Gauged Courant Sigma Models Overview

• Gauged Courant Sigma Models are field theories defined by the geometry of twisted Courant algebroids and Dirac structures that unify various 2D sigma models.
• They implement gauge symmetry by constraining auxiliary fields to Dirac structures, which enables consistent gaugings even with non-closed fluxes and deformed connections.
• The framework extends to quantization via the AKSZ-BV formalism, allowing for higher-dimensional generalizations and models with advanced boundary conditions.

A Gauged Courant Sigma Model (GCSM) is a field theory whose gauge structure and dynamics are determined by the geometry of (possibly twisted) Courant algebroids and their Dirac structures over a target manifold. GCSMs unify and generalize conventional gauged 2D sigma models, including those with Wess–Zumino terms, Poisson and Jacobi structures, and capture models with non-trivial fluxes and non-closed 3-forms. Fundamentally, a GCSM results from imposing that auxiliary or gauge fields take values in a Dirac structure of a Courant algebroid, subsuming all consistent gaugings into a single geometric framework (Chatzistavrakidis et al., 2017, Kotov et al., 2014, Chatzistavrakidis et al., 2016, Chatzistavrakidis et al., 2020, Ikeda, 31 Jan 2026).

1. Courant Algebroids, Dirac Structures, and Universal Sigma Model

Courant algebroids on a manifold $M$ are vector bundles $E\to M$ equipped with a symmetric fiber metric, an anchor map $\rho:E\to TM$, and a Dorfman (or Courant) bracket on $\Gamma(E)$, subject to a set of axioms encoding compatibility between bracket, anchor, and metric. The standard (exact) Courant algebroid is $E=TM\oplus T^*M$ with anchor $\rho(v\oplus\alpha)=v$, metric $$\langle v\oplus\alpha, w\oplus\beta\rangle=\langle v,\beta\rangle+\langle w,\alpha\rangle$$, and (possibly $H$-twisted) Dorfman bracket: $$[v\oplus\alpha, w\oplus\beta]_H = ([v,w], \mathcal{L}_v\beta - \iota_w d\alpha - \iota_w\iota_v H)$$ for closed $H\in\Omega^3(M)$.

A Dirac structure $D\subset E$ is a maximally isotropic, involutive subbundle, i.e., $\langle s, s'\rangle_E=0$ for $s,s'\in\Gamma(D)$ and $[s,s']_E\in\Gamma(D)$. The Dirac structure may arise as the graph of a Poisson bivector, closed 2-form, or an embedding of an action Lie algebroid (Chatzistavrakidis et al., 2017, Chatzistavrakidis et al., 2016).

The universal ungauged 2D sigma model is constructed using auxiliary 1-form fields valued in $E$, allowing for all gaugings: $$S_\text{univ}[X, V, W] = \int_\Sigma \frac{1}{2}g_{ij}(X)\, \mathcal{D}X^i \wedge \star \mathcal{D}X^j + \int_{\widehat\Sigma} X^*H + \int_\Sigma W_i \wedge (dX^i - \frac{1}{2}V^i)$$ where $\mathcal{D}X^i = dX^i - V^i$ [$1705.05007$].

2. From Universal Action to Gauged Courant Sigma Models

To gauge a foliation or symmetry, the auxiliary 1-form fields $V\oplus W$ are constrained to take values in a Dirac structure $D \subset E$. The constraint is implemented by choosing a bundle morphism $\sigma: E_{\text{gauge}}\to E$ whose image is $D$. In local coordinates, this yields gauge fields $A^a\in\Omega^1(\Sigma)$ and the decomposition: $$V^i = \rho^i_a(X) A^a\,,\quad W_i = \theta_{ai}(X) A^a$$ with $\rho^i_a$ the anchor and $\theta_{ai}$ specifying the embedding. Substituting into the universal action gives the GCSM action: $$S_D[X,A] = \int_\Sigma \frac{1}{2}g_{ij}(X) D X^i\wedge\star D X^j + \int_{\widehat\Sigma} X^*H + \int_\Sigma A^a\wedge\theta_{ai}(X)dX^i + \frac{1}{2}\gamma_{ab}(X)A^a\wedge A^b$$ where $D X^i = dX^i - \rho^i_a(X)A^a$, $$\gamma_{ab} = \rho^i_a \theta_{bi} = -\rho^i_b \theta_{ai}$$ (Chatzistavrakidis et al., 2017, Kotov et al., 2014, Chatzistavrakidis et al., 2016).

The resulting theory exhibits local gauge symmetry whose algebra closes under the induced Dirac structure bracket, generalizing the minimal coupling of group-valued gauge theory to arbitrary (possibly singular or non-integrable) foliations.

3. Gauge Symmetries, Connections, and Consistency

The infinitesimal gauge symmetry parameters $\varepsilon^a e_a \in \Gamma(X^*D)$ generate

$\delta_\varepsilon X^i = \rho^i_a(X)\varepsilon^a$

$$\delta_\varepsilon A^a = d\varepsilon^a + C^a_{bc}(X)A^b\varepsilon^c + \omega^a_{b\,i}(X)\varepsilon^b DX^i + \phi^a_{b\,i}(X)\varepsilon^b \star DX^i$$

with $C^a_{bc}$ the structure functions of $[e_a,e_b]_D$, and $\omega,\phi$ encode two connections $\nabla^\pm$ on $D$ entering the non-minimal coupling. Gauge invariance of the action is ensured by compatibility conditions on the background data:

• Courant metric and bracket structure
• Anchor invariance and closure,
• Flatness of $\nabla^\pm$ along $D$ directions,
• Twisted moment map conditions relating $H$-twist to the Dirac structure (Chatzistavrakidis et al., 2016, Chatzistavrakidis et al., 2017, Severa et al., 2019).

These connections encode deformations from the Killing condition; the generators need not correspond to isometries. The gauge algebra is always first-class and closes under the Dirac (restricted Courant) bracket.

4. Examples and Special Cases

Several classes of GCSMs demonstrate the general construction:

• Twisted Poisson sigma models: For a bivector $\Pi$ satisfying $$\frac{1}{2}[\Pi,\Pi]_{SN} = \langle\Pi^{\otimes 3}, H\rangle$$, $D = \mathrm{graph}(\Pi)$. The action recovers the $H$-twisted Poisson sigma model,

$$S[X,A] = \int_\Sigma A_i\wedge dX^i + \frac{1}{2} \Pi^{ij}(X)A_i\wedge A_j + \frac{1}{2}g_{ij}(X)DX^i\star DX^j + \int_{\widehat\Sigma} X^*H$$

(Chatzistavrakidis et al., 2017, Chatzistavrakidis et al., 2016).

• Graph of a closed 2-form: $D = \mathrm{graph}(B_0)$ with $dB_0=0$ gives a coupling of gauge fields via contraction with $B_0$ (Chatzistavrakidis et al., 2017).
• Standard group gauging: The Dirac structure is the image of the action Lie algebroid; $D$ is generated by $k_a \oplus 0$ for Killing fields $k_a$, yielding the conventional gauged sigma model (Chatzistavrakidis et al., 2017, Kotov et al., 2014).
• Transitive Courant algebroids: For non-Abelian gaugings, $E=TM\oplus \mathfrak{g} \oplus T^*M$ supports a Dorfman bracket and compatible pairing, extending standard group-based gaugings to more general targets (Chatzistavrakidis et al., 2020).
• Jacobi sigma models: GCSMs encompass H-twisted Jacobi structures, yielding models controlled by a bivector and an additional vector field, with a corresponding BV–AKSZ description (Chatzistavrakidis et al., 2020).

5. AKSZ/BV Quantization and Higher-Dimensional Generalizations

GCSMs admit a natural description in the AKSZ-BV framework. The Courant sigma model is constructed as the AKSZ model with target a degree 2 QP-manifold, e.g., $T^*[2]E[1]$. The action includes pairing, anchor, Dorfman bracket, and $H$-twist terms: $$S = \int_{T[1]\Sigma} \langle \mathrm{ev}^*\vartheta, D\mathrm{ev}\rangle + \mathrm{ev}^*\Theta$$ where $\Theta$ encodes the Courant algebroid structure, and its self-bracket $\{\Theta,\Theta\}=0$ is the classical master equation. Gauging introduces additional AKSZ superfields, associated to Lie algebroid or Courant algebroid bundles acting on the target; consistency requires flatness conditions and modified Bianchi identities for fluxes (Ikeda, 31 Jan 2026, Chatzistavrakidis et al., 2023). The quantization is implemented via BRST/BV master action, with terms reflecting the Courant structure and its torsions and curvatures.

6. Dirac Structures, Compatibility Conditions, and Hamiltonian Reduction

The possibility and type of GSCSM are governed by the existence of Dirac structures in the relevant Courant algebroid, including "small" (non-maximal) Dirac structures. The presence of a transverse generalized metric compatible with a Dirac structure ensures existence of both the requisite connections for gauge invariance and an appropriate Hamiltonian reduction on the phase space, directly relating the Lagrangian and Hamiltonian formalisms (Severa et al., 2019). The reduction to the symplectic quotient is controlled by the Dirac structure, its orthogonal complement, and the induced metrics.

7. Unification, Extensions, and Distinguished Features

GCSMs unify all consistent 2D gaugings—classical, Poisson, WZW-type, or more general Dirac-linearizable models—under a common Dirac-constraint paradigm. Key advances include:

• The ability to gauge singular foliations and non-group actions, determined by Dirac structures, not necessarily integrable to group actions.
• The universality of the approach: every consistent gauging arises as a constrained pullback from the universal Dirac sigma model functional by a Lie algebroid morphism into a Dirac structure (Kotov et al., 2014, Chatzistavrakidis et al., 2017).
• Gauge-invariance conditions are milder than standard Killing symmetry, controlled instead by connection data on the Dirac structure.
• AKSZ/BV constructions for higher dimensions and models with boundaries (allowing for homotopy moment maps, fluxes, and more sophisticated boundary conditions) (Ikeda, 31 Jan 2026, Chatzistavrakidis et al., 2023).
• Bridging the reduction from 3D Courant sigma models to effective 2D Poisson groupoid/sigma models through coisotropic reduction and the interplay of Lie bialgebroids and Courant geometry (Cabrera et al., 2022).

GCSMs thus serve as a comprehensive framework for gauged 2D sigma models and their generalizations, captured naturally within the formalism of generalized geometry.

---

References:

(Chatzistavrakidis et al., 2017, Kotov et al., 2014, Chatzistavrakidis et al., 2016, Chatzistavrakidis et al., 2020, Severa et al., 2019, Cabrera et al., 2022, Ikeda, 31 Jan 2026, Chatzistavrakidis et al., 2023)