🔢 mathematics

The three- and four-loop equal-mass sunset integrals and local Calabi--Yau fourfolds and fivefolds

This paper establishes that the equal-mass three- and four-loop sunset Feynman integrals in two dimensions are explicitly determined by the genus-zero Gromov–Witten theory of local Calabi–Yau fourfolds and fivefolds, specifically relating the integrals to the one-point potential and its gravitational descendants rather than a prepotential.

Original authors: Hartmut Maennel, Pierre Vanhove

Published 2026-10-09

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Hartmut Maennel, Pierre Vanhove

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

In the subatomic world, particles do not travel in straight lines; they interact, collide, and exchange energy in complex webs of possibility. To predict the outcome of these interactions, physicists use mathematical tools called Feynman integrals. These are not simple calculations but rather vast sums of infinite possibilities, representing every way a particle could move and interact. For decades, scientists have been able to solve these integrals for simple scenarios, but as the interactions become more complex—specifically when more loops of interaction are added—the mathematics becomes incredibly difficult. The results often involve numbers that do not fit into standard categories, defying the usual patterns found in physics. Understanding these difficult integrals is crucial because they describe the fundamental forces of nature with extreme precision. If we cannot calculate them, we cannot fully test our theories against the real universe.

A new study by Hartmut Maennel and Pierre Vanhove tackles one of these stubborn problems: the calculation of specific particle interactions known as "sunset integrals" when they involve three and four loops of complexity. Until now, the mathematical landscape for these specific calculations was largely uncharted territory. The researchers discovered that these complex particle calculations are not isolated mathematical curiosities but are deeply connected to a branch of geometry involving shapes called Calabi–Yau manifolds. These are intricate, multi-dimensional spaces that appear in theories about the extra dimensions of our universe. The team found that the values of the sunset integrals can be determined exactly by counting specific geometric features on these shapes. It is as if the chaotic behavior of subatomic particles is being governed by a hidden, orderly geometric structure.

The paper focuses on two specific cases: the three-loop and four-loop versions of these integrals. In the three-loop case, the researchers proved that the entire calculation can be described by a single, powerful mathematical series derived from the geometry of a four-dimensional Calabi–Yau space. This series acts like a master key, unlocking the values of the integral. The team did not just guess this connection; they derived it independently using two different methods. One method relied on solving a specific type of differential equation that governs how the integral changes, while the other used a geometric approach involving the counting of curves on these shapes. Both methods led to the exact same result, confirming the connection with high certainty. They also proved that the numbers appearing in this series are always whole integers, a property that suggests a deep underlying order in the physics.

For the more complex four-loop case, the situation is slightly different but equally revealing. Here, the calculation cannot be reduced to a single series alone. Instead, the researchers showed that two specific series are sufficient to determine the entire integral. This finding is significant because it establishes a clear limit on how much information is needed to solve the problem, even if it remains an open question whether a single series could eventually suffice. The study also pinpointed exactly how these integrals behave at the very edge of their physical validity, a point known as the threshold. They found that while the values remain continuous, they become jagged and non-smooth at this point, a behavior that was confirmed through rigorous computer-assisted proofs.

A particularly striking discovery in the paper is the role of a specific mathematical constant, the Riemann zeta function evaluated at three, which appears as a boundary term in the calculation. This constant, which arises from the geometry of the shapes, acts as a correction factor that ensures the physical predictions match the mathematical structure. The researchers also demonstrated that the geometry involved in the three-loop case is unique; it is not a simple, symmetric shape like those used in previous two-loop studies, but a more complex variety that required new mathematical tools to analyze. By bridging the gap between the messy, high-energy world of particle physics and the elegant, abstract world of algebraic geometry, this work provides a new way to see the universe. It shows that even the most complicated particle interactions are governed by precise, countable geometric rules, offering a clearer path for future calculations in theoretical physics.

Technical Summary: The Three- and Four-Loop Equal-Mass Sunset Integrals and Local Calabi–Yau Fourfolds and Fivefolds

Problem Statement
The paper addresses the analytic and enumerative structure of equal-mass two-dimensional sunset Feynman integrals, specifically the three-loop ( $I^{(3)}_{\ominus}$ ) and four-loop ( $I^{(4)}_{\ominus}$ ) cases at large Euclidean momentum ( $s \to -\infty$ ). While the two-loop case is known to be governed by the genus-zero prepotential of a local Calabi–Yau threefold, the transition to $L \ge 3$ loops presents significant geometric and arithmetic challenges. For $L \ge 3$ , the associated moduli spaces of unmarked stable maps have positive virtual dimension, meaning no prepotential exists in the traditional sense. Consequently, the Feynman integrals cannot be expressed solely through a single prepotential; instead, they require a more complex structure involving one-point potentials, gravitational descendants, and boundary terms arising from the Gamma class. The paper aims to determine these integrals exactly in terms of the genus-zero Gromov–Witten (GW) theory of specific local Calabi–Yau varieties and to establish the arithmetic properties of the resulting series.

Methodology
The authors employ a multi-faceted approach combining algebraic geometry, differential equations, and computer-assisted analysis:

Geometric Identification: The integrals are linked to the local Calabi–Yau $(L+1)$ -folds $X_L = \text{Tot}(K_{F_L})$ , where $F_L \subset (\mathbb{P}^1)^{L+1}$ is a smooth hypersurface of multidegree $(1, \dots, 1)$ . For $L=3$ and $L=4$ , these correspond to a local Calabi–Yau fourfold and fivefold, respectively.
Mirror Symmetry and Oscillatory Integrals: The authors utilize Iritani's Gamma-class identity, relating the Feynman integral (viewed as a Laplace transform of an oscillatory integral) to the GW theory of $X_L$ . They distinguish between the "local" theory on $X_L$ and the "ambient" theory on the toric variety $Y_L = (\mathbb{P}^1)^{L+1}$ , using Euler twists to handle the non-toric nature of $F_3$ and the non-ambient cohomology of $F_4$ .
Analytic Derivation (Three Loops): An independent analytic derivation for the three-loop case is provided. This involves:
- Solving the inhomogeneous Picard–Fuchs equation governing the integral.
- Using the Frobenius method at infinity to determine the structure of the solution.
- Computing the boundary constants via the polar part of a Mellin transform, identifying the specific $\zeta(3)$ term.
- Establishing a "coefficient bridge" using quantum-Serre identities and logarithmic B-model recurrences to relate the integral's expansion coefficients to the one-point GW potential $F_{\gamma_L}$ .
Reconstruction from Primary Invariants: The paper develops a reconstruction algorithm using genus-zero topological recursion on a compact model of $X_L$ . This allows the expression of descendant series (which appear in the Feynman integral expansion) in terms of primary invariants.
Integrality and Quantum K-Theory: For the three-loop case, the authors prove the integrality of the multiple-cover transforms for arbitrary integral insertions. This is achieved by compactifying the local geometry into a semi-positive projective fourfold and applying the integrality theorem of Ionel and Parker (in the form of Chou), linking the GW invariants to quantum K-theoretic Euler characteristics.
Computer-Assisted Proofs: For the four-loop case and the analysis of endpoint behavior, the authors use rigorous ball arithmetic (via the Arb library) to certify the signs of connection quantities and the non-analyticity of the integrals at physical thresholds.

Key Contributions and Results

Exact Expressions in Terms of GW Invariants:
- Three Loops ( $L=3$ ): The integral $I^{(3)}_{\ominus}$ is expressed entirely in terms of the single one-point potential $F_{\gamma_3}$ (where $\gamma_3 = H^2/2$ ). All descendant series required for the expansion are determined by $F_{\gamma_3}$ via genus-zero topological recursion. The expansion takes the form:
  $I^{(3)}_{\ominus}(s) = -yA \left[ 4R^3 - (\theta_Q F_{\gamma_3})R^2 + \dots - 16\zeta(3) \right]$
  where $R$ is the mirror coordinate and $yA$ is the period.
- Four Loops ( $L=4$ ): The integral $I^{(4)}_{\ominus}$ is determined by two diagonal series: the one-point potential $F_{\gamma_4}$ and the diagonal two-point primary series $T$ (or equivalently the first descendant $D^{\Delta}_1$ ). The expansion includes a $\zeta(3)$ boundary term corrected by $F_{\gamma_4}$ .
Boundary Terms and Gamma Classes:
- The paper explicitly calculates the boundary terms arising from the Gamma class. For $L=3$ , the term is $-16\zeta(3)$ . For $L=4$ , it is $-80\zeta(3)$ multiplied by a logarithmic solution.
- It is shown that for $L \ge 3$ , the unmarked moduli spaces do not define numerical invariants directly; the first numerical observable is the one-point invariant with insertion $\gamma_L = H^{L-1}/(L-1)!$ .
Integrality and Arithmetic Structure:
- The coefficients of the multiple-cover expansion of $F_{\gamma_3}$ are proven to be integers ( $n_m \in \mathbb{Z}$ ), satisfying a specific Gopakumar–Vafa type integrality.
- An explicit quantum-K formula is derived for these transforms, providing a geometric realization of the BPS invariants.
- The paper provides a table of exact values for the coefficients up to degree 19, verified by rational arithmetic.
Analytic Properties and Thresholds:
- The paper determines the exact period connection polynomials at the maximally unipotent point.
- It proves that at the physical threshold ( $s = (L+1)^2$ ), the Feynman integrals are continuous but not $C^1$ (they possess a logarithmic singularity in their derivative).
- Computer-assisted proofs certify the Taylor radii of convergence for the descendant series in the mirror coordinate $Q$ . For $L=3$ , the radius lies between $1/12$ and $Q^* \approx e^{-7\zeta(3)/\pi^2}$ . For $L=4$ , the radius is bounded by $1/20$ and $Q_\infty \approx 0.3366$ .

Significance and Claims
The paper claims to provide the first exact determination of the three- and four-loop equal-mass sunset integrals in terms of the Gromov–Witten theory of local Calabi–Yau fourfolds and fivefolds.

Shift in Observables: It highlights a fundamental shift from $L=2$ to $L \ge 3$ : the governing object changes from a prepotential (unmarked invariants) to a one-point potential with specific insertions, as the virtual dimension of the moduli space becomes positive.
Reduction to Primary Data: A significant result is the demonstration that for $L=3$ , the entire Feynman integral is determined by a single generating series ( $F_{\gamma_3}$ ), whereas for $L=4$ , two diagonal inputs suffice (though minimality is not proven).
Independence of Proofs: The authors provide an independent analytic derivation for the three-loop case that does not rely on the oscillatory Gamma identity used in their companion paper, instead using Picard–Fuchs equations and Mellin transforms.
Arithmetic Rigor: The paper establishes the integrality of the BPS transforms for arbitrary integral insertions in the three-loop case, grounding the physical coefficients in rigorous algebraic geometry (quantum K-theory).
Computational Verification: The paper utilizes computer-assisted proofs to certify the non-analytic behavior at thresholds and the signs of connection constants, ensuring the results are not merely numerical coincidences but mathematically certified facts.

The work serves as a bridge between high-loop perturbative quantum field theory and the enumerative geometry of higher-dimensional Calabi–Yau manifolds, clarifying the specific role of Gamma classes and descendant invariants in the asymptotic expansion of Feynman integrals.

More like this