⚛️ high-energy theory

How to choose a good rational basis for elliptic Feynman integrals?

This paper extends a method using elliptic generalizations of leading singularities to construct a rational basis for elliptic Feynman integrals, demonstrating that the off-diagonal blocks of their differential equations possess a universal $\epsilon$ -dependent structure preserved under decoupling transformations.

Original authors: Ekta Chaubey, Vasily Sotnikov

Published 2026-09-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ekta Chaubey, Vasily Sotnikov

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 high-energy world of particle physics, scientists study the collisions of subatomic particles to understand the fundamental forces that shape our universe. When two particles smash together, they scatter, creating a spray of new particles. To predict exactly what happens in these collisions, physicists use complex mathematical objects called scattering amplitudes. These are essentially the blueprints that tell us the probability of every possible outcome. Calculating these blueprints is notoriously difficult, especially when the interactions involve loops of virtual particles that pop in and out of existence. For decades, physicists have found that the most efficient way to handle these calculations is to split the problem into two distinct parts: a rational part, which involves simple fractions and polynomials, and a transcendental part, which involves complex, multi-valued functions that capture the deep geometric structure of the interaction. Keeping these two parts separate is crucial because it allows researchers to use powerful modern techniques to reconstruct the full answer from limited data. However, this separation becomes incredibly tricky when the underlying geometry of the problem involves elliptic curves, a type of mathematical shape more complex than a simple circle or sphere.

For a long time, physicists struggled to find a clean way to separate the rational and transcendental parts of these elliptic calculations. The standard approach involved transforming the equations into a form where the complexity was neatly packaged, but this often required introducing new, artificial mathematical functions that carried their own confusing ambiguities. It was like trying to organize a messy room by moving the clutter into a new box that was just as messy as the original. The researchers in this study, working at the University of Bonn and Johannes Gutenberg University Mainz, asked a fundamental question: is there a better way to choose the starting point for these calculations that naturally keeps the rational and transcendental parts distinct, even when elliptic curves are involved? They proposed a new method for selecting a specific set of starting values, or a "basis," that respects the underlying geometry of the problem without introducing unnecessary confusion.

The team focused on a technique inspired by "leading singularities," which are essentially the most extreme, dominant features of a mathematical integral. In simpler cases, physicists could use these features to build a perfect starting point. The researchers extended this idea to the elliptic case, constructing a new set of rational bases that are guided by the geometry of the elliptic curve itself. They found that by matching the integrals to specific algebraic shapes defined by the curve, they could create a basis where the dependence on a small parameter, used to regulate the calculations, follows a very simple and predictable pattern. This pattern is linear, meaning the complexity grows in a straight, manageable line rather than spiraling out of control. This discovery suggests that the rational part of the calculation can be kept clean and free from the confusing ambiguities that usually plague these problems.

A major part of their work involved looking at how different parts of the calculation interact with each other. In these complex systems, the main elliptic part is connected to smaller, simpler parts called subsectors. The researchers investigated the "off-diagonal" blocks, which are the mathematical terms that describe how these different parts talk to one another. They discovered that even in these connecting regions, the complexity follows a universal structure. Specifically, the way the calculations depend on the small parameter can be organized into a simple form that remains stable even when the equations are transformed to solve them. This finding is significant because it shows that the simplicity observed in the main part of the problem extends to the connections between all the different pieces. It implies that the new rational basis they constructed is not just a local fix but a robust framework that works across the entire system.

The researchers also explained the origin of a specific integer number that appears in their equations, a number that dictates how the complexity scales. They showed that this number is not arbitrary but is determined by the local behavior of the mathematical integrand at a specific point, much like how the steepness of a hill determines how fast a ball rolls down it. By understanding this connection, they provided a way to predict this number before doing the full calculation. Their work suggests that for a wide range of complex particle physics problems, there is a preferred way to set up the equations that keeps the rational coefficients simple and the transcendental functions well-behaved. This approach offers a systematic strategy for tackling multiscale problems, which are calculations involving many different energy scales at once.

While the paper does not claim to have solved every possible elliptic problem, it provides strong evidence that this new method works for the families of integrals they tested. They found that for some complex cases, it is impossible to make the equations perfectly simple using only rational numbers, but their method ensures that the remaining complexity is as minimal and structured as possible. This allows physicists to separate the rational information, which encodes the specific details of the particle interactions, from the transcendental information, which encodes the universal geometric structure. The result is a clearer path to calculating the scattering amplitudes needed for precision experiments at facilities like the Large Hadron Collider. By establishing a clean separation between the rational and multivalued parts of the amplitude, this work paves the way for more efficient analytic computations and more stable numerical solutions, helping physicists to extract the most precise possible predictions from their theories.

Technical Summary: How to choose a good rational basis for elliptic Feynman integrals?

Problem Statement
The analytic computation of multi-loop scattering amplitudes relies on expressing results as linear combinations of transcendental functions with rational coefficients. While this separation is well-understood for rational and algebraic cases (involving square roots), it becomes significantly more complex for elliptic Feynman integrals. In the elliptic case, constructing an $\epsilon$ -factorized basis (where the dimensional regulator $\epsilon$ factors out of the differential equations) typically requires transformations involving complete elliptic integrals. These transformations introduce auxiliary multivalued functions and monodromy ambiguities that are not fixed by physical boundary conditions. Consequently, there is no universally known, simple relation between a rational basis (where coefficients are rational) and an $\epsilon$ -factorized algebraic/elliptic basis. The central question addressed is whether a preferred rational basis exists for elliptic integrals that serves as a natural starting point for constructing monodromy-invariant transcendental function bases, thereby maintaining a clean separation between rational and multivalued parts of the amplitude.

Methodology
The authors build upon a method proposed in Ref. [1], which constructs rational bases using an elliptic generalization of leading singularities (LS). The methodology proceeds in three main stages:

Elliptic Leading Singularity Construction:
- The authors generalize the $d\log$ integrand construction used in the rational case. Instead of requiring the integrand to be a logarithmic form, they match the algebraic part of the integrand (after maximal cut) to a basis of algebraic one-forms naturally selected by the cohomology of the underlying elliptic curve.
- These basis forms correspond to Abelian differentials of the first, second, and third kinds ( $\omega_\psi, \omega_\phi, \omega_{\pi}$ ).
- Rational coefficients are determined such that the ansatz matches these specific one-forms while vanishing on standard logarithmic leading singularities. This yields a rational basis $J_R$ .
Analysis of Homogeneous Blocks:
- The differential equations (DEs) for this rational basis $J_R$ are shown to have a diagonal elliptic block that is linear in $\epsilon$ : $dJ_R = [M^{(0)}_R(s) + \epsilon M^{(1)}_R(s)] J_R$ .
- A universal transformation $T_C$ , constructed from the elliptic leading singularities (periods and quasi-periods), is applied to $J_R$ . This transforms the system into a new basis $J_C$ satisfying a DE where the $\epsilon$ -dependence of matrix elements factorizes as either $\epsilon$ or $1-n\epsilon$ , with $n$ being an integer.
- The leading-order (LO) connection in this new basis is nilpotent, allowing solutions to be constructed order-by-order in $\epsilon$ as iterated integrals.
Extension to Off-Diagonal Blocks:
- The paper extends the analysis beyond the maximal cut (diagonal blocks) to the off-diagonal blocks that couple the elliptic sector to its subsectors.
- The authors investigate whether rational shifts of the form $J'_{R,\phi} = J_{R,\phi} + \sum r_i(s, \epsilon) J_{R,i}$ can simplify the off-diagonal DE blocks.
- They demonstrate that while complete $\epsilon$ -factorization of off-diagonal blocks is not always achievable via rational shifts alone, a specific restricted structure can always be enforced.

Key Contributions and Results

Origin of the Integer $n$ : The paper provides a derivation for the integer $n$ appearing in the factor $(1-n\epsilon)$ . It is shown that $n$ is determined by the local $\epsilon$ -dependent scaling of the integrand at the double pole associated with the second-kind differential. Specifically, $n = \sum (b_i \nu_i)/2$ , where $b_i$ are the $\epsilon$ -coefficients of the exponents of irreducible polynomials in the Baikov representation, and $\nu_i$ are their orders at the double pole.
Structure of Off-Diagonal Blocks: A primary result is the characterization of the off-diagonal DE blocks for the rational basis constructed via elliptic LS. The authors find that for rational bases constructed according to their prescription, the $\epsilon$ $ϵ$ dependence of these blocks can be organized into a universal structure:
- Entries coupling to first-kind periods ( $\psi$ ) and third-kind periods ( $\pi_a$ ) can be made proportional to $\epsilon$ .
- Entries coupling to the second-kind period ( $\phi$ ) take the form $\frac{\epsilon}{1-n\epsilon} \omega^1 + \frac{\epsilon^2}{1-n\epsilon} \omega^2$ .
- This structure is preserved under the transformation $T_C$ used to decouple the homogeneous elliptic block.
Monodromy Invariance: The paper argues that the rational basis $J_R$ provides a natural framework for defining monodromy-invariant transcendental functions. By applying the inverse transformation $T_C^{-1}$ to the iterated integral solutions of the decoupled system, one reconstructs functions that are invariant under the monodromy group of the homogeneous DEs. This ensures the clean separation of rational coefficients from multivalued transcendental functions.
Limitations and Observations: The authors note that while the restricted $\epsilon$ dependence holds for the rational basis, achieving a fully $\epsilon$ -factorized basis with a constant intersection matrix generally requires enlarging the set of transcendental functions $E(s)$ and introducing non-rational transformations. The paper does not claim that the rational basis itself is $\epsilon$ -factorized, but rather that it is the optimal rational starting point.

Significance
The paper claims that the proposed rational basis, constructed via elliptic leading singularities, offers a systematic strategy for multiscale elliptic Feynman integrals. Its significance lies in:

Geometric Consistency: It incorporates information about the underlying elliptic geometry directly into the rational basis selection.
Monodromy Invariance: It facilitates the construction of transcendental function bases where the rational and multivalued parts of the amplitude are cleanly separated, avoiding the gauge-like ambiguities associated with arbitrary choices of auxiliary homogeneous solutions.
Universality of Structure: The discovery that the $\epsilon$ dependence of off-diagonal blocks follows a universal, restricted form (preserved under decoupling) suggests that the simplicity observed on the maximal cut extends to the full system.
Practical Utility: This approach provides a concrete interface between iterated-integral representations (encoding geometry) and function bases suitable for scattering amplitudes with rational coefficients, potentially aiding both analytic reconstruction and numerical evaluation of complex multiscale amplitudes.

The authors conclude that while further work is needed to understand residual freedoms in off-diagonal blocks and extensions to higher-genus geometries, the proposed method establishes a robust foundation for handling elliptic sectors in precision phenomenology.

More like this