The Dark Halo That Never Lit Up
Galaxies announce themselves through the light of billions of stars, all embedded in vast clumps, or “halos,” of dark matter. But researchers may have spotted, for the first time, a starless halo of dark matter—containing only a gas cloud. The result was announced by Rachael Beaton of the Space Telescope Science Institute in Maryland at the meeting of the American Astronomical Society in Phoenix, Arizona. Using observations from the Hubble Space Telescope, Beaton and her collaborators showed that the object, known as Cloud-9, contains a negligible amount of stars [1]. “There is nothing like this that we have found so far in the Universe,” Beaton said in a press conference last week.
Cloud-9’s makeup—as inferred from radio and optical observations—would qualify it as the first example of a REionization-Limited H I Cloud (RELHIC), a starless dark matter halo filled with neutral hydrogen gas (H I). RELHICs are thought to be leftovers of dark matter clumps that couldn’t accrue a sufficient amount of gas to form stars, says the project’s principal investigator Alejandro Benítez-Llambay of the University of Milano-Bicocca in Italy. A RELHIC is “a tale of a failed galaxy,” he says.
Starless halos arise naturally within the standard paradigm of cosmology: the lambda cold dark matter (ΛCDM) model, where Λ refers to a “cosmological constant” that describes dark energy. According to ΛCDM, dark matter can cluster into halos that provide the gravitational backbone for galaxy formation. The model also predicts that there is a critical mass below which halos would be too small to ever form stars. Spotting unlit halos might sound hopeless, but simulations by Benítez-Llambay and collaborators in 2017 suggested that halos within a narrow mass range may exist as RELHICs (a term they coined) [2]. According to their calculations, RELHICs would have masses close to the critical value for galaxy formation. Crucially, the compact, hydrogen-filled cores of these objects provide a potential observational window, since hydrogen clouds have a characteristic radio emission.
To search for possible RELHICs, Beaton and her colleagues first had to find suitable hydrogen gas clouds—ones looking like galaxies at radio frequencies. Among hundreds of detected hydrogen clouds, a particularly promising one was Cloud-9, named for being the ninth cloud recorded around the nearby galaxy Messier 94 (astronomers would be “on cloud nine” if a RELHIC were confirmed, but the association with that idiomatic expression is just a coincidence). The object was first spotted in 2023 by the Five-hundred-meter Aperture Spherical Telescope in China, and later by the Green Bank Telescope and the Very Large Array (VLA) in the US. The VLA observations pinpointed the cloud’s position with high accuracy and also mapped its shape in detail.
These observations revealed that the cloud contains 1 million solar masses of hydrogen gas, has a slightly distorted spherical shape about 4900 light-years across, and lies in a seemingly starless region (according to previous surveys with ground-based optical telescopes). What’s more, the cloud doesn’t appear to rotate (rotation can be measured by mapping the radio-emission frequency from different parts of the cloud). The lack of motion in such a large cloud led the researchers to conclude that the pressure of the hydrogen gas must be counterbalanced by the gravitational pull of about 5 billion solar masses of dark matter.
But is Cloud-9 really a starless RELHIC, or could it be a routine dwarf galaxy that is too faint to be seen by ground-based telescopes? The researchers ruled out the galaxy hypothesis using the Hubble Space Telescope. Inspecting the Cloud-9 location on eight orbits, Hubble delivered no evidence of starlight. To quantify this assessment, the researchers compared Hubble observations with simulations of images expected from galaxies of varying mass. The analysis concluded, with 95% confidence, that the cloud cannot contain more than 3000 solar masses in stars.
The dearth of stellar detections in Cloud-9 marks a stark difference with other recently detected clouds with similar radio-emission patterns, which were found to be faint dwarf galaxies in follow-up optical-telescope observations. “Whatever we have seen here, it’s not a galaxy,” Benítez-Llambay says. He adds that the team analyzed and ruled out several alternative scenarios in which Cloud-9’s observational features are produced by more conventional types of astronomical objects. “Taken together, [the observed features] provide compelling evidence that Cloud-9 is a RELHIC,” says Haibo Yu of the University of California, Riverside, who wasn’t involved in the study.
“RELHICs have important implications for our understanding of the nature of dark matter,” Yu says. More massive, galaxy-hosting halos are affected by complex processes involving normal, or baryonic, matter. Since they lack stars, RELHICs avoid this complication, he says. Their features, such as their mass distribution, are therefore more directly associated with dark matter properties, Benítez-Llambay adds. “I see them as astrophysical laboratories for probing the nature of dark matter,” he says.
If the RELHIC interpretation is confirmed, it would verify a cornerstone prediction of standard cosmology, says Gianfranco Bertone of the University of Amsterdam. “But its real impact will emerge once we identify a population of such objects,” he says. Both Yu and Benítez-Llambay hope that more objects like Cloud-9 will be discovered by upcoming surveys. Benítez-Llambay anticipates that the Square Kilometer Array radio telescope being built in Australia and South Africa could be transformative for the field, as it will have the sensitivity and resolution to systematically sift the sky for RELHIC candidates.
–Matteo Rini
Matteo Rini is the Editor of Physics Magazine.
References
- G. S. Anand et al., “The first RELHIC? Cloud-9 is a starless gas cloud,” Astrophys. J., Lett. 993, L55 (2025).
- A. Benítez-Llambay et al., “The properties of ‘dark’ ΛCDM haloes in the Local Group,” Mon. Not. R. Astron. Soc. 465, 3913 (2016).