LZ dark matter detector records a single unexplained xenon recoil

Physicists on the LUX-ZEPLIN experiment told a conference in Tendo, Japan, on 1 September that a single particle strike in their underground xenon tank does not match any known background they have catalogued. They stopped short of claiming a dark matter detection. The collaboration will send the paper to Physical Review Letters.

LZ sits about a mile down in the former Homestake gold mine at the Sanford Underground Research Facility in Lead, South Dakota. Lawrence Berkeley National Laboratory manages the project. Some 250 scientists from 39 institutions work on the data. The detector holds ten tonnes of liquid xenon watched by photomultiplier arrays that pick up faint light and charge when a nucleus recoils.

The event happened on 16 June 2023. After the team finished a blinded analysis of 220 days of data from March 2023 to April 2024 and "unsalted" the sample, one recoil remained. It deposited 248 keV. Rick Gaitskell of Brown University, the spokesperson, said the leftover event sat in the region where a weakly interacting massive particle would be expected and where competing backgrounds are low.

If the strike was a WIMP, the implied mass is at least 200 gigaelectronvolts and more likely near 1,000 GeV, or roughly a thousand times the mass of a proton. Ordinary matter is about 15 percent of the universe's matter budget. The rest is inferred from gravity on galaxies and does not emit or reflect light.

The statistical claim is modest. The collaboration quotes about 2.6 sigma, or a chance of roughly 1 in 200 that known physics produced the flash. Some write-ups put it near 3 sigma and 1 in 400. Discovery in particle physics is conventionally 5 sigma. Aaron Manalaysay of Berkeley Lab said it was the first outlier he had seen on any experiment that still looked valid after every background test.

Two mundane explanations were checked. Two unrelated flashes arriving close in time is treated as very unlikely. Residual radioactivity and neutrino scatters were modelled and subtracted before the box was opened. The collaboration still wants other xenon detectors to hunt the same mass window in their own archives.

LZ has already set world-leading limits on lighter WIMPs. This result does not replace those limits. It is one candidate event in a detector designed to run about 1,000 live days. More exposure will either add similar recoils or wash this one into the noise.

For readers who have followed decades of null dark-matter searches, the useful fact is the energy and the date: 248 keV on 16 June 2023, in a 10-tonne xenon target a mile underground, at a significance that is interesting and not yet decisive. The next data set will decide whether that sentence survives.