Geometric Hodge filtered complex cobordism (Q6076333): Difference between revisions

Given a topological rationally even spectrum \(E\) and a prime \(p\), \textit{M. J. Hopkins} and \textit{G. Quick} [J. Topol. 8, No. 1, 147--183 (2015; Zbl 1349.32009)] have defined Hodge filtered \(E\)-cohomology groups with twist \(p\) for complex manifolds \(X\). For each \(n\), these are \(E^n_{\mathcal D}(p) (X) = \mathrm{Hom}_{hoSp(\mathbf{ sPre_*})}(\Sigma^\infty(X_+), \Sigma^nE_{\mathcal D}(p))\), where \(\mathbf{ sPre_*}\) is the homotopy category of presheaves of spectra on \(\mathbf{Man}_{\mathbb{C}}\), the category of complex manifolds and holomorphic maps. \(E_{\mathcal D}(p)\) is the homotopy pullback of \(H(F^p\mathcal{A}^*(\mathcal{V}_*)) \rightarrow H(\mathcal{A}^*(\mathcal{V}_*)) \leftarrow \mathrm{sing}(E)\) as maps of presheaves of spectra on \(\mathbf{Man}_{\mathbb{C}}\). Here, \(\mathcal{A}^*(\mathcal{V}_*)\) denotes the space of smooth forms on \(\mathcal{V}_* = E_* \otimes_{\mathbb{Z}} \mathbb{C}\), the \((F^p\mathcal{A}^*(\mathcal{V}_*))\) represent the \textit{Hodge filtration on forms} (see Definition 2.8), and \(\mathrm{sing}(E)\) is the spectrum of singular simplicial sets for \(E\). The \(E^n_{\mathcal D}(p) (X)\) are an example of a Hodge filtered extension of a rationally even cohomology theory, in line e.g. with complex analytic Deligne cohomology or Karoubi's multiplicative \(K\)-theory. The purpose of this paper is to obtain a more concrete and geometric description of the \(E^n_{\mathcal D}(p) (X)\) whenever \(E=MU\), the complex cobordism spectrum. The authors define thus their \textit{geometric Hodge filtered complex cobordism groups} \(MU^n(p)(X)\), but do it not simply for complex manifolds but for smooth manifolds (together with a descending filtration \(F^*\) of \(\mathcal{A}^*\) on the manifold). In their words, this generality makes it easier to work with products of a complex manifold with a smooth one, an advantage for future applications of the results. On the other hand, it is reasonable to expect that Hopkins-Quick's definition of the \(E^n_{\mathcal D}(p) (X)\) might be extended to the category of smooth manifolds. The \(MU^n(p)(X)\) are obtained in Definition 2.18 as a quotient between \(ZMU^n(p)(X)\), the group of \textit{Hodge filtered cycles} of Definition 2.13, and \(BMU^n(p)(X)\), the group of \textit{Hodge filtered cobordism relations} stemming from Proposition 2.17 (and using the so-called \textit{nullbordant} cycles). Some of the properties of the \(MU^n(p)(X)\) are then presented, and special mention should be given to the long exact sequence in Theorem 2.21, and to the attempt at an axiomatic interpretation of Hodge filtered cohomology theories in Section 2.9. For the main result of the paper, Theorem 1.2, the authors consider again just complex manifolds. This result states that, given one such manifold \(X\) with a Hodge filtration, the two theories agree, that is, there is an isomorphism between the cohomology groups \(MU^n_{\mathcal D}(p) (X)\) and \(MU^n(p)(X)\) (for each \(n\) and \(p\)). The proof of this theorem uses a form of Pontryagin-Thom construction that takes into account the geometric data given by the Hodge filtration. The isomorphism is in fact constructed as a composition of two, and is achieved by means of an intermediate theory \(MU_{hs}(p)\) which shares the homotopy-theoretic flavor of \(MU^n_{\mathcal D}(p) (X)\) but is more accessible. This new theory (for smooth manifolds with the Hodge filtration) appears as Definition 3.15 and uses Mathai-Quillen forms. Section 4.2 shows that the new model is represented by a presheaf of spectra which fits into a homotopy pullback similar to the one defining \(MU^n_{\mathcal D}(p)\), and which permits the direct comparison of the two models in Theorem 4.7, leading to the isomorphism \(MU^n_{\mathcal D}(p) (X) \cong MU^n_{hs}(p) (X)\) of Theorem 4.9. Finally, Section 5 relates the \(MU^n_{hs}(p) (X)\) to the \(MU^n(p)(X)\) by a geometric Pontryagin-Thom map that induces the isomorphism of Definition 5.7, later allowing for Theorem 5.9, which gives \(MU^n_{hs}(p) (X) \cong MU^n(p)(X)\) for each \(X\), \(n\) and \(p\). This result, combined with Theorem 4.7, then gives the main result of Theorem 1.2.