Quantum Chromodynamics
Where the Strong Force Gets Strange
Picking Up Where Strong Field Left Off
The Strong Nuclear Field page walked through color charge, flux tubes, asymptotic freedom, and quark-gluon plasma – the main phenomena. This page is what those ideas look like once they are turned into a precise theory called quantum chromodynamics. The phenomena all stay the same. What the theory adds is a handful of things that nobody would have guessed from the phenomena alone: a small surprise in the number of gluons, a computational method that predicts hadron masses from first principles, a parameter the Lagrangian allows but nature refuses to use, a missing fraction of the proton's spin, and a zoo of exotic bound states that experimentalists are still sorting out.
QCD was assembled in the early 1970s and has held up under every experimental test since. It is a gauge theory built on a symmetry called SU(3), and the specific mathematical structure of that symmetry is what makes everything else about the theory strange. Gross, Politzer, and Wilczek shared the 2004 Nobel Prize for proving that such a theory could even exist. Everything below follows from the same Lagrangian, less than a single line long.
Why Exactly Eight Gluons
Quarks carry one of three color charges. Gluons mediate the strong force by carrying color too, in the form of a color and an anti-color simultaneously. Three colors and three anti-colors should give nine combinations: red-antired, red-antigreen, red-antiblue, and so on. Count the gluons, though, and there are only eight. One combination is missing, and the missing one is not random.
The combination that drops out is the color singlet: an equal mix of red-antired, green-antigreen, and blue-antiblue. Unlike the other eight, this mixture looks identical under every possible color rotation. It is color-neutral in every reference frame simultaneously. If a singlet gluon existed, it would couple to color-neutral objects like protons and mediate a force between them reaching across arbitrary distances – a second long-range force on top of electromagnetism and gravity. No such force has ever been observed. SU(3) removes the singlet because the mathematics demands the gauge bosons correspond to the eight non-trivial generators of the symmetry, not to the one trivial one.
The number of gauge bosons always equals the number of non-trivial generators of the underlying symmetry. Electromagnetism is based on U(1) and has one generator, so one photon. Weak force is based on SU(2) and has three generators, so three bosons (W+, W−, Z). Strong force is based on SU(3) and has eight, so eight gluons. The pattern looks arbitrary until you realize it is the same pattern as writing down every independent way to rotate a three-dimensional complex object while preserving its length. There are exactly eight. Physics inherited the count from pure mathematics.
Lattice QCD
Asymptotic freedom saves QCD at high energies. At low energies – where all of nuclear physics happens – it does not. The coupling becomes large, perturbation theory stops converging, and writing down Feynman diagrams no longer produces useful answers. For decades this meant that QCD could predict almost nothing about the everyday physics it was supposed to describe. The proton mass. The binding energy of deuterium. The spectrum of mesons. The theory had the rules, and nobody could extract the consequences.
Lattice QCD is the solution, and it works. The idea: replace continuous spacetime with a cubic grid of discrete points, put quark fields at the sites, put gluon field values on the links between sites, and write down the action as a sum over the grid. The path integral, which in the continuum is an impossible infinite-dimensional sum over every possible field configuration, becomes an ordinary – if astronomical – finite integral over the values at every site and link. Monte Carlo sampling does the rest. Generate many random configurations weighted by the QCD action, average whatever physical quantity you care about, and let statistical errors shrink with more samples.
What comes out is quantitative. Modern lattice QCD computes the proton mass from first principles, with no free parameters other than the quark masses that enter as inputs, and gets about 938 MeV – matching experiment to roughly 1%. The spectrum of light hadrons follows. The string tension of the flux tube comes out near 1 GeV per fermi of separation, agreeing with what the meson spectrum implies. The phase diagram of QCD at finite temperature matches heavy-ion collision data. Quark-antiquark potentials, form factors, decay constants, the chiral condensate – all extractable, all matching.
Lattice QCD is the only non-perturbative calculation in the entire Standard Model that has been rigorously confronted with experiment and passed. Running it takes some of the world's largest supercomputers and lattices of millions of sites, and getting the continuum and infinite-volume limits right requires years of methodological care. But it turns QCD from "the theory that explains confinement qualitatively" into "the theory that predicts the proton mass to sub-percent accuracy." The difference is the difference between a framework and a working calculational tool.
The Strong CP Problem
The QCD Lagrangian as written permits one additional term that the phenomena on the Strong Field page never needed. It is parameterized by an angle called theta, which by construction can take any value between zero and 2π. A nonzero theta would mean the strong force violates CP symmetry – it would treat matter and antimatter slightly differently, similar to how the weak force does but through a completely independent mechanism. If theta were near 1, the neutron would carry a measurable electric dipole moment on the order of 10−16 e·cm, well within reach of current experiments.
Experiments searching for a neutron electric dipole moment have set an upper bound of about 10−26 e·cm. Back-propagating, this forces theta to be smaller than about 10−10. A parameter the theory allows to be anything between zero and 2π is experimentally constrained to sit at almost exactly zero. Nothing in the QCD Lagrangian explains why. The other free parameters of the Standard Model take a range of values, small and large, none obviously near a fixed point. Theta is uniquely close to zero, and uniquely hard to account for.
The most widely accepted proposed solution, due to Peccei and Quinn in 1977, promotes theta from a fixed parameter to a dynamical field. A new symmetry drives the field to relax toward zero automatically, and the fluctuations of that field manifest as a new particle: the axion. If axions exist at the right mass, they also happen to be an excellent cold dark matter candidate, tying together two open problems in one framework. The axion connection is explored further on the Dark Matter page. If axions are ruled out in the theoretically favored mass range, the strong CP problem becomes one of the cleanest unsolved puzzles in particle physics. For now it is a problem waiting for an experimental verdict.
The Proton Spin Crisis
A proton has spin 1/2. For most of the twentieth century, the standard explanation was simple: two of the three valence quarks have opposite spins that cancel, and the third quark's spin of 1/2 accounts for the proton. Textbook arithmetic. In 1988, the European Muon Collaboration at CERN scattered polarized muons off polarized protons and measured which fraction of the proton's spin came from the quarks' own spins. The answer was about 30%, not 100%. The other 70% had to come from somewhere else entirely.
Decades of follow-up experiments have filled in some of the missing pieces. Gluon spin, measured at RHIC, contributes roughly 30% by current estimates, though with significant uncertainty. Orbital angular momentum of quarks and gluons – their motion around the proton's center – contributes much of the remainder. Sea quarks, the short-lived quark-antiquark pairs constantly appearing and disappearing in the proton's interior, also contribute a small amount. None of these pieces is surprising individually; QCD has always allowed them. What was surprising was that the valence quarks alone do not come close to accounting for the total.
Nailing down the breakdown precisely is still active research. Lattice QCD calculations have recently reached the point of computing the decomposition directly from the theory, and the numbers are converging with experiment. The Electron-Ion Collider being built at Brookhaven will be dedicated in part to mapping proton spin with enough resolution to finally close the ledger. QCD is doing its job – the spin is conserved and has to live somewhere – but extracting exactly how it is distributed among the internal degrees of freedom has taken forty years and counting.
Exotic Hadrons
The simplest color-neutral combinations are a quark with an antiquark (a meson) and three quarks together (a baryon, like a proton). QCD allows more. Gluons self-interact, so pure glue can in principle form bound states with no quarks at all – glueballs. Four-quark combinations, tetraquarks, can arrange themselves as a diquark paired with an antidiquark. Five-quark combinations, pentaquarks, are allowed. Hybrids combine a quark-antiquark pair with an excited gluon. None of these violates any rule of QCD; the theory just requires that the total color combination come out neutral, and there are many ways to achieve that.
For decades the exotic spectrum was theoretical. Experimental candidates appeared and faded. Then in 2015 the LHCb experiment at CERN confirmed the first pentaquarks, and has since found several more. Tetraquark candidates have been confirmed at BESIII in China, Belle in Japan, and LHCb. Many of these are tightly bound, long-lived enough to be unambiguous, and their decay patterns match QCD predictions for multi-quark states better than for ordinary mesons. Exotic hadron spectroscopy has become a full branch of particle physics in less than a decade.
Glueballs are the holdout. Candidates exist in the 1.5 to 2.5 GeV mass region, but distinguishing a true glueball from an ordinary meson with similar quantum numbers is hard because the states can mix. Recent BESIII data on the X(2370) resonance has been interpreted as a likely glueball candidate, though the community has not fully converged. If glueballs are confirmed, it would be the first direct observation of a bound state made purely of force-carrying particles – something with no analog in electromagnetism or the weak force, because their gauge bosons do not self-interact. QCD's self-coupling is what makes glueballs possible, and finding one would be a direct confirmation of that self-coupling in a way no other experiment can provide.
What QCD Still Cannot Prove
Confinement – the fact that no free quark has ever been observed – is confirmed by every experiment and reproduced by every lattice QCD simulation. It has never been proven analytically from the QCD Lagrangian. Proving it, along with a related statement about the mass gap between the vacuum and the lightest hadron, is one of the seven Clay Millennium Prize Problems. The problem is worth a million dollars and has been open since the prizes were established in 2000. Nobody has collected.
Other gaps are more technical but real. The full decomposition of the proton's spin into quark spin, gluon spin, and orbital contributions is still an active research program, not a settled answer. Nuclear physics computed directly from QCD is barely beginning; lattice calculations now reach deuterium and helium-4, but heavier nuclei are still beyond what available computing can handle. The QCD vacuum has nontrivial topological structure – instantons, theta vacua – that ties into the strong CP problem and has not been fully mapped out. These are not signs that QCD is wrong. They are signs that extracting its full predictions from the Lagrangian is genuinely hard, and will remain hard for a long time.
The Bigger Picture
QCD is a strange theory. It predicts particles that cannot exist in isolation. Its coupling runs the wrong way with energy. It contains a parameter nature refuses to use and has so far not explained. It ties together a phenomenon (confinement) that has never been derived analytically with calculations (lattice QCD) that reproduce the proton mass to sub-percent accuracy. Most of the rules on the Strong Field page are simple consequences of the one-line Lagrangian that sits under everything above.
A century from now, QCD will still be the theory any successor framework has to reduce to at accessible energies. Its quantitative successes are not going to be undone. The remaining open problems – the strong CP puzzle, the confinement proof, the full nuclear physics program – are not signs that QCD is provisional. They are signs that the theory is deeper than the calculations we can currently do with it, and that some of those calculations will take another generation of physicists to push through.
