Joint measurement of the galaxy cluster pressure profile with Planck and SPT-SZ

We measured the average Compton profile of 461 clusters detected jointly by the South Pole Telescope (SPT) and Planck. The number of clusters included in this analysis is about one order of magnitude larger than in previous analyses. We propose an innovative method developed in Fourier space to combine optimally the Planck and SPT-SZ data, allowing us to perform a clean deconvolution of the point spread and transfer functions while simultaneously rescaling by the characteristic radial scale R₅₀₀ with respect to the critical density. The method additionally corrects for the selection bias of SPT clusters in the SPT-SZ data. We undertake a generalised Navarro-Frenk-White (gNFW) fit to the profile with only one parameter fixed, allowing us to constrain the other four parameters with excellent precision. The best-fitting profile is in good agreement with the universal pressure profile based on REXCESS in the inner region and with the Planck intermediate Paper V profile based on Planck and the XMM-Newton archive in the outer region. We investigate trends with redshift and mass, finding no indication of redshift evolution but detecting a significant difference in the pressure profile of the low- versus high-mass subsamples, in the sense that the low mass subsample has a profile that is more centrally peaked than that of the high mass subsample. We also scaled the average Compton profile by the mean Universe density (R_200m) and provide the best-fitting gNFW profile. Using the profiles scaled by both the critical (R₅₀₀) and the mean Universe density (R_200m), we studied the outskirt regions by reconstructing the average Compton parameter profile in real space. These profiles show multiple pressure drops at θ > 2θ₅₀₀, but these cannot clearly be identified with the accretion shocks predicted by hydrodynamical simulations. This is most probably due to our having reached the noise floor in the outer parts of the average profile with the current data sets.