Quantum Master Equation and Anomaly Obstructions

The classical master equation becomes quantum only after choosing measure or renormalization data. In a finite-dimensional BV manifold with compatible density, the BV Laplacian $\Delta$ enters $\frac12(W,W)-i\hbar\Delta W=0$. In local quantum field theory the coincident functional derivatives in $\Delta$ are singular, so a renormalized master Ward identity replaces the naive formula. At each loop order its defect is a local ghost-number-one cocycle; a local counterterm removes the defect exactly when its cohomology class vanishes. Formal solvability is still not a construction of a positive, convergent quantum theory.

Required background. The BV complex and classical master equation fix the bracket and classical differential, while what an anomaly is distinguishes a genuine obstruction from a removable symmetry-breaking term.

Helpful background. BV quantization and obstruction classes give the effective-field-theory formulation, and Ward identities and anomalous obstructions give the causal-renormalization counterpart.

From a compatible measure to the QME

In finite-dimensional Darboux coordinates with a compatible Berezin density, define

$$\Delta F=(-1)^{\epsilon_A} \frac{\partial_r}{\partial\Phi^A} \frac{\partial_r F}{\partial\Phi_A^*}.$$

The operator has ghost number one, is second order, and satisfies $\Delta^2=0$. Its failure to be a derivation generates the antibracket:

$$\Delta(FG)=(\Delta F)G+(-1)^{\epsilon_F}F\Delta G +(-1)^{\epsilon_F}(F,G).$$

For the Lorentzian weight $e^{iW/\hbar}$, the condition $\Delta e^{iW/\hbar}=0$ is

$$\frac12(W,W)-i\hbar\Delta W=0.$$

This equation implies nilpotence of the quantum BV differential

$$\widehat sF=(W,F)-i\hbar\Delta F.$$

The implication uses the graded Jacobi identity, $\Delta^2=0$, and compatibility of $\Delta$ with the bracket. Conversely, nilpotence on a restricted set of observables does not establish the full QME or the existence of the density. Gomis, París, and Samuel derive the QME, its gauge-fixing role, and the loopwise anomaly equation in Gomis, París, and Samuel 1995, §§8.2 and 8.5, pp. 108–116.

Loopwise obstruction and counterterm freedom

Write a formal quantum action

$$W=S+\hbar M_1+\hbar^2M_2+\cdots ,$$

where $S$ solves the CME and every $M_k$ is required to be a local functional in the declared renormalization scheme. At first order,

$$sM_1=i\Delta S,\qquad s=(S,\,\cdot\,).$$

If a regulator or renormalization prescription instead produces a defect $\mathcal A_1$, consistency implies $s\mathcal A_1=0$. Changing the local counterterm $M_1\mapsto M_1+B_1$ shifts $\mathcal A_1$ by an $s$-exact term. Therefore the invariant obstruction is

$$[\mathcal A_1]\in H^1_{\mathrm{loc}}(s),$$

or, for local densities modulo integrations by parts, $H^{1,n}(s\mid d)$. If the class vanishes, choose a local $B_1$ and continue to the next loop order. If it is nonzero, no local counterterm restores the QME with the same fields and symmetry. Vanishing at one loop is not the converse of all-order solvability; a new cocycle may appear later unless an independent nonrenormalization or cohomological theorem excludes it.

In field theory, the displayed $\Delta S$ contains products of distributions at coincident points and is not defined before renormalization. One must not infer $\Delta^2=0$ for that formal expression. Perturbative AQFT constructs renormalized time-ordered products and an anomaly map in the master Ward identity; the QME is then a local relation in the algebraic adiabatic limit Fredenhagen and Rejzner 2013, §§4.2–4.3, pp. 716–721. This formulation retains compact switching functions and does not assert a global adiabatic limit or series convergence.

One-loop chiral gauge test

For four-dimensional left-handed Weyl fermions of Abelian charges $q_i$, the parity-odd one-loop defect has a representative proportional to

$$\mathcal A_1(c,A)\propto \left(\sum_i q_i^3\right)\int c\,F\wedge F ,$$

with normalization fixed by the chirality, generator, and effective-action conventions. The representative has ghost number one and obeys $s\mathcal A_1=0$ because $sc=0$ and $sF=0$. A change of regulator may add $sB+dC$, but it cannot remove a nonzero class using a local counterterm. When the relevant charge sum vanishes, this particular perturbative gauge-anomaly class cancels; mixed and global anomalies require separate tests.

This is the exact first application returned to Standard-Model anomaly cancellation: evaluate the one-loop master-equation defect, type it as a ghost-number-one local class, and ask whether a local counterterm removes the representative. The application is a cohomological and perturbative statement, not a nonperturbative existence theorem.

An independent check differentiates the consistency equation: applying $s$ to the order-$\hbar$ defect gives zero by $s^2=0$ and the compatible finite-dimensional identities, or by the renormalized Wess–Zumino relation in QFT. A second check changes $M_1$ by a local functional and verifies that only the representative, not the class, changes.

Adversarial failures

A nonlocal functional such as one containing $\Box^{-1}$ may cancel a local expression algebraically, but it is not an allowed local counterterm and therefore does not trivialize the local anomaly class. Likewise, writing $\Delta^2=0$ for unrenormalized functional derivatives merely assumes away the ultraviolet problem. The strongest valid conclusion in those cases is that a formal manipulation cancels a chosen representative; quantum gauge consistency has not been established.

Even an all-order renormalized QME constructs a formal deformation of local observables. It does not by itself give a convergent series, an infrared limit, reflection positivity, or a gauge-invariant physical Hilbert space.

Exercises

Show that changing $M_1$ by $B_1$ changes the defect by an exact term.

Solution

At order $\hbar$, the QME defect is $i\Delta S-sM_1$ in the displayed convention. Replacing $M_1$ by $M_1+B_1$ changes it by $-sB_1$. Hence its cohomology class is independent of the finite local counterterm.

Check the ghost number of the Abelian chiral representative.

Solution

$c$ has ghost number one and $F$ has ghost number zero, so $\int cF\wedge F$ has ghost number one. Its form degree is four, as required for integration in four spacetime dimensions.

References

• Batalin, Igor A., and Grigori A. Vilkovisky. “Gauge Algebra and Quantization.” Physics Letters B 102 (1981): 27–31. DOI.
• Fredenhagen, Klaus, and Katarzyna Rejzner. “Batalin–Vilkovisky Formalism in Perturbative Algebraic Quantum Field Theory.” Communications in Mathematical Physics 317 (2013): 697–725. DOI; Open PDF.
• Gomis, Joaquim, Jordi París, and Stuart Samuel. “Antibracket, Antifields and Gauge-Theory Quantization.” Physics Reports 259 (1995): 1–145. DOI; Open PDF.