Suzaku Measurements of Gas Bulk Motions in a Galaxy Cluster Takayuki Tamura (ISAS/JAXA) In collaboration with K. Hayashida, S. Ueda, and M. Nagai (Osaka) July2011, Boston

Motivations Cluster evolutions via mergers Galaxies Dark Matter Gas Dynamics Cluster evolutions via mergers Cosmic Rays (Non-thermal Processes) X-ray Mass estimates (Precise Cosmology) Galaxies Dark Matter Move in concert or separately ? Shock, turbulence, particle acceleration Kinetic pressure ? What have we already observed ? As highlighted by this observation of the Bullet cluster, X-ray imaging have clearly shown that some systems are really forming. And we believe that we are witnessing dynamic features like shocks and cold-front. Using these X-ray imaging, however, we could only estimate gas motion in the plane of sky in a limited number of objects which happen to move in the plane of the sky. As far as I know, there is NO direct measurement of the gas motion so far. In my talk, I will try to convince you that we detected gas motion for the 1st time. July2011, Boston

How to measure Dynamical motion (1) X-ray imaging have shown cluster dynamics in the plane of the sky. (2) Doppler Mapping of X-ray lines Goals for future missions What have we already observed ? As highlighted by this observation of the Bullet cluster, X-ray imaging have clearly shown that some systems are really forming. And we believe that we are witnessing dynamic features like shocks and cold-front. Using these X-ray imaging, however, we could only estimate gas motion in the plane of sky in a limited number of objects which happen to move in the plane of the sky. As far as I know, there is NO direct measurement of the gas motion so far. In my talk, I will try to convince you that we detected gas motion for the 1st time. Markevitch & Vikhlinin 2007 July2011, Boston

Previous Attempts Suzaku Limits ASCA Perseus: Dupke and Bregman (2001) claimed 4100 (+2200, -3100) km/s, but not confirmed by later study (Ezawa et al. 2001) Centaurs: Dupke and Bregman (2001) claimed 1600 ± 320 km/s, but not confirmed by Suzaku 3000 km/s There are some attempts previously. Using ASCA, Dupke and Bregman pionearlingly suggested … Not confirmed by later study with the same data. They also suggested similar variation Centaurus cluster. Some Suzaku observations have improved these measurements. One of the first report is from Centaurus, Ota et al. put strong limit on the possible velocity shift. They divided image into 8x8 cells and measure redshifts from each region and find no significant variation. There are other reports on the limits but not detection so far. Suzaku Limits Centaurus: Δv < 1400 km/s (Ota et al. 2007) July2011, Boston

Velocity shift vs. Energy Resolution (km/s) E/ΔE Shift @ Fe-K line 300 1000 0.1 % = 7 eV 0.3 % = 20 eV CCD energy resolution (FWHM) 60 120 eV July2011, Boston

Suzaku XIS energy gain calibration Accuracy of the XIS energy scale in % (0.1% = 7eV = 300 km s-1) -0.1 0.1 0.2 0.3 0.4 0.5 Mn I Ka, time-averaged accuracy of the absolute energy (Ozawa et al. 2009) Mn I Ka, during the A2256 observation. Fe-K, Galactic center (Koyama et al. 2007) Fe-K, Perseus CL. Variation among 2’x 2’ cells (Ota et al. 2007) Fe-K, Perseus CL. Difference between the two regions. Mn I K lines from the Cal. source a good calibration of the instrument, SUZAKU XIS, is a key in our analysis. Here we summaries accuracies of the energy gain studied in a various way. There are some calibration papers from the instrumental team and they reported that the uncertainty is about +-0.1% based on Mn line from the calibration source from a continus monitoring as shown here. We have also checked cal-source energy during the A2256 observation. Similar levels of accuracy are reported from GC and the Perseus clusters. This range indicates the observed shift. It is clear that compared with the cal accuracy, this detected signal is large. I will explain how we can get this shift. Observed shift in A2256 and statistical error Good calibration is a key July2011, Boston

A2256, X-ray bright, double peaked merging cluster SUB Peak SUB MAIN ΔV 〜 2000 km/s Here is our target. By looking into several SUZAKU data, I believe that this could be the best one to search the bulk motion. That is because X-ray bright Two peaks in X-ray brightness separated about 3’ Galaxy velocity distribution also shows two peaks separated by 2000 km/s The question is whether if we can detect this velocity shift in X-ray emitting gas. MAIN Peak Berrington et al. 2002 (galaxy radial velocities) Sun et al. 2002 (Chandra) July2011, Boston

Suzaku Fe-line Image of A2256(1) He-like Fe band H-like Fe band Sub, cooler In order to examine the spectral variation over the A~2256 central region, we extracted two images in different energy bands including He-like and H-like iron line emission as shown in figure There appear at least two emission components which corresponds to the main cluster at east and the sub component at west discovered by ROSAT. We noticed a clear difference in the distribution between the two images. In the He-like iron image, the sub component exhibits brightness comparable to that of the main component. On the other hand, in the H-like iron image, the sub has lower brightness. This clearly indicates that the sub has cooler emission compared with the main. Based on these contrasting spatial distributions, we define centers of the two components and extract spectra from these two regions. Main July2011, Boston

Suzaku Results on A2256 (2) Main, 7.1 keV Sub, 5.0 keV Here are examples of spectral fits. We use the energy range of the 5.5--7.3~keV around the iron line complex. We use a CIE component to model the spectra. There are three CCD data shown in different colors. Models for the three XISs are assumed to have different redshifts to compensate inter-CCD calibration uncertainties. Other parameters, temperature and abundance, are fixed to be common among the CCDs. July2011, Boston

Redshifts from X-ray and Optical Statistical χ2 Sub Redshift Sub Left panel shows obtained z of the main and sub components from three CCD. There are some difference among the CCD, but in all cases there are shift in the redshift. Right: we tried to correct these CCD-to-CCD gain, combined the 3 CCD spectra. These are chi^2 distribution as a function of redshift. These dashed lines indicate redshift from optical galaxies, which is independent of X-ray data. Redshift July2011, Boston

Sub Region Spectra Δv fixed to 0 Best fit redshift Χ2/d.o.f. = 40.4/42 To visualize the redshift shift in the sub region more directly, we show spectral fittings by combining the two front-illuminated XIS (0 and 3) data in these plots Left free redshift, acceptable Right redshift is fixed to the main com. Redshift. This comparison shows that we can reject that the sub region has the same redshift as the main cluster. These are results by which we believe we detected the gas motion directly for the first time in clusters. Χ2/d.o.f. = 40.4/42 July2011, Boston Χ2/d.o.f. = 80.8/43

Interpretations (1) X-ray mass estimation Departs from hydrostatic equilibrium around the sub component. Need to consider to weight the total cluster mass. (No significant effect on the mas of the primary) (2) A new method to study the gas dynamics. Complementary with X-ray imaging studies. (3) Merger state in A2256 Before the final collision July2011, Boston

Discussion : Dynamics of A2256 (1) The X-ray mass estimation * The determined radial velocity, Vr ~1500 km/s, gives an lower limit of V = Vr/sinα, this velocity corresponds to a Mach number M> 1.4. α（ the angle between the motion and the plane of the sky) * Around the sub component, the gas kinematic pressure (or energy) can be 1.42 ~ 2 x larger than the thermal one. →　the gas departs from hydrostatic equilibrium. * does this motion affect the estimation of the mass of the primary cluster ? →　depends on the physical separation between the two components. * In the case of A2256, the two are not likely too closely connected to disturb the hydrostatic condition around the primary. * However, to weight the total mass within the larger volume including the sub component we should consider not only the mass of the sub itself but also the dynamical pressure associated with the relative motion. July2011, Boston

(2) Methods to measure ICM Shock/Cold front Assumptions: adiabatic flow Velocity in the plane of the sky →　rare case Requires high spatial resolution Shock in 1E~0657-56, M~3±0.4 (Markevitch et al. 2002). Cold front in A3667, M~1±0.2 (Vikhlinin et al. 2001). X-ray Doppler shift Direct measurement of the radial velocity Require high energy spectral resolution In A2256, M>1.4　 Combination: These two are complementary. If applied to a merging system simultaneously, →　 a direct measurement of three dimensional motion July2011, Boston

(3) Merger state in A2256 Suzaku Measurement: Vr = 1500 km/s Cold front condition (Sun et al. 2002): Vsky ~ 1500 km/s → V ~ 2000 km/s Assume the two system start from rest. → V ~ (2GMtot/R)1/2 , Mtot ~ 8x1014 Msun Time to the collision at R=1 Mpc →　(2-1 Mpc)/ V ~ 0.4 Gy “Before the final collision”, Consistent with no strong disturbance in the X-ray structure July2011, Boston

ASTRO-H, an International X-ray Observatory - To be launched in 2014 from Japan. - SXS is a key instrument to reveal the large scale structure and its evolution of the Universe. (check astro-h.isas.jaxa.jp) Soft X-ray Spectrometer (SXS; microcalorimeter) July2011, Boston

ASTRO-H/SXS simulation (1) the two components in A2256 SXS 100 ks Using the SXS with an energy resolution better than 7eV, we could measure gas bulk motions in a fair number of X-ray bright clusters. Note the Suzaku-observed energy shift is about 30 eV ~ 1500 km/s. July2011, Boston

SXS simulation (2) The brightest cluster core: The Perseus Detect and locate the gas turbulence. Combined with hard X-ray imaging, gas dynamics, particle acceleration, shocks and non-thermal processes will be investigated. A-H, which will be launched in a few years. We can … July2011, Boston

Summary X-ray Doppler mapping of the ICM is a next major step to study the cluster dynamics. Suzaku observation of the merging system A2256 demonstrated this. A significant shift of the redshift of the sub component was detected. The gas moves is pair with galaxies. Bulk motions and turbulences will be measured by the ASTRO-H (SXS) more robustly and in a systematic way. Double peaked in X-ray image and galaxy velocity distribution Δv is ～ 2000 km s-1. X-ray: Δv=1500 ±300(sta.) ±300(sys.) km s-1. See Tamura et al. 2011, PASJ, arXiv:1104.2667 July2011, Boston 19

Supplements July2011, Boston

Interactions in clusters focusing on the ICM SZ Stars in galaxies CMB Galaxy motion, Winds Thermal X-ray X-ray emission and absorption ICM Metals Accretion, cooling flow? WHIM, Filament, Voids Thermal conduction Ionization Magnetic Field IC (X-ray) Synchrotron Jet, bubble shock Black Hole in AGN Cosmic rays Turbulence Gas Motion acceleration Here is a map of interplay of physics in a cluster, focusing on the gas dynamics. The gas bulk motion and turbulence are mainly driven by the gravity of dark matter. The gas dynamics should tell something about dark matter. For example, how dark matter distribute, or interact with each other, and most importantly its nature are questions to be answered. The gas dynamics in turn accelerates cosmic-ray at shocks. There non-thermal processes are observed in radio synchrotron emission and will be observed by hard X-ray imaging. In addition, the central AGN could develop different kinds of gas dynamics around the cluster center. In any cases, we need to measure how much energies are in kinematic form in the gas. In short, without gas dynamic measurements, we can not draw the picture of formation and evolution of the large scale structure in the universe. Keeping these motivations, we try to detect the gas motion. G-lensing Interactions in clusters focusing on the ICM Dark matter heating Observation Interaction July2011, Boston

Spectral Resolution and Features 1000 km/s CCD This plot shows velocity shifts and other spectroscopic features and spectral resolution to resolve those features. Here typical CCD resolution is shown. This is 1000 km/s And 1000 km/s with 100 photons. This means that if we have more than 100 photons we can detect velocity structure larger than 1000 km/s. Of course, in this comparison, No systematic uncertainty is considered. 1000 km/s (100 photons) CCD CCD July2011, Boston Mitsuda et al. 2010

Suzaku, ASTRO-H, and IXO Spectrometers Suzaku XIS ASTRO-H SXS (best estimate) IXO XMS Effective Area cm2 @ 7 keV 460 220 6000 FOV 18’ x 18’ 3’ x 3’ 5’ x 5 ’ Spatial Resolution, HPD 120” 90” 5” Energy Resolution E/ΔE @ 7 keV 50 1000 3000 So how we can improve our measurements with feature missions. Here is comparison of ins. Performances among SUZAKU/XIS, A-H/SXS, IXO/XMS. The A-H/SXS will be the 1st calo. In orbit providing significant improvement of spectral resolution. Following A-H, IXO should give us extreme improvements not only in spectral resolution but also in area, fov, spatial resolutions. Using these instruments, How we can improve our understanding ? July2011, Boston