Gauge Instantons, Fermion Zero Modes, and Condensates

A gauge instanton determines an exact chiral term only when its collective-coordinate integral is controlled and precisely two fermion zero modes remain unlifted for a superpotential insertion. In pure four-dimensional SYM a unit instanton has too many gaugino zero modes and an infrared-sensitive size integral; in Higgsed SQCD with $N_f=N_c-1$ , Yukawa lifting and the Higgs scale leave exactly two modes and make the ADS calculation reliable. Compactification supplies a third, distinct regulator by fractionalizing the instanton into monopole events.

Required background. Instanton zero modes and selection rules supplies the index theorem and supersymmetric measure. Gaugino condensation supplies the chiral observable whose normalization is being tested.

Helpful background. Instanton measures, zero modes, and determinants develops the bosonic collective coordinates and determinant factors in general gauge theory.

Euclidean saddle and index convention

Continue to oriented Euclidean $\mathbb R^4$ and take a self-dual $SU(N_c)$ instanton of topological charge $k=+1$ . Chiral and antichiral Weyl fields are independent Euclidean variables; one does not impose the Lorentzian Majorana condition during the saddle calculation. With $T(\mathbf{N_c})=1/2$ , the chiral Dirac index is

n_{\rm zero}(R)=2T(R)k.

Consequently,

n_\lambda=2N_c, \qquad n_{Q}+n_{\widetilde Q}=2N_f

for a unit instanton in SQCD. Which Euclidean chirality carries these zero modes reverses for an anti-instanton or the opposite orientation convention, but the number and the resulting selection rule do not.

A local Wilsonian superpotential contribution appears as $\int d^4x\,d^2\theta\,W_{\rm np}$ . The two $d^2\theta$ modes are the universal supersymmetry zero modes. Any additional fermion zero modes must be absorbed by operator insertions or lifted by interactions. This criterion is necessary; convergence of every bosonic collective-coordinate integral is also required.

Pure SYM: the correlator is allowed, a direct superpotential is not

In pure $SU(N_c)$ SYM, $2N_c$ gaugino zero modes forbid a one-instanton contribution to a local superpotential for an otherwise empty low-energy theory. They can instead saturate

\left\langle S(x_1)S(x_2)\cdots S(x_{N_c}) \right\rangle_{k=1},

where each $S\propto\lambda\lambda$ absorbs two modes. Supersymmetric Ward identities make separated chiral correlators position-independent, subject to contact terms. If one first selects a single gapped vacuum and then separates all $x_i$ , cluster decomposition suggests

\langle S^{N_c}\rangle=\langle S\rangle^{N_c} =\Lambda^{3N_c}.

This correctly anticipates the $N_c$ roots. It does not by itself give a controlled normalization. The four-dimensional instanton has a size modulus $\rho$ , and the semiclassical measure samples $\rho\Lambda\gtrsim1$ , where the running coupling is strong. A finite-volume path integral can also average over all discrete vacua; cluster decomposition applies only after a vacuum-selecting source and the infinite-volume limit are ordered correctly.

The historical mismatch between strong-coupling-instanton and weak-coupling/compactified normalizations is therefore a diagnostic, not a paradox to hide. The controlled small-circle result of Davies et al. 1999, §§ II–V agrees with holomorphic decoupling and shows why the unregulated $\mathbb R^4$ size integral is not the normalization anchor.

Higgsed SQCD: how $2N_c+2N_f$ modes become two

Now take $SU(N_c)$ SQCD with

N_f=N_c-1, \qquad W_{\rm tree}=0,

at a generic D-flat point where quark expectation values of scale $v$ completely break the gauge group. If $v\gg\Lambda$ , the coupling at the instanton scale is weak. The scalar profile adds a positive action of order $\rho^2v^2$ , so the size integral is suppressed for $\rho\gtrsim v^{-1}$ .

Before lifting, there are $2N_c$ gaugino and $2(N_c-1)$ matter zero modes. Gauge Yukawa interactions schematically contain

\sqrt2g\,q^\dagger\lambda\psi_Q +\sqrt2g\,\widetilde q^\dagger\lambda\psi_{\widetilde Q} +\text{h.c.}

Each insertion pairs one matter zero mode with one gaugino zero mode and a scalar expectation value. Using all $2(N_c-1)$ matter modes also lifts $2(N_c-1)$ gaugino modes, leaving

2N_c-2(N_c-1)=2

unlifted gaugino modes. The instanton can therefore generate a superpotential.

Flavor symmetry, R-charge, dimension, and the semiclassical determinant give

W_{k=1}=\frac{\Lambda^{2N_c+1}}{\det M}, \qquad M_i{}^{\tilde\jmath}=\widetilde Q_a{}^{\tilde\jmath}Q^a{}_i.

Here $b_0=3N_c-(N_c-1)=2N_c+1$ . The direct one-instanton calculation fixes the coefficient to one in this composite and scale convention. The original analysis is Affleck, Dine, and Seiberg 1984, pp. 493–534; an explicit $SU(2)$ , one-flavor calculation is given in Shifman 2022, § 10.20, pp. 542–549.

Holomorphic continuation is doing real work

The instanton calculation is controlled only on the large- $v$ Higgs branch. The resulting Wilsonian F-term is holomorphic in $M$ and $\Lambda$ and has a unique symmetry-allowed form. It can therefore be continued to smaller $M$ as long as no singularity or extra branch invalidates the chosen low-energy variables.

This does not mean the small- $M$ region becomes semiclassical. It means a protected holomorphic function determined in one open region extends analytically. The Kähler potential, instanton-size distribution, and ordinary scattering observables do not inherit this exact continuation.

For $N_f<N_c-1$ , a single four-dimensional instanton leaves more than two zero modes in the corresponding Higgs background because an $SU(N_c-N_f)$ gauge subgroup remains. The general ADS superpotential is still exact, but its coefficient is transported by holomorphic decoupling or derived from the unbroken gauge sector—not by pretending the same one-instanton saddle applies directly.

Compactification changes the saddle, not the selection rule

On $\mathbb R^3\times S^1$ with periodic gauginos and center-symmetric holonomy, an $SU(N)$ instanton splits into $N$ fundamental monopole instantons, including one Kaluza–Klein monopole. Each has two gaugino zero modes and can contribute directly to a three-dimensional superpotential. Their product carries unit four-dimensional topological charge and the full $2N$ zero modes.

This is a controlled calculation when $NL\Lambda\ll1$ . It is not the same saddle as an isolated BPST instanton on $\mathbb R^4$ . Agreement of its protected condensate with decoupling supplies independent evidence; it does not retroactively make the strong-coupling $\rho$ integral weakly coupled.

A stop rule for instanton claims

Before accepting an instanton-generated condensate or F-term, record:

topological sector and orientation;
every bosonic and fermionic collective coordinate;
the interaction that lifts each non-universal zero mode;
the infrared regulator of the size or separation integral;
the regime in which the running coupling is small;
the composite, scale, and measure normalization;
the order of infinite-volume, source-removal, Higgs, radius, and mass limits; and
the holomorphic step, if any, that extends the result beyond the saddle regime.

If the bosonic integral reaches strong coupling without a regulator, zero-mode saturation alone is not a controlled calculation.

Exercises

In $SU(4)$ SQCD with $N_f=3$ , count the zero modes before and after Yukawa lifting.

Solution

The instanton has $2N_c=8$ gaugino zero modes and $2N_f=6$ matter zero modes. Six Yukawa insertions pair all matter modes with six gaugino modes, leaving two universal gaugino modes. A superpotential contribution is therefore allowed at a generic completely Higgsed point.

Explain why the factor $e^{-c\rho^2v^2}$ improves semiclassical control.