Hadronization is the process through which quarks and gluons form hadrons, the composite particles governed by the strong interaction. It connects the elementary constituents produced in high-energy reactions with the color-neutral particles observed experimentally. Within quantum chromodynamics (QCD), this conversion involves long-distance, nonperturbative dynamics. It is often called fragmentation, although some simulation conventions use fragmentation specifically for hadron formation and hadronization more broadly for subsequent particle decays and related processes. (pythia.org)
Physical basis
Quarks and gluons carry color charge, whereas observable isolated hadrons are color singlets. Color confinement prevents ordinary colored partons from appearing as free asymptotic particles. Hadronization is the dynamical formation of hadrons from a colored system; confinement explains why the final observable states cannot simply remain isolated quarks and gluons. The two concepts are therefore closely related but not interchangeable. (pythia.org)
At large momentum-transfer scales, asymptotic freedom makes the QCD coupling relatively small, allowing calculations based on perturbation theory. At scales of roughly a gigaelectronvolt and below, the coupling grows and this expansion loses its usefulness for describing hadron formation. Collision simulations therefore supplement perturbative calculations with phenomenological models rather than calculate every final hadron directly from the QCD Lagrangian. (pythia.org)
The resulting particles include mesons, whose simplest valence structure is a quark–antiquark pair, and baryons, whose ordinary valence structure contains three quarks, together with their antiparticles. Their production involves additional quark–antiquark creation and the redistribution of the original system’s energy and momentum. A high-energy parton does not generally become one identifiable hadron: a connected colored system produces a collection of hadrons. (pythia.org)
From a hard collision to hadrons
A useful example is electron–positron annihilation into a quark–antiquark pair. Before hadron formation, the energetic partons radiate gluons, which can radiate further gluons or split into quark–antiquark pairs. This cascade is represented by a parton shower. Once its evolution reaches a low-scale cutoff, a hadronization model converts the remaining colored configuration into primary hadrons. Unstable hadrons subsequently decay, changing the particle species and multiplicities ultimately recorded. (pythia.org)
The particles often remain concentrated around the directions of the energetic parent partons, forming particle jets. Nevertheless, hadronization is not necessarily independent for each jet: color connections can extend between outgoing partons and, in hadron collisions, beam remnants. The separation between shower evolution and hadronization is a modeling convention, and predictions depend on how these stages are matched. (arxiv.org)
String and cluster models
The Lund string model represents the confining color field as a flux tube between connected color charges. Its energy approximately increases with separation. Support for this picture comes from lattice QCD calculations of the confining potential. As the system stretches, quark–antiquark pairs are created and the string breaks into color-singlet segments that become hadrons. Gluons appear as kinks along the string rather than as independent endpoints. This framework underlies ordinary string fragmentation in PYTHIA. (pythia.org)
String fragmentation proceeds iteratively. Probabilistic rules select the flavors created at breaks, the species of the resulting hadrons, their transverse momenta, and their shares of the remaining longitudinal energy–momentum. Baryon production requires additional mechanisms, such as diquark–antidiquark creation or string-junction configurations. Parameters controlling these choices are constrained by experimental distributions rather than uniquely fixed by perturbative QCD. (pythia.org)
The cluster hadronization approach instead groups color-connected partons into color-singlet clusters. It is motivated by preconfinement: suitable perturbative showers organize color connections into systems with characteristic masses governed mainly by the shower cutoff. Remaining gluons are split into quark–antiquark pairs; clusters then decay into hadrons, while sufficiently massive clusters undergo further splitting. Herwig implements this general approach. String and cluster models describe the same physical conversion using different effective constructions. (lss.fnal.gov)
Fragmentation functions
An inclusive description uses fragmentation functions, conventionally written . These encode the distribution of a specified hadron carrying a fraction of a parent parton’s longitudinal momentum at scale . They are not complete event-by-event pictures of string breaking or cluster decay. Instead, they provide nonperturbative quantities that enter factorized predictions for identified-hadron production. (pdgweb.lbl.gov)
Their input distributions are determined from data, while QCD evolution equations describe their scale dependence. Within the conditions of factorization, the same functions can be applied across different reactions, including electron–positron annihilation, deep inelastic scattering, and hadron collisions. This universality allows measurements from different experiments to constrain a shared description, subject to the relevant approximations and power corrections. (pdgweb.lbl.gov)
Dense matter and experimental constraints
Hadronization from a quark–gluon plasma introduces a surrounding medium containing many interacting partons. Quark coalescence, also called recombination, models hadron formation through the combination of nearby quarks and antiquarks. Such models describe aspects of baryon-to-meson yields and collective-flow patterns in heavy-ion collisions. They complement fragmentation descriptions, with their relative importance depending on momentum and the medium’s properties. (arxiv.org)
Experimental constraints include identified-particle spectra, charged-particle multiplicities, baryon and strange-hadron yields, and correlations among particles. Monte Carlo event generators adjust model parameters to reproduce these measurements and test their applicability across collision energies and processes. Hadronization uncertainty remains relevant when translating parton-level calculations into measurable hadronic observables; comparisons between model constructions and parameter variations help characterize that uncertainty. (arxiv.org)