Infrared and Collinear Safety in Jet Substructure

This direction is split into other directions: infrared and collinear safety. Its last version stays readable and citable here.

Version 1 of 117 papers30 conceptsfrozen 2026-10-04

In the high-energy environment of the Large Hadron Collider (LHC), quarks and gluons never appear in isolation. Instead, they manifest as "jets"—tightly packed sprays of particles. To understand the fundamental physics occurring at the heart of these collisions, physicists must look inside these jets to determine their origin. However, the quantum nature of these sprays introduces a significant mathematical hurdle: many intuitive ways of measuring a jet's shape or composition lead to infinite results in theoretical calculations.

The solution to this problem is a design principle known as Infrared and Collinear (IRC) safety. This principle ensures that a measurement remains stable even when the underlying physical process undergoes minor, inevitable fluctuations.

The Stability Requirement

At its core, IRC safety is about robustness against two specific types of perturbations:

  1. Infrared (Soft) Emissions: The addition of an extremely low-energy particle should not change the value of the observable.
  2. Collinear Splittings: Replacing one particle with two particles traveling in the exact same direction (sharing the original energy) should not change the value of the observable.

As noted by C. M. Buttar, Gavin P. Salam, and colleagues in their benchmarking report, legacy algorithms often failed this test. If a single soft particle can "bridge" two separate jets and cause them to merge, the algorithm is IRC unsafe. This instability makes it impossible to perform high-precision perturbative calculations, as the theory predicts divergent (infinite) probabilities for these soft and collinear emissions.

Building Robust Observables

To move beyond simple jet counting, researchers developed observables that characterize the "internal" shape of a jet while maintaining IRC safety. A landmark example is $N$-subjettiness, introduced by Jesse Thaler and Ken Van Tilburg in Identifying boosted objects with N-subjettiness.

Instead of trying to find discrete sub-jets, $N$-subjettiness calculates an inclusive sum over all jet constituents. Because it uses an energy-weighted sum of distances to candidate axes, the addition of a zero-energy particle adds zero to the sum, and a collinear split doesn't change the energy-weighted distance. This design ensures the metric is robust against the "noise" of the quantum vacuum.

Similarly, Andrew J. Larkoski, Gavin P. Salam, and Jesse Thaler developed Energy Correlation Functions (ECFs). These functions avoid the need for jet axes entirely, instead summing over products of particle energies and their angular separations. This approach provides a "prong-like" signature for heavy particles (like $W$ or Higgs bosons) that is inherently safe from soft radiation.

Grooming and Sudakov Safety

Sometimes, the most interesting physics is hidden behind "unsafe" variables. For instance, the number of particles in a jet (multiplicity) is highly sensitive to soft emissions and is thus IRC unsafe.

To bridge this gap, researchers use grooming. The Soft Drop algorithm, developed by Andrew J. Larkoski and colleagues in Soft drop, recursively prunes wide-angle, low-energy radiation from a jet. By removing the "soft" components that cause divergences, grooming allows for the calculation of properties that were previously considered theoretically inaccessible.

This leads to the concept of Sudakov safety, where an observable might be unsafe in standard perturbation theory but becomes well-defined when grooming or all-orders resummation is applied. As demonstrated in Exposing the QCD Splitting Function with CMS Open Data, grooming allows physicists to directly measure fundamental constants of the Standard Model by creating a "clean" environment where theoretical infinities cancel out.

Geometry and Machine Learning

As the field moves toward more complex analysis techniques, the requirement for IRC safety remains the primary constraint.

  • Optimal Transport: In SHAPER, Demba E. Ba, Jesse Thaler, and others use the Wasserstein metric (Earth Mover's Distance) to compare jets to idealized shapes. Because this metric naturally accounts for the "work" required to move energy, it is inherently IRC safe.
  • Spectral Clustering: When adapting machine learning for jet reconstruction, G. Cerro and colleagues in Spectral Clustering for Jet Reconstruction found that a specific mathematical scaling factor was necessary to ensure the algorithm didn't produce erratic results when a particle split.
  • The Black Box Problem: Alexis Romero and D. Whiteson have warned in Safety of Quark/Gluon Jet Classification that high-performing deep learning models often rely on IRC-unsafe features. If a model learns to distinguish quarks from gluons based on "hit counts" in a detector, it may be exploiting simulation artifacts rather than fundamental physics.

Hidden Instabilities

Even well-established algorithms can harbor secrets. Recent work by Fabrizio Caola, Gavin P. Salam, and Jesse Thaler in Flavoured jets with exact anti-kt kinematics used automated testing to show that jet "flavor" (identifying the specific type of quark) is notoriously difficult to define safely. They found that while an algorithm might appear safe at low orders of calculation, complex radiation patterns at higher orders can trigger hidden instabilities, necessitating even more rigorous mathematical frameworks for the next generation of LHC physics.

Papers in this version

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_infrared_and_collinear_safety_in_jet_sub_9ac8912c,
  title = {Infrared and Collinear Safety in Jet Substructure},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/infrared-and-collinear-safety-in-jet-substructure/rdr_9ac8912c80654bfc9d833a6e09297b96},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-22},
}

This exact version (version 1, frozen text):

@misc{lacuna_infrared_and_collinear_safety_in_jet_sub_9ac8912c_v1,
  title = {Infrared and Collinear Safety in Jet Substructure},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/infrared-and-collinear-safety-in-jet-substructure/rdr_9ac8912c80654bfc9d833a6e09297b96/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-22},
}