CERN physicists coax the universe's primordial matter from the smallest atomic nuclei yet

Physicists at CERN's Large Hadron Collider have generated quark-gluon plasma — the extremely hot, dense state of matter thought to have filled the universe for a few millionths of a second after the Big Bang — by colliding some of the smallest atomic nuclei ever used for the purpose. The result, described in a pair of papers published this week, upends an assumption that had guided nuclear physics for two decades: that only heavy nuclei, such as lead, carried enough material to briefly "melt" protons and neutrons into their constituent quarks and gluons.

Researchers working across all four of the LHC's major experiments — ALICE, ATLAS, CMS and LHCb — found evidence of the plasma in collisions between oxygen-16 nuclei, and, in follow-up analyses, between the somewhat larger neon-20 nuclei, at a collision energy of 5.36 trillion electron volts per pair of nucleons.

Oxygen's sphere, neon's bowling pin

The clearest new evidence comes from a study led by physicists at the Niels Bohr Institute at the University of Copenhagen, published in Physical Review Letters as an Editors' Suggestion. The team, working within the LHC's ALICE collaboration, showed that the shape of the debris flying out of these collisions directly reflects the shape of the nuclei that collided. Oxygen-16 nuclei are close to spherical, and their collisions produced a rounded pattern of particles. Neon-20 nuclei, by contrast, are more elongated, and their collisions produced a distinctive "bowling pin" pattern — a geometric signature that let researchers effectively read the shape of the nucleus back out of the plasma it briefly formed.

• Collision energy: 5.36 TeV per nucleon pair, generated in the LHC's 2025 light-ion run
• Oxygen-16 nuclei produced rounded, near-spherical particle-flow patterns
• Neon-20 nuclei produced elongated, "bowling pin"-shaped flow patterns
• All four major LHC experiments reported independent signs of quark-gluon plasma from the run

"We have pushed the boundary for how small the atomic nuclei can be while still recreating this primordial matter," said You Zhou, an associate professor at the Niels Bohr Institute who led the analysis. His colleague, postdoctoral researcher Emil Gorm Dahlbæk Nielsen, said the method effectively turns particle detectors into an imaging tool for the atomic nucleus itself: "The particles from the primordial matter are directly governed by the geometric shape of the atomic nucleus."

Why size matters

Quark-gluon plasma has previously been studied almost exclusively through collisions of heavy nuclei, such as lead-lead collisions at the LHC or gold-gold collisions at Brookhaven National Laboratory's Relativistic Heavy Ion Collider, on the theory that only a large-enough soup of nucleons could sustain the state, even briefly, before it cools and the quarks and gluons lock back into ordinary protons and neutrons. A separate analysis from ALICE, presented at a CERN seminar and detailed on the ALICE collaboration's website, found evidence of "parton energy loss" — a hallmark of quark-gluon plasma in which fast-moving quarks and gluons lose energy as they plow through the hot medium — in the oxygen-oxygen collisions, at a statistical significance of 4.9 sigma, corresponding to roughly a one-in-2-million chance the pattern arose by accident.

"With many measurements still emerging, the results promise an unprecedented view of how strongly interacting matter evolves under extreme conditions from the smallest to the largest systems we can create in the laboratory," said Kai Schweda, spokesperson for the ALICE collaboration.

A related paper from the LHC's CMS collaboration, also published in Physical Review Letters, measured how strongly particles flowed outward in patterns consistent with the two nuclei's different internal geometries, adding independent support to the picture built by ALICE. Together, the studies suggest that quark-gluon plasma can form in systems only about a fifth the mass of the lead nuclei traditionally used, a much lower bar than most nuclear physicists expected even a few years ago. The 2025 run that produced these results was itself part of a broader light-ion program at the LHC, which also collected proton-oxygen collisions intended as a reference point for disentangling collective, plasma-like effects from ordinary particle-physics processes that occur even without forming a hot medium. Researchers say comparing the neon and oxygen results against that proton-oxygen baseline was central to building confidence that the observed effects were genuinely due to quark-gluon plasma formation rather than some more mundane explanation for the particle-flow patterns. The CMS collaboration's own account of the broader light-ion program, including its earlier findings from the same 2025 run, is described on the CMS experiment's website.

A new tool, not just a curiosity

Beyond confirming that small systems can form the plasma, physicists say the technique gives them a genuinely new way to study nuclear structure — a long-standing puzzle in physics, since individual nuclei are far too small to image directly. By colliding nuclei of known, contrasting shapes and reading out the resulting flow patterns, researchers can test predictions about how protons and neutrons arrange themselves inside different nuclei, work that has implications for understanding nuclear stability, radioactive decay and the physics of neutron stars. Additional detail on the findings, including background on how the geometric signal is extracted from particle-flow data, was compiled following the papers' publication.

What comes next

The LHC is now being upgraded into the "High-Luminosity LHC," a project intended to sharply increase the rate of particle collisions the accelerator can deliver, which physicists say should allow far more precise follow-up studies of small-system quark-gluon plasma once it resumes operation later this decade. Researchers involved in the current analyses say they plan additional runs with other light nuclei to map out more precisely where the lower size limit for plasma formation actually lies, and to refine models of nuclear geometry that so far have not fully matched some of the finer details in the new data, including certain triangular-flow measurements that diverged from theoretical predictions.