An anomalous shock at 248 keV emerges from LUX-ZEPLIN data: it passed the background noise tests, but it is still not enough for a discovery

Only one impact, recorded in 220 days of observation, managed to emerge from a region of the detector where the expected background noise is extremely low. The event deposited an energy of 248 kiloelectronvolts and is consistent with the recoil of a xenon atom's nucleus. It could have been produced by a dark matter particle, but the available data do not yet allow us to establish this.

The signal comes from LUX-ZEPLIN, or LZ, one of the most sensitive instruments built to search for interactions between ordinary matter and hypothetical Weakly Interacting Massive Particles, the weakly interacting massive particles known as WIMPs. The detector operates nearly a mile below the surface, in the Sanford Underground Research Facility of Lead, South Dakota.

La LZ Collaboration has published the result in preprint form, therefore before peer review, after having extended the analysis to nuclear recoils much more energetic than conventional research. Statistical significance reaches a local maximum of 3,4 sigma, but it goes down to 2,6 sigma on a global scale considering the number of models tested. It's an anomaly worthy of attention, not the announcement of the first direct observation of dark matter.

Seven tons of xenon transform shocks into light

Astronomical evidence indicates that dark matter makes up approximately 85 percent of the matter in the universeIts presence is inferred from the gravitational effects it exerts on galaxies, clusters, and the light passing through large cosmic structures. Since it neither emits nor absorbs electromagnetic radiation in an observable way, the substance remains invisible to telescopes, and its microscopic composition is unknown.

LZ attempts a direct measurement. At the center of the apparatus is a dual-phase time projection chamber containing approximately seven active tons of liquid xenon, part of a total inventory of nearly ten tons. Xenon offers heavy nuclei and an extremely pure medium in which even minimal energy stores can be reconstructed.

When a particle interacts in the liquid, it produces an initial scintillation pulse, called S1, and releases electrons. An electric field pushes them toward the top of the chamber, where they are extracted into a layer of xenon gas and generate a second light signal, S2The relative intensity of the two pulses, their time interval, and the distribution of light allow us to estimate the energy and three-dimensional location of the interaction.

This information also helps distinguish two families of phenomena. Gamma and beta particles tend to interact with xenon's electrons, producing electronic recoils; neutrons or possible WIMPs can instead directly hit the nuclei, generating nuclear recoilsStatistical separation is very effective, but not absolute: rare events, random coincidences, and non-standard device responses can mimic the desired signal.

The detector is located deep underground to reduce the flux of cosmic rays and is protected by a water tank and external systems capable of identifying neutrons and other particles. The search for dark matter thus depends not only on the target's mass, but on the ability to identify and reject nearly every possible background source.

The energy window increases from 55 to about 270 keV

The analysis uses data collected between the March 27, 2023 and April 1, 2024, for an overall exposure of 2,84 tons per yearThe same sample had already been studied in the search for the simplest WIMP interactions, concentrated mainly at low energies. The new work has expanded the region of nuclear recoils from 5,4 up to about 270 keV, against a previous limit near 55 keV.

The choice allows us to explore models in which the expected energy distribution does not decrease as rapidly as in the traditional spin-independent elastic interaction. The collaboration considered operators of the non-relativistic effective field theory, a mathematical tool that describes different ways a WIMP might couple with protons and neutrons.

Scenarios of were also examined inelastic dark matterIn such models, the particle changes state during the collision, and some of the energy is used to produce a slightly heavier version of itself. To make this process possible, WIMPs need to be sufficiently fast and massive; the resulting recoils can therefore occupy energy regions ignored by standard analyses.

The anomalous event was reconstructed at 248 keV, with a statistical uncertainty of 23 keV and a systematic uncertainty of the same magnitude. Its morphology appears consistent with that of a single nuclear recoil, while its position falls within the internal volume selected to limit contamination from the walls and other components.

If the collision was produced by dark matter, the interpretations considered would generally require a heavy WIMP, with a mass of at least 200 GeV/c², two hundred times greater than that of the proton. Furthermore, it would not be the most frequently sought elementary interaction, but rather a coupling dependent on momentum, spin, or an inelastic transition. The signal, therefore, does not identify a specific particle: it is compatible with several theoretical constructions subjected to analysis.

From the local maximum to the global figure of about 2,6 sigma

For some models, the incompatibility between the observed event and the background-only hypothesis reaches locally 3,4 sigma. Taken in isolation, such a value corresponds to a very low probability of obtaining a result at least as extreme through fluctuation. But the experiment didn't verify a single predetermined prediction: it explored numerous combinations of operators, masses, and energy separations.

The more hypotheses are examined, the greater the chance that at least one will randomly produce an interesting discrepancy. This is the so-called look-elsewhere effectThe collaboration quantified it by building simulated samples with Monte Carlo techniques and virtually repeating the entire research. After this correction, the overall significance was found to be equal to 2,6 sigma, associated with a p-value in the order of 0,5 percent.

This does not mean there is a 99,5 percent probability that the event is dark matter. The p-value measures how unusual it would be to observe such a discrepancy assuming the underlying model is correct; it does not directly assign a probability to the different physical causes. The conventional threshold for claiming a discovery in particle physics is five sigma, much more restrictive.

The caution is accentuated by the fact that the entire excess depends on a single eventAnother interaction with similar characteristics would increase interest, but the significance would not increase automatically: the energy, the position in the detector, the compatibility with the same theoretical spectrum and the evolution of the expected background with the new exposure would be determining factors.

Rick Gaitskell, professor at Brown University and spokesperson for LZ, excluded premature interpretations in the note released by University of Sheffield:.

"We don't want to prejudge the conclusions with just one event. We're not claiming to have observed dark matter. But we have seen something interesting that we want to share with the scientific community to gather their input."

Inspections at the bottom have not yet found a cause

The collaboration spent months testing more commonplace explanations. The model includes internal beta decays, neutrinos, xenon isotopes, traces of tritium and carbon-14, radiogenic and cosmogenic neutrons, accidental coincidences between S1 and S2 signals, and configurations in which multiple scintillations can be associated with a single ionization.

Events produced near surfaces, multiple interactions erroneously reconstructed as single, and possible waveform effects were also evaluated. The 248 keV shock passed the applied selections and was not accompanied by signals in the veto systems that would have suggested a neutron had passed through. This reduces some known explanations, without demonstrating that any rare process or modeling flaw could have generated it.

Aaron Manalaysay, Lawrence physicist Berkeley National Laboratory and chairman of the LZ Institutional Board, noted that anomalous events are not unexpected, but usually reveal their origin after more thorough inspection.

"This is the first example, among all the experiments I've worked on, of an anomalous event that appears valid in every respect. We're still trying to understand whether we may have missed a rare underlying mechanism, but it's exciting to wonder if this could be the first hint of an observation of dark matter."

The situation illustrates a structural limitation of very low-count experiments. When the desired phenomenon is expected to manifest itself only a few times in years of operation, every single interaction acquires scientific weight; at the same time, a population of a single element does not allow us to reconstruct a spectrum, verify a temporal distribution, or compare multiple events with each other.

A single event is not enough to reconstruct a physical signal

Sam Eriksen, senior researcher atUniversity of Bristol and first author of the study, explained that the analysis explored a previously unexamined region within the same data sample. The team spent additional months reconstructing all known causes capable of producing underlying events.

“We know our detector and its components so well that even a single unexplained event, like the one we found, is important. We expect dark matter interactions to be extremely rare, so even a few events could mark the first detection of WIMP dark matter.”

Independent verification could come from other xenon detectors, including XENONnT at the Gran Sasso National Laboratories. Comparability, however, will depend on the size of the respective energy windows, thresholds, exposures, and models analyzed. A similar event observed by a different apparatus would be particularly valuable because it would reduce the probability of a specific instrumental origin of the LZ.

Dan Tovey, professor and head of the LZ group of theSheffield University, emphasized that the signal has so far passed all tests conducted to determine whether it was due to more common processes. However, the only way to determine its nature is to examine a larger amount of data.

"With just one event, it's impossible to conclude that we're actually observing the first signs of new physics. A tremendous amount of work has been done to verify that the event isn't the result of more common processes, and so far it has passed every test. The only way to be certain is to analyze more data."

LZ continues to accumulate observations in South Dakota and already has the largest sample ever produced by the WIMP-hunting experiment. Future data will reveal whether the anomaly will remain an isolated event, fade in significance with more precise knowledge of the background, or be followed by other recoils of comparable energy and characteristics. Only this possible repetition will transform a single xenon flash into a finding on the nature of dark matter.

Here are three insights that might interest you:

SNOLAB, two kilometers underground, science passes through data
Euclid, the ESA mission to hunt for the dark forces of the universe
Those "dark realities" full of (new) mysteries about the cosmos