Multi-wavelength studies of substructures and inhomogeneities in galaxy clusters

Transcription

1 Multi-wavelength studies of substructures and inhomogeneities in galaxy clusters SZE (RX J1347) ΔS X (A3667) Tetsu Kitayama Toho University, Japan Infrared (Coma)

2 1. How can we study ICM physics with the SZE?

3 RX J z=0.45 Carlstrom et al Mason et al Contours: SZE, OVRO/BIMA 30GHz) Color: X-ray, ROSAT (FWHM~ 5 ) Appeared to be a relaxed cluster Contours: SZE, MUSTANG 90GHZ) Color: X-ray, Chandra (FWHM~0.5, smoothed to 10 ) X-ray: ~100 clusters SZE: only 1 cluster so far

4 Why SZE? For the same thermal plasma, I SZ n e T e dl independent of z : Inverse Compton I X n e2 Λ(T e ) dl /(1+z) 4 : Bremsstrahlung Λ(T e )~T e 1/2 for bolometric Spatially resolved X-ray spectroscopy (& weak lensing) become difficult at high z Higher frequency of major mergers at z>0.5 Unique probe of thermal structures in distant clusters - gravitational potential - merger shocks - missing SZE relative to X-rays? etc.

5 SZE & X-ray: Temperature profile T e & n e deprojection without X-ray spectroscopy Assuming spherical sym., I SZ n(r)t(r) dl I X n 2 (r) Λ[T(r)] dl (Yoshikawa et al. 1999) RX J excluding subclump & point source (TK, Komatsu, Ota et al. 2004) Applicable to X-ray faint gas - distant clusters -outskirts

6 High resolution SZE images of RX J (largest L X, strongest SZE: y max ~10-3, z=0.45) point source 100, 580 kpc sources + 90GHz (Mason+10) MUSTANG on GBT 100m 9 beam + 4 smoothing 1σ=0.3 mjy/beam 150GHz (Komatsu+01; TK+04) NOBA on Nobeyama 45m 13 beam + 15 smoothing 1σ=0.7 mjy/beam 350GHz (Komatsu+99; TK+04) SCUBA on JCMT 15m 15 beam + 15 smoothing 1σ=3.0 mjy/beam Contours: 1~5σ (Mason s talk on Tue.) Contours: Chandra kev (Allen+02)

7 SZE as a probe of violent mergers RX J at 150GHz If merger shock, Rankine-Hugoniot relation: Mach number~2.1 (γ=5/3) V preshock ~3900 km/s V postshock ~1600 km/s (relative to the shock front) SZE & Chandra joint analysis (TK+04) n excess = (1.45 ± 0.58) x 10-2 cm -3 L excess = 250 ±190 kpc (l.o.s. extent) kt excess = 28.5 ± 7.3 kev cf. Chandra only: kt excess >21.5 kev cf. Bullet cluster (Markevitch 2006) Chandra only: T 2 ~35 kev, T 1 ~9keV Mach number ~ 3.0 (γ=5/3) V preshock ~4700 km/s V postshock ~1600 km/s

8 Further evidences of merger in RX J kev Suzaku kev spectrum 150ks (Ota et al. 2008) Combined with spatially resolved Chandra kev data, Contours: Radio halo Color: XMM Contours: Lensing mass Grey scale: VLT (Gitti et al. 2007) (Miranda et al. 2008) kt excess = kev (X-ray only) cf. kt excess =28.5 ± 7.3 kev (SZE+Chandra)

9 Hard X-ray missions in near future NuStar (2012-): Hard X (E=5 80 kev), HPD=43 ASTRO-H (2014-): Micro-calorimeter: E= kev, ΔE=7eV, HPD=1.7 Hard X : E=5 80 kev, HPD=1.7 etc. IXO (2021-): E=0.3-7 kev, HPD= kev 30 High resolution (<10 ) SZE obs. are important and complementary.

10 SZE imaging with ALMA (Yamada et al.) Merits: Spatial resolutions (<10 ) Removal of point sources 12m 50 Higher resolutions Challenges: Limited FOV, sensitivity to extended emission combination of 12m & ACA, mosaicing ACA (Atacama Compact Arrays) 7m 12 & 12mSD 4 Lower resol. Mock simulations

11 Mock Obs. : Set up INPUT: Simulated merging cluster from Takizawa (2005) Place it at z=1, Dec=-23 deg Add noise & point sources 100 =0.8Mpc at z=1 90GHz 3 sources 3 mjy 1 mjy 0.3 mjy

12 Mock obs. with point 90GHz Raw map (Dirty beam uncorrected) Most compact configuration Mixed side lobes 12m long baseline map remove sources self-consistently Completely blind Final map

13 2. What about even smaller scales?

14 Inhomogeneities in simulated clusters (Kawahara et al. 2007, ApJ 659, 257) Local Universe Simulations (SPH) by Dolag et al. (2005) Smoothing IRAS map to linear scale initial conditions + cooling, SF, SN feedback, etc. Left: δ n = n(r,θ,φ)/<n>(r) Right: δ T = T(r,θ,φ)/<T>(r) red: simulation data for an entire cluster black:log-normal distribution Both n e & T e have fluctuations (nearly log-normal).

15 Observed inhomogeneities in S X S X (cnt/s/cm 2 /arcmin 2 ) Data Lognormal Poisson Gaussian (Gaussian) 2 Chandra images of A3667 at z=0.06, after masking point sources and smoothing over 4 x 4 = 3 kpc/h (Non-cooling flow cluster with highest counts) Observed S X has fluctuations (nearly log-normal). inhomogeneities in n e with Δlog n e ~0.4 (Kawahara et al ApJ 687, 936)

16 Observed inhomogeneities in T e Emission measure vs T Hydra A, XMM spectroscopy Broad T e distribution, nearly lognormal with Δlog T e ~0.2 (Simionescu et al. 2009)

17 Systematic errors in H 0 (d A ) from SZE/X-ray Sources for overestimation clumpiness <n 2 >/<n> 2 unresolved radio sources Sources for underestimation inhomogeneous T e radial gradient of <T e > if central I SZ is used. Others asphericity if R // > R, underestimate < overestimate The effects marked in red should yield, in total, 10~15% underestimate in H 0, if isothermal βmodel is used. (Kawahara et al. 2008, ApJ 674, 11) Another probe of SZE/X-ray intensity ratio

18 What about the real measurements? In general, X-ray/SZE used to give low H 0 : ROSAT/ASCA + OVRO/BIMA, 18 clusters (Reese et al. 2002) H 0 =60± km/s/mpc using isothermal βmodel On the other hand, Chandra + OVRO/BIMA, 38 clusters (Bonamente et al. 2006) H 0 = km/s/mpc using isothermal βmodel cf. SNIa + Cepheid,,, (Riess et al. 2009) H 0 =74.3±3.6 km/s/mpc CMB (Komatsu et al. 2010) H 0 =71.0±2.5 km/s/mpc

19 Impact of Chandra calibration (Reese et al., arxiv: ) (used in Bonamente+06) 38 clusters from Bonamente+06 CALDB versions (2005), (2009) vs (2010) - effective area of mirror - ACIS contaminations model etc. SZ (OVRO/BIMA) + X-ray (Chandra) ~10% change in T spec ~15% change in H 0 1/d A T 1.5 v3.1: H 0 =70.0±3.7 v4.1: H 0 =55.4±2.9 v4.2: H 0 =63.7±3.3 (stat. error) Should affect cluster cosmology widely!

20 3. Are there dust grains in ICM?

21 Search for intracluster/intergalactic dust Extinction of background sources 1962 Zwicky Coma A V ~ Chelouche et al SDSS clusters E(g-i)=0.008±0.003 A V =0.013±0.004 at d>r 200, z~ Menard et al SDSS galaxies E(g-i)~0.01 & A V ~0.02 (both field & cluster) at d=20kpc/h, z~0.3 IR emission 2002 Stickel et al. Coma I(120μ)~0.2 MJy/sr (ISO) 5 Abell clusters no detection 2005 Montier, Giard clusters I(100μ)~0.03MJy/sr (IRAS) Galaxies are not removed! If real, dust should be newly supplied (τ sputt ~10 8 yr for grain size 0.1μm). dust can contaminate the SZE, SNIa,,,

22 Limits on intracluster dust in Coma by Spitzer (Kitayama et al. 2009, ApJ 695, 1191) Source are masked. Coma at z=0.023 Color: ROSAT X-ray Boxes: Scan paths Spitzer/MIPS, 40 ksec 24μ 70μ 160μ 24μ(FWHM 6 ) : < MJy/sr within 100kpc 70μ( 18 ) : < MJy/sr 160μ( 40 ) : < MJy/sr M dust /M gas <10-5, A V <0.02 at r<100kpc red: bin size = 3.6 = 100 kpc green: 7.2 = 200 kpc black: collisional heating model that matches the ISO result I(120μm)=0.2 MJy/sr

23 Summary 1. High resolution (<10 ) SZE obs. - Unique probe of the ICM physics, especially at high z - Sample size and precision will improve, e.g., by ALMA 2. ICM has small-scale fluctuations &X-ray calibration uncertainty is not negligible: ΔT spec ~10% Impacts on cluster cosmology (H0, σ8, scaling relations,,,) 3. Upper limit on the intracluster dust in Coma M dust /M gas <1/1000 of Galactic ISM at r<100pkc