Research > Projects > Exploring intergalactic magnetic fields to unveil the cosmic web

Cosmic web filaments connecting galaxy clusters are fundamental structures in the Universe. Yet, understanding their thermal and non-thermal constituents remains a challenge. We propose using magnetic fields to explore the warm-hot intergalactic medium within filaments, starting with the one between Abell 3667 and Abell 3651, recently observed by eROSITA. Leveraging Faraday rotation measures from the Polarization Sky Survey of the Universe Magnetism (POSSUM), we will gain insights into intergalactic magnetic fields and explore new methods for detecting filaments with radio data. This project lays the groundwork for exploiting forthcoming radio surveys by testing machine-learning algorithms in data analysis.

Planned activities:
• find signatures of magnetic fields in the Abell 3667-51 filament. The filament region has already been observed by EMU and POSSUM and the applicant has asked and obtained to lead the analysis of this field within the POSSUM collaboration. The Abell 3667-51 system is a perfect test study, because using the thermal gas density derived from X-ray analysis, we can infer IGMF strength. In the case of detection with the POSSUM RM grid, we will measure the IGMF in the first individual filament ever detected extending outside the virial radius of galaxy clusters. In the case of a non-detection, we will impose upper limits on IGMF’s strength. These values will be compared with different magnetogenesis models to test their predictions. Furthermore, we will search for direct emission from the filament, using both total intensity and polarization;
• development of machine learning (ML) tools to detect filaments in radio surveys. POSSUM will cover the entire Southern sky and the partial overlap with the SRG/eROSITA All-Sky Survey area makes it possible to search for new targets in the X-ray superclusters catalog recently released. Manual cataloging methodologies are anticipated to be insufficient in the era of these big surveys. Hence, with this project we will extend the search to other cosmic filaments, integrating classical methods with new ML techniques applied to survey data. In particular, we will use and further develop Radio U-Net, a fully convolutional U-Net architecture, which performs the automated segmentation of radio images to detect diffuse structures. This novel tool has shown promising results in terms of accuracy and computational efficiency in detecting diffuse emissions in clusters. Radio U-Net was specifically designed for total intensity low-frequency data. We will assess its effectiveness in analyzing ASKAP EMU data and explore its potential for detecting filaments, possibly also using POSSUM polarization intensity images.

Start: 15/05/2025
Duration: 2 years
Financing program: Ricerca fondamentale INAF
Referent: Chiara Stuardi (INAF)