Nuclear transmutation without a head-on collision
The ALICE collaboration at CERN reported in May 2025 that it had systematically measured lead nuclei transforming into gold during near-miss encounters in the Large Hadron Collider. The phrase sounds like alchemy, but the mechanism is standard nuclear physics: change the number of protons in a nucleus and the element changes. Lead has 82 protons, while gold has 79, so ejecting three protons from a lead nucleus produces a gold isotope.
These events are not the violent central collisions for which the LHC is best known. In an ultraperipheral collision, two lead nuclei pass close enough for their electromagnetic fields to interact but remain too far apart for the strong nuclear force to make them collide directly. At almost the speed of light, each nucleus's intense electromagnetic field appears as a short pulse of photons to the other nucleus.
How a pulse of light changes a nucleus
A photon can excite the lead nucleus and trigger electromagnetic dissociation. The disturbed nucleus then releases neutrons and sometimes protons. Removing no protons leaves lead; removing one, two or three produces thallium, mercury or gold respectively. Different neutron counts select different isotopes of those elements.
ALICE identified the outgoing fragments with zero degree calorimeters positioned far forward along the beam line. These detectors are suited to particles that continue close to the original ion direction. The collaboration measured channels in which zero through three protons were accompanied by at least one neutron during lead-lead running at a centre-of-mass energy of 5.02 teraelectronvolts per nucleon pair.
The Physical Review C paper compares the measured cross sections with the RELDIS electromagnetic-dissociation model. The model described the zero-proton and three-proton channels within measurement uncertainties but underestimated the one- and two-proton rates by roughly 17 to 25 percent. Those differences give theorists specific nuclear dynamics to improve.
How much gold is actually made
At the ALICE interaction point, current lead running can produce gold at a maximum rate of about 89,000 nuclei per second. That large-sounding count corresponds to an almost unimaginably small mass. CERN estimates that all four major LHC experiments created about 86 billion gold nuclei during Run 2, only 29 picograms in total.
The nuclei also do not accumulate in a recoverable nugget. They leave the interaction with nearly the beam energy, strike accelerator components downstream and fragment almost immediately. Producing them requires a major accelerator, so their economic value is negligible compared with the operating effort. The result is a sensitive physics measurement, not a manufacturing method.
Why the measurement matters to the collider
Electromagnetic dissociation is one source of beam loss in heavy-ion accelerators. When a lead ion changes its charge or mass, magnets bend it differently from the main beam. The altered ion eventually leaves the designed orbit and deposits energy in the beam pipe or a collimator. Accurate production rates therefore help engineers predict where losses occur and protect present and future colliders.
The experiment also tests how an intense, rapidly varying electromagnetic field couples to a heavy nucleus. ALICE can study both central collisions that create thousands of particles and ultraperipheral events with only a few forward fragments. That range turns equipment designed for quark-gluon-plasma physics into a precise probe of photonuclear reactions.
What was and was not discovered
Artificial gold was made before 2025 using other nuclear reactions. ALICE's advance was the first systematic experimental detection and quantification of gold production by this near-miss mechanism at the LHC. The collaboration inferred the elemental products from measured proton and neutron emission with a tested reaction model; it did not collect and chemically assay a piece of metal.
The playful alchemy comparison makes the result memorable, while the durable contribution is quantitative. Cross sections for rare proton-emission channels now provide new constraints on electromagnetic-dissociation models and on a practical limit to heavy-ion accelerator performance.
