Ghost Particles May Decide Whether Dying Stars Explode — or Vanish Into Black Holes

Subtitle: Copenhagen simulations of 195 collapsing stars show that neutrino flavor conversion can tip massive stars toward failed supernovas — and help explain why the universe seems short on explosions.

When a massive star dies, almost everything that matters is invisible. Nuclear fuel runs out. The iron core implodes. Protons and electrons crush into neutrons. And in a fraction of a second, the star dumps about 99% of the released gravitational energy not into light, shock waves, or fireworks — but into a flood of neutrinos, the ghost particles that barely notice ordinary matter.

That flood is not a side show. It is the engine. Neutrino heating behind a stalled shock can revive the blast and leave a neutron star. Fail to heat enough, and the star collapses into a black hole with little or no optical bang.

On September 20, 2026, Space.com amplified a Physical Review D study that puts a sharper edge on that story: the way neutrinos change flavor — electron, muon, or tau — may decide which fate wins. Authors Mariam Gogilashvili and Irene Tamborra of the University of Copenhagen’s Niels Bohr Institute simulated the collapse of 195 stellar progenitors from 9 to 120 solar masses, folding in a schematic treatment of neutrino flavor conversion. Their conclusion is blunt. Flavor change can make dying massive stars significantly less likely to explode and more likely to form black holes — especially in the awkward middleweight band of 16–30 solar masses, where standard runs often succeed and nature often seems to vanish.

“Neutrinos are not a side detail in supernovas,” Gogilashvili told Space.com. “They carry away about 99% of the energy released when the core collapses, and a small change in how they behave can decide the fate of the whole star.”

Three flavors, one deadly hinge

Neutrinos come in three flavors, named for the charged partners they couple to: electron, muon, and tau. Since 1998 — work that later earned a Nobel Prize — physicists have known these particles oscillate, swapping identities as they travel. Electron-flavor neutrinos and antineutrinos are the ones that deposit energy through charged-current interactions behind the shock. Muon and tau flavors interact far more weakly with ordinary matter. Swap flavors in the dense heart of a dying star, and you can change how efficiently the “heating battery” runs.

For years, modelers treated flavor conversion as a curiosity that would not matter much where the explosion decision is made. Then dense-neutrino physics got stranger. When neutrinos are packed tightly enough to interact with each other, collective flavor conversion can fire up near the proto-neutron star — not only far away in vacuum. That is the regime Gogilashvili and Tamborra lean on: not a slow textbook oscillation across light-years, but a schematic instantaneous flavor equipartition triggered below a chosen baryon-density threshold, spanning from near the neutrinosphere out toward the stalled shock.

They ran the open-source, spherically symmetric code GR1D with energy-dependent neutrino transport, the SFHo equation of state, and a reduced-dimensionality turbulence model (STIR) to stand in for convection that full 3-D codes resolve more honestly. Flavor conversion is applied shortly after bounce in regions below density cutoffs from 10⁹ to 10¹³ g/cm³. Success means the shock expands past 1,000 km; otherwise the model counts as a failed supernova.

The numbers that flip the sky’s ledger

Without flavor conversion, 25.6% of their models fail to explode — about 27% when weighted by a Salpeter initial mass function, in the ballpark of observational hints that a large share of massive-star deaths may be optically dark or faint.

Turn flavor conversion on, and the failed fraction climbs. IMF-weighted failure rates run from 47.6% at the lowest density threshold they tested to 88.3% at the highest. Raw model failure fractions go even higher in some cases (up to 96.4% of the simulated set). The paper’s own table is the cleanest scoreboard: deeper, denser conversion generally means more cooling into heavy-lepton flavors, less efficient electron-flavor heating, and more black-hole endings.

Tamborra’s quote to Space.com lands the surprise: stars between 16 and 30 solar masses — “many of which explode comfortably in our standard simulations” — become particularly sensitive. A large number of them fail once flavor conversion is included.

That matters because astronomy has two long-running headaches in exactly this mass neighborhood. The red-supergiant problem: identified Type IIP supernova progenitors seem to cut off below the heaviest red supergiants we see in the wild. The supernova-rate problem: the cosmic core-collapse rate inferred from explosions sits systematically below the rate you would expect from star formation. More failed explosions in the 16–30 solar-mass window is not a full solution — but it is a physically motivated way to ease both tensions without inventing a new class of stars.

Lighter neutron stars when the blast still wins

Flavor conversion does not only kill explosions. When conversion happens near the stalled shock in a way that accelerates shock revival, successful models leave less massive compact remnants. Earlier revival means less post-bounce accretion onto the proto-neutron star. The paper reports baryonic remnant masses for successful explosions spanning roughly 1.2–1.9 solar masses, systematically lower with flavor conversion than without — after the usual baryonic-to-gravitational correction, a better match to the observed neutron-star mass peak around 1.2–1.4 solar masses.

In other words: the same microphysics that can erase a supernova can also shave the survivors. That dual effect is why the authors call flavor conversion a fundamental ingredient in predicting neutron-star and black-hole populations, not an optional garnish on the neutrino transport module.

What this is — and what it is not

It is not a claim that every 20-solar-mass star is doomed, that black holes suddenly outnumber neutron stars by a known factor, or that 3-D supernova theory is finished. The treatment of flavor conversion is schematic — instantaneous equipartition below a density cut, not a fully dynamical, multi-angle quantum-kinetic evolution. The hydrodynamics are 1-D+ with a turbulence surrogate. High failure fractions at the densest thresholds (~74–88% IMF-weighted) sit in tension with some observational caps; the authors say so plainly and flag the need for subgrid models that evolve in time and space inside multi-dimensional simulations. Nuclear equation of state, progenitor set, and missing proto-neutron-star convection can all move the percentages.

It is the first broad survey — across nearly two hundred progenitors — arguing that neutrino flavor conversion systematically reshapes which stars explode, which fail, and how heavy the leftover neutron stars are. Peer-reviewed acceptance in Physical Review D (accepted August 5, 2026; DOI 10.1103/pz3y-3lv5), with an earlier arXiv posting (2605.16504), and a same-day public science wave on September 20, 2026.

Ghosts with veto power

Supernovas built the periodic table’s heavy half and seeded the dust that becomes planets. Black holes rewrite spacetime and dominate gravitational-wave catalogs. The hinge between those outcomes has always been energy transport in a region denser than a nucleus and hotter than anything on Earth. Gogilashvili and Tamborra’s contribution is to show — in a controlled, if idealized, simulation campaign — that identity swaps among neutrinos can tip that hinge.

Future work they flag is the obvious next chapter: more realistic flavor physics that starts and evolves as the star dies, inside 3-D cores that already strain the world’s supercomputers. Until then, the public takeaway is sharp enough for a Sunday night share. The universe’s most plentiful messengers may also be its quietest executioners — deciding, particle by particle, whether a dying star gets a fireworks display or simply disappears.

Sources

Gogilashvili, M., & Tamborra, I. (2026). Neutrino flavor conversion shapes the rate of failed core-collapse supernovae. Physical Review D (accepted Aug. 5, 2026). https://doi.org/10.1103/pz3y-3lv5
Gogilashvili, M., & Tamborra, I. (2026). Same title (preprint). arXiv:2605.16504. https://ar5iv.labs.arxiv.org/html/2605.16504
Choi, C. Q. / Space.com. (2026, September 20). Flavor-changing 'ghost particles' may reveal which stars go supernova — and which turn into black holes. https://www.space.com/astronomy/stars/flavor-changing-ghost-particles-may-reveal-which-stars-go-supernova-and-which-turn-into-black-holes