Fractional-Charge Instantons in Gauge Theories

• Fractional-charge instantons are finite-action, semiclassical topological configurations with rational charge induced by twisted boundary conditions and nontrivial holonomy.
• They appear in diverse settings such as Yang–Mills on tori, sigma models, and spin chains, influencing confinement, mass gaps, and vacuum structure.
• Analytical and lattice studies connect these fractional instantons to semiclassical resurgence, duality, and large-N dynamics, offering insights into nonperturbative physics.

Fractional-charge instantons are finite-action, or more generally semiclassically relevant, topological configurations whose charge is not an integer but a rational fraction fixed by global data such as twisted boundary conditions, nontrivial holonomy, or compactification. In four-dimensional $SU(N)$ Yang–Mills theory on tori with ’t Hooft twists, the canonical fractional values are $Q=k/N$, with minimal-action sectors often realizing $Q=1/N$; in $SU(2)$ twisted compactifications the basic smooth self-dual lump carries $Q=1/2$ (González-Arroyo, 2019, Bergner et al., 13 Feb 2025). Related objects also appear in lower-dimensional sigma models, principal chiral models, Schwinger models, and $SU(n)$ spin chains, where they govern mass generation, vacuum structure, and resurgence-inspired semiclassics (Wamer et al., 2020, Misumi et al., 2019, Nitta, 2015).

1. Topological origin of fractionalization

On simply connected manifolds or compact spaces with periodic boundary conditions, the Yang–Mills topological charge

$$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$

is integer-valued. Fractionalization begins when the gauge bundle is nontrivial, typically through ’t Hooft twisted boundary conditions on a torus. In $SU(N)$, the transition functions $\Omega_\alpha(x)$ obey cocycle relations only up to a center phase $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$, with $Q=k/N$0, and the twist data can be repackaged into electric and magnetic flux vectors $Q=k/N$1 (González-Arroyo, 2019).

In that setting, the instanton number is shifted from integrality. A standard formula is

$Q=k/N$2

or equivalently $Q=k/N$3 with $Q=k/N$4. For non-orthogonal twists this allows $Q=k/N$5 in the minimal sector, and the BPS bound then gives

$Q=k/N$6

For $Q=k/N$7 with suitable twists in two orthogonal planes, the basic self-dual solution instead has

$Q=k/N$8

and is smooth rather than singular (González-Arroyo, 2019, Bergner et al., 13 Feb 2025).

This mechanism has close analogues outside four-dimensional Yang–Mills. In compactified sigma models and principal chiral models, twisted boundary conditions fractionalize the relevant homotopy classes so that constituents carry $Q=k/N$9 or $Q=1/N$0 of the ordinary instanton number; in the charge-$Q=1/N$1, $Q=1/N$2-flavor Schwinger model, the relevant semiclassical saddles on $Q=1/N$3 carry $Q=1/N$4 or $Q=1/N$5 and generate fractional $Q=1/N$6-dependence $Q=1/N$7 or $Q=1/N$8 (Nitta, 2014, Nitta, 2015, Misumi et al., 2019).

2. Geometric realizations in Yang–Mills theory

A central realization is $Q=1/N$9 Yang–Mills on $SU(2)$0 with twist. Constant-field-strength solutions, generalizing ’t Hooft’s original constructions, provide explicit self-dual backgrounds in the $SU(2)$1 sector for special torus metrics; away from those special metrics one can construct non-constant self-dual solutions by expanding the gauge field in a deformation parameter of the metric. In that construction the gauge potential is decomposed into block-diagonal and off-diagonal sectors, and the leading nontrivial deformation is expressed through theta-function solutions of the twisted abelianized equations (González-Arroyo, 2019).

A second realization, emphasized in recent lattice work, is $SU(2)$2 Yang–Mills on $SU(2)$3, implemented numerically as a four-torus with two short and two long directions. The small two-torus carries twisted boundary conditions that preserve center symmetry, enforce vanishing expectation value of short Polyakov loops, and allow smooth $SU(2)$4 objects. These fractional instantons are localized in the large plane, essentially fill the small torus, and behave as $SU(2)$5 center vortices at long distance, with large Wilson loops approaching $SU(2)$6 when they enclose the lump (Bergner et al., 13 Feb 2025).

A third realization is $SU(2)$7 Yang–Mills on $SU(2)$8 with a specially chosen asymmetric torus and ’t Hooft flux. For $SU(2)$9 and magnetic flux $Q=1/2$0 from the Fibonacci sequence, one obtains self-dual solutions with

$Q=1/2$1

on $Q=1/2$2, $Q=1/2$3, $Q=1/2$4. In the Hamiltonian limit these are vacuum-to-vacuum tunneling events between inequivalent pure-gauge configurations whose $Q=1/2$5 Polyakov loop phases differ by a center element (Golán et al., 2022, Golan et al., 2022).

Related weak-coupling evidence exists for $Q=1/2$6 on $Q=1/2$7 with $Q=1/2$8 twists. There the total topological charge on the compact lattice remains integer, but the local charge density splits into peaks whose integrated charges are close to integer multiples of $Q=1/2$9, and those peaks correlate with rotations of the Polyakov-loop phase between degenerate $SU(n)$0-broken vacua (Itou, 2018).

3. Semiclassical structure, lattice diagnostics, and confinement

The $SU(n)$1 program gives the clearest recent identification of fractional-charge instantons as semiclassical constituents of confinement. Numerically, $SU(n)$2 gauge fields are generated with Wilson action on lattices $SU(n)$3, $SU(n)$4, then smoothed with gradient flow. The topological density is integrated over the short torus,

$SU(n)$5

and local extrema of $SU(n)$6 are fitted to a BPST-like profile

$SU(n)$7

The resulting histograms show peaks at $SU(n)$8, identifying individual fractional instantons and anti-fractional instantons (Bergner et al., 13 Feb 2025).

Their density is summarized by the diluteness

$SU(n)$9

and semiclassics predicts

$$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$0

The simulations confirm this form over a range of $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$1, including the expected $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$2-dependence of $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$3. This supports the interpretation of the $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$4 lumps as genuine semiclassical saddles rather than filtering artefacts (Bergner et al., 13 Feb 2025).

Confinement is probed through Wilson loops in the large plane and Creutz ratios. In the thin Abelian vortex approximation, a dilute Poisson gas of center-vortex-like objects with areal density $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$5 yields

$$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$6

The measured string tension $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$7 tracks the fractional-instanton density $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$8, with $$Q=\frac{1}{32\pi^2}\int d^4x\,F_{\mu\nu}^a\tilde F_{\mu\nu}^a$$9, and the ratio remains approximately constant along gradient flow even though both quantities decrease. At small $SU(N)$0 this is a dilute two-dimensional gas; as $SU(N)$1 increases toward $SU(N)$2–$SU(N)$3 fm, the density rises, the nearest-neighbor distance falls toward a constant physical scale of about $SU(N)$4 fm, and the system crosses over to a fractional-instanton liquid picture of the Yang–Mills vacuum (Bergner et al., 13 Feb 2025).

This suggests that, in this geometry, confinement is mediated by vortex-like fractional instantons whose size is controlled by the small torus in the dilute regime and by inter-object spacing in the dense regime. The same work presents this as a roadmap from controlled semiclassics to the infinite-volume $SU(N)$5 vacuum (Bergner et al., 13 Feb 2025).

4. Large-$SU(N)$6, duality, and constituent pictures

The large-$SU(N)$7 regime makes fractional-charge instantons structurally distinctive because $SU(N)$8 implies an action $SU(N)$9 at fixed ’t Hooft coupling $\Omega_\alpha(x)$0, so their semiclassical weight remains $\Omega_\alpha(x)$1-independent. In the Fibonacci-twist construction on $\Omega_\alpha(x)$2, rescaled action-density profiles and Polyakov-loop observables approach universal large-$\Omega_\alpha(x)$3 forms, while the asymmetric small torus remains consistent with twisted volume-independence arguments (Golán et al., 2022, Golan et al., 2022).

The Nahm transform gives a precise duality statement for multi-fractional instantons on $\Omega_\alpha(x)$4. Embedding the twisted $\Omega_\alpha(x)$5 bundle into a $\Omega_\alpha(x)$6 bundle with integer-quantized $\Omega_\alpha(x)$7 fluxes $\Omega_\alpha(x)$8, an $\Omega_\alpha(x)$9 fractional instanton of charge

$z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$0

is mapped to an $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$1 fractional instanton of charge

$z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$2

For constant-field-strength backgrounds both the original and dual configurations are self-dual when the torus periods satisfy the appropriate tuning conditions, and a detuning parameter $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$3 is mapped to $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$4 on the dual torus (Anber et al., 2024).

A different constituent interpretation arises in five-dimensional maximally supersymmetric $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$5 Yang–Mills as the circle reduction of the six-dimensional $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$6 theory. There, instanton solitons carry Kaluza–Klein momentum, and the conjectured “instanton partons” have

$z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$7

The paper identifies realizations as intersections of center-valued magnetic flux sheets, as fractional monopole constituents of calorons on $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$8, and as BPS monopole-like objects in the symmetric phase when only adjoint matter is present (Bolognesi et al., 2011). A cautious reading is required here: the same source stresses that smooth localized $z_{\alpha\beta}=e^{2\pi i n_{\alpha\beta}/N}$9 instantons on $Q=k/N$00 are elusive, and the most concrete constructions rely on compactification, center-valued gauge transformations, or singular flux-sheet configurations.

5. Fractional instantons outside four-dimensional Yang–Mills

Fractional-charge instantons are not unique to non-Abelian gauge theory. The following settings realize closely related structures, though the precise meaning of “charge” and the relevant homotopy differ.

Setting	Fractional object	Role
$Q=k/N$01 spin chains	$Q=k/N$02 species with charge vectors $Q=k/N$03	Mass-gap generation except when $Q=k/N$04 (Wamer et al., 2020)
Charge-$Q=k/N$05, $Q=k/N$06-flavor Schwinger model	“Quantum instantons” with $Q=k/N$07 or $Q=k/N$08	Fractional $Q=k/N$09-dependence and $Q=k/N$10-branch vacuum structure (Misumi et al., 2019)
$Q=k/N$11 sigma models on $Q=k/N$12	Half-lumps, half-Skyrmions, vortices or monopoles with twisted core moduli	Neutral bions and fractionalized homotopy on twisted compactifications (Nitta, 2014)
$Q=k/N$13 principal chiral model on $Q=k/N$14	Global vortices with twisted $Q=k/N$15 moduli; $Q=k/N$16 or irrational baryon number	Neutral and charged bions; Yang–Mills correspondence via non-Abelian Josephson junctions (Nitta, 2015)
Sigma models with chemical potential	Complex BPS-like constituents with complex fractional charge	Lefschetz-thimble-compatible saddles and Stokes jumps at $Q=k/N$17 (Bruckmann et al., 2018)

In $Q=k/N$18 spin chains, the fractional instantons are vortex-like defects in the flag-manifold sigma model, and the interference phase $Q=k/N$19 determines whether their fugacity vanishes. This reproduces the generalized Haldane criterion: when $Q=k/N$20 and $Q=k/N$21 are coprime, the leading fractional instantons cancel, whereas for $Q=k/N$22 some sector survives and generates a mass gap (Wamer et al., 2020).

In the Schwinger model, the relevant semiclassical saddles are explicitly distinguished from classical instantons: they are “quantum instantons” or fractons of the holonomy effective quantum mechanics, with action scaling like $Q=k/N$23 rather than $Q=k/N$24. Their charges directly explain chiral condensates proportional to $Q=k/N$25 and vacuum energies with $Q=k/N$26 branches under flavor-twisted boundary conditions (Misumi et al., 2019).

In the principal chiral and $Q=k/N$27 models, fractionalization is geometrically transparent. One first obtains a host soliton from the fixed manifold of the twist, then a daughter soliton on its core moduli space, with the daughter twisted by half or $Q=k/N$28 of a full winding along $Q=k/N$29. The composite carries the fractional instanton number, while neutral or charged bions are composites of such constituents (Nitta, 2014, Nitta, 2015).

6. Conceptual distinctions and open problems

A recurrent misconception is that fractional topological charge is generic on compact space. It is not. In every construction above, fractionalization requires additional global data: twisted boundary conditions, nontrivial bundle topology, or nontrivial holonomy. On periodic untwisted $Q=k/N$30, one recovers ordinary integer instanton sectors (González-Arroyo, 2019).

A second distinction concerns local versus global fractionalization. In the $Q=k/N$31 weak-coupling $Q=k/N$32 lattice study, the total topological charge remains integer, but the local density splits into constituents carrying approximately $Q=k/N$33 and the Polyakov-loop phase rotates across those regions. Fractional charge there is a constituent property of the configuration, not a non-integer total $Q=k/N$34 on the compact lattice (Itou, 2018).

A third distinction is between classical and quantum instantons. In the Schwinger model, the objects generating fractional $Q=k/N$35-dependence are not ordinary classical instantons of the original two-dimensional action but BPS-saturated saddles of the holonomy effective dynamics. In sigma models with chemical potential, the relevant saddles live on complexified field space and their charges can be genuinely complex; the real part retains the familiar topological interpretation, while the imaginary part reflects noncompact directions in the complexified gauge orbit (Misumi et al., 2019, Bruckmann et al., 2018).

Several open problems remain explicit in the recent literature. On $Q=k/N$36, extending the analysis deeper into the dense regime requires identification methods that no longer rely on integrating over the short torus, and it remains necessary to test whether the same liquid picture accounts quantitatively for observables such as topological susceptibility or glueball scales (Bergner et al., 13 Feb 2025). In weak-coupling $Q=k/N$37 lattice studies, continuum and infinite-volume extrapolations are still needed to decide which configuration classes dominate and how bions emerge in the $Q=k/N$38 sector (Itou, 2018). In the Nahm-transform and D-brane picture, the endpoint of tachyon condensation for detuned tori remains unresolved, although the small-$Q=k/N$39 expansion suggests a feasible route (Anber et al., 2024).

Taken together, these results support a unified view: fractional-charge instantons are semiclassical or quasi-semiclassical constituents whose charges are set by global structure rather than local singular behavior. In Yang–Mills theory they interpolate between twisted vacua, act as center-vortex-like objects, and can underpin confinement or large-$Q=k/N$40 tunneling; in lower-dimensional theories they organize mass generation, anomaly matching, and resurgence. The common thread is not a single universal profile, but a universal mechanism: compactification plus global twisting decomposes ordinary topological sectors into fractional constituents.