Line Operators, Screening, and Generalized-Symmetry Diagnostics

Line operators diagnose a phase only after the theory tells us which electric, magnetic, and dyonic charges are genuine and which dynamical particles can screen them. String breaking is the physical manifestation of this arithmetic: if matter can neutralize an external charge, the corresponding line cannot retain an asymptotic area law.

Required background. Confinement definitions and their non-equivalence separates line behavior from flux tubes, spectra, and mass gaps.

Helpful background. Topological sectors and global form supplies the bundle data, while genuine line spectra and discrete theta angles gives the full theory specification.

Shared comparison. The diagnostic and counterexample map shows where line criteria cease to imply other notions of confinement; the claim and evidence comparison keeps the symmetry statement distinct from a proposed mechanism.

Genuine lines and mutually local charges

For a representation $R$ of the gauge group, the Wilson line on a closed curve $C$ is

$$W_R(C)=\operatorname{tr}_R\,\mathrm P \exp\!\left(ig\oint_C A_\mu\,\mathrm d x^\mu\right).$$

Its role as a nonlocal confinement probe originates in Wilson 1974, §§IV–V, pp. 2450–2456.

An ’t Hooft line is instead a disorder insertion: on a small two-sphere linking the line, the gauge field has a prescribed magnetic flux. A dyonic line carries both charges. If electric and magnetic weights are denoted $(e,m)$, two lines can both be genuine only when their Dirac pairing obeys the appropriate integrality condition,

$$\langle(e,m),(e',m')\rangle =e(m')-e'(m)\in\mathbb Z.$$

The Lie algebra does not select a unique maximal mutually local set. The global form of the gauge group and any discrete theta parameter do. For pure $SU(N)$ theory, all Wilson representations are allowed and the fundamental Wilson line is genuine; the basic center-charged magnetic line that is genuine in a $PSU(N)$-type spectrum requires extra surface data in the $SU(N)$ theory, although coroot-lattice magnetic lines can be genuine there. For $PSU(N)$, the basic fundamental Wilson line is not an operator of the theory, whereas suitable magnetic or dyonic lines are genuine. This exchange is developed explicitly in Gaiotto, Kapustin, Seiberg, and Willett 2015, §4.2, pp. 17–19.

Consequently, “the Wilson loop has an area law” is incomplete. It must identify a genuine representation of a declared global form. Testing a non-genuine line can still be useful after its attached surface is specified, but it is not a standalone order parameter.

Screening is charge-lattice arithmetic

Let $\Lambda_{\mathrm{probe}}$ be the lattice of allowed external electric charges and $\Lambda_{\mathrm{dyn}}$ the sublattice generated by dynamical matter. Long-distance electric superselection sectors are classes

$$\mathcal Q_e=\Lambda_{\mathrm{probe}}/\Lambda_{\mathrm{dyn}}.$$

A line whose charge is trivial in $\mathcal Q_e$ can end on a dynamical excitation. Its flux tube may be long-lived when the matter is heavy, but at sufficiently large separation the energy gained by shortening the tube exceeds the pair-creation cost. The string then breaks.

For $SU(N)$, the useful coarse charge is $N$-ality. Adjoint gluons have zero $N$-ality and screen representations down to the same center charge, so pure Yang–Mills retains $\mathbb Z_N$ sectors. A dynamical fundamental has $N$-ality one and generates the whole center charge group, making $\mathcal Q_e$ trivial. More generally, if matter has $N$-alities $k_i$, the surviving electric one-form symmetry is

$$\mathbb Z_{\gcd(N,k_1,k_2,\ldots)}.$$

This group is the subgroup of the center acting trivially on every dynamical field. Gaiotto and collaborators formulate the same result as electric charges modulo matter charges Gaiotto et al. 2015, §4.2, pp. 18–19.

Example: SU(6) with two-index matter

A two-index antisymmetric field has $N$-ality $2$. With only adjoint and this matter, the unbroken electric one-form symmetry is $\mathbb Z_{\gcd(6,2)}=\mathbb Z_2$. A source of $N$-ality $2$ can be screened, while a source of odd $N$-ality cannot. Thus some nontrivial Wilson loops may have perimeter behavior while others retain area behavior. A binary “confining or not” label discards this mixed structure.

One-form symmetry diagnoses, but does not explain

For an exact electric one-form symmetry $\Gamma$, a genuine line charged under $\Gamma$ is its order parameter. After subtracting local perimeter counterterms:

• an area law for lines charged under an unbroken subgroup means that subgroup is unbroken;
• perimeter or Coulomb behavior for a charged line signals spontaneous breaking, with low-dimensional infrared qualifications;
• partial breaking produces charge-dependent mixed behavior.

This is a clean symmetry classification Gaiotto et al. 2015, §5, pp. 27–31. It does not identify monopoles, vortices, or any other microscopic mechanism. An anomaly can further rule out a trivially gapped symmetric infrared phase, but it likewise leaves several compatible realizations.

At finite temperature, compact Euclidean time splits the four-dimensional electric one-form symmetry into a lower-dimensional zero-form symmetry acting on temporal Polyakov loops and a one-form symmetry acting on spatial loops. A vanishing Polyakov loop in pure Yang–Mills constrains the free energy of an isolated temporal probe; a spatial area law concerns a spatial string tension. Do not use one as a synonym for the other.

String breaking as a spectral crossing

Place a static source and antisource a distance $R$ apart. In a theory with heavy dynamical fundamentals, two schematic states share the same gauge-invariant quantum numbers:

$$E_{\mathrm{string}}(R)\simeq 2\delta m+\sigma R, \qquad E_{\mathrm{broken}}(R)\simeq 2M_{Qq}.$$

Mixing turns their crossing into an avoided crossing. The exact ground-state energy approaches the screened branch, while a Wilson loop built from a thin connecting transporter can have poor overlap with that branch and continue to display an apparently linear excited energy at accessible Euclidean time. This is why string breaking tests both the operator basis and the order of limits, not merely the plotted shape of a potential.

Common pitfalls

Using the center of the Lie algebra as the symmetry. The center symmetry depends on the global group and the matter representations. $SU(N)$ and $PSU(N)$ theories have the same algebra but different genuine lines.

Calling every perimeter law screening. A perimeter law can arise from charge screening or from spontaneous breaking of a one-form symmetry. Coulomb behavior also requires its power-law term to be distinguished from removable perimeter renormalization.

Treating a heavy string as stable. A large matter mass can push the breaking distance far out, but finite mass still permits asymptotic screening when charges match.

Exercises

1. Residual symmetry. Find the surviving electric one-form symmetry of $SU(8)$ gauge theory with dynamical matter of $N$-alities $2$ and $4$. Which center charges can remain unscreened?

Solution

The subgroup acting trivially on both matter representations is $\mathbb Z_{\gcd(8,2,4)}=\mathbb Z_2$. Matter generates the even subgroup of $\mathbb Z_8$, so charges are classified by parity. Odd $N$-ality remains nontrivial; even $N$-ality can be screened.

2. Global-form check. Why is comparing a “fundamental Wilson area law” between $SU(N)$ and $PSU(N)$ ill posed?

Solution

The fundamental Wilson line is a genuine operator in the $SU(N)$ theory but not in the $PSU(N)$ theory. A valid comparison must first choose genuine lines in each theory and match their charge lattices, possibly including a discrete theta choice. Otherwise the two quantities do not belong to the same operator class.

References

• Gaiotto, Davide, Anton Kapustin, Nathan Seiberg, and Brian Willett. “Generalized Global Symmetries.” Journal of High Energy Physics 2015, no. 2 (2015): 172. DOI. Open PDF.
• Wilson, Kenneth G. “Confinement of Quarks.” Physical Review D 10 (1974): 2445–2459. DOI.