The LUX-ZEPLIN experiment recorded a single interaction in 220 live days that could have come from a WIMP. Here is how the detector works and why 2.6 sigma is not a discovery.
The LUX-ZEPLIN xenon detector, known as LZ, is a 10-tonne tank of liquid xenon nearly a mile beneath South Dakota. On Tuesday 1 September 2026 its collaboration announced one particle interaction in 220 live days of data that could potentially have been caused by a WIMP, a leading dark matter candidate. Known backgrounds would mimic it only about 0.5% of the time. That works out to a significance of 2.6 sigma, well short of the 5 sigma physicists demand before they use the word discovery.
The analysis was presented at the 2026 TeV Particle Astrophysics conference in Japan by a collaboration of about 250 scientists and engineers from 39 institutions in six countries. The lead author is Sam Eriksen of the University of Bristol. LZ spokesperson Rick Gaitskell of Brown University put the message plainly: “We are not claiming to have seen dark matter.”
So why is a single event making news? Because of where it landed, and because dark matter is the largest hole in the physical account of the cosmos. This article explains how the detector registers an interaction, what kind of search produced the event, what 2.6 sigma means in practice, and what the paper will have to show. The facts here come from the Brown University and University of Bristol announcements of 1 September. I have not seen the paper, so the background estimate and the event’s measured properties are not covered.
The result in numbers
| Quantity | Reported value |
|---|---|
| Events of interest | One |
| Exposure | 220 live days, March 2023 to April 2024 |
| Significance | 2.6 sigma, below the 5 sigma discovery threshold |
| Chance known backgrounds explain the event | Roughly 0.5%, about 1 in 200 |
| Implied WIMP mass if it is dark matter | Likely at least 200 GeV/c², over 200 times the proton mass |
| Detector target | 10 tonnes of ultrapure liquid xenon |
| Collaboration | About 250 scientists and engineers, 39 institutions, six countries |
| Where it was presented | 2026 TeV Particle Astrophysics conference, Japan, 1 September 2026 |
Why physicists believe in dark matter at all
The case rests on gravity doing things that visible matter cannot explain. Stars at the edges of galaxies orbit faster than the glow of the galaxy’s own gas and stars should allow. Clusters of galaxies bend the light of objects behind them more strongly than their visible mass can account for. The pattern of the cosmic microwave background, the afterglow of the early universe, also fits only if most of the matter is of a kind that does not interact with light.
All of that evidence is gravitational. Nobody has yet caught a dark matter particle interacting in a laboratory, which is what makes a detector like LZ so consequential. A confirmed interaction would turn an inference into an observation.
What LZ is and where it sits
LZ stands for LUX-ZEPLIN. Berkeley Lab manages the experiment for the US Department of Energy, and the detector sits nearly a mile underground at the Sanford Underground Research Facility in South Dakota. The rock overhead shields it from cosmic radiation, which would otherwise swamp the faint signals it is looking for. A water tank and outer detectors guard against stray neutrons.
The target is dark matter, which Bristol’s release puts at around 85% of the universe’s mass. The term is a placeholder for something that tugs on galaxies and bends light yet has never been seen to emit or absorb it. WIMPs, weakly interacting massive particles, are one leading candidate. LZ has been running since 2021, and the collaboration says it holds the largest dark matter dataset in the world.
How a tank of xenon registers an interaction
Picture a collision between a WIMP and a xenon nucleus. The nucleus recoils, and two things happen in rapid sequence. The recoil produces a flash of light, picked up by light sensors at the top and bottom of the liquid xenon chamber. It also knocks loose electrons, which an electric field drags upward to the surface of the liquid, where they produce a second light flash.
Two flashes carry a lot of information. The gap between them reveals how deep in the tank the collision occurred, and the sizes of the flashes say how much energy was deposited and what kind of particle did it. Brown University built the light sensors.
Why xenon, and why so much of it
A WIMP, if it exists, would interact so rarely that a typical detector would see nothing for years. Two things help. A heavy target nucleus gives a heavy particle a better chance of a recoil that can be measured, and xenon is a heavy element. And the more target there is, the more chances there are, which is why LZ holds 10 tonnes of it. This is general detector physics rather than anything specific to the new result.
Purity matters as much as mass. Any radioactive contaminant in the xenon or the surrounding materials produces events that look a little like the ones being hunted. Much of the work of an experiment like this lies in knowing, to a high degree of precision, how many such impostors to expect.
The kind of search that found the event
Earlier, the collaboration searched this same dataset for the simplest WIMP interactions. The new analysis looked for a broader range of interactions, including ones that deposit more energy in the detector. The single event sits in that high-energy region, where backgrounds are very low, in Gaitskell’s words.
That is why one event can carry weight. In a quiet corner of the data, even a lone event stands out, in the way a single cough stands out in an empty library. In a noisy corner, the same event would vanish into the crowd.
Why 2.6 sigma is intriguing but not a discovery
Sigma is shorthand for how unlikely a result is if nothing new is happening. A 2.6 sigma excess corresponds to a chance of roughly 0.5% that known backgrounds alone would produce it, which is the 1 in 200 figure the collaboration quotes. Five sigma, the discovery bar, corresponds to roughly 1 chance in 3.5 million.
Those numbers sound very different because they are. The gap between a 1-in-200 fluke and a 1-in-3.5-million one is the gap between an interesting anomaly and a result that rewrites textbooks. With one event, there is also no shape to examine, no energy spectrum to compare against a prediction, and no second event to confirm the first. Particle physics has seen plenty of 2 and 3 sigma bumps fade as more data arrived. Berkeley Lab’s Manalaysay said this is the first outlier he has seen in any experiment that looks sound on every count he can check.
The questions the paper will have to answer
- The background budget. How many events did known sources, such as radioactivity in detector materials, neutrons, and rare processes, predict in this region? The announcements give only the 0.5% figure.
- The look-elsewhere effect. A search over a wide range of energies and interaction types gives many chances for a fluke to appear somewhere. Whether the 2.6 sigma figure accounts for that is a statistics question the announcements do not address.
- The event’s properties. Where in the tank it happened, how its two flashes compare, and whether anything unusual was happening in the detector at the time.
- Consistency across the dataset. A real particle signal should behave like one in every respect that can be checked.
What it would mean if it were dark matter
If the event were caused by dark matter, the particle would likely have a mass of at least 200 GeV/c², more than 200 times the proton mass. Bristol’s release adds that it would imply an interaction beyond the simplest model. In other words, the usual textbook version of a WIMP would not do, and theorists would have to go back to their blackboards.
That is a large “if.” The collaboration itself calls the result intriguing, not a detection, and it wants the wider community’s scrutiny before anyone draws conclusions.
What happens next
More data will settle it. If the event is a statistical fluke, further running should show nothing nearby. If it points to something real, similar events should start to accumulate. Gaitskell is scheduled to discuss the result in a Brown Physics colloquium on 11 September.
Longer term, Bristol is contributing to XLZD, a proposed next-generation xenon detector that may be hosted at the Boulby Underground Laboratory run by STFC. Any signal that survives LZ would be a prime target for it.
A short glossary
- WIMP: weakly interacting massive particle, a hypothesized dark matter particle that would rarely touch ordinary matter.
- Sigma: a measure of statistical significance. Higher means less likely to be a fluke.
- Live days: days the detector was actually taking good data, not total calendar time.
- Background: ordinary events that can mimic a dark matter interaction.
- GeV/c²: a unit of mass used in particle physics. A proton is about 1 GeV/c².
Frequently asked questions
Did LZ detect dark matter?
No. The collaboration says it is not claiming to have seen dark matter. It found one event that is unlikely, at about 1 in 200, to be explained by known backgrounds.
What is 2.6 sigma?
A significance level. It means there is roughly a 0.5% chance that known backgrounds alone would produce the observation, well below the 5 sigma standard for a discovery.
Why is the detector underground?
About a mile of rock blocks cosmic radiation that would otherwise overwhelm the signals LZ is hunting.
Is the paper available?
Bristol’s release says a preprint is available and the work has been submitted to a journal, while Brown’s says it will be posted to arXiv and submitted to Physical Review Letters. Check both before relying on either.
Sources
