Searching for the Synchrotron Cosmic Web Again: A replication attempt

TL;DR

This study attempts to replicate the detection of the synchrotron cosmic web using new, more sensitive low-frequency radio observations but finds no significant evidence of the predicted emission, challenging previous claims.

Contribution

The paper provides a rigorous replication of prior detection claims with improved data sensitivity and finds no supporting evidence, highlighting potential issues in earlier results.

Findings

No statistically significant detection of excess radio emission along LRG pairs.

Reproduction of excess X-ray emission from ROSAT data.

Failure to reproduce extended intercluster filamentary emission in original data.

Abstract

We follow up on the surprising recent announcement by Vernstrom et al. (2021) of the detection of the synchrotron cosmic web. We attempt to reproduce their detection with new observations with the Phase II, extended configuration of the Murchison Widefield Array at \SI{118.5}{\mega \hertz}. We reproduce their detection methodology by stacking pairs of nearby luminous red galaxies (LRGs) -- used as tracers for clusters and galaxy groups -- contained in our low frequency radio observations. We show that our observations are significantly more sensitive than those used in Vernstrom et al., and that our angular sensitivity is sufficient. And yet, we make no statistically significant detection of excess radio emission along the bridge spanning the LRG pairs. This non-detection is true both for the original LRG pair catalogue as used in Vernstrom et al., as well as for other larger catalogues with modified selection criteria. Finally, we return to the original data sets used in Vernstrom et al., and find that whilst we clearly reproduce the excess X-ray emission from ROSAT, we are not able to reproduce any kind of broad and extended excess intercluster filamentary emission using the original 118.5 MHz MWA survey data. In the interests of understanding this result, as part of this paper we release images of the 14 fields used in this study, the final stacked images, as well as key components of our stacking and modelling code.