Torsion models for tensor-triangulated categories: the one-step case
Scott Balchin, J. P. C. Greenlees, Luca Pol, Jordan Williamson
TL;DR
This work develops a comprehensive framework to reconstruct tensor-triangulated categories from local torsion data via an adelic torsion model. Central to the approach is a one-step Tate-type diagram that expresses the unit as a homotopy pullback of localized and completed pieces, yielding a Quillen equivalence between the original model category and a cellularization of a diagram of module categories built from $L ext{1}$, $oldsymbol{ extLambda} ext{1}$, and their smash products. The paper introduces the pre-torsion model and its extension to the full torsion model, showing how adelic data enforces algebraicity and enabling abelian/differential models in special cases such as rational $ ext{T}$-spectra. Along the way, it develops a robust theory of localization, completion, and support in finite-dimensional Noetherian Balmer spectra, with concrete applications to derived algebra, chromatic homotopy, and rational equivariant spectra. The results culminate in a Quillen equivalence between rational $ ext{T}$-spectra and a differential model, establishing a concrete, algebraic foothold for the torsion-based reconstruction program in tensor-triangulated settings.
Abstract
Given a suitable stable monoidal model category $\mathscr{C}$ and a specialization closed subset $V$ of its Balmer spectrum one can produce a Tate square for decomposing objects into the part supported over $V$ and the part supported over $V^c$ spliced with the Tate object. Using this one can show that $\mathscr{C}$ is Quillen equivalent to a model built from the data of local torsion objects, and the splicing data lies in a rather rich category. As an application, we promote the torsion model for the homotopy category of rational circle-equivariant spectra from [18] to a Quillen equivalence. In addition, a close analysis of the one step case highlights important features needed for general torsion models which we will return to in future work.
