A comparative test of different pressure profile models in clusters of galaxies using recent ACT data

TL;DR

This work tests whether a single universal radial pressure profile can describe the intracluster medium across a large cluster population by comparing four parametric forms using ACT SZ data. It adopts a population-level approach, stacking 3,496–4,946 clusters across mass and redshift ranges to measure angular Compton-y profiles and fits them with four pressure prescriptions (gNFW/UPP, β-model, polytropic, and exponential universal). All models reproduce the measured profiles within uncertainties and show no decisive preference, though hints of higher amplitudes and steeper outskirts arise for high-mass, low-redshift systems, suggesting residual population-level trends and a breakdown of strict universality at high precision. The study underscores the limitations of SZ-only data for constraining universal pressure models and highlights that physically motivated alternatives like BMP and EUP perform comparably to the canonical gNFW form, motivating joint analyses with X-ray data or larger SZ datasets for tighter discrimination.

Abstract

Context. The electron pressure profile is a convenient tool to characterize the thermodynamical state of a galaxy cluster, with several studies adopting a "universal" functional form. Aims. This study aims at using Sunyaev-Zel'dovich (SZ) data to test four different functional forms for the cluster pressure profile: generalized Navarro-Frenk-White (gNFW), $β$-model, polytropic, and exponential. The goal is to assess to what level they are universal over a population-level cluster sample. Methods. A set of 3496 ACT-DR4 galaxy clusters, spanning the mass range $[10^{14},10^{15.1}]\,\text{M}_{\odot}$ and the redshift range $[0,2]$, is stacked on the ACT-DR6 Compton parameter $y$ map over $\sim13,000\,\text{deg}^2$. An angular Compton profile is then extracted and modeled using the theoretical pressure recipes, whose free parameters are constrained against the measurement via a multi-stage MCMC approach. The analysis is repeated over cluster subsamples spanning smaller mass and redshift ranges. Results. All functional forms are effective in reproducing the measured $y$ profiles within their error bars, without a clearly favored model. While best-fit estimates are in broad agreement with previous findings, hints of residual subsample dependency are detected favoring higher amplitudes and steeper profiles in high-mass, low-redshift clusters. Conclusions. Population-level cluster studies based on SZ data alone are likely unable to accurately constrain different pressure profile models. Residual trends at population level and scatter at individual cluster level undermine the universal pressure model assumption whenever high precision is required. Finally, functional forms different from the gNFW prove equally effective while being more physically motivated.

Figures (16)

• Figure 1: Visual representation of the data set employed in this study. The ACT DR6 Compton parameter map is plotted on an equirectangular projection, the white region highlighting the mask footprint. Blue spots show the positions of the 4195 clusters from the ACT DR5 catalog.
• Figure 2: Left and middle panel show, respectively, the distribution in mass and redshift of the full cluster sample adopted in this study. The right panel shows instead the catalog scatter across these two quantities. In all panels, dashed lines mark the edges of the chosen bins in the corresponding quantity.
• Figure 3: Summary plot showing the stacks on ACT $y$-map of all the 16 cluster samples described in Sec. \ref{['ssec:samples']}, with labels clarifying the associated mass and redshift intervals. Although the final stack maps are $30'\times30'$ in size, this figure zooms in the central $12'\times12'$ square to better show details in the cluster signal.
• Figure 4: Estimation of the mean background level in the stacks. For each mass and redshift bin, the colored bundle is the set of all 500 profiles extracted from the shuffled stacks; the solid darker line shows instead their average, taken as the zero-level profile, and the black dashed lines the 68% confidence interval around the mean.
• Figure 5: Correlation matrices associated to the reconstructed Compton parameter profiles over the relevant angular span, for all mass and redshift bins.