A differential graded model for derived analytic geometry (Q2281339)

Derived analytic geometry has seen a rise of interest in recent years. It has lagged behind the development of derived algebraic geometry partly due to a number of technical obstacle to the foundations. In fact, there is not quite a consensus what foundations to use and this paper contributes a very interesting new model. The basic local object in any theory of derived analytic geometry is the ring of functions on a derived Stein space (or affinoid, in the rigid analytic setting). For this Pridham considers commutative dg algebras equipped with an \emph{entire functional calculus} (EFC) on their degree 0 part. To understand the meaning of an EFC, let us recall the basic technical difficulty in describing analytic geometry using algebra. It is not enough to consider polynomals in the analytic function (which arise directly from any algebra structure) but there are also analytic functions whose arguments are analytic functions. Thus the ring of functions on a derived analytic space has an action of holomorphic functions. One approach based on analytic rings, see [\textit{E. Dubuc} and \textit{G. Taubin}, Cah. Topologie Géom. Différ. Catégoriques 24, 225--265 (1983; Zbl 0575.32004)] and is used by \textit{J. Lurie} [``Derived algebraic geometry V: structured spaces'', Preprint, \url{arXiv:0905.0459}] and \textit{M. Porta} and \textit{T. Y. Yu} [``Derived non-Archimedean analytic spaces'', \url{arXiv:1601.00859}]. It uses a holomorphic functional calculus, based on the action of partial holomorphic functions. The work under consideration in contrast is based on the entire functional calculus of [\textit{A. Yu. Pirkovskii}, J. Noncommut. Geom. 9, No. 1, 215--264 (2015; Zbl 1314.32012)], where only entire holomorphic functions appear. Formally, Pridham considers dg algebras with EFC, which are dg algebras for a certain Fermat theory, similar to the construction of dg \(\mathcal C^\infty\)-rings as they appear in derived differential geometry. This simplifies the theory (as less data needs to be considered) and there is little loss of information: The author provides comparison results with important constructions in structured spaces. In particular the category of derived Stein spaces (in the sense of Lurie/Porta) embeds fully faithfully in the opposite of the category of complex dg algebras with EFC. As opposed to the very general theory of structured spaces, the constructions in this paper make explicit use of the theory of Stein spaces (respectively affinoid spaces), notably classical embedding theorems. The article develops the foundational constructions in a theory of derived analytic geometry, in particular the cotangent complex and an analytification functor. An important technical ingredient introduced by the author is a model structure on a subclass of dg algebras with EFC, whose weak equivalences are quasi-isomorphisms and such that the EFC-algebras of holomorphic funcions on Stein submanifolds of \(\mathbb C^n\) are cofibrant objects. This model category is built starting from a class of maps between dg algebras with EFC that come from open immersions of Stein spaces (so-called \emph{finite localizations}). An alternative model structure may be constructed starting with a class of maps coming from local biholomorphisms (étale maps). In the last section, the gluing of derived analytic spaces, Deligne Mumford and Artin stacks from the opposite category of dg algebras with ETF is discussed. It is an application of the theory of hypergroupoids as developped in [\textit{J. P. Pridham}, Adv. Math. 238, 184--245 (2013; Zbl 1328.14028)]. The starting point is that any \(\mathbb C\)-analytic space has an open cover by Stein spaces, and thus the main algebraic tool are localizations of dg algebras with EFC. Most of the results in this work apply in derived analytic geometry over \(\mathbb C\) and over non-archimedean fields, but there are some caveats when applying this approach to rigid analytic geometry. The main difficulty arises with gluing. As admissible open immersions in rigid analytic geometry do not correspond to localisation of their algebras of functions one has to restrict to partially proper spaces or consider derived dagger analytic spaces (rigid spaces with overconvergent structure sheaf) as developped in [\textit{E. Grosse-Klönne}, J. Reine Angew. Math. 519, 73--95 (2000; Zbl 0945.14013)]. In this regard the approach in this article is not as powerful as that of Lurie and Porta and Yu.