The refined local Donaldson-Thomas theory of curves

This paper solves the $K$-theoretically refined Donaldson-Thomas theory of local curves by employing direct localization methods to express partition functions via universal series and skew nested Hilbert schemes, thereby establishing the Aganagic-Schaeffer formula, proving the Nekrasov-Okounkov DT/PT correspondence in arbitrary genus, and providing key results for the refined GW/PT correspondence.

The Gist

Imagine you are a cosmic accountant trying to count the number of ways you can build specific structures out of "quantum Lego bricks" inside a complex, multi-dimensional universe. This is essentially what Donaldson-Thomas (DT) theory does: it counts stable "sheaves" (which are like intricate, invisible scaffolding structures) on complex geometric shapes called Calabi-Yau threefolds.

For a long time, mathematicians could only do this counting for very simple, "flat" shapes or by using a method called "degeneration," which is like taking a complex sculpture, smashing it into smaller, simpler pieces, counting those, and then trying to glue the answers back together. It's messy, prone to errors, and often loses the "refined" details of the original shape.

Sergej Monavari's paper is a breakthrough because it solves this counting problem for a specific type of universe called a "local curve" (think of a long, flexible tube with extra dimensions wrapped around it) without smashing it apart. Instead, the author uses a clever "laser scan" technique called localisation.

Here is a simple breakdown of the paper's key ideas using everyday analogies:

1. The Problem: Counting in the Dark

Imagine trying to count how many different ways you can arrange a pile of sand on a wobbly, shifting table. If the table is unstable, you can't just count the sand directly.

• The Old Way: Break the table into small, stable tiles, count the sand on each tile, and add them up. This is the "degeneration" method.
• The New Way (This Paper): Shine a special laser (the Torus action) on the table. The laser only illuminates specific "fixed points" where the sand settles into perfect, stable patterns. The author realized that if you only look at these illuminated spots, you can count everything perfectly without ever breaking the table.

2. The "Skew Nested Hilbert Schemes": The Russian Dolls

The author introduces a new tool called Skew Nested Hilbert Schemes.

• The Analogy: Imagine a set of Russian nesting dolls, but instead of just one inside another, they are arranged in a complex grid (like a 3D chessboard). Some dolls are missing, creating a "skew" shape.
• The Magic: The author proved that the complex "quantum sand" structures on the big tube universe are mathematically identical to these simpler, nested Russian doll arrangements on a simple curve (a line).
• Why it matters: Instead of counting sand on a 3D tube, you just count the ways to stack these Russian dolls on a 1D line. It turns a 3D nightmare into a 1D puzzle.

3. The "Refined" Count: Adding Color and Texture

Standard counting just gives you a number (e.g., "There are 5 ways"). Refined counting is like adding a color palette to your count.

• The Analogy: Imagine counting the ways to build a tower with Lego.
  • Standard: "I can build 5 towers."
  • Refined: "I can build 5 towers: 2 are red, 2 are blue, and 1 is a mix."
• The paper calculates these "colored" counts (using variables $t_1$ and $t_2$) for any shape of the tube, any number of bricks, and any "twist" in the universe.

4. The "Vertex": The Universal Building Block

The paper discovers that all these complex counts are actually built from just three universal "building blocks" (called Universal Series).

• The Analogy: Think of a massive library of books. The author realized that every single book in the library is just a combination of three specific "alphabet letters" (A, B, and C) arranged in different ways.
• Once you know the formulas for these three letters, you can instantly write down the answer for any curve, any degree, and any genus (shape complexity). You don't need to re-calculate from scratch every time.

5. The "DT/PT Correspondence": Two Languages, One Story

There are two different ways mathematicians describe these structures:

1. DT (Donaldson-Thomas): Counting "sand piles" (sub-schemes).
2. PT (Pandharipande-Thomas): Counting "stable pairs" (a specific type of map).

• The Analogy: It's like describing a cake. One person counts the number of layers (DT), and another counts the number of frosting swirls (PT). They look different, but they describe the same cake.
• The Result: The paper proves a strict mathematical rule connecting them: DT = (Base Constant) × PT. This confirms a famous conjecture by Nekrasov and Okounkov. It's like proving that if you know the number of frosting swirls, you can instantly calculate the number of layers, no matter how big the cake is.

6. The "Refined Limit": The String Theory Connection

Finally, the paper looks at what happens when you stretch the "laser" parameters to infinity (the Refined Limit).

• The Analogy: Imagine zooming out so far on a map that the roads disappear, and you only see the major highways.
• The Result: In this limit, the complex formulas simplify into a beautiful, clean expression that matches a prediction made by string theorists (Aganagic and Schaeffer) years ago. This connects pure math (counting sheaves) directly to theoretical physics (topological string theory), suggesting that the "quantum Lego" structures are actually describing the vibrations of strings in the universe.

Summary

In short, this paper is a masterclass in simplification through symmetry.

• Old approach: Smash the problem into pieces, count the pieces, and hope they fit back together.
• New approach: Shine a light on the problem, realize the complex 3D shapes are just fancy versions of simple 1D stacking puzzles, and use a universal "alphabet" to write down the answer instantly.

This work not only solves a decades-old counting problem for a wide class of shapes but also provides a new, powerful toolkit that physicists and mathematicians hope will help unlock the secrets of the "refined" universe, potentially leading to a proof of the GW/PT correspondence for all Calabi-Yau threefolds.

Technical Summary

1. Problem Statement and Context

The paper addresses the K-theoretically refined Donaldson-Thomas (DT) theory for local curves. A local curve $X$ is defined as the total space of the sum of two line bundles over a smooth projective curve $C$ of genus $g$:
$$X = \text{Tot}_C(L_1 \oplus L_2)$$
The goal is to compute the refined DT partition function $\widehat{DT}_d(X, q)$, which counts stable sheaves (or subschemes) on $X$ with fixed curve class $\beta = d[C]$ and Euler characteristic $n$, weighted by the equivariant parameters $t_1, t_2$ of the torus action scaling the fibers.

Key Challenges:

• Non-compactness: The moduli spaces (Hilbert schemes) are generally not proper, requiring virtual localization techniques.
• Refinement: Moving from cohomological DT theory (counting with Euler characteristics) to K-theoretic DT theory (counting with structure sheaves and virtual canonical bundles) introduces complex dependence on equivariant parameters.
• Arbitrary Genus: Previous results often relied on degeneration techniques (breaking the curve into $\mathbb{P}^1$s) or were limited to specific genera or degrees.
• Conjectural Relations: There are deep conjectures linking DT theory to Pandharipande-Thomas (PT) theory, Gromov-Witten (GW) theory, and topological string theory, particularly in the refined setting.

2. Methodology

The author avoids traditional degeneration techniques. Instead, the proof relies on direct localization methods and the geometry of skew nested Hilbert schemes.

A. Skew Nested Hilbert Schemes

The core geometric innovation is the introduction of skew nested Hilbert schemes $C[n]$.

• Let $\lambda$ be a Young diagram and $n$ be a skew plane partition of shape $\mathbb{Z}^2_{\geq 0} \setminus \lambda$.
• $C[n]$ parametrizes flags of zero-dimensional subschemes $(Z_\square)_{\square \in \mathbb{Z}^2_{\geq 0} \setminus \lambda}$ on the curve $C$, nested according to the opposite poset structure of the lattice.
• The author proves that the connected components of the $T$-fixed locus of the Hilbert scheme $\text{Hilb}_n(X, \beta)$ are isomorphic to these skew nested Hilbert schemes $C[n]$.

B. Virtual Localization and Universality

1. Fixed Locus Analysis: Using the $T = (\mathbb{C}^*)^2$ action on $X$, the fixed locus is decomposed into disjoint unions of $C[n]$.
2. Virtual Classes: The paper establishes that the virtual fundamental class on these fixed components coincides with the ordinary fundamental class (they are reduced local complete intersections).
3. Universality: The generating series of the invariants are shown to depend only on the Chern numbers $(g, \deg L_1, \deg L_2)$. This allows the reduction of the computation to three "universal series" ($A_\lambda, B_\lambda, C_\lambda$) determined by evaluating the theory on specific toric cases: $(\mathbb{P}^1, \mathcal{O}, \mathcal{O})$, $(\mathbb{P}^1, \mathcal{O}, \mathcal{O}(-2))$, and $(\mathbb{P}^1, \mathcal{O}(-2), \mathcal{O})$.

C. Combinatorial Realization (The Vertex)

The universal series are expressed in terms of 1-leg K-theoretic equivariant vertices ($\widehat{V}_\lambda(q)$).

• These vertices are combinatorially realized as sums over skew plane partitions.
• The author connects these to the vertex/edge formalism (originally in equivariant cohomology) but lifts them to K-theory.
• The computation involves summing over all skew plane partitions of a fixed shape, effectively reducing the problem to the combinatorics of the Hilbert scheme of points on $\mathbb{A}^2$.

3. Key Contributions and Results

A. Full Computation of the Refined DT Partition Function

The main theorem (Theorem 1.1 / Definition 5.6) provides a closed formula for the refined DT partition function for any degree $d \geq 0$ and genus $g \geq 0$:
$$\widehat{DT}_d(X, q) = \sum_{|\lambda|=d} \left( q^{|\lambda|} \widehat{A}_\lambda(q) \right)^{1-g} \cdot \left( q^{-n(\lambda)} \widehat{B}_\lambda(q) \right)^{\deg L_1} \cdot \left( q^{-n(\lambda)} \widehat{C}_\lambda(q) \right)^{\deg L_2}$$
where $\widehat{A}, \widehat{B}, \widehat{C}$ are explicit universal series involving the 1-leg vertex $\widehat{V}_\lambda(q)$ and products over boxes in the Young diagram $\lambda$ involving arm/leg lengths and equivariant parameters.

B. Refined Limit and Topological Strings

In the Calabi-Yau case ($L_1 \otimes L_2 \cong \omega_C$), the author takes the refined limit (scaling $t_1, t_2 \to \infty$ while keeping $t_1 t_2$ constant).

• Result: The resulting formula (Theorem 1.4) matches a conjectural formula for the refined topological string partition function proposed by Aganagic-Schaeffer.
• This establishes a rigorous mathematical derivation of a string-theoretic prediction using algebraic geometry.

C. Anti-Diagonal Restriction

The paper computes the partition function under the restriction $t_1 t_2 = 1$.

• Result: The refined invariants reduce to a signed topological Euler characteristic partition function (Corollary 1.6).
• This confirms that under this specialization, the K-theoretic invariants become purely topological, relating to Behrend's function.

D. DT/PT Correspondence

The paper extends these methods to Pandharipande-Thomas (PT) theory (stable pairs).

• Main Result: The author proves the K-theoretic DT/PT correspondence for local curves in arbitrary genus (Corollary 1.8):
  $$\widehat{DT}_d(X, q) = \widehat{DT}_0(X, q) \cdot \widehat{PT}_d(X, q)$$
• This confirms a conjecture by Nekrasov-Okounkov. The proof relies on the vertex correspondence $\widehat{V}_\lambda = \widehat{V}_\emptyset \cdot \widehat{V}^{PT}_\lambda$ established by Kuhn-Liu-Thimm.

E. Degree 0 and 1 Formulas

• Degree 0: A compact plethystic exponential formula is derived (Corollary 1.2), generalizing Okounkov's results for $A_3$.
• Degree 1: Explicit formulas are provided for irreducible curve classes, recovering and generalizing known results in GW theory via the GW/DT correspondence.

4. Significance and Impact

1. Methodological Breakthrough: By avoiding degeneration techniques, the paper provides a direct, local proof for local curves. This avoids the cumbersome machinery of relative invariants and degeneration formulas, offering a clearer geometric picture of the fixed loci.
2. Unification of Theories: The work unifies K-theoretic DT theory, PT theory, and topological string theory. It validates the Aganagic-Schaeffer conjecture and the Nekrasov-Okounkov DT/PT correspondence in a non-toric, arbitrary genus setting.
3. Foundation for GW/PT: The author argues that these explicit results on local curves are a crucial step toward proving the refined GW/PT correspondence (Brini-Schuler) for all smooth Calabi-Yau threefolds, leveraging Pardon's machinery which suggests curve-counting theories are universally determined by their local curve analogues.
4. Explicit Combinatorics: The paper provides explicit, computable formulas involving Young diagrams, hook lengths, and MacMahon functions, making the refined invariants accessible for further physical and mathematical exploration.

In summary, Monavari's work solves the refined local DT theory completely, bridging the gap between abstract enumerative geometry and concrete string-theoretic predictions through a novel application of skew nested Hilbert schemes and localization.