Accidental Hubble find — a stellar stream in a dwarf galaxy — may unlock dark matter secrets

Full text

In late June 2023, on a morning train commuting to work in New York City, David Hendel was scrolling through a recent astronomy paper on his phone when a half-page image caught his eye.

The image, captured a year earlier by the Hubble Space Telescope , showed UGC 9050-Dw1, a dwarf galaxy some 115 million light-years from Earth . It appeared as little more than a faint, diffuse smudge of bluish-white light that astronomers call an ultra-diffuse galaxy — a type of galaxy so sparsely populated with stars that most of its mass is thought to be dark matter , the invisible substance that makes up more than 80% of the matter in the universe but has yet to be directly detected.

What made Hendel pause was not the galaxy itself but a thin, fragile ribbon of stars trailing off from it, so dim it nearly disappeared among the other smudges of light crowding the frame. Hendel, a former astronomer who had studied thousands of similar structures, immediately recognized the subtle interplay of the arc's features, including the shared color of the stars and their curved trajectory pointing toward a tiny, matching bright knot along its track. Together, those clues matched a stellar stream tearing from a tight ball of stars called a globular cluster, pulled apart by the gravity of UGC 9050-Dw1.

If Hendel's reading was correct, it would be the first such stream ever found outside the Milky Way .

"It almost seemed too good to be true," Hendel told Space.com. "At this point I felt the excitement of a discovery to such a degree that I recall looking up and around the train for someone to tell!"

This discovery would give researchers a pristine new laboratory for probing dark matter, and one far less contaminated by competing astrophysical forces that muddy similar searches within our own galaxy . Hendel, who now works in finance but still follows research from his former field, ran a quick literature search to see whether anyone had already documented the feature. No one had.

The right person to tell was not on the train, of course. Before reaching his stop, Hendel took a screenshot, sketched a rough circle around the arc and sent it to Sarah Pearson, an associate professor of astrophysics at Technical University of Denmark's DTU Space, and his friend since their graduate school days at Columbia University.

That message kicked off the collaborative effort that would ultimately confirm the find, led by Pearson and Julie Kiel Holm, a Ph.D. student at the Niels Bohr Institute at Denmark's University of Copenhagen.

The team's paper gives the stream the name Oyashio (pronounced "oh-yah-shee-oh"), Japanese for a cold Pacific current, following a tradition of naming these faint stellar trails after bodies of water in Indigenous languages to honor cultures worldwide with long astronomical traditions, including in Australia, Chile and India.

An image of the Hubble Space Telescope. (Image credit: NASA) Within the Milky Way, astronomers have so far identified roughly 100 stellar stream candidates shed by globular clusters, but finding one around another galaxy is extraordinarily rare. Streams that form when a whole dwarf galaxy is torn apart are relatively easy to spot, since they emerge in the sparse outskirts of a larger galaxy. But globular cluster streams form deep inside their host galaxy, where they are washed out by brighter, more crowded starlight. Oyashio's detection was possible in part because UGC 9050-Dw1's already-sparse stellar population meant a dimmer backdrop gave the stream just enough contrast to be seen, Holm told Space.com.

Oyashio's discovery is "a very exciting step forward," Emily Cunningham, an assistant professor at Boston University who studies stellar streams and dark matter halos and was not involved in the new study, told Space.com. "Using techniques like those described in this paper, we will detect many more streams around other galaxies with upcoming observations from facilities such as the Vera Rubin Observatory and the Nancy Grace Roman Space Telescope."

A sufficiently large catalog of such streams in different galactic environments would allow them to tease out the subtler perturbations that dark matter is predicted to cause at the smallest scales, the researchers say.

"This is one of the most promising avenues for new astrophysical constraints on dark matter's nature," said Hendel.

No Eureka! moment

UGC 9050-Dw1 was first identified in 2017 in images taken by the Canada–France–Hawaii Telescope atop Mauna Kea in Hawaii, as part of a survey for diffuse dwarf galaxies. A follow-up study in 2023, drawing on the same Hubble images in which Hendel later spotted Oyashio, catalogued an "exceptionally high" population of globular clusters near the galaxy. Its authors concluded that UGC 9050-Dw1 likely formed from the merger of two comparably sized galaxies roughly 10 billion years ago, one gentle enough that gas from both survived and ignited a burst of star formation, which would also explain the clusters' uniform color. That 2023 study did not flag Oyashio.

In the new study, Holm and her colleagues recovered Oyashio independently in archival data from both Hubble and the Canada-France-Hawaii Telescope, ruling out the possibility that it was an artifact of telescope imaging or subsequent data processing.

The team then ran thousands of simulations, testing which combinations of globular cluster properties and dark matter distributions could reproduce the stream's observed brightness and shape. The best-fitting models produced new estimates of the UGC 9050-Dw1's total mass and how that mass is spread throughout it, suggesting the galaxy harbors far more dark matter than its sparse stellar population would imply, consistent with expectations for such a diffuse galaxy.

A ring of dark matter (seen by the Hubble Space Telescope in false color), its location inferred from its gravity, around the galaxy cluster ZwCl0024+1652, which is 5 billion light-years away. (Image credit: NASA/ESA/M. Jee and H. Ford (Johns Hopkins University).) In the study, the team also weighs a range of alternative explanations and lays out why none quite fit the evidence. The arc could be, for instance, a chance alignment of unrelated stars, the product of gravitational lensing from a background galaxy, or a dusty region that happens to mimic a stream's shape. None of those scenarios, the team says, accounts on its own for the key pieces of evidence, chiefly the stars' consistent color, their alignment along a single curved track and that track's connection to what appears to be the globular cluster progenitor.

"It was an interesting process in the sense that there wasn't a specific moment where we're like, 'Okay, now we know what it is,'" Holm told Space.com. "We started with the assumption that just by looking at it, it looks like it could be something. How many different ways can we try to prove that it can't be?"

Still, in the interest of rigor, the team acknowledges it cannot fully exclude every alternative explanation. "There's still the possibility that some stars just randomly fall in this pattern," said Holm, "But again ... for all these things to line up as nicely as they do would require a lot of coincidence."

Dark matter clues

The payoff of that caution lies in what Oyashio can reveal about the universe's invisible scaffolding. As a globular cluster is shredded over billions of years, the stars it sheds fan out along its former orbit rather than scattering randomly, preserving, in effect, a pristine record of the gravitational forces that shaped them. If a clump of dark matter too small to host stars of its own — a "subhalo," in the language of cold dark matter theory — passes through that stream, its gravity can tug stream stars off course and carve out a gap.

"Streams that form from globular clusters are particularly sensitive to these perturbations," said Cunningham.

The challenge lies in isolating the cause, and one stream is not enough to draw firm conclusions about the astrophysical process responsible. Inside the Milky Way, any disturbance in a stream of stars could just as easily come from the galaxy's central bar of stars or from nearby dwarf galaxies falling into the Milky Way, like the Magellanic Clouds. UGC 9050-Dw1 has none of those features, so any gap found in Oyashio would stand as a more confident signature of dark matter at work, Holm said.

"If a gap is observed in a single stream, it is very difficult to say for certain which astrophysical process is responsible," said Cunningham. "However, if we can build up a sufficiently large sample of streams in different galactic environments, we can start to disentangle these effects."

"To do this, we need to look beyond the Milky Way," she added.

New instruments coming online could help do just that. Astronomers are particularly excited for the launch of the Nancy Grace Roman Space Telescope , scheduled for launch Aug. 30 from the Kennedy Space Center in Florida, which has a field of view at least 100 times larger than Hubble’s. In its five-year primary mission, with a possible five-year extension, Roman is expected to measure light from a billion galaxies, according to NASA .

"You would be able to do this kind of research in different galaxies at once with the same exposure, basically," said Holm. She hopes Roman, along with the European Space Agency's Euclid space telescope, which launched in 2023, will turn up more galaxies like Oyashio, "so we can start to put together a more generalized image of how these things work and what they can tell us."

Some of those discoveries could even come to light the way Oyashio did, as a product of pure scientific curiosity applied to data already sitting in an archive. Hendel said he read the 2023 paper mostly because he was intrigued by the team's conclusion that UGC 9050-Dw1 likely formed through a dwarf galaxy merger. "I had no aim when opening it except to satisfy my curiosity," he said.

That he was able to spot a rare feature hiding inside an image collected for a completely different research project speaks to the value of the scientific community's long-standing practice of making data publicly available, Holm said.

"Anyone can come along with another idea and actually find something also interesting with the same data that you didn't know was hiding in it," said Holm. "I think that's also a really cool part of this."

The team's paper was published on Aug. 12 in the journal Nature .