Cosmological constraints from thermal Sunyaev–Zeldovich power spectrum revisited

Abstract

Thermal Sunyaev-Zeldovich (tSZ) power spectrum is one of the most sensitive methods to constrain cosmological parameters, scaling as the amplitude $\sigma_8^8$. It is determined by the integral over the halo mass function multiplied by the total pressure content of clusters, and further convolved by the cluster gas pressure profile. It has been shown that various feedback effects can change significantly the pressure profile, possibly even pushing the gas out to the virial radius and beyond, strongly affecting the tSZ power spectrum at high $l$. Energetics arguments and SZ-halo mass scaling relations suggest feedback is unlikely to significantly change the total pressure content, making low $l$ tSZ power spectrum more robust against feedback effects. Furthermore, the separation between the cosmic infrared background (CIB) and tSZ is more reliable at low $l$. Low $l$ modes are however probing very small volumes, giving rise to very large non-gaussian sampling variance errors. By computing the trispectrum contribution we identify $90<l<350$ as the minimum variance scale where the combined error is minimized. Using the measurement at this $l$ provides constraints that are almost as strong as using the entire power spectrum. We extend the work by the Planck collaboration by including a full trispectrum and feedback effects in the analysis. We perform a Markov-Chain Monte Carlo analysis over the two dimensional parameter space and find constraints on $\sigma_8$ by marginalizing over the feedback nuisance parameter.We obtain $\sigma_8 =0.820^{+0.021}_{-0.009}\left(\Omega_m/0.31\right)^{0.4}$, when fixing other parameters to Planck cosmology values. Our results suggest it is possible to obtain competitive cosmological constraints from tSZ without cluster redshift information, and that the current tSZ power spectrum shows no evidence for a low amplitude of $\sigma_8$.