🔢 mathematics

Perfect generation for regular algebraic stacks

The paper proves that the derived category of quasi-coherent complexes on a regular Noetherian algebraic stack with quasi-finite diagonal is generated by a single perfect complex, utilizing techniques such as gluing generators along recollement and suitable filtrations.

Original authors: Pat Lank

Published 2026-03-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Pat Lank

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 understand a massive, complex city. In mathematics, this "city" is called an Algebraic Stack. It's a place where shapes can overlap, twist, and have hidden layers, making it much harder to study than a simple flat map (which mathematicians call a "scheme").

The mathematicians who study these cities have a powerful tool called a Derived Category. Think of this as a giant library containing every possible "snapshot" or "movie" of the city's geometry. The problem is, this library is so huge that it's impossible to navigate unless you have a Master Key.

In math terms, a "Master Key" is a single object (a specific complex shape) that can be used to build or detect every other object in the library. If you have this one key, you can understand the whole city.

The Big Discovery

The author of this paper, Pat Lank, has found a way to create a single Master Key for a very specific, tricky type of city: a Regular Algebraic Stack.

Here is the breakdown of what he did, using simple analogies:

1. The Problem: The Library is Too Big

For simple cities (schemes), mathematicians have known for a long time that one Master Key exists. But for these complex, overlapping cities (stacks), the rules were messy. Sometimes, the keys you found were too heavy (not "compact") to be useful, or you needed a whole bunch of keys to do the job of one.

Lank wanted to prove that for "Regular" cities (cities that are smooth and well-behaved, like a perfectly paved road), one single, perfect key is enough to unlock the entire library.

2. The Strategy: The "Splitting" Trick

How do you find a key for a whole city? You don't try to open the whole thing at once. You break it down.

Lank used a technique called a "Monomorphic Splitting Sequence."

The Analogy: Imagine the city is a giant onion. You can't eat the whole onion at once. Instead, you peel it layer by layer.
The Math: He showed that any of these complex cities can be peeled apart into a sequence of smaller, simpler layers (like open neighborhoods).
The Process:
1. Start with the innermost layer (the core).
2. Find a Master Key for that core.
3. Move to the next layer. Use a mathematical "glue" (called Recollement) to attach the key from the inner layer to the new outer layer.
4. Repeat until you have covered the whole city.

3. The Glue: Recollement

This is the most creative part of the paper. Recollement is like a sophisticated 3-way zipper.

Imagine you have a jacket (the whole city).
You have a zipper that separates the jacket into a Left Side (an open area), a Right Side (a closed area), and the Seam where they join.
Lank proved that if you have a key for the Left Side and a key for the Right Side, you can "zip them together" to create a new key for the whole jacket.
He used this to stitch together the keys from the different layers of the onion, eventually creating one giant key for the entire stack.

4. The "Perfect" Key

The key he found is called a Perfect Complex.

Analogy: Think of a "Perfect" key as a key made of pure gold. It's finite, manageable, and high-quality.
In the past, people weren't sure if a single "gold" key existed for these complex cities. Lank proved that yes, for smooth, regular cities, you can always find one single gold key that works for everything.

Why Does This Matter?

Before this paper, if you wanted to study a complex, overlapping mathematical city, you might have needed a whole toolbox of keys, or you might have been stuck because the city was too "wild."

Lank's result says: "If the city is regular (smooth), you only need one tool to understand it all."

This simplifies the work for everyone else. It's like discovering that instead of needing a different key for every door in a massive mansion, there is actually one master key that opens every single door, no matter how many rooms or secret passages there are.

Summary in Plain English

The Goal: Prove that one single mathematical object can generate (create/understand) all the complex shapes in a specific type of geometric space.
The Method: Break the space into smaller, manageable layers (peeling an onion).
The Tool: Use a special "gluing" technique (Recollement) to combine the solutions of the small layers into one big solution.
The Result: For smooth, well-behaved geometric spaces, one perfect key is enough to unlock the entire universe of shapes within them.

This is a major step forward because it turns a chaotic, multi-key problem into a simple, one-key solution.

1. Problem Statement and Context

The Core Problem:
The paper addresses the generation of the derived category of quasi-coherent complexes, denoted $D_{qc}(X)$ , on algebraic stacks. Specifically, it investigates whether this category is singly compactly generated.

Background: For quasi-compact quasi-separated (qcqs) schemes, $D_{qc}(X)$ is known to be singly compactly generated (generated by a single perfect complex).
The Gap: For general algebraic stacks, perfect complexes are not necessarily compact objects. While recent work by Hall and Rydh established compact generation for specific classes of stacks (e.g., Deligne–Mumford stacks with quasi-finite separated diagonals), the question of single object generation (by a single perfect complex, which may or may not be compact) remained open for broader classes of regular stacks.
Specific Challenge: Previous results often required strong conditions like separated diagonals, finite Krull dimension, or the stack being "concentrated" (where compact objects coincide with perfect complexes). Lank aims to remove these restrictions for regular Noetherian algebraic stacks.

The Main Question:
Does there exist a single perfect complex $P \in \text{Perf}(X)$ that generates $D_{qc}(X)$ for a quasi-compact quasi-separated regular algebraic stack $X$ with a quasi-finite diagonal?

2. Methodology and Technical Framework

The proof strategy relies on a combination of inductive arguments on Nisnevich coverings, recollement techniques, and duality arguments.

A. Recollements and Gluing Generators

The author utilizes the theory of recollements for triangulated categories. Given a qcqs algebraic stack $X$ and a quasi-compact open immersion $j: U \hookrightarrow X$ with closed complement $Z$ , there exists a recollement:
$D_{qc}(U) \rightleftarrows D_{qc}(X) \rightleftarrows D_{qc, Z}(X)$

Key Lemma (Proposition 3.7): If a subcategory $\mathcal{G}$ generates the "open" part ( $D_{qc}(U)$ ) and a subcategory $\mathcal{G}'$ generates the "closed" part ( $D_{qc, Z}(X)$ ), then the union of their images generates the whole category $D_{qc}(X)$ . This allows the author to "glue" generators from local pieces.

B. Nisnevich Coverings and Splitting Sequences

To handle the global structure of the stack, the paper employs the presentation results of Hall and Rydh [HR18].

For a regular stack $X$ with quasi-finite diagonal, there exists a Nisnevich covering $f: Y \to X$ and a monomorphic splitting sequence of open immersions:
$\emptyset = X_0 \hookrightarrow X_1 \hookrightarrow \dots \hookrightarrow X_n = X$
This sequence allows for an inductive proof on the length $n$ . The base cases involve locally closed substacks where the covering admits a section.

C. Finite Duality and Support Conditions

The proof leverages the regularity of the stack to utilize finite duality.

Lemma 4.1 & 4.6: If $f: Y \to X$ is a finite morphism and $D_{qc}(Y)$ is generated by a set $\mathcal{B}$ , then $D_{qc, f(|Y|)}(X)$ is generated by the pushforward of $\mathcal{B}$ , provided the dualizing object behaves well.
Regularity Hypothesis: For regular Noetherian stacks, $\text{Perf}(X) \cong D^b_{\text{coh}}(X)$ . This equivalence is crucial for ensuring that the generators constructed are perfect complexes.

D. Thomason Conditions

The paper introduces and utilizes the $\beta$ -Thomason condition, which requires that for every closed subset $Z$ with quasi-compact complement, there exists a perfect complex supported exactly on $Z$ . The author proves that regular stacks with quasi-finite diagonals satisfy the 1-Thomason condition (generated by a single object).

3. Key Contributions and Results

Main Theorem (Theorem 1.1)

Statement: Let $X$ be a quasi-compact quasi-separated regular algebraic stack with quasi-finite diagonal. Then there exists a single perfect complex $P \in \text{Perf}(X)$ that generates $D_{qc}(X)$ .

Significance of Conditions:

Regularity: Essential for the duality arguments and the identification of perfect complexes with bounded coherent complexes.
Quasi-finite Diagonal: Replaces the stronger "separated diagonal" requirement found in previous literature.
No "Concentrated" Assumption: The result holds even if the stack is not concentrated (i.e., compact objects may not equal perfect complexes), though the generator itself is perfect.

Corollary 1.2 (Singly Compactly Generated)

Statement: If $X$ is a concentrated regular algebraic stack with quasi-finite diagonal, then $D_{qc}(X)$ is singly compactly generated.

This improves upon previous results (e.g., [Hal22]) by removing the requirement for finite Krull dimension and extending the result to a broader class of stacks (including those with affine stabilizers over DVRs).

Technical Lemmas

Proposition 4.3: Establishes that the $\beta$ -Thomason condition is preserved under finite, flat, surjective morphisms.
Lemma 4.4: Shows that the $\beta$ -Thomason condition is preserved under quasi-affine monomorphisms (injective on topological spaces).
Recollement Gluing: The paper formalizes how to construct generators for the whole stack by combining generators of open and closed substacks (Proposition 3.7).

4. Significance and Impact

Generalization of Known Results: The paper significantly extends the scope of single object generation from schemes and specific DM stacks to all regular Noetherian algebraic stacks with quasi-finite diagonals.
New Cases: It covers cases previously unknown, such as:
- Deligne–Mumford stacks in arbitrary characteristic (e.g., smooth stacks over fields or DVRs).
- Stacks that are not necessarily separated or concentrated.
Methodological Innovation: The use of recollement to glue generators along a Nisnevich splitting sequence provides a robust framework for handling the local-to-global transition in derived categories of stacks. This technique is highlighted as being of independent interest.
Distinction between Generation and Compact Generation: The author clarifies that while a single perfect complex generates the category, proving it is compact (which would imply single compact generation) remains an open problem in the non-concentrated case (Remark 4.5). However, for concentrated stacks, the result is fully established.

Conclusion

Pat Lank's paper resolves the question of single object generation for a wide class of regular algebraic stacks. By combining the structural properties of Nisnevich coverings with the powerful machinery of recollements and duality, the author demonstrates that the derived category of quasi-coherent complexes on these stacks is generated by a single perfect complex, thereby unifying and extending the theory of derived categories in algebraic geometry.