Topological Hochschild homology and integral \(p\)-adic Hodge theory (Q2418568): Difference between revisions

Let \(C\) denote a \(p\)-completed algebraic closure of \(\mathbb{Q}_p\), and let \(\mathcal{O}_C\) denote its ring of integers. The paper under review is a natural continuation of the article by \textit{B. Bhatt} et al. [Publ. Math., Inst. Hautes Étud. Sci. 128, 219--397 (2018; Zbl 1446.14011)], where the authors construct a cohomology theory for smooth, proper formal schemes \(\mathfrak{X}\) over \(\mathcal{O}_C\), with coefficients in Fontaine's period ring \(A_{\textrm{inf}}\). The theory, aptly named \(A_{\textrm{inf}}\)-cohomology, is constructed in the prior article using perfectoid spaces, the \(L\eta\) operator, and other tools more in the realm of arithmetic geometry. The article under review gives an alternative construction using methods of homotopy theory, largely utilizing the language of \(\infty\)-categories and spectra. Moreover, it builds a more general theory with Breuil-Kisin modules as output. As mentioned by the authors in the introduction, this more homotopical viewpoint was historically the first construction of the \(A_{\textrm{inf}}\)-cohomology theory. The same theory would find a third construction in the later work of \textit{B. Bhatt} and \textit{P. Scholze} [``Prisms and prismatic cohomology'', Preprint, \url{arXiv:1905.08229}] as a particular instance of prismatic cohomology, which acts as a universal \(p\)-adic cohomology theory by obtaining other well-known \(p\)-adic theories (e.g., de Rham, crystalline, etale) through functorial means. The most basic relationship between the \(A_{\textrm{inf}}\)-theory and topological Hochschild homology (\(\textrm{THH}\)) is roughly as follows: given a commutative ring \(R\), its topological Hochschild homology \(\textrm{THH}(R)\) is a spectrum equipped with a circle action. Let \(\textrm{TC}^{-}(R)\) denote the homotopy fixed points of the circle action on \(\textrm{THH}(R)\), and let \(\textrm{TC}^{-}(R;\mathbb{Z}_p)\) denote its \(p\)-completion. Let \(A\Omega_R\) denote the complex of \(A_\textrm{inf}\)-modules that yields the \(A_\textrm{inf}\)-cohomology theory for \(R\). The first statement is that there is an appropriate site (the quasi-syntomic site) relative to \(R\), whose objects are quasi-syntomic \(R\)-algebras \(S\), with a topology given by quasi-syntomic covers (where quasi-syntomic morally means syntomic, in a derived sense). Second, the association \(S \mapsto \pi_0(\textrm{TC}^{-}(S);\mathbb{Z}_p)\) is a sheaf of rings on a basis of this site, given by quasiregular semiperfectoid rings (which are in the authors' own words, ''essentially ... quotients of perfectoid rings by regular sequences''). Finally, for \(R\) the \(p\)-completion of a smooth \(\mathcal{O}_C\)-algebra, the authors prove that \(A\Omega_R \simeq R\Gamma(R,\pi_0(\textrm{TC}^{-}(-;\mathbb{Z}_p))\) functorially as \(E_\infty\)-\(A_{\textrm{inf}}\)-algebras, and equivariant with respect to the Frobenius action. Moreover, using the homotopy fixed point spectral sequence, the authors show that \(\pi_0(\textrm{TC}^{-}(-;\mathbb{Z}_p))\) comes equipped with a natural abutment filtration. This filtration manifests itself in cohomology as the Nygaard filtration of prismatic cohomology, which agrees with the classical Nygaard filtration when the theory is used to obtain crystalline cohomology. The authors moreover show that there is a filtration of \(\textrm{TC}^{-}(-;\mathbb{Z}_p)\) itself, resembling the motivic filtration on algebraic \(K\)-theory whose graded pieces yield motivic cohomology, constructed using flat descent to semiperfectoid rings. The paper ultimately unveils deep relationships between \(p\)-adic Hodge theory and algebraic \(K\)-theory, using results of the former to obtain new results of the latter, and provides an invaluable (homotopical) perspective on prismatic cohomology.