| Lattice QCD | |
|---|---|
| Name | Lattice QCD |
| Field | Quantum chromodynamics |
| Introduced | 1974 |
| Contributors | Kenneth G. Wilson, Michael Creutz, John Kogut, Leonard Susskind |
| Institutions | CERN, Brookhaven National Laboratory, Fermilab, RIKEN |
Lattice QCD
Lattice QCD is a non-perturbative computational formulation of quantum chromodynamics (QCD) defined on a discrete spacetime lattice. It enables ab initio calculations of hadronic properties, phase structure, and weak matrix elements, providing vital links between the Standard Model and experimental results in particle physics and nuclear physics.
Lattice QCD was proposed to regularize QCD and make its dynamics amenable to numerical evaluation. Pioneered by Kenneth G. Wilson in the context of the renormalization group, it replaces continuous spacetime by a hypercubic lattice and represents gluon fields as link variables in the gauge group SU(3). The approach addresses confinement, chiral symmetry breaking, and the non-perturbative regime where perturbation theory fails. Major motivations include computing hadron masses, decay constants, and form factors relevant to experiments at CERN, BNL and Fermilab and testing the Cabibbo–Kobayashi–Maskawa matrix via lattice determinations of weak matrix elements.
The lattice action discretizes the continuum Yang–Mills theory and fermion fields. The standard gluon action is the Wilson plaquette action, constructed from elementary Wilson loops (plaquettes) of link variables U in SU(3). Fermions require special treatments to avoid the fermion doubling problem identified by the Nielsen–Ninomiya theorem; common discretizations include Wilson fermions, staggered fermions (Kogut–Susskind), domain wall fermions, and overlap fermions based on the Ginsparg–Wilson relation. Lattice observables are expectation values computed via the Euclidean path integral; Monte Carlo importance sampling yields ensembles of gauge configurations weighted by the action. Gauge fixing (e.g., Landau gauge) is sometimes used for specific operators, while most hadronic quantities are extracted from gauge-invariant correlation functions.
Numerical evaluation relies on Monte Carlo algorithms and linear solvers. The seminal algorithmic developments include the Metropolis algorithm, heatbath and overrelaxation updates, and the Hybrid Monte Carlo (HMC) algorithm for dynamical fermions. Solving the Dirac equation on gauge backgrounds employs iterative Krylov solvers such as Conjugate Gradient and BiCGStab, accelerated by preconditioners and multigrid methods. High-performance computing hardware — from dedicated supercomputers to GPU clusters and specialized machines like the QCDOC — and software suites such as Chroma and the USQCD software stack are central. Algorithmic advances from collaborations (e.g., MILC Collaboration, RBC and UKQCD collaborations) reduced critical slowing down and enabled physical-point simulations with light quark masses.
Lattice QCD provides a regulator; renormalization removes lattice artifacts to recover continuum QCD. Approaches include perturbative matching, nonperturbative renormalization schemes such as the Rome–Southampton RI/MOM method, and step-scaling techniques developed by the Alpha Collaboration. Extrapolation to zero lattice spacing (a → 0) requires Symanzik effective theory to classify cutoff effects and improved actions (e.g., clover improvement) to reduce O(a) errors. Chiral extrapolations use chiral perturbation theory to connect simulations at heavier pion masses to the physical point, while finite-volume effects are treated with analytical corrections (e.g., Lüscher's method) and explicit volume studies.
Lattice QCD has produced ab initio determinations of the light hadron spectrum, reproducing observed masses within controlled uncertainties. It calculates decay constants (e.g., f_pi, f_K), semileptonic form factors for weak decays (relevant to extractions of |V_{ub}| and |V_{cb}|), and neutral-meson mixing parameters (e.g., B_K, B_{B}). Lattice inputs are essential for precision tests of the Standard Model and searches for new physics in flavor observables reported by experiments such as LHCb and Belle II. Finite-temperature lattice studies of QCD thermodynamics map the crossover from hadronic matter to the quark–gluon plasma, informing heavy-ion programs at RHIC and CERN's ALICE experiment. Studies of nucleon structure, parton distribution functions via the quasi-PDF approach, and matrix elements for beyond-the-Standard-Model operators extend lattice impact into nuclear physics and astrophysics.
Controlled uncertainties distinguish modern lattice results: statistical errors from finite Monte Carlo samples, and systematic errors from finite lattice spacing, finite volume, unphysical quark masses, renormalization, and excited-state contamination. Ensembles at multiple lattice spacings and volumes enable continuum and infinite-volume extrapolations; partially quenched and fully dynamical simulations address sea-quark effects. Blind analysis practices and global averaging by groups such as the Flavor Lattice Averaging Group (FLAG) provide community assessments of systematic reliability.
Recent progress includes simulations at the physical pion mass, applications of multigrid solvers and machine learning accelerations, and computations of increasingly challenging quantities such as long-distance contributions to rare decays and hadronic light-by-light scattering for the muon g−2 anomaly. Efforts continue to reduce systematic errors, incorporate isospin-breaking and electromagnetic effects, and compute real-time properties via analytic continuation and novel approaches. Future directions rely on exascale computing, collaborations among national laboratories and universities, and synergies with experimental programs at CERN, J-PARC, and Jefferson Lab to further constrain the Standard Model and explore non-perturbative QCD phenomena.
Category:Quantum chromodynamics Category:Computational physics Category:Lattice gauge theory