A hidden Y-shaped junction of gluons inside protons may carry one of matter’s most fundamental properties, challenging a decades-old textbook assumption.
New findings from the STAR detector at the Relativistic Heavy Ion Collider (RHIC) suggest that gluons, the particles that act like glue between quarks inside protons, may play a much bigger role in preserving baryon number than scientists once thought. Baryon number is a fundamental part of a particle’s quantum identity.
The results come from high-energy particle collisions at RHIC, a U.S. Department of Energy (DOE) Office of Science user facility for nuclear physics research that operated at DOE’s Brookhaven National Laboratory from 2000 to early 2026. The new study, published in Science, suggests that baryon number may be carried by a Y-shaped “junction” of gluons linking the proton’s three main quarks.
If confirmed, the result would challenge the long-standing assumption that baryon number is carried only by those three quarks.
“Traditionally, scientists have assumed that each of the three main ‘valence’ quarks inside a proton or neutron carries one-third of the baryon number,” said Zhangbu Xu, a professor at Kent State University with a joint appointment at Brookhaven Lab.
A Decades-Old Idea About Gluon Junctions
Physicists first proposed the baryon junction, also called a gluon junction, in the 1970s as a way to describe how gluons connect the valence quarks inside protons.
Then, in 1996, four years before RHIC began operating, theoretical physicist Dmitri Kharzeev of Stony Brook University and Brookhaven Lab proposed that this junction might do more than hold the quarks together. He suggested that it could actually carry baryon number instead of the valence quarks themselves.
The STAR collaboration has now developed a way to test that possibility using several types of particle collisions at RHIC.
“Using data collected from different types of particle collisions at RHIC, our results suggest that the baryon number is not simply carried by individual quarks,” Xu added. “Our findings strongly support the idea that baryon number is more favorably carried and transported by gluons, the particles that hold quarks together, when arranged in this special configuration.”
Why Baryon Number Matters
Determining what actually carries baryon number could have consequences far beyond the collisions produced at RHIC.
In those experiments, baryon number conservation means that the total number of baryons, particles made from three quarks such as protons and neutrons, must remain consistent before and after a collision. But this conservation law also applies on a much larger scale across the universe.
“Since the Big Bang, the number of protons and neutrons all together never changes as a function of time,” said Nicole Lewis, a STAR physicist at Rice University who started this project as a postdoc at Brookhaven Lab in 2020. “The reasons for this conservation are not well understood. It’s one of the mysteries of the universe, related to why we have more matter than antimatter,” she said.
Baryon number conservation is also tied to the extraordinary stability of protons, which form a central part of atomic nuclei and do not appear to decay under ordinary conditions.
“It’s believed that the lifetime of a proton is longer than the lifespan of the universe,” Lewis said. “This allows atomic nuclei to form and be stable — which means matter, as we interact with it in the universe, can exist.”
A Textbook Picture Comes Under Pressure
The possibility that gluons help carry baryon number conflicts with the conventional explanation found in many physics textbooks.
A proton has a baryon number of plus one. In the traditional picture, that value is divided evenly among its three primary valence quarks, giving each quark a baryon number of plus one third. The idea is similar to how a proton’s electric charge is distributed among those same valence quarks.
But a real proton is far more complicated than a simple collection of three quarks.
“In the naïve quark model, there are three quarks inside a proton, but nothing else,” said Tommy Tsang, formerly a postdoc at Kent State University, now at DOE’s Argonne National Laboratory. “But if we look at details inside, there are not only three quarks but also a lot of gluons interacting, connecting between those quarks, and there are also quarks and antiquarks that pop up from the vacuum, so it’s actually a really complex object.”
That complex environment is described by quantum chromodynamics (QCD), the theory of the strong force that governs interactions among quarks and gluons. QCD has been extremely successful, but models inspired by it sometimes require extra assumptions to explain what scientists observe when nuclei moving close to the speed of light collide at RHIC.
An Unexpected Excess of Baryons
One observation has been especially puzzling.
“In the STAR detector, we consistently see an excess of baryons coming out of the collisions perpendicular to the direction of the colliding beams,” Tsang said. “The fact that we end up with more baryons than antibaryons — or more matter than antimatter — is not surprising since our collisions start with matter,” he said.
These extremely energetic collisions convert enormous amounts of energy into thousands of newly created particles.
What caught researchers’ attention was where the extra baryons appeared. STAR detected a net excess of baryons emerging away from the beamline.
If valence quarks alone were responsible for carrying baryon number, all three quarks from an incoming proton would need to stop during the collision. They would then have to convert from matter into energy and eventually back into matter near the center of the detector, producing the excess baryons seen moving away from the beamline.
The STAR team suspected that another mechanism might be involved.
Electric Charge Provides a Critical Test
The researchers found a way to test the idea by taking advantage of a property valence quarks definitely carry: electric charge.
They compared the net baryon number produced in different types of RHIC collisions with the way electric charge was redistributed in those same events.
“Measuring the electric charge coming out perpendicular to the collision gives you a definitive way of measuring how many quarks are stopped and transformed into new particles,” said Zebo Tang, a professor at the University of Science and Technology of China who led a group of students performing data analyses and model simulations.
The comparison revealed a striking mismatch.
Researchers observed about twice as many baryons as should have been produced based on the electric charge associated with stopped quarks. According to QCD-based models, the number of quarks being stopped simply was not large enough to account for the baryons appearing in the detector.
That raised an obvious question: if the quarks were not producing all of the excess baryons, what was?
The STAR team argues that the answer may lie with gluons, especially the three-pronged gluon junction connecting a proton’s valence quarks.
How a Gluon Junction Could Carry Baryon Number
The researchers propose that when nuclei collide at RHIC, the quarks and the gluon junction inside their protons do not necessarily behave the same way.
The “gluon junction” or “baryon junction” may be much easier to stop during the collision than the three fast-moving valence quarks. If the junction slows down while the quarks keep traveling along the beam direction, the junction’s energy can instead be converted into new baryons that emerge in other directions.
Understanding how this could happen requires looking at what occurs inside a proton as it is accelerated to extremely high energies.
“The baryon junction is always there even as protons are accelerated to higher and higher energy,” said Prithwish Tribedy, a STAR physicist at Brookhaven Lab. “But at high energy, gluons within the proton split and multiply.”
As the number of gluons increases, each individual gluon carries a smaller share of the proton’s overall momentum. That includes the gluons that make up the junction. The valence quarks, meanwhile, continue to carry much of the proton’s forward momentum.
As a result, when the collision happens, the relatively slower three-pronged gluon structure may be easier to stop and convert into new particles than the rapidly moving quarks.
The fact that only one structure needs to be stopped also makes the process more likely than a scenario requiring all three valence quarks to stop independently, Tribedy said.
“In the collision, the baryon junction gets held behind, and the quarks continue on,” he noted.
A Y-Shaped Junction Rebuilds Into New Matter
Quarks and gluons cannot exist freely on their own, so after the collision they rapidly combine with other particles.
In a simplified example, each quark continuing down the beampipe could pair with an antiquark to form a two-quark particle called a meson. The three-pronged gluon junction, meanwhile, could behave somewhat like a Y-shaped magnet, drawing in three new quarks from the vacuum and forming a new baryon.
Real RHIC collisions are far more complex.
“Even though we start with nuclei that contain roughly 100 protons and 100 neutrons, these collisions create thousands of new particles; 99% of the energy is transformed into new particles,” said Rongrong Ma, a Brookhaven Lab physicist.
The researchers found that collisions producing larger numbers of particles also showed a greater excess of “midrapidity” baryons compared with what would be expected if quarks alone carried baryon number.
The large number of baryons appearing perpendicular to the beamline therefore provides strong support, the researchers say, for the existence and importance of the baryon junction.
“Our research challenges the long-held idea that baryon number is simply divided among and carried by the three quarks,” said Ma. “This new understanding reshapes how we think about the structure of matter and deepens our knowledge of the most fundamental element that is responsible for the universe in its current form.”
Reference: “Tracking the baryon number with nuclear collisions” by STAR Collaboration*, B. E. Aboona, J. Adam, L. Adamczyk, I. Aggarwal, M. M. Aggarwal, Z. Ahammed, A. K. Alshammri, E. C. Aschenauer, S. Aslam, J. Atchison, V. Bairathi, X. Bao, P. Barik, K. Barish, S. Behera, R. Bellwied, P. Bhagat, A. Bhasin, S. Bhatta, S. R. Bhosale, J. Bielcik, J. Bielcikova, J. D. Brandenburg, C. Broodo, X. Z. Cai, H. Caines, M. Calderon de la Barca Sánchez, D. Cebra, J. Ceska, I. Chakaberia, P. Chaloupka, Y. S. Chang, Z. Chang, A. Chatterjee, D. Chen, J. Chen, J. H. Chen, L. Chen, Q. Chen, W. Chen, Z. Chen, J. Cheng, Y. Cheng, W. Christie, X. Chu, S. Corey, H. J. Crawford, M. Csanad, G. Dale-Gau, A. Das, D. De Souza Lemos, I. M. Deppner, A. Deshpande, A. Dhamija, A. Dimri, P. Dixit, X. Dong, J. L. Drachenberg, E. Duckworth, J. C. Dunlop, …, H. Zbroszczyk, W. Zha, C. Zhang, D. Zhang, J. Zhang, K. Zhang, L. Zhang, S. Zhang, W. Zhang, X. Zhang, Y. Zhang, Y. Zhang, Y. Zhang, Y. Zhang, Z. Zhang, Z. Zhang, F. Zhao, J. Zhao, S. Zhou, Y. Zhou, C. Zhu, X. Zhu, M. Zurek and M. Zyzak, 13 August 2026, STAR Collaboration.
DOI: 10.1126/science.ads5962
The research was supported by the DOE Office of Science, the U.S. National Science Foundation (NSF), and several international agencies and organizations listed in the scientific paper. The team also used the Open Science Grid, supported directly by NSF, along with computing resources from the Scientific Data and Computing Facilities at Brookhaven Lab and the National Energy Research Scientific Computing Center (NERSC), another DOE Office of Science user facility located at DOE’s Lawrence Berkeley National Laboratory.
Never miss a breakthrough: Join the SciTechDaily newsletter.
Follow us on Google and Google News.