On August 18, *Science* published an experimental result that may rewrite particle physics textbooks. The STAR international collaboration—led by the University of Science and Technology of China, Kent State University, and Brookhaven National Laboratory—used high-energy nucleus-nucleus collisions at the Relativistic Heavy Ion Collider (RHIC) to precisely measure net baryon number and net charge. The data indicate that nucleons require a fundamental structure that "carries baryon number but no electric or color charge," posing a major challenge to the naive quark model.
The experiment did not directly "see" inside the proton. STAR used RHIC as a particle-scale stress-testing platform: it analyzed several types of high-energy collision events, including isobar collisions and photon-nucleus collisions, systematically compared the spatial distributions of baryons versus their antimatter counterparts after collisions, and cross-validated the new results with earlier gold-gold collision measurements. The key clue came from an unexpected "baryon excess": the STAR detector persistently observed large numbers of baryons emitted perpendicular to the collision beam direction.
The naive quark model gives a simple picture: a proton consists of three valence quarks (uud), each carrying baryon number +1/3 and charge +2/3 or −1/3, summing to baryon number +1 and charge +1. If baryon number belonged only to these three valence quarks, the total number of slowed quarks could be inferred from the number of slowed baryons, yielding an expected net charge. But the measured charge was only half the expectation—meaning the nucleon must contain, besides quarks and gluons, a structure carrying baryon number without charge. Neither valence quarks nor single gluons in the traditional model have this property.
The only picture that consistently describes both sets of results is the Y-shaped gluon baryon junction, a topological structure theoretically predicted in the 1970s: gluons cannot merely connect the three valence quarks pairwise into a triangle; they must form a Y-shaped field structure whose central junction is the baryon junction. In 1996, theorist Dmitri Kharzeev proposed that what truly carries baryon number may not be the quarks but the junction itself—it can pull three new quarks out of the vacuum to form a new baryon, thereby preserving baryon number. STAR's results provide the first systematic experimental support for this hypothesis.
Why the "baryon excess" is critical
RHIC's isobar and photon-nucleus collisions cover different modes of nucleon deceleration. If the three valence quarks alone carried baryon number, the number of slowed baryons and slowed charge would follow a fixed ratio (the per-quark B/Q ratio follows from the quark model). The measured B/Q far exceeds the naive prediction—something non-charged is being carried along when baryons decelerate. The Y-shaped junction geometry fits exactly: the gluon junction carries no charge but can "repackage" a baryon by pulling three quarks from the vacuum, and those pulled quarks contribute no systematic charge.
Why baryon number conservation matters
Baryon number (B) is one of the oldest and most stable conserved quantities in particle physics. Since the Big Bang, the total number of baryons in the visible universe has barely changed—the material world of protons and neutrons we live in exists because this conservation law holds. But baryon number conservation remains an observational fact without the dynamical explanation that charge conservation enjoys. The Y-junction picture shifts the mechanism from "intrinsic quark property" to "gluon topological structure"—if consistently confirmed, the explanation would be rooted in QCD's non-Abelian symmetry rather than quark statistics.
A capstone of 25 years at Brookhaven
RHIC began operating in 2000 and will complete its final collision run in early 2026. This STAR result is not an isolated highlight but the culmination of 25 years of data—gold-gold, copper-copper, uranium-uranium, isobar, and photon-nucleus collision modes all yield a consistent picture. Such cross-mode consistency carries statistical strength no single experiment could fake, and for the first time brings the decades-old question "is the naive quark model sufficient?" close to an answer.
What it changes
The Y-shaped gluon junction evidence opens a window: baryon number conservation may not be a byproduct of quark statistics but a geometric necessity of gluon topology. Future heavy-ion colliders, the Electron-Ion Collider (EIC), and even high-energy cosmic-ray observations will need to revisit the question of where baryon number comes from. The experiment, led by Tang Zebo's team at USTC, moves Chinese fundamental physics on core QCD questions from "major participant" to "leading role"—arguably the most memorable result of RHIC's 25-year finale.