJ, 806, 1019 103 Arnaud, M., Pointecouteau, E., & Pratt, G. W. 2005, A&A, 441, Roncarelli, M., Ettori, S., Borgani, S., et al. 2013, MNRAS, 432, 893 3030 Avestruz, C., Nagai, D., Lau, E. T., & Nelson, K. 2015, ApJ, 808, Sanders, J. S., Fabian, A. C., Churazov, E., et al. 2013, Science, 176 341, 1365 Bahe,´ Y. M. & McCarthy, I. G. 2015, MNRAS, 447, 969 Sayers, J., Czakon, N. G., Mantz, A., et al. 2013, ApJ, 768, 177 Bonamente, M., Landry, D., Maughan, B., et al. 2013, MNRAS, Sembolini, F., Yepes, G., Pearce, F. R., et al. 2016, MNRAS, 457, 428, 2812 4063 Borgani, S., Finoguenov, A., Kay, S. T., et al. 2005, MNRAS, 361, Simionescu, A., Allen, S. W., Mantz, A., et al. 2011, Science, 331, 233 1576 Cappellari, M. & Copin, Y. 2003, MNRAS, 342, 345 Sunyaev, R. A. & Zeldovich, Y. B. 1972, Comments on Astro- Cavagnolo, K. W., Donahue, M., Voit, G. M., & Sun, M. 2009, physics and Space Physics, 4, 173 ApJS, 182, 12 Tchernin, C., Eckert, D., Ettori, S., et al. 2016, A&A, 595, A42 Cen, R. & Ostriker, J. P. 1999, ApJ, 514, 1 Tozzi, P., Scharf, C., & Norman, C. 2000, ApJ, 542, 106 Chaudhuri, A., Nath, B. B., & Majumdar, S. 2012, ApJ, 759, 87 Umetsu, K., Birkinshaw, M., Liu, G.-C., et al. 2009, ApJ, 694, Dave,´ R., Cen, R., Ostriker, J. P., et al. 2001, ApJ, 552, 473 1643 David, L. P., Jones, C., & Forman, W. 1996, ApJ, 473, 692 Vazza, F., Brunetti, G., Gheller, C., & Brunino, R. 2010, New As- De Grandi, S., Eckert, D., Molendi, S., et al. 2016, A&A, 592, tronomy, 15, 695 A154 Vazza, F., Brunetti, G., Gheller, C., Brunino, R., & Bruggen,¨ M. Eckert, D., Jauzac, M., Shan, H., et al. 2015a, Nature, 528, 105 2011, A&A, 529, A17+ Eckert, D., Molendi, S., Owers, M., et al. 2014, A&A, 570, A119 Vazza, F., Eckert, D., Simionescu, A., Bruggen,¨ M., & Ettori, S. Eckert, D., Roncarelli, M., Ettori, S., et al. 2015b, MNRAS, 447, 2013, MNRAS, 429, 799 2198 Voit, G. M., Kay, S. T., & Bryan, G. L. 2005, MNRAS, 364, 909 Eckert, D., Vazza, F., Ettori, S., et al. 2012, A&A, 541, A57 Vollmer, B., Cayatte, V., Balkowski, C., & Duschl, W. J. 2001, Ettori, S., Gastaldello, F., Leccardi, A., et al. 2010, A&A, 524, ApJ, 561, 708 A68+ Zhuravleva, I., Churazov, E., Kravtsov, A., et al. 2013, MNRAS, Fujita, Y., Hayashida, K., Nagai, M., et al. 2008, PASJ, 60, 1133 428, 3274 Gaspari, M. & Churazov, E. 2013, A&A, 559, A78 Gunn, J. E. & Gott, III, J. R. 1972, ApJ, 176, 1 Heinz, S., Churazov, E., Forman, W., Jones, C., & Briel, U. G. 2003, MNRAS, 346, 13 Hoshino, A., Patrick Henry, J., Sato, K., et al. 2010, PASJ, 62, 371 Hudson, D. S., Mittal, R., Reiprich, T. H., et al. 2010, A&A, 513, A37+ Ichikawa, K., Matsushita, K., Okabe, N., et al. 2013, ApJ, 766, 90 Kawaharada, M., Okabe, N., Umetsu, K., et al. 2010, ApJ, 714, 423 Kravtsov, A. V. & Borgani, S. 2012, ARA&A, 50, 353 Lapi, A., Fusco-Femiano, R., & Cavaliere, A. 2010, A&A, 516, A34+ Lau, E. T., Nagai, D., Avestruz, C., Nelson, K., & Vikhlinin, A. 2015, ApJ, 806, 68 Leccardi, A. & Molendi, S. 2008, A&A, 486, 359 Mathiesen, B., Evrard, A. E., & Mohr, J. J. 1999, ApJ L, 520, L21 Molendi, S., De Grandi, S., Fusco-Femiano, R., et al. 1999, ApJ L, 525, L73 Munari, E., Biviano, A., & Mamon, G. A. 2014, A&A, 566, A68 Nagai, D. & Lau, E. T. 2011, ApJ L, 731, L10+ Owers, M. S., Nulsen, P. E. J., & Couch, W. J. 2011, ApJ, 741, 122 Planck Collaboration V. 2013, A&A, 550, A131 Planck Collaboration XXIX. 2014, A&A, 571, A29 Planck Collaboration XXVII. 2015, arXiv:1502.01598 Ponman, T. J., Cannon, D. B., & Navarro, J. F. 1999, Nature, 397, 135 Pratt, G. W., Arnaud, M., Piffaretti, R., et al. 2010, A&A, 511, A85+ Pratt, G. W., Bohringer,¨ H., Croston, J. H., et al. 2007, A&A, 461, 71 Reiprich, T. H., Basu, K., Ettori, S., et al. 2013, SSRv,

c 0000 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.an-journal.org