Dagger geometry as Banach algebraic geometry (Q907494): Difference between revisions

The aim of the work under review is to describe some techniques which are useful to extend dagger geometry over a valued field (as introduced by Große-Klönne) to the relative setting. It is a natural extension of the first author's Ph.D. thesis, where he introduced bornological algebraic geometry, and borrows vastly from there. The main new input is the interpretation of some category of bornological algebras in terms of Ind-objects in the category of Banach algebras, which is well-suited for the relative geometric settings the authors have in mind. The starting point of the paper (see \S 2.3) is that, in order to study ``relative geometry'' over a category \(\mathbb{C}\), one needs to look at the category \(\mathsf{Comm}(\mathbb{C})\) of commutative monoids; it is technically crucial to insist here that \(\mathbb{C}\) be a closed, symmetric, monoidal category with unit object. To get a nice category of modules over an object \(A\in\mathsf{C}\) it is better to require that \(\mathbb{C}\) is quasi-abelian. After the introduction, Section 2 contains background material mainly about Ind-objects and quasi-abelian categories, and the authors can introduce in Section 3 the main object of study. Fix a Banach ring \(R\); the authors prove in Propositions 3.15 and 3.17 that the category \(\mathsf{Ban}_R^A\) of Banach modules over \(R\) is a closed, symmetric, monoidal, quasi-abelian category: they can therefore apply the general machinery introduced in Section 2 to it. Many of the results in this section are taken from a preprint by F. Prosmans and the authors often refer to her work for most proofs. A technical but crucial point that should be mentioned here is that the authors distinguish, when \(R\) is non-Archimedean, between the categories \(\mathsf{Ban}_R^A\) and \(\mathsf{Ban}_R^{nA}\) which is the one commonly considered in non-Archimedean settings: as a matter of fact, most results of this paper work both when the base Banach ring \(R\) is Archimedean and when it is non-Archimedean. Section 3 continues with some analysis of the category \(\mathsf{Ind}(\mathsf{Ban}_R^{A})\) (\textit{resp.} \(\mathsf{Ind}(\mathsf{Ban}_R^{nA})\)) and then introduces the notion of bornological spaces (and algebras) culminating in the definition of the category \(\mathsf{CBorn}_k\) (see Definition 3.43) of complete bornological spaces over complete, non-trivially valued field \(k\). We call the attention on the useful commutative diagram on page 425 where relations about suitable subcategories of bornological and Banach \(k\)-algebras are summarized. An important result which the authors prove is Proposition 3.60 which shows that the category \(\mathsf{CBorn}_k\) is equivalent to the sub-category of \(\mathsf{Ind}(\mathsf{Ban}_k)\) of essential monomorphic objects: this allows them to use many results contained in the first author's Ph.D thesis. The connection with dagger geometry starts in Section 4. Fix a non-trivially valued complete field \(k\). The main classical reference for dagger geometry is \textit{E. Große-Klönne}'s work [J. Reine Angew. Math. 519, 73--95 (2000; Zbl 0945.14013)] whose underlying idea is to perform (rigid) analytic geometry replacing classical Tate algebras with dagger \(k\)-algebras, composed of \textit{overconvergent} functions; then the usual yoga of localizing at suitable Weierstraß or rational domains can start, leading to a \(G\)-topology on the category of (spectra of) dagger \(k\)-algebras. What the authors do here is to look at a dagger \(k\)-algebra \(A\) as an object of \(\mathsf{Comm}(\mathsf{Ind}(\mathsf{Ban}_k))\) (see Definition 4.8, to be compared with \S1.2 in Große-Klönne's work) and they define the category \(\mathsf{Afnd}_k\) of dagger affinoid \(k\)-algebras; the remark preceding Theorem 4.9, observing that to any dagger affinoid algebra (seen as an Ind-object) one can attach a meaningful underlying ring, can be helpful in what follows. By endowing the category of (Banach) spectra of dagger affinoid \(k\)-algebras with a suitable \(G\)-topology, they can finally define in Definition 4.15 the category of dagger affinoid spaces over \(k\). Probably the most relevant result of this section is Theorem 4.20 showing that dagger affinoid spaces satisfy Gerritzen-Grauert's result on locally closed immersions, Tate's acyclicity and Kiehl's result on coherent sheaves. Section 5 contains the most relevant geometric results, and focuses to a large extent on describing (open) immersions in the category of dagger affinoid spaces over \(k\). The most relevant result is probably Theorem 5.15 showing that the \(G\)-topology introduced here on a dagger affinoid space is compatible with the formal homotopy Zariski topology with covers given by dagger affinoid algebras. Although most of the material in Sections 2 and 3 is developed over an arbitrary base Banach ring \(R\), the hard results about dagger affinoid algebras of Sections 4 and 5 are only given over a field \(k\). The aim of Section 6 is to sketch some ideas to merge the two approaches so as to give a meaningful definition of what relative dagger geometry over any Banach base ring should be. The main construction is that of an object \(W_R^n(\rho)\in\mathsf{Comm}(\mathsf{Ind}(\mathsf{Ban}_R)))\) corresponding to the ``polydisk of multiradi \(\rho\) and dimension \(n\) over \(R\)'', mimicking what has been done in case \(R=k\) is a field. This can be applied, for instance, when \(R=\mathbb{Z}\), endowed with the trivial norm. The paper finished with an appendix about a non-Archimidification functor which enables one, in particular, to attach to every dagger affinoid space a classical rigid analytic space. Some properties of this functor are studied.