Product-Group and Quiver Gauge Dynamics

In a product-group theory, one multiplet can be charged under several gauge nodes. That simple fact couples their anomaly conditions, holomorphic scales, center symmetries, and strong-coupling limits. Sequential confinement or duality is reliable only when a declared hierarchy makes one node strong while its neighbors are weakly gauged flavor symmetries; the resulting composites must then be re-entered into the neighboring theory card.

Required background. Quantum-modified moduli and s-confinement supplies the exact single-node constraints, and holomorphic decoupling and scale matching fixes threshold powers. Helpful background. Chiral-theory anomaly constraints provides the node-by-node consistency workflow.

A two-node bifundamental theory

Consider four-dimensional $\mathcal N=1$ theory with

$$G=SU(N)_1\times SU(N)_2, \qquad N\geq3,$$

and a vectorlike bifundamental pair

$$X:({\bf N},\overline{\bf N}), \qquad \widetilde X:(\overline{\bf N},{\bf N}), \qquad W_{\mathrm{tree}}=0.$$

The cubic anomaly of each node cancels. Each node sees $N$ flavors, since the other node’s index supplies $N$ copies, and therefore

$$b_1=b_2=3N-N=2N, \qquad \Lambda_a^{2N}=\mu^{2N}e^{2\pi i\tau_a(\mu)}.$$

There are two independent theta angles and hence two complex holomorphic scales. Treating “the strong scale” as a single number already discards physical information.

The diagonal center element $(z,z)\in\mathbb Z_N\times\mathbb Z_N$ acts trivially on $X$ and $\widetilde X$. For the simply connected product gauge group it generates a diagonal electric $\mathbb Z_N^{(1)}$ one-form symmetry. Quotienting the gauge group by that diagonal center instead changes the genuine line lattice and introduces discrete-theta choices; it is a different theory even though the local fields are unchanged.

A baryon symmetry may be normalized as $B(X)=1$, $B(\widetilde X)=-1$. Its order-$N$ subgroup overlaps a gauge-center transformation, so its faithful action is $U(1)_B/\mathbb Z_N$. The anomaly-free scalar $R$ charges are $R(X)=R(\widetilde X)=0$: for either node, $N+N(R-1)=0$. These finite quotients and the diagonal one-form symmetry must be carried through any claimed infrared description.

Sequential strong dynamics

Assume

$$\lvert\Lambda_1\rvert\gg\lvert\Lambda_2\rvert.$$

Near the scale $\Lambda_1$, node 2 is weak and may be treated as a weakly gauged subgroup of the $N_f=N_c=N$ flavor symmetry of node 1. Node 1 then has the exact composite coordinates

$$M=\widetilde X X, \qquad B_1=\det X, \qquad \widetilde B_1=\det\widetilde X,$$

obeying

$$\boxed{\det M-B_1\widetilde B_1=\Lambda_1^{2N}.}$$

$M$ transforms by conjugation under $SU(N)_2$ and decomposes into an adjoint plus a singlet; the two baryons are node-2 singlets. This is the first mandatory update to the theory card. Below $\Lambda_1$, node 2 no longer couples to $N$ elementary flavor pairs; it couples to constrained composites with a noncalculable Kähler metric near the quantum region. The exact $N_f=N_c$ constraint and its mass-deformation checks are established in Seiberg 1994, pp. 6857–6863.

The branches are physically distinct. At a generic diagonalizable $M$ with distinct eigenvalues, its adjoint expectation value breaks $SU(N)_2$ to its maximal torus. At $M\propto\mathbf1$, node 2 remains unbroken. At baryonic points the constraint can be satisfied with rank-deficient $M$ and nonzero $B_1\widetilde B_1$. A formula derived on one stabilizer stratum cannot be extended through an enhanced-symmetry locus without adding the degrees of freedom that become light there.

Why two independent constraints are generally wrong

If one instead starts with $\lvert\Lambda_2\rvert\gg\lvert\Lambda_1\rvert$, the same reasoning gives a node-2 quantum constraint in an oppositely ordered description. The two descriptions concern the same microscopic fields and overlapping composites. When $\Lambda_1$ and $\Lambda_2$ are comparable, it is not valid to impose two single-node constraints as if the nodes had independent flavor fields. Holomorphic quantities may depend on the dimensionless ratio $\Lambda_1^{2N}/\Lambda_2^{2N}$, and additional singular loci can occur when the hierarchy is removed.

The correct statement is limited but useful: each sequential description is exact in its holomorphic variables within a regime where the neighboring gauge coupling is parametrically weak at the first strong scale. Continuing between the two regimes requires anomaly matching, branch tracking, and control of every singularity; it is not guaranteed by writing both limiting answers on the same line. Product-group dualities built by successively dualizing nodes pass many such tests, as shown in Poppitz, Shadmi, and Trivedi 1996, §§ 2–4.

A clean mass-threshold check

Add the gauge-invariant deformation

$$W_m=m\,\operatorname{Tr}(\widetilde X X).$$

For $\lvert m\rvert\gg\lvert\Lambda_1\rvert,\lvert\Lambda_2\rvert$, all $N$ flavor pairs are integrated out from both nodes. Each node becomes pure $SU(N)$ SYM. Since $b_{\mathrm{high}}=2N$ and $b_{\mathrm{low}}=3N$, holomorphic matching gives independently

$$\boxed{ \Lambda_{a,\mathrm{pure}}^{3N} =m^N\Lambda_a^{2N}, \qquad a=1,2.}$$

The exponent $N$ counts massive flavor pairs at that node. This deformation yields $N^2$ semiclassical choices of pure-theory condensate phases when the two nodes are well separated after decoupling, before identifications from any alternative global-form choice are imposed.

The same threshold can be approached from the confined variables when, for example, $\Lambda_1$ is crossed first. Then $m\operatorname{Tr}M$ lifts composite directions and the quantum constraint selects isolated branches. Matching the final pure-node scales must reproduce the boxed formula, but intermediate masses depend on the normalization of $M$ and its Kähler potential. Using the elementary mass $m$ in the holomorphic relation is robust; declaring a canonically normalized composite mass without the Kähler metric is not.

Node duality changes every neighboring card

For a general quiver, dualizing an $SU(N_i)$ node with $F_i$ effective flavors replaces it by $SU(F_i-N_i)$, adds mesons made from every length-two path through that node, and adds superpotential couplings between those mesons and magnetic bifundamentals. Four updates are compulsory:

1. Recompute the cubic and mixed anomalies of the neighboring gauge nodes.
2. Recompute each neighboring beta-function coefficient using the new charged fields.
3. Match holomorphic scales with the duality normalization scale and every massive threshold.
4. Recompute the faithful ordinary symmetry and genuine line operators; local anomaly matching alone does not fix global form.

Different dualization orders can give different ultraviolet quivers that are proposed to share an infrared limit. Agreement of anomalies, moduli spaces, deformations, and scale relations is strong duality evidence, but it does not make either ultraviolet Lagrangian a weakly coupled description everywhere. Explicit product-group examples and their renormalization-flow qualifications are analyzed in Poppitz, Shadmi, and Trivedi 1996, pp. 125–169.

Common pitfalls

Counting a bifundamental once. For the $SU(N)_1$ beta function, its $SU(N)_2$ index gives $N$ copies, and conversely. Missing this multiplicity changes both $b_a$ and anomaly coefficients.

Forgetting the diagonal one-form symmetry. Matter breaks the two center symmetries down to the subgroup acting trivially on every bifundamental. Quotienting by that subgroup changes the theory’s line operators.

Sequentially confining without a hierarchy. The first node may be treated as an isolated strong sector only if every neighboring gauge coupling is weak at that scale. Comparable scales require a simultaneous analysis or a dual description with its own stated regime.

Exercises

1. Center and beta functions

For the two-node theory, derive $b_1=b_2=2N$ and find the subgroup of $\mathbb Z_N\times\mathbb Z_N$ left unbroken by the matter.

Solution

Each node sees $N$ fundamentals and $N$ antifundamentals, whose total Dynkin index is $N$, hence $b_a=3N-N=2N$. A center element $(z_1,z_2)$ acts on $X$ as $z_1z_2^{-1}$ and on $\widetilde X$ inversely. Both are invariant precisely when $z_1=z_2$, leaving the diagonal $\mathbb Z_N^{(1)}$ for the simply connected product.

2. Match the massive theory

Derive the pure-node scale relation after adding $m\operatorname{Tr}(\widetilde X X)$, and explain why it holds for both nodes even when $\Lambda_1\neq\Lambda_2$.

Solution

At either node, integrating out $N$ flavor pairs changes $b$ from $2N$ to $3N$. Holomorphic matching at the common mass threshold gives $\Lambda_{a,\mathrm{pure}}^{3N}=m^{3N-2N}\Lambda_a^{2N}=m^N\Lambda_a^{2N}$. The two complex scales remain independent; the shared mass does not identify their couplings or theta angles.

References

• Poppitz, Erich, Yael Shadmi, and Sandip P. Trivedi. “Duality and Exact Results in Product Group Theories.” Nuclear Physics B 480 (1996): 125–169. DOI; arXiv.
• Seiberg, Nathan. “Exact Results on the Space of Vacua of Four-Dimensional SUSY Gauge Theories.” Physical Review D 49 (1994): 6857–6863. DOI; arXiv.