Among physicists, those studying elusive particles called neutrinos may set the standard for dogged determination—or obdurate stubbornness. For 12 years, scientists in Japan have fired trillions of neutrinos hundreds of kilometers through Earth to a gigantic subterranean detector called Super-Kamiokande (Super-K) to study their shifting properties. Yet the nearly massless particles interact with other matter so feebly that the experiment, known as T2K, has captured fewer than 600 of them.
Nevertheless, so alluring are neutrinos that physicists are not just persisting, they are planning to vastly scale up efforts to make and trap them. At stake may be insight into one of the most profound questions in physics: how the newborn universe generated more matter than antimatter, so that it is filled with something instead of nothing.
That prospect, among others, has sparked a race to build two massive subterranean detectors, at costs ranging from hundreds of millions to billions of dollars. In an old zinc mine near the former town of Kamioka in Japan, physicists are gearing up to build Hyper-Kamiokande (Hyper-K), a gargantuan successor to Super-K, which will scrutinize neutrinos fired from a particle accelerator at the Japan Proton Accelerator Research Complex (J-PARC) in Tokai 295 kilometers away. In the United States, scientists are developing the Deep Underground Neutrino Experiment (DUNE) in a former gold mine in Lead, South Dakota, which will snare neutrinos from Fermi National Accelerator Laboratory (Fermilab) 1300 kilometers away in Batavia, Illinois.
Researchers with both experiments acknowledge they’re in competition—and that Hyper-K may have an advantage because it will likely start to take data a year or two before DUNE. Yet aside from their goals, “Hyper-K and DUNE are vastly different,” says Chang Kee Jung, a neutrino physicist at Stony Brook University and a T2K member who now also works on DUNE.
Hyper-K, which will be bigger but cheaper than DUNE, represents the next in a series of ever larger neutrino detectors of the same basic design developed over 40 years by Japanese physicists. It is all but certain to work as expected, says Masato Shiozawa, a particle physicist at the University of Tokyo and co-spokesperson for the 500-member Hyper-K collaboration. “Hyper-K is a more established technology than DUNE,” he says. “That is why I proposed it.”
DUNE will employ a relatively new technology that promises to reveal neutrino interactions in stunning detail and allow physicists to test their understanding of the particles with unprecedented rigor. “Without bragging too much, we are best in class,” says Sergio Bertolucci, a particle physicist at the University of Bologna and Italy’s National Institute for Nuclear Physics and co-spokesperson for the 1300-member DUNE collaboration. However, that technological edge comes with a hefty price tag and, Bertolucci acknowledges, more risk.
How the rivalry plays out will depend on factors as mundane as the cost of underground excavation and as exciting as the possibility that neutrinos, always quirky, hold some surprise that will transform physicists’ understanding of nature.
The most common particles in the universe besides photons, neutrinos exert no effect on the everyday objects around us. Yet they could carry clues to deep mysteries. Neutrinos and their antimatter counterparts both come in three types or flavors—electron, muon, and tau—depending on how they’re generated. For example, electron neutrinos emerge from the radioactive decay of some atomic nuclei. Muon neutrinos fly from the decays of fleeting particles called pi-plus mesons, which can be produced by smashing a beam of protons into a target. These identities aren’t fixed: A neutrino of one type can change into another, chameleonlike, as it zips along at near–light-speed.
Weirdly, a neutrino of a definite flavor has no definite mass. Rather, it is a quantum mechanical combination of three different “mass states.” For example, a decaying pi-plus spits out the combination of mass states that makes a muon neutrino. However, like gears turning at different speeds, the mass states evolve at different rates, changing that combination. So, a particle that began as an electron neutrino might later appear as a tau neutrino—a phenomenon known as neutrino oscillation.
Theorists can explain all of this with a mathematical clockwork known as the three-flavor model. It has just a handful of parameters: roughly speaking, the probabilities with which one flavor will oscillate into another and the differences among the three mass states. The picture has gaps. Experiments show two mass states are close, but not whether the two similar states are lighter or heavier than the third—a puzzle known as the hierarchy problem.
Moreover, neutrinos and antineutrinos might oscillate by different amounts, an asymmetry called charge-parity (CP) violation. Measuring that asymmetry is the prize physicists seek, as it could help explain how the soup of fundamental particles in the early universe generated more matter than antimatter.
The wispy neutrinos themselves did not tilt the matter-antimatter balance. Rather, according to some theories, the familiar neutrinos are mirrored by vastly heavier “sterile” neutrinos that would interact with nothing except neutrinos. If sterile neutrinos and antineutrinos also behave asymmetrically, then in the early universe their decays could have generated more electrons than antielectrons (also called positrons), seeding the dominance of matter.
Seeing CP violation in ordinary neutrinos wouldn’t prove this scenario played out, notes Patrick Huber, a theorist at Virginia Polytechnic Institute and State University. But not seeing CP violation among ordinary neutrinos would render it much less likely that the hypothetical heavyweights possess the key asymmetry, Huber says. “It’s not impossible, but it’s implausible,” he says.
But, first, scientists must determine whether neutrinos really exhibit this asymmetry. The teams in Japan and the United States will both employ a well-established technique to probe neutrino behavior. By smashing energetic protons from a particle accelerator into a target to produce pipluses, they will generate a beam of muon neutrinos and shoot it toward a distant underground detector. There, researchers will count the surviving muon neutrinos and the electron neutrinos that have emerged along the way. Then they will switch to producing a beam of muon antineutrinos, by collecting pi-minuses instead of pi-pluses from the target. They’ll repeat the measurements, looking for any differences.
The experiment is much harder than it sounds, as several other factors could create a spurious asymmetry. For example, the neutrino and antineutrino beams will inevitably differ slightly, both in intensity and in their energy spectrum. To account for such differences, the researchers must sample the particles as they start their journey by placing a small detector, preferably with a design as similar as possible to the distant detector, in front of the beam source.
The physics of neutrinos themselves could also skew the results. For example, either neutrinos or antineutrinos will be absorbed more strongly by the matter they traverse on their flight to the detector. The direction of that effect depends on the solution to the hierarchy problem. So, to spot CP violation, physicists will most likely have to solve the hierarchy problem, too.
The biggest barrier to sorting all of this out, however, has been the measly harvest of neutrinos from even the biggest experiments. Like their counterparts in Japan, U.S. physicists already have a neutrino-oscillation experiment, NOνA, which shoots neutrinos from Fermilab to a detector 810 kilometers north in Minnesota. Like T2K, it has netted just several hundred neutrinos.
Hyper-K will tackle the scarcity primarily by providing a much bigger target for the neutrinos to hit. Proposed a decade ago, it’s a scaled-up version of the storied Super-K detector and will consist of a cylindrical stainless steel tank 78 meters tall and 74 meters wide, holding 260,000 tons of ultrapure water—five times as much as Super-K.
To spot neutrinos, the detector will rely on the optical equivalent of a sonic boom. Rarely, a muon neutrino zipping through the water will knock a neutron out of an oxygen atom and change it into a proton, while the neutrino itself morphs into a high-energy muon. The fleeing muon will actually exceed the speed of light in water, which is 25% slower than in a vacuum, and generate a shock wave of so-called Cherenkov light, just as a supersonic jet creates a shock wave of sound. That conelike shock wave will cast a ring of light on the tank’s side, which is lined with photodetectors.
Similarly, an electron neutrino can strike a neutron to produce a high-speed electron, which is lighter than a muon and will be buffeted more by the water molecules. The result will be a fuzzier light ring. Muon and electron antineutrinos can spawn detectable antimuons and antielectrons by striking protons, although with about half the efficiency of the neutrino interactions.
Hyper-K will be Japan’s third great detector, all in the same mining area. From 1983 to 1995, the Kamioka Nucleon Decay Experiment (Kamiokande), a 3000-ton detector, tried to spot the ultrarare decays of protons that some theories predict. Instead, in 1987, it glimpsed neutrinos from a supernova—an advance that won a share of the Nobel Prize in Physics in 2002. In 1996, Super-K came online. It proved neutrinos oscillate by studying muon neutrinos generated when cosmic rays strike the atmosphere. Fewer come up from the ground than down from the sky, showing that those traversing Earth change flavor along the way. The discovery shared the Nobel in 2015. “It’s spectacular what [Japanese physicists] have done,” says Erin O’Sullivan, a neutrino astrophysicist at Uppsala University and a Hyper-K member who was drawn by “the dynasty of Super-K.”
Hyper-K will reuse J-PARC’s neutrino beam, which is now being upgraded to increase its power by a factor of 2.5. Overall, it should collect data at 20 times the rate of T2K, says Stephen Playfer, a particle physicist at the University of Edinburgh and the University of Tokyo and the project’s lead technical coordinator. Before joining Hyper-K in 2014, he and his Edinburgh colleagues also considered joining DUNE. “When it came to comparing who was going to be first to see something, we thought Hyper-K was in a good position, just because it would have the statistics and it had a well-known technology,” he says.
Hyper-K will have limitations. In particular, it won’t measure the neutrinos’ energies precisely. That matters because the rate at which a neutrino oscillates depends on its energy, and a beam contains neutrinos with a range or spectrum of energies. Without a way to pinpoint each neutrino’s energy, the experiment would be unable to make sense of the oscillation rates.
To avoid this problem, Hyper-K, like current experiments, will rely on a trick. A neutrino beam naturally diverges, with lower energy neutrinos spreading more than higher energy ones. Thus, if a detector sits slightly to the side of the beam’s path, it will see neutrinos with a narrower range of energies that should oscillate at roughly the same rate. So, like Super-K, Hyper-K will sit off the beam axis by an angle of 2.5°.
Physicists can then tune the energy of the beam so the neutrinos reach the detector when the oscillation is at its maximum. With the neutrinos’ energy constrained, physicists basically count the number of arriving muon neutrinos, electron neutrinos, and their antimatter counterparts. Hyper-K’s CP measurement comes down to comparing two ratios: electron neutrinos with muon neutrinos and electron antineutrinos with muon antineutrinos.
Workers have already begun the excavation for Hyper-K, which should take 2 years, Shiozawa says. The whole project will cost Japan about $600 million, with international partners chipping in an additional $100 million to $200 million, he says. The detector will be complete in 2027, Shiozawa says, and will start taking data a year later. So confident are Hyper-K researchers in their technology that they say the trickiest part of the project is the digging. “We need to construct probably the largest underground cavern” in the world, Shiozawa says. “In terms of technology and also cost, this is the biggest challenge.”
If, technologically, Hyper-K amounts to much more of the same, DUNE aims to be something almost completely different. It will employ a technology that, until recently, was used in only one other large experiment but that should enable physicists to see neutrino interactions as never before. “To me, the draw of DUNE is its precision,” says Chris Marshall, a particle physicist at the University of Rochester and DUNE’s physics coordinator. “This is an experiment that will be world leading in just about everything that it measures.”
Hunkering 1480 meters deep in a repurposed gold mine, DUNE will consist of two rectangular tanks 66 meters long, 19 meters wide, and 18 meters tall. Each will contain 17,000 tons of frigid liquid argon cooled to below –186°C. Just as in a water-filled detector, a neutrino can blast a neutron—in this case in an argon nucleus—to create a muon or an electron. But the neutrinos reaching DUNE from Fermilab will pack up to 10 times more energy than those flowing to Hyper-K. So, in addition to the muon or electron, a collision will typically produce a spurt of other particles such as pions, kaons, protons, and neutrons.
DUNE aims to track all those particles—or at least the charged ones—with a technology called a liquid argon time projection chamber. As a charged particle streaks through the argon, it will ionize some of the atoms, freeing their electrons. A strong electric field will push the electrons sideways until they hit three closely spaced planes of parallel wires, each plane oriented in a different direction. By noting when the electrons strike the wires, physicists can reconstruct with millimeter precision the original particle’s 3D trajectory. And from the amount of ionization it produces, they can determine its type and energy.
The details are mind-boggling. The electrons will have to drift as far as 3.5 meters, driven by a voltage of 180 kilovolts. And unlike Hyper-K, DUNE will sit directly in the beam from Fermilab. So, it will capture a bigger but messier harvest of neutrinos, with energies ranging from less than 1 giga-electron volt to more than 5 GeV.
DUNE’s ability to precisely track all the particles should enable it to do something unprecedented in neutrino physics: Measure the energy of each incoming neutrino to construct energy spectra for each flavor of neutrino and antineutrino. Because of the flavor changing, a plot of each spectrum should itself exhibit a distinct wiggle or oscillation. By analyzing all of the spectra, physicists should be able to nail down the entire three-flavor model, including the amount of CP violation and the hierarchy, in one fell swoop, Bertolucci says. “It can measure all the parameters in the same experiment,” he says.
Until now, the technology has never been fully developed. Italian Nobel laureate Carlo Rubbia dreamed up the liquid argon detector in 1977. But it wasn’t until 2010 that one called ICARUS in Italy’s subterranean Gran Sasso National Laboratory snared a few neutrinos shot from the European particle physics laboratory, CERN, near Geneva, Switzerland. Researchers at Fermilab and CERN have embarked on a crash program to build prototypes, which have worked even better than expected, says Kate Scholberg, a neutrino physicist and a DUNE team member at Duke University. “It’s kind of an entrancing thing to look at the event displays” coming in, she says. “It’s just incredible detail.”
That precision comes at a price. For accounting purposes, the U.S. Department of Energy (DOE) splits the project in two. One piece, the Long Baseline Neutrino Facility (LBNF), includes the new neutrino beam at Fermilab and all the infrastructure. The second, DUNE, is an international collaboration that will build just the guts of the detectors. In 2015, DOE estimated LBNF/DUNE would cost $1.5 billion and come online in 2027. Last year, however, DOE reported that unexpected construction costs had raised the bill to $3.1 billion. The detector should be completed in 2028, says Christopher Mossey, project director for LBNF/DUNE-U.S. But the beam will lag until early 2031, potentially giving Hyper-K a head start of more than 2 years.
With contracts in hand and construction underway, DUNE developers are confident that the new cost and timeline will hold. Excavation has passed 40% and should be completed in May 2023, Mossey says. “We really are accomplishing big, tangible things.” Still, DUNE physicists acknowledge that the project is riskier than Hyper-K. “It’s a leap into the unknown, and that’s the trade-off you make,” Scholberg says. “Something that is more transformative is going to certainly entail more scariness.”
Both experiments have other scientific goals, such as searching for proton decay. There, Hyper-K has an advantage, Huber says, as it is simply bigger and contains lots of lone protons in the hearts of the hydrogen atoms in water molecules. Another tantalizing payoff could come if a giant star collapses and explodes as a supernova near our Galaxy, as one did in 1987. The experiments would provide complementary observations, Scholberg says, as DUNE would see the electron neutrinos produced just as the core implodes and Hyper-K would see mainly electron antineutrinos released later in the explosion.
Nevertheless, the raison d’être for both experiments remains deciphering neutrino oscillations and searching for CP violation. So, how would a gambler handicap this race?
Given their head start, Hyper-K physicists could score major discoveries before DUNE even finds its feet. If CP violation is as big as it can possibly be, “then we may discover it in 3 years,” Shiozawa says. “And also, we may discover proton decay in 3 years.” But, he says, “It really depends on nature.”
Hyper-K is optimized to measure CP violation assuming it’s big and the three-flavor model is the final word on neutrino oscillations, Huber notes. Neither assumption may hold. And with its simpler counting technique and shorter baseline, the experiment may struggle to distinguish CP violation from the matter effect unless some other experiment independently solves the hierarchy problem. “Hyper-K certainly requires more external inputs,” Huber says.
DUNE, in contrast, should be able to disentangle the whole mess on its own. Shiozawa, for one, is not counting out his rival. The Japanese project has experienced growing pains of its own, he notes, including being scaled back from an initial 1-million-ton design. And the Japanese government won’t countenance any cost increase, putting project leaders in constant tension with contractors, he says. “The situation is not so different between the two projects.”
Ultimately, the rivalry between Hyper-K and DUNE may be less a dash for glory than a decadeslong slog through uncertainty. If so, the two teams could end up collaborating as much as they compete, at least informally. “We’ll have a long time where the most accurate results will come from a combination of the two [experiments],” Huber predicts.
Most tantalizing, instead of completing the current theory, the results could upend it. They might reveal deviations from the three-flavor model that could hint at new particles and phenomena lurking in the vacuum. After all, neutrinos have repeatedly surprised physicists, who once assumed that the particles came in only one type and were completely massless and inert. “Previously, neutrino experiments have taught us that we very rarely take data in a neutrino beam and get exactly what we expect,” Marshall says.
The unexpected may be a long shot worth betting on.