The standard cohomology of regular Courant algebroids (Q2105746)

This paper aims to introduce a minimal model of regular Courant algebroids, exploiting it in order to compute the standard cohomology of regular Courant algebroids. The concept of Courant algebroids started with Courant and Weinstein's study on Dirac manifolds via what is known as Courant bracket [\textit{T. J. Courant}, Trans. Am. Math. Soc. 319, No. 2, 631--661 (1990; Zbl 0850.70212); \textit{T. Courant} and \textit{A. Weinstein}, in: Actions hamiltoniennes de groupes. Troisième théorème de Lie, Sémin. Sud-Rhodan. Géom. VIII, Lyon/France 1986, Trav. Cours 27, 39--49 (1988; Zbl 0698.58020)], which eventually led to the definition of Courant algebroids by \textit{Z. Liu} et al. [J. Differ. Geom. 45, No. 3, 547--574 (1997; Zbl 0885.58030)]. Roughly speaking, a \textit{Courant algebroid} \((E,g,\rho,\circ)\) is a pseudo-Euclidean vector bundle \((E,g)\) over a smooth manifold \(M\) together with a morphism of vector bundles \(\rho:E\rightarrow TM\), called \textit{anchor}, and a Leibniz algebra structure \(\circ\), called \textit{Dorfman bracket} [\textit{I. Ya. Dorfman}, in: Nonlinear evolution equations: integrability and spectral methods, Proc. Workshop, Como/Italy 1988, Proc. Nonlinear Sci., 425--431 (1990; Zbl 0717.58026)] on the section space \(\Gamma(E)\), satisfying some compatibility conditions. There is another approach from differential graded (dg for short) geometry [\textit{D. Roytenberg}, Contemp. Math. 315, 169--185 (2002; Zbl 1036.53057); \textit{P. Ševera}, Trav. Math. 16, 121--137 (2005; Zbl 1102.58010)]. Given a pseudo-Euclidean sector bundle \((E,g)\), the pair \((E[1],g)\) becomes a degree \((-2)\) Poisson manifold. By choosing a metric connection \(\nabla\) on \((E,g)\), the \textit{minimal symplectic realization} \(\mathcal{E}\) of \((E[1],g)\) is identified with an exact degree \(2\) symplectic manifold \((T^{\ast}[2]M\oplus E[1],\omega_{g},\nabla)\) [\textit{M. Rothstein}, Lect. Notes Phys. 375, 331--343 (1991; Zbl 0747.58010)]. The associated \((-2)\) Poisson algebra \((\mathcal{C}^{\infty}(\mathcal{E}),\left\{ -,-\right\} )\) is known as the \textit{Rothstein algebra}, where \(\left\{ -,-\right\} \) denotes the degree \((-2)\) Poisson bracket known as the big bracket [\textit{Y. Kosmann-Schwarzbach}, in: Mathematical aspects of classical field theory. Proceedings of the AMS- IMS-SIAM joint summer research conference, held in Seattle, Washington, USA, July 20-26, 1991. Providence, RI: American Mathematical Society. 459--489 (1992; Zbl 0847.17020)]. A Courant algebroid structure on \((E,g)\) is encoded by a degree \(3\) function \(\Theta\) (known as the \textit{generating Hamiltonian}) on \(\mathcal{E}\), subject to the classical master equation \(\left\{ \Theta,\Theta\right\} =0\). The anchor \(\rho\) and the Dorfman bracket \(\circ\) of \(E\) are recovered from \(\Theta\) via a Vinogradov bracket [\textit{A. Cabras} and \textit{A. M. Vinogradov}, J. Geom. Phys. 9, No. 1, 75--100 (1992; Zbl 0748.58008)]. This approach leads to the definition of the standard complex and cohomology of a Courant algebroid \(E\) [\textit{D. Roytenberg}, Contemp. Math. 315, 169--185 (2002; Zbl 1036.53057)]. The Hamiltonian vector field \(X_{\Theta}\) \(:=(\Theta,-)\) is a homological vector field on the graded manifold \(\mathcal{E}\), defining a differential \(d_{E}:=X_{\Theta}\) on the algebra \(\mathcal{C}^{\infty}(\mathcal{E})\) of functions on \(\mathcal{E}\). The cochain complex \ is called the \textit{standard complex} of , and the cohomology \ is called the \textit{standard cohomology} of the Courant algebroid \(E\). It is very difficult to compute the standard cohomology \(H_{\mathrm{CE} }^{\bullet}(E)\) of a Courant algebroid \(E\). \textit{M. Stiénon} and \textit{P. Xu} [C. R., Math., Acad. Sci. Paris 346, No. 3--4, 193--198 (2008; Zbl 1132.22004)] defined the naive complex \((C_{\mathrm{naive} }^{\bullet}(E),d)\) of a Courant algebroid \(E\), establishing that the corresponding degree \(1\) naive cohomology \(H_{\mathrm{naive}}^{1}(E)\) is isomorphic to the standard cohomology \(H_{\mathrm{st}}^{1}(E)\). \textit{G. Ginot} and \textit{M. Grützmann} [J. Symplectic Geom. 7, No. 3, 311--335 (2009; Zbl 1209.53063)] established that the naive cohomoloy of any transitive Courant algebroid is isomorphic to its standard cohomology, which was first conjectured by \textit{M. Stiénon} and \textit{P. Xu} [C. R., Math., Acad. Sci. Paris 346, No. 3--4, 193--198 (2008; Zbl 1132.22004)]. Nevertheless, for general Courant algebroids, the two cohomology theories are quite distinct. Another achievement of \textit{G. Ginot} and \textit{M. Grützmann} [J. Symplectic Geom. 7, No. 3, 311--335 (2009; Zbl 1209.53063)] is the construction of a Leray-type spectral sequence, called the naive ideal spectral sequence, which is used to calculate the standard cohomology \(H_{\mathrm{st}}^{\bullet }(E)\) under the asumption that \(E\) has a split base. This paper focuses on regular Courant algebroids. A Courant algebroid \(E\) is said to be \textit{regular} if its anchor \(\rho\) has constant rank, where \(F:=\rho E\subseteq TM\) defines an integrable distribution, called the \textit{characteristic distribution} of \(E\). The synopsis of the paper goes as follows. \begin{itemize} \item[\S 2] collects some standard facts of Courant algebroids and their formulation via degree \(2\) symplectic manifolds with generating Hamiltonian functions [\textit{D. Roytenberg}, Contemp. Math. 315, 169--185 (2002; Zbl 1036.53057); Lett. Math. Phys. 79, No. 2, 143--159 (2007; Zbl 1125.81040)] after [\textit{M. Grützmann} et al., J. Math. Pures Appl. (9) 154, 67--107 (2021; Zbl 1481.53105)]. \item[\S 3] defines the regular dg algebra of a regular Courant algebroid, constructing a semifull algebra contraction from the regular dg algebra onto the Chevalley-Eilenberg dg algebra of the ample Lie algebroid. Given a regular Courant algebroid \((E,g,\rho,\circ)\), some foundational facts and notations are inherited from the joint work [Zbl 1273.53067] of the second author with Stiénon and Xu. To see the relation between the regular Courant algebroid \(E\) and its \textit{ample Lie algebroid} \(A_{E}\) (called the associated Lie algebroid in [\textit{Z. Chen} et al., J. Symplectic Geom. 11, No. 1, 1--24 (2013; Zbl 1273.53067)]) sharing the same characteristic distribution \(F\subseteq TM\) with that of \(E\), the authors consider the graded manifold \[ \mathcal{R}=F^{\ast}[2]\oplus E[1] \] It turns out that the standard differential \(d_{E}\) on \(E\) defines a differential on the algebra \(C^{\infty}(\mathcal{R})\) of smooth functions on \(\mathcal{R}\) as well. The pair \((C^{\infty }(\mathcal{R}),d_{E})\) is called the \textit{regular dg algebra} of , being a dg subalgebra of the Rothstein dg algebra \((C^{\infty}(\mathcal{E}),d_{E})\). It is established (Theorem 3.14) that there exists a semifull algebra contraction from the regular dg algebra of \(E\) onto the Chevalley-Eilenberg dg algebra of the ample Lie algebroid \(A_{E}\). Therefore, the corresponding cohomology \(H^{\ast }(C^{\infty}(\mathcal{R}),d_{E})\), called the \textit{regular cohomology} of \(E\), is isomorphic to the Chevalley-Eilenberg cohomology \(H_{\mathrm{CE}}^{\ast}(A_{E})\) of the ample Lie algebroid \(A_{E}\), which is in turn isomorphic to the naive cohomology of \(E\). \item[\S 4] first establishes the existence of a minimal model for each regular Courant algebroid by extending the semifull algebra contraction for the regular dg algebra to a semifull algebra contraction for the whole Rothstein dg algebra. It is then established that the minimal model \(\mathcal{M}_{E}\) of \(E\) indeed admits a \(2\)-shifted derived Poisson manifold structure. The authors finally compute the standard cohomology of a regular Courant algebroid \(E\) by analyzing the Chevalley-Eilenberg-to-standard spectral sequence for the minimal model \(\mathcal{M}_{E}\) of \(E\). \item[\S 5] considers applications to three types of regular Courant algebroids. \begin{itemize} \item[1.] The first type consists of Courant algebroids with split base. \item[2.] The authors consider generalized exact Courant algebroids, i.e., those fitting into the following exact sequence of vector bundles over \(M\) \[ 0\rightarrow F^{\ast}\overset{\rho^{\ast}}{\rightarrow}E\underset{} {\overset{\rho}{\rightarrow}}F\rightarrow0 \] where \(F\subseteq TM\) is the characteristic distribution of \(E\), and \(F^{\ast }\) is its dual bundle. It was established in [\textit{G. Ginot} and \textit{M. Grützmann}, J. Symplectic Geom. 7, No. 3, 311--335 (2009; Zbl 1209.53063)] that generalized exact Courant algebroids with characteristic distribution \(F\) are classified up to isomorphisms by a degree \(3\) class \([C]\), called Ševera class in \(H_{\mathrm{CE}}^{3}(F)\). The explicit description of the standard cohomology of generalized exact Courant algebroids is obtained (Theorem 5.3). \item[3.] Regular Courant algebroids arising from regular Lie algebroids \((A,\rho_{A},)\) are studied. It is well known that the Whitney sum \(K:=A\oplus A^{\ast}\) admits regular Courant algebroid structure, where the cohomology class \([d_{T}]\in H_{\mathrm{CE}}^{3}(A_{E};B^{\ast}[2])\) is isomorphic to the cohomology class \([\omega]\in H_{\mathrm{CE}}^{2}(A;\mathrm{Hom} (B,K))\) introduced by Gracia-Saz in [] to characterize the obstruction to a local trivializability of \(A\) around any leaf \(L\) of \(F\), so that we obtain an explicit description of the standard cohomology of \(E=A\oplus A^{\ast}\), which is indeed isomorphic to the cohomology of the shifted cotangent bundle \(T^{\ast}[2](A[1])\) of the dg manifold \((A1,d_{A})\) arising from the regular Lie algebroid \(A\). \end{itemize} \item[Appendix A] gives a proof of Lemma 3.26, which plays a significant role in the proof of Theorem 3.14. \item[Appendix B] gives a proof of Proposition 4.8 claiming that for each regular Courant algebroid \(E\) there exists a canonical cohomology class \([d_{T}]\in H_{\mathrm{CE}}^{3}(A_{E},B^{\ast}[2])\) of total degree \(1\), which characterizes whether the Courant algebroid structure around each leaf \(L\) of \(F\) is determined by its restriction \(E\mid_{L}\), where the quotient bundle \(B:=TM/F\) is canonically a Lie algebroid \(A_{E}\)-module, the Chevalley-Eilenberg differential \(d_{\mathrm{CE}}\) of which determines a homological vector field on the graded manifold \(A_{E}[1]\oplus B^{\ast}[2]\). \end{itemize} To conclude this review, I must give some comments on related works. First, I should note that many insights and clues for this paper have already appeared in the authors' [\textit{X. Cai} et al., Commun. Algebra 48, No. 8, 3315--3341 (2020; Zbl 1461.17004)] on cohomology of hemistrict Lie \(2\)-algebras, i.e. \(2\)-term dg Leibniz algebras whose bracket is skew-symmetric up to homotopy. \textit{M. Cueca} and \textit{R. A. Mehta} [Commun. Math. Phys. 381, No. 3, 1091--1113 (2021; Zbl 1465.53039)] found out that the differential of the Keller-Waldmann construction [\textit{F. Keller} and \textit{S. Waldmann}, J. Pure Appl. Algebra 219, No. 8, 3391--3426 (2015; Zbl 1315.53103)] of a canonical dg Poisson algebra from a given Courant algebroid satisfies a Cartan-type formula, providing a way to relate Courant algebroid cohomology to classical de Rham or Chevalley-Eilenberg cohomology theories and leading to a thorough study of characteristic classes of Courant algebroids. Finally, I would like to mention the work [J. Math. Pures Appl. (9) 116, 1--39 (2018; Zbl 1395.53014)] of \textit{M. Jotz Lean} on Dorfman connections and their applications to linear splittings of the standard Courant algebroids over vector bundles, which could be related to this paper.