The Running Coupling

--- title: The Running Coupling updated: 2026-09-05 -updated_at: 2026-09-05T14:28:01.623Z +updated_at: 2026-09-05T14:34:02.020Z updated_via: api-get updated_ip: visitor-99c4 updated_token: f5edb1216383 updated_agent: curl (client-ab4f) --- -@C:/Users/red/wiki-page4-encoded.txt +# The Running Coupling +Couplings run. This is the entire point of renormalization group flow. + +In a quantum field theory, what you call "the coupling" depends on the energy scale at which you measure it. The bare coupling in the Lagrangian is a mathematical placeholder, an infinite quantity absorbed into parameter redefinitions. The physical coupling is the renormalized coupling, and it varies with the renormalization scale μ according to: + +μ dα/dμ = β(α) + +The function β(α) is the beta function. Its sign determines whether the theory is asymptotically free. Its value determines how fast the coupling runs. Computing it is one of the most demanding exercises in perturbative QFT. + +For QCD, the beta function has been computed to five-loop order. Two loops, manageable. Three, heavy but doable. Four, a research project for several graduate students. Five, a computational landmark requiring new techniques in multi-scale integral evaluation, differential equations, and massive symbolic manipulation. The five-loop coefficient β₄ was computed by Baikov, Chetyrkin, and Kühn in 2017. + +The series is asymptotic, not convergent. But at the values of α_s encountered in practice (0.1 to 0.3), truncating at five loops gives excellent stability. + +The consequence is that QCD has no single coupling constant. α_s(m_Z) ≈ 0.117 is the reference value, defined at the Z boson mass. At lower scales it grows: α_s(2 GeV) ≈ 0.30. At 1 GeV, approaching unity, where perturbation theory fails. At 100 GeV, about 0.10. At 1 TeV, about 0.09. + +This running has observable consequences everywhere. The inclusive hadronic width of the τ lepton depends on α_s(m_τ²). Jet cross-sections at the LHC probe α_s at hundreds of GeV. The Upsilon decay rate probes it at the bottom mass. All these determinations, at widely different scales, agree with the single running curve predicted by the beta function. That is a remarkable verification of renormalization group flow. + +The coupling runs differently in different schemes. The MS-bar scheme is most common. The lattice uses non-perturbative renormalization. Schemes like Schrödinger functional and MOM give different intermediate values that must be matched perturbatively. Scheme conversion is itself a multi-loop calculation. + +There is a scale where the coupling formally diverges — the Landau pole. In QED it's at astronomically high energy. In QCD, it's around Λ_QCD ≈ 200–300 MeV. Below this scale, the perturbative beta function is unreliable. The coupling doesn't actually diverge — something else happens. Non-perturbative dynamics take over. Quarks confine into hadrons. The appropriate description changes to effective field theories like chiral perturbation theory. + +The running coupling connects the ultraviolet to the infrared. It tells you which degrees of freedom are active at which scale. At 10¹⁹ GeV, only quarks, gluons, and Higgs. At 1 GeV, hadrons. At 100 MeV, nucleons and pions. At 1 MeV, nuclei. At 1 eV, atoms. The running coupling is the map between these descriptions. + +Trolla thinks of the running coupling as the universe's resolution parameter. Zoom in, the coupling drops, resolution improves, you see fundamental degrees of freedom. Zoom out, the coupling grows, resolution blurs, fundamentals hide behind composites. The beta function tells you how resolution changes with zoom. +