A Detector a Mile Under South Dakota Saw One Event It Cannot Explain, and Its Own Scientists Say Do Not Get Excited
The LUX-ZEPLIN collaboration reported a single particle interaction in ten tonnes of liquid xenon that matches no known background. It is a 2.6 sigma result, one chance in two hundred of being a fluke, and far from a discovery. Why it matters anyway.
Outspoken Digest Technology Desk
Tuesday, September 8, 2026/3 min read
On 16 June 2023, at 3:22 in the afternoon and thirty-nine seconds, something struck an atomic nucleus inside a tank of liquid xenon buried a mile beneath the Black Hills of South Dakota and produced a flash of light and a pulse of charge that nobody has been able to account for since. The LUX-ZEPLIN collaboration, thirty-nine institutions running the most sensitive dark matter detector ever built, announced the event on 1 September at a particle astrophysics conference in Japan. The paper is going to the arXiv and to Physical Review Letters.
Every sentence about it since has been a fight between two true things. The event is genuinely strange. And one event proves nothing.
What the detector is for
Most of the matter in the universe is invisible, known only by its gravity, and for four decades the leading guess has been that it consists of weakly interacting massive particles, WIMPs, that drift through ordinary matter almost without touching it. Almost is the point. Very rarely one should collide with a nucleus, and a large, quiet, cold volume of xenon, watched by photomultipliers, is the instrument built to see that collision. LZ sits in a former gold mine at the Sanford Underground Research Facility so that a mile of rock filters out cosmic rays, inside a water tank that stops neutrons, with software that models every remaining source of ordinary background.
The result comes from 220 live days of data taken between March 2023 and April 2024. In an expanded search of that data, one interaction stood out as far from any of the known backgrounds. It was, in the words of one collaborator quoted by Scientific American, far too explosive for a conventional WIMP signal, which is part of why it was not caught earlier: the team had been looking for something gentler.
2.6 sigma
The collaboration puts the significance at 2.6 sigma, meaning roughly a one in two hundred chance that known backgrounds produced it. Particle physics declares a discovery at five sigma, about one in three and a half million. The gap between those numbers is the whole story. Two point six sigma events appear and vanish in physics all the time; the 750 GeV bump at the Large Hadron Collider in 2015 was more significant than this and turned out to be nothing.
There is a further caveat the team has been candid about. The event was found in an analysis of data that had not been blinded, the standard protocol in which researchers work without knowing where a signal would appear, precisely to stop them talking themselves into one. Richard Gaitskell of Brown University, the collaboration's spokesperson, said: with only one event, we don't want to get ahead of ourselves. 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. Jingke Xu of Lawrence Livermore noted that the event could be described by one of the dark matter models the team had explored, which is a different statement from saying it was one.
Why anyone is excited at all
Because of who is being careful. Katherine Freese of the University of Texas at Austin, who has worked on dark matter detection since the field began, told Scientific American she had been waiting for a positive result for forty years. Hugh Lippincott of UC Santa Barbara, an LZ member, said simply that it is far from any of the known backgrounds. The people best placed to find a mundane explanation have not found one yet, and they have spent two years looking.
And because the test is already designed. LZ holds more than 700 days of blinded data that has not yet been analysed. If the June 2023 event was a real signal, its cousins are in that data at a rate the models predict; if the blinded analysis finds nothing, the event goes into the long file of anomalies that meant nothing. Two rival xenon experiments, XENONnT in Italy and PandaX-4T in China, can look for the same thing independently. The answer, unusually for a claim about the nature of the universe, is on a schedule measured in months rather than decades.
Our earlier reporting on a different kind of result that was oversold on release, and how to read a significance figure, is in the piece on Saturn's new decagon and the eight-letter DNA story from last week. This one is being sold honestly by its own authors. That is worth noticing, and it is worth waiting for the blinded data before deciding what happened at 3:22 on a June afternoon three years ago.
Published in The Outspoken Digest
Editorial desk
Outspoken Digest Technology DeskSoftware, hardware, artificial intelligence and what they change for everyone else.
Newsletter
The Digest, in your inbox
One edition, sent when it is ready. No noise, and your address is never passed on.
Read Next
More Technology →
Nvidia Is Buying Hugging Face for 12.93 Billion Dollars, Putting the Chipmaker in Charge of Where Everyone's Models Live
Sep 8, 2026/3 min read
Apple's Event Is Tomorrow, John Ternus's First as Chief Executive. Only the Foldable Would Count as a Surprise
Sep 8, 2026/3 min read
MIT Wired Bacteria Into Transistors. Its Biggest Circuit Is Twenty-Four Colonies on a Dish That Takes Eight Hours to Add
Sep 8, 2026/3 min read
Saturn Has Grown a Ten-Sided Storm at Its South Pole, and Nobody Saw It Start
Sep 6, 2026/3 min read