Chiral equivariant cohomology of a point: A first look (Q639380): Difference between revisions

This paper investigates the conformal vertex algebra structure given by the \(G\)-equivariant chiral cohomology \(\mathbf{H}_G^*(\mathbb{C})\) of a point, in the case where \(G=SU(2)\). Equivariant chiral cohomology was introduced by the author together with Lian and Song in a series of 3 papers (entitled ``Chiral Equivariant Cohomology'', [I: \textit{B. H. Lian} and \textit{A. R. Linshaw}, Adv. Math. 209, No. 1, 99--161 (2007; Zbl 1172.17021), II: \textit{B. H. Lian, A. R. Linshaw} and \textit{B. Song}, Trans. Am. Math. Soc. 360, No. 9, 4739--4776 (2008; Zbl 1221.17035) and III: ibid., Am. J. Math. 132, No. 6, 1549-1590 (2010; Zbl 1235.17022)]). It is a new equivariant cohomology (for manifolds with an action by a compact Lie group) with value in vertex algebras which contains the classical equivariant cohomology (which has value in commutative superalgebras with a differentiable \(G\)-action). Some familiarity with this previous work is indeed very helpful in order to study the paper under review. The chiral equivariant cohomology is a finer invariant; indeed, it was proved in [loc. cit., III] that, for any simple connected Lie group \(G\), there is a sphere with infinitely many actions of \(G\) which have distinct chiral equivariant cohomology. Nevertheless, the algebraic structure of \(\mathbf{H}_G^*(M)\) is largely unknown in general. The case where \(M\) is a point is very interesting since, as for the usual equivariant cohomology, there is a map \(\mathbf{H}_G^*(\mathbb{C})\to \mathbf{H}_G^*(M)\) (inducing the classical Chern-Weil map for equivariant cohomology) which makes \(\mathbf{H}_G^*(M)\) a module over \(\mathbf{H}_G^*(\mathbb{C})\) and defines universal characteristic classes in the chiral setting. If the group \(G\) is abelian, then the chiral equivariant cohomology of a point \(\mathbf{H}_G^*(\mathbb{C})\) has a trivial structure, that is, it is the abelian vertex algebra generated by the classical equivariant cohomology \(H^*_G(pt)\). However, new structures arise for semi-simple Lie groups and \(SU(2)\) is somehow the first non-trivial example. The main result of the paper under review is to construct a map \(\psi: U(\mathfrak{sl}_2)\to \mathbf{H}_{SU(2)}^*(\mathbb{C})\) from the universal enveloping algebra of the Lie algebra \(\mathfrak{sl}_2\). The author proves that \(\psi\) is injective and has images consisting entirely of non-classical elements. It is further conjectured that the data of \(\Psi\) and the classical equivariant cohomology \(H_{SU(2)}^*(pt)\) generates the whole chiral equivariant cohomology. The core of the paper deals with combinatorial techniques to study the algebraic structure of \(\mathbf{H}_{SU(2)}^*(\mathbb{C})\), which is a commutant vertex algebra, using a good filtration whose associated graded object is supercommutative, hence amenable by algebraic geometry techniques.