Physicists have spotted elusive subatomic particles called neutrinos pinging off atomic nuclei in a way predicted more than 40 years ago, but never before observed. Even more remarkably, they spotted the scattering effect not with a giant detector weighing thousands of tons, but with a device the size of a milk jug. The advance could open the way to portable neutrino detectors that could monitor nuclear facilities and, for example, sniff out neutrinos created in the production of plutonium.

“It’s a real thrill that something that I predicted 43 years ago has been realized experimentally,” says Daniel Freedman, a theoretical physicist emeritus at the Massachusetts Institute of Technology in Cambridge, who in 1974 laid out the theory of the effect, called coherent elastic neutrino-nucleus scattering. The observation doesn’t change physicists’ understanding of the nucleus or neutrinos, says Natalie Roe, an experimentalist at Lawrence Berkeley National Laboratory in Berkeley, California, who was not involved in the study. Still, she says, “it’s a tour de force to dig this tiny signal out.”

Neutrinos come in three types—electron, muon, and tau—and interact with atomic nuclei in a few ways. For example, a muon neutrino can strike a neutron in a nucleus, transforming it into a proton while itself turning into a muon—a heavier cousin of the electron—in so-called “quasi-elastic scattering.” Or a neutrino can simply bounce off a nucleus while retaining its type in plain “elastic scattering.” All such interactions are exceedingly rare, but they provide the only means to observe neutrinos. To detect just a few of the trillions of electron neutrinos that pass from the sun through every square meter of Earth’s surface every second, physicists deploy detectors weighing kilotons, upping the number of nuclei in them—and the chances that the neutrinos will strike one. As a rule, the probability of a neutrino interaction increases with the number of protons and neutrons a nucleus.

However, Freedman realized there should be an exception to the rule. Like any quantum particle, a neutrino acts like a wave with a wavelength that grows shorter as the energy of the particle increases. If the neutrino's energy is high, the neutrino will interact with a single proton or neutron. But if a low-energy neutrino has a wavelength that’s as long as the nucleus is wide, it will interact with all the protons and neutrons in concert. Thanks to that “coherence,” the probability that the neutrino will bounce off the nucleus increases, roughly speaking, with the number of protons and neutrons squared, leading to a big increase in the number of interactions.

That means there should be a lot more elastic scattering at low energies, and for decades physicist have tried to spot it in experiments at nuclear reactors. But there’s a catch. The only signal is the recoil of the nucleus, and the low-energy neutrino gives it only a feeble, hard-to-detect kick. “It’s like hitting a bowling ball with a ping pong ball,” says Kate Scholberg, a physicist at Duke University in Durham, North Carolina. “You can hit the bowling ball pretty easily, but it isn’t going to roll very fast.”

Nevertheless, Scholberg and the 80 other members of the COHERENT collaboration have now detected coherent neutrino scattering, as they report today in Science. To do so, they employed a sensitive detector consisting of crystals of cesium iodide doped with sodium and weighing just 14.6 kilograms. When something sends a nucleus recoiling in a crystal, even just a bit, the crystal produces a small, but detectable flash of light. To generate the neutrinos, COHERENT physicists relied on the Spallation Neutron Source (SNS) at Oak Ridge National Laboratory in Tennessee.

The SNS generates the world’s most intense beams of neutrons for material science research, but it also radiates copious neutrinos as a byproduct. The neutrinos have a slightly higher energy than those from a nuclear reactor, making the SNS just the right source for the experiment, Scholberg says. The neutrinos’ energy is still low enough to produce coherent scattering, but high enough to produce detectable signals in the crystals, she says. Using more than 461 days’ worth of data, the researchers observed 134 neutrino scattering events, in good agreement with the predictions for coherent scattering.

The COHERENT team has beaten others to the punch. That includes a group led by Leo Stodolsky, a physicist and director emeritus at the Max Planck Institute for Physics in Munich, Germany, which is developing cryogenic detectors that can spot nuclear recoils with far lower energies. They hope to detect coherent scattering at a nuclear reactor. “My colleagues and I have been going over this paper, hoping to find something wrong with it,” Stodolsky quips. “But we haven’t been able to find anything.”

The new observation won’t rewrite the textbooks on nuclear or particle physics. Indeed, it would have been far more revolutionary if physicists had somehow proved that coherent scattering didn’t exist, Freedman says, as that would have meant that the bedrock rules of quantum mechanics were somehow wrong.

The real significance of coherent scattering may lie in the potential applications of small portable neutrino detectors, Stodolsky says. Such detectors might be used to monitor nuclear reactors for safety and security. For example, from the details of the neutrino flux, observers might be able to tell whether a reactor is being used to generate plutonium, he says.

But slimmed-down detectors won’t work in all big neutrino experiments, however. For example, 2 weeks ago physicists started excavation for the Deep Underground Neutrino Experiment (DUNE), a 70,000-ton detector in South Dakota that will study neutrinos shot from a particle accelerator 1300 kilometers away in Illinois. The DUNE will study how the three types of neutrinos morph into one another as the particles zip along—a phenomenon called neutrino mixing that could ultimately help explain how the universe generated so much matter and so little antimatter. There is no hope of replacing the DUNE with a smaller detector based on coherent scattering, because the interaction doesn’t reveal what type of neutrino bounced off a nucleus—essential information for studying neutrino mixing.