🔢 mathematics

A Semi-Orthogonal Decomposition Theorem for Weighted Blowups

This paper generalizes Orlov's classic result by establishing a semi-orthogonal decomposition for the weighted blowup of an algebraic stack along a Koszul-regular weighted centre, utilizing the framework developed by Bergh and Schnürer.

Original authors: Oliver Li

Published 2026-05-04

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Oliver Li

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

The Big Picture: Reshaping a Shape

Imagine you have a complex, multi-dimensional shape (an "algebraic stack"). Sometimes, this shape has a rough spot, a singularity, or a specific feature you want to fix or study more closely. In mathematics, a common way to fix or examine a rough spot is to perform a blowup.

Think of a standard blowup like taking a piece of clay with a dent in it and inflating that dent into a smooth, round bubble. You replace the rough point with a whole new surface (a "divisor") so you can work with it more easily.

Weighted Blowups are a more sophisticated version of this. Instead of inflating the dent evenly, you inflate it with different "weights" or speeds in different directions. It's like blowing up a balloon that is stretched more in one direction than another, or inflating a balloon inside a box where the walls push back with different strengths. This is a powerful tool used to solve difficult problems in geometry, such as smoothing out singularities in shapes.

The Main Question: What Happens to the "Library"?

In modern mathematics, we don't just look at the shape itself; we look at its "library" of data. This library is called a derived category. Think of this library as a massive collection of all the possible patterns, waves, and structures that can exist on your shape.

When you perform a blowup (changing the shape), you naturally wonder: "How does the library change?"

If you take a shape $X$ and blow it up to get a new shape $\tilde{X}$ , the library for $\tilde{X}$ is bigger and more complex. The paper asks: Can we break this new, complex library down into smaller, manageable pieces that we already understand?

The Discovery: A "Semi-Orthogonal Decomposition"

The author, Oliver Li, proves that the answer is yes. He shows that the library of the new shape ( $\tilde{X}$ ) can be split into a stack of distinct, non-overlapping layers.

He calls this a Semi-Orthogonal Decomposition (SOD).

The Analogy: Imagine the library of the new shape is a giant, messy bookshelf. Li proves that you can organize this bookshelf into a neat stack of smaller bookcases.
The Layers:
1. The Original Library: One of the layers is just the library of the original shape ( $X$ ), pulled into the new space.
2. The "New" Layers: The other layers come from the "bubble" you created (the exceptional divisor). These layers are copies of the library of the center of the blowup, but they are "twisted" or shifted in specific ways (like rotating a book on a shelf).

The theorem says: The new library is exactly equal to the original library plus a specific number of twisted copies of the center's library.

Why Was This Hard? (The "Weighted" Problem)

Mathematicians already knew how to do this for:

Standard Blowups: Where the inflation is even.
Root Stacks: A specific type of weighted blowup used in other contexts.

However, Weighted Blowups are a general mix of these. The problem was that the old mathematical "blueprints" (arguments) used to prove this for standard blowups relied on the shape fitting into a specific, simple grid (a projective space).

In the weighted case, the shape doesn't fit neatly into that simple grid. It's like trying to use a square peg in a round hole, or trying to fit a complex, twisted knot into a straight line. The old proofs broke down because they couldn't handle the "weights" (the different speeds of inflation).

The Solution: The "Extended Rees Algebra"

Li's clever trick was to stop trying to force the weighted blowup into the old grid. Instead, he looked at the blowup through a different lens: The Extended Rees Algebra.

The Metaphor: Imagine you are trying to describe a complex machine. The old way was to take it apart and look at the gears one by one. Li's way was to look at the instruction manual (the algebra) that built the machine.
By viewing the weighted blowup as a construction built from this specific algebra, he could use explicit, step-by-step calculations (Koszul complexes) to show exactly how the library pieces fit together. It was like realizing that if you follow the instruction manual correctly, the messy knot actually untangles into a predictable pattern.

The Second Result: The "Mirror" (Dualizing Sheaf)

The paper also calculates something called the relative dualizing sheaf.

The Analogy: In geometry, every shape has a "mirror" or a "shadow" that tells you how light and volume behave on it. When you change the shape (via a blowup), you need to know how this mirror changes.
Li provides a precise formula for this new mirror. He shows that the mirror of the weighted blowup is just the original mirror, but stretched by a specific amount determined by the "weights" of the blowup. This generalizes a famous result from standard blowups to this more complex, weighted world.

Summary

The Goal: Understand how the mathematical "library" of a shape changes when you perform a complex, weighted inflation (blowup).
The Result: The new library is a perfect stack of the old library plus several twisted copies of the center's library.
The Innovation: Previous methods failed because the shape didn't fit standard grids. The author solved this by using a specific algebraic "instruction manual" (Extended Rees Algebra) to untangle the complexity.
The Bonus: He also figured out exactly how the shape's "mirror" (dualizing sheaf) changes during this process.

This work unifies several known mathematical results into one powerful theorem, allowing mathematicians to handle these complex "weighted" shapes with the same confidence they have for standard shapes.

1. Problem Statement

The paper addresses the behavior of derived categories of coherent sheaves under weighted blowups of algebraic stacks. While the semi-orthogonal decomposition (SOD) of the derived category is a classical result for standard blowups (Orlov's Theorem) and root stacks (Ishii-Ueda, Bergh-Lunts-Schnürer), a general theorem for weighted blowups along Koszul-regular weighted centers was missing.

The core difficulty lies in the fact that standard proofs for SODs rely on the blowup embedding into a projective bundle $\mathbb{P}^{c-1}_X$ (where $c$ is the codimension). In the weighted case, the exceptional divisor is a weighted projective stack bundle, not a standard projective bundle, and the embedding into a standard projective space does not hold. Consequently, the arguments used by Bergh and Schnürer for standard blowups and root stacks cannot be directly applied.

2. Methodology

Li develops a new approach that bypasses the need for a global embedding into a standard projective space. The methodology relies on three main pillars:

Extended Rees Algebras: The author defines the weighted blowup $\tilde{X}$ $\tilde{X}$ as the stacky Proj of an extended Rees algebra $\mathcal{A}$ $A$ . Unlike the standard Rees algebra (which is non-negatively graded), the extended Rees algebra includes negative degrees, allowing for a precise description of the deformation to the weighted normal cone.
- Locally, for a center defined by $(f_i, d_i)$ , the algebra is isomorphic to $R[t^{-1}, x_0, \dots, x_n] / (f_i - x_i t^{-d_i})$ .
Conservative Descent: To handle the global case on algebraic stacks, the paper utilizes the theory of Fourier-Mukai transforms and conservative descent (developed by Bergh and Schnürer). This allows the reduction of the problem to the affine case. The paper verifies that the relevant functors form Fourier-Mukai transforms that satisfy the necessary base-change properties.
Explicit Cohomological Computations: The proof involves detailed computations of cohomology groups for weighted projective stacks and the weighted blowup itself.
- The author proves that the pullback functor $L\pi^*$ is fully faithful by showing $R\pi_* \mathcal{O}_{\tilde{X}} \cong \mathcal{O}_X$ .
- The author establishes that the derived category of the weighted blowup is generated by a specific set of objects derived from the exceptional divisor and the base.

3. Key Contributions and Results

A. Main Theorem: Semi-Orthogonal Decomposition (Theorem 1.1)

The paper establishes a semi-orthogonal decomposition for the derived category $D(\tilde{X})$ of a weighted blowup $\pi: \tilde{X} \to X$ along a Koszul-regular weighted center $I_\bullet = \sum (f_i, d_i)$ .

Let $|d| = \sum d_i$ . The decomposition is:
$D(\tilde{X}) = \langle \text{im } \Phi_1, \dots, \text{im } \Phi_{-1}, \text{im } \Phi_0 \rangle$
where:

$\Phi_0 = L\pi^*: D(X) \to D(\tilde{X})$ .
For $r < 0$ $r < 0$ , $\Phi_r: D(Z_1) \to D(\tilde{X})$ $Φ_{r} : D (Z_{1}) \to D (\tilde{X})$ is defined by $F \mapsto j_*(p^*(F) \otimes \mathcal{O}_E(r))$ $F \mapsto j_{*} (p^{*} (F) \otimes O_{E} (r))$ .
- Here, $Z_1 = V(I_1)$ is the center, $E$ is the exceptional divisor, $j: E \hookrightarrow \tilde{X}$ is the inclusion, and $p: E \to Z_1$ is the projection.
The decomposition has $|d|$ summands.
This holds for $D = D_{qc}$ (quasi-coherent), $Perf$ (perfect complexes), and, if $X$ is locally Noetherian, for $D^-_{Coh}$ and $D^b_{Coh}$ .

B. Relative Dualizing Sheaf (Theorem 1.3)

The paper computes the right adjoint $\pi^!$ to the pushforward functor $R\pi_*$ (the relative dualizing complex).

Result: $\pi^! F \cong L\pi^* F \otimes \mathcal{O}_{\tilde{X}}((\sum d_i - 1)E)$ .
This generalizes the classical formula for standard blowups (where all $d_i=1$ , yielding $\mathcal{O}((c-1)E)$ ).

C. Affine Case and Generation

A significant portion of the technical work is proving that the derived category of the weighted blowup (in the affine case) is generated by the specific objects appearing in the SOD.

The author proves that the object $\bigoplus_{r=1-|d|}^0 \mathcal{O}_{\tilde{X}}(r)$ generates $D_{qc}(\tilde{X})$ (Lemma 4.2).
This generation result is crucial for proving the "fullness" of the semi-orthogonal sequence using Lemma 2.22.

4. Significance and Implications

Unification of Results: The theorem unifies previous results for standard blowups (Orlov) and root stacks (Ishii-Ueda, Bergh-Schnürer) into a single framework for weighted blowups.
Algorithmic Applications: Weighted blowups are critical tools in modern resolution of singularities algorithms (e.g., by Abramovich, Temkin, and Włodarczyk). Understanding the derived category structure is essential for studying invariants (like the integral Chow ring or K-theory) under these transformations.
Overcoming Geometric Obstacles: The paper demonstrates that the lack of a standard projective embedding does not prevent the existence of an SOD. By utilizing the extended Rees algebra and explicit Koszul complex computations, the author provides a robust method for handling "stacky" geometry where standard vector bundle intuition fails.
Dualizing Sheaf: The explicit calculation of the relative dualizing sheaf is a valuable byproduct, facilitating the study of Grothendieck duality in the context of weighted blowups.

5. Conclusion

Oliver Li's paper successfully generalizes Orlov's semi-orthogonal decomposition theorem to the setting of weighted blowups on algebraic stacks. By leveraging the extended Rees algebra formalism and conservative descent, the author overcomes the geometric limitations of the weighted case, providing a complete description of the derived category structure and the associated duality functors. This work bridges the gap between classical blowup theory and the increasingly important theory of root stacks and weighted resolutions.