Astronomers detect first confirmed minutes-long X-ray flash from a neutron-star merger

An international team of astronomers has identified a previously undetected phase of X-ray activity in the aftermath of a neutron-star collision, a finding researchers say exposes a hidden stage of these violent mergers that gamma-ray instruments alone have never been able to see. The event, cataloged as EP250704a in X-rays and GRB 250704B in gamma rays, produced a short, blinding flash of gamma rays lasting under half a second, then kept glowing in fainter, softer X-rays for nearly ten minutes afterward — far longer than theory predicted and long enough to be captured in detail for the first time.

The discovery, published this month in Science Bulletin, is based on observations from three satellites that happened to be watching the same patch of sky on July 4, 2025: China's Einstein Probe, the French-Chinese SVOM mission, and the Insight Hard X-ray Modulation Telescope. Researchers from the Hong Kong Institute for Astronomy and Astrophysics at the University of Hong Kong led the analysis together with Nanjing University, Beijing Normal University and the University of Rome Tor Vergata, drawing on a 78-member collaboration.

What the satellites recorded

The gamma-ray burst itself was unremarkable by the standards of these events: a spike of high-energy radiation lasting about 0.4 seconds, the signature already associated with the collision of two neutron stars or a neutron star and a black hole. What set EP250704a apart was what came next. Einstein Probe's wide-field X-ray telescope, designed specifically to catch faint, fleeting sources across a tenth of the sky at once, recorded a soft X-ray glow in the 0.5 to 4 kiloelectronvolt range that persisted for roughly 560 seconds — close to ten minutes — before fading into a more conventional afterglow.

• Gamma-ray burst duration: about 0.4 seconds
• Soft X-ray flash duration: about 560 seconds, the longest confirmed prompt X-ray phase from this class of event
• Distance: redshift 0.661, more than 6 billion light-years away
• Satellites involved: Einstein Probe, SVOM and Insight-HXMT

Follow-up observations with the European Southern Observatory's Very Large Telescope, using its X-Shooter and FORS2 instruments, found no trace of a supernova at the location — evidence against a massive star's collapse and in favor of a merger between two compact, dead stars. The team's paper describing the analysis argues that the variability and spectral shape of the X-ray emission do not match the standard picture of a brief, hard burst of accretion followed by a slowly fading afterglow shock. Instead, it points to a distinct, previously unconfirmed stage of the explosion itself.

Why a ten-minute afterglow matters

Short gamma-ray bursts, lasting less than about two seconds, have long been linked to neutron-star mergers, the same kind of collision that produces the gravitational waves detected by instruments such as LIGO. For years, roughly 30 percent of these bursts have shown signs of unusually long-lived, engine-driven X-ray emission in their afterglows, but no one had captured the prompt soft X-ray signal directly enough to confirm what was producing it. Earlier candidate detections were inconclusive. The HKU-led team argues that EP250704a changes that, because Einstein Probe's sensitivity to soft X-rays, a band largely invisible to gamma-ray detectors, caught the signal while it was still happening rather than inferring it after the fact.

The emission's behavior is consistent with a magnetar: a neutron star left behind by the merger that spins extremely fast and carries an immensely powerful magnetic field, and that continues dumping energy into its surroundings well after the initial collision. That would mean at least some neutron-star mergers do not immediately collapse into a black hole but instead leave behind a short-lived, highly active stellar remnant whose magnetic field keeps powering radiation for minutes rather than milliseconds.

"Einstein Probe is allowing us to uncover a part of compact star mergers that was hidden from previous gamma-ray observations," said Yi-Han Iris Yin, a PhD student in HKU's Department of Physics and a co-corresponding author of the study.

What researchers say

Eleonora Troja, an astrophysicist at the University of Rome Tor Vergata and one of the paper's corresponding authors, described the physical picture behind the signal in terms of the remnant's magnetic behavior. "Magnetars are rapidly spinning neutron stars with huge magnetic fields," she said. "When they damp their magnetic power into the surroundings, they can make any explosion brighter and longer-lasting." Co-author Niccolò Passaleva, who carried out the Very Large Telescope follow-up that ruled out a supernova origin, said the absence of any stellar explosion at the site strengthened the case for a compact-object merger rather than a collapsing massive star.

Bing Zhang, founding director of the Hong Kong Institute for Astronomy and Astrophysics and a co-author on the paper, framed the result as part of a broader push to put early-career scientists at the center of major discoveries. "One of our goals at HKIAA is to create an environment where talented young researchers can take scientific leadership and work at the forefront of international astronomy," he said. The event was also logged in real time through the astronomical community's rapid-alert system, with a NASA-hosted circular notifying observers worldwide within hours of the detection so that telescopes could be pointed at the location before the afterglow faded.

What happens next

The research team says the detection of a prompt soft X-ray counterpart in what otherwise looked like an ordinary short gamma-ray burst suggests this kind of extended emission is common in merger-driven explosions, not a rare exception — meaning many similar signals may have been missed by instruments that were not sensitive enough, or not looking in the right energy band, to catch them. That has direct consequences for the search for electromagnetic counterparts to gravitational-wave events, since a reliable X-ray signature lasting minutes rather than milliseconds gives observatories a far easier target to find and follow up quickly after a gravitational-wave alert.

Einstein Probe, launched in January 2024 as a collaboration between the Chinese Academy of Sciences, the European Space Agency and the Max Planck Institute for Extraterrestrial Physics, is expected to keep monitoring roughly a tenth of the sky at a time for faint, fast-changing X-ray sources through at least 2027. Researchers involved in the EP250704a study say the next step is a systematic search through the mission's existing archive for similar long, soft X-ray flashes that may have gone unrecognized in other short gamma-ray bursts, alongside coordinated efforts to catch the next such event with both gravitational-wave detectors and X-ray telescopes pointed at the same patch of sky at the same time.