Oxygen Collisions at RHIC Recreate Early-Universe Matter at Lower Energies Than Ever Before

On February 6, 2026, the Relativistic Heavy Ion Collider at Brookhaven National Laboratory ended its run the way it had been spending its final months: colliding oxygen nuclei, with both the STAR and sPHENIX detectors recording. The machine had operated since 2000. Now a new analysis of earlier oxygen data, published in Physical Review Letters, shows that those small collisions carry a signature of the matter that filled the universe in its first microseconds.

That matter is quark-gluon plasma, or QGP, believed to have existed before protons and neutrons formed. RHIC was built to make it. The usual method is to accelerate gold nuclei, each with 79 protons and 118 neutrons, and smash them together. At sufficient energy, the boundaries of the protons and neutrons break down and release their quarks and gluons, while the impact converts energy into thousands more. The result is a droplet smaller than an atomic nucleus that lasts less than a billionth of a billionth of a second before recombining into ordinary particles.

Because the droplet vanishes so fast, no one can examine it directly. Physicists read the debris instead. One key clue is jet quenching: narrow sprays of energetic particles, called jets, lose energy as they cross the plasma. Li Yi, a Shandong University physics professor who oversaw the STAR analysis, compared it to a bullet slowing as it passes through water. Jets that travel farther through the material lose more, because they meet more quarks and gluons. Jets formed near the outer edge cross less.

Jet quenching has been seen for decades in gold collisions at RHIC. A second signature, collective flow, was a surprise: low-energy particles moved in coordinated patterns, as though the plasma were an almost perfect liquid with extraordinarily little internal resistance. Later, similar flow turned up in much smaller systems, including protons, deuterons and helium-3 nuclei striking gold or lead, and even proton-proton collisions at the Large Hadron Collider. That raised the question of whether QGP was forming there or something else explained it. Rongrong Ma, a STAR physicist at Brookhaven, said no unambiguous jet-quenching signal had been seen in such systems, so the matter was still debated. The goal was to find the smallest system that still shows one.

Oxygen was the candidate. Its nucleus has just eight protons and eight neutrons and a fairly spherical shape, which simplifies the geometry. Maowu Nie, an associate professor at Shandong University, said this makes it easier to look for lopsided quenching in pairs of back-to-back jets, or di-jets, made near the edge of the interaction region in head-on collisions. The team could also compare those central collisions, which produce the most particles and are most likely to form QGP, with grazing ones that produce far fewer particles and are not expected to form it.

Sijie Zhang, who did the work for her master's thesis at Shandong and is now a Ph.D. student at Yale, called the measurements essentially counting experiments: an algorithm finds the jets and tallies how many particles above an energy threshold emerge. Yang He, a former Shandong Ph.D. student who built the analysis framework and now studies oxygen collisions at the LHC, described the comparisons against edge-produced jets and non-central collisions. The team used STAR data from oxygen collisions at RHIC in 2021, at 200 GeV. They found a 20% suppression of high-energy particles. That is smaller than typical in gold collisions, as expected for a smaller system, but strong enough to count as clear evidence that QGP formed.

The energy is the notable part. Oxygen is the only species besides protons studied at both RHIC and the LHC, allowing an apples-to-apples comparison. The LHC ran its first oxygen collisions in 2025, and CERN reported both collective flow and jet quenching, at energies above 5,000 GeV. Yi noted that higher energy should produce a hotter, denser QGP and a larger quenching signal. Ma said the RHIC result pushes in two directions at once, toward smaller nuclei and lower energy, and still sees the signature.

More data are waiting. The oxygen runs recorded by STAR and sPHENIX at the end of RHIC's operations could show how small and how low in energy a collision can be before quark-gluon plasma stops forming. The work was supported by the DOE Office of Science, the National Science Foundation and other agencies named in the paper.