On a Batalin-Vilkovisky operator generating higher Koszul brackets on differential forms (Q829956)

Consider first a Poisson bivector field \(P\) on a manifold \(M\). It is well known that it induces a Lie algebroid structure on the cotangent bundle \(T^*M\), in particular a Lie algebra bracket on the one-forms, which extends to the Koszul bracket on all differential forms \(\Omega(M)\). Further, the Koszul bracket is generated by a differential operator, namely the commutator \([d,i(P)]\) of the de Rham differential and the contraction with \(P\). Here ``generated'' is meant in the following sense: the Koszul bracket is the failure of \([d,i(P)]\) to be a derivation of the wedge product. The first part of the paper at hand extends these results from Poisson structures to \(P_{\infty}\)-structures, defined as even functions on the parity-reversed cotangent bundle \(\Pi T^*M\), where \(M\) is now allowed to be a supermanifold. In that case, by earlier results of H. Khudaverdian and Th. Voronov, one obtains a sequence of higher Koszul brackets on \(\Omega(M)\), satisfying higher Jacobi identities. As \(\Omega(M)\) consists of the functions on \(\Pi TM\), this gives rise to an \(S_{\infty}\)-structure on \(\Pi TM\). The paper shows that the higher Koszul brackets are ``generated'' by a formal \(\hbar\)-differential operator \(\Delta_P\) on \(\Omega(M)\), via a derived brackets formula (Theorem 3.1). Even more, this operator is a BV operator for the \(S_{\infty}\)-structure on \(\Pi TM\), in particular it squares to zero (Theorem 3.3). In the case of classical Poisson structures, the passage from the Lie algebroid structure on \(T^*M\) and the linear Poisson bivector field on its dual \(TM\) (which is closely related to the Koszul bracket) is obtained by the Mackenzie-Xu transformation (also known as Legendre transformation) between the corresponding parity-reversed bundles; further, the pair \((T^*M,TM)\) forms a Lie bialgebroid. Similarly, a \(P_{\infty}\)-structure on a supermanifold \(M\) gives rise to a cotangent \(L_{\infty}\)-algebroid, which actually forms a \(L_{\infty}\)-bialgebroid. However the symmetry between data on the cotangent and tangent bundle is not preserved by a naive extension of the Mackenzie-Xu transformation: the obstruction is the modular class of the \(P_{\infty}\)-structure. The second half of the paper is devoted to establishing a symmetry in a refined way, namely by means of a Fourier-transform between spaces of densities.