Technical Summary: Open Strings on Knot Complements

Problem Statement

The paper addresses the computation of open topological string partition functions on Lagrangian knot complements $M_K = S^3 \setminus K$ within the cotangent bundle $T^*S^3$ and the resolved conifold $X$. While the duality between Chern-Simons theory and topological strings is well-established for knot conormals (where the partition function generates colored HOMFLYPT polynomials), the case of the knot complement itself presents significant challenges. Specifically, for general fibered knots, the set of holomorphic curves (and their associated flow loops) is infinite and complex, making direct enumeration difficult. The paper seeks to derive explicit, deformation-invariant formulas for these partition functions using skein-valued curve counting techniques and to connect these results to the quantization of augmentation varieties and $\hat{Z}$-invariants.

Methodology

The authors employ a combination of symplectic geometry, Floer theory, and topological string theory:

1. Skein-Valued Curve Counting: Building on the framework established in [5], the authors utilize the fact that counting holomorphic curves with boundaries on a Lagrangian $L$ by their values in the HOMFLYPT skein module $Sk(L)$ yields a deformation-invariant quantity. This approach resolves the wall-crossing phenomena associated with boundary nodes.
2. Flow Loop Localization: For fibered knots, the complement $M_K$ admits a nowhere-vanishing 1-form $\beta$. The authors establish a correspondence between the flow loops of $\beta$ and holomorphic annuli stretching between the knot complement and the zero section $S^3$. They prove that the partition function localizes on these flow loops, allowing the infinite sum over curves to be expressed as a product over a finite (or manageable) set of flow loops.
3. Skein-Valued Annuli: The contribution of each flow loop is determined by the type of holomorphic annulus it supports (bosonic/fermionic, annulus/anti-annulus), which depends on the orientation and spin structures of the Lagrangians and the eigenvalues of the linearized return map of the flow.
4. Gl$_1$-Skein Specialization: To obtain concrete formulas, the authors specialize the skein-valued partition function to the $gl_1$-skein. In this limit, the skein module becomes a power series ring generated by the meridian of the knot, and the partition function reduces to a sum over multiplicities of flow loops weighted by HOMFLYPT polynomials of the associated links in $S^3$.
5. Cobordism and Quantization: The authors utilize Lagrangian cobordisms (specifically relating knot complements to Hopf links) to derive augmentation polynomials. They further analyze the consistency of quantization schemes across different branches of the augmentation variety (conormal vs. complement) by comparing the resulting $q$-difference equations.

Key Contributions and Results

1. Flow Loop Formula for Fibered Knots

The primary theoretical result is Theorem 1.1, which expresses the skein-valued partition function of a fibered knot complement $M_K$ as a product over the set of flow loops $\mathcal{F}$ of a generic 1-form on $M_K$:
$$Z(T^*S^3, M_K) = \prod_{\gamma \in \mathcal{F}} \iota_{\partial A(\gamma)}(\Psi^{(\sigma(\gamma), \epsilon(\gamma))}_{ann})$$
where $\Psi_{ann}$ represents the skein-valued partition function of a basic holomorphic annulus, and $\iota$ denotes insertion into the skein of $M_K \sqcup S^3$. This formula demonstrates that the partition function is organized by contributions from (generalized) annuli, implying a lift to M-theory with M2-branes wrapping these annuli.

2. Explicit Formulas for Torus Knots

While general fibered knots yield infinite sets of flow loops, the authors show that for negative torus knots $T_{2, 2p+1}$, the flow loop structure simplifies drastically.

• Proposition 3.2: The complement of a $T_{2, 2p+1}$ knot admits a 1-form with exactly two basic flow loops: an elliptic loop of degree 2 and a negative hyperbolic loop of degree $2p+1$. These loops form a negative Hopf link.
• Corollary 1.2 & Proposition 4.2: This localization leads to an explicit, finite-sum formula for the $gl_1$ partition function $\psi(X, M_K)$ involving the HOMFLYPT polynomial of the negative Hopf link.
• Quiver Structure: The partition function for $T_{2, 2p+1}$ complements is shown to admit a generalized quiver structure (Equation 4.23) involving six basic curves (four disks, one annulus, one anti-annulus). This implies integrality structures for M2-brane counts on the knot complement.

3. Quantization and Augmentation Curves

The paper investigates the relationship between the partition function on the complement branch and the conormal branch of the augmentation variety.

• Consistency of Quantization: The authors verify that the partition function on the complement branch satisfies the same $q$-difference equation (quantized augmentation polynomial) as the conormal branch, up to a specific change of variables ($a \to aq$) and coordinate shifts.
• Derivation of Augmentation Polynomials: Using the explicit flow loop formulas and Lagrangian cobordisms, the authors provide new derivations of the augmentation polynomials for $T_{2, 2p+1}$ knots (Sections 6.3 and 7.3).
• Tropicalization: The paper analyzes the tropical limits of augmentation curves, defining "phases" corresponding to different Lagrangian fillings. It demonstrates how the decompactification limit ($R \to \infty$) leads to tropical curves and discusses the consistency of quantization schemes across these phases.

4. Connection to $\hat{Z}$-Invariants

The authors connect their results to the $\hat{Z}$-invariants for knot complements. They argue that the partition function $\psi(X, M_K)$ computed via curve counting corresponds to the $\hat{Z}$-invariant (specifically the $F_K$ invariant) after appropriate variable identifications and framing adjustments. This provides a geometric interpretation of the $\hat{Z}$-invariant as the partition function of open topological strings on the knot complement.

Significance and Claims

The paper claims to provide a direct geometric interpretation for the wave function on the complement branch of the augmentation variety, which is related to the $\hat{Z}$-invariant. By establishing a flow loop localization formula, the authors bridge the gap between the infinite complexity of general knot complements and the computable structures of torus knots.

Key claims regarding significance include:

• Generalization of Knot-Quivers Correspondence: The work extends the known knot-quivers correspondence (previously established for knot conormals) to knot complements, showing that curve counts on complements are also organized by quiver structures.
• Integrality: The quiver structure implies integrality of M2-brane counts on knot complements, offering a geometric basis for conjectured integrality structures in complex Chern-Simons theory.
• Unification of Branches: The paper demonstrates that the quantization of the augmentation curve is consistent across the conormal and complement branches (up to known shifts), reinforcing the idea that these distinct geometric phases describe the same underlying quantum system.
• Computational Tool: The derived formulas offer a concrete method for computing partition functions and augmentation polynomials for torus knots without relying solely on recursive relations or dualities, providing a new geometric perspective on these invariants.

The authors remain modest regarding general fibered knots, acknowledging that for non-torus fibered knots, the set of flow loops is infinite and the resulting expressions are complex, though substantial cancellations are expected to yield polynomial recursion relations. The primary success is demonstrated for the class of $T_{2, 2p+1}$ torus knots.