The first neutrino image of our galaxy

IceCube at the South pole with the Milky Way in the background. Photo: Martin Wolf, IceCube/NSF

For ages, the view of our Milky Way galaxy has inspired awe, visible with the naked eye as a hazy band of stars that stretches across the sky. Now IceCube researchers are able to see the Milky Way using neutrinos – tiny, ghostlike particles that fly freely through matter and space.

Neutrino telescope at the South Pole

The neutrinos were detected using the IceCube Neutrino Observatory, a neutrino telescope built at the South Pole and monitoring 1 billion tons of the Antarctic ice for the rare neutrino interactions.  Members of the IceCube Collaboration at Stockholm University and Uppsala University have been involved in the analysis of the data acquired during ten years of observations and in the interpretation of the results.

Chad Finley.
Photo: Sören Andersson

“Seeing our galaxy with neutrinos is something that we dreamed of, but which seemed out of reach for our project for many years to come,” says Chad Finley, associate professor at the Department of Physics, Stockholm University and one of the IceCube team members who worked closely on the paper. “What made this result possible today is the revolution in Machine Learning, allowing us to explore much deeper into our data than before.”

Interaction of cosmic rays, high-energy protons and nuclei

During the last century, astronomers began to study the Milky Way in all wavelengths of light, from radio waves to gamma rays. High-energy gamma rays in our galaxy are believed to be created mainly by the interaction of cosmic rays, high-energy protons and nuclei, with galactic gas and dust. This process should produce neutrinos as well. However, large uncertainties about the intensity of cosmic ray interactions in different parts of our galaxy made predictions for neutrinos a challenge.  

The analysis method used for the latest results was originally developed in 2017 at Stockholm University by Jonathan Dumm, a postdoctoral researcher at the Oskar Klein Centre. “Jon realized that if the upper range of these neutrino predictions were correct, they might have been faintly detectable in the IceCube data we had at the time,” says Chad Finley.

New computing techniques

Milky Way depicted with visible light and with neutrinos. Photo: IceCube/NSF

To really map out the neutrino contours of the Milky Way, however, it was expected that many more years of data-taking would be needed. While the IceCube Observatory records billions of events each year, only a very tiny fraction (one in one-hundred million recorded events) is due to neutrinos from space. Identifying these neutrino events is a difficult computational task. Thanks to the development of new computing techniques called Deep Neural Networks, it has become possible to identify these neutrino events with 20 times higher efficiency than before.

"Previous discoveries with IceCube have involved neutrino emission from much greater distances and it is really exciting that we can now see a neutrino image of our own galaxy," says Klas Hultqvist, Professor at  the Department of Physics, Stockholm University.

"This opens the door to studying further processes using neutrinos. The first step is to try to identify objects in the Milky Way as neutrino sources. IceCube-Gen2, an upgrade of IceCube that we are now planning, would give us much greater sensitivity for further studies of events in the galaxy."

The article Observation of High-Energy Neutrinos from the Galactic Plane is published in the journal Science.