Researchers grapple with XENON1T excess

2 July 2020
The XENON1T detector
XENON1T differentiates nuclear and electronic recoils by comparing two signals: initial scintillation light and ionisation charges drifting to the top of the detector. This photograph shows the lower array of photomultiplier tubes and the copper coils which generate the drift field. Credit: XENON collaboration.

An intriguing low-energy excess of background events recorded by the world’s most sensitive WIMP dark-matter experiment has sparked a series of preprints speculating on its underlying cause. On 17 June, the XENON collaboration, which searches for excess nuclear recoils in the XENON1T detector, a one-tonne liquid-xenon time-projection chamber (TPC) located underground at Gran Sasso National Laboratory in Italy, reported an unexpected excess in electronic recoils at energies of a few keV, just above its detection threshold. Though acknowledging that the excess could be due to a difficult-to-constrain tritium background, the collaboration says solar axions and solar neutrinos with a Majorana nature, both of which would signal physics beyond the Standard Model, are credible explanations for the approximately 3σ effect.

Who needs the WIMP if we can have the axion?

Elena Aprile

“Thanks to our unprecedented low event rate in electronic recoils background, and thanks to our large exposure, both in detector mass and time, we could afford to look for signatures of rare and new phenomena expected at the lowest energies where one usually finds lots of background,” says XENON spokesperson Elena Aprile, of Columbia University in New York. “I am especially intrigued by the possibility to detect axions produced in the Sun,” she says. “Who needs the WIMP if we can have the axion?”

The XENON collaboration has been in pursuit of WIMPs, a leading bosonic cold-dark-matter candidate, since 2005 with a programme of 10 kg, 100 kg and now 1 tonne liquid-xenon TPCs. Particles scattering in the liquid xenon create both scintillation light and ionisation electrons; the latter drift upwards in an electric field towards a gaseous phase where electroluminescence amplifies the charge signal into a light signal. Photomultiplier tubes record both the initial scintillation light and the later electroluminescence, to reveal 3D particle tracks, and the relative magnitudes of the two signals allows nuclear and electronic recoils to be differentiated. XENON1T derives its world-leading limit on WIMPs – the strictest 90% confidence limit being a cross-section of 4.1×10−47 cm2 for WIMPs with a mass of 30 GeV – from the very low rate of nuclear recoils observed by XENON1T from February 2017 to February 2018.

XENON1T low-energy electronic recoils

A surprise was in store, however, in the same data set, which also revealed 285 electronic recoils at the lower end of XENON1T’s energy acceptance, from 1 to 7 keV, over the expected background of 232±15. The sole background-modelling explanation for the excess that the collaboration has not been able to rule out is a minute concentration of tritium in the liquid xenon. With a half-life of 12.3 years and a relatively low amount of energy liberated in the decay of 18.6 keV, an unexpected contribution of tritium decays is favoured over XENON1T’s baseline background model at approximately 3σ. “We can measure extremely tiny amounts of various potential background sources, but unfortunately, we are not sensitive to a handful of tritium atoms per kilogram,” explains deputy XENON1T spokesperson Manfred Lindner, of the Max Planck Institute for Nuclear Physics in Heidelberg. Cryogenic distillation plus running the liquid xenon through a getter is expected to remove any tritium below the level that would be relevant, he says, but this needs to be cross-checked. The question is whether a minute amount of tritium could somehow remain in liquid xenon or if some makes it from the detector materials into the liquified xenon in the detector. “I personally think that the observed excess could equally well be a new background or new physics. About 3σ implies of course a certain statistical chance for a fluctuation, but I find it intriguing to have this excess not at some random place, but towards the lower end of the spectrum. This is interesting since many new-physics scenarios generically lead to a 1/E or 1/E2 enhancement which would be cut off by our detection threshold.”

Solar axions

One solution proposed by the collaboration is solar axions. Axions are a consequence of a new U(1) symmetry proposed in 1977 to explain the immeasurably small degree of CP violation in quantum chromodynamics – the so-called strong CP problem — and are also a dark-matter candidate. Though XENON1T is not expected to be sensitive to dark-matter axions, should they exist they would be produced by the sun at energies consistent with the XENON1T excess. According to this hypothesis, the axions would be detected via the “axioelectric” effect, an axion analogue of the photoelectric effect. Though a good fit phenomenologically, and like tritium favoured over the background-only hypothesis at approximately 3σ, the solar-axion explanation is disfavoured by astrophysical constraints. For example, it would lead to a significant extra energy loss in stars.

Axion helioscopes such as the CERN Axion Solar Telescope (CAST) experiment, which directs a prototype LHC dipole magnet at the Sun and could convert solar axions into X-ray photons, will help in testing the hypothesis. “It is not impossible to have an axion model that shows up in XENON but not in CAST,” says deputy spokesperson Igor Garcia Irastorza of University of Zaragoza, “but CAST already constraints part of the axion interpretation of the XENON signal.” Its successor, the International Axion Observatory (IAXO), which is set to begin data taking in 2024, will have improved sensitivity. “If the XENON1T signal is indeed an axion, IAXO will find it within the first hours of running,” says Garcia Irastorza.

A second new-physics explanation cited for XENON1T’s low-energy excess is an enhanced rate of solar neutrinos interacting in the detector. In the Standard Model, neutrinos have a negligibly small magnetic moment, however, should they be Majorana rather than Dirac fermions, and identical to their antiparticles, their magnetic moment should be larger, and proportional to their mass, though still not detectable. New physics beyond the Standard Model could, however, enhance the magnetic moment further. This leads to a larger interaction cross section at low energies and an excess of low-energy electron recoils. XENON1T fits indicate that solar Majorana neutrinos with an enhanced magnetic moment are also favoured over the background-only hypothesis at the level of 3σ.

The absorption of dark photons could explain the observed excess.

Joachim Kopp

The community has quickly chimed in with additional ideas, with around 40 papers appearing on the arXiv preprint server since the result was released. One possibility is a heavy dark-matter particle that annihilates or decays to a second, much lighter, “boosted dark-matter” particle which could scatter on electrons via some new interaction, notes CERN theorist Joachim Kopp. Another class of dark-matter model that has been proposed, he says, is “inelastic dark matter”, where dark-matter particles down-scatter in the detector into another dark-matter state just a few keV below the original one, with the liberated energy then seen in the detector. “An explanation I like a lot is in terms of dark photons,” he says. “The Standard Model would be augmented by a new U(1) gauge symmetry whose corresponding gauge boson, the dark photon, would mix with the Standard-Model photon. Dark photons could be abundant in the Universe, possibly even making up all the dark matter. Their absorption in the XENON1T detector could explain the observed excess.”

“The strongest asset we have is our new detector, XENONnT,” says Aprile. Despite COVID-19, the collaboration is on track to take first data before the end of 2020, she says. XENONnT will boast three times the fiducial volume of XENON1T and a factor six reduction in backgrounds, and should be able to verify or refute the signal within a few months of data taking. “An important question is if the signal has an annual modulation of about 7% correlated to the distance of the sun,” notes Lindner. “This would be a strong hint that it could be connected to new physics with solar neutrinos or solar axions.”

bright-rec iop pub iop-science physcis connect