(Notes from K. Basu)

The aim was to drive the discussion along the different observational signatures of the various physical processes in the cluster outskirts (r > r_500) affecting the X-ray and SZ measurements. An important question is how to disentangle these effects with the current observational capabilities.

Today's discussion focused primarily on clumping (as could be expected!), as several of the participants were not present during the conference where Suzaku results for Perseus cluster were presented. The main points on this discussion were:

  1. Simulators agree it is difficult to get clumping of the order ~10 inside r_200, what is seen in the Suzaku results. The current simulations show large scatter and very small statistic. Suzaku results of Perseus showing C = (n_obs / n_true)^2 ~ 10-20 near r_200.
  2. S. Molendi pointed out the issue with soft X-ray background modeling near Galactic plane, which can make Suzaku results biased. They have shown the comparison with the ROSAT PSPC measurement with Suzaku in a recent paper (Eckert, Moelndi et al., arXiv:1103.2455), for PKS0745. Although the Suzaku works claims consistency with the ROSAT measurement of Perseus from Ettori, Fabian & White 1998, MNRAS. (Update from A. Simionescu: The Suzaku analysis of Perseus cluster included a hot foreground component (kT ~ 0.6 keV) for the Galaxy due to its low Galactic latitude. They estimate the local background/foreground component from Suzaku+ROSAT data, so the issue raised in Eckert et al. paper should not apply.)
  3. It was pointed out that the Suzaku measurement of increasing f_gas (which is the main motivation behind invoking clumping argument) also depends on the extrapolation of the NFW profile to the outskirts (near r200 and beyond). The validity of this has been argued over several times in this workshop.
  4. It is agreed that HSE mass bias is only a small fraction of what Suzaku is showing.

Other highlights from today's discussion:

  • A. Bykov raised the interesting possibility that the surface excess can be a result from non-thermal X-ray emission (see e.g. Bykov et al. 2008, ApJ). Their simulations for SN remnants has similar magnetic fields and shock Mach numbers as can be expected in the cluster outskirts, although densities in the outskirts are much lower. The expected gas cooling time is ~1 Myr for the shocked gas. One example of both thermal and non-thermal X-ray emission with stochastic magnetic fields is Cas A. The power of the synchrotron emission is roughly nu^2-3.

  • C. Pfrommer discussed his latest simulation results on gas clumping in the outskirts, where he also sees significant pressure clumping. The density and pressure clumping are of very similar magnitude, which will mean there should be no steepening of the temperature profile. It is unclear right now what relates the density and pressure clumping, or why they are similar magnitude. The pressure clumping appears to extend towards more inner regions. The recent work by Nagai & Lau (2011, ApJ) does not mention pressure clumping, and their sample size is small with large scatter. UPDATE: A more comprehensive study of the scatter in the gas density and X-ray brightness due to clumping is presented by Vazza, Roncarelli, Ettori & Dolag, arXiv:1011.5783.

  • SZ vs. X-ray test to detect gas climb will fail to deliver unambiguous answer to this problem if there is pressure clumping. One possibility raised by K. Basu would be to measure X-ray surface brightness in some hard energy band so that it goes roughly as ~T^2, where T is the gas temperature. In such case the ratio of X-ray brightness to SZ squared will give the pressure clumping. But it will be very difficult to get a measurement of the X-ray brightness above 4 keV (exponential fall-off + diminishing effective area + stronger background), particularly from the low temp cluster outskirts.


(18 Apr. 2011)

In today's discussion we started with the Planck results and the dependence of Planck's Ysz measurements on the cluster pressure profile, especially in the outskirts. J. Bartlett gave a detailed summary of the matched filtering technique used by Planck. The main points in this discussion were:
  1. The Planck algorithm assumes the Arnaud et al. 2010 universal pressure profile to match SZ clusters. The scale radius r_s is varied till the maximum in S/N is reached, thereafter r_s is kept fixed and fitted for the peak amplitude P0. (But we know P0 and r_s are degenerate! M. Voit also pointed out that since we're trying to maximize the S/N in an aperture, the size will always be upwards biased.)
  2. A particular problem for Planck is that most galaxy clusters are roughly the size of its beam, or slightly larger. This means the analysis is not simple as finding point source amplitudes, and also getting a handle on the cluster shapes is difficult since it's only marginally resolved.
  3. Planck cluster detection algorithms were previously tested on WMAP data (ref. Melin et al. 2011), but its even worse resolution makes checking the detailed dependence on the assumed pressure profile more difficult.

  • M. Voit pointed out a good exercise for Planck would be to take selected high S/N resolved clusters (e.g. Perseus) and reconstruct the SZ signal out to r_200 without a parametric model or mask, just bin-by-bin reconstruction. This is already what is being attempted by ground based bolometer experiments (e.g. APEX-SZ or BOLOCAM), but the unique spectral coverage and low noise should allow Planck to go to larger radii. (The contribution of SZ shot noise, even for Planck beam, is much smaller than the instrumental white noise. Only issue is a careful subtraction of CMB and possible Galactic foregrounds.)

  • We discussed the relative contribution of electron-ion non-equilibrium as compared to cosmic ray injection of non-thermal electrons in the cluster outskirts, to cause a reduction of the SZ signal. D. Rudd pointed out the strong temperature dependence on the equilibrium timescale, for small mass halos the effect of non-equilibrium is negligible. This can be in contrast to the cosmic rays contribution, but the percentage of high energy electrons in any cosmic ray models is negligible to cause any impact (CR are mostly protons).

  • We discussed the prospect of X-ray missions to measure the thermal line widths in X-ray emission lines directly, to measure T_ion and check for non-equilibrium against T_e. Fox & Loeb (1992) estimated that for the 7 keV Fe line one would need delta E / E ~ 1000 to measure the line width. The proposed Astro-H mission promises a 4 eV resolution across its spectrum, so measuring this thermal broadening for the 7 keV line should be possible.

  • Any other wavelength absorption line through the cluster outskirts should also be able to measure in principle the T_ion through line broadening. Of particular interest is high resolution UV spectroscopy from Hubble's Cosmic Origins Spectrograph. But M. Voit pointed out that there are simply not enough bright sources (~100 sources in the whole sky!) to do such high resolution spectroscopy. See this document, page 3.

One remark concerning the flattening of entropy profile in the cluster outskirts was that all Suzaku measurements are limited to high temperature clusters. No entropy flattening has been measured in low mass systems.

Update from D. Nagai concerning the results from Nagai & Lau (2011, ApJ): their analysis has now been extended to a sample of more than 3000 clusters/groups. The current analysis includes evidence for pressure clumping. The amplitude and onset radius for both density and pressure clumping are very similar.