A Precise Determination of $\alpha_s$ from the Heavy Jet Mass Distribution

This paper presents a precise determination of the strong coupling constant $\alpha_s(m_Z) = 0.1148^{+ 0.0015}_{-0.0022}$ through a global fit of $e^+e^-$ heavy jet mass data, utilizing state-of-the-art theoretical predictions that combine fixed-order calculations, multiple resummation orders, and first-principles power corrections to demonstrate the critical role of resummation in achieving robust results and revealing evidence for negative power corrections in the trijet region.

The Gist

Imagine you are trying to measure the strength of a very sticky glue that holds the tiniest building blocks of the universe together. In physics, this "glue" is called the strong force, and the measure of its strength is a number called $\alpha_s$ (alpha-s).

For decades, physicists have tried to measure this number by smashing particles together and watching how they fly apart. One way to do this is to look at a specific shape formed by the debris, called the "Heavy Jet Mass." Think of it like watching a firework explode and measuring how heavy the main chunk of the explosion is compared to the sparks flying off.

However, there was a problem. When physicists looked at the Heavy Jet Mass, their calculations kept giving them a value for the glue's strength that was too low compared to other methods. It was like trying to weigh a gold bar using a scale that kept telling you it was lighter than it actually was.

The Problem: The "Bumpy Road" of Math

The universe at this tiny scale is messy. To calculate the Heavy Jet Mass, physicists use a mathematical map. But this map has two tricky areas:

1. The "Two-Lane" Highway (Dijet Region): This is where the particles mostly fly in two main streams. The math here is tricky because of huge, swirling numbers (logarithms) that make the prediction wobble.
2. The "Three-Lane" Shoulder (Trijet Region): Sometimes, the particles split into three streams. This creates a weird "bump" or "shoulder" in the data. Previous calculations ignored this bump, causing the map to be inaccurate.

Furthermore, the particles don't just exist as pure math; they are made of real stuff (hadrons) that have mass. This adds a "power correction"—a small nudge to the data that previous models got wrong.

The Solution: A Better Map and a New Strategy

The authors of this paper built a much more sophisticated map to fix these issues. Here is how they did it, using some everyday analogies:

1. Smoothing the Bumpy Road (Resummation)
Imagine driving down a road with huge potholes (the mathematical "logarithms"). If you drive fast (standard math), you crash. The authors used a technique called "Resummation." Think of this as building a smooth, paved bridge over the potholes. This allowed them to drive smoothly through the "Two-Lane" and "Three-Lane" regions without the math breaking down.

2. Accounting for the "Shoulder"
They realized that the "Three-Lane" bump (the shoulder) was real and important. They added a special section to their map specifically for this bump. Without this, the map was missing a whole neighborhood.

3. The "Random Walk" for Uncertainty
In science, we never know the exact answer; we only know a range of possibilities. Usually, physicists guess the range by changing their numbers a few times. The authors used a smarter method called a "Flat Random Scan."

• The Analogy: Imagine trying to find the highest point in a foggy mountain range. Instead of just checking a few spots, they generated 5,000 random paths across the entire mountain. By looking at all these paths, they could create a perfect "fog map" showing exactly where the uncertainty lies and how different parts of the map are connected. This ensured they didn't miss any hidden valleys or peaks in their error estimates.

The Discovery: A Negative Nudge

One of the most surprising findings was about the "power correction" (the nudge from the real mass of the particles).

• Old Idea: Everyone thought this nudge pushed the data in one direction (to the right).
• New Discovery: When they included the "Shoulder" math, they found that in the "Three-Lane" region, the nudge actually pushes the data in the opposite direction (to the left).
• The Metaphor: It's like driving a car. In the straight part of the road, the wind pushes you slightly right. But when you hit the curve (the shoulder), the wind suddenly pushes you left. If you ignore the curve, you'll crash. The authors found that you must include the curve to see this leftward push.

The Result: A Precise Measurement

By combining the smooth bridge (resummation), the shoulder map, and the 5,000-path fog scan, they finally got a clear picture.

• The Value: They determined the strength of the strong force to be 0.1148.
• The Confidence: This number is very precise and matches what other methods (like measuring "Thrust") have found.
• The Lesson: The most important takeaway is that you cannot get a good answer without the "bridge" (resummation). Without it, the answer changes wildly depending on which part of the road you choose to measure. With the bridge, the answer stays the same no matter where you look.

Summary

This paper is like a team of cartographers who finally fixed a broken map of a complex city. They realized that previous maps missed a specific neighborhood (the shoulder) and didn't account for the terrain's bumps (resummation). By building a better map and using a new method to estimate the fog (uncertainty), they finally found the exact location of the "Strong Force" treasure, confirming that the universe is consistent, provided you look at it with the right tools.

Technical Summary

1. Problem Statement

The primary goal of the study is to perform a high-precision determination of the strong coupling constant, $\alpha_s(m_Z)$, using Heavy Jet Mass (HJM) event shape data from $e^+e^-$ collisions.

• Historical Context: Previous attempts to extract $\alpha_s$ from HJM data consistently yielded values significantly lower than those derived from Thrust distributions and the Particle Data Group (PDG) average.
• Theoretical Discrepancies:
  1. Sudakov Shoulder: Unlike Thrust, the HJM distribution exhibits a "left-Sudakov shoulder" as the variable $\rho \to 1/3$ due to large logarithms.
  2. Power Corrections: Recent studies suggested that non-perturbative power corrections shift the HJM distribution leftward (negative shift), whereas Thrust shifts rightward (positive shift).
  3. Factorization Differences: Power corrections in HJM moments depend on different non-perturbative parameters than the tail Operator Product Expansion (OPE), unlike in Thrust where they are identical.
• Challenge: Standard fixed-order perturbation theory (FO) fits to HJM data are highly sensitive to the chosen fit range (specifically the lower cutoff), leading to unreliable extractions of $\alpha_s$.

2. Methodology

The authors employ a global fit framework incorporating state-of-the-art theoretical predictions and a sophisticated treatment of uncertainties.

A. Theoretical Framework

The differential cross-section is modeled by factorizing the distribution into distinct kinematic regions:

1. Dijet Region ($\rho \to 0$):
   • Factorized into a hard function, two jet functions, and a soft/shape function.
   • Includes $N^3LL'$ (Next-to-Next-to-Next-to-Leading Logarithmic) resummation.
   • Non-perturbative effects are modeled via a shape function parameter $\Omega_1^\rho$ in the R-gap scheme (cancelling leading renormalons).
2. Trijet/Sudakov Shoulder Region ($\rho \to 1/3$):
   • Factorized into three jet functions and a soft function.
   • Includes $N^2LL$ resummation of Sudakov shoulder logarithms.
   • A numerical Fourier transform is required for this region, making it computationally expensive.
   • Power corrections are modeled via a shift parameter $\Theta_1$. The authors hypothesize a negative shift ($\Theta_1 < 0$) in this region.
3. Fixed-Order Matching:
   • The full prediction combines the resummed dijet and shoulder regions with $O(\alpha_s^3)$ (NNLO) fixed-order results.
   • Overlaps are handled using profile functions ($\mu_i(\rho)$) to smoothly transition scales between regions, avoiding double-counting of singular terms.

B. Uncertainty Treatment (Innovation)

A major methodological innovation is the treatment of theoretical uncertainties:

• Random-Scan Covariance Matrix: Instead of standard scale variations, the authors use a flat random scan over 16 theory parameters (profile parameters, renormalization scales, etc.).
• Procedure:
  1. Generate 5,000 sets of theory parameters sampled from uniform distributions.
  2. Compute predictions for all data points.
  3. Derive the mean prediction ($\bar{x}_i$) and the 1-$\sigma$ uncertainty ($\Delta_i^{theo}$).
  4. Calculate the correlation matrix $r_{ij}^{theo}$ based on the covariance of these random sets.
• Total Covariance: The theoretical covariance matrix is added to the experimental covariance matrix (which uses a minimal overlap model for systematic uncertainties) to form the total covariance used in the $\chi^2$ minimization.

C. Data Selection

• Datasets: 50 datasets from ALEPH, DELPHI, JADE, L3, OPAL, and SLD collaborations.
• Range: Center-of-mass energies $Q$ from 35 GeV to 207 GeV.
• Fit Region: Restricted to $3a_{peak}/Q \leq \rho \leq 0.3$.
  • Lower bound: Avoids the fully non-perturbative peak region.
  • Upper bound: Avoids unknown higher-order fixed-order contributions and non-perturbative effects in the shoulder.

3. Key Contributions

1. First-Principles Treatment of HJM: This is the first global fit to HJM data that simultaneously includes $N^3LL'$ dijet resummation, $N^2LL$ Sudakov shoulder resummation, and a first-principles treatment of power corrections in the dijet region.
2. Resolution of Fit-Range Sensitivity: The study demonstrates that resummation is essential. Without it, the extracted $\alpha_s$ depends linearly and overwhelmingly on the lower cutoff of the fit range. With resummation, the result is robust across a wide range of fit boundaries.
3. Evidence for Negative Power Correction: The analysis provides evidence for a negative power correction ($\Theta_1 < 0$) in the trijet region, but only when Sudakov shoulder resummation is included. This validates the theoretical prediction that HJM and Thrust have opposite non-perturbative shifts.
4. Advanced Uncertainty Quantification: The implementation of the flat random-scan covariance matrix allows for a more robust inclusion of correlated theoretical uncertainties, preventing bias in the fit results.

4. Results

The global fit yields the following value for the strong coupling constant:
$$\alpha_s(m_Z) = 0.1148^{+0.0015}_{-0.0022}$$

• Comparison: This result is compatible with determinations from Thrust ($\alpha_s \approx 0.1145$) and the C-parameter, and consistent with the PDG world average at the $1.8\sigma$ level.
• Non-Perturbative Parameters:
  • Dijet shape parameter: $\Omega_1^\rho = 0.61 \pm 0.08$ GeV.
  • Trijet shift parameter: $\Theta_1 = -0.46 \pm 0.17$ GeV.
• Model Comparison:
  • Fixed-Order Only: Yields $\alpha_s \approx 0.1166$ with large fit-range uncertainty.
  • Dijet Resummation Only: Reduces uncertainty but still shows some sensitivity.
  • Full Model (FO + Dijet + Shoulder): Provides the best $\chi^2/\text{dof} = 1.039$ and the most stable result.
• Uncertainty Breakdown: The dominant uncertainty arises from the non-perturbative dijet parameter $\Omega_1^\rho$. The uncertainty due to the fit range choice is subdominant in the resummed models.

5. Significance

• Validation of QCD Factorization: The successful extraction of $\alpha_s$ using a complex factorization formula that unifies dijet and trijet physics confirms the validity of Soft-Collinear Effective Theory (SCET) descriptions for HJM.
• Resolution of the "HJM Puzzle": The paper resolves the long-standing discrepancy where HJM data yielded low $\alpha_s$ values. The discrepancy was caused by the omission of Sudakov shoulder resummation and the incorrect assumption of power correction signs in previous analyses.
• Methodological Benchmark: The use of random-scan covariance matrices for theoretical uncertainties sets a new standard for precision QCD fits, ensuring that correlated theory errors do not artificially inflate or deflate the significance of the fit.
• Future Directions: The detection of a negative power correction in the trijet region motivates further first-principles studies of non-perturbative effects in the symmetric three-jet limit.

In conclusion, this work represents a state-of-the-art determination of $\alpha_s$ from HJM, demonstrating that with complete resummation and proper uncertainty handling, HJM is a competitive and precise observable for testing the Standard Model.