Out there in the Universe, a similar cosmic story plays out time and time again: the stellar life cycles. A massive, cold cloud of gaseous material, under the power of its own self-gravitation, begins to contract. As clumps within that gas get denser, the cloud fragments, with each fragment collapsing on its way to become a protostar. The most massive and dense clumps, initially, attract surrounding matter the fastest, growing into the most massive of these new stars, where around 50% of all such stars wind up developing in multi-star (binary, trinary, or higher multiplicity) systems. When a star forms with more than about 8-10 solar masses, initially, it’s destined to die in a core-collapse supernova explosion: one of the most energetic cataclysms that occurs within our Universe. And yet, despite the fact that: binary systems are common, stars of similar masses are often found together (especially at high masses), and that the highest-mass stars are the shortest-lived ones before dying, we’ve never witnessed or traced back a “sibling supernova” event: where two core-collapse supernovae occurred from the same original system. In a remarkable new study led by Stanford scientist Miltiadis Michailidis, two different supernovae have just been linked — through their gamma-rays — across the same great molecular cloud complex. We may have just witnessed the first pair of sibling stars, separated when the first went supernova, reunited when the second went supernova just a little bit later. Here’s the story. The dense cores of protostar cluster G333.23–0.06, as identified by ALMA, show strong evidence for large levels of multiplicity within these cores. Binary cores are common, and groups of multiple binaries, forming quaternary systems, are also quite common. Triplet and quintuplet systems are also found inside, while, for these high-mass clumps, singlet stars turn out to be quite rare. It is expected that the stars forming in nebulae all throughout the Universe, including in the Eagle Nebula, have similar clumpy, fragmented properties. Credit: S. Li et al., Nature Astronomy, 2024 What you see, above, is an actual image of many protostellar cores forming within the same obscured star-forming region. This is one of the highest-resolution images of protostars still being formed within a dusty, obscured region of a star-forming nebula: only revealed because of the incredible power of the ALMA observatory. Later on in life, we can see the newborn star clusters that emerge, as well as the singlet, binary, trinary, and higher multiplicity systems that eventually emerge from those dense stellar nurseries. Of the stars that form, it’s the highest-mass ones that burn through their fuel the fastest, and that die in a core-collapse supernova after the smallest amount of time elapses. For a high-mass binary that forms, that means the more massive star dies first, with its supernova blast wave expected to “kick” its companion across interstellar space. However, if the less massive star is close to the more massive one in mass and lifetime, only a short amount of time might elapse before it goes supernova, too. If the remnant from the first supernova hasn’t faded away, these two supernova remnants — from stars that were initially siblings — could overlap, leading to a magnificent reunion. With over 15 years of Fermi/LAT data now available to cover the gamma-ray sky, high-resolution and sensitive maps can be uncovered. Over on the far right of the image, located almost 180 degrees opposite to the location of the galactic center, the region containing IC 443 is located: blown up in the inset, and the subject of a recent study that reveals two overlapping supernova remnants in the same patch of sky. Credit: NASA/DOE/Fermi LAT Collaboration The way to look for such an event isn’t visually, with the type of light that our eyes can see, but rather at the highest energies of all: through gamma-rays. The Fermi Gamma-Ray Space Telescope, launched by NASA in 2008, is our greatest, most sensitive, highest-resolution gamma-ray telescope to ever open its eyes on the Universe, giving us an unprecedented view of our galaxy (and beyond) at the highest energies of all. Pulsars, black holes, and supernova remnants stand out at these energies, with one specific region pointed almost directly away from the galactic center containing several notable gamma-ray sources within it: IC 443, a bright supernova remnant also known as the Jellyfish Nebula, the Crab nebula, a pulsar inside of a supernova remnant, and Geminga, one of the brightest and closest pulsars of all. If you take a close look at IC 443 and the region around it in a variety of wavelengths, however, you’ll find that it isn’t just an isolated supernova remnant out there in deep space, but rather is interacting with its environment. There’s a nearby star-forming region filled with ionized hydrogen, Sharpless 249, that the supernova remnant interacts with, as well as another nearby supernova shell: G189.6+3.3. This view of the Jellyfish Nebula, also known as IC 443, is a prominent and recent supernova remnant in the direction opposite to the galactic center. Nearby the Crab Nebula and the Geminga pulsar in space, it turns out to be a part of a much larger complex, whose features are revealed by multiwavelength studies of this region. Credit: Siderevs nuncivs/Wikimedia Commons With not just Fermi data, but also including data from Planck and MWISP in the microwave/radio, from the Digitized Sky Survey in the optical, from WISE in the infrared, from Swift in the ultraviolet, and from eROSITA in the X-ray, we can see the vast scale at which IC 443 interacts with its surroundings. The greater the number of different data sets we fold in, the greater the amount of detail emerges not only within this region of space, but the greater the details we can reconstruct about the interactions between different sources of energy and different regions within this volume of our Milky Way. Because there are multiple supernova remnants within this one region of space — remnants that may be separated by many light-years in space but only by a few tens of thousands of years in terms of time-of-detonation — there’s a remarkable scenario that’s worth considering. Could it be possible that: there was, initially, a binary system of high-mass stars, where the first, more massive star exploded, sending the companion journeying rapidly away from it through space, where, tens of thousands of years later, the second star exploded, and finally, the remnants expanded and interacted, reuniting these companions in the aftermath of their demise? A multiwavelength view of this region is certainly suggestive of precisely that scenario. As more and more data gets added to the initial optical data set (yellow), additional features start to emerge. In reds and browns, long wavelength radio and infrared data is added in, In teal, higher-energy light is added in: localized atop IC 443. And then, in an extended fashion, additional high-energy photons are seen, particularly in the gamma-ray part of the spectrum. It’s the analysis from this data that first revealed the hidden presence of a second, earlier supernova event. Credit: NASA Goddard Space Flight Center and M. Michailidis et al. 2026; orange, brown: radio, ESA/Planck and MWISP; yellow: optical, DSS; red: infrared, NASA/ WISE; violet: ultraviolet, NASA/Swift; teal: X-rays, SRG/eROSITA It isn’t generally appreciated why this is the case, or why gamma-rays are such an important puzzle piece in putting this picture together. Even with the incredible energies that go into a core-collapse supernova, including nuclear processes like electron capture, neutrino production, and a high-energy blast wave, none of those events produce radiation at the highest energies: gamma-rays. Instead, there are only two mechanisms that can get you there: high-energy proton collisions with interstellar gas, or high-energy electrons that collide with low energy photons and boost them to gamma-ray energies. In addition, gamma-rays come in both low-energy and high-energy varieties, and those different energies encode different pieces of information. Low-energy gamma-rays are thought to emerge primarily from one bright gamma-ray beacon in the sky: IC 443, in this case. This is because, at those low energies, the entire region looks blended together, with IC 443 being the location where the emission signals appear strongest. However, at higher gamma-ray energies, you can see separate, distinct structures: structures that become even more apparent when you subtract the gamma-ray light from IC443 out. This animation shows low-energy gamma-rays (left) and high-energy gamma-rays (right) in this complex, with the emissions from the IC 443 supernova remnant subtracted out where the main pink light disappears. What remains provides evidence for a second, older supernova remnant located to the left of the first, coincident with the bright gamma-ray emissions that still persist. Credit: NASA Goddard Space Flight Center and M. Michailidis et al. 2026; orange, brown: radio, ESA/Planck and MWISP; yellow: optical, DSS; red: infrared, NASA/WISE; violet: ultraviolet, NASA/Swift; teal: X-rays, SRG/eROSITA; magenta: gamma rays, NASA/DOE/Fermi LAT Collaboration Over on the left of the image above, you can see the low-energy gamma-ray signature, while the right shows the high-energy gamma-ray signature. Where the bright pink glow appears around the bubble (IC 443), that shows the cumulative gamma-ray emission, including IC 443. However, when the pink glow disappears, that showcases the signal from IC 443 subtracted out: leading to a wealth of structure that’s apparent, particularly at the highest gamma-ray energies. It isn’t just one instrument that sees this, either, as both NASA’s Fermi and ESA’s eROSITA detect these important signals. When they take a closer look, they see that IC 443 isn’t the only structure that contributes to the signal. Instead — and very importantly — they indicate the presence of a second, independent supernova remnant: G189.6+3.3, a source that, up until recently, was hidden in the shadow of the much brighter remnant in its vicinity: IC 443. eROSITA detects a shell from this supernova remnant, while the shape traced out by the high-energy gamma-ray emissions indicates an origin from within the remnant. In particular, a very striking feature that emerges is a bright, ionized wave at the edge of the bubble: consistent with a shock front. As much as there can be one, this is what we can consider “smoking gun” evidence for a second, recent supernova. This enhanced view of the filament located within the IC 443 complex shows the main supernova remnant (at right), the interacting neutral/ionized gas cloud (at center), and the second, older supernova remnant located to the left of the first. The shocked gas along the filament, in purple, appears in visible, ultraviolet, X-ray, and gamma-ray light. Credit: NASA Goddard Space Flight Center and M. Michailidis et al. 2026; orange, brown: radio, ESA/Planck and MWISP; yellow: optical, DSS; red: infrared, NASA/ WISE; violet: ultraviolet, NASA/Swift; teal: X-rays, SRG/eROSITA The gamma-ray emission that you see above — prominently in purple — clusters at the northern boundary of the second supernova remnant: exactly where it overlaps with the dense cloud of interstellar material. There’s a filament that emerges, and that filament is also apparent in ultraviolet light and in visible light, particularly in its hydrogen-alpha emissions. The fact that this filamentary feature appears in a slew of energetic forms of light: in gamma-rays and X-rays, in the ultraviolet, and from the de-excitation of excited atoms in visible light, all points in one direction: this is a radiative shock. The ejecta from the second supernova departs from the core-collapse event extremely rapidly, with the fastest-moving ejecta leading the way and colliding with material in the interstellar medium first. When it collides, it slows down: producing gamma-ray particles (because they’re fast-moving initially) at the start, but then ceasing to produce gamma-ray particles after the slowdown. So if we’re seeing gamma-rays coming from this region after all, that teaches us something: there must be a re-acceleration of material — perhaps from previously accelerated protons — that’s producing this bright signature gamma-rays. This illustration shows the shock front of the long, energized filament located at the northern boundary of the older core-collapse supernova. The highest gamma-ray intensity comes from the edge of the shock front, and overlaps with where the shock front and the Sharpless 249 molecular cloud (shown in dashed blue) overlap. Credit: Miltiadis Michailidis, Stanford This was validated in the new paper by Michailidis et al., which explored a crash cloud model for this scenario. All throughout the remnant, you have cosmic ray electrons that collide with low energy photons — from the CMB, from infrared (thermal) radiation, and from starlight directly — which can boost photons up to high, gamma-ray energies. This produces small numbers of gamma-rays at low energies, but as you look to greater and greater energies, the gamma-ray flux increases. This leads to a relatively even flux of gamma-rays spatially, but one that’s energy-dependent: with few low-energy gamma-rays but more high-energy gamma-rays. However, at the northern boundary, where you have that filamentary feature resulting from a shock wave colliding with a population of dense molecular gas, the interaction of cosmic ray protons with that gas produces gamma-rays: but a big excess of gamma-rays at low and medium energies over the other (inverse Compton scattering) process. The clouds and the filament provide abundant targets for proton-proton interactions, which leads to an increased overall flux of gamma-rays. That explains why, when we look in gamma-ray energies, we see a much greater overall count of gamma-ray photons where the filament and dense molecular cloud overlap. The Fermi-LAT gamma-ray counts (center) are shown for the newfound (and earlier) supernova remnant, G189.6+3.3, which has substantial gamma-ray signatures throughout but a massive enhancement at the northern boundary: where a dense cloud and the observed filament are seen. While the entire nebula is energized by the process (inverse Compton scattering) drawn out at right, only the northern boundary is affected by proton-proton enhancement, shown at left, enhancing the total number of gamma-ray photons significantly. Credit: Miltiadis Michailidis, Stanford One worry you might have is this: how confident are we, even if we see evidence for two supernova remnants, that these remnants actually overlap in three-dimensional space, rather than just coexist along the same line-of-sight? After all, the pulsar Geminga, close by in the sky, is only around 800 light-years away, whereas IC 443 and the other features within that gas cloud are more like 5000-6000 light-years distant. The way we can be confident of their nearby nature is that even if the two remnants aren’t interacting with and crashing into one another, the new supernova remnant is definitely interacting with the same molecular cloud that IC 433 interacts with, and therefore, they’re at comparable distances. You might also worry that, just possibly, these are two unrelated supernova remnants, rather than being remnants whose progenitor stars have a common, shared history with one another. This is not something we can answer definitively, unfortunately, but what we can do (and what Michailidis and his team did in the paper) is to run simulations that allow us to estimate the probabilities that this event arose from one of two different scenarios. Can two unrelated supernovae occur, leading to two remnants that appear this close to one another on the sky: whether co-located in 3D space or just along the same line of sight? Can a binary pair of massive stars form of sufficient mass that one of them goes supernova, kicks the other one, and then, just tens of thousands of years later, the second one goes supernova as well? As it turns out, both scenarios are plausible, but with vastly different probabilities and likelihoods. The first scenario, in all of its permutations, has a likelihood between 0.1% up to 1.2%: not negligible, but not particularly likely. The second scenario, on the other hand, is extremely well-motivated on physical grounds, and could in particular result from a star of between 30 and 40 solar masses exploding, creating the newfound (and older) remnant, and then, between 20,000 and 100,000 years later, a star that was initially about 5 solar masses lighter (between 25 and 35 solar masses) would go supernova, creating IC 443. That second scenario, for the two supernova remnants to have been born in a binary, has a probability of approximately 99%. This four-panel illustration begins from a massive binary system, with two high-mass but comparably massed stars in orbit around one another. When the first star dies in a core-collapse event, the lower-mass companion gets kicked away. After tens of thousands of years, the companion evolves and explodes, and the two supernova remnants then overlap: definitely along the same line-of-sight, and possibly in three-dimensional space as well. Credit: Miltiadis Michailidis, Stanford What isn’t supported by the data — and this is interesting to consider — is a scenario where the older remnant exploded first, and as the second star hurtles through space and then explodes, that the second explosion’s shockwave catches up with the first. However, that would lead to a different morphology, or shape for the remnants, than what we observe. The bi-directional shape of the ejecta instead indicates that the shockwave from IC 443 has not yet caught up with the earlier supernova: it doesn’t show up in gamma-rays, X-rays, or the ultraviolet. There’s a questionable (albeit, probably spurious) signature that could point to those interactions, but it has yet to be confirmed. On the other hand, two related supernova remnants, that overlap and that originate from the same set of initially bound stars, has never been found before. This system, therefore, represents the first candidate scenario for stars that were: born together, were separated by the death of the first one, then the second one died, and now, tens of thousands of years later, are reunited by the ejecta of their remnants. We have never observed two supernova remnants that originated from the progenitors within a binary system before. As our views of the Universe improve — in resolution, in sensitivity, in observing time, and in wavelength coverage — we can be sure that more and more rare events will appear. In time, we may either confirm or refute this picture, but as long as we’re willing to look, we’ll continue to come closer and closer to piecing together how the Universe actually behaves: when it comes to common, rare, and even unique events. This article A cosmic first: separated sibling stars reunited by supernovae is featured on Big Think.