A particle detected in a Chinese accelerator fifteen years ago could consist mainly of gluons, the particles that hold quarks together in protons and neutrons. The latest measurements from the BESIII experiment strengthen this interpretation for X(2370), an object whose nature had remained uncertain.
Usually, particles formed by the strong nuclear interaction contain quarks. Gluons bind them together, somewhat like glue on a scale infinitely smaller than the atom. Yet theory predicts that gluons can also bind to one another without valence quarks. Such an assembly is called a “glueball,” literally a ball of gluons.

A conventional atomic nucleus consists of quarks (shown here as small colored spheres), held together in groups of three by gluons (the white links). Each group of three quarks bound in this way forms a proton or a neutron.
Source: FIJ PAN
The principle has been predicted for decades, but identifying a glueball in experiments is difficult. Particles produced in accelerators disappear almost immediately and are identified through the particles into which they transform. Several candidates may also share similar characteristics.
BESIII studies it through the decays of the J/ψ meson, a particle produced in large numbers at the Beijing electron-positron collider. The experiment now has around 10 billion J/ψ events. This quantity makes it possible to examine very rare transformations and precisely compare the different ways X(2370) can decay.
A recent piece of the puzzle comes from a transformation that physicists expected to be able to detect under certain interpretations of X(2370). They searched for a decay proceeding through a particle called K*(892), but found no evidence of one. This absence provides information about the composition of X(2370).
Specifically, the data indicate that X(2370) behaves like a state that does not favor any of the three light quark families. This property joins several other clues already measured: its mass, quantum characteristics, production in J/ψ meson decays, and the ways in which it transforms all match expectations for the lightest glueball in this category.
The authors describe a glueball as the dominant component, rather than a particle necessarily made of pure gluons. This distinction matters because states produced by the strong interaction can mix. Measurements must therefore distinguish a strong gluonic component from other possible structures that still contain quarks.
Future measurements will be able to test this interpretation through other decay channels and clarify the proportion genuinely attributable to gluons. BESIII has its immense sample of J/ψ events for this purpose, enabling researchers to search for even rarer transformations of X(2370) and compare each of them with predictions.