🔬 physics

The double-real part of NNLO matching with logarithmically accurate showers for $e^+e^-$ collisions

This paper presents a logarithmically accurate NNLO matching procedure for $e^+e^-$ collisions using PanScales showers that incorporates leading-colour double-real contributions and new inversion techniques to preserve exponentiation while avoiding spurious super-leading logarithms.

Original authors: Melissa van Beekveld, Keith Hamilton, Gavin P. Salam, Nicolas Schalch, Gregory Soyez, Silvia Zanoli

Published 2026-10-30

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Melissa van Beekveld, Keith Hamilton, Gavin P. Salam, Nicolas Schalch, Gregory Soyez, Silvia Zanoli

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 smash particles together at speeds close to that of light to uncover the fundamental rules of nature. To understand what happens in these collisions, researchers rely on two powerful tools that usually work in separate lanes. The first is a precise calculation of the initial, violent crash, which tells them exactly how many new particles are created and how much energy they carry. The second is a simulation of the aftermath, a process where these new particles rapidly split and multiply into sprays of other particles before settling down. This splitting process is chaotic and happens in billions of tiny steps, so physicists use a method called a "parton shower" to model it, much like a weather forecast predicts a storm by tracking how air moves in layers. The challenge has always been to stitch these two tools together seamlessly. If the precise calculation and the simulation do not agree perfectly at the point where they meet, the entire prediction can develop hidden errors that grow larger and larger, eventually making the results unreliable.

A team of researchers has now developed a new way to join these two tools for electron-positron collisions, ensuring that the transition between the precise calculation and the simulation is smooth and free of these hidden errors. Their work focuses on a specific and difficult scenario: when the initial crash produces not just one or two new particles, but a complex spray of four. In the past, trying to match the precise math for these four-particle events with the simulation often introduced subtle distortions that broke the mathematical rules governing how the particles behave. The researchers found that to fix this, the simulation must be forced to mimic the precise calculation's behavior in every possible way, even in the most extreme and rare conditions where particles move very slowly or very close together. They achieved this by creating a new set of instructions that allow the simulation to run backward, reconstructing the exact path the particles took to get to their current state. This backward step is crucial because it lets the researchers check if the simulation is telling the truth at every single moment.

The team applied these new instructions to a sophisticated simulation known as the PanScales shower, which is designed to be highly accurate. They constructed a matching procedure that handles the first two emissions of particles with perfect precision. By using a technique called "disordering," they allowed the simulation to explore paths that it would normally skip, ensuring that no part of the possible outcome was left out. They also introduced a system of "swaps," where the simulation can occasionally relabel particles to ensure that the final mix of outcomes matches the exact mathematical prediction. This approach prevents the simulation from generating fake signals that look like real physics but are actually just artifacts of a poor match. The result is a simulation that is accurate for events with two, three, or four jets of particles, all while preserving the correct logarithmic accuracy that ensures the predictions remain reliable even when looking at very rare, soft emissions.

While this work represents a major step forward, the researchers are careful to note that it is not the final solution for every possible scenario. Their method works perfectly for the leading contributions to the four-particle events, but it currently cannot handle a specific, more complex interaction involving three particles splitting at once, a situation that existing simulations do not yet fully understand. Because of this missing piece, the match is not yet perfect in every single corner of the physical landscape. However, the team has proven that their method works as a powerful proof of concept. They have demonstrated that by carefully aligning the simulation with the precise calculation, it is possible to eliminate the spurious errors that have plagued previous attempts. This achievement lays the necessary groundwork for future calculations that will eventually include the missing pieces, bringing physicists closer to a complete and flawless description of how matter behaves at the smallest scales.

Problem Statement

The central challenge addressed is the construction of a matching procedure between fixed-order calculations and parton showers that preserves the logarithmic accuracy of the shower while correctly reproducing higher-order matrix elements. Specifically, the authors focus on matching the first two real emissions (the double-real, $O(\alpha_s^2)$ , contribution) in $e^+e^-$ collisions.

Previous work established that if a matching scheme (specifically one using a Hard Emission Generator, or HEG) does not perfectly align with the parton shower's behavior in the infrared (soft and collinear) limits, it generates spurious "super-leading logarithms" (SLLs). These are unphysical terms of the form $\alpha_s^n L^m$ where $m > n$ , which can break exponentiation and spoil the accuracy of single-logarithmic observables. While this issue was understood for NLO matching (one emission), extending this to NNLO (two emissions) with logarithmically accurate showers (like the PanScales showers) presents new technical hurdles:

Phase Space Coverage: Standard ordered shower histories often fail to cover the entire phase space for multi-parton final states (e.g., ~20% of the 4-parton phase space is inaccessible via strictly ordered emissions in PanGlobal).
Boundedness: Reweighting factors required to match the exact matrix element can become unbounded, leading to inefficiencies or failures in standard Monte Carlo acceptance steps.
Triple-Collinear Limits: Current logarithmically accurate showers (including PanScales) do not yet incorporate the correct triple-collinear splitting structure. This creates a fundamental misalignment with the double-real matrix elements in that specific limit, which the authors acknowledge as a source of potential SLLs.

Methodology

The authors develop a nested multiplicative matching procedure for the PanGlobal and PanLocal showers. The core philosophy is to let the shower generate emissions and then apply an acceptance probability to reproduce the exact leading-colour tree-level matrix elements for four-parton final states.

Key technical components include:

1. Inversion of Shower Maps and Jacobians:
To perform multiplicative matching, one must reconstruct the pre-branching configuration from the post-branching momenta to calculate the shower's effective matrix element. The paper derives:

Analytic inversion formulas for the PanGlobal and PanLocal kinematic maps for any number of emissions.
Explicit expressions for the Jacobian relating the shower variables (ordering variable, rapidity, azimuth) to the Lorentz-invariant phase space measure.
These allow the calculation of the shower's branching probability density for any given phase-space point.

2. Handling Phase Space Coverage (Disordering):
To address the issue of inaccessible phase-space regions, the authors implement disordering. Instead of strictly enforcing the shower ordering variable $v_{i+1} < v_i$ , the algorithm allows emissions to be generated with a slightly "disordered" scale ( $v_{i+1} < v_i + \Delta$ ).

The exact matrix element is partitioned into "ordered" ( $e_{ord}$ ) and "disordered" ( $e_{dis}$ ) contributions.
If the ordered shower path underestimates the matrix element (or cannot reach the point), the disordered path fills the gap.
This ensures full phase-space coverage without double-counting, provided the disordering parameter $\Delta$ is chosen carefully to remain logarithmically safe.

3. Handling Boundedness (Grouping and Swaps):
To ensure the reweighting factors remain bounded and to correct for flavor/momentum permutations, the authors employ grouping and swaps:

Grouping: Instead of treating a single color-ordered configuration in isolation, the algorithm considers a group of configurations (e.g., permutations of momenta or flavors like $\{g_a, g_b\}$ or $\{q'\bar{q}'\}$ ).
Swaps: The algorithm generates events based on the sum of the group's shower matrix elements. If a specific channel within the group is overpopulated by the shower relative to the exact matrix element, events are probabilistically "swapped" to underpopulated channels.
Minimal Swaps: For the PanLocal shower, the authors introduce a "minimal swap" scheme. Unlike the "always active" swaps used for PanGlobal, these are only activated when necessary to prevent the swapping of distinct topologies that differ in their Sudakov form factors at the single-logarithmic level, which would otherwise break logarithmic accuracy.

Key Contributions

New Inversion Formulas: The paper provides the first analytic inversion and Jacobian calculations for the PanGlobal and PanLocal showers for arbitrary numbers of emissions, enabling them to function as full phase-space generators.
Double-Real Matching Framework: It constructs a working framework for matching the first two real emissions of a shower to the exact leading-colour $O(\alpha_s^2)$ matrix elements.
Resolution of Technical Obstacles: It successfully integrates disordering and swap mechanisms to handle phase-space coverage and boundedness issues in a way that is compatible with logarithmic accuracy.
Validation of Principles: The work validates the principle that perfect alignment between the matching generator and the shower in the infrared is necessary to avoid spurious SLLs.

Results

The authors validate their procedure through several tests:

Fixed-Order Tests: The matched shower reproduces the exact leading-colour tree-level matrix elements for four-parton final states.
Double-Unresolved Limits: The procedure correctly reproduces the double-soft limits of the shower.
All-Order Tests: The matching preserves the shower's double-soft evolution.
Limitations: The authors explicitly demonstrate that because the PanScales showers lack the triple-collinear splitting structure, the matching does not perfectly align in the triple-collinear limit. Consequently, while the method avoids SLLs in the double-soft limit, sources of SLLs remain in the triple-collinear limit.

Significance and Claims

The paper positions itself as a "powerful proof of principle" and an "important stepping stone" towards full NNLO+NNLL matching.

Modest Scope: The authors explicitly state that this is not the final word on 4-parton matching. They acknowledge that a complete solution requires showers that incorporate full double-soft and triple-collinear structures, which are not yet available.
Achievement: Despite the caveat regarding triple-collinear limits, the work achieves a significant advance by establishing a matching procedure that reproduces the exact leading-colour double-real contribution while preserving the shower's double-soft evolution.
Outcome: The resulting shower is simultaneously NLO accurate for 2-jet events, LO accurate for 3-jet events, and LO accurate for 4-jet events (for unoriented events).
Future Work: The authors note that phenomenological results and a discussion of shower uncertainties will be presented in a forthcoming companion paper.

In summary, the paper provides the necessary technical infrastructure (inversions, Jacobians, and matching algorithms) to move beyond NLO matching with logarithmically accurate showers, while honestly delineating the current limitations imposed by the lack of triple-collinear accuracy in existing shower algorithms.

Problem Statement

Methodology

Key Contributions

Results

Significance and Claims

More like this