🔢 mathematics

Perfectly generated $t$ -structures for algebraic stacks

This paper establishes that the standard $t$ -structure on the derived category of quasi-coherent sheaves over suitable algebraic stacks is compactly generated, thereby enabling the classification of compactly generated tensor $t$ -structures via Thomason filtrations.

Original authors: Michal Hrbek, Pat Lank, Simone Pizzirani

Published 2026-07-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Michal Hrbek, Pat Lank, Simone Pizzirani

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 the universe of algebraic geometry as a massive, chaotic library. Inside this library, the books aren't made of paper; they are "complexes"—intricate, multi-layered structures that hold deep information about shapes and spaces. Mathematicians have long tried to organize this library using a system called a t-structure. Think of a t-structure as a set of magical shelves that sort these complex books into "past" (negative degrees) and "future" (positive degrees), allowing researchers to pull out specific information without getting lost in the noise.

For a long time, mathematicians knew how to build these shelves for simple, tidy libraries (like standard geometric shapes called schemes). But when they tried to organize the more wild, twisted libraries known as algebraic stacks, the shelves kept collapsing. It was a mystery: could you even build a stable, compact shelf system for these messy stacks?

The Big Discovery: Building the Shelves

In this paper, the authors Michal Hrbek, Pat Lank, and Simone Pizzirani prove that yes, you absolutely can build these shelves for a huge class of these wild libraries.

Their main finding is that for "concentrated" algebraic stacks (a specific type of well-behaved but complex stack), the standard way of sorting these books is compactly generated.

To understand what "compactly generated" means, imagine trying to describe a giant, infinite wall of bricks. If you can describe the entire wall just by pointing to a finite, manageable collection of "master bricks" and saying, "Everything else is built from these," then the wall is compactly generated. The authors show that for these algebraic stacks, you don't need an infinite list of rules to sort the books; you only need a specific, finite collection of "perfect" building blocks (called perfect complexes) to generate the entire sorting system.

What They Explicitly Rule Out

Before this paper, there was a nagging doubt. Some earlier research suggested that for many algebraic stacks, the standard sorting system might not be compactly generated. In fact, for some specific types of stacks, it was known that the "master bricks" approach failed.

The authors do not claim that every possible algebraic stack in existence has this property. They specifically focus on stacks with a "quasi-finite and separated diagonal" (a technical way of saying the stack doesn't have too much chaotic overlap) or "Deligne–Mumford Q-stacks." They prove that within this specific, well-behaved group, the shelves hold up. They do not claim to have solved the problem for every single weird stack imaginable, but they have cleared the fog for a massive, important chunk of them.

The "Pseudoapproximation" Trick

How did they build these shelves? They invented a new tool they call pseudoapproximation.

Imagine you have a broken, jagged rock (a complex mathematical object) and you want to smooth it out using a set of perfect, polished marbles (perfect complexes). Usually, you can't just swap the rock for a marble; the rock is too weird. But the authors found a way to "pseudo-approximate" the rock. They showed that even if you can't replace the whole rock with a marble, you can at least find a marble that matches the rock's most important feature (its "highest nonvanishing cohomology sheaf") perfectly.

They used a technique called étale dévissage, which is like taking a complex, knotted rope and carefully untying it into smaller, simpler loops that you can handle one by one. By gluing these smaller loops together along "étale neighborhoods" (which are like zooming in on a map to see the details), they proved that the "master bricks" (perfect complexes) are enough to build the whole shelf system.

The Grand Classification: The Thomason Filtration

Once they proved the shelves were stable, they did something even cooler: they created a one-to-one map between these shelf systems and something called Thomason filtrations.

Think of a Thomason filtration as a "shadow map" of the library. It's a rule that assigns a specific "shadow" (a closed subset of the library's floor plan) to every integer number.

If you have a specific way of sorting the books (a $\otimes$ -aisle), the authors show you can draw a unique shadow map for it.
Conversely, if you draw a valid shadow map (a Thomason filtration), you can build a unique sorting system for it.

This is a huge deal because, until now, no one had a complete map for these wild stacks. The authors proved that for concentrated stacks, the relationship is perfect: One sorting system = One shadow map.

How Sure Are They?

The authors are proven certain. This isn't a guess, a simulation, or a "maybe." They have provided rigorous mathematical proofs (Theorem 1.3, Proposition 1.1, and a series of lemmas) that demonstrate this relationship holds true. They didn't just suggest it; they constructed the logic step-by-step, showing that if you have a concentrated stack, the shelves must be compactly generated, and the map to the shadow filters must exist.

The Takeaway

In short, Hrbek, Lank, and Pizzirani took a chaotic, messy library of algebraic stacks and proved that, for a large and important section of it, the books can be neatly organized using a finite set of perfect tools. They didn't just organize the books; they drew a complete, foolproof map showing exactly how every possible way of organizing those books corresponds to a specific pattern on the library floor. It's a solid, proven foundation that turns a mystery into a manageable, organized system.

Technical Summary: Perfectly Generated t-Structures for Algebraic Stacks

Problem Statement
The paper addresses the structural understanding of $t$ -structures within the derived category of quasi-coherent sheaves, $D_{qc}(X)$ , on algebraic stacks. While $t$ -structures are fundamental tools in algebraic geometry, a complete classification is generally impossible even for affine schemes. Consequently, research has focused on specific classes, such as compactly generated $t$ -structures.

A central open question concerns the "standard aisle" $D^{\leq 0}_{qc}(X)$ (the subcategory of complexes with cohomology vanishing in positive degrees). For Noetherian schemes, it is known that this aisle is compactly generated by perfect complexes. However, for general algebraic stacks, it was recently shown that the standard aisle need not be singly compactly generated. The authors investigate whether the standard aisle is compactly generated (i.e., generated by a set of compact objects, not necessarily a single one) for a broad class of algebraic stacks, and whether this property enables a classification of tensor-compatible $t$ -structures analogous to the affine case.

Methodology
The authors employ a combination of descent techniques, approximation theory, and the theory of Thomason filtrations.

Pseudoapproximation: A core technical innovation is the introduction of "pseudoapproximation by compact complexes." This notion allows for the approximation of bounded-above complexes (with finitely presented highest non-vanishing cohomology) by compact objects. This condition is strictly weaker than approximation by perfect complexes or by compact objects generally.
Étale Dévissage: The proof of the generation result proceeds via étale dévissage, following the framework established in [HR18]. The authors utilize flat Mayer–Vietoris squares and étale neighborhoods to glue local approximation properties to the global stack.
Reduction to Affine Schemes: For the classification results, the authors leverage the existence of smooth covers by affine schemes. They establish that the classification of tensor aisles on the stack can be "bootstrapped" from the known classification on the affine cover, provided the stack satisfies specific concentration and diagonal conditions.
Support Theory: The classification relies on Thomason filtrations (decreasing functions from $\mathbb{Z}$ to Thomason subsets of the topological space). The authors prove that on the relevant stacks, every such filtration is "supported by perfects," meaning it can be realized by the supports of cohomology sheaves of perfect complexes.

Key Contributions and Results

Generation of the Standard Aisle (Proposition 1.1 & Corollary 1.2):
The primary result establishes that for a concentrated algebraic stack $X$ with a quasi-finite and separated diagonal (or a Deligne–Mumford Q-stack), the standard aisle $D^{\leq 0}_{qc}(X)$ is compactly generated.
- Notably, this result is new even for quasi-compact quasi-separated schemes, where it was previously unknown whether the standard aisle is compactly generated.
- The proof relies on showing that such stacks satisfy the pseudoapproximation property, which implies compact generation via Lemma 3.3.
Classification of Tensor t-Structures (Theorem 1.3):
Building on the generation result, the authors provide a classification of compactly generated $\otimes$ -aisles (tensor-compatible aisles) on $D_{qc}(X)$ .
- There is a one-to-one correspondence between the set of compactly generated $\otimes$ -aisles on $D_{qc}(X)$ and the set of Thomason filtrations on the underlying topological space $|X|$ .
- The correspondence is explicit: an aisle generated by a set of perfect complexes $P \subseteq \text{Perf}(X)$ corresponds to the filtration $n \mapsto \bigcup_{j \geq n} \bigcup_{P \in P} \text{supp}(H^j(P))$ .
- This classification extends the affine results of [Hrb20] directly to algebraic stacks without relying on the intermediate results of [DS23] (which required Noetherian hypotheses).

Significance and Claims
The paper claims to resolve the question of whether the standard aisle is compactly generated for a wide class of algebraic stacks, a property that was previously ambiguous even for schemes. By establishing this generation property, the authors are able to extend the classification of tensor $t$ -structures from affine schemes to algebraic stacks.

The significance lies in the unification of the affine and stacky cases:

It demonstrates that the "standard" $t$ -structure behaves well (is compactly generated) under relatively mild geometric conditions (concentrated with quasi-finite separated diagonal).
It provides the first classification of compactly generated tensor $t$ -structures for algebraic stacks, filling a gap in the literature where no such classification was available.
The methodology of using pseudoapproximation and étale dévissage offers a robust framework for transferring properties from affine covers to stacks, bypassing the need for stronger Noetherian assumptions in certain contexts.

The authors do not claim to solve the classification for all $t$ -structures (which is impossible in full generality) nor do they claim to remove all geometric constraints (e.g., the concentrated and diagonal conditions are necessary for their proofs). The results are presented as a definitive step forward for the specific class of "suitable" algebraic stacks defined in the text.