Holomorphic Couplings and Background Superfields

A mass, Yukawa coupling, or complexified gauge coupling can be promoted to a nondynamical chiral superfield. The Wilsonian superpotential must then be holomorphic in both dynamical chiral fields and these background sources. Combining that holomorphy with spurionic symmetries, dimensions, regular limits, and one controlled normalization input sharply constrains exact F-terms—but holomorphy alone rarely fixes an arbitrary function.

Required background. Wess–Zumino models supplies the chiral action and auxiliary-field equations. Coupling to background gauge fields explains why nondynamical sources are useful probes of Ward identities.

Helpful background. Branches, sheets, and monodromy supplies the analytic language needed when exact terms contain logarithms or fractional powers.

Couplings as chiral sources

Consider chiral fields $\Phi_i$ and a Wilsonian action at scale $\mu$,

$$S_\mu=\int d^4x\,d^4\theta\;K_\mu +\left[\int d^4x\,d^2\theta\; W_\mu(\Phi_i,\lambda_A)+\text{h.c.}\right].$$

Promoting a constant coupling $\lambda_A$ to a background chiral superfield $\Lambda_A(x,\theta)$ means imposing

$$\bar D_{\dot\alpha}\Lambda_A=0$$

while omitting a kinetic term and not integrating over $\Lambda_A$ in the path integral. At the end one sets its scalar component to the desired constant and its higher components to zero. The promotion is a source construction, not the addition of a new propagating particle.

Local supersymmetry then permits the Wilsonian F-term to depend holomorphically on the chiral variables,

$$W_\mu=W_\mu(\Phi_i,\Lambda_A),$$

but not on $\Lambda_A^dagger$. Real wavefunction factors and the Kähler potential are not subject to this holomorphic restriction. The statement assumes a supersymmetric regulator, a local Wilsonian expansion at nonzero cutoff, and a patch on field/source space where the selected variables are nonsingular. Seiberg formulated the modern background-coupling argument in precisely this Wilsonian setting Seiberg 1993, pp. 470–472.

The source may be assigned a transformation law that makes a classical term formally invariant. Such a spurionic symmetry says that theories at different values of the coupling belong to one covariant family. It need not be a symmetry of a single theory after the source is frozen.

A Wess–Zumino constraint, step by step

Take one chiral field with

$$W_{\mathrm{tree}}=\frac12m\Phi^2+\frac13y\Phi^3.$$

Treat $m$ and $y$ as chiral sources. Assign an auxiliary $U(1)_A$ charge

$$q_A(\Phi)=1, \qquad q_A(m)=-2, \qquad q_A(y)=-3,$$

and an R-charge assignment

$$R(\Phi)=1, \qquad R(m)=0, \qquad R(y)=-1, \qquad R(W)=2.$$

Both tree terms are invariant. Since $[\Phi]=1$, $[m]=1$, and $[y]=0$, any holomorphic superpotential compatible with these data can be written on the patch $m\neq0$ as

$$W_{\mathrm{eff}}=m\Phi^2F(z), \qquad z=\frac{y\Phi}{m},$$

because $z$ is dimensionless and neutral under both symmetries. This is already a strong result: every allowed term is contained in one holomorphic function. It is not yet a nonrenormalization theorem. Symmetry and holomorphy permit, for example, $F(z)=1/2+z/3+c_2z^2+\cdots$.

Now add the missing information in the right order:

1. Domain. Work with a Wilsonian cutoff $\mu>0$ and with $m$ in a punctured neighborhood where no field being integrated out becomes massless.
2. Weak-coupling boundary value. As $y\to0$ at fixed $m$, the action becomes Gaussian, so $F(0)=1/2$.
3. Perturbative theorem. Superspace locality and $D$-algebra imply that perturbative loops do not generate a local Wilsonian superpotential. Hence no perturbative coefficient beyond the tree value is present.
4. Nonperturbative sectors. This ungauged four-dimensional model has no finite-action instanton sector supplying terms nonanalytic in $y$.

Consequently, in the Wilsonian holomorphic variables,

$$W_{\mathrm{eff}}=\frac12m\Phi^2+\frac13y\Phi^3.$$

This is the standard local Wilsonian result obtained by the superspace nonrenormalization theorem Weinberg 2000, §§ 27.6 and 30.3. It does not say that all amplitudes are tree-level. The Kähler potential and wavefunction normalization run; after $\Phi_c=Z^{1/2}\Phi$, the canonically normalized parameters inherit $Z$ dependence.

What a spurionic symmetry actually proves

For a proposed exact term, a complete constraint chain has five logically different links:

Input	What it controls	What it cannot supply by itself
chirality and locality	holomorphic dependence in a Wilsonian F-term	infrared behavior of the 1PI action
ordinary and spurionic symmetries	charges and invariant ratios	an arbitrary neutral holomorphic function
anomalous Jacobians	transformation of $\tau$ or $\Lambda_h$	a convention-independent overall coefficient
regularity and limits	excluded singularities and boundary values	behavior across a genuine singular locus
controlled dynamics	residual constants or branch data	global validity beyond the controlled domain

The order matters. If a dimensionless neutral invariant exists, the most honest intermediate answer is “a holomorphic function of that invariant.” One may reduce it further only with stated information about regularity, asymptotics, weakly coupled regions, or other dynamics.

An anomalous transformation is not discarded. Suppose a chiral rotation of matter changes the regulated measure by a shift of the gauge angle. Assigning the corresponding transformation to the background complex coupling makes the family of theories covariant. Equivalently, the holomorphic scale $\Lambda_h^{b_0}=\mu^{b_0}e^{2\pi i\tau}$ carries a spurionic charge. The anomaly therefore adds a constraint rather than erasing one; the normalization and sign are developed on R-symmetry, anomalies, and the holomorphic scale.

Singularities, branches, and frozen sources

Three qualifications prevent common overclaims.

A holomorphic function may be singular. Integrating out a field produces an effective description valid only where its mass matrix is invertible. Poles or logarithms at the excluded locus can be physically required. Regularity may be imposed only where the effective variables remain valid.

Fractional powers require a branch. A term such as $(\Lambda^b/X)^{1/k}$ describes $k$ local branches. Analytic continuation can permute them. Symmetry fixes how the collection transforms; it does not select a vacuum.

Freezing a source reduces the manifest symmetry. Spurionic covariance relates different constant values of a coupling. Once $m$ is fixed, a transformation that changes $m$ is not an internal symmetry of that one theory. Ward identities must keep this distinction explicit.

The component $F_{\Lambda_A}$ of a background chiral source can be used to probe operator insertions or introduce controlled soft breaking. Doing so changes the question: the exact supersymmetric F-term statement applies before such a component is turned on.

Common pitfalls

Holomorphy determines the function. It determines the analytic type of the dependence. Neutral dimensionless ratios can support arbitrary holomorphic functions until limits or dynamics remove them.

A spurion is a new dynamical field. It is a nondynamical source unless a kinetic term and path integration are explicitly added. Confusing the two changes both the spectrum and the effective action.

Regular at one point means regular everywhere. Coordinates on moduli space and field content of an effective theory can fail at discriminant loci. Regularity is always a statement on a declared patch.

Exercises

In the Wess–Zumino example, choose instead $R(\Phi)=2/3$, so that $R(y)=0$ and $R(m)=2/3$. Show that the same invariant $z=y\Phi/m$ and the same functional form follow.

Solution

The ratio has

$$R(z)=R(y)+R(\Phi)-R(m)=0+\frac23-\frac23=0,$$

and it is also neutral under $U(1)_A$. The prefactor has $R(m\Phi^2)=2$ and mass dimension three. Thus $W=m\Phi^2F(z)$ again. Different spurionic charge bases encode the same invariant information.

Why is the limit $m\to0$ at fixed $y$ not enough to demand that $F(z)$ be regular at $z=\infty$?

Solution

At $m=0$ the spectrum and the convenient coordinate patch change: the quadratic mass vanishes, and the expansion that treated $m$ as an invertible source no longer applies. Regularity at $z=\infty$ would therefore be an additional physical assumption, not a consequence of holomorphy on the $m\neq0$ patch.

References

• Nathan Seiberg, “Naturalness versus Supersymmetric Non-renormalization Theorems,” Physics Letters B 318 (1993), 469–475, arXiv, DOI.
• Steven Weinberg, The Quantum Theory of Fields, Volume III: Supersymmetry, Cambridge University Press (2000), §§ 27.6 and 30.3, DOI.