An international team found an isolated pulsar in 15 years of observations from NASA’s Fermi Gamma-ray Space Telescope using the equivalent of just 10 hours on a standard processor. The star, PSR J1622−7204, spins 261 times a second. By treating the search as a geometry problem, the researchers developed a way to find pulsars about a million times faster than before. Their results appeared Thursday in Physical Review Letters.

Finding a rhythm when most rotations go unseen

A pulsar is a spinning neutron star whose radiation beam sweeps past Earth. Fermi catches few photons from any one pulsar, and those arrive mixed with gamma rays from other sources. To pick out the rhythm, researchers must work out where the star is, how fast it spins and how quickly its spin slows, all at once. Lead author Daniel Pook-Kolb of Germany’s Max Planck Institute for Gravitational Physics said an exhaustive search would take millions of trials in each direction of a four-dimensional grid, beyond the computing power available on Earth.

Earlier searches split the observations into shorter stretches and enlisted volunteers around the world to analyze them on their computers. Even with that help, a single discovery typically took hundreds of computer-years of processing.

Dotan Gazith and Barak Zackay of Israel’s Weizmann Institute of Science recast the search as a problem with a mathematical solution: finding the shortest vector in a lattice, a regular array of points in many dimensions. Cryptographers had developed algorithms for that task to test encryption designed to withstand future quantum computers. As Gazith explained, the astronomers borrowed the mathematics, not to break a code but to find a signal. The lattice technique has also helped make satellite navigation more efficient.

One trial took about five minutes on a single processor core, well within the capacity of an ordinary laptop. The lattice search identified a candidate. When the team checked it against the full set of Fermi observations used in the study, the signal was strong enough to rule out chance.

An open archive and the search for pulsars in pairs

Since its launch in 2008, Fermi has surveyed the whole sky every three hours in gamma rays, the most energetic form of light. After a year of calibration, NASA opened the mission’s data to everyone and has continued to release observations as soon as they are processed. That policy let scientists outside the mission try a new approach, and NASA’s open software, including Fermitools, helped the team prepare the observations for analysis.

The discovery used only part of the archive: study co-author Colin Clark of the Max Planck Institute said all 18 years of Fermi observations are public. The mission’s catalog has grown from a few hundred gamma-ray sources to thousands. Elizabeth Hays, Fermi project scientist at NASA’s Goddard Space Flight Center, noted that independent teams can repeat a search and check its results.

The team’s next target is pulsars that orbit companion stars. A companion’s gravity distorts the pulsar signal, making these systems all but impossible to find with conventional searches. Clark estimates that tens or perhaps hundreds of such pulsars may be hidden in the data. He hopes to pair the lattice algorithm with computing power donated by Einstein@Home volunteers to find them.