The oldest experiment, run again.

For well over a thousand years, alchemists pursued chrysopoeia — the making of gold. Their guiding fiction was the philosopher’s stone: an imagined agent that could perfect base matter and, some hoped, the person who held it.

Mercury sat at the centre of the work. Liquid at room temperature, mirror-bright, impossible to hold, it was treated as a principle of metals itself. The stone was never found, and alchemy was never a science — but its furnaces, distillations and patient notebooks helped prepare the ground for chemistry.

Chemistry cannot turn one element into another: an element is defined by the protons in its nucleus. Nuclear reactions can. In 1919 Ernest Rutherford reported the first artificial transmutation, turning nitrogen into oxygen with alpha particles.

Mercury (80 protons) and gold (79) are neighbours on the periodic table. In 1941, physicists at Harvard made radioactive gold by bombarding mercury with fast neutrons. One route to stable gold: a neutron strikes mercury-198 and knocks two neutrons out, leaving mercury-197, which decays into gold-197 over a few days.

The physics is real. As a way of making gold, it has so far been wildly uneconomic — in 1980, a team led by Glenn Seaborg produced a vanishingly small number of gold atoms from bismuth at enormous cost.

In 2025 the ALICE collaboration at CERN reported that near-miss collisions between lead nuclei in the Large Hadron Collider briefly create gold nuclei — around 29 trillionths of a gram across several years of operation, each nucleus breaking apart almost at once.

The same year, the fusion start-up Marathon Fusion published a preprint proposing that the neutron blankets of future fusion power plants could convert mercury-198 into gold-197 as a by-product. That proposal was theoretical: its production figures were company-reported and based on simulations, and no power plant of that kind exists yet.