For the first time, scientists have seen neutrinos originating from the central disk of the Milky Way.
Neutrinos are tiny, deeply weird particles that can zip through matter with nary a ripple. Because of their lack of interactions, they’re hard to detect—but also promising for revealing new secrets of the universe. In particular, Milky Way neutrinos may help scientists understand the origin of high-energy particles known as cosmic rays, which kick off the formation of neutrinos. And because neutrinos are particles outside of the electromagnetic spectrum, they’re like a new light-independent lens through which to study the galaxy’s structure, says Naoko Kurahashi Neilson, a physicist at Drexel University, who came up with the new method that allowed researchers to make the discovery.
“Now we see, for the first time, our galaxy in something other than light,” Kurahashi Neilson says. The team reported its findings today in the journal Science.
Neutrinos can be formed when the high-energy radiation that makes up most cosmic rays interacts with matter, creating charged particles called pions, which produce neutrinos as they decay. This process, it’s thought, is constantly churning out neutrinos in the matter-dense, cosmic-ray-bathed disk of the Milky Way. Like all neutrinos, those thought to emerge from the Milky Way’s disk are so insubstantial as to be ghostlike: They have a neutral charge, a mass that is so small that scientists still don’t know exactly how much these particles weigh, and they barely interact with matter or electromagnetic fields at all, even as they travel long distances across the universe at nearly the speed of light. This lack of interaction makes investigating neutrinos a promising way to study the cosmic rays that birthed them. Many cosmic rays are in fact extremely high-energy photons—gamma rays—that can be absorbed by interstellar or intergalactic matter as they travel through space. In contrast, neutrinos are like a time capsule of their own formation, bearing few if any imprints from their subsequent travels but hopefully some lingering evidence to enlighten scientists seeking cosmic rays’ deepest astrophysical origins—which are still unknown.
That’s where the IceCube experiment comes in. For the past 10 years, an array of small light sensors drilled into Antarctic ice has been detecting neutrinos as they zip through our planet. IceCube is an actual cube of these sensors, a kilometer long on each side, that was sunk between 1.5 and 2.5 km deep in the ice. In this translucent medium, the sensors pick up tiny flashes of so-called Cherenkov radiation that forms when a vanishingly rare neutrino hits the ice and creates a shower of secondary particles. Physicists can also create neutrinos in particle accelerators on Earth to inform their studies, says Anthony Ezeribe, a physicist at the University of Sheffield in England, who was not involved in the new paper. Some neutrinos from space zing in at higher energy levels than any from a lab, however, making their physics important to study.
IceCube had already definitively detected neutrinos streaming in from outside the Milky Way, but it couldn’t be said with certainty that any of them came from within the galaxy, says Francis Halzen, lead investigator of the project and a physicist at the University of Wisconsin–Madison. This was rather strange, considering the proximity of the Milky Way’s disk (in fact, our solar system is embedded in it) and the high likelihood that neutrinos form there.
The problem, though, was one of location. Most of the neutrinos that zip through IceCube are homegrown particles that form when cosmic rays hit Earth’s atmosphere. These atmospheric neutrinos trigger the detector a few thousand times a second, says Stephen Sclafani, now a postdoctoral researcher at the University of Maryland, who worked on the IceCube collaboration when he was a doctoral student at Drexel. In comparison, the interesting astrophysical neutrinos only pop up about once a day.
IceCube is in the Southern Hemisphere, and Earth’s bulk actually filters out a lot of this atmospheric noise when it comes from the northern half of the sky. But the Milky Way’s disk is largely situated in the Southern Hemisphere’s skies, too, making for a very noisy environment—the equivalent of trying to pick out a single voice from a football stadium’s worth of shouting. Kurahashi Neilson, Sclafani and their team’s key advance was to find a way to filter out all that noise using the type of machine learning that’s now common in image-recognition software.
Analyzing a decade’s worth of IceCube data, they first set aside certain signals called tracks, which are long streaks that originate outside the detector. Tracks are useful because they have a clear direction and origin point, Kurahashi Neilson says, yet a lot of them are made by atmospheric neutrinos. To catch more neutrinos formed in space, she and her team focused on another type of signal called cascades, which look like a blob of light. Cascades are harder to find an origin point for, Kurahashi Neilson says, but they’re the signals that are more likely to be important. “We can actually see the southern sky better using cascades rather than tracks,” she adds.
Sclafani developed a deep neural net and trained it to recognize cascade events that arose deep within the detector—those that were most likely to be astrophysical neutrinos rather than atmospheric ones. By letting the neural net recognize these complex patterns of features, the researchers were able to glean 30 times the number of promising events from the dataset than previous methods. It would have taken an estimated 75 years to observe that number of events the old-fashioned way, Sclafani says.
“It’s a bit like putting a pair of glasses on,” says Kathrin Valerius, a physicist at the Karlsruhe Institute of Technology in Germany, who was not involved in the new study. “Everything seems much sharper with machine learning.”
The researchers were then able to compare the neutrino information to data on high-energy gamma rays in the Milky Way to see that they had the same origin, indicating that these neutrinos were the consequence of cosmic rays that originated in and around the Milky Way’s central disk.
“This is like a quantum leap to be able to say this has finally happened,” Valerius says. “People a few years ago cannot have imagined it would be done.”
Moving forward, researchers may be able to pick apart the neutrino data to answer long-standing questions, namely the fundamental origins of cosmic rays. They may come from supernovae remnants, active galactic nuclei or something else entirely—or, perhaps most probably, a mix of all these sources. So far, it’s impossible to tell whether any given neutrino forms at the source of a cosmic ray or during that cosmic ray’s travels through space, says Luigi Antonio Fusco, an astroparticle physicist at the University of Salerno in Italy. Future studies might be able to distinguish between these two scenarios, he says. “This emission is like a haze along the galactic plane,” Fusco says. “But inside that haze, we should really see individual point sources, individual emitters, and that would be groundbreaking again.”
Neutrinos could also be harnessed to delve into dark matter, the mysterious substance that is invisible on the electromagnetic spectrum but seems to gravitationally dominate galaxies and other large cosmic structures. It’s possible the neutrinos could form during collisions between dark matter particles, Valerius says. Detecting anomalous neutrino signals could lead to the indirect detection of dark matter.
“Every time you look at [something] in a new way, you pick out different things,” Kurahashi Neilson says, “and you’re able to construct a more full picture of what it is. This is a very powerful and completely new way to look.”
