Lensing-like tensions in the Planck legacy release

TL;DR

This paper analyzes Planck data to investigate a lensing-like anomaly, finding a persistent 2-3 sigma tension with expectations that cannot be fully explained by current models or additional measurements.

Contribution

It provides a model-independent analysis of lensing in Planck data, revealing a consistent anomaly that challenges the standard cosmological model and its extensions.

Findings

Lensing potential is 2σ to 2.8σ higher than expectations.

The anomaly persists even after including BAO and supernova data.

Constraints on neutrino mass are only marginally affected.

Abstract

We analyze the final release of the Planck satellite data to constrain the gravitational lensing potential in a model-independent manner. The amount of lensing determined from the smoothing of the acoustic peaks in the temperature and polarization power spectra is 2$\sigma$ too high when compared with the measurements using the lensing reconstruction and 2.8$\sigma$ too high when compared with $\Lambda$CDM expectation based on the "unlensed" portion of the temperature and polarization power spectra. The largest change from the previous data release is the $\Lambda$CDM expectation, driven by improved constraints to the optical depth to reionization. The anomaly still is inconsistent with actual gravitational lensing, given that the lensing reconstruction constraints are discrepant independent of the model. Within the context of $\Lambda$CDM, improvements in its parameter constraints from lensing reconstruction bring this tension to 2.1$\sigma$ and from further adding baryon acoustic oscillation and Pantheon supernova data to a marginally higher 2.2$\sigma$. Once these other measurements are included, marginalizing this lensing-like anomaly cannot substantially resolve tensions with low-redshift measurements of $H_0$ and $S_8$ in $\Lambda$CDM, $\Lambda$CDM+$N_\mathrm{eff}$ or $\Lambda$CDM+$\sum m_\nu$; furthermore the artificial strengthening of constraints on $\sum m_\nu$ is less than 20%.