MeerKAT detects large-scale clustering of cosmic hydrogen using 21-cm intensity mapping

The MeerKLASS Collaboration has reported its first measurement of how neutral hydrogen clusters on large scales in the distant universe. Using the South African MeerKAT radio telescope, the team detected a signal that follows the structure of the cosmic web, at a significance of more than six times the statistical noise. The study, led by Matilde Barberi-Squarotti and colleagues, appears in Monthly Notices of the Royal Astronomical Society.

Mapping hydrogen without resolving galaxies

The result rests on 21-cm intensity mapping. Rather than picking out individual galaxies, the method adds up the radio emission from neutral hydrogen across large patches of sky. The collective pattern traces how matter is distributed, even when single galaxies are too faint to detect. In principle, this allows three-dimensional maps of the universe to be built quickly over very large volumes.

The team analysed MeerKAT observations covering roughly 250 square degrees, comparable to the area of 1,250 full moons. The detected signal dates from about 4.5 billion years ago, when the universe was a little over 9 billion years old.

A faint signal beneath bright foregrounds

The main difficulty is that astrophysical foreground emission is about 10,000 times stronger than the hydrogen signal. Isolating the cosmological component therefore demands careful modelling and removal of contaminants, as well as control of instrumental effects.

Dr Steven Cunnington, a senior research fellow at the University of Portsmouth's Institute of Cosmology and Gravitation, led development of the analysis pipeline that made the detection possible. He is also co-leading one of the largest working groups in the collaboration.

A different way of using the telescope

The measurement depended on a new observing mode. Instead of combining the 64 MeerKAT dishes into a single interferometer, the team used each dish as an independent telescope. This sacrifices angular resolution but greatly increases survey speed over wide areas.

The approach yielded many independent maps, made with different dishes at different times. Genuine cosmic structure should appear in all of them, whereas instrumental artefacts that vary between dishes and observations are suppressed when the maps are cross-compared. According to the researchers, combining the maps was central to reducing systematic errors and confirming the detection.

Why it matters and what comes next

Professor Mário Santos of the University of the Western Cape, principal investigator of MeerKLASS, described the result as a milestone in establishing 21-cm intensity mapping as a cosmological tool. The work involved researchers in seven countries across three continents and took about ten years, and it relied on the South African Radio Astronomy Observatory, which built and operates MeerKAT.

For cosmologists, the detection shows that the technique can recover large-scale structure from real data despite the foreground and systematics challenges. The source does not describe specific follow-up analyses. A natural next step would be to use such measurements to constrain cosmological models, though that would need larger datasets and further control of residual systematic effects.

The detection is a first measurement rather than a finished cosmological survey, and it covers a single epoch and a limited sky area. Even so, it provides a working demonstration of a method that could map the universe over far larger volumes than galaxy-by-galaxy surveys allow.

Source: Cosmic hydrogen measurement marks breakthrough in mapping the universe (Phys.org, 5 November 2026).

Image: AI-generated illustration.