In 2017, the LIGO and Virgo detectors caught gravitational waves from the merger of two neutron stars. A kilonova flared: a brilliant cosmic explosion thought to forge the heaviest elements in the periodic table. Gold in jewelry, platinum in catalysts, europium in smartphone screens: all of it, according to a beautiful legend, came to us from such catastrophes. But when astronomers tallied the reserves, the legend began to crack.

Imagine a forge where heavy elements are hammered out in the furnace of stellar catastrophes. The hammers are neutron star mergers. But the hammers strike too rarely and too late to forge as much metal as we see in the disk of the Milky Way. So astrophysicists decided to test other, more exotic hammers: fast mergers that occur almost immediately after the birth of stars, and mergers of neutron stars with black holes.

The scientists built a unified Bayesian model of the Galaxy's chemical evolution that accounts for uncertainties in merger rates, delay time distributions, ejecta masses, and observational data. The model included three scenarios: fast mergers visible to gravitational-wave detectors; fast mergers that detectors miss; and neutron star-black hole mergers as an additional channel. For each scenario, they computed the probability of reproducing the observed [Eu/Fe] versus [Fe/H] relation for disk stars.

The result was harsh. If fast mergers are visible to detectors, they must make up more than 98% of all neutron star mergers: that is almost every hammer in the forge. If they are invisible, their rate must be 17 times higher than the rate of delayed mergers. And neutron star-black hole mergers, even under optimistic assumptions, contribute less than 10%. In other words, more than 90% of r-process enrichment must come from an unknown channel that tracks the star formation rate.

So where have the missing hammers gone? Perhaps they are magnetorotational supernovae: rare explosions of massive stars with strong magnetic fields, or collapsars. These exotic events happen quickly, right after star birth, and could forge heavy elements at the required rate. Future observations of the most metal-poor stars and improved supernova models will help narrow down the candidates. For now, the mystery remains: the Universe's forge is missing hammers, and we are only beginning to search for where they are hidden.