Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Higher dimensional formal loop spaces

Higher dimensional formal loop spaces 8 1 If M is a symplectic manifold then the space of smooth loops C pS , Mq inherits of a quasi- symplectic form. We will focus in this article on an algebraic analogue of that result. In their article [KV1], Kapranov and Vasserot introduced and studied the formal loop space of a scheme X. We generalize their construction to higher dimensional loops. To any scheme X – not neces- sarily smooth – we associate L pXq, the space of loops of dimension d. We prove it has a structure of (derived) Tate scheme – ie its tangent is a Tate module: it is infinite dimensional but behaves nicely enough regarding duality. We also define the bubble space B pXq, a variation of the loop space. We prove that B pXq is endowed with a natural symplectic form as soon as X has one (in the sense of [PTVV]). Throughout this paper, we will use the tools of p8, 1q-categories and symplectic derived alge- braic geometry. Résumé 8 1 Espaces des lacets formels de dimension supérieure : L’espace des lacets C pS , Mq associé à une variété symplectique M se voit doté d’une structure (quasi-)symplectique induite par celle de M. Nous traiterons dans cet article d’un analogue algébrique de cet énoncé. Dans leur article [KV1], Kapranov et Vasserot ont introduit l’espace des lacets formels associé à un schéma. Nous généralisons leur construction à des lacets de dimension supérieure. Nous associons à tout schéma X – pas forcément lisse – l’espace L pXq de ses lacets formels de dimension d. Nous démontrerons que ce dernier admet une structure de schéma (dérivé) de Tate : son espace tangent est de Tate : de dimension infinie mais suffisamment structuré pour se soumettre à la dualité. Nous définirons également l’espace B pXq des bulles de X, une variante de l’espace des lacets, et nous montrerons que le cas échéant, il hérite de la structure symplectique de X. Contents Introduction 2 Preliminaries 5 0.1 A few tools from higher category theory . . . . . . . . . . . . . . . . . . . . . . . . . . 5 0.2 Derived algebraic geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1 Ind-pro-stacks 12 1.1 Cotangent complex of a pro-stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.2 Cotangent complex of an ind-pro-stack . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.3 Uniqueness of pro-structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 1.4 Uniqueness of ind-pro-structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 arXiv:2010.10318v1 [math.AG] 20 Oct 2020 2 Symplectic Tate stacks 27 2.1 Tate stacks: definition and first properties . . . . . . . . . . . . . . . . . . . . . . . . . 27 2.2 Shifted symplectic Tate stacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.3 Mapping stacks admit closed forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.4 Mapping stacks have a Tate structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3 Formal loops 34 3.1 Dehydrated algebras and de Rham stacks . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.2 Higher dimensional formal loop spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.3 Tate structure and determinantal anomaly . . . . . . . . . . . . . . . . . . . . . . . . . 39 4 Bubble spaces 42 4.1 Two lemmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 4.2 Definition and properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.3 Its tangent is a Tate module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 4.4 A symplectic structure (shifted by d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 References 49 Introduction Considering a differential manifold M, one can build the space of smooth loops LpMq in M. It is a central object of string theory. Moreover, if M is symplectic then so is LpMq – more precisely quasi-symplectic since it is not of finite dimension – see for instance [MP]. We will be interested here in an algebraic analogue of that result. The first question is then the following: what is an algebraic analogue of the space of smooth loops? An answer appeared in 1994 in Carlos Contou-Carrère’s work (see [CC]). He studies there G pCpptqqq, some sort of holomorphic functions in the multiplicative group scheme, and defines the famous Contou- Carrère symbol. This is the first occurrence of a formal loop space known by the author. This idea was then generalised to algebraic groups as the affine Grassmannian Gr “ GpCpptqqq showed G GpCrrtssq up and got involved in the geometric Langlands program. In their paper [KV1], Mikhail Kapranov and Éric Vasserot introduced and studied the formal loop space of a smooth scheme X. It is an ind-scheme LpXq which we can think of as the space of maps SpecCpptqq Ñ X. This construction strongly inspired the one presented in this article. There are at least two ways to build higher dimensional formal loops. The most studied one consists in using higher dimensional local fields kppt qq . . . ppt qq and is linked to Beilinson’s adèles. There 1 d is also a generalisation of Contou-Carrère symbol in higher dimensions using those higher dimensional local fields – see [OZ] and [BGW1]. If we had adopted this angle, we would have considered maps from some torus Specpkppt qq . . . ppt qqq to X. 1 d The approach we will follow in this work is different. We generalize here the definition of Kapranov and Vasserot to higher dimensional loops in the following way. For X a scheme of finite presentation, not necessarily smooth, we define L pXq, the space of formal loops of dimension d in X. We define d d L pXq the space of maps from the formal neighbourhood of 0 in A to X. This is a higher dimensional version of the space of germs of arcs as studied by Jan Denef and François Loeser in [DL]. Let also d d L pXq denote the space of maps from a punctured formal neighbourhood of 0 in A to X. The formal d d d loop space L pXq is the formal completion of L pXq in L pXq. Understanding those three items is V U the main goal of this work. The problem is mainly to give a meaningful definition of the punctured The variable t , . . . , t are actually ordered. The author likes to think of Specpkppt qq . . . ppt qqq as a formal torus 1 d 1 d equipped with a flag representing this order. 2 formal neighbourhood of dimension d. We can describe what its cohomology should be: ’krrX , . . . , X ss if n “ 0 1 d n d ´1 ´1 ´1 H pA r t0uq “ pX . . . X q krX , . . ., X s if n “ d ´ 1 1 d 0 otherwise but defining this punctured formal neighbourhood with all its structure is actually not an easy task. Nevertheless, we can describe what maps out of it are, hence the definition of L pXq and the formal loop space. This geometric object is of infinite dimension, and part of this study is aimed at identifying some structure. Here comes the first result in that direction. Theorem 1 (see Proposition 3.3.4). The formal loop space of dimension d in a scheme X is rep- resented by a derived ind-pro-scheme. Moreover, the functor X ÞÑ L pXq satisfies the étale descent condition. We use here methods from derived algebraic geometry as developed by Bertrand Toën and Gabriele Vezzosi in [TV]. The author would like to emphasize here that the derived structure is d d d necessary since, when X is a scheme, the underlying schemes of L pXq, L pXq and L pXq are iso- U V morphic as soon as d ě 2. Let us also note that derived algebraic geometry allowed us to define L pXq for more general X’s, namely any derived stack. In this case, the formal loop space L pXq is no longer a derived ind-pro-scheme but an ind-pro-stack. The case d “ 1 and X is a smooth scheme gives a derived enhancement of Kapranov and Vasserot’s definition. This derived enhancement is conjectured to be trivial when X is a smooth affine scheme in [GR, 9.2.10]. Gaitsgory and Rozenblyum also prove in loc. cit. their conjecture holds when X is an algebraic group. The proof of Theorem 1 is based on an important lemma. We identify a full sub-category C of the category of ind-pro-stacks such that the realisation functor C Ñ dSt is fully faithful. We then prove that whenever X is a derived affine scheme, the stack L pXq is in the essential image of C and is thus endowed with an essentially unique ind-pro-structure satisfying some properties. The generalisation to any X is made using a descent argument. Note that for general X’s, the ind-pro-structure is not known to satisfy nice properties one could want to have, for instance on the transition maps of the diagrams. We then focus on the following problem: can we build a symplectic form on L pXq when X is symplectic? Again, this question requires the tools of derived algebraic geometry and shifted symplectic structures as in [PTVV]. A key feature of derived algebraic geometry is the cotangent complex L of any geometric object X. A (n-shifted) symplectic structure on X is a closed 2-form O r´ns Ñ L ^ L which is non degenerate – ie induces an equivalence X X T Ñ L rns X X Because L pXq is not finite, linking its cotangent to its dual – through an alleged symplectic form – requires to identify once more some structure. We already know that it is an ind-pro-scheme but the proper context seems to be what we call Tate stacks. Before saying what a Tate stack is, let us talk about Tate modules. They define a convenient context for infinite dimensional vector spaces. They where studied by Lefschetz, Beilinson and Drin- feld, among others, and more recently by Bräunling, Gröchenig and Wolfson [BGW2]. We will use here the notion of Tate objects in the context of stable p8, 1q-categories as developed in [Hen2]. If C is a stable p8, 1q-category – playing the role of the category of finite dimensional vector spaces, the category TatepCq is the full subcategory of the p8, 1q-category of pro-ind-objects Pro IndpCq in C containing both IndpCq and PropCq and stable by extensions and retracts. We will define the derived category of Tate modules on a scheme – and more generally on a derived ind-pro-stack. An Artin ind-pro-stack X – meaning an ind-pro-object in derived Artin stacks – is then gifted with a cotangent complex L . This cotangent complex inherits a natural structure of pro-ind-module on X. This allows us to define a Tate stack as an Artin ind-pro-stack whose cotangent 3 d complex is a Tate module. The formal loop space L pXq is then a Tate stack as soon as X is a finitely presented derived affine scheme. For a more general X, what precedes makes L pXq some kind of locally Tate stack. This structure suffices to define a determinantal anomaly ´ ¯ “ ‰ 2 d ˆ Det d P H L pXq,O L pXq d L pXq for any quasi-compact quasi-separated (derived) scheme X – this construction also works for slightly more general X’s, namely Deligne-Mumford stacks with algebraisable diagonal, see Definition 2.1.3. Kapranov and Vasserot proved in [KV3] that in dimension 1, the determinantal anomaly governs the existence of sheaves of chiral differential operators on X. One could expect to have a similar result in higher dimensions, with higher dimensional analogues of chiral operators and vertex algebras. The author plans on studying this in a future work. Another feature of Tate modules is duality. It makes perfect sense and behaves properly. Using the theory of symplectic derived stacks developed by Pantev, Toën, Vaquié and Vezzosi in [PTVV], we are then able to build a notion of symplectic Tate stack: a Tate stack Z equipped with a (n-shifted) closed 2-form which induces an equivalence T Ñ L rns Z Z of Tate modules over Z between the tangent and (shifted) cotangent complexes of Z. To make a step toward proving that L pXq is a symplectic Tate stack, we actually study the bubble space B pXq – see Definition 4.2.3. When X is affine, we get an equivalence d d d B pXq » L pXq ˆ L pXq V V L pXq Note that the fibre product above is a derived intersection. We then prove the following result Theorem 2 (see Theorem 4.4.1). If X is an n-shifted symplectic stack then the bubble space B pXq is endowed with a structure of pn ´ dq-shifted symplectic Tate stack. The proof of this result is based on a classical method. The bubble space is in fact, as an ind- pro-stack, the mapping stack from what we call the formal sphere S of dimension d to X. There are therefore two maps pr ev d d oo ˆ // B pXq B pXq ˆ S X d ˚ The symplectic form on B pXq is then ev ω , where ω is the symplectic form on X. The key X X ˆd argument is the construction of this integration on the formal sphere, ie on an oriented pro-ind-stack of dimension d. The orientation is given by a residue map. On the level of cohomology, it is the morphism d d ´1 ´1 ´1 H pS q » pX . . .X q krX , . . . , X s Ñ k 1 d 1 d ´1 mapping pX . . . X q to 1. 1 d This integration method would not work on L pXq, since the punctured formal neighbourhood does not have as much structure as the formal sphere: it is not known to be a pro-ind-scheme. Nevertheless, Theorem 2 is a first step toward proving that L pXq is symplectic. We can consider the d d nerve Z of the map L pXq Ñ L pXq. It is a groupoid object in ind-pro-stacks whose space of maps V U is B pXq. The author expects that this groupoid is compatible in some sense with the symplectic structure so that L pXq would inherit a symplectic form from realising this groupoid. One the other hand, if L pXq was proven to be symplectic, then the fibre product defining B pXq should be a Lagrangian intersection. The bubble space would then inherit a symplectic structure from that on L pXq. 4 Techniques and conventions Throughout this work, we will use the techniques of p8, 1q-category theory. We will once in a while use explicitly the model of quasi-categories developed by Joyal and Lurie (see [HTT]). That being said, the results should be true with any equivalent model. Let us fix now two universes U P V to deal with size issues. Every algebra, module or so will implicitly be U-small. The first part will consist of reminders about p8, 1q-categories. We will fix there some notations. Note that we will often refer to [Hen1] for some p8, 1q-categorical results needed in this article. We will also use derived algebraic geometry, as introduced in [TV]. We refer to [Toë2] for a recent survey of this theory. We will denote by k a base field and by dSt the p8, 1q-category of (U-small) derived stacks over k. In the first section, we will dedicate a few page to introduce derived algebraic geometry. Outline This article begins with a few paragraphs, recalling some notions we will use. Among them are p8, 1q- categories and derived algebraic geometry. In section 1, we set up a theory of geometric ind-pro-stacks. We then define in section 2 symplectic Tate stacks and give a few properties, including the construction of the determinantal anomaly (see Definition 2.1.3). Comes section 3 where we finally define higher dimensional loop spaces and prove Theorem 1 (see Proposition 3.3.4). We finally introduce the bubble space and prove Theorem 2 (see Theorem 4.4.1) in section 4. Aknowledgements I would like to thank Bertrand Toën, Damien Calaque and Marco Robalo for the many discussions we had about the content of this work. I am grateful to Mikhail Kapranov, James Wallbridge and Giovanni Faonte for inviting me at the IPMU. My stay there was very fruitful and the discussions we had were very interesting. I learned after writing down this article that Kapranov had an unpublished document in which higher dimensional formal loops are studied. I am very grateful to Kapranov for letting me read those notes, both inspired and inspiring. This work is extracted from my PhD thesis [Hen1] under the advisement of Bertrand Toën. I am very grateful to him for those amazing few years. Preliminaries In this part, we recall some results and definitions from p8, 1q-category theory and derived algebraic geometry. 0.1 A few tools from higher category theory In the last decades, theory of p8, 1q-categories has tremendously grown. The core idea is to consider categories enriched over spaces, so that every object or morphism is considered up to higher homotopy. The typical example of such a category is the category of topological spaces itself: for any topological spaces X and Y , the set of maps X Ñ Y inherits a topology. It is often useful to talk about topological spaces up to homotopy equivalences. Doing so, one must also consider maps up to homotopy. To do so, one can of course formally invert every homotopy equivalence and get a set of morphisms rX, Y s. This process loses information and mathematicians tried to keep trace of the space of morphisms. The first fully equipped theory handy enough to work with such examples, called model categories, was introduced by Quillen. A model category is a category with three collections of maps – weak equivalences (typically homotopy equivalences), fibrations and cofibrations – satisfying a bunch of conditions. The datum of such collections allows us to compute limits and colimits up to homotopy. We refer to [Hov] for a comprehensive review of the subject. 5 Using model categories, several mathematicians developed theories of p8, 1q-categories. Let us name here Joyal’s quasi-categories, complete Segal spaces or simplicial categories. Each one of those theories is actually a model category and they are all equivalent one to another – see [Ber] for a review. In [HTT], Lurie developed the theory of quasi-categories. In this book, he builds everything necessary so that we can think of p8, 1q-categories as we do usual categories. To prove something in this context still requires extra care though. We will use throughout this work the language as developed by Lurie, but we will try to keep in mind the 1-categorical intuition. In this section, we will fix a few notations and recall some results to which we will often refer. Notations: Let us first fix a few notations, borrowed from [HTT]. • We will denote by Cat the p8, 1q-category of U-small p8, 1q-categories – see [HTT, 3.0.0.1]; L,U • Let Pr denote the p8, 1q-category of U-presentable (and thus V-small) p8, 1q-categories with left adjoint functors – see [HTT, 5.5.3.1]; • The symbol sSets will denote the p8, 1q-category of U-small simplicial sets up to homotopy equivalences (this is equivalent to the category of (nice) topological spaces up to homotopy); • For any p8, 1q-categories C and D we will write FctpC,Dq for the p8, 1q-category of functors from op C to D (see [HTT, 1.2.7.3]). The category of presheaves will be denoted PpCq “ FctpC , sSetsq. • For any p8, 1q-category C and any objects c and d in C, we will denote by Map pc, dq the space of maps from c to d. • For any simplicial set K, we will denote by K the simplicial set obtained from K by formally adding a final object. This final object will be called the cone point of K . The following theorem is a concatenation of results from Lurie. Theorem 0.1.1 (Lurie). Let C be a V-small p8, 1q-category. There is an p8, 1q-category Ind pCq and a functor j : C Ñ Ind pCq such that (i) The p8, 1q-category Ind pCq is V-small; (ii) The p8, 1q-category Ind pCq admits U-small filtered colimits and is generated by U-small filtered colimits of objects in jpCq; (iii) The functor j is fully faithful and preserves finite limits and finite colimits which exist in C; (iv) For any c P C, its image jpcq is U-small compact in Ind pCq; (v) For every p8, 1q-category D with every U-small filtered colimits, the functor j induces an equiv- alence U´c U Fct pInd pCq,Dq Ñ FctpC,Dq U´c U U where Fct pInd pCq,Dq denote the full subcategory of FctpInd pCq,Dq spanned by functors preserving U-small filtered colimits. (vi) If C is U-small and admits all finite colimits then Ind pCq is U-presentable; (vii) If C is endowed with a symmetric monoidal structure then there exists such a structure on Ind pCq such that the monoidal product preserves U-small filtered colimits in each variable. Proof. Let us use the notations of [HTT, 5.3.6.2]. Let K denote the collection of U-small filtered U K simplicial sets. We then set Ind pCq “ P pCq. It satisfies the required properties because of loc. cit. 5.3.6.2 and 5.5.1.1. We also need tiny modifications of the proofs of loc. cit. 5.3.5.14 and 5.3.5.5. The last item is proved in [HAlg, 6.3.1.10]. 6 U Remark 0.1.2. Note that when C admits finite colimits then the category Ind pCq embeds in the V-presentable category Ind pCq. Definition 0.1.3. Let C be a V-small 8-category. We define Pro pCq as the p8, 1q-category ´ ¯ op U U op Pro pCq “ Ind pC q It satisfies properties dual to those of Ind pCq. Definition 0.1.4. Let C be a V-small p8, 1q-category. Let V U V i: FctpC, Cat q Ñ FctpInd pCq, Cat q 8 8 denote the left Kan extension functor. We will denote by Ind the composite functor Ind ˝´ V U V U V // // FctpC, Cat q FctpInd pCq, Cat q FctpInd pCq, Cat q 8 8 8 We will denote by Pro the composite functor Pro ˝´ V V U V // // FctpC, Cat q FctpC, Cat q FctpPro pCq, Cat q 8 8 8 We define the same way V V V Ind : FctpC, Cat q Ñ FctpInd pCq, Cat q C 8 V V V Pro : FctpC, Cat q Ñ FctpPro pCq, Cat q C 8 Remark 0.1.5. The Definition 0.1.4 can be expanded as follows. To any functor f : C Ñ Cat and any ind-object c colimit of a diagram c¯ // // K C Ind pCq we construct an p8, 1q-category U U Ind pfqpcq » Ind pcolimfpc¯qq To any pro-object d limit of a diagram op // // K C Pro pCq we associate an p8, 1q-category U U Pro pfqpdq » lim Pro pfpdqq V,st V Definition 0.1.6. Let Cat denote the subcategory of Cat spanned by stable categories with 8 8 V,st,id V,st exact functors between them – see [HAlg, 1.1.4]. Let Cat denote the full subcategory of Cat 8 8 spanned by idempotent complete stable categories. Remark 0.1.7. It follows from [HAlg, 1.1.4.6, 1.1.3.6, 1.1.1.13 and 1.1.4.4] that the functors Ind and Pro restricts to functors U U V,st V,st Ind : FctpC, Cat q Ñ FctpInd pCq, Cat q C 8 8 U V,st U V,st Pro : FctpC, Cat q Ñ FctpPro pCq, Cat q C 8 8 7 Symmetric monoidal p8, 1q-categories: We will make use in the last chapter of the theory of symmetric monoidal p8, 1q-categories as developed in [HAlg]. Let us give a (very) quick review of those objects. Definition 0.1.8. Let Fin denote the category of pointed finite sets. For any n P N, we will denote by xny the set t˚, 1, . . ., nu pointed at ˚. For any n and i ď n, the Segal map δ : xny Ñ x1y is defined i i by δ pjq “ 1 if j “ i and δ pjq “ ˚ otherwise. Definition 0.1.9. (see [HAlg, 2.0.0.7]) Let C be an p8, 1q-category. A symmetric monoidal structure on C is the datum of a coCartesian fibration p: C Ñ Fin such that • The fibre category C is equivalent to C and x1y b b n n • For any n, the Segal maps induce an equivalence C Ñ pC q » C . xny x1y b b,V where C denote the fibre of p at xny. We will denote by Cat the p8, 1q-category of V-small xny symmetric monoidal p8, 1q-categories – see [HAlg, 2.1.4.13]. ˚ V Such a coCartesian fibration is classified by a functor φ: Fin Ñ Cat – see [HTT, 3.3.2.2] – such that φpxnyq » C . The tensor product on C is induced by the map of pointed finite sets µ : x2y Ñ x1y mapping both 1 and 2 to 1 b “ φpµ q: C Ñ C b,V V Remark 0.1.10. The forgetful functor Cat Ñ Cat preserves all limits as well as filtered colimits 8 8 – see [HAlg, 3.2.2.4 and 3.2.3.2]. Moreover, it follows from Theorem 0.1.1 - (vii) that the functor Ind induces a functor U b,V b,V Ind : Cat Ñ Cat 8 8 U U U The same holds for Pro . The constructions Ind and Pro therefore restrict to U b,V U b,V Ind : FctpC, Cat q Ñ FctpInd pCq, Cat q C 8 8 U b,V U b,V Pro : FctpC, Cat q Ñ FctpPro pCq, Cat q C 8 8 Tate objects: We now recall the definition and a few properties of Tate objects in a stable and idempotent complete p8, 1q-category. The content of this paragraph comes from [Hen2]. See also [Hen1]. Definition 0.1.11. Let C be a stable and idempotent complete p8, 1q-category. Let Tate pCq denote U U U U the smallest full subcategory of Pro Ind pCq containing Ind pCq and Pro pCq, and both stable and idempotent complete. U U U The category Tate pCq naturally embeds into Ind Pro pCq as well. op Proposition 0.1.12. If moreover C is endowed with a duality equivalence C Ñ C then the induced functor ´ ¯ op U U U U U U Pro Ind pCq Ñ Pro Ind pCq » Ind Pro pCq U U op preserves Tate objects and induces an equivalence Tate pCq » Tate pCq . Definition 0.1.13. Let C be a V-small p8, 1q-category. We define the functor Tate ˝´ V,st U V,st U V,st,id // // Tate : FctpC, Cat q FctpInd pCq, Cat q FctpInd pCq, Cat q 8 8 8 8 0.2 Derived algebraic geometry We present here some background results about derived algebraic geometry. Let us assume k is a field of characteristic 0. First introduced by Toën and Vezzosi in [TV], derived algebraic geometry is a generalisation of algebraic geometry in which we replace commutative algebras over k by commutative differential graded algebras (or cdga’s). We refer to [Toë2] for a recent survey of this theory. ď0 Generalities on derived stacks: We will denote by cdga the p8, 1q-category of cdga’s over k concentrated in non-positive cohomological degree. It is the p8, 1q-localisation of a model category ď0 along weak equivalences. Let us denote dAff the opposite p8, 1q-category of cdga . It is the category of derived affine schemes over k. In this work, we will adopt a cohomological convention for cdga’s. op ď0 A derived prestack is a presheaf dAff » cdga Ñ sSets. We will thus write PpdAff q for k k the p8, 1q-category of derived prestacks. A derived stack is a prestack satisfying the étale descent condition. We will denote by dSt the p8, 1q-category of derived stacks. It comes with an adjunction p´q : PpdAff q Õ dSt k k where the left adjoint p´q is called the stackification functor. Remark 0.2.1. The categories of varieties, schemes or (non derived) stacks embed into dSt . Definition 0.2.2. The p8, 1q-category of derived stacks admits an internal hom MappX, Y q between ď0 two stacks X and Y . It is the functor cdga Ñ sSets defined by A ÞÑ Map pX ˆ Spec A, Y q dSt We will call it the mapping stack from X to Y . There is a derived version of Artin stacks of which we first give a recursive definition. Definition 0.2.3. (see for instance [Toë1, 5.2.2]) Let X be a derived stack. • We say that X is a derived 0-Artin stack if it is a derived affine scheme ; • We say that X is a derived n-Artin stack if there is a family pT q of derived affine schemes and a smooth atlas u: T Ñ X such that the nerve of u has values in derived pn ´ 1q-Artin stacks ; • We say that X is a derived Artin (or algebraic) stack if it is an n-Artin stack for some n. • We say that X is locally of finite presentation if there exists a smooth atlas T Ñ X as above, such that the derived affine schemes T are all finitely presented (ie their cdga of functions is finitely presented, or equivalently compact is the category of cdga’s). We also say that X is finitely presented if there is such an atlas with a finite number of T ’s. Art We will denote by dSt the full subcategory of dSt spanned by derived Artin stacks. Definition 0.2.4. A morphism X Ñ Y of derived stacks is called algebraic if for any cdga A and any map Spec A Ñ Y , the derived intersection X ˆ Spec A is an algebraic stack. To any cdga A we associate the category dgMod of dg-modules over A. Similarly, to any derived stack X we can associate a derived category QcohpXq of quasicoherent sheaves. It is a U-presentable p8, 1q-category given by the formula QcohpXq » lim dgMod SpecAÑX Moreover, for any map f : X Ñ Y , there is a natural pull back functor f : QcohpY q Ñ QcohpXq. This functor admits a right adjoint, which we will denote by f . This construction is actually a functor of p8, 1q-categories. 9 Definition 0.2.5. Let us denote by Qcoh the functor op L,U Qcoh: dSt Ñ Pr For any X we can identify a full subcategory PerfpXq Ă QcohpXq of perfect complexes. This defines a functor op Perf : dSt Ñ Cat Remark 0.2.6. For any derived stack X the categories QcohpXq and PerfpXq are actually stable and idempotent complete p8, 1q-categories. The inclusion PerfpXq Ñ QcohpXq is exact. Moreover, for any map f : X Ñ Y the pull back functor f preserves perfect modules and is also exact. Definition 0.2.7 (see for instance the appendix of [HLP]). Let X be a derived stack and let π: X Ñ ˚ denote the projection. We say that X is of finite cohomological dimension if there is a non-negative integer d such that the complex π O “ RΓpX,O q P dgMod is concentrated in degree lower or ˚ X X equal to d. Example 0.2.8. Any derived affine scheme is of finite cohomological dimension (take d “ 0). Any quasi-compact quasi-separated derived stack (ie a finite colimit of derived affine schemes) is of finite cohomological dimension. Any derived Artin stack X over a basis S admits a cotangent complex L P QcohpXq. If X is X{S locally of finite presentation, then the its cotangent complex is perfect L P PerfpXq X{S Symplectic structures: Following [PTVV], to any derived stack X we associate two complexes p p,cl A pXq and A pXq in dgMod , respectively of p-forms and closed p-forms on X. They come with p,cl p 2 a natural morphism A pXq Ñ A pXq forgetting the lock closing the forms . This actually glues into a natural transformation p,cl '' dSt dgMod Let us emphasize that, as soon as X is Artin, the complex A pXq is canonically equivalent to the global sections complex of L ^ L . In particular, any n-shifted 2-forms kr´ns Ñ A pXq induces a X X morphism O r´ns Ñ L ^ L in QcohpXq. If X is locally of finite presentation, the cotangent L X X X X is perfect and we then get a map T r´ns Ñ L X X Definition 0.2.9. Let X be a derived Artin stack locally of finite presentation. • An n-shifted 2-form ω : kr´ns Ñ A pXq is called non-degenerated if the induced morphism T r´ns Ñ L is an equivalence; X X 2,cl • An n-shifted symplectic form on X is a n-shifted closed 2-form ω : kr´ns Ñ A pXq such 2,cl 2 that underlying 2-form kr´ns Ñ A pXq Ñ A pXq is non degenerate. ď0 ď´1 Obstruction theory: Let A P cdga and let M P dgMod be an A-module concentrated in k A negative cohomological degrees. Let d be a derivation A Ñ A ‘ M and s: A Ñ A ‘ M be the trivial derivation. The square zero extension of A by Mr´1s twisted by d is the fibre product // A ‘ Mr´1s A // A A ‘ M This lock is a structure on the form: being closed in not a property in this context. 10 ď´1 Let now X be a derived stack and M P QcohpXq . We will denote by XrMs the trivial square zero extension of X by M. Let also d: XrMs Ñ X be a derivation – ie a retract of the natural map X Ñ XrMs. We define the square zero extension of X by Mr´1s twisted by d as the colimit X rMr´1ss “ colim SpecpA ‘ ˚ f Mr´1sq d f d f : SpecAÑX It is endowed with a natural morphism X Ñ X rMr´1ss induced by the projections p as above. Proposition 0.2.10 (Obstruction theory on stacks). Let F Ñ G be an algebraic morphism of derived ď´1 stacks. Let X be a derived stack and let M P QcohpXq . Let d be a derivation d P Map pXrMs, Xq X{´ We consider the map of simplicial sets ψ: MappX rMr´1ss, Fq Ñ MappX, Fq ˆ MappX rMr´1ss, Gq d d MappX,Gq Let y P MappX, Fq ˆ MappX rMr´1ss, Gq and let x P MappX, Fq be the induced map. There MappX,Gq d exists a point αpyq P Mappx L , Mq such that the fibre ψ of ψ at y is equivalent to the space of F{G y paths from 0 to αpyq in Mappx L , Mq F{G ψ » Ω Mappx L , Mq y 0,αpyq F{G Proof. This is a simple generalisation of [TV, 1.4.2.6]. The proof is very similar. We have a natural commutative square // XrMs X // X X rMr´1ss It induces a map α: MappX, Fq ˆ MappX rMr´1ss, Gq Ñ Map pXrMs, Fq » Mappx L , Mq d F{G X{´{G MappX,Gq Let Ω Map pXrMs, Fq denote the space of paths from 0 to αpyq. It is the fibre product 0,αpyq X{´{G // Ω Map pXrMs, Fq ˚ 0,αpyq X{´{G αpyq // ˚ Map pXrMs, Fq X{´{G The composite map αψ is by definition homotopic to the 0 map. This defines a morphism f : Ω Map pXrMs, Fq Ñ ψ 0,αpyq y X{´{G It now suffices to see that the category of X’s for which f is an equivalence contains every derived affine scheme and is stable by colimits. The first assertion is exactly [TV, 1.4.2.6] and the second one is trivial. 11 Postnikov towers: To any cdga A, one can associate its n-truncation A for some n. It is, by ďn definition, the universal cdga with vanishing cohomology H pA q for p ă ´n associated to A. The ďn truncation comes with a canonical map A Ñ A so that one can form the diagram ďn A Ð A Ð . . . ď0 ď1 This induces a canonical morphism A Ñ lim A which is an equivalence. n ďn r´n,0s This phenomenon has a counterpart when dealing with derived stacks. We denote by cdga the category of cdga’s with cohomology concentrated in degrees ´n to 0. It comes with the fully r´n,0s ď0 faithful embedding i : cdga Ñ cdga . k k ď0 For any prestack X : cdga Ñ sSets, we define its truncation τ X as the restriction of X to ďn r´n,0s the category cdga . We will abuse notations and also denote by τ X the functor obtained after ďn left Kan extending along i . The prestack τ X comes with a canonical morphism τ X Ñ X which, n ďn ďn as n varies, assembles to define a canonical map colimτ X Ñ X ďn Remark that this morphism is not necessarily an equivalence. We will study it in Lemma 1.3.1. Algebraisable stacks: Let X be a derived stack and A be a cdga. Let a “ pa , . . . , a q be a 1 p 0 0 n n sequence of elements of A forming a regular sequence in H pAq. Let A denote the Kozsul a , . . . , a 1 p n n complex associated with the regular sequence pa , . . ., a q. It is endowed with a cdga structure. There 1 p is a canonical map ´ ¯ ´ ¯ n n n n ψpAq : colim X A Ñ X lim A { { a a , . . ., a a , . . . , a 1 p 1 p n n This map is usually not an equivalence. Definition 0.2.11. A derived stack X is called algebraisable if for any A and any regular sequence a the map ψpAq is an equivalence. A map f : X Ñ Y is called algebraisable if for any derived affine scheme T and any map T Ñ Y , the fibre product X ˆ T is algebraisable. We will say that a derived stack X has algebraisable diagonal if the diagonal morphism X Ñ XˆX is algebraisable. Remark 0.2.12. A derived stack X has algebraisable diagonal if for any A and a the map ψpAq is fully faithful. One could also rephrase the definition of being algebraisable as follows. A stack is algebraisable if it does not detect the difference between ´ ¯ ´ ¯ n n n n colim Spec A and Spec lim A { { a , . . ., a a , . . ., a 1 p 1 p n n Example 0.2.13. Any derived affine scheme is algebraisable. Another important example of alge- braisable stack is the stack of perfect complexes. In [Bha], Bhargav Bhatt gives some more examples of algebraisable (non-derived) stacks – although our definition slightly differs from his. He proves that any quasi-compact quasi-separated algebraic space is algebraisable and also provides with examples of non-algebraisable stacks. Let us name KpG , 2q – the Eilenberg-Maclane classifying stack of G – m m as an example of non-algebraisable stack. Algebraisability of Deligne-Mumford stacks is also look at in [Lur]. 1 Ind-pro-stacks Throughout this section, we will denote by S a derived stack over some base field k and by dSt the category of derived stack over the base S. 12 1.1 Cotangent complex of a pro-stack Definition 1.1.1. A pro-stack over S an object of Pro dSt . Remark 1.1.2. Note that the category Pro dSt is equivalent to the category of pro-stacks over k with a morphism to S. op U Definition 1.1.3. Let Perf : dSt Ñ Cat denote the functor mapping a stack to its category of perfect complexes. We will denote by IPerf the functor U U op L op IPerf “ Ind pPerfq: pPro dSt q Ñ Pr dSt 8 where Ind was defined in Definition 0.1.4. Whenever X is a pro-stack, we will call IPerfpXq the derived category of ind-complexes on X. It is U-presentable. If f : X Ñ Y is a map of pro-stacks, then the functor IPerfpfq: IPerfpY q Ñ IPerfpXq admits a right adjoint (as both the involved categories are presentable and the functor preserves all ˚ I colimits). We will denote f “ IPerfpfq and f its right adjoint. op Remark 1.1.4. Let X be a pro-stack and let X : K Ñ dSt denote a U-small cofiltered diagram of whom X is a limit in Pro dSt . The derived category of ind-perfect complexes on X is by definition the category IPerfpXq “ Ind pcolim PerfpXqq It thus follows from [HAlg, 1.1.4.6 and 1.1.3.6] that IPerfpXq is stable. Note that it is also equivalent to the colimit L,V IPerfpXq “ colim IPerfpXq P Pr It is therefore equivalent to the limit of the diagram op L,V R,V op IPerf pXq: K Ñ dSt Ñ Pr » pPr q 8 8 An object E in IPerfpXq is therefore the datum of an object E of IPerfpXpkqq for each k P K and of some compatibilities between them. We will then have E » p E where p : X Ñ Xpkq is the k k k natural projection. Definition 1.1.5. Let X be a pro-stack. We define its derived category of pro-perfect complexes op PPerfpXq “ pIPerfpXqq Recall that perfect complexes are precisely the dualizable objects in the category of quasi-coherent op complexes. They therefore come with a duality equivalence Perfp´q Ñ pPerfp´qq . This gives rise to the equivalence PPerfpXq » Pro pcolimPerfpXqq op U whenever X : K Ñ dSt is a cofiltered diagram of whom X is a limit in Pro dSt . S S U U op V,st,id Definition 1.1.6. Let us define the functor Tate : pPro dSt q Ñ Cat P 8 U U Tate “ Tate oppPerfq P dSt U op Remark 1.1.7. The functor Tate maps a pro-stack X given by a diagram X : K Ñ dSt to the stable p8, 1q-category U U Tate pXq “ Tate pcolim PerfpXqq There is a canonical fully faithful natural transformation U U Tate Ñ Pro ˝ IPerf We also get a fully faithful U U Tate Ñ Ind ˝ PPerf 13 op V Definition 1.1.8. Let Qcoh: dSt Ñ Cat denote the functor mapping a derived stack to its S 8 derived category of quasi-coherent sheaves. It maps a morphism between stacks to the appropriate pullback functor. We will denote by IQcoh the functor U U V op op IQcoh “ Ind pQcohq: pPro dSt q Ñ Cat dSt 8 If f : X Ñ Y is a map of pro-stacks, we will denote by f the functor IQcohpfq. We also define ď0 U ď0 op IQcoh “ Ind pQcoh q dSt the functor of connective modules. Remark 1.1.9. There is a fully faithful natural transformation IPerf Ñ IQcoh ; for any map f : X Ñ Y of pro-stacks, there is therefore a commutative diagram // IPerfpY q IQcohpY q ˚ ˚ f f I I // IPerfpXq IQcohpXq The two functors denoted by f are thus compatible. Let us also say that the functor f : IQcohpY q Ñ IQcohpXq does not need to have a right adjoint. We next show that it sometimes has one. Proposition 1.1.10. Let f : X Ñ Y be a map of pro-stacks. If Y is actually a stack then the functor f : IQcohpY q Ñ IQcohpXq admits a right adjoint. op Proof (sketch of). For a complete proof, we refer to [Hen1, 1.2.0.8]. Let us denote by X : K Ñ dSt a cofiltered diagram of whom X is a limit in Pro dSt . The map X Ñ Y factors through the S S projection X Ñ Xpkq for some k P K. The right adjoint of f is then (informally) given by the limit lim fplq kÑl ¯ ¯ ¯ where fplq is the right adjoint to the induced functor fplq : IQcohpY q Ñ IQcohpXplqq. IQ Definition 1.1.11. Let f : X Ñ Y be a map of pro-stacks. We will denote by f the right adjoint to f : IQcohpY q Ñ IQcohpXq if it exists. Remark 1.1.12. In the situation of Proposition 1.1.10, there is a natural transformation // IPerfpXq IQcohpXq 4< ♣♣ I IQ ♣♣ f ♣ f ˚ ˚ ♣♣ ♣♣ // IPerfpY q IQcohpY q It does not need to be an equivalence. Definition 1.1.13. Let X be a pro-stack over S. The structural sheaf O of X is the pull-back of O along the structural map X Ñ S. op Example 1.1.14. Let X be a pro-stack over S and X : K Ñ dSt be a U-small cofiltered diagram of whom X is a limit in Pro dSt . Let k be a vertex of K, let X denote Xpkq and let p denote S k k the induced map of pro-stacks X Ñ X . If f : k Ñ l is an arrow in K, we will also denote by f the map of stacks Xpfq. We have IQ pp q pO q » colim f O k ˚ X ˚ X f : kÑl One can see this using [Hen1, 1.2.0.7]. 14 1 Definition 1.1.15. For any category C with finite colimits, we will denote by B the functor C Ñ Cat mapping a morphism φ: c Ñ d to the category of factorizations c Ñ e Ñ d of φ. For a formal definition in the context of p8, 1q-categories, we refer to [Hen1, 1.3.0.14]. Definition 1.1.16. Let T be a stack over S. Let us consider the functor ´ ¯ op ď0 > T op QcohpTq Ñ B pid q » dSt T T dSt mapping a quasi-coherent sheaf E to the square zero extension T Ñ TrEs Ñ T . This construction is functorial in T and actually comes from a natural transformation ď0 > Ex: Qcoh Ñ B oppid q dSt op V Pro of functors dSt Ñ Cat . We will denote by Ex the natural transformation Pro U ď0 U > > Ex “ Ind oppExq: IQcoh Ñ Ind oppB oppid qq » B pid q U op ´ ´ dSt dSt dSt pPro dSt q S S S op between functors pPro dSt q Ñ Cat . The equivalence on the right is the one from [Hen1, S 8 ď0 1.3.0.18]. If X is a pro-stack and E P IQcohpXq then we will denote by X Ñ XrEs Ñ X the image Pro of E by the functor Ex pXq. op Remark 1.1.17. Let us give a description of this functor. Let X be a pro-stack and let X : K Ñ dSt denote a U-small cofiltered diagram of whom X is a limit in Pro dSt . For every k P K we S S can compose the functor mentioned above with the base change functor ´ˆ X X r´s X op // // X X k{ { pQcohpX qq dSt Pro dSt k X X ` ˘ op This is functorial in k and we get a functor colimQcohpXq Ñ Pro dSt which we extend and obtain a more explicit description of the square zero extension functor op X U Xr´s: pIQcohpXqq Ñ Pro dSt Definition 1.1.18. Let X be a pro-stack. • We finally define the functor of derivations over X : ď0 DerpX, ´q “ Map pXr´s, Xq: IQcohpXq Ñ sSets X{´{S • We say that X admits a cotangent complex if the functor DerpX, ´q is corepresentable – ie there ď0 exists a L P IQcohpXq such that for any E P IQcohpXq X{S DerpX, Eq » MappL , Eq X{S Art Definition 1.1.19. Let dSt be the full sub-category of dSt spanned by derived Artin stacks over U Art Art,lfp Art S. An Artin pro-stack is an object of Pro dSt . Let dSt be the full sub-category of dSt S S spanned by derived Artin stacks locally of finite presentation over S. An Artin pro-stack locally of U Art,lfp finite presentation is an object of Pro dSt Proposition 1.1.20. Any Artin pro-stack X over S admits a cotangent complex L . Let us assume X{S Art U Art op that X : K Ñ dSt is a U-small cofiltered diagram of whom X is a limit in Pro dSt . When S S k is a vertex of K, let us denote by X the derived Artin stack Xpkq. If f : k Ñ l is an arrow in K, we will also denote by f : X Ñ X the map of stacks Xpfq. The cotangent complex is given by the l k formula ` ˘ L “ colimp L P Ind colim QcohpXq » IQcohpXq X{S X {S k k 15 where p is the canonical map X Ñ X . The following formula stands k k IQ p L » colim f L k˚ X{S ˚ X {S f : kÑl If X is moreover locally of finite presentation over S, then its cotangent complex belongs to IPerfpXq. Before proving this proposition, let us fix the following notation Definition 1.1.21. Let C be a full sub-category of an 8-category D. There is a natural transformation from O : d ÞÑ D to the constant functor D: D Ñ Cat . We denote by O the fiber product D { 8 d D O “ O ˆC : D Ñ Cat D 8 Remark 1.1.22. The functor O : D Ñ Cat maps an object d P D to the comma category of objects in C over d C “ pC ˆ tduq ˆ D DˆD Proof (of the proposition). The cotangent complex defines a natural transformation Art op pdSt q λ: O op Ñ Qcohp´q dSt To any stack T and any Artin stack U over S with a map f : T Ñ U, it associates the quasi-coherent ˚ Pro complex f L on T . Applying the functor Ind op we get a natural transformation λ U{S dSt U Art op pPro dSt q Pro U op λ “ Ind pλq: O Ñ IQcohp´q U op dSt pPro dSt q S S Specifying it to X we get a functor ´ ¯ op Pro X U Art λ : Pro dSt Ñ IQcohpXq X S Pro Let us set L “ λ pXq P IQcohpXq. We have by definition the equivalence X{S L » colimp L X{S k X {S Let us now check that it satisfies the required universal property. The functor DerpX, ´q is the limit op ď0 of the diagram K Ñ FctpIQcohpXq , sSetsq Map pXr´s, Xq X{´{S ď0 ď0 Fixing k P K, the functor Map pXr´s, X q: IQcohpXq » Ind pcolim QcohpXq q Ñ sSets X{´{S ď0 preserves filtered colimits. It is hence induced by its restriction to colim QcohpXq . It follows that the diagram Map pXr´s, Xq factors through a diagram X{´{S ` ˘ op ď0 ď0 ¯ ¯ δ: K Ñ Fct colim QcohpXq , sSets » lim FctpQcohpXq , sSetsq Similarly, the functor MappL , ´q is the limit of a diagram X{S op // ď0 // ď0 K lim FctpQcohpXq , sSetsq FctpIQcohpXq , sSetsq The universal property of the usual cotangent complex defines an equivalence between δ and µ . IQ To get the formula for p L , one uses [Hen1, 1.2.0.7] and the last statement is obvious. k˚ X{S Remark 1.1.23. The definition of the derived category of ind-quasi-coherent modules on a pro-stack is build for the above proposition and remark to hold. 16 Remark 1.1.24. We have actually proven that for any pro-stack X, the two functors ď0 X Art IQcohpXq ˆ dSt Ñ sSets defined by pE, Y q ÞÑ Map pXrEs, Y q X{´{S Pro pE, Y q ÞÑ Map pλ pY q, Eq IQcohpXq X are equivalent. 1.2 Cotangent complex of an ind-pro-stack Definition 1.2.1. An ind-pro-stack is an object of the category U U IPdSt “ Ind Pro dSt S S op Definition 1.2.2. Let us define the functor PIPerf : pIPdSt q Ñ Cat as PIPerf “ Pro U pIPerfq op pPro dSt q where Pro was defined in Definition 0.1.4. Whenever we have a morphism f : X Ñ Y of ind-pro- stacks, we will denote by f the functor PI f “ PIPerfpfq: PIPerfpY q Ñ PIPerfpXq PI Remark 1.2.3. Let X be an ind-pro-stack. Let X : K Ñ Pro dSt denote a U-small filtered diagram of whom X is a colimit in IPdSt . We have by definition PIPerfpXq » lim Pro pIPerfpXqq Proposition 1.2.4. Let f : X Ñ Y be a map of ind-pro-stacks. If Y is a pro-stack then the functor f : PIPerfpY q Ñ PIPerfpXq admits a right adjoint. PI Definition 1.2.5. Let f : X Ñ Y be a map of ind-pro-stacks. If the functor f : PIPerfpY q Ñ PIPerfpXq PI PI admits a right adjoint, we will denote it by f . PI I Proof (of the proposition). If both X and Y are pro-stacks, then f “ Pro pf q is right adjoint ˚ ˚ U U ˚ ˚ to f “ Pro pf q. Let now X be an ind-pro-stack and let X : K Ñ Pro dSt denote a U-small PI I filtered diagram of whom X is a colimit in IPdSt . We then have f : PIPerfpY q Ñ PIPerfpXq » lim PIPerfpXq PI The right adjoint is the informally given by the formula PI PI f “ lim fpkq ˚ ˚ where fpkq is the induced map Xpkq Ñ Y . For a formal proof, we refer to [Hen1, 1.2.0.5]. op Definition 1.2.6. Let X P IPdSt . We define IPPerfpXq “ pPIPerfpXqq . If X is the colimit in IPdSt of a filtered diagram K Ñ Pro dSt then we have S S IPPerfpXq » limpInd ˝ PPerf ˝Xq op We will denote by p´q : IPPerfpXq Ñ pPIPerfpXqq the duality functor. 17 U V,st,id op Definition 1.2.7. Let us define the functor Tate : pIPdSt q Ñ Cat as the right Kan IP 8 U U op op extension of Tate along the inclusion pPro dSt q Ñ pIPdSt q . It is by definition endowed S S with a canonical fully faithful natural transformation Tate Ñ PIPerf IP For any X P IPdSt , an object of Tate pXq will be called a Tate module on X. IP Remark 1.2.8. We can characterise Tate objects: a module E P PIPerfpXq is a Tate module if and only if for any pro-stack U and any morphism f : U Ñ X P IPdSt , the pullback f pEq is in IP Tate pUq. Let us also remark here that Lemma 1.2.9. Let X be an ind-pro-stack over S. The fully faithful functors _ ´ ¯ op p´q U op U // oo Tate pXq PIPerfpXq pIPPerfpXqq Tate pXq IP IP have the same essential image. We thus have an equivalence ´ ¯ op _ U U p´q : Tate pXq » Tate pXq IP IP Proof. This is a corollary of Proposition 0.1.12. op Definition 1.2.10. Let us define PIQcoh: pIPdSt q Ñ Cat to be the functor PIQcoh “ Pro U pIQcohq op pPro dSt q From Remark 0.1.10, for any ind-pro-stack X, the category PIQcohpXq admits a natural monoidal structure. We also define the subfunctor ď0 U ď0 PIQcoh “ Pro pIQcoh q U op pPro dSt q Remark 1.2.11. Let us give an informal description of the above definition. To an ind-pro-stack X “ colim lim X we associate the category α β αβ ˆ ˙ U U PIQcohpXq “ lim Pro Ind colim PerfpX q αβ Definition 1.2.12. Let f : X Ñ Y be a map of ind-pro-stacks. We will denote by f the functor PI PIQ PIQcohpfq. Whenever it exists, we will denote by f the right adjoint to f . PI Proposition 1.2.13. Let f : X Ñ Y be a map of ind-pro-stacks. If Y is actually a stack, then the induced functor f admits a right adjoint. PI Proof. This is very similar to the proof of Proposition 1.2.4 but using Proposition 1.1.10. Remark 1.2.14. There is a fully faithful natural transformation PIPerf Ñ PIQcoh. Using the same notation f for the images of a map f : X Ñ Y is therefore only a small abuse. Moreover, PI for any such map f : X Ñ Y , for which the right adjoints drawn below exist, there is a natural tranformation // PIPerfpY q PIQcohpY q 3; ♥♥ ♥♥ ♥♥ PI ♥ PIQ f ♥ f ˚ ♥ ˚ ♥♥ ♥♥ ♥♥ ♥♥ // PIPerfpXq PIQcohpXq It is generally not an equivalence. 18 IP U Pro Definition 1.2.15. Let Ex denote the natural transformation Pro U pEx q op pPro dSt q ´ ¯ IP ď0 U > > Ex : PIQcoh Ñ Pro U B U pid q » B oppid q op op ´ ´ pIPdSt q pPro dSt q pPro dSt q S S S op ď0 of functors pIPdSt q Ñ Cat . If X is an ind-pro-stack and E P PIQcohpXq then we will denote S 8 by X Ñ XrEs Ñ X the image of E by the functor ´ ¯ op IP ď0 X Ex pXq: PIQcohpXq Ñ IPdSt Remark 1.2.16. Let us decipher the above definition. Let X “ colim lim X be an ind-pro-stack α β αβ and let E be a pro-ind-module over it. By definition E is the datum, for every α, of a pro-ind-object α U U ď0 α α E in the category Pro Ind pcolim Qcoh pX qq. Let us denote E “ lim colim E . For any β αβ γ δ γδ α ď0 γ and δ, there is a β pγ, δq such that E is in the essential image of Qcoh pX q. We then αβ pγ,δq γδ 0 have XrEs “ colim lim lim X rE s P IPdSt αβ γδ S α,γ δ βěβ pγ,δq Definition 1.2.17. Let X be an ind-pro-stack. • We define the functor of derivations on X DerpX, ´q “ Map pXr´s, Xq X{´{S • We say that X admits a cotangent complex if there exists L P PIQcohpXq such that for X{S ď0 any E P PIQcohpXq DerpX, Eq » MappL , Eq X{S • Let us assume that f : X Ñ Y is a map of ind-pro-stacks and that Y admits a cotangent complex. We say that f is formally étale if X admits a cotangent complex and the natural map f L Ñ L is an equivalence. Y {S X{S Definition 1.2.18. An Artin ind-pro-stack over S is an object in the category Art U U Art IPdSt “ Ind Pro dSt S S An Artin ind-pro-stack locally of finite presentation is an object of Art,lfp U U Art,lfp IPdSt “ Ind Pro dSt S S Proposition 1.2.19. Any Artin ind-pro-stack X admits a cotangent complex L P PIQcohpXq X{S Art Let us assume that X : K Ñ Pro dSt is a U-small filtered diagram of whom X is a colimit in Art IPdSt . For any vertex k P K we will denote by X the pro-stack Xpkq and by i the structural k k map X Ñ X. For any f : k Ñ l in K, let us also denote by f the induced map X Ñ X . We have k k l for all k P K ˚ ˚ i L » lim f L P PIQcohpX q X{S X {S k k,PI I l f : kÑl If moreover X is locally of finite presentation then L belongs to PIPerfpXq. X{S Pro Proof. Let us recall the natural transformation λ from the proof of Proposition 1.1.20 U Art op pPro dSt q Pro U op λ “ Ind pλq: O Ñ IQcohp´q U op dSt pPro dSt q S S 19 U U op of functors pPro dSt q Ñ Cat . Applying the functor Pro U we define the natural S 8 op pPro dSt q IP transformation λ ` ˘ Art op pIPdSt q IP Pro S λ “ Pro U λ : O Ñ PIQcohp´q op op pPro dSt q pIPdSt q S S op between functors pIPdSt q Ñ Cat . Specifying to X we get a functor S 8 ´ ¯ op Art IP X λ : IPdSt Ñ PIQcohpXq X S IP We now define L “ λ pXq. By definition we have X{S ˚ Pro ˚ i L » lim λ pXq » lim f L X{S X X {S k,PI k I l f : kÑl for every k P K. Let us now prove that it satisfies the expected universal property. It suffices to compare for every k P K the functors Map pX r´s, Xq and Map pi L , ´q k X{S X {´{S PIQcohpX q k,PI k k ď0 ď0 defined on PIQcohpX q . They are both pro-extensions to PIQcohpX q of their restrictions k k ď0 IQcohpX q Ñ sSets. The restricted functor Map pX r´s, Xq is a colimit of the diagram k k X {´{S ´ ¯ op k ď0 Map pX r´s, Xq: K Ñ FctpIQcohpX q , sSetsq k k X {´{S while Map pi L , ´q is a colimit to the diagram PIQcohpX q X{S k k,PI ´ ¯ op Pro k ď0 ¯ { Map pλ pXq, ´q: K Ñ FctpIQcohpX q , sSetsq IQcohpX q X k k We finish the proof with Remark 1.1.24. Let us record here a technical result we will use later on. For a proof, we refer to [Hen1, 2.1.2.20]. Art Proposition 1.2.20. Let X P IPdSt . Let us denote by π: X Ñ S the structural map. Let also IP L denote the functor ´ ¯ op Art U U IPdSt Ñ Pro Ind QcohpSq Art op obtained by extending the functor pdSt q Ñ QcohpSq mapping f : T Ñ S to f L . Then we S T{S PIQ IP have π L » L pXq ˚ X{S Definition 1.2.21. Let X by an Artin ind-pro-stack locally of finite presentation over S. We will call the tangent complex of X the ind-pro-perfect complex on X T “ L P IPPerfpXq X{S X{S 1.3 Uniqueness of pro-structure Lemma 1.3.1. Let Y and Z be derived Artin stacks. The following is true (i) The canonical map MappZ, Y q Ñ lim Mappτ Z, Y q ďn is an equivalence; (ii) If Y is q-Artin and Z is m-truncated then the mapping space MappZ, Y q is pm ` qq-truncated. 20 Proof. We prove both items recursively on the Artin degree of Z. The case of Z affine is proved in [TV, C.0.10 and 2.2.4.6]. We assume that the result is true for n-Artin stacks. Let Z be pn`1q-Artin. There is an atlas u: U Ñ Z. Let us remark that for k P N the truncation τ u: τ U Ñ τ Z is also ďk ďk ďk a smooth atlas — indeed we have τ U » U ˆ τ Z. Let us denote by U the nerve of u and by ďk Z ďk ‚ τ U the nerve of τ u. Because k-truncated stacks are stable by flat pullbacks, the groupoid τ U ďk ‚ ďk ďk ‚ is equivalent to τ pU q. We have ďk ‚ MappZ, Y q » lim MappU , Y q » lim lim Mappτ U , Y q » lim Mappτ Z, Y q p ďk p ďk k k rpsPΔ rpsPΔ That proves item (i). If moreover Z is m-truncated, then we can replace U by τ U. If follows that ďm MappZ, Y q is a limit of pm ` qq-truncated spaces. This finishes the proof of (ii). We will use this well known lemma: Lemma 1.3.2 (see [BK, Chapter XI]). Let S : Δ Ñ sSets be a cosimplicial object in simplicial sets. Let us assume that for any rps P Δ the simplicial set S is n-coconnective. Then the natural morphism lim S Ñ lim S p p rpsPΔ rpsPΔ pďn`1 is an equivalence. op Lemma 1.3.3. Let X : N Ñ dSt be a diagram such that (i) There are two integers m and n such that for any k P N the stack Xpkq is n-Artin, m-truncated and of finite presentation; op 1 op (ii) There exists a diagram u¯: N ˆ Δ Ñ dSt such that the restriction of u¯ to N ˆ t1u is ¯ ¯ equivalent to X, every map u¯pkq: u¯pkqp0q Ñ u¯pkqp1q » Xpkq is a smooth atlas and the limit lim u¯pkq is an epimorphism. If Y is an algebraic derived stack of finite presentation then the canonical morphism ` ˘ ` ˘ ¯ ¯ colim Map X, Y Ñ Map lim X, Y is an equivalence. Proof. Let us prove the statement recursively on the Artin degree n. If n equals 0, this is a simple reformulation of the finite presentation of Y . Let us assume that the statement at hand is true for some n and let Xp0q be pn ` 1q-Artin. Considering the nerves of the epimorphisms u¯pkq, we get a diagram op op Z : N ˆ Δ Ñ dSt op ¯ ¯ Note that Z has values in n-Artin stacks. We observe that the diagram lim Zpkq : Δ Ñ dSt is k ‚ S the nerve of the map lim u¯pkq. Since lim u¯pkq is by assumption an epimorphism (whose target is k k lim X), the natural map ¯ ¯ ¯ colim lim Zpkq Ñ lim X » lim colimZpkq p p rpsPΔ kPN kPN rpsPΔ is an equivalence. We now write ˆ ˙ ` ˘ ¯ ¯ Map lim X, Y » Map colim lim Zpkq , Y rpsPΔ kPN ˆ ˙ » lim Map lim Zpkq , Y rpsPΔ kPN ` ˘ » lim colim Map Zpkq , Y rpsPΔ kPN 21 We also have ` ˘ ` ˘ ¯ ¯ colim Map X, Y » colim lim Map Zpkq , Y kPN rpsPΔ It thus suffices to prove that the canonical morphism of simplicial sets ` ˘ ` ˘ ¯ ¯ colim lim Map Zpkq , Y Ñ lim colim Map Zpkq , Y p p kPN kPN rpsPΔ rpsPΔ is an equivalence. Let us notice that each Zpkq is m-truncated. It is indeed a fibre product of m- truncated derived stacks along flat maps. Let q be an integer such that Y is q-Artin. The simplicial set MappZpkq , Y q is then pm`qq-coconnective (Lemma 1.3.1). It follows from Lemma 1.3.2 that the limit at hand is in fact finite and we have the required equivalence. op ¯ ¯ Lemma 1.3.4. Let M : N Ñ sSets be a diagram. For any i P N and any point x “ px q P lim M, we have the following exact sequence ´ ¯ // ¯ // ¯ // ¯ // 0 lim π pMpnq, x q π lim Mpnq, x lim π pMpnq, x q 0 i`1 n i i n n n n A proof of that lemma can be found for instance in [Hir]. op Lemma 1.3.5. Let M : N ˆ K Ñ sSets denote a diagram, where K is a filtered simplicial set. If for any i P N there exists N such that for any n ě N and any k P K the induced morphism i i Mpn, kq Ñ Mpn ´ 1, kq is an i-equivalence then the canonical map φ: colim lim Mpn, kq Ñ lim colim Mpn, kq kPK nPN nPN kPK is an equivalence. We recall that an i-equivalence of simplicial sets is a morphism which induces isomorphisms on the homotopy groups of dimension lower or equal to i. Proof. We can assume that K admits an initial object k . Let us write M instead of Mpn, kq. 0 nk Let us fix i P N. If i ě 1, we also fix a base point x P lim M . Every homotopy group below is n nk computed at x or at the natural point induced by x. We will omit the reference to the base point. We have a morphism of short exact sequences ´ ¯ // // // // 0 colim lim π pM q colim π lim M colim lim π pM q 0 i`1 nk i nk i nk k n k n k n ˆ ˙ // // // // 0 lim colim π pM q π lim colim M lim colim π pM q 0 i`1 nk i nk i nk n k n k n k We can restrict every limit to n ě N . Using the assumption we see that the limits on the right i`1 hand side are then constant and so are the 1-limits on the left. If follows that the vertical maps on the sides are isomorphisms, and so is the middle map. This begin true for any i, we conclude that φ is an equivalence. op ¯ ¯ Definition 1.3.6. Let X : N Ñ dSt be a diagram. We say that X is a shy diagram if (i) For any k P N the stack Xpkq is algebraic and of finite presentation; ¯ ¯ ¯ (ii) For any k P N the map Xpk Ñ k ` 1q: Xpk ` 1q Ñ Xpkq is affine; (iii) The stack Xp0q is of finite cohomological dimension. ¯ ¯ If X is the limit of X in the category of prostacks, we will also say that X is a shy diagram for X. 22 op Proposition 1.3.7. Let X : N Ñ dSt be a shy diagram. If Y is an algebraic derived stack of finite presentation then the canonical morphism ` ˘ ` ˘ ¯ ¯ colim Map X, Y Ñ Map lim X, Y is an equivalence. Proof. Since for any n, the truncation functor τ preserves shy diagrams, let us use Lemma 1.3.1 ďn and Lemma 1.3.3 ¯ ¯ Mapplim X, Y q » lim Mappτ plim Xq, Y q ďn ¯ ¯ » lim Mapplim τ X, Y q » lim colim Mappτ X, Y q ďn ďn n n On the other hand we have ¯ ¯ colim MappX, Y q » colim lim Mappτ X, Y q ďn and we are to study the canonical map ¯ ¯ φ: colim lim Mappτ X, Y q Ñ lim colim Mappτ X, Y q ďn ďn n n Because of Lemma 1.3.5, it suffices to prove the assertion (1) For any i P N there exists N P N such that for any n ě N and any k P N the map i i ` ˘ ` ˘ ¯ ¯ p : Map τ Xpkq, Y Ñ Map τ Xpkq, Y n,k ďn ďn´1 induces an equivalence on the π ’s for any j ď i. For any map f : τ Xpkq Ñ Y we will denote by F pfq the fibre of p at f. We have to prove ďn´1 n,k n,k that for any such f the simplicial set F pfq is i-connective. Let thus f be one of those maps. The n,k ¯ ¯ derived stack τ Xpkq is a square zero extension of τ Xpkq by a module Mrns, where ďn ďn´1 ´ ¯ M “ ker O Ñ O r´ns ¯ ¯ τ Xpkq τ Xpkq ďn ďn´1 Note that M is concentrated in degree 0. It follows from the obstruction theory of Y –see Proposition 0.2.10 – that F pfq is not empty if and only if the obstruction class n,k αpfq P G pfq “ Map pf L , Mrn ` 1sq n,k Y τ Xpkq ďn´1 of f vanishes. Moreover, if αpfq vanishes, then we have an equivalence F pfq » Map pf L , Mrnsq n,k Y τ Xpkq ďn´1 Using assumptions (iii) and (ii) we have that Xpkq — and therefore its truncation too — is of finite cohomological dimension d. Let us denote by ra, bs the Tor-amplitude of L . We get that G pfq is Y n,k ps ` 1q-connective for s “ a ` n ´ d and that F pfq is s-connective if αpfq vanishes. n,k Let us remark here that d and a do not depend on either k or f and thus neither does N “ i`d´a (we set N “ 0 if this quantity is negative). For any n ě N and any f as above, the simplicial set i i G pfq is at least 1-connective. The obstruction class αpfq therefore vanishes and F pfq is indeed n,k n,k i-connective. This proves (1) and concludes this proof. shy U Definition 1.3.8. Let PdSt denote the full subcategory of Pro dSt spanned by the prostacks shy op which admit shy diagrams. Every object X in PdSt is thus the limit of a shy diagram X : N Ñ dSt . 23 op We will say that X is of cotangent tor-amplitude in ra, bs if there exists a shy diagram X : N Ñ dSt for X such that every cotangent L ¯ is of tor-amplitude in ra, bs. We will also say that X Xpnq is of cohomological dimension at most d if there is a shy diagram X with values in derived stacks of cohomological dimension at most d. The pro-stack X will be called q-Artin if there is a shy diagram ra,bs shy for it, with values in q-Artin derived stacks. Let us denote by C the full subcategory of PdSt d,q S spanned by objects of cotangent tor-amplitude in ra, bs, of cohomological dimension at most d and q-Artin. shy Theorem 1.3.9. The limit functor i : PdSt Ñ dSt is fully faithful and has values in Artin shy S stacks. Proof. This follows directly from Proposition 1.3.7. Definition 1.3.10. A map of pro-stacks f : X Ñ Y if an open immersion if there exists a diagram op 1 f : N ˆ Δ Ñ dSt such that • The limit of f in maps of pro-stacks is f; op op • The restriction N ˆt0u Ñ dSt of f is a shy diagram for X and the restriction N ˆt1u Ñ dSt k k is a shy diagram for Y ; • For any n, the induced map of stacks tnu ˆ Δ Ñ dSt is an open immersion. 1.4 Uniqueness of ind-pro-structures shy,b U shy Definition 1.4.1. Let IPdSt denote the full subcategory of Ind pPdSt q spanned by colimits S S shy ra,bs of U-small filtered diagrams K Ñ PdSt which factors through C for some 4-uplet a, b, d, q. For S d,q shy,b any X P IPdSt we will say that X is of cotangent tor-amplitude in ra, bs and of cohomological ra,bs U shy dimension at most d if it is the colimit (in Ind pPdSt q) of a diagram K Ñ C . S d,q U shy Theorem 1.4.2. The colimit functor Ind pPdSt q Ñ dSt restricts to a full faithful embedding shy,b IPdSt Ñ dSt . shy ra,bs Lemma 1.4.3. Let a, b, d, q be integers with a ď b. Let T P PdSt and X : K Ñ C be a U-small S d,q filtered diagram. For any i P N there exists N such that for any n ě N and any k P K, the induced i i map ¯ ¯ Mappτ T, Xpkqq Ñ Mappτ T, Xpkqq ďn ďn´1 is an i-equivalence. We recall that an i-equivalence of simplicial sets is a morphism which induces isomorphisms on the homotopy groups of dimension lower or equal to i. Remark 1.4.4. For the proof of this lemma, we actually do not need the integer q. Proof. Let us fix i P N. Let k P K and T : N Ñ dSt be a shy diagram for T . We observe here that ¯ ¯ ¯ τ T is a shy diagram whose limit is τ T . Let also Y : N Ñ dSt be a shy diagram for Xpkq. The ďn ďn k S map at hand ¯ ¯ ψ : Mappτ T, Xpkqq Ñ Mappτ T, Xpkqq nk ďn ďn´1 is then the limit of the colimits ¯ ¯ ¯ ¯ lim colim Mappτ Tpqq, Y ppqq Ñ lim colim Mappτ Tpqq, Y ppqq ďn k ďn´1 k pPN qPN pPN qPN 24 ¯ Let now f be a map τ T Ñ Xpkq. It corresponds to a family of morphisms ďn´1 ¯ ¯ f : ˚ Ñ colim Mappτ Tpqq, Y ppqq p ďn´1 k qPN For any p, let F pfq denote the fibre of the map nk ¯ ¯ ¯ ¯ ψ : colim Mappτ Tpqq, Y ppqq Ñ colim Mappτ Tpqq, Y ppqq ďn k ďn´1 k nk qPN qPN over the point f . We also set F pfq “ lim F pfq and observe that F pfq is nothing but the fibre p nk p nk nk of ψ over f. nk To prove the result, it suffices to show that for any such f, the fibre F pfq is i-connective. Using nk the exact sequence of Lemma 1.3.4, it suffices to prove that F pfq is pi ` 1q-connective for any f and nk any p. Fixing such an f and such a p, there exists q P N such that the map f factors through the 0 p canonical map ¯ ¯ ¯ ¯ Mappτ Tpq q, Y ppqq Ñ colim Mappτ Tpqq, Y ppqq ďn´1 0 k ďn´1 k qPN We deduce that F pfq is equivalent to the colimit nk p pq F pfq » colim G pfq nk nk qěq pq where G pfq is the fibre at the point induced by f of the map nk ¯ ¯ ¯ ¯ Mappτ Tpqq, Y ppqq Ñ Mappτ Tpqq, Y ppqq ďn k ďn´1 k The interval ra, bs contains the tor-amplitude of L ¯ and d is an integer greater than the cohomo- Y ppq pq logical dimension of Tpqq. We saw in the proof of Proposition 1.3.7 that G pfq is then pa ` n ´ dq- nk connective. We set N “ i ` d ´ a ` 1. Proof (of Theorem 1.4.2). We will prove the sufficient following assertions U shy (1) The colimit functor Ind pPdSt q Ñ PpdAff q restricts to a fully faithful functor shy,b η: IPdSt Ñ PpdAff q (2) The functor η has values in the full subcategory of stacks. shy Let us focus on assertion (1) first. We consider two U-small filtered diagrams X : K Ñ PdSt and shy Y : L Ñ PdSt . We have ` ˘ ¯ ¯ ¯ ¯ Map shy colimX, colimY » lim Map shy pXpkq, colim Y q U U Ind pPdSt q Ind pPdSt q S k S and ` ˘ ` ˘ ¯ ¯ ¯ ¯ Map colimi X, colim i Y » lim Map i Xpkq, colim i Y shy shy shy shy PpdAffq PpdAffq shy shy We can thus replace the diagram X in PdSt by a simple object X P PdSt . We now assume that S S ra,bs Y factors through C for some a, b, d, q. We have to prove that the following canonical morphism d,q is an equivalence ` ˘ ¯ ¯ φ: colim Mappi X, i Y plqq Ñ Map i X, colimi Y shy shy shy shy lPL where the mapping spaces are computed in prestacks. If i X is affine then φ is an equivalence because shy colimits in PpdAff q are computed pointwise. Let us assume that φ is an equivalence whenever i X S shy is pq´1q-Artin and let us assume that i X is q-Artin. Let u: U Ñ i X be an atlas of i X and let shy shy shy 25 shy Z be the nerve of u in dSt . We saw in the proof of Lemma 1.3.3 that Z factors through PdSt . ‚ S ‚ The map φ is now equivalent to the natural map ¯ ¯ colim Mappi X, i Y plqq Ñ lim colim MappZ , i Y plqq shy shy p shy lPL rpsPΔ lPL ` ˘ ¯ ¯ » lim Map Z , colim i Y » Mappi X, colim i Y q p shy shy shy rpsPΔ Remembering Lemma 1.3.1, it suffices to study the map ¯ ¯ colim lim Mappτ i X, i Y plqq Ñ lim colim lim Mappτ Z , i Y plqq ďn shy shy ďn p shy n n lPL rpsPΔ lPL Applying Lemma 1.4.3 and then Lemma 1.3.5, we see that φ is an equivalence if the natural morphism ¯ ¯ lim colim lim Mappτ Z , i Y plqq Ñ lim lim colim Mappτ Z , i Y plqq ďn p shy ďn p shy n n lPL rpsPΔ rpsPΔ lPL is an equivalence. The stack i Y plq is by assumption q-Artin, where q does not depend on l. Now shy using Lemma 1.3.1 and Lemma 1.3.2, we conclude that φ is an equivalence. This proves (1). We now focus on assertion (2). If suffices to see that the colimit in PpdAff q of the diagram i Y as above S shy op is actually a stack. Let H : Δ Y t´1u Ñ dAff be an hypercovering of an affine SpecpAq “ H . ‚ S ´1 We have to prove the following equivalence ¯ ¯ colim lim MappH , i Y plqq Ñ lim colim MappH , i Y plqq p shy p shy l rpsPΔ rpsPΔ l Using the same arguments as for the proof of (1), we have ¯ ¯ colim lim MappH , i Y plqq » colim lim lim Mappτ H , i Y plqq p shy ďn p shy l rpsPΔ l rpsPΔ » lim colim lim Mappτ H , i Y plqq ďn p shy n l rpsPΔ » lim lim colim Mappτ H , i Y plqq ďn p shy n l rpsPΔ » lim colim lim Mappτ H , i Y plqq ďn p shy l n rpsPΔ » lim colim MappH , i Y plqq p shy rpsPΔ l shy,b We will need one last lemma about that category IPdSt . shy,b Lemma 1.4.5. The fully faithful functor IPdSt XIPdAff Ñ IPdSt Ñ dSt preserves finite S S S limits. shy,b Proof. The case of an empty limit is obvious. Let then X Ñ Y Ð Z be a diagram in IPdSt X IPdAff . There exist a and b and a diagram ´ ¯ ra,bs σ: K Ñ Fct Λ ,C 1 0,0 such that K is a U-small filtered simplicial set and the colimit in IPdSt is X Ñ Y Ð Z. We U U 2 2 can moreover assume that σ has values in FctpΛ , Pro pdAff qq » Pro pFctpΛ , dAff qq. We S S 1 1 deduce that the fibre product X ˆ Z is the realisation of the ind-pro-diagram in derived affine stacks with cotangent complex of tor amplitude in ra ´ 1, b ` 1s. It follows that X ˆ Z is again in shy,b IPdSt X IPdAff . 26 2 Symplectic Tate stacks 2.1 Tate stacks: definition and first properties We can now define what a Tate stack is. Definition 2.1.1. A Tate stack is a derived Artin ind-pro-stack locally of finite presentation whose cotangent complex – see Proposition 1.2.19 – is a Tate module. Equivalently, an Artin ind-pro-stack Tate locally of finite presentation is Tate if its tangent complex is a Tate module. We will denote by dSt the full subcategory of IPdSt spanned by Tate stacks. This notion has several good properties. For instance, using Lemma 1.2.9, if a X is a Tate stack then comparing its tangent T and its cotangent L makes sense, in the category of Tate modules X X over X. We will explore that path below, defining symplectic Tate stacks. Another consequence of Tatity is the existence of a determinantal anomaly as defined in [KV2]. Let us consider the natural morphism of prestacks U Tate θ: Tate Ñ K U U U Tate where Tate denote the prestack A ÞÑ Tate pPerfpAqq and K : A ÞÑ KpTate pPerfpAqqq – K denoting the connective K-theory functor. From [Hen2, Section 5] we have a determinant Tate K Ñ KpG , 2q where KpG , 2q is the Eilenberg-Maclane classifying stack. Definition 2.1.2. We define the Tate determinantal map as the composite map Tate Ñ KpG , 2q To any derived stack X with a Tate module E, we associate the determinantal anomaly rdet s P H pX,O q, image of E by the morphism MappX, Tate q Ñ MappX, KpG , 2qq Let now X be an ind-pro-stack. Let also R denote the realisation functor Pro dSt Ñ dSt k k mapping a pro-stack to its limit in dSt . Let finally X : K Ñ Pro dSt denote a U-small filtered k k diagram whose colimit in IPdSt is X. We have a canonical functor U U U ¯ ¯ F : lim Tate pXq » Tate pXq Ñ lim Tate pRXq P IP Definition 2.1.3. Let X be an ind-pro-stack and E be a Tate module on X. Let X be the realisation of X in Ind dSt and X be its image in dSt . We define the determinantal anomaly of E the image k k of F pEq by the map 1 1 2 Map U pX , Tate q Ñ Map U pX , KpG , 2qq » Map pX , KpG , 2qq m m Ind dSt Ind dSt dSt k k k 2 2 ˆ In particular if X is a Tate stack, we will denote by rdet s P H pX ,O q the determinantal anomaly X 2 associated to its tangent T P Tate pXq. IP The author plans on studying more deeply this determinantal class in future work. For now, let us conclude this section with following Tate Lemma 2.1.4. The inclusion dSt Ñ IPdSt preserves finite limits. or Tateness or Tatitude 27 Proof. Let us first notice that a finite limit of Artin ind-pro-stacks is again an Artin ind-pro-stack. Let now X Ñ Y Ð Z be a diagram of Tate stacks. The fibre product // X ˆ Z X p g // Z Y is an Artin ind-pro-stack. It thus suffices to test if its tangent T is a Tate module. The following Xˆ Z cartesian square concludes // T p T Xˆ Z X Y X ˚ ˚ // ˚ p T p g T Z Y Z X 2.2 Shifted symplectic Tate stacks We assume now that the basis S is the spectrum of a ring k of characteristic zero. Recall from [PTVV] the stack in graded complexes DR mapping a cdga over k to its graded complex of forms. It actually comes with a mixed structure induced by the de Rham differential. The authors also defined there the stack in graded complexes NC mapping a cdga to its graded complex of closed forms. Those two stacks are linked by a morphism NC Ñ DR forgetting the closure. p p,cl ď0 We will denote by A , A : cdga Ñ dgMod the complexes of weight p in DRr´ps and w p p,cl NC r´ps respectively. The stack A will therefore map a cdga to its complexes of p-forms while A will map it to its closed p-forms. For any cdga A, a cocycle of degree n of A pAq is an n-shifted p,cl p p-forms on Spec A. The functors A and A extend to functors op p,cl p A , A : dSt Ñ dgMod p p,cl Definition 2.2.1. Let us denote by A and A the extensions IP IP op U U pIPdSt q Ñ Pro Ind dgMod p,cl p p p,cl of A and A , respectively. They come with a natural projection A Ñ A . IP IP Let X P IPdSt . An n-shifted (closed) p-form on X is a morphism kr´ns Ñ A pXq (resp. IP p,cl p,cl p,cl p A pXq). For any closed form ω : kr´ns Ñ A pXq, the induced map kr´ns Ñ A pXq Ñ A pXq IP IP IP IP is called the underlying form of ω. Remark 2.2.2. In the above definition, we associate to any ind-pro-stack X “ colim lim X its α β αβ complex of forms p U U A pXq “ lim colim A pX q P Pro Ind dgMod αβ IP α β For any ind-pro-stack X, the derived category PIQcohpXq is endowed with a canonical monoidal structure. In particular, one defines a symmetric product E ÞÑ Sym pEq as well as an antisymmetric PI product E ^ E “ Sym pEr´1sqr2s PI PI Theorem 2.2.3. Let X be an Artin ind-pro-stack over k and π: X Ñ ˚ the projection. The push- forward functor PIQ V V π : PIQcohpXq Ñ Pro Ind pdgMod q exists (see Proposition 1.2.13) and maps L ^ L to A pXq. In particular, any 2-form kr´ns Ñ X PI X IP A pXq corresponds to a morphism O r´ns Ñ L ^ L in PIQcohpXq. X X PI X IP 28 Proof. This follows from [PTVV, 1.14], from Proposition 1.2.20 and from the equivalence IP IP U U U U U U λ ^ λ “ Pro Ind pλq ^ Pro Ind pλq » Pro Ind pλ ^ λq PI PI IP where λ is defined in the proof of Proposition 1.2.19. Definition 2.2.4. Let X be a Tate stack. Let ω : kr´ns Ñ A pXq be an n-shifted 2-form on X. It IP induces a map in the category of Tate modules on X ω: T Ñ L rns X X We say that ω is non-degenerate if the map ω is an equivalence. A closed 2-form is non-degenerate if the underlying form is. Definition 2.2.5. A symplectic form on a Tate stack is a non-degenerate closed 2-form. A symplectic Tate stack is a Tate stack equipped with a symplectic form. 2.3 Mapping stacks admit closed forms In this section, we will extend the proof from [PTVV] to ind-pro-stacks. Note that if X is a pro- ind-stack and Y is a stack, then MappX, Y q is an ind-pro-stack. We will then need an evaluation functor MappX, Y q ˆ X Ñ Y . It appears that this evaluation map only lives in the category of ind-pro-ind-pro-stacks colim lim colim lim MappX , Y q ˆ X Ñ Y αζ βξ α β ξ ζ To use this map properly, we will need the following remark. Definition 2.3.1. Let C be a category. There is one natural fully faithful functor φ: PIpCq Ñ pIPq pCq but three IPpCq Ñ pIPq pCq. The first one is given by applying IP to the canonical embedding functor C Ñ IPpCq. The second one by considering the canonical embedding functor D Ñ IPpDq for D “ IPpCq. In this work, we will only consider the third functor ψ: IPpCq Ñ pIPq pCq U U given by applying Ind to the canonical embedding D Ñ PIpDq for D “ Pro pCq. Let us also denote by ξ the natural fully faithful functor C Ñ pIPq pCq. Definition 2.3.2. Let Y be a stack and X be a pro-ind-stack. Let us denote the evaluation map in IP dSt X,Y // ev : ψ Map pX, Y q ˆ φX ξY For a formal definition of this map, we refer to [Hen1, 2.2.3.2]. We assume now that S “ Spec k. Let us recall the following definition from [PTVV, 2.1] Definition 2.3.3. A derived stack X is O-compact if for any derived affine scheme T the following conditions hold • The quasi-coherent sheaf O is compact in QcohpX ˆ Tq ; XˆT • Pushing forward along the projection X ˆ T Ñ T preserves perfect complexes. Let us denote by dSt the full subcategory of dSt spanned by O-compact derived stacks. 29 Definition 2.3.4. An O-compact pro-ind-stack is a pro-ind-object in the category of O-compact derived stacks. We will denote by PIdSt their category. Lemma 2.3.5. There is a functor ` ˘ 1 1 2 op PIdSt Ñ Fct IPdSt ˆ Δ ˆ Δ , pIPq pdgMod q k k defining for any O-compact pro-ind-stack X and any ind-pro-stack F a commutative square p,cl p,cl // A pψF ˆ φXq A pψFq b φO 2 k X IP IP p p // A pψF ˆ φXq A pψFq b φO 2 k X IP IP p,cl p p,cl p where A and A are the extensions of A and A to 2 2 IP IP IP IP op 2 2 pIPq dSt Ñ pIPq pdgMod q Proof. Recall in [PTVV, part 2.1] the construction for any O-compact stack X and any stack F of a commutative diagram (of graded complexes): w w // NC pF ˆ Xq NC pFq b π O k ˚ X // DRpF ˆ Xq DRpFq b π O k ˚ X where π: X Ñ ˚. Taking the part of weight p and shifting, we get p,cl // p,cl A pF ˆ Xq A pFq b π O k ˚ X p p // A pF ˆ Xq A pFq b π O k ˚ X This construction is functorial in both F and X so it corresponds to a functor O 1 1 op dSt Ñ FctpdSt ˆ Δ ˆ Δ , dgMod q k k We can now form the functor ` ˘ O 1 1 op PIdSt Ñ PI Fct Pro dSt ˆ Δ ˆ Δ , PropdgMod q k k ` ˘ op 1 1 Ñ Fct Pro dSt ˆ Δ ˆ Δ , PI PropdgMod q ` ˘ op 1 1 2 Ñ Fct IPdSt ˆ Δ ˆ Δ , pIPq pdgMod q By construction, for any ind-pro-stack F and any O-compact pro-ind-stack, it induces the commutative diagram p,cl p,cl // A pψF ˆ φXq ψA pFq b φO 2 k X IP IP p p // A pψF ˆ φXq ψA pFq b φO 2 k X IP IP 30 Remark 2.3.6. Let us remark that we can informally describe the horizontal maps using the maps from [PTVV]: Θ 2pψF ˆ φXq “ lim colim lim colim ΘpF ˆ X q IP αδ βγ α γ β δ Ñ lim colim lim colim ΘpF q b pO q “ ψΘ pFq b φO αδ X IP X βγ α γ β δ p,cl p where Θ is either A or A . Definition 2.3.7. Let F be an ind-pro-stack and let X be an O-compact pro-ind-stack. Let η: O Ñ p,cl p kr´ds be a map of ind-pro-k-modules. Let finally Θ be either A or A . We define the integration map id bφη // // : Θ 2pψF ˆ φXq ψΘ pFq b φO ψΘ pFqr´ds IP IP X IP Theorem 2.3.8. Let Y be a derived stack and ω be an n-shifted closed 2-form on Y . Let X be an O-compact pro-ind-stack, let π: X Ñ ˚ be the projecton, and let η: π O Ñ kr´ds be a map. The ˚ X mapping ind-pro-stack MappX, Y q admits an pn ´ dq-shifted closed 2-form. Proof. Let us denote by Z the mapping ind-pro-stack MappX, Y q. We consider the diagram ω η Y ev 2,cl 2,cl // 2,cl // // χkr´ns χA pY q A pX ˆ Zq ψA pZqr´ds IP IP ξ ψ op op op where χ: dgMod Ñ IPpdgMod q Ñ pIPq pdgMod q is the canonical inclusion. Note that since k k k op the functor ψ is fully faithful, this induces a map in IPpdgMod q 2,cl // ξk A pZqrn ´ ds IP and therefore a pn´dq-shifted closed 2-form on Z “ MappX, Y q. The underlying form is given by the composition ω η Y ev 2,cl 2 2 2 // // // // χkr´ns χA pY q χA pY q A pX ˆ Zq ψA pZqr´ds IP IP Remark 2.3.9. Let us describe the form issued by Theorem 2.3.8. We set the notations X “ lim colim X and Z “ MappX , Y q. By assumption, we have a map α β αβ αβ αβ η: colim limO Ñ kr´ds αβ For any α, there exists therefore βpαq and a map η : O Ñ kr´ds in dgModpkq. Unwinding αβpαq X αβpαq the definitions, we see that the induced form ω 2 2 // ξk A pMappX, Y qqrn ´ ds » lim colim A pZ qrn ´ ds α β αβ IP is the universal map obtained from the maps αβpαq // 2 // 2 k A pZ qrn ´ ds colim A pZ qrn ´ ds αβpαq β αβ where ω is built using η and the procedure of [PTVV]. Note that ω can be seen as a αβpαq αβpαq αβpαq map T b T Ñ O . We also know from Theorem 2.2.3 that the form ω induces a X X X Y αβpαq αβpαq αβpαq map T b T Ñ O rn ´ ds Z Z Z 31 in IPPpZq. Let us fix α and pull back the map above to Z . We get 0 α ˚ ˚ ˚ colim lim g p pT b T q » i pT b T q Ñ O rn ´ ds Z Z Z Z Z α α αβ αβ αβ α α 0 0 0 αěα 0 β This map is the universal map obtained from the maps ˚ ˚ ˚ ˚ lim g p pT b T q Ñ g p pT b T q Z Z Z Z α0α αβ αβ αβ α0α αβpαq αβpαq αβpαq ˚ ˚ Ñ g p pO qrn ´ ds » O rn ´ ds X X α α α 0 αβpαq αβpαq where g is the structural map Z Ñ Z and p is the projection Z “ lim Z Ñ Z . α α α α αβ α β αβ αβ 0 0 2.4 Mapping stacks have a Tate structure Definition 2.4.1. Let S be an O-compact pro-ind-stack. We say that S is an O-Tate stack if there op exist a poset K and a diagram S : K Ñ Ind dSt such that (i) The limit of S in PIdSt is equivalent to S ; (ii) For any i ď j P K the pro-module over Spiq ` ˘ coker O ¯ Ñ Spi ď jq O ¯ Spiq Spjq is trivial in the pro-direction – ie belong to QcohpSpiqq. (iii) For any i ď j P K the induced map Spi ď jq is represented by a diagram f : L ˆ Δ Ñ dSt such that ¯ ¯ • For any l P L the projections fpl, 0q Ñ ˚ and fpl, 1q Ñ ˚ satisfy the base change formula ; • For any l P L the map fplq satisfies the base change and projection formulae ; • For any m ď l P L the induced map fpm ď l, 0q satisfies the base change and projection formulae. Remark 2.4.2. We will usually work with pro-ind-stacks S given by an explicit diagram already satisfying those assumptions. Proposition 2.4.3. Let us assume that Y is a derived Artin stack locally of finite presentation. Let S be an O-compact pro-ind-stack. If S is an O-Tate stack then the ind-pro-stack MappS, Y q is a Tate stack. op U Proof. Let Z “ MappS, Y q as an ind-pro-stack. Let S : K Ñ Ind dSt be as in Definition 2.4.1. ¯ ¯ ¯ We will denote by Z : K Ñ Pro dSt the induced diagram and for any i P K by s : Zpiq Ñ Z the k i induced map. Let us first remark that Z is an Artin ind-pro-stack locally of finite presentation. It suffices to prove that s L is a Tate module on Zpiq, for any i P K. Let us fix such an i and denote by Z the Z i pro-stack Zpiq. We consider the differential map s L Ñ L Z Z i i It is by definition equivalent to the natural map Pro Pro lim λ pZ| ěiq Ñ λ pZ q K i Z Z i i 32 ěi { ěi where K is the comma category K and Z| is the induced diagram ěi Z K Ñ Pro dSt Let φ denote the diagram ` ˘ op ěi φ : K Ñ IPerfpZ q i i obtained as the kernel of f. It is now enough to prove that φ factors through PerfpZ q. i i Let j ě i in K and let us denote by g the induced map Z Ñ Z of pro-stacks. Let f : LˆΔ Ñ ij i j ¯ ¯ ¯ dSt represents the map Spi ď jq: Spjq Ñ Spiq P Ind dSt as in assumption (i) in Definition 2.4.1. k k Up to a change of L through a cofinal map, we can assume that the induced diagram ` ˘ coker O ¯ Ñ Spi ď jq O ¯ Spiq Spjq op 1 is essentially constant – see assumption (ii). We denote by h: L ˆΔ Ñ dSt the induced diagram, ¯ ¯ so that g is the limit of h in Pro dSt . For any l P L we will denote by h : Z Ñ Z the map hplq. ij k l il jl ¯ ¯ Let us denote by Z the induced diagram l ÞÑ Z and by Z the diagram l ÞÑ Z . Let also p denote i il j jl l the projection Z Ñ Z i il We have an exact sequence ˚ ˚ ˚ φ pjq Ñ colim p h L Ñ colim p L i Z Z l l jl l il l l Let us denote by ψ the diagram obtained as the kernel ij Pro Pro ¯ ¯ ψ Ñ λ pZ q Ñ λ pZ q ij j i Z Z i i so that φ pjq is the colimit colimψ in IPerfpZ q. It suffices to prove that the diagram ψ : L Ñ i ij i ij PerfpZ q is essentially constant (up to a cofinal change of posets). By definition, we have ψ plq » p L r´1s ij l Z {Z il jl Let m Ñ l be a map in L and t the induced map Z Ñ Z . The map ψ pm Ñ lq is equivalent to il im ij the map p ξ where ξ fits in the fibre sequence in PerfpZ q il ˚ ˚ ˚ ˚ // // t L r´1s t h L t L Z {Z Z Z im jm m jm im // // L r´1s h L L Z Z Z {Z jl il il jl l We consider the dual diagram ˚ ˚ ˚ ˚ oo oo t T r1s t h T t T Z {Z Z Z im jm m jm im OO OO OO pσq oo oo T r1s h T T Z Z Z {Z jl il il jl l Using base change along the maps from S , S and S to the point, we get that the square pσq is im jm jl equivalent to ˚ oo ˚ π pid ˆsf q pid ˆsf q E π pid ˆsq pid ˆsq E ˚ m ˚ m ˚ ˚ OO OO oo π pid ˆf q pid ˆf q E π E ˚ l ˚ l ˚ 33 where π: Z ˆ S Ñ Z is the projection, where s: S Ñ S is the map induced by m Ñ l and il il il im il where E » ev T with ev: Z ˆ S Ñ Y the evaluation map. Note that we use here the well known Y il il fact T » pr ev T where MappX,Y q Y pr ev oo // MappX, Y q MappX, Y q ˆ X Y are the canonical maps. Now using the projection and base change formulae along the morphisms s, f and f we get l m that pσq is equivalent to the image by π of the square ˚ oo ˚ E b p s f O E b p s O ˚ m S ˚ S ˚ jm im OO OO ˚ oo ˚ E b p f O E b p O l S S ˚ jl il We therefore focus on the diagram oo s f O s O ˚ m S ˚ S jm im OO OO oo f O O l S S ˚ jl il The map induced between the cofibres is an equivalence, using assumption (ii). It follows that the diagram ψ is essentially constant, and thus that Z is a Tate stack. ij 3 Formal loops In this part, we will at last define and study the higher dimensional formal loop spaces. We will prove it admits a local Tate structure. 3.1 Dehydrated algebras and de Rham stacks In this part, we define a refinement of the reduced algebra associated to a cdga. This allows us to define a well behaved de Rham stack associated to an infinite stack. Indeed, without any noetherian assumption, the nilradical of a ring – the ideal of nilpotent elements – is a priori not nilpotent itself. The construction below gives an alternative definition of the reduced algebra – which we call the dehydrated algebra – associated to any cdga A, so that A is, in some sense, a nilpotent extension of its dehydrated algebra. Whenever A is finitely presented, this construction coincides with the usual reduced algebra. ď0 Definition 3.1.1. Let A P cdga . We define its dehydrated algebra as the ind-algebra A “ deh 0 0 colim H pAq where the colimit is taken over the filtered poset of nilpotent ideals of H pAq. The I { case I “ 0 gives a canonical map A Ñ A in ind-cdga’s. This construction is functorial in A. deh Remark 3.1.2. Whenever A is of finite presentation, then A is equivalent to the reduced algebra deh associated to A. In that case, the nilradical A of A is nilpotent. Moreover, if A is any cdga, it is a filtered colimits of cdga’s A of finite presentation. We then have A » colimpA q in ind-algebras. α deh α red Lemma 3.1.3. The colimit B of the ind-algebra A in the category of algebras is equivalent to the deh reduced algebra A . red Proof. Let us first remark that B is reduced. Indeed any nilpotent element x of B comes from a nilpotent element of A. It therefore belongs to a nilpotent ideal pxq. This define a natural map of algebras A Ñ B. To see that it is an isomorphism, it suffices to say that A is the union of all red nilpotent ideals. 34 Definition 3.1.4. Let X be a prestack. We define its de Rham prestack X as the composition dR p´q Ind pXq deh colim ď0 U ď0 U // // // cdga Ind pcdga q Ind psSetsq sSets k k This defines an endofunctor of p8, 1q-category PpdAff q. We have by definition ´ ¯ X pAq “ colimX H pAq dR Remark 3.1.5. If X is a stack of finite presentation, then it is determined by the images of the cdga’s of finite presentation. The prestack X is then the left Kan extension of the functor dR ď0,fp cdga Ñ sSets A ÞÑ XpA q red Definition 3.1.6. Let f : X Ñ Y be a functor of prestacks. We define the formal completion X of X in Y as the fibre product // X X Y dR // Y Y dR This construction obviously defines a functor FC: PpdAff q Ñ PpdAff q. k k ď0 Remark 3.1.7. The natural map X Ñ Y is formally étale, in the sense that for any A P cdga and any nilpotent ideal I Ă H pAq the morphism ` ˘ ˆ ˆ X pAq Ñ X H pAq ˆ Y pAq Y Y { Y pH pAq{ q is an equivalence. 3.2 Higher dimensional formal loop spaces Here we finally define the higher dimensional formal loop spaces. To any cdga A we associate the d d formal completion V of 0 in A . We see it as a derived affine scheme whose ring of functions ArrX ss 1...d A A is the algebra of formal series in d variables X , . . . , X . Let us denote by U the open subscheme of 1 d ď0 V complementary of the point 0. We then consider the functors dSt ˆ cdga Ñ sSets A k d d L : pX, Aq ÞÑ Map pV , Xq V dSt A d d L : pX, Aq ÞÑ Map pU , Xq U dSt A d d ˜ ˜ Definition 3.2.1. Let us consider the functors L and L as functors dSt Ñ PpdAffq. They come U V d d d d ˜ ˜ ˜ ˜ with a natural morphism L Ñ L . We define L to be the pointwise formal completion of L into V U V L : ´ ¯ d d d ˜ ˜ ˜ L pXq “ FC L pXq Ñ L pXq V U d d d d d d ˜ ˜ ˜ We also define L , L and L as the stackified version of L , L and L respectively. We will call U V U V L pXq the formal loop stack in X. Remark 3.2.2. The stack L pXq is a higher dimensional analogue to the stack of germs in X, as studied for instance by Denef and Loeser in [DL]. 35 d Remark 3.2.3. By definition, the derived scheme U is the (finite) colimit in derived stacks ´ ¯ d ´1 U “ colim colim Spec ArrX ssrX s 1...d i ...i 1 q q i ,...,i 1 q ´1 ´1 ´1 where ArrX ssrX s denote the algebra of formal series localized at the generators X , . . . , X . 1...d i ...i i i 1 q 1 q It follows that the space of A-points of L pXq is equivalent to the simplicial set ´ ´ ¯ ¯ d ´1 I L pXqpAq » colim lim lim Map Spec ArrX ssrX s , X 1...d i ...i 1 q q i ,...,i IĂH pAq 1 q ´1 ´1 where ArrX ssrX s is the sub-cdga of ArrX ssrX s consisting of series 1...d 1...d i ...i i ...i 1 q 1 q n n 1 d a X . . .X n ,...,n 1 d 1 d n ,...,n 1 d where a is in the kernel of the map A Ñ H pAq as soon as at least one of the n ’s is negative. n ,...,n { i 1 d I Recall that in the colimit above, the symbol I denotes a nilpotent ideal of H pAq. Lemma 3.2.4. Let X be a derived Artin stack of finite presentation with algebraisable diagonal (see d d Definition 0.2.11) and let t: T “ SpecpAq Ñ X be a smooth atlas. The induced map L pTq Ñ L pXq V V is an epimorphism of stacks. d d ˜ ˜ Proof. It suffices to study the map L pTq Ñ L pXq. Let B be a cdga. Let us consider a B-point V V x: Spec B Ñ L pXq. It induces a B-point of X Spec B Ñ SpecpBrrX ssq Ñ X 1...d Because t is an epimorphism, there exists an étale map f : Spec C Ñ Spec B and a commutative diagram // Spec C T f t // Spec B X It corresponds to a C-point of Spec B ˆ T . For any n P N, let us denote by C the cdga X n C :“ Crx , . . ., x s n n n 1 d px , . . ., x q 1 d n n and by S the spectrum Spec C We also set B “ Brx , . . . , x s and X “ Spec B . n n n 1 d n n px , . . . , x q 1 d Finally, we define T as the pullback T ˆ X . We will also consider the natural fully faithful functor n X n Δ » t0, . . ., nu Ñ N. We have a natural diagram 2,2 2 0 α : Λ ˆ N > Δ ˆ Δ Ñ dSt 0 k 2,2 0 Λ ˆΔ informally drown has a commutative diagram // // // S . . . S . . . 0 n // // // X . . . X . . . 0 n OO OO !! // // // T . . . T . . . 0 n 36 Let n P N and let us assume we have built a diagram 2,2 2 n α : pΛ ˆ Nq > Δ ˆ Δ Ñ dSt n k 2,2 n Λ ˆΔ extending α . There is a sub-diagram of α n´1 n // S S n n`1 // T T n n`1 n`1 n`1 Since the map t is smooth (it is a pullback of t), we can complete this diagram with a map n`1 S Ñ T and a commutative square. Using the composition in dSt , we get a diagram α n`1 n`1 k n`1 extending α . We get recursively a diagram α: Δ ˆ N Ñ dSt . Taking the colimit along N, we get n k a commutative diagram // // Spec C colim Spec C T f t // // // Spec B colim Spec B SpecpBrrX ssq X n n 1...d This defines a map φ: colim SpecpC q Ñ SpecpBrrX ssq ˆ T . We have the cartesian diagram n 1...d X // SpecpBrrX ssq ˆ T X 1...d X // SpecpBrrX ssq ˆ T X ˆ X 1...d The diagonal of X is algebraisable and thus so is the stack SpecpBrrX ssq ˆ T . The morphism φ 1...d X therefore defines the required map SpecpCrrX ssq Ñ SpecpBrrX ssq ˆ T 1...d 1...d Remark 3.2.5. Let us remark here that if X is an algebraisable stack, then L pXq is a stack, hence the natural map is an equivalence d d L pXq » L pXq V V ď0 Lemma 3.2.6. Let f : X Ñ Y be an étale map of derived Artin stacks. For any cdga A P cdga and any nilpotent ideal I Ă H pAq, the induced map ` ˘ d d 0 d ˜ // ˜ ˜ θ: L pXqpAq L pXq H pAq ˆ L pY qpAq I U U U d 0 L pY qpH pAq q U I is an equivalence. Proof. The map θ is a finite limit of maps ` ` ˘˘ // µ : XpξAq X ξ H pAq Y pξAq { ˆ Y ξ H pAq p p { qq 37 ´1 where ξA “ ArrX ssrX s is obtained from the cdga of formal power series in A with d variables 1...d i ...i 1 p 0 0 by inverting the variables x . Let also ξpH pAq q be defined similarly. The natural map ξpH pAqq Ñ i I ξpH pAq q is also a nilpotent extension. We deduce from the étaleness of f that the map ` ` ˘˘ 0 0 0 // XpξpH pAqqq X ξ H pAq ˆ Y pξpH pAqqq Y pξpH pAq qq is an equivalence. Let now n P N. We assume that the natural map ` ` ˘˘ // XpξpA qq X ξ H pAq ˆ Y pξpA qq ďn { ďn Y pξpH pAq qq ´n´1 is an equivalence. The cdga ξpA q » pξAq is a square zero extension of ξpA q by H pξAq. ďn`1 ďn`1 ďn We thus have the equivalence // XpξpA qq XpξpA qq ˆ Y pξpA qq ďn`1 ďn ďn`1 Y pξpA qq ďn The natural map ` ` ˘˘ // 0 XpξpA qq X ξ H pAq ˆ Y pξpA qq ďn`1 I ďn`1 Y ξ H pAq p p { qq is thus an equivalence too. The stacks X and Y are nilcomplete, hence µ is also an equivalence – recall that a derived stack X is nilcomplete if for any cdga B we have XpBq » lim XpB q ďn Also recall that any Artin stack is nilcomplete. It follows that θ is an equivalence. ď0 Corollary 3.2.7. Let f : X Ñ Y be an étale map of derived Artin stacks. For any cdga A P cdga and any nilpotent ideal I Ă H pAq, the induced map ` ˘ d d 0 d ˜ // ˜ ˜ θ: L pXqpAq L pXq H pAq ˆ L pY qpAq d 0 L pY qpH pAq q is an equivalence. Proposition 3.2.8. Let X be a derived Deligne-Mumford stack of finite presentation with algebraisable d d diagonal. Let t: T Ñ X be an étale atlas. The induced map L pTq Ñ L pXq is an epimorphism of stacks. ď0 d d ˜ ˜ Proof. We can work on the map of prestacks L pTq Ñ L pXq. Let A P cdga . Let x be an A-point of L pXq. It corresponds to a vertex in the simplicial set ` ˘ d 0 d ˜ ˜ colimL pXq H pAq ˆ L pXqpAq V I U d 0 pXq H pAq L p { q U I There exists therefore a nilpotent ideal I such that x comes from a commutative diagram d d // U U H pAq // V 0 X H pAq v 38 Using Lemma 3.2.4 we get an étale morphism ψ: A Ñ B such that the map v lifts to a map u: V Ñ T where J is the image of I by ψ. This defines a point in ` ˘ d 0 ˜ ˜ L pTq H pBq ˆ L pXqpBq U U d 0 L pXq H pBq p { q U J Because of Lemma 3.2.6, we get a point of L pTqpBq. We now observe that this point is compatible with x. In the case of dimension d “ 1, Lemma 3.2.6 can be modified in the following way. Let f : X Ñ Y ď0 be a smooth map of derived Artin stacks. For any cdga A P cdga and any nilpotent ideal I Ă H pAq, the induced map ` ˘ 1 1 0 1 ˜ // ˜ ˜ θ: L pXqpAq L pXq H pAq ˆ L pY qpAq I U U U 1 0 L pY qpH pAq q U I is essentially surjective. The following proposition follows. Proposition 3.2.9. Let X be an Artin derived stack of finite presentation and with algebraisable 1 1 diagonal. Let t: T Ñ X be a smooth atlas. The induced map L pTq Ñ L pXq is an epimorphism of stacks. 1 1 Example 3.2.10. The proposition above implies for instance that L pBGq » BL pGq for any algebraic group G – where BG is the classifying stack of G-bundles. 3.3 Tate structure and determinantal anomaly We saw in subsection 2.1 that to any Tate stack X, we can associate a determinantal anomaly. It a 2 d class in H pX,O q. We will prove in this subsection that the stack L pXq is endowed with a structure of Tate stack as soon as X is affine. We will moreover build a determinantal anomaly on L pXq for any quasi-compact and separated scheme X. ď0 Lemma 3.3.1. For any B P cdga of finite presentation, the functors ď0 d d ˜ ˜ L pSpec Bq,L pSpec Bq: cdga Ñ sSets are in the essential image of the fully faithful functor shy,b IPdSt X IPdAff Ñ IPdSt Ñ dSt Ñ PpdAffq k k k d d d d ˜ ˜ (see Definition 1.4.1). It follows that L pSpec Bq » L pSpec Bq and L pSpec Bq » L pSpec Bq. U U Proof. Let us first remark that Spec B is a retract of a finite limit of copies of the affine line A . It follows that the functor L pSpec Bq is, up to a retract, a finite limit of functors ´ ¯ d ´1 Z : A ÞÑ Map krY s, ArrX ssrX s E 1...d i ...i 1 q where E “ ti , . . . , i u Ă F “ t1, . . ., du. The functor Z is the realisation of an affine ind-pro-scheme 1 q Z » colim lim Specpkra , ´nδ ď α ď psq α ,...,α i i E 1 d n p where δ “ 1 if i P E and δ “ 0 otherwise. The variable a corresponds to the coefficient i i α ,...,α 1 d α α shy,b 1 d d of X . . .X . The functor Z is thus in the category IPdSt X IPdAff . The result about 1 E d d ˜ ˜ L pSpec Bq then follows from Lemma 1.4.5. The case of L pSpec Bq is similar: we decompose it into a finite limit of functors ´ ¯ d ´1 I G : A ÞÑ colim Map krY s, ArrX ssrX s 1...d E i ...i 1 q IĂH pAq 39 0 d where I is a nilpotent ideal of H pAq. We then observe that G is the realisation of the ind-pro-scheme ´ ¯ G » colim lim Spec kra , ´nδ ď α ď ps α ,...,α i i { E 1 d n,m p where J is the ideal generated by the symbols a with at least one of the α ’s negative. α ,...,α 1 d Remark 3.3.2. Let n and p be integers and let kpE, n, pq denote the number of families pα , . . . , α q 1 d such that ´nδ ď α ď p for all i. We have i i d 1 kpE,n,pq Z » colim limpA q n p fp Definition 3.3.3. From Lemma 3.3.1, we get a functor L : dAff Ñ IPdSt . It follows from Proposition 3.2.8 that L is a costack in ind-pro-stacks. We thus define d lfp L : dSt Ñ IPdSt fp lfp lfp to be its left Kan extension along the inclusion dAff Ñ dSt – where dSt is p8, 1q-category of k k k derived stacks locally of finite presentation. This new functor L preserves small colimits by definition. Proposition 3.3.4. There is a natural transformation θ from the composite functor d IP L |´| lfp // // dSt IPdSt dSt k k to the functor L . Moreover, the restriction of θ to derived Deligne-Mumford stacks of finite presen- tation with algebraisable diagonal is an equivalence. Proof. There is by definition a natural transformation d IP d θ: |L p´q| Ñ L p´q Moreover, the restriction of θ to affine derived scheme of finite presentation is an equivalence – see Lemma 3.3.1. The fact that θ is an equivalence for any Deligne-Mumford stack X follows from Proposition 3.2.8. Lemma 3.3.5. Let F be a non-empty finite set. For any family pM q of complexes over k indexed by subsets D of F , we have colim M » M rd ´ 1s D F H‰EĂF H‰DĂE where d is the cardinal of F (the maps in the colimit diagram are the canonical projections). Proof. We can and do assume that F is the finite set t1, . . ., du and we proceed recursively on d. The case d “ 1 is obvious. Let now d ě 2 and let us assume the statement is true for F r tdu. Let pM q be a family as above. We have a cocartesian diagram à à // colim M colim M D D tduĹEĂF H‰EĂFrtdu H‰DĂE H‰DĂE // M colim M tdu H‰EĂF H‰DĂE We have by assumption colim M » M rd ´ 2s D Frtdu H‰EĂFrtdu H‰DĂE 40 and ˜ ¸ ˜ ¸ à à à colim M » M ‘ colim M ‘ colim M D tdu D D tduĹEĂF tduĹEĂF tduĹEĂF H‰DĂE tduĹDĂE H‰DĂErtdu » M ‘ M rd ´ 2s ‘ M rd ´ 2s tdu F Frtdu The result follows. ď0 d Lemma 3.3.6. For any B P cdga of finite presentation, the ind-pro-stack L pSpec Bq is a Tate stack. Proof. Let us first focus on the case of the affine line A . We have to prove that the cotangent p,n complex L d is a Tate module. For any subset D Ă F we define M to be the free k-complex L pA q D generated by the symbols ta , ´n ď α ă 0 if i P D, 0 ď α ď p otherwiseu α ,...,α i i 1 d in degree 0. From the proof of Lemma 3.3.1, we have p,n d d 1 d Z » colim lim Specpkr M sq and L pA q » lim Z E DĂE D U E n p H‰EĂF where F “ t1, . . ., du. If we denote by π the projection L pA q Ñ Spec k, we get ˜ ¸ ˜ ¸ à à ˚ p,n ˚ p,n L d » π colim lim colim M » π lim colim colim M L pA q D D n p n p H‰EĂF H‰EĂF DĂE DĂE Using Lemma 3.3.5 we have ˆ ˙ p,n p,n L d » π lim colimM ‘ M rd ´ 1s L pA q H F n p p,n p,0 p,n 0,n Moreover, we have M » M and M » M . It follows that L d is a Tate module H H F F L pA q d d on the ind-pro-stack L pA q. The case of L pSpec Bq then follows from Lemma 1.4.5 and from U U Lemma 2.1.4. Lemma 3.3.7. Let B Ñ C be an étale map between cdga’s of finite presentation. The induced map d d f : L pSpec Cq Ñ L pSpec Bq is formally étale – see Definition 1.2.17. U U Proof. Let us denote X “ Spec B and Y “ Spec C. We have to prove that the induced map ´ ¯ ´ ¯ d d d d j : Map d L pY qr´s,L pY q Ñ Map d L pY qr´s,L pXq L pY q{´ U U L pY q{´ U U U U d d ď0 is an equivalence of functors PIQcohpL pY qq Ñ sSets. Since L pY q is ind-pro-affine, we can restrict to the study of the morphism ´ ¯ ´ ¯ d d j : Map Zr´s,L pY q Ñ Map Zr´s,L pXq Z{´ U Z{´ U ď0 of functors IQcohpZq Ñ sSets, for any pro-affine scheme Z and any map Z Ñ L pY q. Let us fix ď0 E P IQcohpZq . The pro-stack ZrEs is in fact an affine pro-scheme. Recall that both L pY q and d shy,b L pXq belong to IPdSt . It follows from the proof of Theorem 1.4.2 that the morphism j pEq is U k equivalent to ` ˘ ` ˘ d d |j pEq|: Map |ZrEs|,L pY q Ñ Map |ZrEs|,L pXq |Z|{´ U |Z|{´ U where | ´ | is the realisation functor and the mapping spaces are computed in dSt . It now suffices to see that |ZrEs| is a trivial square zero extension of the derived affine scheme |Z| and to use Lemma 3.2.6. 41 Proposition 3.3.8. Let Spec B be a derived affine scheme of finite presentation. The ind-pro- stack L pSpec Bq admits a cotangent complex. This cotangent complex is moreover a Tate module. d d For any étale map B Ñ C the induced map f : L pSpec Cq Ñ L pSpec Bq is formally étale – see Definition 1.2.17. d d Proof. Let us write Y “ Spec B. Let us denote by i: L pY q Ñ L pY q the natural map. We will prove that the map i is formally étale, the result will then follow from Lemma 3.3.6 and Lemma 3.3.7. To do so, we consider the natural map ´ ¯ ´ ¯ d d d d j : Map d L pY qr´s,L pY q Ñ Map d L pY qr´s,L pY q L pY q{´ L pY q{´ U ď0 of functors PIQcohpL pY qq Ñ sSets. To prove that j is an equivalence, we can consider for every ď0 affine pro-scheme X Ñ L pY q the morphism of functors IQcohpXq Ñ sSets ´ ¯ ´ ¯ d d j : Map Xr´s,L pY q Ñ Map Xr´s,L pY q X{´ X{´ U ď0 Let us fix E P IQcohpXq . The morphism j pEq is equivalent to ` ˘ ` ˘ d d |j pEq|: Map |XrEs|,L pY q Ñ Map |XrEs|,L pY q |X|{´ |X|{´ U where the mapping space are computed in dSt . The map |j pEq| is a pullback of the map k X ` ˘ ` ˘ d d f : Map |XrEs|,L pY q Ñ Map |XrEs|,L pY q dR dR |X|{´ V |X|{´ U It now suffices to see that |XrEs| is a trivial square zero extension of the derived affine scheme |X| and thus f is an equivalence (both of its ends are actually contractible). Let us recall from Definition 2.1.3 the determinantal anomaly ´ ¯ 2 d ˆ rDet d s P H L pSpec Aq,O L pSpecAq L pSpecAq U d It is associated to the tangent T d P Tate pL pSpec Aqq through the determinant map. Using L pSpecAq IP Proposition 3.3.8, we see that this construction is functorial in A, and from Proposition 3.2.8 we get that it satisfies étale descent. Thus, for any quasi-compact and quasi-separated (derived) scheme (or Deligne-Mumford stack with algebraisable diagonal), we have a well-defined determinantal anomaly ´ ¯ 2 d ˆ rDet d s P H L pXq,O L pXq d L pXq Remark 3.3.9. It is known since [KV3] that in dimension d “ 1, if rDet 1 s vanishes, then there L pXq are essentially no non-trivial automorphisms of sheaves of chiral differential operators on X. 4 Bubble spaces In this section, we study the bubble space, an object closely related to the formal loop space. We will then prove the bubble space to admit a symplectic structure. 4.1 Two lemmas In this subsection, we will develop two duality results we will need afterwards. ď0 0 0 Let A P cdga be a cdga over a field k. Let pf , . . . , f q be points of A whose images in H pAq 1 p form a regular sequence. 42 n n Let us denote by A the Kozsul complex associated to the regular sequence pf , . . ., f q for n,k 1 k k ď p. We set A “ A and A “ A for any n. If k ă p, the multiplication by f induces an n,0 n n,p k`1 n n endomorphism ϕ of A . Recall that A is isomorphic to the cone of ϕ : n,k n,k`1 k`1 k`1 k`1 // A A n,k n,k // 0 A n,k`1 Let us now remark that for any couple pn, kq, the A-module A is perfect. n,k Lemma 4.1.1. Let k ď p. The A-linear dual A “ RHom pA , Aq of A is equivalent to n,k n,k n,k A r´ks; n,k Proof. We will prove the statement recursively on the number k. When k “ 0, the result is trivial. Let k ě 0 and let us assume that A is equivalent to A r´ks. Let us also assume that for any n,k n,k a P A, the diagram induced by multiplication by a commutes { „ A A r´ks n,k n,k _ „ A A r´ks n,k n,k We obtain the following equivalence of exact sequences k`1 // // A r´k ´ 1s A r´ks A r´ks n,k`1 n,k n,k „ „ „ pϕ q k`1 _ _ _ { { { A A A // // A A A n,k`1 n,k n,k The statement about multiplication is straightforward. Lemma 4.1.2. Let us assume A is a formal series ring over A : A “ A rrf , . . ., f ss 1 1 p It follows that for any n, the A -module A is free of finite type and that there is map r : A Ñ A 1 n n n 1 n n mapping f . . . f to 1 and any other generator to zero. We deduce an equivalence 1 p „ { A Ñ A “ RHom pA , A q n n 1 given by the pairing ˆ r // // A b A A A n A n n 1 Remark 4.1.3. Note that we can express the inverse A Ñ A of the equivalence above: it maps n n a function α: A Ñ A to the serie n 1 i n´1´i αpf qf i p i 1 where i varies through the uplets pi , . . . , i q and where f “ f . . . f . 1 p p 43 4.2 Definition and properties We define here the bubble space, obtained from the formal loop space. We will prove in the next sections it admits a structure of symplectic Tate stack. Definition 4.2.1. The formal sphere of dimension d is the pro-ind-stack S “ lim colim SpecpA ‘ Hom pA , Aqq » lim colim SpecpA ‘ A r´dsq p n p n n pěn n pěn n n where A “ krx , . . ., x s and A “ A . 1 d n px , . . . , x q 1 d Remark 4.2.2. The notation SpecpA ‘ A r´dsq is slightly abusive. The cdga A ‘ A r´ds is not p n p n concentrated in non positive degrees. In particular, the derived stack SpecpA ‘ A r´dsq is not a p n derived affine scheme. It behaves like one though, regarding its derived category: QcohpSpecpA ‘ A r´dsqq » dgMod p n A ‘A r´ds p n Let us define the ind-pro-algebra O “ colim lim A ‘ A r´ds ˆd p n n pěn where A ‘ A r´ds is the trivial square zero extension of A by the module A r´ds. For any m P N, p n p n let us denote by S the ind-stack S “ colim SpecpA ‘ A r´dsq p m pěm Definition 4.2.3. Let T be a derived Artin stack. We define the d-bubble stack of T as the mapping ind-pro-stack BpTq “ MappS , Tq: Spec B ÞÑ colim lim TpB b pA ‘ A r´dsqq p n n pěn Again, the cdga A ‘A r´ds is not concentrated in non positive degree. This notation is thus slightly p n abusive and by TpB b pA ‘ A r´dsqq we mean p n MappSpecpA ‘ A r´dsq ˆ Spec B, Tq p n We will denote by BpTq the diagram N Ñ Pro dSt of whom BpTq is a colimit in IPdSt . Let us k k also denote by B pTq the mapping pro-stack B pTq “ MappS , Tq: Spec B ÞÑ lim TpB b pA ‘ A r´dsqq p m m m pěm op and B pTq: tp P N|p ě mu Ñ dSt the corresponding diagram. In particular B pTq “ MappS , Tq: Spec B ÞÑ lim TpB b A q 0 0 Those stacks come with natural maps 0 r // // B pTq BpTq B pTq 0 0 // B pTq BpTq Proposition 4.2.4. If T is an affine scheme of finite type, the bubble stack BpTq is the product in ind-pro-stacks // BpTq L pTq d d // L pTq L pTq V U 44 d d Proof. There is a natural map V Ñ S induced by the morphism colim lim A ‘ A r´ds Ñ lim A p n p n pěn p Because T is algebraisable, it induces a map BpTq Ñ L pTq and thus a diagonal morphism d d δ: BpTq Ñ L pTq ˆ L pTq V V L pTq We will prove that δ is an equivalence. Note that because T is a (retract of a) finite limit of copies of A , d d 1 1 1 we can restrict to the case T “ A . Let us first compute the fibre product Z “ L pA qˆ d L pA q. V L pA q V It is the pullback of ind-pro-stacks // Z lim Specpkra , 0 ď α ď psq α ,...,α i // lim Specpkra , 0 ď α ď psq colim lim lim Specpkra , ´nδ ď α ď psq α ,...,α i α ,...,α iPI i 1 d 1 d p n p IĂJ p,n where J “ t1, . . ., du and δ “ 1 if i P I and 0 otherwise. For any subset K Ă J we define M to iPI be the free complex generated by the symbols ta , ´n ď α ă 0 if i P K, 0 ď α ď p otherwiseu α ,...,α i i We then have the cartesian diagram ´ ¯ p,0 // Z lim Spec krM s ´ ¯ p,0 p,n // lim Spec krM s colim lim lim Specpkr M sq p n p IĂJ H KĂI K Using Lemma 3.3.5 we get ´ ” ı¯ p,0 0,n Z » colim lim Spec k M ‘ M rds H J n p Remark 4.2.5. Let us consider the map lim A Ñ A » k mapping a formal serie to its coefficient p p 0 of degree 0. The plim A q-ind-module colimA r´ds is endowed with a natural map to kr´ds. This p n d d d induces a morphism O Ñ k ‘ kr´ds and hence a map S Ñ S , where S is the topological sphere ˆd of dimension d. We then have a rather natural morphism B pXq Ñ MappS , Xq 4.3 Its tangent is a Tate module We will prove in this subsection that the bubble stack is a Tate stack. To do so, we could bluntly apply Proposition 2.4.3 but we will give here a direct proof of that statement. We will get another decomposition of its tangent complex that will be needed when proving B pTq is symplectic. Proposition 4.3.1. Let us assume that the Artin stack T is locally of finite presentation. For any m P N we have an exact sequence ˚ ˚ // ˚ // ˚ s r L d s L d s L d d m B pTq m B pTq m B pTq{B pTq 0 0 where the left hand side is an ind-perfect module and the right hand side is a pro-perfect module. In particular, the middle term is a Tate module, and the ind-pro-stack B pTq is a Tate stack. 45 d d Proof. Throughout this proof, we will write B instead of B pTq and B instead of B pTq for any m. Let us first remark that B is an Artin ind-pro-stack locally of finite presentation. It suffices to prove that s L is a Tate module on B , for any m P N. We will actually prove that it is an m m elementary Tate module. We consider the map ˚ ˚ ˚ s r L Ñ s L B B m m It is by definition equivalent to the natural map Pro Pro λ pB q Ñ lim λ pB pTqq B 0 B ěm m m ¯ ¯ where B pTq is the restriction of BpTq to tn ě mu Ă N. Let φ denote the diagram ěm op φ: tn P N|n ě mu Ñ IPerfpB pTqq obtained as the cokernel of f. It is now enough to prove that φ factors through PerfpB pTqq. Let n ě m be an integer and let g denote the induced map B pTq Ñ B pTq. We have an exact mn m n sequence ˚ ˚ ˚ ˚ ˚ ˚ s r L » g s r L Ñ g L Ñ φpnq B pTq B pTq B pTq m mn n m,n 0 0 n Let us denote by ψpnq the cofiber ˚ ˚ s r L Ñ L Ñ ψpnq B pTq B pTq 0 n so that φpnq » g ψpnq. This sequence is equivalent to the colimit (in IPerfpB pTqq) of a cofiber mn n op sequence of diagrams tp P N|p ě nu Ñ PerfpB pTqq Pro Pro ¯ ¯ ¯ λ pB pTqq Ñ λ pB pTqq Ñ ψpnq B pTq 0 B pTq n n n op It suffices to prove that the diagram ψpnq: tp P N|p ě nu Ñ PerfpB pTqq is (essentially) constant. Let p P N, p ě n. The perfect complex ψpnqppq fits in the exact sequence ˚ ˚ ˚ t ε L Ñ π L Ñ ψpnqppq B pTq B pTq np np n,p 0,p n,p where t : B pTq Ñ B pTq is the canonical projection and ε : B pTq Ñ B pTq is induced by np np n n,p n,p 0,p the augmentation O Ñ O . It follows that ψpnqppq is equivalent to S S n,p 0,p t L np B pTq{B pTq n,p 0,p ¯ ¯ Moreover, for any q ě p ě n, the induced map ψpnqppq Ñ ψpnqpqq is obtained (through t ) from the nq cofiber, in PerfpB pTqq n,q ˚ ˚ // ˚ // ˚ α ε L α L α L B pTq B pTq B pTq{B pTq npq np npq npq 0,p n,p n,p 0,p ˚ ˚ ε α L B pTq nq 0pq 0,p pσq // // ε L L L nq B pTq B pTq B pTq{B pTq 0,q n,q n,q 0,q 46 where α is the map B pTq Ñ B pTq. Let us denote by pσq the square on the left hand side npq n,q n,p above. Let us fix a few more notations oo // B pTq ˆ S B pTq ˆ S B pTq ˆ S 0,p 0,p 0,q n,p n,q n,q 0p r xxr ϕ ϕ np ψ nq S npq 0,p npq oo // np B pTq ˆ S B pTq ˆ S B pTq ˆ S n,p n,p n,q n,p n,q n,q r r r r r r ❏ r r r r r r ❏ ev nq r r r a a r np r r nq ❏ r r r xxr yyr yyr npq %% oo nq S S S T n,p n,p n,q np npq oo B pTq B pTq B pTq n,p n,q “ n,q The diagram pσq is then dual to the diagram ˚ ˚ ˚ ˚ ˚ oo α ε ̟ ev T α ̟ ev T 0p T np T npq np 0p npq np ˚ ˚ OO OO ˚ ˚ ˚ oo ε ̟ ev T ̟ ev T 0q T nq T nq 0q nq ˚ ˚ Moreover, the functor ̟ (for any n and p) satisfies the base change formula. This square is thus np equivalent to the image by ̟ of the square nq ˚ ˚ ˚ ˚ ˚ oo ψ b b ψ ev T b b ev T npq npq T npq T ˚ ˚ npq npq nq ˚ npq nq OO OO ˚ ˚ ˚ oo ϕ ϕ ev T ev T nq T T ˚ nq nq nq Using now the projection and base change formulae along the morphisms ϕ , b and ψ , we see nq npq npq that this last square is again equivalent to ˚ ˚ ˚ ˚ oo pa β ξ O q b pev T q pa β O q b pev T q npq np S T npq S T nq 0,p nq nq n,p nq ˚ ˚ ˚ OO OO ˚ ˚ ˚ ˚ oo pa ξ O q b pev T q pa O q b pev T q nq S T S T nq 0,q nq nq n,q nq We therefore focus on the diagram // O ξ O S nq S n,q 0,q // β O β ξ O npq S npq np S n,p 0,p ˚ ˚ ˚ By definition, the fibres of the horizontal maps are both equivalent to A r´ds and the map induced by the diagram above is an equivalence. We have proven that for any q ě p ě n the induced map ¯ ¯ ψpnqppq Ñ ψpnqpqq is an equivalence. It implies that L is a Tate module. BpTq 47 4.4 A symplectic structure (shifted by d) In this subsection, we will prove the following Theorem 4.4.1. Assume T is q-shifted symplectic. The ind-pro-stack B pTq admits a symplectic Tate structure shifted by q ´ d. Moreover, for any m P N we have an exact sequence ˚ ˚ ˚ ˚ ˚ s r L d Ñ s L d Ñ s r T d rq ´ ds m B pTq m B pTq m B pTq 0 0 Proof. Let us start with the following remark: the residue map r : A Ñ k “ A defined in n n 1 Lemma 4.1.2 defines a map O Ñ kr´ds. From Theorem 2.3.8, we have a pq ´ dq-shifted closed ˆd 2-form on B pTq. We have a morphism from Theorem 2.2.3 O d rq ´ ds Ñ L d b L d B pTq B pTq B pTq in PIPerfpB pTqq. Let m P N. We get a map ˚ ˚ O d rq ´ ds Ñ s L d b s L d B pTq m B pTq m B pTq and then ˚ ˚ s T d b s T d Ñ O d rq ´ ds m B pTq m B pTq B pTqm in IPPerfpB pTq q. We consider the composite map ˚ ˚ ˚ ˚ θ: s T d d b s T d d Ñ s T d b s T d Ñ O d rq ´ ds m B pTq{B pTq m B pTq{B pTq m B pTq m B pTq B pTq 0 0 m Using the Remark 2.3.9 and the proof of Proposition 4.3.1 we see that θ is induced by the morphisms (varying n and p) ´ ¯ ` ˘ A B // // ̟ E b E b ev pT b T q ̟ pE b Erqsq ̟ O d rqs np T T np np B pTq ˆS ˚ np ˚ ˚ np n,p ˚ ! where E “ a ξ h γ O d and the map A is induced by the symplectic form on T . The map B np np A np ˚ ˚ n is induced by the multiplication in O . This sheaf of functions is a trivial square zero extension of Sn,p augmentation ideal ξ h γ O d and B therefore vanishes. It follows that the morphism np np n A ˚ ˚ ˚ ˚ ˚ ˚ s T d b s T d d Ñ s T d b s T d Ñ O d rq ´ ds m B pTq m B pTq{B pTq m B pTq m B pTq B pTq 0 m ˚ ˚ factors through s T d b s T d d . Now using Proposition 4.3.1 we get a map of exact m B pTq m B pTq{B pTq 0 0 sequences in the category of Tate modules over B pTq ˚ ˚ ˚ ˚ // // s T d d s T d s r T d m B pTq{B pTq m B pTq m B pTq 0 0 ˚ ˚ // ˚ // ˚ s r L d rd ´ qs s L d rd ´ qs s L d d rd ´ qs m B pTq0 m B pTq m B pTq{B pTq0 where the maps on the sides are dual one to another. It therefore suffices to see that the map ˚ ˚ ˚ τ : s T d d Ñ s r L d rd ´ qs is an equivalence. We now observe that τ is a colimit m m m B pTq{B pTq m B pTq 0 0 indexed by p ě m of maps ´ ¯ ˚ ˚ ˚ g t ε L d Ñ T d d pm pp pp B pTq B pTq {B pTq 0p pp 0p 48 ˚ ˚ Let us fix p ě m and G “ a ξ O . The map F : T d d Ñ ε L d at hand is pp S p pp 0p B pTq {B pTq pp B pTq ˚ pp 0p 0p induced by the pairing ` ˘ ` ˘ ˚ ˚ T d d b ε T d » ̟ E b ev T b ̟ G b ev T B pTq {B pTq B pTq pp T pp T pp 0p pp 0p pp pp ˚ ˚ ` ˘ ˚ ˚ ̟ E b ev T b G b ev T pp T T pp pp ̟ pE b Gqrqs pp ´ ¯ ̟ O d rqs pp B pTq ˆS ˚ pp pp O d rq ´ ds B pTq pp We can now conclude using Lemma 4.1.2. References [Ber] Julia E. Bergner: A survey of 8-categories. IMA volumes in Math. and Appl., pages 69–83. 2010. [BGW1] Oliver Braunling, Michael Groechenig and Jesse Wolfson: A Generalized Contou- Carrère Symbol and its Reciprocity Laws in Higher Dimensions, 2014. [BGW2] Oliver Braunling, Michael Groechenig and Jesse Wolfson: Tate Objects in Exact Categories (with appendix by Jan Šťovíček and Jan Trlifaj), 2014. [Bha] Bhargav Bhatt: Algebraization and Tannaka duality. Camb. J. Math., 4(4) pp.403–461, [BK] Aldridge Bousfield and Daniel M. Kan: Homotopy limits, completions and localizations, volume 304 of Lecture Notes in Mathematics. Springer-Verlag, Berlin-New York, 1972. [CC] Carlos Contou-Carrère: Jacobienne locale, groupe de bivecteurs de Witt universel et symbole modéré. C. R. Acad. Sci. Paris Sér. I Math., 318(8) pp.743–746, 1994. [DL] Jan Denef and Franois Loeser: Germs of arcs on singular algebraic varieties and motivic integration. Inventiones Mathematica, 135 pp.201–232, 1999. [GR] Dennis Gaitsgory and Nick Rozenblyum: DG Indschemes. In Perspectives in represen- tation theory, volume 610 of Contemp. Math., pages 139–251. AMS, 2014. [HAlg] Jacob Lurie: Higher algebra. 2016. [Hen1] Benjamin Hennion: Formal loop spaces and tangent Lie algebras. PhD thesis, 2015. [Hen2] Benjamin Hennion: Tate objects in stable (8,1)-categories. Homology, Homotopy and Applications, 19(2) pp.373–395, 2017. 49 [Hir] Philip S. Hirschhorn: The homotopy groups of the inverse limit of a tower of fibrations. [HLP] Daniel Halpern-Leistner and Anatoly Preygel: Mapping stacks and categorical notions of properness, 2014. [Hov] Mark Hovey: Model categories, volume 63 of Mathematical Surveys and Monographs. Amer- ican Mathematical Society, Providence, RI, 1999. [HTT] Jacob Lurie: Higher topos theory, volume 170 of Annals of Mathematics Studies. Princeton University Press, Princeton, NJ, 2009. [KV1] Mikhail Kapranov and Éric Vasserot: Vertex algebras and the formal loop space. Publ. Math. Inst. Hautes Études Sci, 100 pp.209–269, 2004. [KV2] Mikhail Kapranov and Éric Vasserot: Formal loops II: A local Riemann-Roch theorem for determinantal gerbes. Ann. Sci. ENS, 40 pp.113–133, 2007. [KV3] Mikhail Kapranov and Éric Vasserot: Formal loops IV: Chiral differential operators, [Lur] Jacob Lurie: Derived algebraic geometry XII: Proper morphisms, com- pletions and the Grothendieck existence theorem. 2011, available at [http://www.math.harvard.edu/~lurie/papers/DAG-XII.pdf]. [MP] Vincente Muñoz and Francisco Presas: Geometric structures on loop and path spaces. Proc. Indian Acad. Sci., 120(4) pp.417–428, 2010. [OZ] Denis Osipov and Xiewen Zhu: The two-dimensional Contou-Carrère symbol and reci- procity laws. to appear in Journal of Alg. geometry, 2013. [PTVV] Tony Pantev, Bertrand Toën, Michel Vaquié and Gabriele Vezzosi: Shifted symplectic structures. Publication de l’IHES, 2011. [Toë1] Bertrand Toën: Simplicial Presheaves and Derived Algebraic Geometry. Advanced Courses in Math. Birkhäuser, 2010. [Toë2] Bertrand Toën: Derived algebraic geometry. EMS Surv. Math. Sci., 1(2) pp.153–240, 2014. [TV] Bertrand Toën and Gabriele Vezzosi: Homotopical Algebraic Geometry II: geometric stacks and applications. Memoirs of the AMS, 193(902) pp.257–372, 2008. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Mathematics arXiv (Cornell University)

Higher dimensional formal loop spaces

Hennion, Benjamin
Mathematics , Volume 2020 (2010) – Oct 20, 2020
Download PDF
50 pages

Loading next page...
 
/lp/arxiv-cornell-university/higher-dimensional-formal-loop-spaces-pJT8cK2oT0

References (24)

ISSN
0012-9593
eISSN
ARCH-3343
DOI
10.24033/asens.2329
Publisher site
See Article on Publisher Site

Abstract

8 1 If M is a symplectic manifold then the space of smooth loops C pS , Mq inherits of a quasi- symplectic form. We will focus in this article on an algebraic analogue of that result. In their article [KV1], Kapranov and Vasserot introduced and studied the formal loop space of a scheme X. We generalize their construction to higher dimensional loops. To any scheme X – not neces- sarily smooth – we associate L pXq, the space of loops of dimension d. We prove it has a structure of (derived) Tate scheme – ie its tangent is a Tate module: it is infinite dimensional but behaves nicely enough regarding duality. We also define the bubble space B pXq, a variation of the loop space. We prove that B pXq is endowed with a natural symplectic form as soon as X has one (in the sense of [PTVV]). Throughout this paper, we will use the tools of p8, 1q-categories and symplectic derived alge- braic geometry. Résumé 8 1 Espaces des lacets formels de dimension supérieure : L’espace des lacets C pS , Mq associé à une variété symplectique M se voit doté d’une structure (quasi-)symplectique induite par celle de M. Nous traiterons dans cet article d’un analogue algébrique de cet énoncé. Dans leur article [KV1], Kapranov et Vasserot ont introduit l’espace des lacets formels associé à un schéma. Nous généralisons leur construction à des lacets de dimension supérieure. Nous associons à tout schéma X – pas forcément lisse – l’espace L pXq de ses lacets formels de dimension d. Nous démontrerons que ce dernier admet une structure de schéma (dérivé) de Tate : son espace tangent est de Tate : de dimension infinie mais suffisamment structuré pour se soumettre à la dualité. Nous définirons également l’espace B pXq des bulles de X, une variante de l’espace des lacets, et nous montrerons que le cas échéant, il hérite de la structure symplectique de X. Contents Introduction 2 Preliminaries 5 0.1 A few tools from higher category theory . . . . . . . . . . . . . . . . . . . . . . . . . . 5 0.2 Derived algebraic geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 1 Ind-pro-stacks 12 1.1 Cotangent complex of a pro-stack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.2 Cotangent complex of an ind-pro-stack . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 1.3 Uniqueness of pro-structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 1.4 Uniqueness of ind-pro-structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 arXiv:2010.10318v1 [math.AG] 20 Oct 2020 2 Symplectic Tate stacks 27 2.1 Tate stacks: definition and first properties . . . . . . . . . . . . . . . . . . . . . . . . . 27 2.2 Shifted symplectic Tate stacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 2.3 Mapping stacks admit closed forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 2.4 Mapping stacks have a Tate structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 3 Formal loops 34 3.1 Dehydrated algebras and de Rham stacks . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.2 Higher dimensional formal loop spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.3 Tate structure and determinantal anomaly . . . . . . . . . . . . . . . . . . . . . . . . . 39 4 Bubble spaces 42 4.1 Two lemmas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 4.2 Definition and properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 4.3 Its tangent is a Tate module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 4.4 A symplectic structure (shifted by d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 References 49 Introduction Considering a differential manifold M, one can build the space of smooth loops LpMq in M. It is a central object of string theory. Moreover, if M is symplectic then so is LpMq – more precisely quasi-symplectic since it is not of finite dimension – see for instance [MP]. We will be interested here in an algebraic analogue of that result. The first question is then the following: what is an algebraic analogue of the space of smooth loops? An answer appeared in 1994 in Carlos Contou-Carrère’s work (see [CC]). He studies there G pCpptqqq, some sort of holomorphic functions in the multiplicative group scheme, and defines the famous Contou- Carrère symbol. This is the first occurrence of a formal loop space known by the author. This idea was then generalised to algebraic groups as the affine Grassmannian Gr “ GpCpptqqq showed G GpCrrtssq up and got involved in the geometric Langlands program. In their paper [KV1], Mikhail Kapranov and Éric Vasserot introduced and studied the formal loop space of a smooth scheme X. It is an ind-scheme LpXq which we can think of as the space of maps SpecCpptqq Ñ X. This construction strongly inspired the one presented in this article. There are at least two ways to build higher dimensional formal loops. The most studied one consists in using higher dimensional local fields kppt qq . . . ppt qq and is linked to Beilinson’s adèles. There 1 d is also a generalisation of Contou-Carrère symbol in higher dimensions using those higher dimensional local fields – see [OZ] and [BGW1]. If we had adopted this angle, we would have considered maps from some torus Specpkppt qq . . . ppt qqq to X. 1 d The approach we will follow in this work is different. We generalize here the definition of Kapranov and Vasserot to higher dimensional loops in the following way. For X a scheme of finite presentation, not necessarily smooth, we define L pXq, the space of formal loops of dimension d in X. We define d d L pXq the space of maps from the formal neighbourhood of 0 in A to X. This is a higher dimensional version of the space of germs of arcs as studied by Jan Denef and François Loeser in [DL]. Let also d d L pXq denote the space of maps from a punctured formal neighbourhood of 0 in A to X. The formal d d d loop space L pXq is the formal completion of L pXq in L pXq. Understanding those three items is V U the main goal of this work. The problem is mainly to give a meaningful definition of the punctured The variable t , . . . , t are actually ordered. The author likes to think of Specpkppt qq . . . ppt qqq as a formal torus 1 d 1 d equipped with a flag representing this order. 2 formal neighbourhood of dimension d. We can describe what its cohomology should be: ’krrX , . . . , X ss if n “ 0 1 d n d ´1 ´1 ´1 H pA r t0uq “ pX . . . X q krX , . . ., X s if n “ d ´ 1 1 d 0 otherwise but defining this punctured formal neighbourhood with all its structure is actually not an easy task. Nevertheless, we can describe what maps out of it are, hence the definition of L pXq and the formal loop space. This geometric object is of infinite dimension, and part of this study is aimed at identifying some structure. Here comes the first result in that direction. Theorem 1 (see Proposition 3.3.4). The formal loop space of dimension d in a scheme X is rep- resented by a derived ind-pro-scheme. Moreover, the functor X ÞÑ L pXq satisfies the étale descent condition. We use here methods from derived algebraic geometry as developed by Bertrand Toën and Gabriele Vezzosi in [TV]. The author would like to emphasize here that the derived structure is d d d necessary since, when X is a scheme, the underlying schemes of L pXq, L pXq and L pXq are iso- U V morphic as soon as d ě 2. Let us also note that derived algebraic geometry allowed us to define L pXq for more general X’s, namely any derived stack. In this case, the formal loop space L pXq is no longer a derived ind-pro-scheme but an ind-pro-stack. The case d “ 1 and X is a smooth scheme gives a derived enhancement of Kapranov and Vasserot’s definition. This derived enhancement is conjectured to be trivial when X is a smooth affine scheme in [GR, 9.2.10]. Gaitsgory and Rozenblyum also prove in loc. cit. their conjecture holds when X is an algebraic group. The proof of Theorem 1 is based on an important lemma. We identify a full sub-category C of the category of ind-pro-stacks such that the realisation functor C Ñ dSt is fully faithful. We then prove that whenever X is a derived affine scheme, the stack L pXq is in the essential image of C and is thus endowed with an essentially unique ind-pro-structure satisfying some properties. The generalisation to any X is made using a descent argument. Note that for general X’s, the ind-pro-structure is not known to satisfy nice properties one could want to have, for instance on the transition maps of the diagrams. We then focus on the following problem: can we build a symplectic form on L pXq when X is symplectic? Again, this question requires the tools of derived algebraic geometry and shifted symplectic structures as in [PTVV]. A key feature of derived algebraic geometry is the cotangent complex L of any geometric object X. A (n-shifted) symplectic structure on X is a closed 2-form O r´ns Ñ L ^ L which is non degenerate – ie induces an equivalence X X T Ñ L rns X X Because L pXq is not finite, linking its cotangent to its dual – through an alleged symplectic form – requires to identify once more some structure. We already know that it is an ind-pro-scheme but the proper context seems to be what we call Tate stacks. Before saying what a Tate stack is, let us talk about Tate modules. They define a convenient context for infinite dimensional vector spaces. They where studied by Lefschetz, Beilinson and Drin- feld, among others, and more recently by Bräunling, Gröchenig and Wolfson [BGW2]. We will use here the notion of Tate objects in the context of stable p8, 1q-categories as developed in [Hen2]. If C is a stable p8, 1q-category – playing the role of the category of finite dimensional vector spaces, the category TatepCq is the full subcategory of the p8, 1q-category of pro-ind-objects Pro IndpCq in C containing both IndpCq and PropCq and stable by extensions and retracts. We will define the derived category of Tate modules on a scheme – and more generally on a derived ind-pro-stack. An Artin ind-pro-stack X – meaning an ind-pro-object in derived Artin stacks – is then gifted with a cotangent complex L . This cotangent complex inherits a natural structure of pro-ind-module on X. This allows us to define a Tate stack as an Artin ind-pro-stack whose cotangent 3 d complex is a Tate module. The formal loop space L pXq is then a Tate stack as soon as X is a finitely presented derived affine scheme. For a more general X, what precedes makes L pXq some kind of locally Tate stack. This structure suffices to define a determinantal anomaly ´ ¯ “ ‰ 2 d ˆ Det d P H L pXq,O L pXq d L pXq for any quasi-compact quasi-separated (derived) scheme X – this construction also works for slightly more general X’s, namely Deligne-Mumford stacks with algebraisable diagonal, see Definition 2.1.3. Kapranov and Vasserot proved in [KV3] that in dimension 1, the determinantal anomaly governs the existence of sheaves of chiral differential operators on X. One could expect to have a similar result in higher dimensions, with higher dimensional analogues of chiral operators and vertex algebras. The author plans on studying this in a future work. Another feature of Tate modules is duality. It makes perfect sense and behaves properly. Using the theory of symplectic derived stacks developed by Pantev, Toën, Vaquié and Vezzosi in [PTVV], we are then able to build a notion of symplectic Tate stack: a Tate stack Z equipped with a (n-shifted) closed 2-form which induces an equivalence T Ñ L rns Z Z of Tate modules over Z between the tangent and (shifted) cotangent complexes of Z. To make a step toward proving that L pXq is a symplectic Tate stack, we actually study the bubble space B pXq – see Definition 4.2.3. When X is affine, we get an equivalence d d d B pXq » L pXq ˆ L pXq V V L pXq Note that the fibre product above is a derived intersection. We then prove the following result Theorem 2 (see Theorem 4.4.1). If X is an n-shifted symplectic stack then the bubble space B pXq is endowed with a structure of pn ´ dq-shifted symplectic Tate stack. The proof of this result is based on a classical method. The bubble space is in fact, as an ind- pro-stack, the mapping stack from what we call the formal sphere S of dimension d to X. There are therefore two maps pr ev d d oo ˆ // B pXq B pXq ˆ S X d ˚ The symplectic form on B pXq is then ev ω , where ω is the symplectic form on X. The key X X ˆd argument is the construction of this integration on the formal sphere, ie on an oriented pro-ind-stack of dimension d. The orientation is given by a residue map. On the level of cohomology, it is the morphism d d ´1 ´1 ´1 H pS q » pX . . .X q krX , . . . , X s Ñ k 1 d 1 d ´1 mapping pX . . . X q to 1. 1 d This integration method would not work on L pXq, since the punctured formal neighbourhood does not have as much structure as the formal sphere: it is not known to be a pro-ind-scheme. Nevertheless, Theorem 2 is a first step toward proving that L pXq is symplectic. We can consider the d d nerve Z of the map L pXq Ñ L pXq. It is a groupoid object in ind-pro-stacks whose space of maps V U is B pXq. The author expects that this groupoid is compatible in some sense with the symplectic structure so that L pXq would inherit a symplectic form from realising this groupoid. One the other hand, if L pXq was proven to be symplectic, then the fibre product defining B pXq should be a Lagrangian intersection. The bubble space would then inherit a symplectic structure from that on L pXq. 4 Techniques and conventions Throughout this work, we will use the techniques of p8, 1q-category theory. We will once in a while use explicitly the model of quasi-categories developed by Joyal and Lurie (see [HTT]). That being said, the results should be true with any equivalent model. Let us fix now two universes U P V to deal with size issues. Every algebra, module or so will implicitly be U-small. The first part will consist of reminders about p8, 1q-categories. We will fix there some notations. Note that we will often refer to [Hen1] for some p8, 1q-categorical results needed in this article. We will also use derived algebraic geometry, as introduced in [TV]. We refer to [Toë2] for a recent survey of this theory. We will denote by k a base field and by dSt the p8, 1q-category of (U-small) derived stacks over k. In the first section, we will dedicate a few page to introduce derived algebraic geometry. Outline This article begins with a few paragraphs, recalling some notions we will use. Among them are p8, 1q- categories and derived algebraic geometry. In section 1, we set up a theory of geometric ind-pro-stacks. We then define in section 2 symplectic Tate stacks and give a few properties, including the construction of the determinantal anomaly (see Definition 2.1.3). Comes section 3 where we finally define higher dimensional loop spaces and prove Theorem 1 (see Proposition 3.3.4). We finally introduce the bubble space and prove Theorem 2 (see Theorem 4.4.1) in section 4. Aknowledgements I would like to thank Bertrand Toën, Damien Calaque and Marco Robalo for the many discussions we had about the content of this work. I am grateful to Mikhail Kapranov, James Wallbridge and Giovanni Faonte for inviting me at the IPMU. My stay there was very fruitful and the discussions we had were very interesting. I learned after writing down this article that Kapranov had an unpublished document in which higher dimensional formal loops are studied. I am very grateful to Kapranov for letting me read those notes, both inspired and inspiring. This work is extracted from my PhD thesis [Hen1] under the advisement of Bertrand Toën. I am very grateful to him for those amazing few years. Preliminaries In this part, we recall some results and definitions from p8, 1q-category theory and derived algebraic geometry. 0.1 A few tools from higher category theory In the last decades, theory of p8, 1q-categories has tremendously grown. The core idea is to consider categories enriched over spaces, so that every object or morphism is considered up to higher homotopy. The typical example of such a category is the category of topological spaces itself: for any topological spaces X and Y , the set of maps X Ñ Y inherits a topology. It is often useful to talk about topological spaces up to homotopy equivalences. Doing so, one must also consider maps up to homotopy. To do so, one can of course formally invert every homotopy equivalence and get a set of morphisms rX, Y s. This process loses information and mathematicians tried to keep trace of the space of morphisms. The first fully equipped theory handy enough to work with such examples, called model categories, was introduced by Quillen. A model category is a category with three collections of maps – weak equivalences (typically homotopy equivalences), fibrations and cofibrations – satisfying a bunch of conditions. The datum of such collections allows us to compute limits and colimits up to homotopy. We refer to [Hov] for a comprehensive review of the subject. 5 Using model categories, several mathematicians developed theories of p8, 1q-categories. Let us name here Joyal’s quasi-categories, complete Segal spaces or simplicial categories. Each one of those theories is actually a model category and they are all equivalent one to another – see [Ber] for a review. In [HTT], Lurie developed the theory of quasi-categories. In this book, he builds everything necessary so that we can think of p8, 1q-categories as we do usual categories. To prove something in this context still requires extra care though. We will use throughout this work the language as developed by Lurie, but we will try to keep in mind the 1-categorical intuition. In this section, we will fix a few notations and recall some results to which we will often refer. Notations: Let us first fix a few notations, borrowed from [HTT]. • We will denote by Cat the p8, 1q-category of U-small p8, 1q-categories – see [HTT, 3.0.0.1]; L,U • Let Pr denote the p8, 1q-category of U-presentable (and thus V-small) p8, 1q-categories with left adjoint functors – see [HTT, 5.5.3.1]; • The symbol sSets will denote the p8, 1q-category of U-small simplicial sets up to homotopy equivalences (this is equivalent to the category of (nice) topological spaces up to homotopy); • For any p8, 1q-categories C and D we will write FctpC,Dq for the p8, 1q-category of functors from op C to D (see [HTT, 1.2.7.3]). The category of presheaves will be denoted PpCq “ FctpC , sSetsq. • For any p8, 1q-category C and any objects c and d in C, we will denote by Map pc, dq the space of maps from c to d. • For any simplicial set K, we will denote by K the simplicial set obtained from K by formally adding a final object. This final object will be called the cone point of K . The following theorem is a concatenation of results from Lurie. Theorem 0.1.1 (Lurie). Let C be a V-small p8, 1q-category. There is an p8, 1q-category Ind pCq and a functor j : C Ñ Ind pCq such that (i) The p8, 1q-category Ind pCq is V-small; (ii) The p8, 1q-category Ind pCq admits U-small filtered colimits and is generated by U-small filtered colimits of objects in jpCq; (iii) The functor j is fully faithful and preserves finite limits and finite colimits which exist in C; (iv) For any c P C, its image jpcq is U-small compact in Ind pCq; (v) For every p8, 1q-category D with every U-small filtered colimits, the functor j induces an equiv- alence U´c U Fct pInd pCq,Dq Ñ FctpC,Dq U´c U U where Fct pInd pCq,Dq denote the full subcategory of FctpInd pCq,Dq spanned by functors preserving U-small filtered colimits. (vi) If C is U-small and admits all finite colimits then Ind pCq is U-presentable; (vii) If C is endowed with a symmetric monoidal structure then there exists such a structure on Ind pCq such that the monoidal product preserves U-small filtered colimits in each variable. Proof. Let us use the notations of [HTT, 5.3.6.2]. Let K denote the collection of U-small filtered U K simplicial sets. We then set Ind pCq “ P pCq. It satisfies the required properties because of loc. cit. 5.3.6.2 and 5.5.1.1. We also need tiny modifications of the proofs of loc. cit. 5.3.5.14 and 5.3.5.5. The last item is proved in [HAlg, 6.3.1.10]. 6 U Remark 0.1.2. Note that when C admits finite colimits then the category Ind pCq embeds in the V-presentable category Ind pCq. Definition 0.1.3. Let C be a V-small 8-category. We define Pro pCq as the p8, 1q-category ´ ¯ op U U op Pro pCq “ Ind pC q It satisfies properties dual to those of Ind pCq. Definition 0.1.4. Let C be a V-small p8, 1q-category. Let V U V i: FctpC, Cat q Ñ FctpInd pCq, Cat q 8 8 denote the left Kan extension functor. We will denote by Ind the composite functor Ind ˝´ V U V U V // // FctpC, Cat q FctpInd pCq, Cat q FctpInd pCq, Cat q 8 8 8 We will denote by Pro the composite functor Pro ˝´ V V U V // // FctpC, Cat q FctpC, Cat q FctpPro pCq, Cat q 8 8 8 We define the same way V V V Ind : FctpC, Cat q Ñ FctpInd pCq, Cat q C 8 V V V Pro : FctpC, Cat q Ñ FctpPro pCq, Cat q C 8 Remark 0.1.5. The Definition 0.1.4 can be expanded as follows. To any functor f : C Ñ Cat and any ind-object c colimit of a diagram c¯ // // K C Ind pCq we construct an p8, 1q-category U U Ind pfqpcq » Ind pcolimfpc¯qq To any pro-object d limit of a diagram op // // K C Pro pCq we associate an p8, 1q-category U U Pro pfqpdq » lim Pro pfpdqq V,st V Definition 0.1.6. Let Cat denote the subcategory of Cat spanned by stable categories with 8 8 V,st,id V,st exact functors between them – see [HAlg, 1.1.4]. Let Cat denote the full subcategory of Cat 8 8 spanned by idempotent complete stable categories. Remark 0.1.7. It follows from [HAlg, 1.1.4.6, 1.1.3.6, 1.1.1.13 and 1.1.4.4] that the functors Ind and Pro restricts to functors U U V,st V,st Ind : FctpC, Cat q Ñ FctpInd pCq, Cat q C 8 8 U V,st U V,st Pro : FctpC, Cat q Ñ FctpPro pCq, Cat q C 8 8 7 Symmetric monoidal p8, 1q-categories: We will make use in the last chapter of the theory of symmetric monoidal p8, 1q-categories as developed in [HAlg]. Let us give a (very) quick review of those objects. Definition 0.1.8. Let Fin denote the category of pointed finite sets. For any n P N, we will denote by xny the set t˚, 1, . . ., nu pointed at ˚. For any n and i ď n, the Segal map δ : xny Ñ x1y is defined i i by δ pjq “ 1 if j “ i and δ pjq “ ˚ otherwise. Definition 0.1.9. (see [HAlg, 2.0.0.7]) Let C be an p8, 1q-category. A symmetric monoidal structure on C is the datum of a coCartesian fibration p: C Ñ Fin such that • The fibre category C is equivalent to C and x1y b b n n • For any n, the Segal maps induce an equivalence C Ñ pC q » C . xny x1y b b,V where C denote the fibre of p at xny. We will denote by Cat the p8, 1q-category of V-small xny symmetric monoidal p8, 1q-categories – see [HAlg, 2.1.4.13]. ˚ V Such a coCartesian fibration is classified by a functor φ: Fin Ñ Cat – see [HTT, 3.3.2.2] – such that φpxnyq » C . The tensor product on C is induced by the map of pointed finite sets µ : x2y Ñ x1y mapping both 1 and 2 to 1 b “ φpµ q: C Ñ C b,V V Remark 0.1.10. The forgetful functor Cat Ñ Cat preserves all limits as well as filtered colimits 8 8 – see [HAlg, 3.2.2.4 and 3.2.3.2]. Moreover, it follows from Theorem 0.1.1 - (vii) that the functor Ind induces a functor U b,V b,V Ind : Cat Ñ Cat 8 8 U U U The same holds for Pro . The constructions Ind and Pro therefore restrict to U b,V U b,V Ind : FctpC, Cat q Ñ FctpInd pCq, Cat q C 8 8 U b,V U b,V Pro : FctpC, Cat q Ñ FctpPro pCq, Cat q C 8 8 Tate objects: We now recall the definition and a few properties of Tate objects in a stable and idempotent complete p8, 1q-category. The content of this paragraph comes from [Hen2]. See also [Hen1]. Definition 0.1.11. Let C be a stable and idempotent complete p8, 1q-category. Let Tate pCq denote U U U U the smallest full subcategory of Pro Ind pCq containing Ind pCq and Pro pCq, and both stable and idempotent complete. U U U The category Tate pCq naturally embeds into Ind Pro pCq as well. op Proposition 0.1.12. If moreover C is endowed with a duality equivalence C Ñ C then the induced functor ´ ¯ op U U U U U U Pro Ind pCq Ñ Pro Ind pCq » Ind Pro pCq U U op preserves Tate objects and induces an equivalence Tate pCq » Tate pCq . Definition 0.1.13. Let C be a V-small p8, 1q-category. We define the functor Tate ˝´ V,st U V,st U V,st,id // // Tate : FctpC, Cat q FctpInd pCq, Cat q FctpInd pCq, Cat q 8 8 8 8 0.2 Derived algebraic geometry We present here some background results about derived algebraic geometry. Let us assume k is a field of characteristic 0. First introduced by Toën and Vezzosi in [TV], derived algebraic geometry is a generalisation of algebraic geometry in which we replace commutative algebras over k by commutative differential graded algebras (or cdga’s). We refer to [Toë2] for a recent survey of this theory. ď0 Generalities on derived stacks: We will denote by cdga the p8, 1q-category of cdga’s over k concentrated in non-positive cohomological degree. It is the p8, 1q-localisation of a model category ď0 along weak equivalences. Let us denote dAff the opposite p8, 1q-category of cdga . It is the category of derived affine schemes over k. In this work, we will adopt a cohomological convention for cdga’s. op ď0 A derived prestack is a presheaf dAff » cdga Ñ sSets. We will thus write PpdAff q for k k the p8, 1q-category of derived prestacks. A derived stack is a prestack satisfying the étale descent condition. We will denote by dSt the p8, 1q-category of derived stacks. It comes with an adjunction p´q : PpdAff q Õ dSt k k where the left adjoint p´q is called the stackification functor. Remark 0.2.1. The categories of varieties, schemes or (non derived) stacks embed into dSt . Definition 0.2.2. The p8, 1q-category of derived stacks admits an internal hom MappX, Y q between ď0 two stacks X and Y . It is the functor cdga Ñ sSets defined by A ÞÑ Map pX ˆ Spec A, Y q dSt We will call it the mapping stack from X to Y . There is a derived version of Artin stacks of which we first give a recursive definition. Definition 0.2.3. (see for instance [Toë1, 5.2.2]) Let X be a derived stack. • We say that X is a derived 0-Artin stack if it is a derived affine scheme ; • We say that X is a derived n-Artin stack if there is a family pT q of derived affine schemes and a smooth atlas u: T Ñ X such that the nerve of u has values in derived pn ´ 1q-Artin stacks ; • We say that X is a derived Artin (or algebraic) stack if it is an n-Artin stack for some n. • We say that X is locally of finite presentation if there exists a smooth atlas T Ñ X as above, such that the derived affine schemes T are all finitely presented (ie their cdga of functions is finitely presented, or equivalently compact is the category of cdga’s). We also say that X is finitely presented if there is such an atlas with a finite number of T ’s. Art We will denote by dSt the full subcategory of dSt spanned by derived Artin stacks. Definition 0.2.4. A morphism X Ñ Y of derived stacks is called algebraic if for any cdga A and any map Spec A Ñ Y , the derived intersection X ˆ Spec A is an algebraic stack. To any cdga A we associate the category dgMod of dg-modules over A. Similarly, to any derived stack X we can associate a derived category QcohpXq of quasicoherent sheaves. It is a U-presentable p8, 1q-category given by the formula QcohpXq » lim dgMod SpecAÑX Moreover, for any map f : X Ñ Y , there is a natural pull back functor f : QcohpY q Ñ QcohpXq. This functor admits a right adjoint, which we will denote by f . This construction is actually a functor of p8, 1q-categories. 9 Definition 0.2.5. Let us denote by Qcoh the functor op L,U Qcoh: dSt Ñ Pr For any X we can identify a full subcategory PerfpXq Ă QcohpXq of perfect complexes. This defines a functor op Perf : dSt Ñ Cat Remark 0.2.6. For any derived stack X the categories QcohpXq and PerfpXq are actually stable and idempotent complete p8, 1q-categories. The inclusion PerfpXq Ñ QcohpXq is exact. Moreover, for any map f : X Ñ Y the pull back functor f preserves perfect modules and is also exact. Definition 0.2.7 (see for instance the appendix of [HLP]). Let X be a derived stack and let π: X Ñ ˚ denote the projection. We say that X is of finite cohomological dimension if there is a non-negative integer d such that the complex π O “ RΓpX,O q P dgMod is concentrated in degree lower or ˚ X X equal to d. Example 0.2.8. Any derived affine scheme is of finite cohomological dimension (take d “ 0). Any quasi-compact quasi-separated derived stack (ie a finite colimit of derived affine schemes) is of finite cohomological dimension. Any derived Artin stack X over a basis S admits a cotangent complex L P QcohpXq. If X is X{S locally of finite presentation, then the its cotangent complex is perfect L P PerfpXq X{S Symplectic structures: Following [PTVV], to any derived stack X we associate two complexes p p,cl A pXq and A pXq in dgMod , respectively of p-forms and closed p-forms on X. They come with p,cl p 2 a natural morphism A pXq Ñ A pXq forgetting the lock closing the forms . This actually glues into a natural transformation p,cl '' dSt dgMod Let us emphasize that, as soon as X is Artin, the complex A pXq is canonically equivalent to the global sections complex of L ^ L . In particular, any n-shifted 2-forms kr´ns Ñ A pXq induces a X X morphism O r´ns Ñ L ^ L in QcohpXq. If X is locally of finite presentation, the cotangent L X X X X is perfect and we then get a map T r´ns Ñ L X X Definition 0.2.9. Let X be a derived Artin stack locally of finite presentation. • An n-shifted 2-form ω : kr´ns Ñ A pXq is called non-degenerated if the induced morphism T r´ns Ñ L is an equivalence; X X 2,cl • An n-shifted symplectic form on X is a n-shifted closed 2-form ω : kr´ns Ñ A pXq such 2,cl 2 that underlying 2-form kr´ns Ñ A pXq Ñ A pXq is non degenerate. ď0 ď´1 Obstruction theory: Let A P cdga and let M P dgMod be an A-module concentrated in k A negative cohomological degrees. Let d be a derivation A Ñ A ‘ M and s: A Ñ A ‘ M be the trivial derivation. The square zero extension of A by Mr´1s twisted by d is the fibre product // A ‘ Mr´1s A // A A ‘ M This lock is a structure on the form: being closed in not a property in this context. 10 ď´1 Let now X be a derived stack and M P QcohpXq . We will denote by XrMs the trivial square zero extension of X by M. Let also d: XrMs Ñ X be a derivation – ie a retract of the natural map X Ñ XrMs. We define the square zero extension of X by Mr´1s twisted by d as the colimit X rMr´1ss “ colim SpecpA ‘ ˚ f Mr´1sq d f d f : SpecAÑX It is endowed with a natural morphism X Ñ X rMr´1ss induced by the projections p as above. Proposition 0.2.10 (Obstruction theory on stacks). Let F Ñ G be an algebraic morphism of derived ď´1 stacks. Let X be a derived stack and let M P QcohpXq . Let d be a derivation d P Map pXrMs, Xq X{´ We consider the map of simplicial sets ψ: MappX rMr´1ss, Fq Ñ MappX, Fq ˆ MappX rMr´1ss, Gq d d MappX,Gq Let y P MappX, Fq ˆ MappX rMr´1ss, Gq and let x P MappX, Fq be the induced map. There MappX,Gq d exists a point αpyq P Mappx L , Mq such that the fibre ψ of ψ at y is equivalent to the space of F{G y paths from 0 to αpyq in Mappx L , Mq F{G ψ » Ω Mappx L , Mq y 0,αpyq F{G Proof. This is a simple generalisation of [TV, 1.4.2.6]. The proof is very similar. We have a natural commutative square // XrMs X // X X rMr´1ss It induces a map α: MappX, Fq ˆ MappX rMr´1ss, Gq Ñ Map pXrMs, Fq » Mappx L , Mq d F{G X{´{G MappX,Gq Let Ω Map pXrMs, Fq denote the space of paths from 0 to αpyq. It is the fibre product 0,αpyq X{´{G // Ω Map pXrMs, Fq ˚ 0,αpyq X{´{G αpyq // ˚ Map pXrMs, Fq X{´{G The composite map αψ is by definition homotopic to the 0 map. This defines a morphism f : Ω Map pXrMs, Fq Ñ ψ 0,αpyq y X{´{G It now suffices to see that the category of X’s for which f is an equivalence contains every derived affine scheme and is stable by colimits. The first assertion is exactly [TV, 1.4.2.6] and the second one is trivial. 11 Postnikov towers: To any cdga A, one can associate its n-truncation A for some n. It is, by ďn definition, the universal cdga with vanishing cohomology H pA q for p ă ´n associated to A. The ďn truncation comes with a canonical map A Ñ A so that one can form the diagram ďn A Ð A Ð . . . ď0 ď1 This induces a canonical morphism A Ñ lim A which is an equivalence. n ďn r´n,0s This phenomenon has a counterpart when dealing with derived stacks. We denote by cdga the category of cdga’s with cohomology concentrated in degrees ´n to 0. It comes with the fully r´n,0s ď0 faithful embedding i : cdga Ñ cdga . k k ď0 For any prestack X : cdga Ñ sSets, we define its truncation τ X as the restriction of X to ďn r´n,0s the category cdga . We will abuse notations and also denote by τ X the functor obtained after ďn left Kan extending along i . The prestack τ X comes with a canonical morphism τ X Ñ X which, n ďn ďn as n varies, assembles to define a canonical map colimτ X Ñ X ďn Remark that this morphism is not necessarily an equivalence. We will study it in Lemma 1.3.1. Algebraisable stacks: Let X be a derived stack and A be a cdga. Let a “ pa , . . . , a q be a 1 p 0 0 n n sequence of elements of A forming a regular sequence in H pAq. Let A denote the Kozsul a , . . . , a 1 p n n complex associated with the regular sequence pa , . . ., a q. It is endowed with a cdga structure. There 1 p is a canonical map ´ ¯ ´ ¯ n n n n ψpAq : colim X A Ñ X lim A { { a a , . . ., a a , . . . , a 1 p 1 p n n This map is usually not an equivalence. Definition 0.2.11. A derived stack X is called algebraisable if for any A and any regular sequence a the map ψpAq is an equivalence. A map f : X Ñ Y is called algebraisable if for any derived affine scheme T and any map T Ñ Y , the fibre product X ˆ T is algebraisable. We will say that a derived stack X has algebraisable diagonal if the diagonal morphism X Ñ XˆX is algebraisable. Remark 0.2.12. A derived stack X has algebraisable diagonal if for any A and a the map ψpAq is fully faithful. One could also rephrase the definition of being algebraisable as follows. A stack is algebraisable if it does not detect the difference between ´ ¯ ´ ¯ n n n n colim Spec A and Spec lim A { { a , . . ., a a , . . ., a 1 p 1 p n n Example 0.2.13. Any derived affine scheme is algebraisable. Another important example of alge- braisable stack is the stack of perfect complexes. In [Bha], Bhargav Bhatt gives some more examples of algebraisable (non-derived) stacks – although our definition slightly differs from his. He proves that any quasi-compact quasi-separated algebraic space is algebraisable and also provides with examples of non-algebraisable stacks. Let us name KpG , 2q – the Eilenberg-Maclane classifying stack of G – m m as an example of non-algebraisable stack. Algebraisability of Deligne-Mumford stacks is also look at in [Lur]. 1 Ind-pro-stacks Throughout this section, we will denote by S a derived stack over some base field k and by dSt the category of derived stack over the base S. 12 1.1 Cotangent complex of a pro-stack Definition 1.1.1. A pro-stack over S an object of Pro dSt . Remark 1.1.2. Note that the category Pro dSt is equivalent to the category of pro-stacks over k with a morphism to S. op U Definition 1.1.3. Let Perf : dSt Ñ Cat denote the functor mapping a stack to its category of perfect complexes. We will denote by IPerf the functor U U op L op IPerf “ Ind pPerfq: pPro dSt q Ñ Pr dSt 8 where Ind was defined in Definition 0.1.4. Whenever X is a pro-stack, we will call IPerfpXq the derived category of ind-complexes on X. It is U-presentable. If f : X Ñ Y is a map of pro-stacks, then the functor IPerfpfq: IPerfpY q Ñ IPerfpXq admits a right adjoint (as both the involved categories are presentable and the functor preserves all ˚ I colimits). We will denote f “ IPerfpfq and f its right adjoint. op Remark 1.1.4. Let X be a pro-stack and let X : K Ñ dSt denote a U-small cofiltered diagram of whom X is a limit in Pro dSt . The derived category of ind-perfect complexes on X is by definition the category IPerfpXq “ Ind pcolim PerfpXqq It thus follows from [HAlg, 1.1.4.6 and 1.1.3.6] that IPerfpXq is stable. Note that it is also equivalent to the colimit L,V IPerfpXq “ colim IPerfpXq P Pr It is therefore equivalent to the limit of the diagram op L,V R,V op IPerf pXq: K Ñ dSt Ñ Pr » pPr q 8 8 An object E in IPerfpXq is therefore the datum of an object E of IPerfpXpkqq for each k P K and of some compatibilities between them. We will then have E » p E where p : X Ñ Xpkq is the k k k natural projection. Definition 1.1.5. Let X be a pro-stack. We define its derived category of pro-perfect complexes op PPerfpXq “ pIPerfpXqq Recall that perfect complexes are precisely the dualizable objects in the category of quasi-coherent op complexes. They therefore come with a duality equivalence Perfp´q Ñ pPerfp´qq . This gives rise to the equivalence PPerfpXq » Pro pcolimPerfpXqq op U whenever X : K Ñ dSt is a cofiltered diagram of whom X is a limit in Pro dSt . S S U U op V,st,id Definition 1.1.6. Let us define the functor Tate : pPro dSt q Ñ Cat P 8 U U Tate “ Tate oppPerfq P dSt U op Remark 1.1.7. The functor Tate maps a pro-stack X given by a diagram X : K Ñ dSt to the stable p8, 1q-category U U Tate pXq “ Tate pcolim PerfpXqq There is a canonical fully faithful natural transformation U U Tate Ñ Pro ˝ IPerf We also get a fully faithful U U Tate Ñ Ind ˝ PPerf 13 op V Definition 1.1.8. Let Qcoh: dSt Ñ Cat denote the functor mapping a derived stack to its S 8 derived category of quasi-coherent sheaves. It maps a morphism between stacks to the appropriate pullback functor. We will denote by IQcoh the functor U U V op op IQcoh “ Ind pQcohq: pPro dSt q Ñ Cat dSt 8 If f : X Ñ Y is a map of pro-stacks, we will denote by f the functor IQcohpfq. We also define ď0 U ď0 op IQcoh “ Ind pQcoh q dSt the functor of connective modules. Remark 1.1.9. There is a fully faithful natural transformation IPerf Ñ IQcoh ; for any map f : X Ñ Y of pro-stacks, there is therefore a commutative diagram // IPerfpY q IQcohpY q ˚ ˚ f f I I // IPerfpXq IQcohpXq The two functors denoted by f are thus compatible. Let us also say that the functor f : IQcohpY q Ñ IQcohpXq does not need to have a right adjoint. We next show that it sometimes has one. Proposition 1.1.10. Let f : X Ñ Y be a map of pro-stacks. If Y is actually a stack then the functor f : IQcohpY q Ñ IQcohpXq admits a right adjoint. op Proof (sketch of). For a complete proof, we refer to [Hen1, 1.2.0.8]. Let us denote by X : K Ñ dSt a cofiltered diagram of whom X is a limit in Pro dSt . The map X Ñ Y factors through the S S projection X Ñ Xpkq for some k P K. The right adjoint of f is then (informally) given by the limit lim fplq kÑl ¯ ¯ ¯ where fplq is the right adjoint to the induced functor fplq : IQcohpY q Ñ IQcohpXplqq. IQ Definition 1.1.11. Let f : X Ñ Y be a map of pro-stacks. We will denote by f the right adjoint to f : IQcohpY q Ñ IQcohpXq if it exists. Remark 1.1.12. In the situation of Proposition 1.1.10, there is a natural transformation // IPerfpXq IQcohpXq 4< ♣♣ I IQ ♣♣ f ♣ f ˚ ˚ ♣♣ ♣♣ // IPerfpY q IQcohpY q It does not need to be an equivalence. Definition 1.1.13. Let X be a pro-stack over S. The structural sheaf O of X is the pull-back of O along the structural map X Ñ S. op Example 1.1.14. Let X be a pro-stack over S and X : K Ñ dSt be a U-small cofiltered diagram of whom X is a limit in Pro dSt . Let k be a vertex of K, let X denote Xpkq and let p denote S k k the induced map of pro-stacks X Ñ X . If f : k Ñ l is an arrow in K, we will also denote by f the map of stacks Xpfq. We have IQ pp q pO q » colim f O k ˚ X ˚ X f : kÑl One can see this using [Hen1, 1.2.0.7]. 14 1 Definition 1.1.15. For any category C with finite colimits, we will denote by B the functor C Ñ Cat mapping a morphism φ: c Ñ d to the category of factorizations c Ñ e Ñ d of φ. For a formal definition in the context of p8, 1q-categories, we refer to [Hen1, 1.3.0.14]. Definition 1.1.16. Let T be a stack over S. Let us consider the functor ´ ¯ op ď0 > T op QcohpTq Ñ B pid q » dSt T T dSt mapping a quasi-coherent sheaf E to the square zero extension T Ñ TrEs Ñ T . This construction is functorial in T and actually comes from a natural transformation ď0 > Ex: Qcoh Ñ B oppid q dSt op V Pro of functors dSt Ñ Cat . We will denote by Ex the natural transformation Pro U ď0 U > > Ex “ Ind oppExq: IQcoh Ñ Ind oppB oppid qq » B pid q U op ´ ´ dSt dSt dSt pPro dSt q S S S op between functors pPro dSt q Ñ Cat . The equivalence on the right is the one from [Hen1, S 8 ď0 1.3.0.18]. If X is a pro-stack and E P IQcohpXq then we will denote by X Ñ XrEs Ñ X the image Pro of E by the functor Ex pXq. op Remark 1.1.17. Let us give a description of this functor. Let X be a pro-stack and let X : K Ñ dSt denote a U-small cofiltered diagram of whom X is a limit in Pro dSt . For every k P K we S S can compose the functor mentioned above with the base change functor ´ˆ X X r´s X op // // X X k{ { pQcohpX qq dSt Pro dSt k X X ` ˘ op This is functorial in k and we get a functor colimQcohpXq Ñ Pro dSt which we extend and obtain a more explicit description of the square zero extension functor op X U Xr´s: pIQcohpXqq Ñ Pro dSt Definition 1.1.18. Let X be a pro-stack. • We finally define the functor of derivations over X : ď0 DerpX, ´q “ Map pXr´s, Xq: IQcohpXq Ñ sSets X{´{S • We say that X admits a cotangent complex if the functor DerpX, ´q is corepresentable – ie there ď0 exists a L P IQcohpXq such that for any E P IQcohpXq X{S DerpX, Eq » MappL , Eq X{S Art Definition 1.1.19. Let dSt be the full sub-category of dSt spanned by derived Artin stacks over U Art Art,lfp Art S. An Artin pro-stack is an object of Pro dSt . Let dSt be the full sub-category of dSt S S spanned by derived Artin stacks locally of finite presentation over S. An Artin pro-stack locally of U Art,lfp finite presentation is an object of Pro dSt Proposition 1.1.20. Any Artin pro-stack X over S admits a cotangent complex L . Let us assume X{S Art U Art op that X : K Ñ dSt is a U-small cofiltered diagram of whom X is a limit in Pro dSt . When S S k is a vertex of K, let us denote by X the derived Artin stack Xpkq. If f : k Ñ l is an arrow in K, we will also denote by f : X Ñ X the map of stacks Xpfq. The cotangent complex is given by the l k formula ` ˘ L “ colimp L P Ind colim QcohpXq » IQcohpXq X{S X {S k k 15 where p is the canonical map X Ñ X . The following formula stands k k IQ p L » colim f L k˚ X{S ˚ X {S f : kÑl If X is moreover locally of finite presentation over S, then its cotangent complex belongs to IPerfpXq. Before proving this proposition, let us fix the following notation Definition 1.1.21. Let C be a full sub-category of an 8-category D. There is a natural transformation from O : d ÞÑ D to the constant functor D: D Ñ Cat . We denote by O the fiber product D { 8 d D O “ O ˆC : D Ñ Cat D 8 Remark 1.1.22. The functor O : D Ñ Cat maps an object d P D to the comma category of objects in C over d C “ pC ˆ tduq ˆ D DˆD Proof (of the proposition). The cotangent complex defines a natural transformation Art op pdSt q λ: O op Ñ Qcohp´q dSt To any stack T and any Artin stack U over S with a map f : T Ñ U, it associates the quasi-coherent ˚ Pro complex f L on T . Applying the functor Ind op we get a natural transformation λ U{S dSt U Art op pPro dSt q Pro U op λ “ Ind pλq: O Ñ IQcohp´q U op dSt pPro dSt q S S Specifying it to X we get a functor ´ ¯ op Pro X U Art λ : Pro dSt Ñ IQcohpXq X S Pro Let us set L “ λ pXq P IQcohpXq. We have by definition the equivalence X{S L » colimp L X{S k X {S Let us now check that it satisfies the required universal property. The functor DerpX, ´q is the limit op ď0 of the diagram K Ñ FctpIQcohpXq , sSetsq Map pXr´s, Xq X{´{S ď0 ď0 Fixing k P K, the functor Map pXr´s, X q: IQcohpXq » Ind pcolim QcohpXq q Ñ sSets X{´{S ď0 preserves filtered colimits. It is hence induced by its restriction to colim QcohpXq . It follows that the diagram Map pXr´s, Xq factors through a diagram X{´{S ` ˘ op ď0 ď0 ¯ ¯ δ: K Ñ Fct colim QcohpXq , sSets » lim FctpQcohpXq , sSetsq Similarly, the functor MappL , ´q is the limit of a diagram X{S op // ď0 // ď0 K lim FctpQcohpXq , sSetsq FctpIQcohpXq , sSetsq The universal property of the usual cotangent complex defines an equivalence between δ and µ . IQ To get the formula for p L , one uses [Hen1, 1.2.0.7] and the last statement is obvious. k˚ X{S Remark 1.1.23. The definition of the derived category of ind-quasi-coherent modules on a pro-stack is build for the above proposition and remark to hold. 16 Remark 1.1.24. We have actually proven that for any pro-stack X, the two functors ď0 X Art IQcohpXq ˆ dSt Ñ sSets defined by pE, Y q ÞÑ Map pXrEs, Y q X{´{S Pro pE, Y q ÞÑ Map pλ pY q, Eq IQcohpXq X are equivalent. 1.2 Cotangent complex of an ind-pro-stack Definition 1.2.1. An ind-pro-stack is an object of the category U U IPdSt “ Ind Pro dSt S S op Definition 1.2.2. Let us define the functor PIPerf : pIPdSt q Ñ Cat as PIPerf “ Pro U pIPerfq op pPro dSt q where Pro was defined in Definition 0.1.4. Whenever we have a morphism f : X Ñ Y of ind-pro- stacks, we will denote by f the functor PI f “ PIPerfpfq: PIPerfpY q Ñ PIPerfpXq PI Remark 1.2.3. Let X be an ind-pro-stack. Let X : K Ñ Pro dSt denote a U-small filtered diagram of whom X is a colimit in IPdSt . We have by definition PIPerfpXq » lim Pro pIPerfpXqq Proposition 1.2.4. Let f : X Ñ Y be a map of ind-pro-stacks. If Y is a pro-stack then the functor f : PIPerfpY q Ñ PIPerfpXq admits a right adjoint. PI Definition 1.2.5. Let f : X Ñ Y be a map of ind-pro-stacks. If the functor f : PIPerfpY q Ñ PIPerfpXq PI PI admits a right adjoint, we will denote it by f . PI I Proof (of the proposition). If both X and Y are pro-stacks, then f “ Pro pf q is right adjoint ˚ ˚ U U ˚ ˚ to f “ Pro pf q. Let now X be an ind-pro-stack and let X : K Ñ Pro dSt denote a U-small PI I filtered diagram of whom X is a colimit in IPdSt . We then have f : PIPerfpY q Ñ PIPerfpXq » lim PIPerfpXq PI The right adjoint is the informally given by the formula PI PI f “ lim fpkq ˚ ˚ where fpkq is the induced map Xpkq Ñ Y . For a formal proof, we refer to [Hen1, 1.2.0.5]. op Definition 1.2.6. Let X P IPdSt . We define IPPerfpXq “ pPIPerfpXqq . If X is the colimit in IPdSt of a filtered diagram K Ñ Pro dSt then we have S S IPPerfpXq » limpInd ˝ PPerf ˝Xq op We will denote by p´q : IPPerfpXq Ñ pPIPerfpXqq the duality functor. 17 U V,st,id op Definition 1.2.7. Let us define the functor Tate : pIPdSt q Ñ Cat as the right Kan IP 8 U U op op extension of Tate along the inclusion pPro dSt q Ñ pIPdSt q . It is by definition endowed S S with a canonical fully faithful natural transformation Tate Ñ PIPerf IP For any X P IPdSt , an object of Tate pXq will be called a Tate module on X. IP Remark 1.2.8. We can characterise Tate objects: a module E P PIPerfpXq is a Tate module if and only if for any pro-stack U and any morphism f : U Ñ X P IPdSt , the pullback f pEq is in IP Tate pUq. Let us also remark here that Lemma 1.2.9. Let X be an ind-pro-stack over S. The fully faithful functors _ ´ ¯ op p´q U op U // oo Tate pXq PIPerfpXq pIPPerfpXqq Tate pXq IP IP have the same essential image. We thus have an equivalence ´ ¯ op _ U U p´q : Tate pXq » Tate pXq IP IP Proof. This is a corollary of Proposition 0.1.12. op Definition 1.2.10. Let us define PIQcoh: pIPdSt q Ñ Cat to be the functor PIQcoh “ Pro U pIQcohq op pPro dSt q From Remark 0.1.10, for any ind-pro-stack X, the category PIQcohpXq admits a natural monoidal structure. We also define the subfunctor ď0 U ď0 PIQcoh “ Pro pIQcoh q U op pPro dSt q Remark 1.2.11. Let us give an informal description of the above definition. To an ind-pro-stack X “ colim lim X we associate the category α β αβ ˆ ˙ U U PIQcohpXq “ lim Pro Ind colim PerfpX q αβ Definition 1.2.12. Let f : X Ñ Y be a map of ind-pro-stacks. We will denote by f the functor PI PIQ PIQcohpfq. Whenever it exists, we will denote by f the right adjoint to f . PI Proposition 1.2.13. Let f : X Ñ Y be a map of ind-pro-stacks. If Y is actually a stack, then the induced functor f admits a right adjoint. PI Proof. This is very similar to the proof of Proposition 1.2.4 but using Proposition 1.1.10. Remark 1.2.14. There is a fully faithful natural transformation PIPerf Ñ PIQcoh. Using the same notation f for the images of a map f : X Ñ Y is therefore only a small abuse. Moreover, PI for any such map f : X Ñ Y , for which the right adjoints drawn below exist, there is a natural tranformation // PIPerfpY q PIQcohpY q 3; ♥♥ ♥♥ ♥♥ PI ♥ PIQ f ♥ f ˚ ♥ ˚ ♥♥ ♥♥ ♥♥ ♥♥ // PIPerfpXq PIQcohpXq It is generally not an equivalence. 18 IP U Pro Definition 1.2.15. Let Ex denote the natural transformation Pro U pEx q op pPro dSt q ´ ¯ IP ď0 U > > Ex : PIQcoh Ñ Pro U B U pid q » B oppid q op op ´ ´ pIPdSt q pPro dSt q pPro dSt q S S S op ď0 of functors pIPdSt q Ñ Cat . If X is an ind-pro-stack and E P PIQcohpXq then we will denote S 8 by X Ñ XrEs Ñ X the image of E by the functor ´ ¯ op IP ď0 X Ex pXq: PIQcohpXq Ñ IPdSt Remark 1.2.16. Let us decipher the above definition. Let X “ colim lim X be an ind-pro-stack α β αβ and let E be a pro-ind-module over it. By definition E is the datum, for every α, of a pro-ind-object α U U ď0 α α E in the category Pro Ind pcolim Qcoh pX qq. Let us denote E “ lim colim E . For any β αβ γ δ γδ α ď0 γ and δ, there is a β pγ, δq such that E is in the essential image of Qcoh pX q. We then αβ pγ,δq γδ 0 have XrEs “ colim lim lim X rE s P IPdSt αβ γδ S α,γ δ βěβ pγ,δq Definition 1.2.17. Let X be an ind-pro-stack. • We define the functor of derivations on X DerpX, ´q “ Map pXr´s, Xq X{´{S • We say that X admits a cotangent complex if there exists L P PIQcohpXq such that for X{S ď0 any E P PIQcohpXq DerpX, Eq » MappL , Eq X{S • Let us assume that f : X Ñ Y is a map of ind-pro-stacks and that Y admits a cotangent complex. We say that f is formally étale if X admits a cotangent complex and the natural map f L Ñ L is an equivalence. Y {S X{S Definition 1.2.18. An Artin ind-pro-stack over S is an object in the category Art U U Art IPdSt “ Ind Pro dSt S S An Artin ind-pro-stack locally of finite presentation is an object of Art,lfp U U Art,lfp IPdSt “ Ind Pro dSt S S Proposition 1.2.19. Any Artin ind-pro-stack X admits a cotangent complex L P PIQcohpXq X{S Art Let us assume that X : K Ñ Pro dSt is a U-small filtered diagram of whom X is a colimit in Art IPdSt . For any vertex k P K we will denote by X the pro-stack Xpkq and by i the structural k k map X Ñ X. For any f : k Ñ l in K, let us also denote by f the induced map X Ñ X . We have k k l for all k P K ˚ ˚ i L » lim f L P PIQcohpX q X{S X {S k k,PI I l f : kÑl If moreover X is locally of finite presentation then L belongs to PIPerfpXq. X{S Pro Proof. Let us recall the natural transformation λ from the proof of Proposition 1.1.20 U Art op pPro dSt q Pro U op λ “ Ind pλq: O Ñ IQcohp´q U op dSt pPro dSt q S S 19 U U op of functors pPro dSt q Ñ Cat . Applying the functor Pro U we define the natural S 8 op pPro dSt q IP transformation λ ` ˘ Art op pIPdSt q IP Pro S λ “ Pro U λ : O Ñ PIQcohp´q op op pPro dSt q pIPdSt q S S op between functors pIPdSt q Ñ Cat . Specifying to X we get a functor S 8 ´ ¯ op Art IP X λ : IPdSt Ñ PIQcohpXq X S IP We now define L “ λ pXq. By definition we have X{S ˚ Pro ˚ i L » lim λ pXq » lim f L X{S X X {S k,PI k I l f : kÑl for every k P K. Let us now prove that it satisfies the expected universal property. It suffices to compare for every k P K the functors Map pX r´s, Xq and Map pi L , ´q k X{S X {´{S PIQcohpX q k,PI k k ď0 ď0 defined on PIQcohpX q . They are both pro-extensions to PIQcohpX q of their restrictions k k ď0 IQcohpX q Ñ sSets. The restricted functor Map pX r´s, Xq is a colimit of the diagram k k X {´{S ´ ¯ op k ď0 Map pX r´s, Xq: K Ñ FctpIQcohpX q , sSetsq k k X {´{S while Map pi L , ´q is a colimit to the diagram PIQcohpX q X{S k k,PI ´ ¯ op Pro k ď0 ¯ { Map pλ pXq, ´q: K Ñ FctpIQcohpX q , sSetsq IQcohpX q X k k We finish the proof with Remark 1.1.24. Let us record here a technical result we will use later on. For a proof, we refer to [Hen1, 2.1.2.20]. Art Proposition 1.2.20. Let X P IPdSt . Let us denote by π: X Ñ S the structural map. Let also IP L denote the functor ´ ¯ op Art U U IPdSt Ñ Pro Ind QcohpSq Art op obtained by extending the functor pdSt q Ñ QcohpSq mapping f : T Ñ S to f L . Then we S T{S PIQ IP have π L » L pXq ˚ X{S Definition 1.2.21. Let X by an Artin ind-pro-stack locally of finite presentation over S. We will call the tangent complex of X the ind-pro-perfect complex on X T “ L P IPPerfpXq X{S X{S 1.3 Uniqueness of pro-structure Lemma 1.3.1. Let Y and Z be derived Artin stacks. The following is true (i) The canonical map MappZ, Y q Ñ lim Mappτ Z, Y q ďn is an equivalence; (ii) If Y is q-Artin and Z is m-truncated then the mapping space MappZ, Y q is pm ` qq-truncated. 20 Proof. We prove both items recursively on the Artin degree of Z. The case of Z affine is proved in [TV, C.0.10 and 2.2.4.6]. We assume that the result is true for n-Artin stacks. Let Z be pn`1q-Artin. There is an atlas u: U Ñ Z. Let us remark that for k P N the truncation τ u: τ U Ñ τ Z is also ďk ďk ďk a smooth atlas — indeed we have τ U » U ˆ τ Z. Let us denote by U the nerve of u and by ďk Z ďk ‚ τ U the nerve of τ u. Because k-truncated stacks are stable by flat pullbacks, the groupoid τ U ďk ‚ ďk ďk ‚ is equivalent to τ pU q. We have ďk ‚ MappZ, Y q » lim MappU , Y q » lim lim Mappτ U , Y q » lim Mappτ Z, Y q p ďk p ďk k k rpsPΔ rpsPΔ That proves item (i). If moreover Z is m-truncated, then we can replace U by τ U. If follows that ďm MappZ, Y q is a limit of pm ` qq-truncated spaces. This finishes the proof of (ii). We will use this well known lemma: Lemma 1.3.2 (see [BK, Chapter XI]). Let S : Δ Ñ sSets be a cosimplicial object in simplicial sets. Let us assume that for any rps P Δ the simplicial set S is n-coconnective. Then the natural morphism lim S Ñ lim S p p rpsPΔ rpsPΔ pďn`1 is an equivalence. op Lemma 1.3.3. Let X : N Ñ dSt be a diagram such that (i) There are two integers m and n such that for any k P N the stack Xpkq is n-Artin, m-truncated and of finite presentation; op 1 op (ii) There exists a diagram u¯: N ˆ Δ Ñ dSt such that the restriction of u¯ to N ˆ t1u is ¯ ¯ equivalent to X, every map u¯pkq: u¯pkqp0q Ñ u¯pkqp1q » Xpkq is a smooth atlas and the limit lim u¯pkq is an epimorphism. If Y is an algebraic derived stack of finite presentation then the canonical morphism ` ˘ ` ˘ ¯ ¯ colim Map X, Y Ñ Map lim X, Y is an equivalence. Proof. Let us prove the statement recursively on the Artin degree n. If n equals 0, this is a simple reformulation of the finite presentation of Y . Let us assume that the statement at hand is true for some n and let Xp0q be pn ` 1q-Artin. Considering the nerves of the epimorphisms u¯pkq, we get a diagram op op Z : N ˆ Δ Ñ dSt op ¯ ¯ Note that Z has values in n-Artin stacks. We observe that the diagram lim Zpkq : Δ Ñ dSt is k ‚ S the nerve of the map lim u¯pkq. Since lim u¯pkq is by assumption an epimorphism (whose target is k k lim X), the natural map ¯ ¯ ¯ colim lim Zpkq Ñ lim X » lim colimZpkq p p rpsPΔ kPN kPN rpsPΔ is an equivalence. We now write ˆ ˙ ` ˘ ¯ ¯ Map lim X, Y » Map colim lim Zpkq , Y rpsPΔ kPN ˆ ˙ » lim Map lim Zpkq , Y rpsPΔ kPN ` ˘ » lim colim Map Zpkq , Y rpsPΔ kPN 21 We also have ` ˘ ` ˘ ¯ ¯ colim Map X, Y » colim lim Map Zpkq , Y kPN rpsPΔ It thus suffices to prove that the canonical morphism of simplicial sets ` ˘ ` ˘ ¯ ¯ colim lim Map Zpkq , Y Ñ lim colim Map Zpkq , Y p p kPN kPN rpsPΔ rpsPΔ is an equivalence. Let us notice that each Zpkq is m-truncated. It is indeed a fibre product of m- truncated derived stacks along flat maps. Let q be an integer such that Y is q-Artin. The simplicial set MappZpkq , Y q is then pm`qq-coconnective (Lemma 1.3.1). It follows from Lemma 1.3.2 that the limit at hand is in fact finite and we have the required equivalence. op ¯ ¯ Lemma 1.3.4. Let M : N Ñ sSets be a diagram. For any i P N and any point x “ px q P lim M, we have the following exact sequence ´ ¯ // ¯ // ¯ // ¯ // 0 lim π pMpnq, x q π lim Mpnq, x lim π pMpnq, x q 0 i`1 n i i n n n n A proof of that lemma can be found for instance in [Hir]. op Lemma 1.3.5. Let M : N ˆ K Ñ sSets denote a diagram, where K is a filtered simplicial set. If for any i P N there exists N such that for any n ě N and any k P K the induced morphism i i Mpn, kq Ñ Mpn ´ 1, kq is an i-equivalence then the canonical map φ: colim lim Mpn, kq Ñ lim colim Mpn, kq kPK nPN nPN kPK is an equivalence. We recall that an i-equivalence of simplicial sets is a morphism which induces isomorphisms on the homotopy groups of dimension lower or equal to i. Proof. We can assume that K admits an initial object k . Let us write M instead of Mpn, kq. 0 nk Let us fix i P N. If i ě 1, we also fix a base point x P lim M . Every homotopy group below is n nk computed at x or at the natural point induced by x. We will omit the reference to the base point. We have a morphism of short exact sequences ´ ¯ // // // // 0 colim lim π pM q colim π lim M colim lim π pM q 0 i`1 nk i nk i nk k n k n k n ˆ ˙ // // // // 0 lim colim π pM q π lim colim M lim colim π pM q 0 i`1 nk i nk i nk n k n k n k We can restrict every limit to n ě N . Using the assumption we see that the limits on the right i`1 hand side are then constant and so are the 1-limits on the left. If follows that the vertical maps on the sides are isomorphisms, and so is the middle map. This begin true for any i, we conclude that φ is an equivalence. op ¯ ¯ Definition 1.3.6. Let X : N Ñ dSt be a diagram. We say that X is a shy diagram if (i) For any k P N the stack Xpkq is algebraic and of finite presentation; ¯ ¯ ¯ (ii) For any k P N the map Xpk Ñ k ` 1q: Xpk ` 1q Ñ Xpkq is affine; (iii) The stack Xp0q is of finite cohomological dimension. ¯ ¯ If X is the limit of X in the category of prostacks, we will also say that X is a shy diagram for X. 22 op Proposition 1.3.7. Let X : N Ñ dSt be a shy diagram. If Y is an algebraic derived stack of finite presentation then the canonical morphism ` ˘ ` ˘ ¯ ¯ colim Map X, Y Ñ Map lim X, Y is an equivalence. Proof. Since for any n, the truncation functor τ preserves shy diagrams, let us use Lemma 1.3.1 ďn and Lemma 1.3.3 ¯ ¯ Mapplim X, Y q » lim Mappτ plim Xq, Y q ďn ¯ ¯ » lim Mapplim τ X, Y q » lim colim Mappτ X, Y q ďn ďn n n On the other hand we have ¯ ¯ colim MappX, Y q » colim lim Mappτ X, Y q ďn and we are to study the canonical map ¯ ¯ φ: colim lim Mappτ X, Y q Ñ lim colim Mappτ X, Y q ďn ďn n n Because of Lemma 1.3.5, it suffices to prove the assertion (1) For any i P N there exists N P N such that for any n ě N and any k P N the map i i ` ˘ ` ˘ ¯ ¯ p : Map τ Xpkq, Y Ñ Map τ Xpkq, Y n,k ďn ďn´1 induces an equivalence on the π ’s for any j ď i. For any map f : τ Xpkq Ñ Y we will denote by F pfq the fibre of p at f. We have to prove ďn´1 n,k n,k that for any such f the simplicial set F pfq is i-connective. Let thus f be one of those maps. The n,k ¯ ¯ derived stack τ Xpkq is a square zero extension of τ Xpkq by a module Mrns, where ďn ďn´1 ´ ¯ M “ ker O Ñ O r´ns ¯ ¯ τ Xpkq τ Xpkq ďn ďn´1 Note that M is concentrated in degree 0. It follows from the obstruction theory of Y –see Proposition 0.2.10 – that F pfq is not empty if and only if the obstruction class n,k αpfq P G pfq “ Map pf L , Mrn ` 1sq n,k Y τ Xpkq ďn´1 of f vanishes. Moreover, if αpfq vanishes, then we have an equivalence F pfq » Map pf L , Mrnsq n,k Y τ Xpkq ďn´1 Using assumptions (iii) and (ii) we have that Xpkq — and therefore its truncation too — is of finite cohomological dimension d. Let us denote by ra, bs the Tor-amplitude of L . We get that G pfq is Y n,k ps ` 1q-connective for s “ a ` n ´ d and that F pfq is s-connective if αpfq vanishes. n,k Let us remark here that d and a do not depend on either k or f and thus neither does N “ i`d´a (we set N “ 0 if this quantity is negative). For any n ě N and any f as above, the simplicial set i i G pfq is at least 1-connective. The obstruction class αpfq therefore vanishes and F pfq is indeed n,k n,k i-connective. This proves (1) and concludes this proof. shy U Definition 1.3.8. Let PdSt denote the full subcategory of Pro dSt spanned by the prostacks shy op which admit shy diagrams. Every object X in PdSt is thus the limit of a shy diagram X : N Ñ dSt . 23 op We will say that X is of cotangent tor-amplitude in ra, bs if there exists a shy diagram X : N Ñ dSt for X such that every cotangent L ¯ is of tor-amplitude in ra, bs. We will also say that X Xpnq is of cohomological dimension at most d if there is a shy diagram X with values in derived stacks of cohomological dimension at most d. The pro-stack X will be called q-Artin if there is a shy diagram ra,bs shy for it, with values in q-Artin derived stacks. Let us denote by C the full subcategory of PdSt d,q S spanned by objects of cotangent tor-amplitude in ra, bs, of cohomological dimension at most d and q-Artin. shy Theorem 1.3.9. The limit functor i : PdSt Ñ dSt is fully faithful and has values in Artin shy S stacks. Proof. This follows directly from Proposition 1.3.7. Definition 1.3.10. A map of pro-stacks f : X Ñ Y if an open immersion if there exists a diagram op 1 f : N ˆ Δ Ñ dSt such that • The limit of f in maps of pro-stacks is f; op op • The restriction N ˆt0u Ñ dSt of f is a shy diagram for X and the restriction N ˆt1u Ñ dSt k k is a shy diagram for Y ; • For any n, the induced map of stacks tnu ˆ Δ Ñ dSt is an open immersion. 1.4 Uniqueness of ind-pro-structures shy,b U shy Definition 1.4.1. Let IPdSt denote the full subcategory of Ind pPdSt q spanned by colimits S S shy ra,bs of U-small filtered diagrams K Ñ PdSt which factors through C for some 4-uplet a, b, d, q. For S d,q shy,b any X P IPdSt we will say that X is of cotangent tor-amplitude in ra, bs and of cohomological ra,bs U shy dimension at most d if it is the colimit (in Ind pPdSt q) of a diagram K Ñ C . S d,q U shy Theorem 1.4.2. The colimit functor Ind pPdSt q Ñ dSt restricts to a full faithful embedding shy,b IPdSt Ñ dSt . shy ra,bs Lemma 1.4.3. Let a, b, d, q be integers with a ď b. Let T P PdSt and X : K Ñ C be a U-small S d,q filtered diagram. For any i P N there exists N such that for any n ě N and any k P K, the induced i i map ¯ ¯ Mappτ T, Xpkqq Ñ Mappτ T, Xpkqq ďn ďn´1 is an i-equivalence. We recall that an i-equivalence of simplicial sets is a morphism which induces isomorphisms on the homotopy groups of dimension lower or equal to i. Remark 1.4.4. For the proof of this lemma, we actually do not need the integer q. Proof. Let us fix i P N. Let k P K and T : N Ñ dSt be a shy diagram for T . We observe here that ¯ ¯ ¯ τ T is a shy diagram whose limit is τ T . Let also Y : N Ñ dSt be a shy diagram for Xpkq. The ďn ďn k S map at hand ¯ ¯ ψ : Mappτ T, Xpkqq Ñ Mappτ T, Xpkqq nk ďn ďn´1 is then the limit of the colimits ¯ ¯ ¯ ¯ lim colim Mappτ Tpqq, Y ppqq Ñ lim colim Mappτ Tpqq, Y ppqq ďn k ďn´1 k pPN qPN pPN qPN 24 ¯ Let now f be a map τ T Ñ Xpkq. It corresponds to a family of morphisms ďn´1 ¯ ¯ f : ˚ Ñ colim Mappτ Tpqq, Y ppqq p ďn´1 k qPN For any p, let F pfq denote the fibre of the map nk ¯ ¯ ¯ ¯ ψ : colim Mappτ Tpqq, Y ppqq Ñ colim Mappτ Tpqq, Y ppqq ďn k ďn´1 k nk qPN qPN over the point f . We also set F pfq “ lim F pfq and observe that F pfq is nothing but the fibre p nk p nk nk of ψ over f. nk To prove the result, it suffices to show that for any such f, the fibre F pfq is i-connective. Using nk the exact sequence of Lemma 1.3.4, it suffices to prove that F pfq is pi ` 1q-connective for any f and nk any p. Fixing such an f and such a p, there exists q P N such that the map f factors through the 0 p canonical map ¯ ¯ ¯ ¯ Mappτ Tpq q, Y ppqq Ñ colim Mappτ Tpqq, Y ppqq ďn´1 0 k ďn´1 k qPN We deduce that F pfq is equivalent to the colimit nk p pq F pfq » colim G pfq nk nk qěq pq where G pfq is the fibre at the point induced by f of the map nk ¯ ¯ ¯ ¯ Mappτ Tpqq, Y ppqq Ñ Mappτ Tpqq, Y ppqq ďn k ďn´1 k The interval ra, bs contains the tor-amplitude of L ¯ and d is an integer greater than the cohomo- Y ppq pq logical dimension of Tpqq. We saw in the proof of Proposition 1.3.7 that G pfq is then pa ` n ´ dq- nk connective. We set N “ i ` d ´ a ` 1. Proof (of Theorem 1.4.2). We will prove the sufficient following assertions U shy (1) The colimit functor Ind pPdSt q Ñ PpdAff q restricts to a fully faithful functor shy,b η: IPdSt Ñ PpdAff q (2) The functor η has values in the full subcategory of stacks. shy Let us focus on assertion (1) first. We consider two U-small filtered diagrams X : K Ñ PdSt and shy Y : L Ñ PdSt . We have ` ˘ ¯ ¯ ¯ ¯ Map shy colimX, colimY » lim Map shy pXpkq, colim Y q U U Ind pPdSt q Ind pPdSt q S k S and ` ˘ ` ˘ ¯ ¯ ¯ ¯ Map colimi X, colim i Y » lim Map i Xpkq, colim i Y shy shy shy shy PpdAffq PpdAffq shy shy We can thus replace the diagram X in PdSt by a simple object X P PdSt . We now assume that S S ra,bs Y factors through C for some a, b, d, q. We have to prove that the following canonical morphism d,q is an equivalence ` ˘ ¯ ¯ φ: colim Mappi X, i Y plqq Ñ Map i X, colimi Y shy shy shy shy lPL where the mapping spaces are computed in prestacks. If i X is affine then φ is an equivalence because shy colimits in PpdAff q are computed pointwise. Let us assume that φ is an equivalence whenever i X S shy is pq´1q-Artin and let us assume that i X is q-Artin. Let u: U Ñ i X be an atlas of i X and let shy shy shy 25 shy Z be the nerve of u in dSt . We saw in the proof of Lemma 1.3.3 that Z factors through PdSt . ‚ S ‚ The map φ is now equivalent to the natural map ¯ ¯ colim Mappi X, i Y plqq Ñ lim colim MappZ , i Y plqq shy shy p shy lPL rpsPΔ lPL ` ˘ ¯ ¯ » lim Map Z , colim i Y » Mappi X, colim i Y q p shy shy shy rpsPΔ Remembering Lemma 1.3.1, it suffices to study the map ¯ ¯ colim lim Mappτ i X, i Y plqq Ñ lim colim lim Mappτ Z , i Y plqq ďn shy shy ďn p shy n n lPL rpsPΔ lPL Applying Lemma 1.4.3 and then Lemma 1.3.5, we see that φ is an equivalence if the natural morphism ¯ ¯ lim colim lim Mappτ Z , i Y plqq Ñ lim lim colim Mappτ Z , i Y plqq ďn p shy ďn p shy n n lPL rpsPΔ rpsPΔ lPL is an equivalence. The stack i Y plq is by assumption q-Artin, where q does not depend on l. Now shy using Lemma 1.3.1 and Lemma 1.3.2, we conclude that φ is an equivalence. This proves (1). We now focus on assertion (2). If suffices to see that the colimit in PpdAff q of the diagram i Y as above S shy op is actually a stack. Let H : Δ Y t´1u Ñ dAff be an hypercovering of an affine SpecpAq “ H . ‚ S ´1 We have to prove the following equivalence ¯ ¯ colim lim MappH , i Y plqq Ñ lim colim MappH , i Y plqq p shy p shy l rpsPΔ rpsPΔ l Using the same arguments as for the proof of (1), we have ¯ ¯ colim lim MappH , i Y plqq » colim lim lim Mappτ H , i Y plqq p shy ďn p shy l rpsPΔ l rpsPΔ » lim colim lim Mappτ H , i Y plqq ďn p shy n l rpsPΔ » lim lim colim Mappτ H , i Y plqq ďn p shy n l rpsPΔ » lim colim lim Mappτ H , i Y plqq ďn p shy l n rpsPΔ » lim colim MappH , i Y plqq p shy rpsPΔ l shy,b We will need one last lemma about that category IPdSt . shy,b Lemma 1.4.5. The fully faithful functor IPdSt XIPdAff Ñ IPdSt Ñ dSt preserves finite S S S limits. shy,b Proof. The case of an empty limit is obvious. Let then X Ñ Y Ð Z be a diagram in IPdSt X IPdAff . There exist a and b and a diagram ´ ¯ ra,bs σ: K Ñ Fct Λ ,C 1 0,0 such that K is a U-small filtered simplicial set and the colimit in IPdSt is X Ñ Y Ð Z. We U U 2 2 can moreover assume that σ has values in FctpΛ , Pro pdAff qq » Pro pFctpΛ , dAff qq. We S S 1 1 deduce that the fibre product X ˆ Z is the realisation of the ind-pro-diagram in derived affine stacks with cotangent complex of tor amplitude in ra ´ 1, b ` 1s. It follows that X ˆ Z is again in shy,b IPdSt X IPdAff . 26 2 Symplectic Tate stacks 2.1 Tate stacks: definition and first properties We can now define what a Tate stack is. Definition 2.1.1. A Tate stack is a derived Artin ind-pro-stack locally of finite presentation whose cotangent complex – see Proposition 1.2.19 – is a Tate module. Equivalently, an Artin ind-pro-stack Tate locally of finite presentation is Tate if its tangent complex is a Tate module. We will denote by dSt the full subcategory of IPdSt spanned by Tate stacks. This notion has several good properties. For instance, using Lemma 1.2.9, if a X is a Tate stack then comparing its tangent T and its cotangent L makes sense, in the category of Tate modules X X over X. We will explore that path below, defining symplectic Tate stacks. Another consequence of Tatity is the existence of a determinantal anomaly as defined in [KV2]. Let us consider the natural morphism of prestacks U Tate θ: Tate Ñ K U U U Tate where Tate denote the prestack A ÞÑ Tate pPerfpAqq and K : A ÞÑ KpTate pPerfpAqqq – K denoting the connective K-theory functor. From [Hen2, Section 5] we have a determinant Tate K Ñ KpG , 2q where KpG , 2q is the Eilenberg-Maclane classifying stack. Definition 2.1.2. We define the Tate determinantal map as the composite map Tate Ñ KpG , 2q To any derived stack X with a Tate module E, we associate the determinantal anomaly rdet s P H pX,O q, image of E by the morphism MappX, Tate q Ñ MappX, KpG , 2qq Let now X be an ind-pro-stack. Let also R denote the realisation functor Pro dSt Ñ dSt k k mapping a pro-stack to its limit in dSt . Let finally X : K Ñ Pro dSt denote a U-small filtered k k diagram whose colimit in IPdSt is X. We have a canonical functor U U U ¯ ¯ F : lim Tate pXq » Tate pXq Ñ lim Tate pRXq P IP Definition 2.1.3. Let X be an ind-pro-stack and E be a Tate module on X. Let X be the realisation of X in Ind dSt and X be its image in dSt . We define the determinantal anomaly of E the image k k of F pEq by the map 1 1 2 Map U pX , Tate q Ñ Map U pX , KpG , 2qq » Map pX , KpG , 2qq m m Ind dSt Ind dSt dSt k k k 2 2 ˆ In particular if X is a Tate stack, we will denote by rdet s P H pX ,O q the determinantal anomaly X 2 associated to its tangent T P Tate pXq. IP The author plans on studying more deeply this determinantal class in future work. For now, let us conclude this section with following Tate Lemma 2.1.4. The inclusion dSt Ñ IPdSt preserves finite limits. or Tateness or Tatitude 27 Proof. Let us first notice that a finite limit of Artin ind-pro-stacks is again an Artin ind-pro-stack. Let now X Ñ Y Ð Z be a diagram of Tate stacks. The fibre product // X ˆ Z X p g // Z Y is an Artin ind-pro-stack. It thus suffices to test if its tangent T is a Tate module. The following Xˆ Z cartesian square concludes // T p T Xˆ Z X Y X ˚ ˚ // ˚ p T p g T Z Y Z X 2.2 Shifted symplectic Tate stacks We assume now that the basis S is the spectrum of a ring k of characteristic zero. Recall from [PTVV] the stack in graded complexes DR mapping a cdga over k to its graded complex of forms. It actually comes with a mixed structure induced by the de Rham differential. The authors also defined there the stack in graded complexes NC mapping a cdga to its graded complex of closed forms. Those two stacks are linked by a morphism NC Ñ DR forgetting the closure. p p,cl ď0 We will denote by A , A : cdga Ñ dgMod the complexes of weight p in DRr´ps and w p p,cl NC r´ps respectively. The stack A will therefore map a cdga to its complexes of p-forms while A will map it to its closed p-forms. For any cdga A, a cocycle of degree n of A pAq is an n-shifted p,cl p p-forms on Spec A. The functors A and A extend to functors op p,cl p A , A : dSt Ñ dgMod p p,cl Definition 2.2.1. Let us denote by A and A the extensions IP IP op U U pIPdSt q Ñ Pro Ind dgMod p,cl p p p,cl of A and A , respectively. They come with a natural projection A Ñ A . IP IP Let X P IPdSt . An n-shifted (closed) p-form on X is a morphism kr´ns Ñ A pXq (resp. IP p,cl p,cl p,cl p A pXq). For any closed form ω : kr´ns Ñ A pXq, the induced map kr´ns Ñ A pXq Ñ A pXq IP IP IP IP is called the underlying form of ω. Remark 2.2.2. In the above definition, we associate to any ind-pro-stack X “ colim lim X its α β αβ complex of forms p U U A pXq “ lim colim A pX q P Pro Ind dgMod αβ IP α β For any ind-pro-stack X, the derived category PIQcohpXq is endowed with a canonical monoidal structure. In particular, one defines a symmetric product E ÞÑ Sym pEq as well as an antisymmetric PI product E ^ E “ Sym pEr´1sqr2s PI PI Theorem 2.2.3. Let X be an Artin ind-pro-stack over k and π: X Ñ ˚ the projection. The push- forward functor PIQ V V π : PIQcohpXq Ñ Pro Ind pdgMod q exists (see Proposition 1.2.13) and maps L ^ L to A pXq. In particular, any 2-form kr´ns Ñ X PI X IP A pXq corresponds to a morphism O r´ns Ñ L ^ L in PIQcohpXq. X X PI X IP 28 Proof. This follows from [PTVV, 1.14], from Proposition 1.2.20 and from the equivalence IP IP U U U U U U λ ^ λ “ Pro Ind pλq ^ Pro Ind pλq » Pro Ind pλ ^ λq PI PI IP where λ is defined in the proof of Proposition 1.2.19. Definition 2.2.4. Let X be a Tate stack. Let ω : kr´ns Ñ A pXq be an n-shifted 2-form on X. It IP induces a map in the category of Tate modules on X ω: T Ñ L rns X X We say that ω is non-degenerate if the map ω is an equivalence. A closed 2-form is non-degenerate if the underlying form is. Definition 2.2.5. A symplectic form on a Tate stack is a non-degenerate closed 2-form. A symplectic Tate stack is a Tate stack equipped with a symplectic form. 2.3 Mapping stacks admit closed forms In this section, we will extend the proof from [PTVV] to ind-pro-stacks. Note that if X is a pro- ind-stack and Y is a stack, then MappX, Y q is an ind-pro-stack. We will then need an evaluation functor MappX, Y q ˆ X Ñ Y . It appears that this evaluation map only lives in the category of ind-pro-ind-pro-stacks colim lim colim lim MappX , Y q ˆ X Ñ Y αζ βξ α β ξ ζ To use this map properly, we will need the following remark. Definition 2.3.1. Let C be a category. There is one natural fully faithful functor φ: PIpCq Ñ pIPq pCq but three IPpCq Ñ pIPq pCq. The first one is given by applying IP to the canonical embedding functor C Ñ IPpCq. The second one by considering the canonical embedding functor D Ñ IPpDq for D “ IPpCq. In this work, we will only consider the third functor ψ: IPpCq Ñ pIPq pCq U U given by applying Ind to the canonical embedding D Ñ PIpDq for D “ Pro pCq. Let us also denote by ξ the natural fully faithful functor C Ñ pIPq pCq. Definition 2.3.2. Let Y be a stack and X be a pro-ind-stack. Let us denote the evaluation map in IP dSt X,Y // ev : ψ Map pX, Y q ˆ φX ξY For a formal definition of this map, we refer to [Hen1, 2.2.3.2]. We assume now that S “ Spec k. Let us recall the following definition from [PTVV, 2.1] Definition 2.3.3. A derived stack X is O-compact if for any derived affine scheme T the following conditions hold • The quasi-coherent sheaf O is compact in QcohpX ˆ Tq ; XˆT • Pushing forward along the projection X ˆ T Ñ T preserves perfect complexes. Let us denote by dSt the full subcategory of dSt spanned by O-compact derived stacks. 29 Definition 2.3.4. An O-compact pro-ind-stack is a pro-ind-object in the category of O-compact derived stacks. We will denote by PIdSt their category. Lemma 2.3.5. There is a functor ` ˘ 1 1 2 op PIdSt Ñ Fct IPdSt ˆ Δ ˆ Δ , pIPq pdgMod q k k defining for any O-compact pro-ind-stack X and any ind-pro-stack F a commutative square p,cl p,cl // A pψF ˆ φXq A pψFq b φO 2 k X IP IP p p // A pψF ˆ φXq A pψFq b φO 2 k X IP IP p,cl p p,cl p where A and A are the extensions of A and A to 2 2 IP IP IP IP op 2 2 pIPq dSt Ñ pIPq pdgMod q Proof. Recall in [PTVV, part 2.1] the construction for any O-compact stack X and any stack F of a commutative diagram (of graded complexes): w w // NC pF ˆ Xq NC pFq b π O k ˚ X // DRpF ˆ Xq DRpFq b π O k ˚ X where π: X Ñ ˚. Taking the part of weight p and shifting, we get p,cl // p,cl A pF ˆ Xq A pFq b π O k ˚ X p p // A pF ˆ Xq A pFq b π O k ˚ X This construction is functorial in both F and X so it corresponds to a functor O 1 1 op dSt Ñ FctpdSt ˆ Δ ˆ Δ , dgMod q k k We can now form the functor ` ˘ O 1 1 op PIdSt Ñ PI Fct Pro dSt ˆ Δ ˆ Δ , PropdgMod q k k ` ˘ op 1 1 Ñ Fct Pro dSt ˆ Δ ˆ Δ , PI PropdgMod q ` ˘ op 1 1 2 Ñ Fct IPdSt ˆ Δ ˆ Δ , pIPq pdgMod q By construction, for any ind-pro-stack F and any O-compact pro-ind-stack, it induces the commutative diagram p,cl p,cl // A pψF ˆ φXq ψA pFq b φO 2 k X IP IP p p // A pψF ˆ φXq ψA pFq b φO 2 k X IP IP 30 Remark 2.3.6. Let us remark that we can informally describe the horizontal maps using the maps from [PTVV]: Θ 2pψF ˆ φXq “ lim colim lim colim ΘpF ˆ X q IP αδ βγ α γ β δ Ñ lim colim lim colim ΘpF q b pO q “ ψΘ pFq b φO αδ X IP X βγ α γ β δ p,cl p where Θ is either A or A . Definition 2.3.7. Let F be an ind-pro-stack and let X be an O-compact pro-ind-stack. Let η: O Ñ p,cl p kr´ds be a map of ind-pro-k-modules. Let finally Θ be either A or A . We define the integration map id bφη // // : Θ 2pψF ˆ φXq ψΘ pFq b φO ψΘ pFqr´ds IP IP X IP Theorem 2.3.8. Let Y be a derived stack and ω be an n-shifted closed 2-form on Y . Let X be an O-compact pro-ind-stack, let π: X Ñ ˚ be the projecton, and let η: π O Ñ kr´ds be a map. The ˚ X mapping ind-pro-stack MappX, Y q admits an pn ´ dq-shifted closed 2-form. Proof. Let us denote by Z the mapping ind-pro-stack MappX, Y q. We consider the diagram ω η Y ev 2,cl 2,cl // 2,cl // // χkr´ns χA pY q A pX ˆ Zq ψA pZqr´ds IP IP ξ ψ op op op where χ: dgMod Ñ IPpdgMod q Ñ pIPq pdgMod q is the canonical inclusion. Note that since k k k op the functor ψ is fully faithful, this induces a map in IPpdgMod q 2,cl // ξk A pZqrn ´ ds IP and therefore a pn´dq-shifted closed 2-form on Z “ MappX, Y q. The underlying form is given by the composition ω η Y ev 2,cl 2 2 2 // // // // χkr´ns χA pY q χA pY q A pX ˆ Zq ψA pZqr´ds IP IP Remark 2.3.9. Let us describe the form issued by Theorem 2.3.8. We set the notations X “ lim colim X and Z “ MappX , Y q. By assumption, we have a map α β αβ αβ αβ η: colim limO Ñ kr´ds αβ For any α, there exists therefore βpαq and a map η : O Ñ kr´ds in dgModpkq. Unwinding αβpαq X αβpαq the definitions, we see that the induced form ω 2 2 // ξk A pMappX, Y qqrn ´ ds » lim colim A pZ qrn ´ ds α β αβ IP is the universal map obtained from the maps αβpαq // 2 // 2 k A pZ qrn ´ ds colim A pZ qrn ´ ds αβpαq β αβ where ω is built using η and the procedure of [PTVV]. Note that ω can be seen as a αβpαq αβpαq αβpαq map T b T Ñ O . We also know from Theorem 2.2.3 that the form ω induces a X X X Y αβpαq αβpαq αβpαq map T b T Ñ O rn ´ ds Z Z Z 31 in IPPpZq. Let us fix α and pull back the map above to Z . We get 0 α ˚ ˚ ˚ colim lim g p pT b T q » i pT b T q Ñ O rn ´ ds Z Z Z Z Z α α αβ αβ αβ α α 0 0 0 αěα 0 β This map is the universal map obtained from the maps ˚ ˚ ˚ ˚ lim g p pT b T q Ñ g p pT b T q Z Z Z Z α0α αβ αβ αβ α0α αβpαq αβpαq αβpαq ˚ ˚ Ñ g p pO qrn ´ ds » O rn ´ ds X X α α α 0 αβpαq αβpαq where g is the structural map Z Ñ Z and p is the projection Z “ lim Z Ñ Z . α α α α αβ α β αβ αβ 0 0 2.4 Mapping stacks have a Tate structure Definition 2.4.1. Let S be an O-compact pro-ind-stack. We say that S is an O-Tate stack if there op exist a poset K and a diagram S : K Ñ Ind dSt such that (i) The limit of S in PIdSt is equivalent to S ; (ii) For any i ď j P K the pro-module over Spiq ` ˘ coker O ¯ Ñ Spi ď jq O ¯ Spiq Spjq is trivial in the pro-direction – ie belong to QcohpSpiqq. (iii) For any i ď j P K the induced map Spi ď jq is represented by a diagram f : L ˆ Δ Ñ dSt such that ¯ ¯ • For any l P L the projections fpl, 0q Ñ ˚ and fpl, 1q Ñ ˚ satisfy the base change formula ; • For any l P L the map fplq satisfies the base change and projection formulae ; • For any m ď l P L the induced map fpm ď l, 0q satisfies the base change and projection formulae. Remark 2.4.2. We will usually work with pro-ind-stacks S given by an explicit diagram already satisfying those assumptions. Proposition 2.4.3. Let us assume that Y is a derived Artin stack locally of finite presentation. Let S be an O-compact pro-ind-stack. If S is an O-Tate stack then the ind-pro-stack MappS, Y q is a Tate stack. op U Proof. Let Z “ MappS, Y q as an ind-pro-stack. Let S : K Ñ Ind dSt be as in Definition 2.4.1. ¯ ¯ ¯ We will denote by Z : K Ñ Pro dSt the induced diagram and for any i P K by s : Zpiq Ñ Z the k i induced map. Let us first remark that Z is an Artin ind-pro-stack locally of finite presentation. It suffices to prove that s L is a Tate module on Zpiq, for any i P K. Let us fix such an i and denote by Z the Z i pro-stack Zpiq. We consider the differential map s L Ñ L Z Z i i It is by definition equivalent to the natural map Pro Pro lim λ pZ| ěiq Ñ λ pZ q K i Z Z i i 32 ěi { ěi where K is the comma category K and Z| is the induced diagram ěi Z K Ñ Pro dSt Let φ denote the diagram ` ˘ op ěi φ : K Ñ IPerfpZ q i i obtained as the kernel of f. It is now enough to prove that φ factors through PerfpZ q. i i Let j ě i in K and let us denote by g the induced map Z Ñ Z of pro-stacks. Let f : LˆΔ Ñ ij i j ¯ ¯ ¯ dSt represents the map Spi ď jq: Spjq Ñ Spiq P Ind dSt as in assumption (i) in Definition 2.4.1. k k Up to a change of L through a cofinal map, we can assume that the induced diagram ` ˘ coker O ¯ Ñ Spi ď jq O ¯ Spiq Spjq op 1 is essentially constant – see assumption (ii). We denote by h: L ˆΔ Ñ dSt the induced diagram, ¯ ¯ so that g is the limit of h in Pro dSt . For any l P L we will denote by h : Z Ñ Z the map hplq. ij k l il jl ¯ ¯ Let us denote by Z the induced diagram l ÞÑ Z and by Z the diagram l ÞÑ Z . Let also p denote i il j jl l the projection Z Ñ Z i il We have an exact sequence ˚ ˚ ˚ φ pjq Ñ colim p h L Ñ colim p L i Z Z l l jl l il l l Let us denote by ψ the diagram obtained as the kernel ij Pro Pro ¯ ¯ ψ Ñ λ pZ q Ñ λ pZ q ij j i Z Z i i so that φ pjq is the colimit colimψ in IPerfpZ q. It suffices to prove that the diagram ψ : L Ñ i ij i ij PerfpZ q is essentially constant (up to a cofinal change of posets). By definition, we have ψ plq » p L r´1s ij l Z {Z il jl Let m Ñ l be a map in L and t the induced map Z Ñ Z . The map ψ pm Ñ lq is equivalent to il im ij the map p ξ where ξ fits in the fibre sequence in PerfpZ q il ˚ ˚ ˚ ˚ // // t L r´1s t h L t L Z {Z Z Z im jm m jm im // // L r´1s h L L Z Z Z {Z jl il il jl l We consider the dual diagram ˚ ˚ ˚ ˚ oo oo t T r1s t h T t T Z {Z Z Z im jm m jm im OO OO OO pσq oo oo T r1s h T T Z Z Z {Z jl il il jl l Using base change along the maps from S , S and S to the point, we get that the square pσq is im jm jl equivalent to ˚ oo ˚ π pid ˆsf q pid ˆsf q E π pid ˆsq pid ˆsq E ˚ m ˚ m ˚ ˚ OO OO oo π pid ˆf q pid ˆf q E π E ˚ l ˚ l ˚ 33 where π: Z ˆ S Ñ Z is the projection, where s: S Ñ S is the map induced by m Ñ l and il il il im il where E » ev T with ev: Z ˆ S Ñ Y the evaluation map. Note that we use here the well known Y il il fact T » pr ev T where MappX,Y q Y pr ev oo // MappX, Y q MappX, Y q ˆ X Y are the canonical maps. Now using the projection and base change formulae along the morphisms s, f and f we get l m that pσq is equivalent to the image by π of the square ˚ oo ˚ E b p s f O E b p s O ˚ m S ˚ S ˚ jm im OO OO ˚ oo ˚ E b p f O E b p O l S S ˚ jl il We therefore focus on the diagram oo s f O s O ˚ m S ˚ S jm im OO OO oo f O O l S S ˚ jl il The map induced between the cofibres is an equivalence, using assumption (ii). It follows that the diagram ψ is essentially constant, and thus that Z is a Tate stack. ij 3 Formal loops In this part, we will at last define and study the higher dimensional formal loop spaces. We will prove it admits a local Tate structure. 3.1 Dehydrated algebras and de Rham stacks In this part, we define a refinement of the reduced algebra associated to a cdga. This allows us to define a well behaved de Rham stack associated to an infinite stack. Indeed, without any noetherian assumption, the nilradical of a ring – the ideal of nilpotent elements – is a priori not nilpotent itself. The construction below gives an alternative definition of the reduced algebra – which we call the dehydrated algebra – associated to any cdga A, so that A is, in some sense, a nilpotent extension of its dehydrated algebra. Whenever A is finitely presented, this construction coincides with the usual reduced algebra. ď0 Definition 3.1.1. Let A P cdga . We define its dehydrated algebra as the ind-algebra A “ deh 0 0 colim H pAq where the colimit is taken over the filtered poset of nilpotent ideals of H pAq. The I { case I “ 0 gives a canonical map A Ñ A in ind-cdga’s. This construction is functorial in A. deh Remark 3.1.2. Whenever A is of finite presentation, then A is equivalent to the reduced algebra deh associated to A. In that case, the nilradical A of A is nilpotent. Moreover, if A is any cdga, it is a filtered colimits of cdga’s A of finite presentation. We then have A » colimpA q in ind-algebras. α deh α red Lemma 3.1.3. The colimit B of the ind-algebra A in the category of algebras is equivalent to the deh reduced algebra A . red Proof. Let us first remark that B is reduced. Indeed any nilpotent element x of B comes from a nilpotent element of A. It therefore belongs to a nilpotent ideal pxq. This define a natural map of algebras A Ñ B. To see that it is an isomorphism, it suffices to say that A is the union of all red nilpotent ideals. 34 Definition 3.1.4. Let X be a prestack. We define its de Rham prestack X as the composition dR p´q Ind pXq deh colim ď0 U ď0 U // // // cdga Ind pcdga q Ind psSetsq sSets k k This defines an endofunctor of p8, 1q-category PpdAff q. We have by definition ´ ¯ X pAq “ colimX H pAq dR Remark 3.1.5. If X is a stack of finite presentation, then it is determined by the images of the cdga’s of finite presentation. The prestack X is then the left Kan extension of the functor dR ď0,fp cdga Ñ sSets A ÞÑ XpA q red Definition 3.1.6. Let f : X Ñ Y be a functor of prestacks. We define the formal completion X of X in Y as the fibre product // X X Y dR // Y Y dR This construction obviously defines a functor FC: PpdAff q Ñ PpdAff q. k k ď0 Remark 3.1.7. The natural map X Ñ Y is formally étale, in the sense that for any A P cdga and any nilpotent ideal I Ă H pAq the morphism ` ˘ ˆ ˆ X pAq Ñ X H pAq ˆ Y pAq Y Y { Y pH pAq{ q is an equivalence. 3.2 Higher dimensional formal loop spaces Here we finally define the higher dimensional formal loop spaces. To any cdga A we associate the d d formal completion V of 0 in A . We see it as a derived affine scheme whose ring of functions ArrX ss 1...d A A is the algebra of formal series in d variables X , . . . , X . Let us denote by U the open subscheme of 1 d ď0 V complementary of the point 0. We then consider the functors dSt ˆ cdga Ñ sSets A k d d L : pX, Aq ÞÑ Map pV , Xq V dSt A d d L : pX, Aq ÞÑ Map pU , Xq U dSt A d d ˜ ˜ Definition 3.2.1. Let us consider the functors L and L as functors dSt Ñ PpdAffq. They come U V d d d d ˜ ˜ ˜ ˜ with a natural morphism L Ñ L . We define L to be the pointwise formal completion of L into V U V L : ´ ¯ d d d ˜ ˜ ˜ L pXq “ FC L pXq Ñ L pXq V U d d d d d d ˜ ˜ ˜ We also define L , L and L as the stackified version of L , L and L respectively. We will call U V U V L pXq the formal loop stack in X. Remark 3.2.2. The stack L pXq is a higher dimensional analogue to the stack of germs in X, as studied for instance by Denef and Loeser in [DL]. 35 d Remark 3.2.3. By definition, the derived scheme U is the (finite) colimit in derived stacks ´ ¯ d ´1 U “ colim colim Spec ArrX ssrX s 1...d i ...i 1 q q i ,...,i 1 q ´1 ´1 ´1 where ArrX ssrX s denote the algebra of formal series localized at the generators X , . . . , X . 1...d i ...i i i 1 q 1 q It follows that the space of A-points of L pXq is equivalent to the simplicial set ´ ´ ¯ ¯ d ´1 I L pXqpAq » colim lim lim Map Spec ArrX ssrX s , X 1...d i ...i 1 q q i ,...,i IĂH pAq 1 q ´1 ´1 where ArrX ssrX s is the sub-cdga of ArrX ssrX s consisting of series 1...d 1...d i ...i i ...i 1 q 1 q n n 1 d a X . . .X n ,...,n 1 d 1 d n ,...,n 1 d where a is in the kernel of the map A Ñ H pAq as soon as at least one of the n ’s is negative. n ,...,n { i 1 d I Recall that in the colimit above, the symbol I denotes a nilpotent ideal of H pAq. Lemma 3.2.4. Let X be a derived Artin stack of finite presentation with algebraisable diagonal (see d d Definition 0.2.11) and let t: T “ SpecpAq Ñ X be a smooth atlas. The induced map L pTq Ñ L pXq V V is an epimorphism of stacks. d d ˜ ˜ Proof. It suffices to study the map L pTq Ñ L pXq. Let B be a cdga. Let us consider a B-point V V x: Spec B Ñ L pXq. It induces a B-point of X Spec B Ñ SpecpBrrX ssq Ñ X 1...d Because t is an epimorphism, there exists an étale map f : Spec C Ñ Spec B and a commutative diagram // Spec C T f t // Spec B X It corresponds to a C-point of Spec B ˆ T . For any n P N, let us denote by C the cdga X n C :“ Crx , . . ., x s n n n 1 d px , . . ., x q 1 d n n and by S the spectrum Spec C We also set B “ Brx , . . . , x s and X “ Spec B . n n n 1 d n n px , . . . , x q 1 d Finally, we define T as the pullback T ˆ X . We will also consider the natural fully faithful functor n X n Δ » t0, . . ., nu Ñ N. We have a natural diagram 2,2 2 0 α : Λ ˆ N > Δ ˆ Δ Ñ dSt 0 k 2,2 0 Λ ˆΔ informally drown has a commutative diagram // // // S . . . S . . . 0 n // // // X . . . X . . . 0 n OO OO !! // // // T . . . T . . . 0 n 36 Let n P N and let us assume we have built a diagram 2,2 2 n α : pΛ ˆ Nq > Δ ˆ Δ Ñ dSt n k 2,2 n Λ ˆΔ extending α . There is a sub-diagram of α n´1 n // S S n n`1 // T T n n`1 n`1 n`1 Since the map t is smooth (it is a pullback of t), we can complete this diagram with a map n`1 S Ñ T and a commutative square. Using the composition in dSt , we get a diagram α n`1 n`1 k n`1 extending α . We get recursively a diagram α: Δ ˆ N Ñ dSt . Taking the colimit along N, we get n k a commutative diagram // // Spec C colim Spec C T f t // // // Spec B colim Spec B SpecpBrrX ssq X n n 1...d This defines a map φ: colim SpecpC q Ñ SpecpBrrX ssq ˆ T . We have the cartesian diagram n 1...d X // SpecpBrrX ssq ˆ T X 1...d X // SpecpBrrX ssq ˆ T X ˆ X 1...d The diagonal of X is algebraisable and thus so is the stack SpecpBrrX ssq ˆ T . The morphism φ 1...d X therefore defines the required map SpecpCrrX ssq Ñ SpecpBrrX ssq ˆ T 1...d 1...d Remark 3.2.5. Let us remark here that if X is an algebraisable stack, then L pXq is a stack, hence the natural map is an equivalence d d L pXq » L pXq V V ď0 Lemma 3.2.6. Let f : X Ñ Y be an étale map of derived Artin stacks. For any cdga A P cdga and any nilpotent ideal I Ă H pAq, the induced map ` ˘ d d 0 d ˜ // ˜ ˜ θ: L pXqpAq L pXq H pAq ˆ L pY qpAq I U U U d 0 L pY qpH pAq q U I is an equivalence. Proof. The map θ is a finite limit of maps ` ` ˘˘ // µ : XpξAq X ξ H pAq Y pξAq { ˆ Y ξ H pAq p p { qq 37 ´1 where ξA “ ArrX ssrX s is obtained from the cdga of formal power series in A with d variables 1...d i ...i 1 p 0 0 by inverting the variables x . Let also ξpH pAq q be defined similarly. The natural map ξpH pAqq Ñ i I ξpH pAq q is also a nilpotent extension. We deduce from the étaleness of f that the map ` ` ˘˘ 0 0 0 // XpξpH pAqqq X ξ H pAq ˆ Y pξpH pAqqq Y pξpH pAq qq is an equivalence. Let now n P N. We assume that the natural map ` ` ˘˘ // XpξpA qq X ξ H pAq ˆ Y pξpA qq ďn { ďn Y pξpH pAq qq ´n´1 is an equivalence. The cdga ξpA q » pξAq is a square zero extension of ξpA q by H pξAq. ďn`1 ďn`1 ďn We thus have the equivalence // XpξpA qq XpξpA qq ˆ Y pξpA qq ďn`1 ďn ďn`1 Y pξpA qq ďn The natural map ` ` ˘˘ // 0 XpξpA qq X ξ H pAq ˆ Y pξpA qq ďn`1 I ďn`1 Y ξ H pAq p p { qq is thus an equivalence too. The stacks X and Y are nilcomplete, hence µ is also an equivalence – recall that a derived stack X is nilcomplete if for any cdga B we have XpBq » lim XpB q ďn Also recall that any Artin stack is nilcomplete. It follows that θ is an equivalence. ď0 Corollary 3.2.7. Let f : X Ñ Y be an étale map of derived Artin stacks. For any cdga A P cdga and any nilpotent ideal I Ă H pAq, the induced map ` ˘ d d 0 d ˜ // ˜ ˜ θ: L pXqpAq L pXq H pAq ˆ L pY qpAq d 0 L pY qpH pAq q is an equivalence. Proposition 3.2.8. Let X be a derived Deligne-Mumford stack of finite presentation with algebraisable d d diagonal. Let t: T Ñ X be an étale atlas. The induced map L pTq Ñ L pXq is an epimorphism of stacks. ď0 d d ˜ ˜ Proof. We can work on the map of prestacks L pTq Ñ L pXq. Let A P cdga . Let x be an A-point of L pXq. It corresponds to a vertex in the simplicial set ` ˘ d 0 d ˜ ˜ colimL pXq H pAq ˆ L pXqpAq V I U d 0 pXq H pAq L p { q U I There exists therefore a nilpotent ideal I such that x comes from a commutative diagram d d // U U H pAq // V 0 X H pAq v 38 Using Lemma 3.2.4 we get an étale morphism ψ: A Ñ B such that the map v lifts to a map u: V Ñ T where J is the image of I by ψ. This defines a point in ` ˘ d 0 ˜ ˜ L pTq H pBq ˆ L pXqpBq U U d 0 L pXq H pBq p { q U J Because of Lemma 3.2.6, we get a point of L pTqpBq. We now observe that this point is compatible with x. In the case of dimension d “ 1, Lemma 3.2.6 can be modified in the following way. Let f : X Ñ Y ď0 be a smooth map of derived Artin stacks. For any cdga A P cdga and any nilpotent ideal I Ă H pAq, the induced map ` ˘ 1 1 0 1 ˜ // ˜ ˜ θ: L pXqpAq L pXq H pAq ˆ L pY qpAq I U U U 1 0 L pY qpH pAq q U I is essentially surjective. The following proposition follows. Proposition 3.2.9. Let X be an Artin derived stack of finite presentation and with algebraisable 1 1 diagonal. Let t: T Ñ X be a smooth atlas. The induced map L pTq Ñ L pXq is an epimorphism of stacks. 1 1 Example 3.2.10. The proposition above implies for instance that L pBGq » BL pGq for any algebraic group G – where BG is the classifying stack of G-bundles. 3.3 Tate structure and determinantal anomaly We saw in subsection 2.1 that to any Tate stack X, we can associate a determinantal anomaly. It a 2 d class in H pX,O q. We will prove in this subsection that the stack L pXq is endowed with a structure of Tate stack as soon as X is affine. We will moreover build a determinantal anomaly on L pXq for any quasi-compact and separated scheme X. ď0 Lemma 3.3.1. For any B P cdga of finite presentation, the functors ď0 d d ˜ ˜ L pSpec Bq,L pSpec Bq: cdga Ñ sSets are in the essential image of the fully faithful functor shy,b IPdSt X IPdAff Ñ IPdSt Ñ dSt Ñ PpdAffq k k k d d d d ˜ ˜ (see Definition 1.4.1). It follows that L pSpec Bq » L pSpec Bq and L pSpec Bq » L pSpec Bq. U U Proof. Let us first remark that Spec B is a retract of a finite limit of copies of the affine line A . It follows that the functor L pSpec Bq is, up to a retract, a finite limit of functors ´ ¯ d ´1 Z : A ÞÑ Map krY s, ArrX ssrX s E 1...d i ...i 1 q where E “ ti , . . . , i u Ă F “ t1, . . ., du. The functor Z is the realisation of an affine ind-pro-scheme 1 q Z » colim lim Specpkra , ´nδ ď α ď psq α ,...,α i i E 1 d n p where δ “ 1 if i P E and δ “ 0 otherwise. The variable a corresponds to the coefficient i i α ,...,α 1 d α α shy,b 1 d d of X . . .X . The functor Z is thus in the category IPdSt X IPdAff . The result about 1 E d d ˜ ˜ L pSpec Bq then follows from Lemma 1.4.5. The case of L pSpec Bq is similar: we decompose it into a finite limit of functors ´ ¯ d ´1 I G : A ÞÑ colim Map krY s, ArrX ssrX s 1...d E i ...i 1 q IĂH pAq 39 0 d where I is a nilpotent ideal of H pAq. We then observe that G is the realisation of the ind-pro-scheme ´ ¯ G » colim lim Spec kra , ´nδ ď α ď ps α ,...,α i i { E 1 d n,m p where J is the ideal generated by the symbols a with at least one of the α ’s negative. α ,...,α 1 d Remark 3.3.2. Let n and p be integers and let kpE, n, pq denote the number of families pα , . . . , α q 1 d such that ´nδ ď α ď p for all i. We have i i d 1 kpE,n,pq Z » colim limpA q n p fp Definition 3.3.3. From Lemma 3.3.1, we get a functor L : dAff Ñ IPdSt . It follows from Proposition 3.2.8 that L is a costack in ind-pro-stacks. We thus define d lfp L : dSt Ñ IPdSt fp lfp lfp to be its left Kan extension along the inclusion dAff Ñ dSt – where dSt is p8, 1q-category of k k k derived stacks locally of finite presentation. This new functor L preserves small colimits by definition. Proposition 3.3.4. There is a natural transformation θ from the composite functor d IP L |´| lfp // // dSt IPdSt dSt k k to the functor L . Moreover, the restriction of θ to derived Deligne-Mumford stacks of finite presen- tation with algebraisable diagonal is an equivalence. Proof. There is by definition a natural transformation d IP d θ: |L p´q| Ñ L p´q Moreover, the restriction of θ to affine derived scheme of finite presentation is an equivalence – see Lemma 3.3.1. The fact that θ is an equivalence for any Deligne-Mumford stack X follows from Proposition 3.2.8. Lemma 3.3.5. Let F be a non-empty finite set. For any family pM q of complexes over k indexed by subsets D of F , we have colim M » M rd ´ 1s D F H‰EĂF H‰DĂE where d is the cardinal of F (the maps in the colimit diagram are the canonical projections). Proof. We can and do assume that F is the finite set t1, . . ., du and we proceed recursively on d. The case d “ 1 is obvious. Let now d ě 2 and let us assume the statement is true for F r tdu. Let pM q be a family as above. We have a cocartesian diagram à à // colim M colim M D D tduĹEĂF H‰EĂFrtdu H‰DĂE H‰DĂE // M colim M tdu H‰EĂF H‰DĂE We have by assumption colim M » M rd ´ 2s D Frtdu H‰EĂFrtdu H‰DĂE 40 and ˜ ¸ ˜ ¸ à à à colim M » M ‘ colim M ‘ colim M D tdu D D tduĹEĂF tduĹEĂF tduĹEĂF H‰DĂE tduĹDĂE H‰DĂErtdu » M ‘ M rd ´ 2s ‘ M rd ´ 2s tdu F Frtdu The result follows. ď0 d Lemma 3.3.6. For any B P cdga of finite presentation, the ind-pro-stack L pSpec Bq is a Tate stack. Proof. Let us first focus on the case of the affine line A . We have to prove that the cotangent p,n complex L d is a Tate module. For any subset D Ă F we define M to be the free k-complex L pA q D generated by the symbols ta , ´n ď α ă 0 if i P D, 0 ď α ď p otherwiseu α ,...,α i i 1 d in degree 0. From the proof of Lemma 3.3.1, we have p,n d d 1 d Z » colim lim Specpkr M sq and L pA q » lim Z E DĂE D U E n p H‰EĂF where F “ t1, . . ., du. If we denote by π the projection L pA q Ñ Spec k, we get ˜ ¸ ˜ ¸ à à ˚ p,n ˚ p,n L d » π colim lim colim M » π lim colim colim M L pA q D D n p n p H‰EĂF H‰EĂF DĂE DĂE Using Lemma 3.3.5 we have ˆ ˙ p,n p,n L d » π lim colimM ‘ M rd ´ 1s L pA q H F n p p,n p,0 p,n 0,n Moreover, we have M » M and M » M . It follows that L d is a Tate module H H F F L pA q d d on the ind-pro-stack L pA q. The case of L pSpec Bq then follows from Lemma 1.4.5 and from U U Lemma 2.1.4. Lemma 3.3.7. Let B Ñ C be an étale map between cdga’s of finite presentation. The induced map d d f : L pSpec Cq Ñ L pSpec Bq is formally étale – see Definition 1.2.17. U U Proof. Let us denote X “ Spec B and Y “ Spec C. We have to prove that the induced map ´ ¯ ´ ¯ d d d d j : Map d L pY qr´s,L pY q Ñ Map d L pY qr´s,L pXq L pY q{´ U U L pY q{´ U U U U d d ď0 is an equivalence of functors PIQcohpL pY qq Ñ sSets. Since L pY q is ind-pro-affine, we can restrict to the study of the morphism ´ ¯ ´ ¯ d d j : Map Zr´s,L pY q Ñ Map Zr´s,L pXq Z{´ U Z{´ U ď0 of functors IQcohpZq Ñ sSets, for any pro-affine scheme Z and any map Z Ñ L pY q. Let us fix ď0 E P IQcohpZq . The pro-stack ZrEs is in fact an affine pro-scheme. Recall that both L pY q and d shy,b L pXq belong to IPdSt . It follows from the proof of Theorem 1.4.2 that the morphism j pEq is U k equivalent to ` ˘ ` ˘ d d |j pEq|: Map |ZrEs|,L pY q Ñ Map |ZrEs|,L pXq |Z|{´ U |Z|{´ U where | ´ | is the realisation functor and the mapping spaces are computed in dSt . It now suffices to see that |ZrEs| is a trivial square zero extension of the derived affine scheme |Z| and to use Lemma 3.2.6. 41 Proposition 3.3.8. Let Spec B be a derived affine scheme of finite presentation. The ind-pro- stack L pSpec Bq admits a cotangent complex. This cotangent complex is moreover a Tate module. d d For any étale map B Ñ C the induced map f : L pSpec Cq Ñ L pSpec Bq is formally étale – see Definition 1.2.17. d d Proof. Let us write Y “ Spec B. Let us denote by i: L pY q Ñ L pY q the natural map. We will prove that the map i is formally étale, the result will then follow from Lemma 3.3.6 and Lemma 3.3.7. To do so, we consider the natural map ´ ¯ ´ ¯ d d d d j : Map d L pY qr´s,L pY q Ñ Map d L pY qr´s,L pY q L pY q{´ L pY q{´ U ď0 of functors PIQcohpL pY qq Ñ sSets. To prove that j is an equivalence, we can consider for every ď0 affine pro-scheme X Ñ L pY q the morphism of functors IQcohpXq Ñ sSets ´ ¯ ´ ¯ d d j : Map Xr´s,L pY q Ñ Map Xr´s,L pY q X{´ X{´ U ď0 Let us fix E P IQcohpXq . The morphism j pEq is equivalent to ` ˘ ` ˘ d d |j pEq|: Map |XrEs|,L pY q Ñ Map |XrEs|,L pY q |X|{´ |X|{´ U where the mapping space are computed in dSt . The map |j pEq| is a pullback of the map k X ` ˘ ` ˘ d d f : Map |XrEs|,L pY q Ñ Map |XrEs|,L pY q dR dR |X|{´ V |X|{´ U It now suffices to see that |XrEs| is a trivial square zero extension of the derived affine scheme |X| and thus f is an equivalence (both of its ends are actually contractible). Let us recall from Definition 2.1.3 the determinantal anomaly ´ ¯ 2 d ˆ rDet d s P H L pSpec Aq,O L pSpecAq L pSpecAq U d It is associated to the tangent T d P Tate pL pSpec Aqq through the determinant map. Using L pSpecAq IP Proposition 3.3.8, we see that this construction is functorial in A, and from Proposition 3.2.8 we get that it satisfies étale descent. Thus, for any quasi-compact and quasi-separated (derived) scheme (or Deligne-Mumford stack with algebraisable diagonal), we have a well-defined determinantal anomaly ´ ¯ 2 d ˆ rDet d s P H L pXq,O L pXq d L pXq Remark 3.3.9. It is known since [KV3] that in dimension d “ 1, if rDet 1 s vanishes, then there L pXq are essentially no non-trivial automorphisms of sheaves of chiral differential operators on X. 4 Bubble spaces In this section, we study the bubble space, an object closely related to the formal loop space. We will then prove the bubble space to admit a symplectic structure. 4.1 Two lemmas In this subsection, we will develop two duality results we will need afterwards. ď0 0 0 Let A P cdga be a cdga over a field k. Let pf , . . . , f q be points of A whose images in H pAq 1 p form a regular sequence. 42 n n Let us denote by A the Kozsul complex associated to the regular sequence pf , . . ., f q for n,k 1 k k ď p. We set A “ A and A “ A for any n. If k ă p, the multiplication by f induces an n,0 n n,p k`1 n n endomorphism ϕ of A . Recall that A is isomorphic to the cone of ϕ : n,k n,k`1 k`1 k`1 k`1 // A A n,k n,k // 0 A n,k`1 Let us now remark that for any couple pn, kq, the A-module A is perfect. n,k Lemma 4.1.1. Let k ď p. The A-linear dual A “ RHom pA , Aq of A is equivalent to n,k n,k n,k A r´ks; n,k Proof. We will prove the statement recursively on the number k. When k “ 0, the result is trivial. Let k ě 0 and let us assume that A is equivalent to A r´ks. Let us also assume that for any n,k n,k a P A, the diagram induced by multiplication by a commutes { „ A A r´ks n,k n,k _ „ A A r´ks n,k n,k We obtain the following equivalence of exact sequences k`1 // // A r´k ´ 1s A r´ks A r´ks n,k`1 n,k n,k „ „ „ pϕ q k`1 _ _ _ { { { A A A // // A A A n,k`1 n,k n,k The statement about multiplication is straightforward. Lemma 4.1.2. Let us assume A is a formal series ring over A : A “ A rrf , . . ., f ss 1 1 p It follows that for any n, the A -module A is free of finite type and that there is map r : A Ñ A 1 n n n 1 n n mapping f . . . f to 1 and any other generator to zero. We deduce an equivalence 1 p „ { A Ñ A “ RHom pA , A q n n 1 given by the pairing ˆ r // // A b A A A n A n n 1 Remark 4.1.3. Note that we can express the inverse A Ñ A of the equivalence above: it maps n n a function α: A Ñ A to the serie n 1 i n´1´i αpf qf i p i 1 where i varies through the uplets pi , . . . , i q and where f “ f . . . f . 1 p p 43 4.2 Definition and properties We define here the bubble space, obtained from the formal loop space. We will prove in the next sections it admits a structure of symplectic Tate stack. Definition 4.2.1. The formal sphere of dimension d is the pro-ind-stack S “ lim colim SpecpA ‘ Hom pA , Aqq » lim colim SpecpA ‘ A r´dsq p n p n n pěn n pěn n n where A “ krx , . . ., x s and A “ A . 1 d n px , . . . , x q 1 d Remark 4.2.2. The notation SpecpA ‘ A r´dsq is slightly abusive. The cdga A ‘ A r´ds is not p n p n concentrated in non positive degrees. In particular, the derived stack SpecpA ‘ A r´dsq is not a p n derived affine scheme. It behaves like one though, regarding its derived category: QcohpSpecpA ‘ A r´dsqq » dgMod p n A ‘A r´ds p n Let us define the ind-pro-algebra O “ colim lim A ‘ A r´ds ˆd p n n pěn where A ‘ A r´ds is the trivial square zero extension of A by the module A r´ds. For any m P N, p n p n let us denote by S the ind-stack S “ colim SpecpA ‘ A r´dsq p m pěm Definition 4.2.3. Let T be a derived Artin stack. We define the d-bubble stack of T as the mapping ind-pro-stack BpTq “ MappS , Tq: Spec B ÞÑ colim lim TpB b pA ‘ A r´dsqq p n n pěn Again, the cdga A ‘A r´ds is not concentrated in non positive degree. This notation is thus slightly p n abusive and by TpB b pA ‘ A r´dsqq we mean p n MappSpecpA ‘ A r´dsq ˆ Spec B, Tq p n We will denote by BpTq the diagram N Ñ Pro dSt of whom BpTq is a colimit in IPdSt . Let us k k also denote by B pTq the mapping pro-stack B pTq “ MappS , Tq: Spec B ÞÑ lim TpB b pA ‘ A r´dsqq p m m m pěm op and B pTq: tp P N|p ě mu Ñ dSt the corresponding diagram. In particular B pTq “ MappS , Tq: Spec B ÞÑ lim TpB b A q 0 0 Those stacks come with natural maps 0 r // // B pTq BpTq B pTq 0 0 // B pTq BpTq Proposition 4.2.4. If T is an affine scheme of finite type, the bubble stack BpTq is the product in ind-pro-stacks // BpTq L pTq d d // L pTq L pTq V U 44 d d Proof. There is a natural map V Ñ S induced by the morphism colim lim A ‘ A r´ds Ñ lim A p n p n pěn p Because T is algebraisable, it induces a map BpTq Ñ L pTq and thus a diagonal morphism d d δ: BpTq Ñ L pTq ˆ L pTq V V L pTq We will prove that δ is an equivalence. Note that because T is a (retract of a) finite limit of copies of A , d d 1 1 1 we can restrict to the case T “ A . Let us first compute the fibre product Z “ L pA qˆ d L pA q. V L pA q V It is the pullback of ind-pro-stacks // Z lim Specpkra , 0 ď α ď psq α ,...,α i // lim Specpkra , 0 ď α ď psq colim lim lim Specpkra , ´nδ ď α ď psq α ,...,α i α ,...,α iPI i 1 d 1 d p n p IĂJ p,n where J “ t1, . . ., du and δ “ 1 if i P I and 0 otherwise. For any subset K Ă J we define M to iPI be the free complex generated by the symbols ta , ´n ď α ă 0 if i P K, 0 ď α ď p otherwiseu α ,...,α i i We then have the cartesian diagram ´ ¯ p,0 // Z lim Spec krM s ´ ¯ p,0 p,n // lim Spec krM s colim lim lim Specpkr M sq p n p IĂJ H KĂI K Using Lemma 3.3.5 we get ´ ” ı¯ p,0 0,n Z » colim lim Spec k M ‘ M rds H J n p Remark 4.2.5. Let us consider the map lim A Ñ A » k mapping a formal serie to its coefficient p p 0 of degree 0. The plim A q-ind-module colimA r´ds is endowed with a natural map to kr´ds. This p n d d d induces a morphism O Ñ k ‘ kr´ds and hence a map S Ñ S , where S is the topological sphere ˆd of dimension d. We then have a rather natural morphism B pXq Ñ MappS , Xq 4.3 Its tangent is a Tate module We will prove in this subsection that the bubble stack is a Tate stack. To do so, we could bluntly apply Proposition 2.4.3 but we will give here a direct proof of that statement. We will get another decomposition of its tangent complex that will be needed when proving B pTq is symplectic. Proposition 4.3.1. Let us assume that the Artin stack T is locally of finite presentation. For any m P N we have an exact sequence ˚ ˚ // ˚ // ˚ s r L d s L d s L d d m B pTq m B pTq m B pTq{B pTq 0 0 where the left hand side is an ind-perfect module and the right hand side is a pro-perfect module. In particular, the middle term is a Tate module, and the ind-pro-stack B pTq is a Tate stack. 45 d d Proof. Throughout this proof, we will write B instead of B pTq and B instead of B pTq for any m. Let us first remark that B is an Artin ind-pro-stack locally of finite presentation. It suffices to prove that s L is a Tate module on B , for any m P N. We will actually prove that it is an m m elementary Tate module. We consider the map ˚ ˚ ˚ s r L Ñ s L B B m m It is by definition equivalent to the natural map Pro Pro λ pB q Ñ lim λ pB pTqq B 0 B ěm m m ¯ ¯ where B pTq is the restriction of BpTq to tn ě mu Ă N. Let φ denote the diagram ěm op φ: tn P N|n ě mu Ñ IPerfpB pTqq obtained as the cokernel of f. It is now enough to prove that φ factors through PerfpB pTqq. Let n ě m be an integer and let g denote the induced map B pTq Ñ B pTq. We have an exact mn m n sequence ˚ ˚ ˚ ˚ ˚ ˚ s r L » g s r L Ñ g L Ñ φpnq B pTq B pTq B pTq m mn n m,n 0 0 n Let us denote by ψpnq the cofiber ˚ ˚ s r L Ñ L Ñ ψpnq B pTq B pTq 0 n so that φpnq » g ψpnq. This sequence is equivalent to the colimit (in IPerfpB pTqq) of a cofiber mn n op sequence of diagrams tp P N|p ě nu Ñ PerfpB pTqq Pro Pro ¯ ¯ ¯ λ pB pTqq Ñ λ pB pTqq Ñ ψpnq B pTq 0 B pTq n n n op It suffices to prove that the diagram ψpnq: tp P N|p ě nu Ñ PerfpB pTqq is (essentially) constant. Let p P N, p ě n. The perfect complex ψpnqppq fits in the exact sequence ˚ ˚ ˚ t ε L Ñ π L Ñ ψpnqppq B pTq B pTq np np n,p 0,p n,p where t : B pTq Ñ B pTq is the canonical projection and ε : B pTq Ñ B pTq is induced by np np n n,p n,p 0,p the augmentation O Ñ O . It follows that ψpnqppq is equivalent to S S n,p 0,p t L np B pTq{B pTq n,p 0,p ¯ ¯ Moreover, for any q ě p ě n, the induced map ψpnqppq Ñ ψpnqpqq is obtained (through t ) from the nq cofiber, in PerfpB pTqq n,q ˚ ˚ // ˚ // ˚ α ε L α L α L B pTq B pTq B pTq{B pTq npq np npq npq 0,p n,p n,p 0,p ˚ ˚ ε α L B pTq nq 0pq 0,p pσq // // ε L L L nq B pTq B pTq B pTq{B pTq 0,q n,q n,q 0,q 46 where α is the map B pTq Ñ B pTq. Let us denote by pσq the square on the left hand side npq n,q n,p above. Let us fix a few more notations oo // B pTq ˆ S B pTq ˆ S B pTq ˆ S 0,p 0,p 0,q n,p n,q n,q 0p r xxr ϕ ϕ np ψ nq S npq 0,p npq oo // np B pTq ˆ S B pTq ˆ S B pTq ˆ S n,p n,p n,q n,p n,q n,q r r r r r r ❏ r r r r r r ❏ ev nq r r r a a r np r r nq ❏ r r r xxr yyr yyr npq %% oo nq S S S T n,p n,p n,q np npq oo B pTq B pTq B pTq n,p n,q “ n,q The diagram pσq is then dual to the diagram ˚ ˚ ˚ ˚ ˚ oo α ε ̟ ev T α ̟ ev T 0p T np T npq np 0p npq np ˚ ˚ OO OO ˚ ˚ ˚ oo ε ̟ ev T ̟ ev T 0q T nq T nq 0q nq ˚ ˚ Moreover, the functor ̟ (for any n and p) satisfies the base change formula. This square is thus np equivalent to the image by ̟ of the square nq ˚ ˚ ˚ ˚ ˚ oo ψ b b ψ ev T b b ev T npq npq T npq T ˚ ˚ npq npq nq ˚ npq nq OO OO ˚ ˚ ˚ oo ϕ ϕ ev T ev T nq T T ˚ nq nq nq Using now the projection and base change formulae along the morphisms ϕ , b and ψ , we see nq npq npq that this last square is again equivalent to ˚ ˚ ˚ ˚ oo pa β ξ O q b pev T q pa β O q b pev T q npq np S T npq S T nq 0,p nq nq n,p nq ˚ ˚ ˚ OO OO ˚ ˚ ˚ ˚ oo pa ξ O q b pev T q pa O q b pev T q nq S T S T nq 0,q nq nq n,q nq We therefore focus on the diagram // O ξ O S nq S n,q 0,q // β O β ξ O npq S npq np S n,p 0,p ˚ ˚ ˚ By definition, the fibres of the horizontal maps are both equivalent to A r´ds and the map induced by the diagram above is an equivalence. We have proven that for any q ě p ě n the induced map ¯ ¯ ψpnqppq Ñ ψpnqpqq is an equivalence. It implies that L is a Tate module. BpTq 47 4.4 A symplectic structure (shifted by d) In this subsection, we will prove the following Theorem 4.4.1. Assume T is q-shifted symplectic. The ind-pro-stack B pTq admits a symplectic Tate structure shifted by q ´ d. Moreover, for any m P N we have an exact sequence ˚ ˚ ˚ ˚ ˚ s r L d Ñ s L d Ñ s r T d rq ´ ds m B pTq m B pTq m B pTq 0 0 Proof. Let us start with the following remark: the residue map r : A Ñ k “ A defined in n n 1 Lemma 4.1.2 defines a map O Ñ kr´ds. From Theorem 2.3.8, we have a pq ´ dq-shifted closed ˆd 2-form on B pTq. We have a morphism from Theorem 2.2.3 O d rq ´ ds Ñ L d b L d B pTq B pTq B pTq in PIPerfpB pTqq. Let m P N. We get a map ˚ ˚ O d rq ´ ds Ñ s L d b s L d B pTq m B pTq m B pTq and then ˚ ˚ s T d b s T d Ñ O d rq ´ ds m B pTq m B pTq B pTqm in IPPerfpB pTq q. We consider the composite map ˚ ˚ ˚ ˚ θ: s T d d b s T d d Ñ s T d b s T d Ñ O d rq ´ ds m B pTq{B pTq m B pTq{B pTq m B pTq m B pTq B pTq 0 0 m Using the Remark 2.3.9 and the proof of Proposition 4.3.1 we see that θ is induced by the morphisms (varying n and p) ´ ¯ ` ˘ A B // // ̟ E b E b ev pT b T q ̟ pE b Erqsq ̟ O d rqs np T T np np B pTq ˆS ˚ np ˚ ˚ np n,p ˚ ! where E “ a ξ h γ O d and the map A is induced by the symplectic form on T . The map B np np A np ˚ ˚ n is induced by the multiplication in O . This sheaf of functions is a trivial square zero extension of Sn,p augmentation ideal ξ h γ O d and B therefore vanishes. It follows that the morphism np np n A ˚ ˚ ˚ ˚ ˚ ˚ s T d b s T d d Ñ s T d b s T d Ñ O d rq ´ ds m B pTq m B pTq{B pTq m B pTq m B pTq B pTq 0 m ˚ ˚ factors through s T d b s T d d . Now using Proposition 4.3.1 we get a map of exact m B pTq m B pTq{B pTq 0 0 sequences in the category of Tate modules over B pTq ˚ ˚ ˚ ˚ // // s T d d s T d s r T d m B pTq{B pTq m B pTq m B pTq 0 0 ˚ ˚ // ˚ // ˚ s r L d rd ´ qs s L d rd ´ qs s L d d rd ´ qs m B pTq0 m B pTq m B pTq{B pTq0 where the maps on the sides are dual one to another. It therefore suffices to see that the map ˚ ˚ ˚ τ : s T d d Ñ s r L d rd ´ qs is an equivalence. We now observe that τ is a colimit m m m B pTq{B pTq m B pTq 0 0 indexed by p ě m of maps ´ ¯ ˚ ˚ ˚ g t ε L d Ñ T d d pm pp pp B pTq B pTq {B pTq 0p pp 0p 48 ˚ ˚ Let us fix p ě m and G “ a ξ O . The map F : T d d Ñ ε L d at hand is pp S p pp 0p B pTq {B pTq pp B pTq ˚ pp 0p 0p induced by the pairing ` ˘ ` ˘ ˚ ˚ T d d b ε T d » ̟ E b ev T b ̟ G b ev T B pTq {B pTq B pTq pp T pp T pp 0p pp 0p pp pp ˚ ˚ ` ˘ ˚ ˚ ̟ E b ev T b G b ev T pp T T pp pp ̟ pE b Gqrqs pp ´ ¯ ̟ O d rqs pp B pTq ˆS ˚ pp pp O d rq ´ ds B pTq pp We can now conclude using Lemma 4.1.2. References [Ber] Julia E. Bergner: A survey of 8-categories. IMA volumes in Math. and Appl., pages 69–83. 2010. [BGW1] Oliver Braunling, Michael Groechenig and Jesse Wolfson: A Generalized Contou- Carrère Symbol and its Reciprocity Laws in Higher Dimensions, 2014. [BGW2] Oliver Braunling, Michael Groechenig and Jesse Wolfson: Tate Objects in Exact Categories (with appendix by Jan Šťovíček and Jan Trlifaj), 2014. [Bha] Bhargav Bhatt: Algebraization and Tannaka duality. Camb. J. Math., 4(4) pp.403–461, [BK] Aldridge Bousfield and Daniel M. Kan: Homotopy limits, completions and localizations, volume 304 of Lecture Notes in Mathematics. Springer-Verlag, Berlin-New York, 1972. [CC] Carlos Contou-Carrère: Jacobienne locale, groupe de bivecteurs de Witt universel et symbole modéré. C. R. Acad. Sci. Paris Sér. I Math., 318(8) pp.743–746, 1994. [DL] Jan Denef and Franois Loeser: Germs of arcs on singular algebraic varieties and motivic integration. Inventiones Mathematica, 135 pp.201–232, 1999. [GR] Dennis Gaitsgory and Nick Rozenblyum: DG Indschemes. In Perspectives in represen- tation theory, volume 610 of Contemp. Math., pages 139–251. AMS, 2014. [HAlg] Jacob Lurie: Higher algebra. 2016. [Hen1] Benjamin Hennion: Formal loop spaces and tangent Lie algebras. PhD thesis, 2015. [Hen2] Benjamin Hennion: Tate objects in stable (8,1)-categories. Homology, Homotopy and Applications, 19(2) pp.373–395, 2017. 49 [Hir] Philip S. Hirschhorn: The homotopy groups of the inverse limit of a tower of fibrations. [HLP] Daniel Halpern-Leistner and Anatoly Preygel: Mapping stacks and categorical notions of properness, 2014. [Hov] Mark Hovey: Model categories, volume 63 of Mathematical Surveys and Monographs. Amer- ican Mathematical Society, Providence, RI, 1999. [HTT] Jacob Lurie: Higher topos theory, volume 170 of Annals of Mathematics Studies. Princeton University Press, Princeton, NJ, 2009. [KV1] Mikhail Kapranov and Éric Vasserot: Vertex algebras and the formal loop space. Publ. Math. Inst. Hautes Études Sci, 100 pp.209–269, 2004. [KV2] Mikhail Kapranov and Éric Vasserot: Formal loops II: A local Riemann-Roch theorem for determinantal gerbes. Ann. Sci. ENS, 40 pp.113–133, 2007. [KV3] Mikhail Kapranov and Éric Vasserot: Formal loops IV: Chiral differential operators, [Lur] Jacob Lurie: Derived algebraic geometry XII: Proper morphisms, com- pletions and the Grothendieck existence theorem. 2011, available at [http://www.math.harvard.edu/~lurie/papers/DAG-XII.pdf]. [MP] Vincente Muñoz and Francisco Presas: Geometric structures on loop and path spaces. Proc. Indian Acad. Sci., 120(4) pp.417–428, 2010. [OZ] Denis Osipov and Xiewen Zhu: The two-dimensional Contou-Carrère symbol and reci- procity laws. to appear in Journal of Alg. geometry, 2013. [PTVV] Tony Pantev, Bertrand Toën, Michel Vaquié and Gabriele Vezzosi: Shifted symplectic structures. Publication de l’IHES, 2011. [Toë1] Bertrand Toën: Simplicial Presheaves and Derived Algebraic Geometry. Advanced Courses in Math. Birkhäuser, 2010. [Toë2] Bertrand Toën: Derived algebraic geometry. EMS Surv. Math. Sci., 1(2) pp.153–240, 2014. [TV] Bertrand Toën and Gabriele Vezzosi: Homotopical Algebraic Geometry II: geometric stacks and applications. Memoirs of the AMS, 193(902) pp.257–372, 2008.

Journal

Mathematics – arXiv (Cornell University)

Published: Oct 20, 2020

There are no references for this article.