Circle-equivariant classifying spaces and the rational equivariant sigma genus (Q658355)

The main objective of this paper is to write down, rationally at least, an equivariant refinement of the sigma orientation from string manifolds to elliptic cohomology theories. It is worth taking a few paragraphs to walk back through the intellectual antecedents of the theorem, as the background will emphasize the heft of this result. In general, a genus is a ring homomorphism out of a cobordism ring. In many examples the target ring is the ring of coefficients of some cohomology theory \(E^\ast\) and the genus is defined by using an \(E^\ast\)-orientation on the tangent bundle of the manifolds in question. However, the sigma genus (or sigma orientation) can be defined in a different manner. Using anomaly cancellation and string theory, Witten assigned to appropriate manifolds \(M\) a power series defined as the index of an operator known as ``the supercharge of the supersymmetric nonlinear sigma model'' on the free loop space on \(M\). Using the evident action of the circle on the free loop space, Witten then argued that this power series was the \(q\)-expansion of a modular form. The basic reference is [\textit{E. Witten}, Commun. Math. Phys. 109, 525--536 (1987; Zbl 0625.57008)]. This inspired a whole generation of algebraic topologists, who set about defining the elliptic cohomology theories, the relevant orientations, and making the entire story homotopy theoretic. An early and often cited paper was by \textit{P. S. Landweber, D. C. Ravenel} and \textit{R. E. Stong} [Contemp. Math. 181, 317--337 (1995; Zbl 0920.55005)]; a breakthrough paper on the orientations themselves was \textit{M. Ando, M. J. Hopkins} and \textit{N. P. Strickland} [Invent. Math. 146, No. 3, 595--687 (2001; Zbl 1031.55005)]. In thinking about the possible targets for this orientation, Hopkins, Miller, and their coworkers were led to the entire theory of topological modular forms, which is itself a centerpiece of the recently emerged field of derived algebraic geometry. Beautiful as this is, many aspects of this field remain incomplete. First, one would surely want to see more of the original physics; \textit{K. Costello} has a recent paper on this question [Proceedings of the international congress of mathematicians (ICM 2010), Hyderabad, India, August 19--27, 2010. Vol. II: Invited lectures. Hackensack, NJ: World Scientific; New Delhi: Hindustan Book Agency. 942--959 (2011; Zbl 1242.58013)]. Second, almost all of the constructions use very difficult homotopy theory; one might hope for a construction that used the geometry and the physics. Finally, if there are geometric constructions, then they should respect any symmetries present in the geometry; that is, there should be equivariant versions. In this direction, the paper of \textit{I. Grojnowski} [Lond. Math. Soc. Lect. Note Ser. 342, 114--121 (2007; Zbl 1236.55008)] has been influential, but not definitive. An elliptic cohomology theory is a triple \((E^\ast,C,\phi)\) where \(E^\ast\) is \(2\)-periodic complex orientable cohomology theory (i.e., it can be given a natural theory of Chern classes), \(C\) is an elliptic curve over \(E^0\), and \(\phi\) is an isomorphism from the formal group of \(C\) to the formal group of \(E^\ast\). Elliptic curves are very rigid geometric objects and support an enormous amount of structure; in particular, there has been a great deal analysis of their finite subgroups. Using that analysis, the second author showed that if \(E^0\) is an algebra over the rationals, then the elliptic cohomology theory can be refined to a equivariant cohomology theory for the group \({\mathbb{T}} = S^1\). See [Topology 44, No. 6, 1213--1279 (2005; Zbl 1085.55002)]. This sets the stage for the paper under review. There is still work to be done, however. The first five sections of this paper are devoted to analyzing the algebraic topology of \(G\)-bundles and their classifying spaces; this is all needed to even define the source of the \({\mathbb{T}}\)-equivariant sigma orientation. The next four sections construct the orientation, and the final part of the paper takes a stab at a differential geometric interpretation. Despite its length, the paper is very densely written. It's imperative to have a copy of the Greenlees paper on \({\mathbb{T}}\)-equivariant elliptic cohomology on hand. This should not be regarded as a drawback, as the Greenlees paper is quite elegant and worth a visit on its own.