Fourier–Mukai equivalences for formal groups and elliptic Hochschild homology

Sarah Scherotzke, Nicolò Sibilla, Paolo Tomasini · 2025

Read Fourier–Mukai equivalences for formal groups and elliptic Hochschild homology by Sarah Scherotzke, Nicolò Sibilla, Paolo Tomasini (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-00172.txt.

Introduction

Sarah Scherotzke, Nicolò Sibilla, and Paolo Tomasini unify several flavors of a homology theory used in modern geometry, in a 2025 paper. Hochschild homology measures how a mathematical structure twists around itself, and twisting it by an elliptic curve gives an elliptic version with two competing definitions. The authors build a bridge using Fourier-Mukai duality, a correspondence trick, extended to formal groups, the infinitesimal shadows of curves.

When the formal group comes from an elliptic curve, they prove the two definitions of elliptic Hochschild homology agree. Two plainer theories, ordinary and Hodge Hochschild homology, fall out as limiting cases tied to degenerate curves. The team then defines global versions of the theory spread over the moduli stacks of elliptic and cubic curves. Those versions hint at a universal form of the theory that interpolates among all the others.

Catalog record

CREATORSarah Scherotzke, Nicolò Sibilla, Paolo Tomasini
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE69 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-00172

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Fourier–Mukai equivalences for formal groups and elliptic Hochschild homology Sarah Scherotzke, Nicolò Sibilla, Paolo Tomasini Abstract . This paper establishes a unifying framework for various forms of twisted Hochschild homology by comparing two definitions of elliptic Hochschild homology: one introduced by Moulinos–Robalo–Toën and the other by Sibilla–Tomasini. Central to our approach is a new Fourier–Mukai duality for formal groups. We prove that when E ^ \widehat{E} is the formal group associated to an elliptic curve E E , the resulting E ^ \widehat{E} -Hochschild homology coincides with the mapping stack construction of Sibilla–Tomasini. This identification also recovers ordinary and Hodge Hochschild homology as degenerate limits corresponding to nodal and cuspidal cubics, respectively. Building on this, we introduce global versions of elliptic Hochschild homology over the moduli stacks of elliptic and cubic curves, which interpolate between these theories and suggest a universal form of TMF-Hochschild homology. 1. Introduction In the beautiful paper [ MRT22 ] the authors investigate the behaviour of Hochschild homology in mixed characteristics. One of their discoveries is that, at least for affine derived schemes, it is possible to define twisted versions of Hochschild homology that depend on the choice of a one-dimensional formal group \mathbb{G} . This gives rise in particular to a notion of elliptic Hochschild homology, when \mathbb{G} is the completion of an elliptic curve E at the identity section. In independent work, the two last named authors gave a definition of elliptic Hochschild homology in [ ST23 ] ; see also the recent paper [ BK25 ] for a very interesting alternative construction which is equivalent to the one from [ ST23 ] . One of the sources of interests of elliptic Hochschild homology as defined in [ ST23 ] is that the authors prove that, when working over \mathbb{C} , it recovers the complexified (equivariant) elliptic cohomology of the analytification, when turning on a natural circle action. One of the main goals of this article is showing that the definition of elliptic Hochschild homology given in [ MRT22 ] and in [ ST23 ] agree. The way we achieve this is by revisiting the construction from [ MRT22 ] through the lenses of a Fourier–Mukai duality for formal groups which we introduce, and which we believe to be of independent interest. This yields an alternative definition \mathbb{G} -Hochschild homology in the sense of [ MRT22 ] , for any formal group \mathbb{G} , which is arguably more concrete. When \mathbb{G}=\widehat{E} is the completion of an elliptic curve, this is the key ingredient in the comparison with the definition from [ ST23 ] . The other technical input is Lurie’s theory of Fourier–Mukai duality for dual abelian varieties over an arbitrary \mathbb{E}_{\infty} -ring. Building on our results, we give a definition of TMF-Hochschild homology as a universal form of elliptic Hochschild homology living over the moduli space of elliptic curves; or, in the tmf-variant, over the moduli of cubic curves. These theories interpolate between elliptic Hochschild homologies for different elliptic curves, and also incorporate degenerations to ordinary Hochschild homology and Hodge Hochschild homology, i.e. the graded algebra of global differential forms. The authors in [ MRT22 ] , and Moulinos in [ Mou24 ] , explore the rich geometry of filtrations of \mathbb{G} -Hochschild homology. The key object in their theory is the \mathbb{G} -filtered circle, a stack over [\mathbb{A}^{1}/\mathbb{G}_{m}] , which encodes a filtration on \mathbb{G} -Hochschild homology whose associated graded is Hodge Hochschild homology; in fact in [ MRT22 ] this is only done when \mathbb{G} is equal to the completion of the multiplicative group, whereas the general picture was established by Moulinos in terms of a general theory of filtered formal groups. We argue that when \mathbb{G} is the completion of either \mathbb{G}_{a} , \mathbb{G}_{m} or an elliptic curve E , it should be possible to model the \mathbb{G} -filtered circle in terms of the moduli theory of cubic curves. A precise conjecture is stated in the last section of the paper. Our results We work over an arbitrary abelian commutative ring R . To explain our results, we need to briefly recall the setting of [ MRT22 ] and [ Mou24 ] . The definition of \mathbb{G} -Hochschild homology depends on Cartier duality theory. Over a field this is classical [ Car62 ] , and can be formulated as an anti-equivalence between the category of abelian affine group schemes and abelian formal groups. Over an arbitrary ring R this is more subtle, but we can nonetheless associate to a one-dimensional abelian formal group \mathbb{G} a one-dimensional affine group scheme \mathbb{G}^{\vee} . Let S^{1}_{\mathbb{G}}:=B\mathbb{G}^{\vee} . The authors in [ MRT22 ] define the \mathbb{G} -Hochschild homology of an affine scheme X as \mathrm{HH}_{*}^{\mathbb{G}}(X):=\mathcal{O}(\mathrm{Map}_{\mathrm{dSt}_{R}}(S^{1}_{\mathbb{G}},X)) where \mathrm{Map}_{\mathrm{dSt}_{R}} denotes the mapping stack. If \mathbb{G}=\widehat{\mathbb{G}_{m}} is the completion of the multiplicative group then \mathrm{HH}_{*}^{\mathbb{G}}(X) is equivalent to ordinary Hochschild homology; whereas when \mathbb{G}=\widehat{\mathbb{G}_{a}} is the completion of the additive group, we obtain Hodge Hochschild homology, i.e. \mathrm{HH}_{*}^{\widehat{\mathbb{G}_{a}}}(X)\simeq\mathcal{O}(\mathbb{T}_{X}[-1])\simeq\bigoplus_{i\in\mathbb{N}}\mathbb{L}^{\wedge^{i}}[i] Our first main result is a kind of Fourier–Mukai duality relating a formal group \mathbb{G} and the \mathbb{G} -twisted affine circle S^{1}_{\mathbb{G}} . Theorem A . There is an equivalence of symmetric monoidal \infty -categories \mathrm{F}_{\mathbb{G}}:\mathrm{QCoh}(\mathbb{G})^{\star}\simeq\mathrm{QCoh}(S^{1}_{\mathbb{G}})^{\otimes} where \star denotes the convolution tensor product, and \otimes the ordinary tensor product. As a consequence, there is an equivalence S^{1}_{\mathbb{G}}\simeq\mathrm{coSpec}(\mathrm{Hom}_{\mathrm{QCoh}(\mathbb{G})}(\mathcal{O}_{e},\mathcal{O}_{e})) where \mathrm{Hom}_{\mathrm{QCoh}(\mathbb{G})}(\mathcal{O}_{e},\mathcal{O}_{e}) is the endomorphism algebra of the skyscraper sheaf at the identity e\in\mathbb{G} equipped with its natural enhancement to an \mathbb{E}_{\infty} -algebra. In the statement, the symbol \mathrm{coSpec} denotes the cospectrum of a coconnective commutative algebra; we refer the reader to the main text for more information on this important construction. Now, let E be an elliptic curve over R . Following [ ST23 ] , the elliptic Hochschild homology of an affine scheme X over R is defined as \mathcal{HH}_{E}(X):=\mathcal{O}(\mathrm{Map}_{\mathrm{dSt}_{R}}(E,X)) Theorem A is the main ingredient in the proof of the following comparison result. Theorem B . Let \widehat{E} be the completion of E at the identity. Then there is a natural equivalence \mathrm{HH}_{*}^{\widehat{E}}(X)\simeq\mathcal{HH}_{E}(X) Additionally, let \mathcal{N} and \mathcal{C} be respectively the nodal and cuspidal cubic curve over R . Then there are natural equivalences \mathrm{HH}_{*}(X)\simeq\mathrm{HH}_{*}^{\widehat{\mathbb{G}_{m}}}(X)\simeq\mathcal{O}(\mathrm{Map}_{\mathrm{dSt}_{R}}(\mathcal{N},X)),\quad\mathrm{HH}_{*}^{\widehat{\mathbb{G}_{a}}}(X)\simeq\mathcal{O}(\mathrm{Map}_{\mathrm{dSt}_{R}}(\mathcal{C},X)) Using Theorem B we define global versions of elliptic Hochschild homology which interpolate between different \widehat{E} -Hochschild homologies, as E varies in a moduli stack. More precisely, we describe three variants of this construction corresponding to the three main deformations theories parametrising elliptic curves and their degenerations: namely the moduli stack of elliptic curves \mathcal{M}_{ell} , its modular compactification \overline{\mathcal{M}_{ell}} , and the moduli stack of cubics \mathcal{M}_{cub} . We denote the resulting global Hochschild homology theories by \mathcal{HH}_{*}^{TMF}(X),\quad\mathcal{HH}_{*}^{Tmf}(X),\quad\text{and}\quad\mathcal{HH}_{*}^{tmf}(X) The notations are meant to be suggestive of the close link between these theories and the R -linearization of TMF and its variants, as studied in the homotopy theory literature. Theorem C . The stalks of \mathcal{HH}_{*}^{TMF}(X) , \mathcal{HH}_{*}^{Tmf}(X) and \mathcal{HH}_{*}^{tmf}(X) at a point C of the corresponding moduli stack are equivalent to • \mathrm{HH}_{*}^{\widehat{E}}(X) if C=E is an elliptic curve • \mathrm{HH}_{*}^{\widehat{\mathbb{G}_{m}}}(X) if C=\mathcal{N} is the nodal cubic curve • and to \mathrm{HH}_{*}^{\widehat{\mathbb{G}_{a}}}(X) if C=\mathcal{C} is the cuspidal cubic curve. 1.1. Future work The theory of \mathbb{G} -Hochschild homology as developed in [ MRT22 ] and [ Mou24 ] is meaningful in the

Read the complete text