Three-Dimensional Supersymmetric Gauge Theory and Duality Webs

Three-dimensional supersymmetric gauge theory is unusually rich because the gauge field can carry a quantized Chern–Simons coupling and a vector multiplet contains a real scalar Aharony et al. 1997, §§2–4. In Euclidean radial quantization, magnetic flux through the small two-sphere linking an insertion defines a local monopole operator Borokhov, Kapustin, and Wu 2002, §§2–3. These ingredients tie vacuum geometry and infrared duality to parity anomalies and background contacts Closset et al. 2012, §§2–3. This chapter develops one disciplined way to analyze them: first specify the global and quantized data, then compute monopole sectors and deformations, and only then state an infrared equivalence.

Helpful background. The chapter uses Chern–Simons level quantization, disorder operators defined by singular boundary conditions, and fixed points and relevant RG flows.

Enter the three-dimensional problem

Unless a page says otherwise, spacetime is oriented Lorentzian $\mathbb R^{1,2}$ with metric $(+--)$ and $\epsilon^{012}=+1$ ; Euclidean formulas are explicitly marked. Gauge fields are Hermitian. For a compact, charge-one $U(1)$ connection,

S_{\mathrm{CS}}[a]=\frac{k}{4\pi}\int a\wedge da, \qquad S_{\mathrm{BF}}[a,b]=\frac{k_{BF}}{2\pi}\int a\wedge db,

and $\frac{1}{2\pi}\int_\Sigma da\in\mathbb Z$ on every closed two-cycle of an allowed compact bundle. The linking two-sphere used above is the Euclidean operator-state construction; the displayed Lorentzian actions are its local dynamical continuation. These local expressions do not determine the required tangential structure—orientation, spin, a correlated spin $_c$ formulation, or a Pin $^\pm$ extension when orientation reversal is gauged—nor the allowed line operators. A duality statement that suppresses that data can get local correlators right while still describing a different quantum theory.

The chapter concentrates on $\mathcal N=2$ theories, with four real supercharges and $U(1)_R$ , and $\mathcal N=4$ theories, with eight real supercharges and R-symmetry Lie algebra $\mathfrak{su}(2)_H\oplus\mathfrak{su}(2)_C$ . The faithfully acting global quotient is theory dependent. “Equality” of two gauge theories below always means an infrared claim after matching relevant deformations, global symmetries, background contact terms, and the spectrum of genuine operators. It does not mean equality of their ultraviolet Lagrangians.

The safest unit of reasoning is a theory card, not a gauge-group label. A reproducible card states:

the required tangential structure—orientation, spin, or a correlated spin $_c$ formulation—and, if orientation-reversing backgrounds are admitted, the relevant Pin $^+$ or Pin $^-$ choice;
the compact global gauge group, allowed bundles, genuine lines, and magnetic cocharacter lattice;
every Yang–Mills, Chern–Simons, BF, and FI normalization;
matter representations, real masses, superpotential, and trial $R$ charges;
dynamical and background counterterms in one regulator convention;
the vacuum or Coulomb chamber in which effective levels are evaluated;
the claimed infrared endpoint and the evidence that supports it.

Two cards can use the same symbol $U(N)_k$ and still define different quantum theories if their Abelian level, quotient, spin dependence, transparent sector, or background contacts differ. Conversely, visibly different ultraviolet cards can flow to the same fixed point. Much of this chapter is the craft of distinguishing those two situations.

A five-minute preparation diagnostic

Before entering a duality calculation, try these three checks.

A charge-one Dirac fermion crosses from positive to negative real mass. By how much does its $U(1)$ Chern–Simons level change?
A compact $U(1)_k$ monopole has flux $m=1$ . What classical electric charge must be cancelled before it can be a local gauge-invariant operator?
Why can two theories have matching classical moduli-space dimensions but fail to be an $\mathcal N=4$ mirror pair?

The answers are, respectively: $\Delta k=-1$ in the convention where a massive fermion contributes $\frac12\operatorname{sgn}M$ ; gauge charge $+k$ for the positive-flux monopole in the displayed $+\frac{k}{4\pi}\int a\wedge da$ convention; and because mirror symmetry must match the quantum branch rings, symmetry actions, deformations, monopoles, and global data, not dimensions alone. If any step feels unfamiliar, begin with the first four pages in order.

Follow the dependency chain

If the question is about …	Start with …	The non-negotiable output is …
supercharges, fields, real masses, FI parameters	algebras and multiplets	reality conditions, off-shell fields, and the topological current
a Lagrangian or vacuum equations	Yang–Mills, Chern–Simons, and matter actions	level normalization, auxiliary-field elimination, and global gauge group
a fermion mass or a duality phase	parity anomalies and contact terms	gauge, flavor, $R$ , and gravitational counterterms modulo their allowed shifts
a Coulomb-branch coordinate	monopole operators	magnetic lattice, induced electric charge, zero modes, and dressing
an $\mathcal N=4$ exchange of branches	$\mathcal N=4$ mirror symmetry	a complete mass–FI and flavor–topological dictionary
a four-supercharge Seiberg-like pair	Aharony and Giveon–Kutasov dualities	ranks, levels, singlets, superpotential, contact terms, and parameter range
a relation between dualities	real-mass, FI, and compactification flows	the selected vacuum, induced levels, surviving monopoles, and order of limits
particle–vortex or bosonization descendants	supersymmetric parent webs	spin/spin $_c$ data, transparent sectors, phase checks, and the unprotected step

This order matters. For example, integrating out a massive Dirac fermion changes a Chern–Simons contact term before any low-energy field is discarded. Likewise, a bare monopole in a Chern–Simons theory is electrically charged and may fail to be a gauge-invariant local operator until it is dressed. These are structural constraints, not refinements appended after a proposed dictionary.

There are three natural routes through the chapter:

Build a theory: algebras and multiplets $\rightarrow$ actions $\rightarrow$ parity/contact terms $\rightarrow$ monopoles.
Test a supersymmetric duality: complete the first route, then choose $\mathcal N=4$ mirror symmetry or the Aharony/Giveon–Kutasov family, and finally follow its deformation flows.
Evaluate a descendant web: complete the duality and flow routes before examining particle–vortex or bosonization claims. The last step is deliberately the least protected.

Localized $F$ -maximization and accidental-symmetry corrections are developed later in the volume. Here one prepares their inputs—trial currents, monopole charges, contact terms, and possible accidental symmetries—without pretending that a trial $R$ charge is already the superconformal answer.

Three benchmark calculations

The same small examples recur across the pages. Keeping them fixed makes sign and normalization errors easier to detect.

One charge-one chiral

Choose a gauge-invariant regulator for a charge-one $\mathcal N=2$ chiral and write the bare dynamical level as $k_{\rm bare}=-\frac12$ . Integrating out its fermion gives

k_{\rm eff}(M)=-\frac12+\frac12\operatorname{sgn}M =\begin{cases} 0,&M>0,\\ -1,&M<0. \end{cases}

The level difference is exact, but the first line does not say that the remaining $U(1)_0$ gauge field is an empty gapped phase. Maxwell, FI, monopole, and global-symmetry data still decide its infrared fate. This distinction prevents a response calculation from being promoted into an unsupported phase claim.

N=4 SQED with one hypermultiplet

The infrared limit of $\mathcal N=4$ $U(1)$ SQED with one hypermultiplet is the basic mirror of a free twisted hypermultiplet. The minimal monopoles $V_\pm$ have dimension $\frac12$ and supply its two free chiral coordinates; the neutral Coulomb operator obeys, up to a nonzero normalization,

V_+V_-\propto\Phi.

Thus the electric quantum Coulomb branch becomes the mirror’s classical Higgs branch, while the SQED Higgs branch and the free theory’s Coulomb branch are points. This operator statement is much sharper than the slogan “Higgs and Coulomb branches exchange” Borokhov, Kapustin, and Wu 2002, PDF §4.2.

Aharony duality and its level-changing flow

For the representative $N_c=2$ , $N_f=3$ Aharony pair, the magnetic group is $U(1)_0$ . If the electric quarks have trial scalar $R$ charge $r$ , the elementary magnetic-theory singlets $v_\pm$ representing the electric monopole operators and the magnetic gauge-theory disorder operators $\widehat V_\pm$ have

R(v_\pm)=2-3r, \qquad R(\widehat V_\pm)=3r,

so each cross-coupling $v_+\widehat V_-+v_-\widehat V_+$ has $R=2$ . This is a zero-mode and root-counting check, not a determination of the exact value of $r$ Aharony 1997, pp. 71–76.

Use the single-trace convention

U(N)_k=\frac{SU(N)_k\times U(1)_{Nk}}{\mathbb Z_N}.

Starting instead from a generic Aharony parent with $N_f+p\geq N_c+1$ flavors and giving $p>0$ pairs a common positive axial mass produces the candidate Giveon–Kutasov endpoints

U(N_c)_p \quad\longleftrightarrow\quad U(N_f+p-N_c)_{-p}

in this convention Giveon and Kutasov 2009, §§2–3.

The displayed ranks and levels are only the visible part of the flow. The heavy determinants also leave flavor, $R$ , mixed, and gravitational contacts, and the magnetic vacuum can contain blocks that must be retained until their topological response is identified Benini, Closset, and Cremonesi 2011, PDF §§2–3 and 5.2.

Read every duality arrow as a record

For each ultraviolet description, record the following data before comparing it with another:

Geometry and global form: dimension, signature, required tangential structure (including spin, spin $_c$ , or Pin data when applicable), gauge group rather than merely its Lie algebra, allowed bundles, and genuine lines.
Quantized couplings: dynamical and background Chern–Simons levels, BF couplings, and gravitational contact terms in one stated regulator convention.
Local fields and interactions: representations, real masses, FI terms, superpotential, and $R$ -charge assignments.
Operators: polynomial gauge invariants, monopole sectors, required dressings, and any singlets introduced on the other side.
Deformations and phases: the map of masses, FI parameters, and superpotential couplings; every semiclassical gapped phase must produce the same residual topological theory and background response.
Infrared qualifications: parameter range, accidental symmetries, decoupled free operators, special small ranks, and which checks establish the claim.

The parity-anomaly calculation and the monopole-charge calculation are independent. Agreement in one does not repair failure in the other. Protected partition functions, supersymmetric indices, moduli spaces, chiral rings, and deformed phases provide complementary evidence, but none licenses dropping the global data Closset et al. 2012, §§2–3.

The following failure signals tell you where to return:

Failure signal	Missing record	Return to …
half-integral dynamical level with no regulator explanation	parity and global quantization	parity anomalies and contacts
“monopole” has unbalanced electric charge	Gauss law, chamber, or dressing	monopole operators
dual ranks match but massive phases do not	background/gravitational terms or vacuum selection	real-mass and compactification flows
branch dimensions match but ring relations do not	quantum branch and operator dictionary	$\mathcal N=4$ mirror symmetry
a nonsupersymmetric transition is called exact	evidence ceiling and relevant-coupling analysis	supersymmetric parent webs

Evidence has layers

An algebraic charge match, a chiral-ring isomorphism, an index identity, an $S^3$ partition-function identity, and a phase-by-phase topological response test probe different structures. A strong duality record uses several of them and says what each cannot see. For example, a local chiral ring may miss a transparent spin sector, while a localized integral may depend on a contour and a chosen counterterm representative.

The evidentiary ceiling drops when supersymmetry is broken. Quantization laws and background-response algebra remain exact, and matching massive phases remains a powerful check, but equality of the intervening finite-rank critical points becomes a dynamical conjecture. Large- $N$ calculations or numerical evidence can strengthen that conjecture without turning it into a theorem.

What this chapter does and does not claim

The supersymmetric dualities treated here have exact or protected tests and controlled deformation chains. Their particle–vortex and bosonization descendants are presented as infrared conjectures with strong phase, symmetry, and response checks—not as theorems Seiberg et al. 2016, §§2–3. The chapter stops at the relativistic field-theory web. Applications to quantum Hall systems, topological-insulator surfaces, and other many-body settings require additional microscopic symmetry and electromagnetic-response data.

At a boundary, a bulk Chern–Simons term is not gauge invariant by itself. One must supply boundary conditions and, when appropriate, boundary degrees of freedom whose anomaly cancels the inflow. All closed-manifold formulas in the chapter should therefore be re-examined before being used on a manifold with boundary.

The chapter also stops before condensed-matter implementation. A relativistic spin or spin $_c$ duality does not specify a lattice Hilbert space, filling constraint, microscopic time-reversal action, electromagnetic normalization, or disorder ensemble. Those data belong to the Volume 12 application bridge.

Review the chapter

Before accepting a proposed three-dimensional duality, answer all of the following.

Which magnetic cocharacters label local monopole sectors for the chosen global gauge group?
Which fermions become massive in each chamber, and what gauge/background contact terms do they induce?
Are the proposed monopoles gauge neutral, or is explicit matter dressing required?
Do masses, FI terms, ordinary flavor currents, and topological currents map consistently?
Do generic relevant deformations lead to the same gapped topological field theories, including gravitational response?
Is the claim supersymmetric and protected, or does it contain an additional nonsupersymmetric dynamical assumption?

If any answer is missing, the duality record is incomplete.

A successful reader should be able to complete two exit tasks without guessing conventions:

Starting from a theory card, compute the allowed minimal monopoles, their zero-mode and Chern–Simons charges, their necessary dressing, and the caveat separating a trial $R$ charge from an exact dimension.
Starting from a proposed dual pair, choose a generic real-mass chamber and show that the two sides produce the same light theory, residual topological sector, and background/gravitational response—or identify the precise point where the claim fails.

Those tasks are the chapter’s practical definition of understanding. Memorizing a list of named dualities is not.

References

Aharony, O. (1997), “IR Duality in $d=3$ $N=2$ Supersymmetric $USp(2N_c)$ and $U(N_c)$ Gauge Theories,” Physics Letters B 404, 71–76. doi:10.1016/S0370-2693(97)00530-3. Open PDF
Aharony, O., Hanany, A., Intriligator, K., Seiberg, N., and Strassler, M. J. (1997), “Aspects of $N=2$ Supersymmetric Gauge Theories in Three Dimensions,” Nuclear Physics B 499, 67–99. doi:10.1016/S0550-3213(97)00323-4. Open PDF
Benini, F., Closset, C., and Cremonesi, S. (2011), “Comments on 3d Seiberg-like Dualities,” Journal of High Energy Physics 2011(10), 075. doi:10.1007/JHEP10(2011)075. Open PDF
Borokhov, V., Kapustin, A., and Wu, X. (2002), “Monopole Operators and Mirror Symmetry in Three Dimensions,” Journal of High Energy Physics 2002(12), 044. doi:10.1088/1126-6708/2002/12/044. Open PDF
Closset, C., Dumitrescu, T. T., Festuccia, G., Komargodski, Z., and Seiberg, N. (2012), “Comments on Chern–Simons Contact Terms in Three Dimensions,” Journal of High Energy Physics 2012(09), 091. doi:10.1007/JHEP09(2012)091. Open PDF
Giveon, A., and Kutasov, D. (2009), “Seiberg Duality in Chern–Simons Theory,” Nuclear Physics B 812, 1–11. doi:10.1016/j.nuclphysb.2008.09.045. Open PDF
Seiberg, N., Senthil, T., Wang, C., and Witten, E. (2016), “A Duality Web in 2+1 Dimensions and Condensed Matter Physics,” Annals of Physics 374, 395–433. doi:10.1016/j.aop.2016.08.007. Open PDF

Original QFT.org content:CC BY 4.0, unless an item supplies different terms. Third-party material retains its own terms.