What is left when a star dies?

It swells, sheds its outer layers, and leaves the exposed core behind: a white dwarf, roughly the size of Earth and held up not by heat but by the refusal of electrons to occupy the same state.

Nothing is burning. It cools, for a very long time, and the oldest ones in our galaxy are useful clocks because that cooling can be modelled. Some are exceptions, holding steady for billions of years, which one model explains by crystals forming inside them and floating up.

Above roughly eight solar masses, fusion continues past carbon into heavier elements until the core is iron. Iron pays nothing, the core gives way in less than a second, and the rebound sends a shock wave out through the star: a supernova.

What survives is either a neutron star, a city-sized object with the mass of the Sun held up by nuclear forces, or a black hole, where nothing known holds it up at all.

The maximum mass of a neutron star cannot yet be calculated, because it depends on how matter behaves at several times the density of an atomic nucleus, and no laboratory experiment reaches that. It is constrained by observation instead, and it sits somewhere around two and a bit solar masses.

Gravitational-wave observations of mergers now contribute to that constraint, which is a new and useful kind of evidence: a measurement of nuclear physics made by listening to spacetime. Combining the binary mass of the 2017 merger with what the light that followed it ruled out, Margalit and Metzger put the ceiling at about 2.17 solar masses with ninety per cent confidence, which they call tighter, or at least less model-dependent, than the constraints that came before it. A later analysis of the same event, adding the conservation of angular momentum, finds the ceiling only weakly constrained, at about 2.3. And a pulsar weighed in 2022 at 2.35 solar masses, give or take 0.17, the heaviest well-measured to that date, pushes from below: with other pulsars like it, its measurers put the ceiling above 2.19 at one standard deviation.

The interior may be neutrons all the way down, or it may contain free quarks, hyperons, or a phase of matter with no laboratory analogue. The candidates make different predictions for how easily the star deforms, which is measurable in a merger.

The current data narrow the field without choosing a winner.

The mapping from initial mass to final object is not clean. Rotation, metallicity, binary companions and the details of how the explosion launches all shift the outcome, and some stars appear to collapse without any explosion at all.

One survey, watching 27 nearby galaxies for more than a decade, has found two stars that faded from view without a bright supernova. Its team reads them as possible failed supernovae; infrared images of the first from JWST also fit two stars merging. It is a difficult thing to prove and an interesting thing to chase.

Each ending doubles as an instrument. White dwarf cooling dates stellar populations. Pulsars test relativity and, timed together across the sky, have produced evidence of gravitational waves whose cycles last years. Merging remnants make heavy elements and measure the distance to their own galaxies.

A dead star is not the end of the astronomy. In several cases it is the beginning of the useful part.