Equal-mass sunset integrals and genus-zero local Gromov--Witten theory of Calabi--Yau $n$-folds

Technical Summary: Equal-Mass Sunset Integrals and Genus-Zero Local Gromov–Witten Theory of Calabi–Yau $n$-Folds

Problem Statement
The paper addresses the analytic and enumerative connection between two-dimensional equal-mass $(n-1)$-loop sunset Feynman integrals ($n \ge 4$) and the genus-zero Gromov–Witten (GW) theory of local Calabi–Yau $n$-folds. Specifically, it seeks to establish an exact identity between the sunset integral, viewed as a function of large Euclidean momentum, and the one-point descendant invariants of the local geometry $X_n = \text{Tot}(K_{F_n})$, where $F_n$ is a smooth $(1, \dots, 1)$ hypersurface in $(\mathbb{P}^1)^n$. While the two-loop case ($n=3$) was previously related to the prepotential of a local Calabi–Yau threefold via mirror symmetry, the general case for $n \ge 4$ involving descendant invariants had not been rigorously established.

Methodology
The proof is uniform in $n$ and integrates four distinct mathematical frameworks:

1. Algebraic Geometry and Virtual Reduction: The authors utilize the fact that stable maps to the local Calabi–Yau $X_n$ factor through the zero section over the Fano hypersurface $F_n$. They establish a "virtual reduction" theorem (Theorem 5.5), showing that the virtual class of $X_n$ is the top Chern class of an obstruction bundle intersected with the virtual class of $F_n$. This allows the extraction of GW invariants of $X_n$ from the ambient cohomology of the toric variety $(\mathbb{P}^1)^n$.
2. Mirror Symmetry and Hypergeometric Series: A "mixed hypergeometric theorem" (Theorem 6.6) is employed, utilizing a hypergeometric modification of the Givental $I$-function with two independent characteristic classes (Euler and inverse Euler). This places the series on a twisted Givental cone. Through monic division and ambient pairing, the authors extract all numerical descendant invariants $N_{a,\beta}$ simultaneously without assuming the ambientness of the full cohomology (though they prove ambientness for the relevant classes separately in Theorem 15.1).
3. Differential Equations: A finite algorithm (Definition 8.1) constructs an order-$(n-1)$ differential operator $P_{n-1}$ over $\mathbb{Q}$. The authors certify that this operator annihilates the relevant period series and is irreducible, thereby identifying it as the minimal operator for the sunset integral. This derivation is independent of the geometric mirror theorem.
4. Analysis and Arithmetic: The physical integral is analyzed via a positive Laplace transform, converting an oscillatory Gamma identity (from Iritani's work on Fano manifolds) into a signed Gamma formula. To address integrality, the authors construct a smooth compact Calabi–Yau double cover $\tilde{Z}$ of a projective bundle over $F_n$. This compactification isolates the local sector on a target with a trivial canonical bundle, allowing the application of quantum K-theory and symplectic integrality theorems (Chou, Ionel–Parker, Siebert) to prove that the transformed primary invariants are integers.

Key Contributions and Results

• Exact Identity (Theorem A/10.1): The paper proves that the equal-mass sunset integral $I^{(n-1)}_\ominus(s)$, as a convergent logarithmic germ at large Euclidean momentum ($s \to -\infty$), is exactly equal to a specific combination of the local mirror coordinate $R$, the Novikov variable $Q$, and the genus-zero one-point descendant invariants $D^\Delta_a(Q)$ of $X_n$. The formula is:
  $$I^{(n-1)}_\ominus(s) = (-1)^{n-1} y A \sum_{k=0}^{n-1} (-1)^k \gamma_k T_{n-1-k}(R, Q)$$
  where $\gamma_k$ are coefficients derived from the Gamma class (involving odd zeta values $\zeta(2r+1)$), and $T_j$ are assembly polynomials involving the descendant series and their derivatives.
• Explicit Descendant Extraction (Theorem B/6.9): The authors provide a closed-form expression for the descendant series $D^\Delta_a(Q)$ in terms of the coefficients of the ambient hypergeometric series $I(y,z)$ and the inverse mirror map. This resolves the extraction of all numerical descendants from the mirror series.
• Global Univalence and Convergence (Theorem 11.3/11.4): The paper proves the global univalence of the principal mirror map $y \mapsto Q$ on the slit plane $\mathbb{C} \setminus (-\infty, -n^{-2}]$. It establishes explicit upper and lower bounds for the Taylor radius of the inverse mirror map, refining the domain of convergence for the descendant series.
• Integrality of Transformed Primaries (Corollary 14.10): By lifting the problem to the compact Calabi–Yau cover $\tilde{Z}$, the authors prove that the "primary cover transform" coefficients $b_\beta$ (defined via the multiple-cover formula $N_{0,\beta} = \sum b_{\alpha/k}/k^2$) are integers for all $n \ge 4$. This is achieved via two independent routes: an algebraic proof using quantum K-theory on $\tilde{Z}$ and a symplectic proof using the semipositive condition on $\tilde{Z}$.
• Differential Operator Certificate: The paper provides a uniform construction of the differential operator governing the sunset integral for all $n$, certifying its action on the full series without relying on finite coefficient fitting.

Significance and Claims
The paper claims to provide the first rigorous enumerative identity for the coefficients of the equal-mass sunset integral expansion for all $n \ge 4$. It identifies the "instanton coefficients" (previously obtained as rational numbers via residue calculus) as genus-zero one-point descendant invariants of a local Calabi–Yau $n$-fold.

Crucially, the authors are modest about the enumerative interpretation of the resulting integers. They explicitly state that:

• The integers $b_\beta$ are not claimed to be unsigned counts of embedded rational curves, sheaf-theoretic multiplicities, or physical BPS-state degeneracies. They are the output of a specific arithmetic transform applied to a fixed virtual invariant.
• No multiple-cover law is asserted for the descendant series with $a > 0$; only the primary sector ($a=0$) is transformed.
• The standard "ideal geometry" hypotheses (unobstructedness of simple rational curves, moving-cover conditions) required for higher-dimensional covering behavior are not assumed and, in fact, fail for $X_n$. The proofs are purely algebraic and do not rely on these geometric conditions.

The work bridges perturbative quantum field theory (Feynman integrals) and algebraic geometry (Gromov–Witten theory) by demonstrating that the analytic structure of the sunset integral is governed by the mirror symmetry of a local Calabi–Yau geometry, with the coefficients of the expansion being deeply rooted in the arithmetic of the Gamma class and the integrality of quantum K-theory invariants.