The vacuum was supposed to be empty — a magnetar’s X-rays just proved it isn’t

The space between a magnetar and Earth is not empty. According to quantum electrodynamics, the fabric of the vacuum is threaded with virtual particles that wink in and out of existence, and in the presence of an extraordinary magnetic field, they exert a measurable pressure on passing light. Werner Heisenberg first worked out the mathematics of this effect in 1936. Eight decades later, no telescope had found direct evidence it was real.

NASA’s IXPE has now changed that. By pointing its X-ray polarimetry instruments at magnetar 1E 1547.0-5408 for more than 140 hours, a team led by Rachael Stewart at George Washington University and Hoa Dinh Thi at Rice University found polarization values that nothing in standard astrophysics can explain — unless the vacuum itself is doing something to the light.

What IXPE measured — and why the number matters

IXPE does not just count X-ray photons; it measures which way they are oscillating. That orientation is polarization. Most astrophysical X-ray sources produce modest polarization, because the emitting plasma is turbulent and the photons leave in mixed directions. Magnetars are different: their fields are so ordered that even without vacuum effects, the emitted X-rays should carry a detectable polarization signature.

Recommended ReadingA planet nursery was caught spinning for the first time, and parts ignore the rules

The team measured 40% polarization from the upper emission cone and 80% from the lower — values almost three times higher than what any comparable magnetar has produced in previous IXPE observations. The researchers also combined IXPE data with simultaneous observations from NASA’s NICER telescope aboard the International Space Station and CSIRO’s Parkes radio telescope in Australia. That coordinated triple-instrument approach was the first of its kind for any magnetar and gave the team an unusually complete picture of where the radio and X-ray emissions originate and how they relate to each other.

The geometry matters here. Magnetar 1E 1547.0-5408 happens to have its magnetic and rotational axes nearly aligned, and observers on Earth happen to be looking almost straight down the magnetic pole. That alignment amplifies the vacuum signal: instead of seeing the light at a glancing angle, IXPE is looking along the axis where the field is strongest and where vacuum birefringence should be most pronounced.

The 90-year-old prediction

Vacuum birefringence was worked out in 1936 by Heisenberg and Hans Heinrich Euler as a consequence of quantum electrodynamics. The theory holds that the vacuum is not a passive backdrop to events but a dynamic medium filled with virtual particles. In regions of ordinary magnetic field strength — including the strongest fields ever produced in a laboratory — these virtual particles have no detectable effect on light. But around a magnetar, whose field reaches approximately one trillion times the strength of Earth’s, they become impossible to ignore.

The mechanism is filtering. Light oscillating parallel to the magnetic field travels through the virtual-particle medium at a slightly different speed than light oscillating perpendicular to it. The result is that one polarization direction is preferentially preserved and the other is suppressed, so the light that escapes is far more polarized than what was emitted. Astronomers call this birefringence, the same word used for double-refracting crystals — but the medium here is nominally empty space.

Previous attempts to detect this effect have run into two problems. Other magnetars either have magnetic axes oriented at angles that dilute the signal, or their emission geometry is poorly constrained. Magnetar 1E 1547.0-5408 sidesteps both problems.

What this magnetar has that others don’t

The coordinated radio and X-ray observations gave the team information about the magnetar’s geometry that would have been impossible from X-rays alone. Radio emission from magnetars traces the magnetic field lines precisely; CSIRO’s Parkes telescope tracked how the radio polarization rotated across the 2.1-second rotation period, pinning down the orientation of both the magnetic and rotational axes. That geometric model then fed into the X-ray analysis, letting Stewart and colleagues disentangle the baseline emission polarization from the vacuum-birefringence enhancement on top of it.

The result matches theoretical predictions both in magnitude and in angular dependence across the rotation period. It is not just the polarization level that fits the vacuum birefringence signature — it is the way that level changes as the magnetar rotates.

Why the finding isn’t settled yet

Science is cautious about extraordinary claims, and this one qualifies. A companion analysis published in April 2026, by an independent team working with the same IXPE data but a different geometric model for the emission region, concluded that the data cannot yet be regarded as compelling evidence for vacuum birefringence. The disagreement hinges on assumptions about where exactly the X-rays originate and how the emission geometry relates to the magnetic field. Stewart and colleagues used the simultaneous radio data to constrain that geometry; the April team did not have coordinated radio measurements.

The discrepancy is not a sign that one result is wrong. It is a sign that the measurement is at the edge of what current observations can resolve, and that the physical interpretation depends critically on how well the geometry is known. Additional IXPE observations of 1E 1547.0-5408, along with improved simulations of magnetar emission physics, are needed before the community will treat vacuum birefringence as settled.

Common questions about vacuum birefringence

What does it mean that empty space isn’t empty? Quantum field theory predicts that the vacuum contains a constant churning sea of virtual particle pairs that appear and annihilate on timescales too short to detect directly. In ordinary conditions, their effect on light is undetectable. In the magnetic field of a magnetar, their effect on polarization is large enough for IXPE to measure.

Why use a magnetar and not a laboratory magnet? Laboratory magnets cap out at around 45 tesla — powerful, but many orders of magnitude below the trillion-tesla fields of a magnetar. The vacuum birefringence effect scales with field strength squared, which means going from 45 tesla to a magnetar’s field makes the signal roughly a septillion times larger.

What is IXPE and why is it the right instrument? NASA’s Imaging X-ray Polarimetry Explorer launched in 2021 specifically to measure the polarization direction of X-ray light. Earlier X-ray telescopes could measure brightness and spectrum but not polarization, so the vacuum birefringence signal was always hidden inside data that couldn’t resolve it.

Does this change how we think about empty space? If confirmed by further observations, it would be the first direct demonstration that the quantum vacuum has physical effects on light in free space. It would also provide a new experimental test for quantum electrodynamics in an energy regime that no particle accelerator can reach.

The IXPE team plans further observations of 1E 1547.0-5408 and other well-positioned magnetars over the next observing cycle. The critical next step is extended simultaneous radio and X-ray coverage, giving the geometric models better constraints. If the polarization excess holds up across additional observations and the geometric debate is resolved, it will mark the first experimentally confirmed prediction from the vacuum sector of QED that goes beyond what laboratory physics can test.

Published in August 2026.

Reference: Stewart et al., “Vacuum birefringence and the polarized X-ray emission from a radio magnetar,” Nature, 2026. DOI: 10.1038/s41586-026-10859-z

Tags: astronomy, physics, IXPE, magnetar, quantum electrodynamics, vacuum birefringence

More Like This

A speck of metal with 10,000 atoms was held in two places at once

Huge Rings Around a Black Hole

Jupiter flings electrons to near light speed the way supernovae make cosmic rays

Hubble caught the largest planet nursery — 40 times wider than our solar system

Quasar Tsunamis Rip Across Galaxies

The Universe’s Ghostly Hard Drives: Why the Largest Black Holes Are Made of Air