MeerKAT Radio Telescope Detects Cosmic Hydrogen Signal Independently for First Time at Dark-Energy Scales

After a decade of painstaking work, an international team has cleared the most significant technical hurdle in radio cosmology: they have directly measured the statistical clustering of cosmic neutral hydrogen — without any help from optical galaxy catalogs — at the large scales where dark energy and the geometry of spacetime leave their most distinctive imprints. The result, published today in MNRAS by the MeerKLASS Collaboration, represents the first confirmed detection of the hydrogen 21-cm auto-power spectrum at genuinely cosmological scales. It is not merely a new data point. It is a proof of concept for a new way of mapping the Universe — one that the multi-billion-dollar Square Kilometre Array Observatory (SKAO) is counting on.

What Was Detected — and Why the Scales Matter

The story centers on neutral hydrogen's signature radio emission, known as the 21-centimeter (about 8.3 inches) line. When a hydrogen atom's electron flips its spin, it releases a photon at exactly this wavelength — a property so reliable that physicists call the 21-cm line nature's most precise clock. Because the Universe is expanding, this signal is stretched to longer wavelengths during its journey to Earth, encoding in its arrival frequency the age of the Universe at the moment of emission.

The MeerKLASS team detected hydrogen light emitted roughly 4.5 billion years ago, when the Universe was just over nine billion years old. What they extracted from MeerKAT's data is not an image of individual hydrogen clouds or galaxies — the signal is far too faint for that at cosmological distances. Instead, they measured the auto-power spectrum: a statistical description of how the collective hydrogen glow varies from place to place across the sky, capturing the web-like skeleton of large-scale cosmic structure.

The detection achieved a signal-to-noise ratio of 6.2 after surviving a battery of robustness tests — a threshold the community considers a firm detection.

The specific scales detected — covering a Fourier wavenumber range of 0.095 to 0.245 h/Mpc⁻¹, corresponding to comoving separations of roughly 25 to 65 million light-years — are scientifically critical. The signal from baryon acoustic oscillations (BAOs), the frozen imprint of primordial sound waves that serves as cosmology's standard measuring ruler, is encoded right at the lower end of this range. No previous standalone hydrogen detection had reached these scales. The Canadian Hydrogen Intensity Mapping Experiment (CHIME), which made the first standalone auto-spectrum detection in a 2025 paper, did so at smaller, non-linear scales where dark energy signatures are weaker and BAO features are washed out. The MeerKLASS result pushes the field into the domain where the most scientifically valuable measurements live.

"This is a milestone in establishing 21-cm intensity mapping as a novel tool for cosmology," said Prof. Mário Santos of the University of the Western Cape, South Africa, the project's scientific lead.

Why "No Galaxy Catalog" Is the Headline

That phrase — without any help from optical galaxy catalogs — is what elevates the result above several earlier milestones.

Intensity mapping researchers have long used a workaround: they cross-correlate their radio maps with independently compiled optical surveys of visible galaxies. If the radio intensity fluctuations match the locations of real galaxies, the cross-correlation confirms the presence of a hydrogen signal even when the radio data alone are too noisy to claim a detection. The method has worked well, and MeerKLASS itself published multiple cross-correlation detections over the past several years.

But a cross-correlation detection is ultimately a crutch. It demonstrates that hydrogen is there, but it does not prove that the auto-power spectrum — the statistical fingerprint derived from the hydrogen maps alone — is free from instrumental systematics. A standalone auto-spectrum detection requires the radio data to tell the whole story by itself.

The MeerKLASS team achieved this by applying what might be called a time-split nulling technique. Rather than measuring the power spectrum of a single combined map, they divided the full dataset into independent subsets — different antenna groups, different observation epochs — and computed the cross-spectrum between pairs of subsets. The key insight: the cosmic hydrogen signal is present in every subset, because it is the same Universe being observed at the same frequency. Instrumental noise and time-varying systematic artifacts, by contrast, are uncorrelated between subsets and vanish in the cross-spectrum. The cosmological signal survives; the noise cancels. The resulting detection, at a signal-to-noise ratio of 6.2, comes from the hydrogen radio maps and the hydrogen radio maps alone.

Battling a Million-to-One Foreground

The deepest technical challenge in 21-cm intensity mapping is not sensitivity — it is contamination. At the same radio frequencies as the redshifted hydrogen signal, other astrophysical sources shine enormously brighter. The Milky Way's synchrotron radiation and extragalactic radio sources can outshine the cosmological hydrogen signal by a factor of 100,000 to one million. Isabella Paola Carucci of INAF's Trieste Astronomical Observatory, a co-author on the paper, described isolating the signal as like hearing a whispered conversation during a concert.

Separating signal from foreground requires removing spectrally smooth components — the Milky Way's synchrotron glow is spectrally smooth; the cosmological hydrogen signal is not — using techniques such as principal component analysis (PCA). Crucially, this cleaning inevitably removes some of the cosmological signal as well. The team quantifies and corrects for this signal loss using foreground transfer functions, derived by injecting mock signals into the real data and measuring how much is removed by the cleaning process. Getting this correction right at cosmological scales, without over-subtracting signal, was a key technical achievement of the decade-long program.

The Data and the Team

The measurement used MeerKAT observations made in 2021 covering approximately 236 square degrees of sky — roughly half the apparent area of the Orion constellation — using frequencies between 971 and 1,024 MHz. Mário Santos of the University of the Western Cape is the program's scientific lead. The collaboration's Italian contributors include Matilde Barberi-Squarotti (Università Statale di Milano), Stefano Camera (Università di Torino), and Isabella Paola Carucci (INAF Trieste). Dr. Steven Cunnington, a Stephen Hawking Fellow at Portsmouth, led development of the analysis pipeline. The work involved researchers across multiple countries and three continents, and the South African Radio Astronomy Observatory (SARAO), which built and operates MeerKAT, provided essential operational and calibration support.

The full author list: MeerKLASS Collaboration, Matilde Barberi-Squarotti, José L. Bernal, Philip Bull, Stefano Camera, Isabella P. Carucci, Zhaoting Chen, Steven Cunnington, Brandon N. Engelbrecht, José Fonseca, Keith Grainge, Wenkai Hu, Melis O. Irfan, Yichao Li, Mário G. Santos, Marta Spinelli, Jingying Wang, Amadeus Witzemann, and Laura Wolz.

Where This Fits: Fifteen Years of Incremental Progress

The 21-cm intensity mapping field has been building toward this result for fifteen years. Early work in 2009–2010 — using the Green Bank Telescope (GBT) in West Virginia and the Parkes Observatory in Australia — produced the first faint hints of cosmic hydrogen structure, but only via cross-correlation with optical surveys. CHIME's 2025 auto-spectrum detection was a genuine milestone, but at non-linear scales where the most valuable cosmological signals are not accessible.

The MeerKLASS result pushes the detection into genuinely cosmological territory: the domain where BAO features and dark energy signatures imprint themselves on the matter power spectrum. That distinction matters because the whole scientific case for large-scale hydrogen intensity mapping rests on the ability to measure these large-scale modes. A detection at non-linear scales validates the instrument; a detection at cosmological scales validates the science.

MeerKAT's 64-dish array, now proven as a single-dish intensity mapping instrument, operates on the same Karoo Desert site where the SKA-Mid is currently under construction. The single-dish mode trades angular resolution for field of view, making it ideally suited to catching the large-scale fluctuations that encode BAOs and dark energy signatures — and effectively optimizing for the same measurement class the SKA's cosmological surveys will pursue at a vastly larger scale.

What Comes Next

The immediate next step is scale. The 236 square degrees analyzed in this paper represents a proof-of-concept volume. The team's current data collection aims to reach earlier cosmic epochs and a sky coverage more than 1,000 times larger. At that scale, the BAO standard ruler — the 490-million-light-year preferred separation between matter clumps, imprinted by primordial sound waves — becomes detectable in the hydrogen maps themselves, potentially yielding independent constraints on dark energy and the expansion history of the Universe from radio waves alone.

Longer term, the MeerKLASS detection delivers something arguably as valuable as the measurement itself: validated pipelines, characterized systematics, and a proven foreground methodology. These are the essential tools the SKA Observatory will need when it begins its own cosmological hydrogen surveys, expected later this decade. The SKA-Mid's 197 dishes — built around the existing MeerKAT array and extending outward across South Africa — will scan the sky in single-dish mode, covering thousands of square degrees and mapping the three-dimensional distribution of hydrogen across billions of years of cosmic history.

For now, the signal that took ten years to isolate from a million-to-one foreground represents the clearest demonstration yet that the Universe's most abundant element can serve as a standalone cosmological instrument. The whisper, it turns out, can be heard.