In a cosmic event that unfolded over nearly a decade, astronomers have witnessed something long predicted but rarely observed: a massive star collapsing directly into a black hole without the spectacular fireworks of a supernova explosion. The discovery, published in the journal Science, represents the most complete observational record ever made of a star’s transformation into a black hole and challenges assumptions about how these enigmatic objects form.
The star, designated M31-2014-DS1, was located approximately 2.5 million light-years away in the Andromeda Galaxy. When it was newly formed, the star weighed around 13 times the mass of our sun. By the time of its death, it had shed most of its mass through powerful stellar winds, leaving behind roughly five solar masses that would ultimately collapse into a black hole.
“This has probably been the most surprising discovery of my life,” said Kishalay De, a Columbia University astronomy professor who led the research team. “The evidence of the disappearance of the star was lying in public archival data and nobody noticed for years until we picked it out.”
The discovery came from an exhaustive analysis of archival data from NASA’s NEOWISE mission, combined with observations from ground-based and space-based telescopes spanning 2005 to 2023. In 2014, the star’s infrared light began brightening, glowing more intensely for approximately three years. Then, starting in 2017, M31-2014-DS1 began to fade dramatically. By 2019, it had dimmed by a factor of more than 100 in optical light. When astronomers looked for the star in 2022 and 2023, it had essentially vanished in visible and near-infrared wavelengths, becoming one ten-thousandth as bright as before.
“This star used to be one of the most luminous stars in the Andromeda Galaxy, and now it was nowhere to be seen,” De explained. “Imagine if the star Betelgeuse suddenly disappeared. Everybody would lose their minds! The same kind of thing [was] happening with this star in the Andromeda Galaxy.”
The dramatic fading provides strong evidence for what astronomers call a “failed supernova”—a process where a star’s core collapses but fails to produce the explosive ejection of outer layers characteristic of a typical supernova. Instead, most of the stellar material falls back onto the collapsing core, forming a black hole.
For decades, theoretical models have predicted that some massive stars should undergo this quieter death. When a massive star—roughly 10 or more times heavier than our sun—runs out of nuclear fuel, the delicate balance between gravity pulling inward and pressure pushing outward is disrupted. The core collapses first, forming a dense neutron star. Typically, the emission of neutrinos during this collapse generates a powerful shock wave explosive enough to tear the star apart in a supernova. But if this neutrino-powered shock fails to eject the stellar envelope, theory suggests most of the material should fall back, creating a stellar-mass black hole.
“We’ve known for almost 50 years now that black holes exist,” De noted, “yet we are barely scratching the surface of understanding which stars turn into black holes and how they do it.”
What makes this discovery particularly significant is not just the observation of the disappearance itself, but the comprehensive data that allowed researchers to construct a detailed physical picture of the process. The key breakthrough involved understanding the role of convection in the star’s outer layers.
When the star’s core collapsed, the gas in its outer layers was still moving rapidly due to convection—the churning motion caused by vast temperature differences between the star’s hot interior and cooler exterior regions. Andrea Antoni, a Flatiron Research Fellow who developed theoretical predictions for these convection models, explained the critical insight: “the accretion rate—the rate of material falling in—is much slower than if the star imploded directly in. This convective material has angular momentum, so it circularizes around the black hole. Instead of taking months or a year to fall in, it’s taking decades.”
Similar to water swirling around a bathtub drain rather than flowing straight down, the convective motion prevented the entire star from collapsing directly into the newly formed black hole. Instead, the researchers estimate that only about one percent of the original stellar envelope gas actually falls into the black hole. The remaining material was partially ejected outward, where it cooled and formed dust. This dusty debris shell explains the infrared brightening observed in 2014 and the lingering mid-infrared glow that remains visible today.
The discovery also allowed researchers to reinterpret observations of a similar star, NGC 6946-BH1, which had been categorized as a candidate failed supernova about a decade ago but whose exact nature remained unclear and debated. NGC 6946-BH1, located in the galaxy NGC 6946 approximately 10 times farther away than M31-2014-DS1, exhibited similar behavior but was 100 times fainter with lower-quality observational data.
“It’s only with these individual jewels of discovery that we start putting together a picture like this,” De said.
Morgan MacLeod, a lecturer on astronomy at Harvard University and co-author on the paper, emphasized the broader significance: “We’ve known that black holes must come from stars. With these two new events, we’re getting to watch it happen, and are learning a huge amount about how that process works along the way.”
The findings have important implications for understanding the population of stellar deaths in the universe. Previous research suggested that between 10 and 30 percent of massive stars might undergo failed supernovae, but observational evidence has been scarce.
“Unlike finding supernovae which is easy because the supernova outshines its entire galaxy for a few weeks, finding individual stars that disappear without producing an explosion is remarkably difficult,” De explained. “It comes as a shock to know that a massive star basically disappeared (and died) without an explosion and nobody noticed it for more than five years. It really impacts our understanding of the inventory of massive stellar deaths in the universe. It says that these things may be quietly happening out there and easily going unnoticed.”
The discovery also raises intriguing questions about which stars become black holes and why. The stellar evolution of M31-2014-DS1 was similar to that of many core collapse supernova progenitors, suggesting what De described as a “complex (possibly chaotic) relationship between stellar birth mass and black hole formation” for stars with initial masses greater than about 12 solar masses.
The story of M31-2014-DS1 is far from over. The dusty debris surrounding the newborn black hole will continue to emit infrared light for decades to come, providing astronomers with an unprecedented opportunity for long-term study.
“This is just the beginning of the story,” De said. Light from the dusty debris “is going to be visible for decades at the sensitivity level of telescopes like the James Webb Space Telescope, because it’s going to continue to fade very slowly. And this may end up being a benchmark for understanding how stellar black holes form in the universe.”
Endnotes:
- De, K., MacLeod, M., Jencson, J.E., et al. (2026). “Disappearance of a massive star in the Andromeda Galaxy due to formation of a black hole.” Science, 391(6786). DOI: 10.1126/science.adt4853
- Kochanek, C.S., et al. (2008). “On the Rate of Black Hole Binary Mergers in Milky Way-like Galaxies.” The Astrophysical Journal, 684, 1336-1342.
- Adams, S.M., Kochanek, C.S., Gerke, J.R., Stanek, K.Z., & Dai, X. (2017). “The search for failed supernovae with the Large Binocular Telescope: confirmation of a disappearing star.” Monthly Notices of the Royal Astronomical Society, 468, 4968-4981.
- Smartt, S.J. (2009). “Progenitors of Core-Collapse Supernovae.” Annual Review of Astronomy and Astrophysics, 47, 63-106.
- Neustadt, J.M.M., et al. (2021). “The search for failed supernovae with the Large Binocular Telescope: constraints from 7 yr of data.” Monthly Notices of the Royal Astronomical Society, 508, 516-528.
IMAGE CREDIT: Keith Miller, Caltech/IPAC – SELab