Gauge Fixing and the Localization Deformation Complex

In a gauge theory, the localizing supercharge commonly squares to a field-dependent gauge transformation as well as rigid bosonic symmetries. A one-loop determinant cannot be read from the ungauge-fixed Hessian, whose gauge directions are null. Supersymmetry must be combined with BRST gauge fixing into one odd differential, and the resulting fluctuation complex—not a list of component operators—determines the determinant grading and zero modes.

Required background. $Q$-cohomology and path-integral deformation supplies the deformation argument. BRST cohomology and physical observables supplies ghosts and gauge-orbit cohomology.

Helpful background. Master equations and BV gauge fixing supplies the extension required for open or reducible gauge algebras.

Combining supersymmetry with BRST

Assume that on covariant physical fields

$$Q^2=\mathcal H+\delta_{\rm gauge}(\Lambda),$$

where $\mathcal H$ collects background isometries, R rotations, and flavor transformations. Introduce the odd ghost $c$ and define a combined differential $\widehat Q$ by

$$\widehat QX=QX+\delta_cX.$$

In a standard graded convention,

$$\widehat Q^2X =Q^2X+\delta_{\widehat Qc+\frac12[c,c]}X.$$

Therefore the ghost transformation begins with

$$\widehat Qc=-\Lambda-\frac12[c,c],$$

and its action on $\Lambda$ is fixed by consistency, so that

$$\widehat Q^2=\mathcal H$$

on physical fields and ghosts. Signs change if the gauge variation or Lie bracket is normalized differently, but the invariant requirement is exact: the field-dependent gauge transformation in $Q^2$ must be absorbed, and the same residual $\mathcal H$ must act throughout the extended complex. This is the construction used in gauge-theory localization; see Pestun 2012, §4.2.

For an ordinary irreducible gauge condition $G(A)=0$, add antighost $\bar c$ and multiplier $B$ with

$$\widehat Q\bar c=B, \qquad \widehat QB=\mathcal H\bar c.$$

Then $\widehat Q^2=\mathcal H$ also on the nonminimal pair. A gauge fermion may be chosen as

$$\Psi_{\rm gf}=\int_M\operatorname{Tr} \left[\bar c\left(G(A)+\frac\xi2B\right)\right],$$

provided $\Psi_{\rm gf}$ is $\mathcal H$ invariant and respects the boundary conditions. The full deformation uses $\widehat Q(V_{\rm loc}+\Psi_{\rm gf})$.

Linearizing at a BPS configuration

Let $\Phi_0$ solve the localization equations. Separate bosonic and fermionic fluctuations into graded bundles $E_0$ and $E_1$. The linearized supersymmetry and gauge-fixing equations define a symbol sequence schematically of the form

$$\Gamma(E_{-1}) \xrightarrow{\ D_{\rm gauge}\ } \Gamma(E_0) \xrightarrow{\ D_{10}\ } \Gamma(E_1) \xrightarrow{\ D_{21}\ } \Gamma(E_2).$$

$E_{-1}$ contains infinitesimal gauge parameters; $D_{10}$ contains the linearized BPS equations together with the gauge condition; and $D_{21}$ records identities among those equations when present. The one-loop problem is controlled by the cohomology of this complex:

$$H^0=\frac{\ker D_{10}}{\operatorname{im}D_{\rm gauge}}, \qquad H^1=\frac{\ker D_{21}}{\operatorname{im}D_{10}}.$$

The first space contains physical tangent zero modes to the BPS moduli space. The second contains obstructions or unpaired fermionic modes. Ghost zero modes encode stabilizers of $\Phi_0$ and must be treated as gauge-group volume or residual group integration, not included in a primed determinant.

Elliptic and transversely elliptic complexes

If the symbol sequence is exact for every nonzero cotangent vector, the complex is elliptic. On a compact manifold its cohomology is finite dimensional and its index is stable under lower-order deformations. When $\widehat Q^2$ generates a torus action, the symbol may be invertible only in directions transverse to the group orbits. The complex is then transversely elliptic: its equivariant index is generally a distribution or a formal character, and an expansion chamber must be specified.

This distinction matters for determinants. Treating a transversely elliptic operator as elliptic can discard infinitely many modes along the orbit or choose an unannounced series expansion. Five-dimensional contact localization gives a concrete transversely elliptic example Källén, Qiu, and Zabzine 2012, §§3–4.

Reducible saddles and boundaries

At a reducible connection the stabilizer subgroup $G_{\Phi_0}$ is nontrivial. Then the Faddeev–Popov operator has zero modes, the local quotient is singular, and simply writing $1/\operatorname{Vol}G$ is incorrect. One must stratify the saddle space, divide by the actual stabilizer, and include any induced measure on its collective coordinates. If gauge transformations themselves have relations, ghosts-for-ghosts or the BV complex are required.

With a boundary, the differential operator and its domain are inseparable. Boundary conditions must make the variational problem well posed, be preserved by $\widehat Q$, and render the complex Fredholm (or supply edge modes restoring it). Absolute and relative boundary conditions, for example, generally lead to different cohomology and determinants even when the bulk symbol is identical.

A determinant-ready checklist

Before diagonalizing modes, verify:

• the full $Q^2$, including its field-dependent gauge parameter;
• the transformations of ghosts, antighosts, multipliers, and auxiliaries;
• $\widehat Q^2=\mathcal H$ on every field;
• the gauge fermion and field-space cycle are $\mathcal H$ invariant;
• the symbol and whether it is elliptic or transversely elliptic;
• stabilizers, reducible strata, and ghost zero modes;
• boundary domains and adjoint boundary conditions;
• orientation and fermion Pfaffian conventions.

Exercises

1. Absorbing the gauge square. Use the displayed graded identity to show that $\widehat Qc=-\Lambda-\frac12[c,c]$ removes $\delta_{\rm gauge}(\Lambda)$ from $\widehat Q^2X$.

Solution

Substitution gives $\widehat Qc+\frac12[c,c]=-\Lambda$. Hence $\widehat Q^2X=\mathcal H X+\delta_\Lambda X-\delta_\Lambda X=\mathcal H X$.

2. Stabilizer zero mode. Let $d_A\epsilon=0$ for a nonzero gauge parameter $\epsilon$ at a saddle. Why should this mode not enter $\det' d_A$?

Solution

$\epsilon$ generates a gauge transformation that leaves the saddle fixed, so it is a stabilizer direction rather than a Gaussian fluctuation. Its zero eigenvalue must be removed from the determinant and replaced by division by, or integration over, the residual stabilizer with the correct measure.

References

• Källén, Johan, Jian Qiu, and Maxim Zabzine. “The Perturbative Partition Function of Supersymmetric 5D Yang–Mills Theory with Matter on the Five-Sphere.” Journal of High Energy Physics 2012, no. 8 (2012): 157. doi:10.1007/JHEP08(2012)157. Open preprint.
• Pestun, Vasily. “Localization of Gauge Theory on a Four-Sphere and Supersymmetric Wilson Loops.” Communications in Mathematical Physics 313 (2012): 71–129. doi:10.1007/s00220-012-1485-0. Open preprint.

Next step

Use the completed complex to derive the BPS loci, collective-coordinate measure, and one-loop determinant.