3D N=2 Gauged Supergravity: Off-Shell Branches

• 3D N=2 gauged supergravity is a family of theories with inequivalent off-shell formulations distinguished by compensator choices and distinct AdS branches.
• It employs superspace geometry and Chern–Simons formulations to construct various vacua, including ghost-free massive deformations in the (1,1) branch.
• Applications extend to matter-coupled U(1)-gauged models and string-derived reductions that facilitate c-extremization and holographic dualities.

Searching arXiv for recent and foundational papers on 3D $\mathcal N=2$ gauged supergravity and related AdS / Chern–Simons formulations. In the literature represented here, $D=3,\; \mathcal N=2$ gauged supergravity is not a single theory but a family of inequivalent constructions whose distinction is already visible off shell, in superspace, in higher-derivative completions, in matter-coupled $U(1)$-gauged models, and in Chern–Simons formulations. A central organizing fact is that three-dimensional $\mathcal N=2$ supergravity admits two inequivalent AdS branches, $\mathcal N=(1,1)$ and $\mathcal N=(2,0)$, associated with different compensators and different realizations of $R$-symmetry. In addition, there are matter-coupled $U(1)$-gauged models with Fayet–Iliopoulos extensions, string-derived gauged truncations in which $c$-extremization is realized as extremization of a $T$-tensor, and asymptotically flat $D=3,\; \mathcal N=2$0 Chern–Simons supergravities whose “gauged” description refers instead to a boundary chiral Wess–Zumino–Witten reduction (Kuzenko et al., 2011).

1. Off-shell branches and the meaning of the $D=3,\; \mathcal N=2$1 labels

Three-dimensional $D=3,\; \mathcal N=2$2 supergravity comes in two distinct off-shell versions, $D=3,\; \mathcal N=2$3 and $D=3,\; \mathcal N=2$4. The $D=3,\; \mathcal N=2$5 notation refers to the underlying AdS superalgebras $D=3,\; \mathcal N=2$6, with bosonic $D=3,\; \mathcal N=2$7 isometry $D=3,\; \mathcal N=2$8 and $D=3,\; \mathcal N=2$9-symmetry $U(1)$0. This means that the total supersymmetry count does not determine the theory off shell; the left/right split does (Alkac et al., 2014).

Both theories are obtained from the same $U(1)$1 Weyl multiplet by coupling a compensating multiplet and then fixing conformal symmetries. The $U(1)$2 theory uses a compensating scalar multiplet and is also referred to as Type I minimal or 3D old minimal. The $U(1)$3 theory uses a compensating vector multiplet and is also referred to as Type II minimal or 3D new minimal. In superspace there is also an improved complex linear compensator, but it furnishes a dual realization of $U(1)$4 AdS supergravity rather than a third inequivalent AdS branch (Kuzenko et al., 2011).

Formulation	Compensator	AdS realization
Type I minimal	chiral scalar multiplet	$U(1)$5
Type II minimal	vector multiplet	$U(1)$6
Improved non-minimal	improved complex linear multiplet	dual to $U(1)$7

The distinction has direct implications for gauging. In the $U(1)$8 construction, the compensating scalar is complex, so gauge fixing removes both dilatations and $U(1)$9. In the $\mathcal N=2$0 construction, the compensating scalar is real, so $\mathcal N=2$1 is not gauge-fixed away. This is the branch in which the $\mathcal N=2$2-symmetry gauge field remains explicitly in the Poincaré multiplet and can be viewed dynamically (Alkac et al., 2014).

2. Superspace geometry, cosmological terms, and $\mathcal N=2$3 gauging

The superspace structure group is

$\mathcal N=2$4

with covariant derivatives

$\mathcal N=2$5

The geometry is encoded in the dimension-1 torsion superfields $\mathcal N=2$6, $\mathcal N=2$7, $\mathcal N=2$8, and $\mathcal N=2$9, together with their Bianchi identities and super-Weyl transformations. The local symmetries are local superspace diffeomorphisms, local Lorentz $\mathcal N=(1,1)$0, local $\mathcal N=(1,1)$1, and super-Weyl symmetry before gauge fixing (Kuzenko et al., 2011).

In the Type I formulation the compensator is a covariantly chiral scalar $\mathcal N=(1,1)$2,

$\mathcal N=(1,1)$3

and with cosmological term the action is

$\mathcal N=(1,1)$4

In the gauge $\mathcal N=(1,1)$5, the on-shell AdS conditions are

$\mathcal N=(1,1)$6

which is precisely $\mathcal N=(1,1)$7 AdS superspace.

In the Type II formulation the compensator is a real prepotential $\mathcal N=(1,1)$8 with gauge-invariant field strength

$\mathcal N=(1,1)$9

and the AdS deformation is a supersymmetric Chern–Simons term,

$\mathcal N=(2,0)$0

In the gauge $\mathcal N=(2,0)$1, the background is characterized by

$\mathcal N=(2,0)$2

which is $\mathcal N=(2,0)$3 AdS superspace. Here the $\mathcal N=(2,0)$4 generator appears directly in the supersymmetry anticommutator, making the gauged $\mathcal N=(2,0)$5 structure explicit (Kuzenko et al., 2011).

This split controls matter couplings and supercurrents. In $\mathcal N=(2,0)$6 AdS, every theory admits a Ferrara–Zumino multiplet by improvement, while in $\mathcal N=(2,0)$7 AdS the natural supercurrent is $\mathcal N=(2,0)$8-multiplet-like. The rigid matter sectors also differ sharply. In $\mathcal N=(2,0)$9 AdS, the sigma-model target space must have exact Kähler form and therefore be noncompact. In $R$0 AdS, only $R$1-invariant models are allowed in the generic case, Fayet–Iliopoulos terms are allowed, compact Kähler targets are allowed in the $R$2 class, and a frozen vector multiplet exists, enabling central-charge real masses (Kuzenko et al., 2011).

3. Component multiplets and higher-derivative massive deformations

The $R$3 Weyl multiplet contains

$R$4

where $R$5 is the $R$6 $R$7-symmetry gauge field. The off-shell scalar compensator has components

$R$8

while the off-shell vector compensator has

$R$9

After gauge fixing, the $U(1)$0 Poincaré multiplet is

$U(1)$1

with $U(1)$2, whereas the $U(1)$3 Poincaré multiplet is

$U(1)$4

The corresponding bosonic Einstein–Hilbert Lagrangians are

$U(1)$5

for $U(1)$6, and

$U(1)$7

for $U(1)$8 (Alkac et al., 2014).

All off-shell invariants up to four derivatives were constructed for both branches. A common ingredient is the superconformal Lorentz Chern–Simons invariant

$U(1)$9

In the $c$0 branch, the invariant structure includes an off-diagonal $c$1-type term, and this is the crucial ingredient permitting a special higher-derivative combination with a supersymmetric AdS vacuum and a ghost-free massive spin-2 multiplet. For the parameter choice

$c$2

the theory becomes the supersymmetric extension of generalized massive gravity, and the paper identifies a maximally supersymmetric AdS background together with a ghost-free and unitary massive spectrum for suitable noncritical parameters (Alkac et al., 2014).

The $c$3 branch behaves differently. Its four-derivative invariants include $c$4, $c$5, and Ricci-squared structures, but the analysis does not find an independent $c$6-type invariant without curvature-squared contamination. This prevents the cancellation mechanism that is available in $c$7, and the conclusion is explicit: the $c$8 model does not seem to have a supersymmetric AdS vacuum with ghost-free spectrum. The same branch does, however, admit a ghost-free massive theory around Minkowski. For the parameter choice

$c$9

the maximally supersymmetric Minkowski vacuum has

$T$0

and the bosonic spectrum organizes into two massive spin-2 multiplets of $T$1 (Alkac et al., 2014).

4. Matter-coupled $T$2-gauged models and Fayet–Iliopoulos deformations

A distinct and more conventional gauged-supergravity sector is provided by the $T$3 $T$4-gauged matter-coupled supergravity of Abou-Zeid–Nash–Roček with Fayet–Iliopoulos extension. In the model studied in (Deger et al., 29 Oct 2025), the bosonic field content is the dreibein $T$5, one $T$6 gauge field $T$7, and $T$8 complex scalars; the analysis specializes to $T$9 with

$D=3,\; \mathcal N=2$00

The scalar target space must be Kähler, and the maximally symmetric two-dimensional choices considered are

$D=3,\; \mathcal N=2$01

The bosonic Lagrangian is

$D=3,\; \mathcal N=2$02

with

$D=3,\; \mathcal N=2$03

The scalar potential is generated by

$D=3,\; \mathcal N=2$04

The FI extension introduces the additional constant $D=3,\; \mathcal N=2$05, while $D=3,\; \mathcal N=2$06 parameterizes freedom in defining the superpotential. For $D=3,\; \mathcal N=2$07 and $D=3,\; \mathcal N=2$08, the parameters may be taken as $D=3,\; \mathcal N=2$09 or equivalently $D=3,\; \mathcal N=2$10, with

$D=3,\; \mathcal N=2$11

Supersymmetric extrema satisfy

$D=3,\; \mathcal N=2$12

There is always an extremum at $D=3,\; \mathcal N=2$13, and for $D=3,\; \mathcal N=2$14 there is an additional pair at $D=3,\; \mathcal N=2$15. The FI term therefore creates extra supersymmetric extrema and is crucial for interpolating string solutions between two vacua (Deger et al., 29 Oct 2025).

The supersymmetric analysis uses Killing-spinor bilinears and splits all solutions into null and timelike classes. In the null class, the general supersymmetric metric is

$D=3,\; \mathcal N=2$16

The solution space includes pp-waves in Minkowski or AdS, electromagnetic waves, null $D=3,\; \mathcal N=2$17-warped $D=3,\; \mathcal N=2$18, and string solutions with waves. For constant nonzero $D=3,\; \mathcal N=2$19 at $D=3,\; \mathcal N=2$20, the null warped geometry takes the form

$D=3,\; \mathcal N=2$21

with

$D=3,\; \mathcal N=2$22

For suitable parameter ranges, periodic identification $D=3,\; \mathcal N=2$23 yields a physically sensible supersymmetric null $D=3,\; \mathcal N=2$24-warped AdS black hole (Deger et al., 29 Oct 2025).

In the timelike class, the metric is a timelike fibration over a two-dimensional base,

$D=3,\; \mathcal N=2$25

The solution space includes global $D=3,\; \mathcal N=2$26, Ricci-flat or Euclidean Rindler-type geometries, timelike warped flat, timelike stretched warped $D=3,\; \mathcal N=2$27, and string-like solutions. The timelike stretched warped $D=3,\; \mathcal N=2$28 solution has

$D=3,\; \mathcal N=2$29

Charged string and string-like solutions interpolate between supersymmetric AdS or Minkowski extrema of the scalar potential; one endpoint controls the asymptotic geometry, while the other is a horizon, and the FI term is essential for the existence of the second vacuum and therefore for the horizon (Deger et al., 29 Oct 2025).

5. String-derived $D=3,\; \mathcal N=2$30 gauged supergravities and $D=3,\; \mathcal N=2$31-extremization

Wrapped D3-brane $D=3,\; \mathcal N=2$32 solutions in type IIB provide a string-theoretic origin for a broad class of three-dimensional gauged supergravities. For D3-branes wrapping Kähler two-cycles in a $D=3,\; \mathcal N=2$33, the dual two-dimensional SCFT has $D=3,\; \mathcal N=2$34, and the reduced three-dimensional theory preserves four real supercharges, exactly those of $D=3,\; \mathcal N=2$35 supergravity (Karndumri et al., 2013).

In the cleanest example, one starts from $D=3,\; \mathcal N=2$36 gauged supergravity and reduces on a Riemann surface $D=3,\; \mathcal N=2$37 with

$D=3,\; \mathcal N=2$38

After dualizing the three-dimensional vectors, one defines

$D=3,\; \mathcal N=2$39

and the scalar target manifold becomes

$D=3,\; \mathcal N=2$40

with Kähler potential

$D=3,\; \mathcal N=2$41

This realizes the standard $D=3,\; \mathcal N=2$42 requirement that the scalar manifold be Kähler (Karndumri et al., 2013).

The scalar potential is expressed entirely in terms of a $D=3,\; \mathcal N=2$43-tensor,

$D=3,\; \mathcal N=2$44

because when the $D=3,\; \mathcal N=2$45-symmetry is gauged, consistency requires the holomorphic superpotential to vanish,

$D=3,\; \mathcal N=2$46

Supersymmetry imposes the twist condition

$D=3,\; \mathcal N=2$47

and supersymmetric $D=3,\; \mathcal N=2$48 vacua are critical points of $D=3,\; \mathcal N=2$49, located at

$D=3,\; \mathcal N=2$50

The same framework yields a direct supergravity realization of $D=3,\; \mathcal N=2$51-extremization: $D=3,\; \mathcal N=2$52 In this sense, the exact central charge and the exact $D=3,\; \mathcal N=2$53-symmetry of the dual $D=3,\; \mathcal N=2$54 SCFT are determined by extremizing the $D=3,\; \mathcal N=2$55-tensor (Karndumri et al., 2013).

The same paper also studies less generic but explicit reductions. For an $D=3,\; \mathcal N=2$56 geometry descending from a universal $D=3,\; \mathcal N=2$57 truncation, a consistent $D=3,\; \mathcal N=2$58 truncation yields scalar manifold $D=3,\; \mathcal N=2$59, while for intersecting D3-branes or $D=3,\; \mathcal N=2$60 one can extract $D=3,\; \mathcal N=2$61 subsectors with scalar manifolds $D=3,\; \mathcal N=2$62 and $D=3,\; \mathcal N=2$63. The same three-dimensional framework also accommodates null-warped $D=3,\; \mathcal N=2$64 sectors in suitable truncations (Karndumri et al., 2013).

6. Asymptotically flat $D=3,\; \mathcal N=2$65 Chern–Simons supergravity and boundary gauging

A separate line of development concerns asymptotically flat $D=3,\; \mathcal N=2$66 Chern–Simons supergravity with internal $D=3,\; \mathcal N=2$67-symmetry. Here the bulk gauge superalgebra is the $D=3,\; \mathcal N=2$68 super-Poincaré algebra with bosonic generators

$D=3,\; \mathcal N=2$69

fermionic generators $D=3,\; \mathcal N=2$70, and defining brackets

$D=3,\; \mathcal N=2$71

$D=3,\; \mathcal N=2$72

$D=3,\; \mathcal N=2$73

The $D=3,\; \mathcal N=2$74-symmetry generator $D=3,\; \mathcal N=2$75 acts on the two supercharges, and the central element $D=3,\; \mathcal N=2$76 is required for the existence of a nondegenerate invariant bilinear form (Banerjee et al., 2019).

The Chern–Simons connection is

$D=3,\; \mathcal N=2$77

and the action is

$D=3,\; \mathcal N=2$78

In component form,

$D=3,\; \mathcal N=2$79

The field equation is the flatness condition

$D=3,\; \mathcal N=2$80

The paper’s main result is not the construction of a conventional matter-coupled gauged supergravity with scalar potentials and AdS vacua. Rather, it studies asymptotically flat $D=3,\; \mathcal N=2$81 supergravity at null infinity and derives equivalent reduced descriptions: a chiral Wess–Zumino–Witten model, a gauged chiral Wess–Zumino–Witten model, and a reduced phase-space theory that is the flat limit of a generalized Liouville theory, up to zero modes. The word “gauged” therefore refers to boundary gauging of a subalgebra of chiral WZW symmetries associated with first-class constraints, not to gauged supergravity in the standard higher-dimensional sense (Banerjee et al., 2019).

The asymptotic symmetry algebra is quantum $D=3,\; \mathcal N=2$82 super-BMS$D=3,\; \mathcal N=2$83. The gauged boundary action introduces Lagrange multipliers that enforce current constraints, while the reduced phase-space action becomes

$D=3,\; \mathcal N=2$84

The gauged chiral WZW model and the reduced Liouville-like theory possess identical global and gauge symmetries. This sharpens a common terminological distinction: the presence of an internal $D=3,\; \mathcal N=2$85-symmetry gauge field in the bulk Chern–Simons superalgebra is a genuine bulk gauge feature, but the “gauged” model constructed in the dual description is a gauged boundary WZW theory rather than a standard bulk gauged supergravity (Banerjee et al., 2019).