The U(1)A Problem and QCD Topology

The apparent $U(1)_A$ puzzle is resolved because the singlet axial current is not conserved in quantum QCD, even when every quark mass vanishes. Its anomalous divergence is tied to gauge-field topology and Dirac zero modes; at large $N_c$, the pure-Yang–Mills topological susceptibility controls the leading singlet pseudoscalar mass relation. These are exact identities and controlled limits, not evidence that a dilute instanton gas universally describes the QCD vacuum.

Required background. Chiral Symmetry in QCD supplies the singlet current and the non-singlet breaking pattern; Regulated Jacobians and Measure Variation supplies the measure anomaly; Theta Dependence in Yang–Mills Theory and QCD supplies vacuum-angle conventions.

Helpful background. Theta Terms, Periodicity, and Vacuum Sectors supplies the global topological-sector interpretation.

The singlet axial Ward identity

Define

$$J_5^\mu=\sum_{f=1}^{N_f}\bar q_f\gamma^\mu\gamma_5q_f, \qquad q_{\rm top}(x)=\frac{g_s^2}{32\pi^2} G_{\mu\nu}^a\widetilde G^{a\mu\nu},$$

where $\widetilde G^{a\mu\nu}=\epsilon^{\mu\nu\rho\sigma}G^a_{\rho\sigma}/2$. The renormalized anomalous Ward identity is

$$\partial_\mu J_5^\mu =2i\bar qM\gamma_5q+2N_fq_{\rm top}(x).$$

The triangle calculation fixes the local anomaly coefficient Adler 1969, pp. 2426–2433, while the regulated change of the fermion measure explains why it persists nonperturbatively Fujikawa 1979, pp. 1195–1198. The separate composite operators can mix under renormalization; the correctly renormalized Ward identity is the invariant statement.

In massless classical QCD, spontaneous breaking of the full $U(N_f)_L\times U(N_f)_R$ would appear to require $N_f^2$ Goldstone bosons. Quantum mechanically, the continuous singlet axial factor is anomalous, so Goldstone’s theorem applies only to the $N_f^2-1$ non-singlet axial generators. The absence of an additional light singlet is therefore not a contradiction. Topological dynamics determine how the singlet channel acquires its spectrum; they do not retroactively turn $U(1)_A$ into an exact symmetry.

Topological charge and Dirac zero modes

For a smooth Euclidean gauge field of finite action, compactifying the boundary of $\mathbb R^4$ makes the topological charge

$$Q=\int d^4x\,q_{\rm top}(x)$$

an integer. Choose the orientation and chirality convention in which the index theorem reads

$$\operatorname{ind}\mathcal D=n_- - n_+=Q,$$

where $n_\mp$ counts normalizable zero modes of chirality $\gamma_5=\mp1$. Reversing the orientation or the definition of $\gamma_5$ reverses both signs, leaving every physical conclusion unchanged. The mathematical index relation is exact for the stated smooth-field hypotheses Atiyah and Singer 1968, pp. 546–556.

Under $q\mapsto e^{i\alpha\gamma_5}q$, the fermion measure in sector $Q$ changes by the phase associated with $2N_f\alpha Q$. Equivalently, a $Q=1$ background supplies one chiral zero mode per massless flavor. Saturating those zero modes generates the schematic multi-fermion interaction

$$\mathcal L_{\rm zero\ mode} \propto e^{i\theta}\det_{f,g}(\bar q_R^{,f}q_L^{,g}) +e^{-i\theta}\det_{f,g}(\bar q_L^{,f}q_R^{,g}).$$

It is invariant only for the discrete axial rotations with $e^{2iN_f\alpha}=1$. The zero-mode selection rule is exact sector by sector. Deriving the displayed local effective interaction by a dilute instanton calculation adds semiclassical size, separation, and coupling assumptions; ’t Hooft 1976, §§IV–VI is the canonical calculation. Instanton dominance in four-dimensional QCD is therefore a mechanism hypothesis, not part of the anomaly identity.

Topological susceptibility

Define the vacuum functional by

$$Z(\theta)=\sum_Q e^{i\theta Q}Z_Q, \qquad E(\theta)=-\lim_{V_4\to\infty}\frac1{V_4}\ln Z(\theta).$$

At $\theta=0$, the Euclidean topological susceptibility is most safely defined through the vacuum energy,

$$\chi_{\rm top} =\left.\frac{\partial^2E}{\partial\theta^2}\right|_{\theta=0} =\lim_{V_4\to\infty}\frac{\langle Q^2\rangle_c}{V_4} =\int d^4x\,\langle q_{\rm top}(x)q_{\rm top}(0)\rangle_c.$$

The last equality is formal until contact terms and the renormalization of the local density product are specified; the vacuum-energy derivative fixes those ambiguities. Susceptibilities in pure Yang–Mills and in full QCD are different quantities.

If one quark is exactly massless, an axial rotation of that flavor can shift away $\theta$ without changing a mass term. Then the full-QCD vacuum energy is $\theta$ independent and

$$\chi_{\rm top}^{\rm QCD}=0$$

in that limit, despite a generally nonzero pure-Yang–Mills susceptibility. At small nonzero masses, anomalous Ward identities relate the full-QCD susceptibility to the condensate and quark masses; it must not be replaced by the pure-gauge value.

The large-N singlet mass relation

In the ’t Hooft large-$N_c$ limit with fixed $N_f$, introduce the normalized singlet generator and decay constant by

$$T^0=\frac{\mathbf1}{\sqrt{2N_f}}, \qquad A_0^\mu=\bar q\gamma^\mu\gamma_5T^0q, \qquad \langle0|A_0^\mu|\eta_0(p)\rangle=iF_0p^\mu.$$

Then $F_0^2=O(N_c)$ while the pure-Yang–Mills susceptibility is $O(1)$. The anomaly-induced singlet mass is therefore $O(1/N_c)$. Matching the anomalous chiral effective theory to the pure-gauge vacuum energy gives the Witten–Veneziano relation

$$m_{\eta'}^2+m_\eta^2-2m_K^2 =\frac{2N_f}{F_0^2}\chi_{\rm YM} +\text{corrections subleading in }1/N_c \text{ and the chiral expansion}.$$

Here $F_0$ is the singlet decay constant in the stated normalization and $\chi_{\rm YM}$ is evaluated in pure Yang–Mills theory. Replacing it with the full-QCD susceptibility would contradict the massless-quark check above. The relation follows in a controlled combined large-$N_c$ and chiral analysis Witten 1979, pp. 269–276 and Veneziano 1979, pp. 213–220; it is not an exact finite-$N_c$ mass formula.

What each statement establishes

Statement	Type	Scope and limit
$\partial_\mu J_5^\mu=2i\bar qM\gamma_5q+2N_fq_{\rm top}$	Renormalized operator identity	Exact after regulator and operator conventions are fixed
$n_- - n_+=Q$	Index theorem	Smooth finite-action Euclidean fields with the stated orientation
Zero-mode determinant vertex	Semiclassical mechanism plus exact selection rule	Selection rule is topological; local dilute-gas coefficient needs semiclassical control
Witten–Veneziano relation	Controlled asymptotic relation	Leading combined large-$N_c$ and chiral orders; uses pure-Yang–Mills susceptibility
Numerical $\chi_{\rm top}$	Regulated evidence	Requires density definition, volume, spacing, continuum limit, and uncertainties
Singlet meson spectrum	Phenomenological consequence	Requires mixing, decay-constant, quark-mass, and higher-order analysis

This hierarchy prevents three common overclaims: the anomaly equation is not a proof of instanton dominance; a measured singlet mass is not a direct measurement of a regulator-independent local $q(x)q(0)$ correlator; and the large-$N_c$ relation is not a finite-$N_c$ identity.

For the numerical index and topology diagnostics, continue to Topology, Lattice Index, and Diagnostics.

Common pitfalls

Using the full-QCD susceptibility in Witten–Veneziano. The leading relation contains the pure-Yang–Mills susceptibility. Full QCD has fermion-induced Ward identities and vanishes when any quark is exactly massless.

Identifying the anomaly with one microscopic mechanism. The anomalous Ward identity is regulator independent; a dilute instanton gas is a particular semiclassical realization whose applicability must be demonstrated.

Dropping contact terms in $\langle q(x)q(0)\rangle$. Define the susceptibility through $E''(0)$ or state the density-product prescription. The unsmeared integral is not self-defining.

Exercise

Use the large-$N_c$ scalings $F_0^2=O(N_c)$ and $\chi_{\rm YM}=O(1)$ to find the scaling of the anomaly-induced singlet mass. Then explain why the result is compatible with a ninth Goldstone boson as $N_c\to\infty$.

Solution

The Witten–Veneziano term scales as

$$m_0^2\sim\frac{2N_f\chi_{\rm YM}}{F_0^2}=O(N_c^{-1}).$$

Thus the anomalous singlet mass vanishes in the strict large-$N_c$ limit at fixed $N_f$. In that limit the anomaly is parametrically suppressed in singlet meson dynamics, so the singlet joins the nonet of Goldstone modes. At finite $N_c$, the anomaly lifts it.

References

• Adler, Stephen L. “Axial-Vector Vertex in Spinor Electrodynamics.” Physical Review 177 (1969): 2426–2438. DOI.
• Atiyah, M. F., and I. M. Singer. “The Index of Elliptic Operators: III.” Annals of Mathematics 87 (1968): 546–604. DOI.
• Fujikawa, Kazuo. “Path-Integral Measure for Gauge-Invariant Fermion Theories.” Physical Review Letters 42 (1979): 1195–1198. DOI.
• ’t Hooft, Gerard. “Computation of the Quantum Effects Due to a Four-Dimensional Pseudoparticle.” Physical Review D 14 (1976): 3432–3450; erratum 18 (1978): 2199. DOI.
• Veneziano, Gabriele. “$U(1)$ Without Instantons.” Nuclear Physics B 159 (1979): 213–224. DOI.
• Witten, Edward. “Current Algebra Theorems for the $U(1)$ Goldstone Boson.” Nuclear Physics B 156 (1979): 269–283. DOI.