A mysterious signal could reveal the invisible matter that makes up the universe - critical summary review

On September 1, 2026, physicists working on a detector buried under South Dakota, in the United States, reported that they had recorded a particle interaction they cannot explain. It looks like what a dark matter particle should produce. It is also one single event, and the team has been careful to say that is not the same thing as finding dark matter. Understanding why requires knowing what they were looking for in the first place.

What dark matter is, and why most of the universe is invisible

Dark matter is the name given to a substance that has never been observed directly. It emits no light and reflects none, so no telescope can photograph it. What betrays it is gravity: galaxies spin as though far more mass were holding them together than anything visible can account for, and light bends around large clusters in ways that only make sense if something unseen is there.

The accounting is roughly this. Ordinary matter, everything that can be seen or touched, makes up about 5% of the universe. Dark matter accounts for about 27%, and dark energy, a separate force linked to the accelerating expansion of the universe, about 68%. Together the two invisible components come to about 95%. Counting matter alone, dark matter is around 85% of it.

The first clue came in the 1930s, when the Swiss astronomer Fritz Zwicky noticed that galaxies in the Coma Cluster were moving too fast for the cluster's visible mass to hold them in place. He coined the term "dark matter" in 1933. The strongest evidence today comes from the Bullet Cluster, formed when two galaxy clusters collided: the mass revealed in the collision was distributed differently from the hot gas, and researchers said that mass could be dark matter.

What it is made of remains open. The longest-standing candidate is the WIMP, short for weakly interacting massive particle, heavy and almost never interacting with ordinary matter. Another is the axion, lighter and behaving more coherently. Nearly a century after Zwicky, neither has been caught.

One flash in a tank of xenon, deep in an old gold mine

Detectors like LUX-ZEPLIN, known as LZ, do not go looking for anything. They sit and wait. They are buried deep underground, where the surrounding rock blocks cosmic radiation that would otherwise drown the signal, and they are built from materials chosen for very low levels of trace radioactivity, so that a genuine detection is not lost in background noise.

LZ sits in an old gold mine beneath the Black Hills, at the Sanford Underground Research Facility, the deepest laboratory in the United States at nearly a mile below the surface. At its centre is a tank of ultra-pure liquid xenon surrounded by hundreds of light sensors. The theory being tested is specific: if a WIMP strikes the nucleus of a xenon atom, it should release a faint flash of light and a pulse of electric charge, producing two flashes at a particular energy level.

The collaboration, 250 scientists and engineers, reviewed 220 days of data collected between March 2023 and April 2024. They found one interaction of that kind, recorded on June 16, 2023. Sam Eriksen, a senior research associate at the University of Bristol in the United Kingdom and lead scientist on the study, presented the findings at the 2026 TeV Particle Astrophysics conference in Japan.

"We understand our detectors and the backgrounds so well," Eriksen said in a media release, "that even a single outstanding event, like the one we found, is important." Theresa Fruth, a physicist at the University of Sydney who worked on the study, told the Australian broadcaster ABC News that the event is interesting precisely because it has survived repeated scrutiny: "This event just won't go away, even after many, many checks."

Why 2.6 sigma is not a discovery

Particle physics measures how confident it is in a signal using a scale called sigma. The LZ result was rated 2.6 sigma. The normal threshold for calling something a discovery is five. The findings have also not yet been through peer review by scientists outside the collaboration.

The researchers themselves have been the loudest voices for restraint. "With only one event, we don't want to get ahead of ourselves," said Rick Gaitskell, the LZ spokesman and a physicist at Brown University in the United States. "We are not claiming to have seen dark matter. But we have seen something interesting that we want to share with the scientific community for their input."

The difficulty is separating a real signal from a rare background. Henning Flaecher, an experimental particle physicist at the University of Bristol, said a large amount of work had gone into checking previously known explanations for the reaction, and that "to date none provide a convincing explanation". Eriksen put the problem to the news outlet Al Jazeera this way: "The backgrounds we have occur very rarely, but so does dark matter." In a detector of this kind, fewer than five events a year are expected in total. Dark matter passes through human bodies constantly with no measurable effect: billions of candidate particles cross a person every second, and only about ten strike a nucleus in the body over a year.

That is the scale of the thing being hunted, and why a single flash carries so much weight and so little certainty at the same time.

The larger puzzle the event sits inside

While one team tries to catch a particle, cosmologists are questioning whether the invisible universe is even two separate things. In 2024, the Dark Energy Spectroscopic Instrument collaboration, known as DESI, found evidence that the strength of dark energy was not constant. A 2025 study using more than twice as much data reached the same conclusion.

Those results pointed to a period in which dark energy appeared to grow stronger, something researchers call the "phantom regime" because it seems to defy energy conservation. A growing number of theorists suspect the explanation is that dark energy and dark matter are coupled. "The notion that you can compute dark energy independently of dark matter is wrong," said Cumrun Vafa, a physicist at Harvard University, who with collaborators proposed in 2022 that both could be linked through an extra "dark dimension". Some of these models also ease the Hubble tension, a persistent disagreement of roughly 9% between two ways of measuring how fast the universe is expanding.

None of this bears directly on what LZ saw. It is the reason the field treats the question as unsettled at every level at once.

Why dark matter is back in public view

The timing helped. The LZ observation came days after the second season of the science-fiction thriller Dark Matter, starring Joel Edgerton and Jennifer Connelly, and around the same time NASA announced the launch of a roughly $4bn project, the Nancy Grace Roman Space Telescope, built to investigate both dark matter and dark energy. Three unrelated events pushed the same two words into circulation within a week, which is why a cautious conference talk landed like a breakthrough.

What is known, and what is still unknown

What is established: dark matter's gravitational effects have been observed for decades, from the Coma Cluster in the 1930s to the Bullet Cluster; it accounts for about 27% of the universe; it has never been detected directly. What LZ reported is one particle interaction, recorded in June 2023, that survived months of checks and that the team has been unable to attribute to any known background.

What is unknown: whether that event is a WIMP, a rare background nobody has characterised yet, or something else entirely. At 2.6 sigma it is roughly half the statistical strength physics demands before using the word discovery, and the work has not been peer-reviewed.

So the thing to follow is not this event. It is whether the analysis clears independent review, and, more decisively, whether further candidate events show up in the next batches of LZ data. Flaecher said the collaboration cannot wait to analyse more data to see if additional candidates appear. Eriksen told Al Jazeera that the key next step is confirming the signal is "something new" rather than background. One flash proves nothing. A handful of them, at the same energy, would be a different story, and that is the only thing that will settle which way this goes.