Aspects of Scattering Amplitudes and Moduli Space Localization

Sebastian MizeraSpringer Theses 2020 · DOI · arXiv · OpenAlex

Introduction

Calculating scattering amplitudes—the probabilities of particle interactions—is a cornerstone of theoretical physics. Traditionally, this involves summing thousands of Feynman diagrams or evaluating intricate string theory integrals. Recent developments have proposed a more elegant geometric alternative: viewing these amplitudes as "intersection numbers" of twisted cohomology classes on the moduli space of punctured spheres ($M_{0,n}$).

This paper establishes a unified geometric framework that bridges the gap between string theory and point-particle physics. By treating amplitudes as geometric objects, the author provides a powerful tool for computing tree-level interactions, even for theories involving massive states.

Geometric Localization and the Singularity Structure

A central challenge in this field is the "0/0 problem": integrals over the moduli space technically vanish in the "bulk" and diverge at the boundaries. The author resolves this by using "compactly-supported" forms, which force the mathematical focus onto the boundaries of the space.

Nodal Riemann surfaces representing the localization of integrals onto boundaries corresponding to physical factorization channels, where the geometry mirrors the singularity structure of scattering amplitudes.

As shown in the figure above, the topology of the moduli space—specifically how punctures collide—perfectly mirrors the physical factorization of amplitudes. These boundaries correspond one-to-one with trivalent Feynman diagrams, allowing the complex integration to "localize" or collapse into a sum over physical channels.

The Fibration and Recursion Method

To make these abstract intersection numbers computable for a large number of particles, the author introduces a puncture-by-puncture recursion method. This approach treats the multi-dimensional moduli space as a sequence of one-dimensional "fibers."

The fibration of the moduli space $\mathcal{M}_{0,n}$ as a sequence of punctured spheres, allowing the problem to be reduced to a sequence of one-dimensional residue calculations.

By "peeling off" punctures one by one, the integration is reduced to a series of linear algebra steps involving "braid matrices." These matrices describe how punctures wrap around each other, with their eigenvalues corresponding to physical factorization channels where intermediate particles go "on-shell."

Key Insights: From Massless to Massive Particles

One of the most significant contributions of this work is the treatment of massive particles. While existing worldsheet approaches like the CHY formalism excel at massless particles (represented by logarithmic forms or simple poles), this framework extends to massive states by utilizing forms with higher-order poles. This provides a clear geometric dictionary: the degree of the pole in the mathematical form corresponds to the mass of the physical particle.

This method has been validated by reproducing known tree-level amplitudes for Yang-Mills theory (gluons) and Einstein Gravity (gravitons), and it has successfully computed planar amplitudes for massive cubic scalar theories—a major step beyond previous massless-only constraints.

Future Directions

While the recursion method is a significant advancement, challenges remain. As the number of particles ($n$) increases, inverting the braid matrices becomes computationally expensive. Furthermore, this framework is currently limited to genus-zero surfaces (tree-level amplitudes). Extending this localization and recursion to higher-genus surfaces (loop-level amplitudes) and incorporating supersymmetry more directly into the twisted forms are the next frontiers in this geometric approach to quantum field theory.

Summary updated

BibTeX
@inproceedings{mizera2020aspects,
  author = {Sebastian Mizera},
  title = {{Aspects of Scattering Amplitudes and Moduli Space Localization}},
  booktitle = {Springer Theses},
  year = {2020},
  doi = {10.1007/978-3-030-53010-5},
}

Loading related content...

Figure