XENON
XENON is a project searching for one of the most elusive substances in the universe: dark matter. Deep beneath the Gran Sasso mountains in Italy, physicists have spent years building progressively larger and more sensitive detectors, each one pushing closer to a signal that has never been directly observed. The collaboration is led by Elena Aprile, an Italian professor of physics at Columbia University. What they are hunting is a hypothetical particle called a WIMP, a weakly interacting massive particle, thought to be a prime candidate for the invisible matter that makes up most of the universe's mass. In June 2020, one of their detectors picked up an excess of electron recoils, 285 events when only 232 were expected, with a statistical significance of 3.5 sigma. The physics world held its breath. Was this the long-awaited signal? The answer, and the technology behind the question, reveals how scientists search for something that has never been seen.
Gran Sasso National Laboratory sits beneath roughly 3,100 meters of water-equivalent rock shielding, which blocks the constant rain of cosmic rays that would otherwise swamp any sensitive detector. The XENON experiments use liquid xenon as their detection medium, choosing it partly because natural xenon contains isotopes with odd nuclear spin states. Xenon-129 is present at 26% abundance with a spin of one-half; xenon-131 runs at 21% with a spin of three-halves. That spin structure makes xenon sensitive not just to the most common type of WIMP interaction but also to spin-dependent couplings, expanding the range of dark matter models the detectors can probe. When a particle passes through the liquid xenon volume, it can collide with a xenon nucleus and produce a nuclear recoil. That recoil is the signature researchers are watching for, buried beneath a far larger number of ordinary radioactive background events.
A particle interaction in the liquid xenon produces two distinct signals, and it is their relationship that gives the detector its discriminating power. The first is prompt scintillation light at 178 nanometers in the ultraviolet, collected by photomultiplier tubes and called the S1 signal. An applied electric field then prevents electrons produced in the collision from recombining; instead the field drifts them upward through the liquid toward a gas layer above. There, a stronger electric field accelerates the electrons until they produce a secondary flash of proportional scintillation light, called the S2 signal. The technique is sensitive enough to detect the S2 produced by a single drifting electron. The time delay between S1 and S2 reveals how deep in the detector the collision occurred, because electrons travel through liquid xenon at a uniform drift velocity. The pattern of photons across the individual photomultiplier tubes pins down the horizontal position. That full three-dimensional location matters because the innermost volume of the detector is surrounded by xenon that acts as its own shield, reducing background rates through self-shielding. Backgrounds from electronic recoils can be suppressed by more than 99% while retaining 50% of the nuclear recoil events, using the ratio of S2 to S1 as a discrimination parameter.
XENON10 was installed at Gran Sasso in March 2006, carrying just 15 kilograms of liquid xenon in a sensitive volume 20 centimeters in diameter and 15 centimeters in height. It was intended as a prototype to prove the design concept, check the achievable detection threshold, and measure background rejection. An analysis covering 59 live days of data taken between October 2006 and February 2007 found no WIMP signatures, but it did set the world's most stringent restrictions on spin-dependent WIMP coupling to neutrons at that time. Its successor, XENON100, arrived at Gran Sasso in 2008, installed in the same shield. It held 165 kilograms of liquid xenon in total, with 62 kilograms in the active target and the remainder serving as an active veto surrounding that target. Before construction, every component was screened for radioactivity using high-purity germanium detectors, and in some cases mass spectrometry was applied to low-mass plastic samples. That painstaking campaign achieved a background rate below 10 to the minus 2 events per kilogram per day per keV, at the time the lowest rate ever achieved in a dark matter search. By 2012 the experiment had set the most stringent spin-independent WIMP-nucleon limit then known, with a minimum at a cross section of 2.0 for a WIMP mass of 65 GeV. A separate result published in 2014 set a new best limit on axions, a different class of hypothetical particle.
Construction of XENON1T began in Hall B of Gran Sasso in 2014, and the detector represents a dramatic scale-up: 3.2 tons of ultra radio-pure liquid xenon in a time projection chamber 1 meter in diameter and 1 meter in height, sitting inside a 10-meter water tank that serves as a muon veto. The fiducial volume, the inner low-background region used for the actual dark matter search, contains roughly 2 tons. The XENON Collaboration that built and operates it includes 135 investigators from 22 institutions spanning Europe, the Middle East, and the United States. First results, released on the 18th of May 2017, came from just 34 days of data taken between November 2016 and January 2017. No dark matter signal appeared, but the detector posted a record low background and its exclusion limits surpassed those of the previous world leader, the LUX experiment, excluding cross sections larger than 7.7 times 10 to the minus 47 square centimeters for WIMP masses of 35 GeV. By September 2018, with 278.8 days of data in hand, the collaboration set a new record minimum of 4.1 times 10 to the minus 47 square centimeters at a WIMP mass of 30 GeV. Then in April 2019, published in Nature, the collaboration announced something unexpected: the first direct observation of two-neutrino double electron capture in xenon-124 nuclei. The measured half-life of that process is several orders of magnitude longer than the age of the universe, demonstrating that the detectors can catch extraordinarily rare events far beyond the dark matter search itself.
In June 2020, with data-taking for XENON1T already concluded in preparation for its successor, the collaboration published a striking anomaly. The detector had recorded 285 electron recoil events in a particular energy range, against an expected background of 232, a surplus of 53 events at 3.5 sigma significance. Three explanations dominated early discussion: solar axions, an unexpectedly large magnetic moment for neutrinos, and tritium contamination in the xenon. Later, other groups put forward additional interpretations, including candidates called chameleons, particles associated with dark energy rather than dark matter. The successor detector, XENONnT, resolved the question in July 2022 when a new analysis found no excess, effectively ruling out a new-physics origin for the signal seen by XENON1T.
XENONnT is an upgrade of XENON1T built in the same underground hall at Gran Sasso. Its total xenon mass exceeds 8 tonnes, and beyond simply scaling the target it incorporates new systems designed to further reduce or tag background radiation that would otherwise obscure the signal. The upgrade was still under way in early 2019 when first light was anticipated for 2020. Even with the disruptions of the COVID-19 pandemic, the project completed construction by mid-2020 and began full detector operations by the end of that year. By September 2021 the detector was collecting science data in its first science run. On the 28th of July 2023, XENONnT published its first WIMP search results, excluding spin-independent cross sections above a certain threshold at 28 GeV with 90% confidence; on that same date the competing LZ experiment published its own first results, excluding cross sections above a comparable threshold at 36 GeV. The XENONnT design is ambitious enough that at some point in the mass range it probes, neutrinos themselves become a significant source of background, marking a frontier where the instruments are sensitive enough that the universe's own neutrino flux begins to obscure the search for something even more elusive.
Common questions
What is the XENON dark matter experiment and where is it located?
XENON is a deep underground detector facility at the Gran Sasso National Laboratory in Italy, designed to search for dark matter particles called WIMPs (weakly interacting massive particles) using liquid xenon as a target. The underground location provides 3,100 meters of water-equivalent shielding against cosmic ray backgrounds. The project is led by Elena Aprile, an Italian professor of physics at Columbia University.
How does the XENON detector work to find dark matter?
The XENON detector uses a dual phase time projection chamber filled with liquid xenon. When a particle collides inside the liquid, it produces two signals: a prompt ultraviolet scintillation flash (S1) and a secondary electroluminescence flash (S2) created by electrons drifted into a gas layer above. The ratio of S2 to S1 distinguishes nuclear recoils (expected from WIMPs) from electronic recoils (background), suppressing electronic recoil backgrounds by more than 99% while retaining 50% of nuclear recoil events.
What did XENON1T discover with its 2020 electron recoil excess?
In June 2020, XENON1T reported 285 electron recoil events where only 232 were expected, a surplus of 53 events at 3.5 sigma significance. The three main candidate explanations were solar axions, an unexpectedly large neutrino magnetic moment, and tritium contamination. In July 2022, the successor detector XENONnT found no such excess in new data, discarding the anomaly.
What rare nuclear process did XENON1T observe for the first time?
In April 2019, the XENON Collaboration reported in Nature the first direct observation of two-neutrino double electron capture in xenon-124 nuclei. The measured half-life of this process is several orders of magnitude larger than the age of the universe, demonstrating the sensitivity of liquid xenon detectors to extremely rare events.
How much xenon does the XENONnT detector contain compared to earlier XENON experiments?
XENONnT contains more than 8 tonnes of xenon in total. This compares to 3.2 tonnes in XENON1T, 165 kilograms in XENON100, and just 15 kilograms in the original XENON10 prototype installed in 2006.
When did XENONnT publish its first WIMP search results?
XENONnT published its first WIMP search results on the 28th of July 2023, excluding spin-independent cross sections above a set threshold at 28 GeV with 90% confidence. On the same date, the competing LZ experiment independently published its own first results, excluding cross sections above a comparable threshold at 36 GeV.