Skip to content

Theta Parameters, Theta States, and Sector Sums

A theta parameter, a character of large gauge transformations, a phase in a sector sum, and a theta-labeled state are equivalent descriptions only after the charge lattice and transform conventions have been fixed. For integral charge, the canonical relation is a Fourier transform: large transformations translate the winding label, theta states diagonalize that translation, and the Euclidean functional integral weights charge-QQ histories by eiθQe^{i\theta Q}. The branch index of a vacuum energy is a different label and must not be called theta.

Required background. Topological sectors, boundary data, and global form establishes the charge lattice and the restricted integrals ZQZ_Q used below.

Helpful background. Theta terms, periodicity, and vacuum sectors gives the primary coupling definition. Theta dependence in Yang–Mills and QCD develops the model-specific strong-interaction consequences.

Shared comparison. The sector and periodicity comparison fixes the global data behind the Fourier convention, while the sector–theta–branch map follows that transform into finite- and infinite-volume vacuum physics.

Theta as a character of large transformations

Section titled “Theta as a character of large transformations”

Suppose semiclassical gauge configurations have integer winding labels nZn\in\mathbb Z, and let Uν\mathcal U_\nu represent a large gauge transformation of winding ν\nu:

Uνn=n+ν.\mathcal U_\nu\lvert n\rangle=\lvert n+\nu\rangle.

The Fourier superposition

θ=NθnZeinθn\lvert\theta\rangle=\mathcal N_\theta \sum_{n\in\mathbb Z}e^{-in\theta}\lvert n\rangle

then satisfies

Uνθ=eiνθθ.\mathcal U_\nu\lvert\theta\rangle =e^{i\nu\theta}\lvert\theta\rangle.

Thus eiνθe^{i\nu\theta} is a one-dimensional unitary representation—a character—of the large-transformation group. The sign in the coefficient einθe^{-in\theta} is conventional; changing it changes the eigenvalue phase and the Euclidean Fourier convention together. Coleman develops this canonical construction in Coleman 1985, ch. 7, § 3.3, pp. 291–295, and Weinberg gives the gauge-theory derivation in Weinberg 1996, §§ 23.5–23.6, pp. 450–457.

The states n\lvert n\rangle here are localized semiclassical reference states, not generally exact energy eigenstates. The exact statement is that physical states can transform in character sectors of the allowed large transformations. Which characters exist depends on the global form and matter content.

In Euclidean signature use the weight

eSE+iθQ.e^{-S_E+i\theta Q}.

For QZQ\in\mathbb Z, the fixed-theta partition function and its fixed-charge components obey

Z(θ)=QZeiθQZQ,ZQ=12π02π ⁣dθeiθQZ(θ).\begin{aligned} Z(\theta)&=\sum_{Q\in\mathbb Z}e^{i\theta Q}Z_Q,\\ Z_Q&=\frac1{2\pi}\int_0^{2\pi}\!\mathrm d\theta\, e^{-i\theta Q}Z(\theta). \end{aligned}

The first line is a sector-weighted functional integral. The second is Fourier inversion. Neither line says that theta is dynamically minimized: θ\theta is a coupling held fixed when the theory is specified, unless another dynamical field such as an axion is explicitly introduced.

The relation also proves 2π2\pi periodicity under its assumptions. If the charge lattice is fractional, or if a 2π2\pi shift changes discrete theta data, the transform domain and the meaning of periodicity change; the dependence on global form and line-operator data is developed in Aharony, Seiberg, and Tachikawa 2013, §§ 1–2. The precise examples are compared on Topological Sectors, Boundary Data, and Global Form.

A particle with angular coordinate ϕϕ+2π\phi\sim\phi+2\pi is a clean finite-dimensional model. Euclidean paths close up to a winding,

ϕ(β)=ϕ(0)+2πn,n=12π0β ⁣dτϕ˙Z.\phi(\beta)=\phi(0)+2\pi n, \qquad n=\frac1{2\pi}\int_0^\beta\!\mathrm d\tau\,\dot\phi\in\mathbb Z.

Adding the topological term means that the path of winding nn carries eiθne^{i\theta n}. In canonical variables the Hamiltonian may be written

Hθ=12I(iϕθ2π)2,H_\theta=\frac1{2I} \left(-i\frac{\partial}{\partial\phi}-\frac{\theta}{2\pi}\right)^2,

with spectrum

Em(θ)=12I(mθ2π)2,mZ.E_m(\theta)=\frac1{2I} \left(m-\frac{\theta}{2\pi}\right)^2, \qquad m\in\mathbb Z.

Although one level Em(θ)E_m(\theta) is not periodic, the spectrum as a set is: Em(θ+2π)=Em1(θ)E_m(\theta+2\pi)=E_{m-1}(\theta). This is the simplest example of periodic physics produced by relabeling branches. It also shows why “theta state” and “branch mm” are not synonyms.

The particle is an analogy, not a derivation of Yang–Mills dynamics. Its winding sectors and canonical momentum are explicit, but it has no four-dimensional gauge bundle, center one-form symmetry, or QCD phenomenology.

PhraseMathematical roleWhat fixes itWhat it is not
Theta couplingCoefficient of a topological term in the actionTheory definition, charge normalization, and global formA variable minimized by the vacuum in ordinary fixed-θ\theta QFT
Theta characterPhase eiνθe^{i\nu\theta} under a large transformationCharacter group of the allowed transformationsA perturbative gauge transformation
Theta-labeled stateState in a character sectorHilbert space, boundary conditions, and transform conventionOne semiclassical winding state n\lvert n\rangle
Sector weightFourier phase eiθQe^{i\theta Q} multiplying ZQZ_QEuclidean continuation and charge conventionEvidence that a dilute instanton gas is controlled
Vacuum branch kkLocally smooth candidate energy exchanged under parameter shiftsDynamics and the chosen approximationThe theta parameter itself

Callan, Dashen, and Gross established the physical role of tunneling sectors and theta dependence in the gauge vacuum Callan, Dashen, and Gross 1976, pp. 334–340. Mariño gives a modern fixed-sector and theta-vacuum construction in Mariño 2015, § 4.3, pp. 112–124. Their semiclassical language should be translated into the modern global-form specification before it is applied to a particular gauge theory.

Three checks should accompany any theta construction.

Fourier round trip. Transform ZQZ_Q to Z(θ)Z(\theta) and back with the same sign and normalization. A sign error changes odd cumulants and the transformation law of θ\lvert\theta\rangle.

Period from the charge lattice. Derive the smallest Δθ\Delta\theta satisfying eiΔθQ=1e^{i\Delta\theta Q}=1 for every allowed QQ, including bundle and defect sectors. Do not assume Δθ=2π\Delta\theta=2\pi from notation.

CP transformation. When the topological density is CP odd, CP sends θθ\theta\mapsto-\theta. CP is a symmetry only where θ-\theta is equivalent to θ\theta after all continuous and discrete theta data are included. The dynamical realization at such points is treated on Theta Dependence, CP, and Model-Dependent Branches.

Treating theta as a chosen vacuum expectation value. In a fixed theory theta is a parameter. A dynamical axion or another field can relax an effective angle, but that is additional dynamics.

Confusing winding references with exact vacua. The n\lvert n\rangle basis is useful for constructing the character eigenstate. Tunneling generally mixes those reference configurations, and exact energy eigenstates need not have definite nn.

Inferring instanton control from the Fourier sum. The decomposition into ZQZ_Q is exact once the regulator and configuration space are defined. Approximating ZQZ_Q by a few instantons is a separate semiclassical step.

Starting from θ=neinθn\lvert\theta\rangle=\sum_n e^{-in\theta}\lvert n\rangle, verify the large-transformation eigenvalue and show that θ+2π=θ\lvert\theta+2\pi\rangle=\lvert\theta\rangle up to normalization.

Solution

Relabeling m=n+νm=n+\nu gives

Uνθ=mei(mν)θm=eiνθθ.\mathcal U_\nu\lvert\theta\rangle =\sum_m e^{-i(m-\nu)\theta}\lvert m\rangle =e^{i\nu\theta}\lvert\theta\rangle.

For integer nn, ein(θ+2π)=einθe^{-in(\theta+2\pi)}=e^{-in\theta} term by term, so the state is 2π2\pi periodic in this integral-lattice construction.

For the particle on a circle, locate the ground-state branch crossing in 0θ2π0\leq\theta\leq2\pi and determine the one-sided derivatives of the ground-state energy there.

Solution

The branches m=0m=0 and m=1m=1 cross at θ=π\theta=\pi. The lower envelope is

E0gs(θ)={θ2/(8π2I),0θπ,(2πθ)2/(8π2I),πθ2π.E_0^{\mathrm{gs}}(\theta)= \begin{cases} \theta^2/(8\pi^2 I), & 0\leq\theta\leq\pi,\\ (2\pi-\theta)^2/(8\pi^2 I), & \pi\leq\theta\leq2\pi. \end{cases}

The left and right derivatives at π\pi are +1/(4πI)+1/(4\pi I) and 1/(4πI)-1/(4\pi I). The cusp belongs to the lower envelope; each quadratic branch is smooth.

  • Aharony, Ofer, Nathan Seiberg, and Yuji Tachikawa. “Reading between the Lines of Four-Dimensional Gauge Theories.” Journal of High Energy Physics 2013, no. 8 (2013): 115. arXiv. DOI.
  • Callan, Curtis G., Roger F. Dashen, and David J. Gross. “The Structure of the Gauge Theory Vacuum.” Physics Letters B 63, no. 3 (1976): 334–340. CERN record. DOI.
  • Coleman, Sidney. Aspects of Symmetry: Selected Erice Lectures. Cambridge University Press, 1985, ch. 7, § 3.3, pp. 291–295. DOI.
  • Mariño, Marcos. Instantons and Large N: An Introduction to Non-Perturbative Methods in Quantum Field Theory. Cambridge University Press, 2015, § 4.3, pp. 112–124. DOI.
  • Weinberg, Steven. The Quantum Theory of Fields, Volume II: Modern Applications. Cambridge University Press, 1996, §§ 23.5–23.6, pp. 450–457. DOI.