🔢 mathematics

Tensor triangular geometry -- Notes for an Oberwolfach Seminar

These notes from an October 2025 Oberwolfach Seminar develop notions of support for triangulated categories and apply them to classify thick and localising tensor ideals arising in the modular representation theory of finite groups.

Original authors: Henning Krause

Published 2026-05-20

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Henning Krause

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to organize a massive, chaotic library. This library doesn't just contain books; it contains entire universes of mathematical objects called "triangulated categories." These objects are complex, interconnected, and often impossible to see clearly all at once.

The goal of Henning Krause's notes is to provide a map and a filing system for this library. The title, "Tensor Triangular Geometry," sounds intimidating, but the core idea is about finding the "address" of every object in this library so we can sort them into neat, logical piles.

Here is the breakdown of the paper's journey, using everyday analogies:

1. The Problem: A Messy Lattice

Imagine the library has a rule: if you have a specific book, you must also include all its sequels, prequels, and any book that can be built from it. Mathematicians call this a "thick subcategory."

The paper starts by looking at how these groups of books relate to each other. Do they overlap? Can you combine two groups to make a bigger one? The author shows that these groups form a lattice (a structured grid). Sometimes this grid is messy and doesn't follow simple rules, but if the groups "commute" (meaning they play nicely together without getting tangled), the grid becomes a perfect, orderly distributive lattice. Think of it like sorting LEGO bricks: if you sort by color and then by size, you get a clean system. If the sorting rules clash, you get a mess.

2. The Solution: Giving Every Object an Address (Support)

How do we know which pile a specific object belongs to? The paper introduces the concept of Support.

Imagine every object in the library has a hidden "GPS coordinate."

The Map: This coordinate is a point on a geometric space (like a map of a city).
The Action: The library is "acted upon" by a ring (a set of mathematical rules, like a coordinate system).
The Address: The "support" of an object is simply the set of points on the map where that object "exists" or is "alive." If you zoom in on a specific point on the map, you can see exactly which objects are there.

The paper proves that if you know the address (support) of an object, you can often figure out exactly which pile (subcategory) it belongs to. It's like saying, "If I know a book is in the 'Science Fiction' section, I know it's not in 'History'."

3. The Big Breakthrough: Stratification

The ultimate goal is Stratification. This is the paper's main "Aha!" moment.

Imagine the library is a multi-layered cake.

Layer 1: The bottom layer contains the most basic, fundamental objects.
Layer 2: The next layer contains objects built from the first.
Layer 3: And so on.

Stratification means the library is perfectly layered. The paper claims that if the library is "stratified," you can classify every single possible pile of books just by looking at the map. There is a perfect one-to-one match between the "subsets of the map" and the "piles of books."

If the library is stratified, you don't need to look at the books individually. You just look at their addresses on the map, and the map tells you exactly how they are organized.

4. The Real-World Example: Finite Groups

The paper isn't just abstract theory; it applies this to Modular Representations of Finite Groups.

The Group: Think of a finite group as a specific set of rules for shuffling a deck of cards (or symmetries of a shape).
The Cohomology Ring: This is a mathematical "fingerprint" of the group. It's a ring of numbers that captures the group's hidden structure.
The Connection: The paper shows that the "map" for the library of this group's representations is actually the spectrum of its cohomology ring.

In simple terms: To understand the complex behavior of a group's representations (the books), you just need to look at the geometry of its cohomology ring (the map). The paper proves that the "support" of a representation is determined by where it lives on this map.

5. The "Subgroup" Shortcut

How do we handle huge, complicated groups? The paper uses a clever trick involving Elementary Abelian Subgroups.

Think of a giant, complex machine. Instead of trying to understand the whole machine at once, the paper says: "Break it down into its smallest, simplest gears."

It turns out that the behavior of the whole group is determined by its "elementary abelian" subgroups (the simplest gears).
By understanding how the library is organized for these simple subgroups, we can reconstruct the organization for the entire group.
This relies on Quillen's Stratification, a famous result that says the "fingerprint" of a group is built from the fingerprints of its simplest subgroups.

Summary

The paper is a guidebook for organizing a chaotic mathematical universe.

Identify the rules: It defines how groups of objects relate (Lattices).
Create a map: It assigns a "support" (an address) to every object based on a ring action.
Prove the system works: It shows that if the library is "stratified," the map perfectly predicts every possible organization of objects.
Apply it: It proves that for finite groups, this stratification works perfectly. The complex world of group representations is completely classified by the geometry of their cohomology rings.

The Bottom Line: The paper tells us that even in the most complex mathematical structures, there is an underlying geometric order. If you know the "address" (support) of an object, you know its place in the universe.

Technical Summary: Tensor Triangular Geometry Notes for an Oberwolfach Seminar

Problem and Scope
The paper addresses the classification of thick and localising tensor ideals within triangulated categories arising in modular representation theory, specifically those associated with finite groups. The central problem is to establish a "stratification" of these categories via an appropriate topological space, thereby reducing the global classification of subcategories to a local problem. The author, Henning Krause, develops a unified framework for support theory that applies to two primary settings: triangulated categories equipped with a monoidal (tensor) structure, and those admitting an action by a graded-commutative ring. The ultimate goal is to classify these subcategories by computing their support, with a specific focus on the derived categories of modular representations of finite groups.

Methodology
The methodology relies on the interplay between lattice theory, local cohomology, and the geometry of prime ideal spectra. The approach is structured as follows:

Lattice of Thick Subcategories: The author first establishes the lattice structure of thick subcategories. A key concept introduced is "commutativity" between pairs of thick subcategories, defined via the factorization of morphisms or the isomorphism of localisation functors ( $\Gamma_U \Gamma_V \cong \Gamma_{U \cap V}$ ). It is shown that central subcategories (those commuting with all others) form a distributive sublattice, and under certain conditions, a frame.
Support via Central Actions: The paper introduces cohomological support for objects in a triangulated category $\mathcal{T}$ acted upon by a graded-commutative ring $R$ . Support is defined as the set of prime ideals $\mathfrak{p} \in \text{Spec } R$ where the localisation $x_{\mathfrak{p}}$ is non-zero. The author utilizes Koszul objects ( $x//r$ ) to construct central thick subcategories $T_{V(\mathfrak{a})}$ associated with ideals $\mathfrak{a} \subseteq R$ .
Local Cohomology and Fibred Categories: For compactly generated triangulated categories, the theory is extended using local cohomology functors $\Gamma_Y$ supported on locally closed subsets $Y$ of a space $X$ . The category is said to be "fibred" over $X$ if there is a lattice morphism from the open subsets of $X$ to the thick subcategories of compact objects. This allows for the definition of a support for arbitrary objects via the non-vanishing of local cohomology functors $\Gamma_{\mathfrak{p}}(x)$ .
Stratification: A category is defined as "stratified" if it satisfies a local-global principle (every localising subcategory is the join of its local components) and a minimality condition (there are no proper non-zero localising subcategories in the local components). Under stratification, the map from localising subcategories to subsets of the spectrum becomes a lattice isomorphism.
Comparison of Supports: The paper compares the cohomological support (via ring action) with the tensor triangular support (via the Balmer spectrum). A comparison map $\rho_T: \text{Spc } \mathcal{T}^c \to \text{Spec } R$ is constructed, and conditions are established under which this map is a homeomorphism, linking the two support theories.

Key Contributions and Results

General Stratification Theory: The paper provides a general criterion (Proposition 4.15, 4.18) for a rigidly-compactly generated tensor triangulated category to be stratified. It proves that stratification is equivalent to the support map inducing a lattice isomorphism between localising tensor ideals and subsets of the spectrum.
Differential Graded (DG) Algebras: The theory is applied to DG algebras. Theorem 5.1 establishes that if $A$ is a formal DG algebra with noetherian cohomology $H^*(A)$ , then $D(A)$ is stratified via the canonical $H^*(A)$ -action.
Elementary Abelian $p$ -Groups: Using a DG version of the Bernstein–Gelfand–Gelfand (BGG) correspondence, the paper classifies thick tensor ideals for the group algebra of an elementary abelian $p$ -group. It shows that cohomological support induces a lattice isomorphism between thick tensor ideals and Thomason subsets of the cohomology ring spectrum (Theorem 5.9).
Finite Groups (Thick Ideals): By reducing the problem to elementary abelian subgroups via Quillen's stratification of group cohomology, the paper classifies thick tensor ideals in the bounded derived category of finitely generated modules for an arbitrary finite group $G$ (Theorem 5.14). The classification is given by the cohomological support over $H^*(G, k)$ .
Finite Groups (Localising Ideals): The results are extended to the compactly generated category $K(\text{Inj } kG)$ (the homotopy category of complexes of injective modules). Theorem 6.17 establishes that $K(\text{Inj } kG)$ is stratified via the action of $H^*(G, k)$ , yielding a complete classification of localising tensor ideals by subsets of $\text{Spec } H^*(G, k)$ .
Stable Module Categories: As a consequence, the stable module category $\text{StMod } kG$ is shown to be stratified via the projective spectrum $\text{Proj } H^*(G, k)$ (Remark 6.18).

Significance and Claims
The paper claims to provide a comprehensive framework for understanding the structure of triangulated categories in modular representation theory. Its primary significance lies in the "categorification" of Quillen's stratification of group cohomology. By establishing that the derived category of modular representations is stratified, the paper demonstrates that the classification of thick and localising tensor ideals is entirely determined by the cohomology ring of the group.

The author notes that these notes are not exhaustive but are designed to be self-contained, developing the necessary concepts (such as the lattice of thick subcategories, central actions, and local cohomology) specifically to facilitate applications in modular representation theory. The work builds upon foundational results by Quillen, Benson, Carlson, Rickard, Neeman, and Balmer, synthesizing them into a coherent theory of support and stratification. The results confirm that for finite groups, the "spectrum" governing the tensor triangular geometry is precisely the spectrum of the group cohomology ring.