Threshold resummation for $Z$-boson pair production at NNLO+NNLL

This paper presents the threshold resummation of large logarithms up to NNLL accuracy for on-shell $Z$-boson pair production at the LHC, demonstrating that matching these resummed predictions to NNLO fixed-order results significantly reduces scale uncertainties and provides a precise description of the invariant mass distribution in the high-energy regime.

The Gist

Imagine the Large Hadron Collider (LHC) as a massive, high-speed particle racetrack. Scientists smash protons together at incredible speeds to see what happens. One of the most important things they look for is a pair of "Z-bosons" (think of them as heavy, invisible messengers that carry the weak nuclear force). Finding these pairs helps scientists check if the Standard Model of physics is working correctly and look for any hidden "new physics" that might be lurking.

However, predicting exactly how often these Z-boson pairs appear is like trying to predict the exact weather in a hurricane. The math is incredibly complex.

Here is a simple breakdown of what this paper does, using some everyday analogies:

1. The Problem: The "Traffic Jam" at the Edge

When two protons collide, they produce Z-bosons. Sometimes, the collision happens right at the very limit of what the energy allows. In physics terms, this is called the "threshold."

Imagine you are driving a car up a hill. If you have just enough gas to reach the very top, you might sputter and stall right at the peak. In the world of particle physics, when the collision energy is just barely enough to create the heavy Z-bosons, the math gets messy. You get huge "logarithms" (mathematical numbers that get very large) that make the predictions unreliable. It's like trying to hear a whisper in a room full of screaming fans; the signal gets drowned out by the noise.

2. The Solution: "Resummation" (Cleaning Up the Noise)

The authors of this paper developed a method called threshold resummation.

Think of the calculation as a recipe.

• The Old Way (Fixed Order): Scientists used to calculate the recipe step-by-step. They calculated the main ingredients (Leading Order), then added a pinch of spice (Next-to-Leading Order), and then a dash more (Next-to-Next-to-Leading Order or NNLO). But at the very top of the energy hill, the "noise" (the large logarithms) was so loud that even adding more spices didn't fix the taste.
• The New Way (Resummation): Instead of just adding spices one by one, the authors realized that the "noise" follows a pattern. They figured out how to group all those noisy terms together and "resum" them (add them up in a smarter way) to cancel out the chaos. They did this up to a very high level of precision called NNLL (Next-to-Next-to-Leading Logarithmic).

It's like realizing that the screaming fans are actually singing a specific song. Once you know the song, you can tune your radio to cancel out the noise and hear the whisper clearly.

3. The Challenge: A Heavy Load

The authors note that doing this for Z-boson pairs is much harder than for other particles (like the Higgs boson or pairs of light electrons).

• The Analogy: Imagine trying to balance a stack of plates. Balancing one plate (a single particle) is hard. Balancing two heavy plates (two Z-bosons) that are wobbling is much harder.
• Because there are two heavy particles in the final state, the math requires "two-loop" calculations (very complex virtual interactions). This made the numerical computation a "non-trivial task," meaning it required significant computer power and clever coding to get right.

4. The Results: Sharper Predictions

After doing all this heavy lifting, the authors compared their new, super-precise predictions with the older, less precise ones.

• The "K-Factor" (The Boost): They found that at high energies (around 1 TeV, which is 1,000 times the mass of a proton), the old calculations underestimated the production rate by a lot (up to 83% higher than the simplest guess). Their new method added a little extra boost on top of that, increasing the predicted number of Z-boson pairs by about 4%.
• The "Uncertainty" (The Margin of Error): In science, every prediction comes with a "margin of error."
  • The old method had an uncertainty of about 3.4%.
  • The new method (NNLO+NNLL) reduced this uncertainty to about 2.6%.
  • Analogy: Imagine trying to hit a target with a bow and arrow. The old method said, "You'll hit within 3.4 meters of the bullseye." The new method says, "You'll hit within 2.6 meters." It's a small difference, but in the world of high-energy physics, that extra precision is huge.

5. Why It Matters

The paper concludes that their new, more precise predictions match what the ATLAS and CMS experiments (the giant detectors at the LHC) are actually seeing.

• The Takeaway: By cleaning up the mathematical "noise" at the energy limits, the scientists have provided a clearer map for future experiments. This helps ensure that if scientists do find something weird or new in the future, they can be sure it's not just a mistake in their math.

In short: The authors took a very messy, difficult math problem involving heavy particle collisions, figured out a way to clean up the noise at the energy limits, and produced a sharper, more reliable prediction that matches what we see in the real world.

Technical Summary

Technical Summary: Threshold Resummation for Z-boson Pair Production at NNLO+NNLL

Problem Statement
The production of on-shell Z-boson pairs ($ZZ$) at the Large Hadron Collider (LHC) is a critical process for precision Standard Model (SM) tests, electroweak symmetry breaking studies, and searches for new physics. While Leading Order (LO) predictions are unreliable due to large theoretical uncertainties, Next-to-Leading Order (NLO) and Next-to-Next-to-Leading Order (NNLO) calculations have been established. However, in the partonic threshold limit ($z \to 1$, where $z = Q^2/s$), large logarithmic terms of the form $\ln^k(1-z)$ arise. These terms dominate the cross-section in the high invariant mass region and require resummation to all orders in the strong coupling constant $\alpha_s$ to achieve high precision. While threshold resummation for Drell-Yan (DY) and Higgs production has reached N$^3$LO+N$^3$LL accuracy, achieving similar precision for $ZZ$ production is more challenging due to the presence of two massive particles in the final state, which complicates the virtual corrections compared to massless or single-mass processes.

Methodology
The authors perform threshold resummation for $ZZ$ production up to Next-to-Next-to-Leading Logarithmic (NNLL) accuracy in QCD. The theoretical framework involves:

1. Factorization: The hadronic cross-section is expressed as a convolution of parton distribution functions (PDFs) and partonic cross-sections. The partonic cross-section is separated into a soft-virtual (SV) part, containing singular terms as $z \to 1$, and a regular part.
2. Mellin Space Resummation: The resummation is performed in Mellin ($N$) space, where convolutions become products. The SV part is organized as $\hat{\sigma}_N \equiv g_0 \exp(G_N)$, where $G_N$ resums the large logarithms ($\ln N$) and $g_0$ is a process-dependent coefficient.
3. Calculation of Coefficients:
   • The universal exponent $G_N$ is constructed using known anomalous dimensions (cusp and non-cusp) up to NNLL.
   • The process-dependent coefficient $g_0$ is computed up to two-loop order ($g_{02}$). This requires the evaluation of two-loop virtual amplitudes ($M^{(0,2)}$) and one-loop squared amplitudes ($M^{(1,1)}$). The authors utilized the VVamp package for two-loop amplitudes and in-house FORM codes for algebraic manipulation and handyG for numerical evaluation of generalized polylogarithms.
   • The infrared pole structure was verified using Catani's $I$-operator formalism.
4. Matching: The resummed predictions are matched to the fixed-order NNLO results (calculated using the MATRIX package) to avoid double counting. The matching formula subtracts the truncated resummed expansion from the fixed-order result.
5. Numerical Implementation: The Mellin inversion is performed using the minimal prescription to handle the Landau pole. The analysis uses MSHT20 PDFs and considers center-of-mass energies of 13.0 TeV, 13.6 TeV, and a future 100 TeV collider.

Key Contributions

• First NNLO+NNLL Calculation: This work presents the first calculation of $ZZ$ production cross-sections matched to NNLO fixed-order results with NNLL threshold resummation.
• Handling Massive Final States: The authors successfully navigated the computational complexity introduced by the two massive Z-bosons, specifically the non-trivial two-loop virtual contributions required for the $g_0$ coefficient, which are more difficult than in DY or Higgs production.
• Phenomenological Analysis: The paper provides detailed predictions for the invariant mass distribution ($d\sigma/dQ$) and total inclusive cross-sections, including a systematic study of scale uncertainties and the impact of the gluon fusion channel.

Results

• Cross-Section Enhancement: In the high invariant mass region ($Q = 1$ TeV), NNLO corrections relative to LO are substantial (approx. 83% at 13.6 TeV). The NNLL resummation provides an additional enhancement of approximately 4% to the NNLO cross-section at 13.6 TeV.
• Scale Uncertainty Reduction: The inclusion of NNLL resummation reduces the theoretical scale uncertainties. For the invariant mass region $Q = 1.3$ TeV, the uncertainty decreases from 3.4% at NNLO to approximately 2.6% at NNLO+NNLL. The authors note that this reduction is expected to be more significant at higher $Q$ values.
• Convergence: The K-factors indicate a good convergence of the perturbative series. The gap between fixed-order and resummed results narrows as the order increases, suggesting the stability of the prediction.
• Comparison with Experiment: The NNLO+NNLL predictions for the total cross-section at 13 TeV agree well with experimental measurements from ATLAS and CMS within reported uncertainties.
• Gluon Fusion: The study quantifies the contribution of the gluon fusion channel ($gg \to ZZ$), noting it contributes about 9.3% of the LO cross-section at 13.6 TeV. The scale uncertainties for the pure quark-annihilation channel ($NNLO_{q\bar{q}}$) are found to be lower than the full NNLO results due to the opening of the gluon channel at higher orders.

Significance
The paper claims that these results provide a necessary foundation for future precision studies of diboson processes at the LHC and future high-energy colliders (such as the FCC-hh). By reducing theoretical scale uncertainties and improving the accuracy of predictions in the high invariant mass region, the work enables more stringent tests of the SM and better sensitivity to potential new physics signals. The authors emphasize that while the NNLL contribution is a few percent, it is non-negligible for the high-precision era of the High-Luminosity LHC (HL-LHC) and future colliders where integrated luminosities will reach $3000-4000 \text{ fb}^{-1}$. The work also serves as a stepping stone toward even higher-order calculations (N$^3$LO) for processes involving massive final states.