Phase-space integrals through Mellin-Barnes representation

This paper presents an analytical computation of angular phase-space integrals with three and four denominators using dimensional regularisation and the Mellin-Barnes representation, expressing the results in terms of Goncharov polylogarithms and deriving recursion relations to facilitate the solution of full phase-space integrals.

The Gist

Imagine you are trying to bake the perfect cake (a prediction for a particle physics experiment). To do this, you need to mix two main ingredients: the recipe (how particles interact) and the kitchen space (where the particles are allowed to move).

In the world of quantum physics, calculating exactly how particles scatter and interact is like trying to measure every possible path a fly could take inside a giant, invisible, multi-dimensional room. This "room" is called Phase Space.

The paper you shared is about a new, super-efficient way to calculate the volume of this room when there are three or four specific obstacles (denominators) inside it. Here is how the authors did it, explained simply:

1. The Problem: A Messy, Infinite Room

Physicists need to calculate these "room volumes" to predict what happens when particles smash together (like at the Large Hadron Collider).

• The Difficulty: The math involved is incredibly messy. It's like trying to count every grain of sand on a beach while the beach is constantly shifting shape. The equations often blow up into infinity (singularities) unless you handle them with extreme care.
• The Old Way: Previous methods were like trying to solve a giant jigsaw puzzle by looking at one piece at a time. It worked for simple puzzles (few obstacles), but as soon as you added more pieces (more particles), the puzzle became impossible to finish analytically.

2. The Solution: The "Mellin-Barnes" Magic Lens

The authors used a mathematical tool called the Mellin-Barnes (MB) representation.

• The Analogy: Imagine you have a tangled ball of yarn (the complex integral). Trying to untangle it by pulling on the ends is a nightmare. The MB representation is like a magic lens that, when you look through it, instantly untangles the yarn into a neat, straight line of beads.
• What it does: It converts a terrifying, multi-dimensional math problem into a series of simpler, one-dimensional problems that look like a string of beads (integrals).

3. The Four-Step Cooking Process

The authors developed a step-by-step recipe to turn these "beads" into a final, usable answer:

1. The Safety Check (Analytic Continuation): As they zoom in on the math, some "beads" (poles) might try to jump over the line where they aren't supposed to be. The authors carefully track these jumps and catch them, turning them into smaller, manageable sub-problems. It's like a traffic controller rerouting cars that are about to crash.
2. The Zoom-In (Expansion): Once the path is clear, they expand the math to see the fine details. This is like zooming in on a map to see the street names.
3. The Transformation (Real Integrals): They convert the abstract "beads" back into real-world numbers (integrals over simple ranges like 0 to 1). This turns the abstract math into something you can actually calculate on a computer.
4. The Final Polish (GPLs): The result is expressed in a special language called Goncharov Polylogarithms (GPLs).
   • Why this matters: Think of GPLs as "Lego bricks." Once you have your answer built out of these specific bricks, you can easily snap them together with other parts of the physics problem (the "radial" part) to build the full cake. Other methods used "clay" (Clausen functions), which is hard to shape and snap together.

4. The Results: Solving the Impossible

The team successfully solved the math for rooms with 3 obstacles and 4 obstacles.

• 3 Obstacles: They solved it completely, even for cases where particles have mass (weight).
• 4 Obstacles: This was a massive breakthrough. They solved a 6-fold and 7-fold integral (math with 6 or 7 dimensions of complexity). Before this, no one had managed to write down the exact answer for this in the entire history of physics literature.

5. The "Partial Fraction" Shortcut

What if there are more than 4 obstacles?

• The Trick: The authors realized that if you have a room with 5 or 6 obstacles, you can break it down into a sum of simpler rooms with just 1 or 2 heavy obstacles. It's like saying, "To measure a giant, complex house, I don't need a new tool; I just need to measure a few simple rooms and add them up."

6. Why Should You Care?

• Speed: The old way of checking these numbers took about 30 minutes per calculation. Their new method, using these "Lego bricks" (GPLs), takes about 1 second. That's a 1,800x speedup!
• Precision: Because the math is now "analytic" (exact formulas) rather than just approximations, physicists can make much more precise predictions about the universe.
• Future Proof: This method is scalable. If they need to solve a room with 5 or 6 obstacles later, they just need more computer power, not a new theory.

In a nutshell: The authors found a way to untangle the most complex knots in particle physics math, turning a 30-minute nightmare into a 1-second calculation, and provided a universal "Lego set" that allows physicists to build perfect models of how the universe works.

Technical Summary

1. Problem Statement

In perturbative Quantum Chromodynamics (QCD), high-order predictions require the evaluation of real-emission contributions. These involve integrating matrix elements over the phase space (PS), which generically produces infrared singularities (regulated via dimensional regularisation, $d = 4 - 2\epsilon$) and finite pieces.

While multi-loop virtual corrections are well-understood, the analytic treatment of real-emission phase-space integrals remains challenging. In a specific reference frame, these integrals factorize into a process-dependent radial part and a universal angular part. The angular integral with $n$ denominators is defined as:
$$\Omega_{j_1, \dots, j_n}(\{p_i^\mu\}, d) \equiv \int d\Omega_{d-1}(q) \frac{1}{(p_1 \cdot q)^{j_1} \cdots (p_n \cdot q)^{j_n}}$$
where $p_i$ are external reference momenta (massless or massive) and $q$ is a massless vector.

• Current State: Exact results for $n=1, 2$ are known in terms of hypergeometric functions. For $n \ge 3$, exact all-orders-in-$\epsilon$ representations exist via multivariable $H$-functions, but expanding them into a Laurent series in $\epsilon$ for general kinematics is non-trivial.
• Gap: Previous results for $n=3$ existed in different function bases (Clausen functions vs. Goncharov Polylogarithms). Clausen functions lack the iterative integration properties required to easily convolve the angular part with the radial component. There were no known analytic results for $n=4$ angular integrals in terms of Goncharov Polylogarithms (GPLs).

2. Methodology: Mellin-Barnes (MB) Representation

The authors employ a fully algorithmic Mellin-Barnes (MB) representation approach to convert multifold MB integrals into real parametric integrals, ultimately expressing results in terms of GPLs.

The Four-Step Strategy:

1. Analytic Continuation: The MB integral involves $\Gamma$-functions whose poles can cross the integration contour as $\epsilon \to 0$. The authors track these pole crossings, collecting residues (which become lower-fold MB integrals) iteratively until the integrand is safe to expand.
2. $\epsilon$-Expansion: Once contours are clear, the integrand is expanded in a Laurent series in $\epsilon$. Coefficients are simpler MB integrals.
3. Reduction to Real Parametric Integrals: The resulting MB integrals are "balanced" (equal numbers of $\Gamma(a+z)$ and $\Gamma(a-z)$ factors). These are converted into Euler beta functions and then integral representations. The remaining MB integrations collapse via delta-function constraints ($\int (dz/2\pi i) A^z = \delta(1-A)$), leaving ordinary integrals over parameters $x_i \in [0, 1]$.
4. Expression in Terms of GPLs: The resulting integrands are rational functions and logarithms.
   • Rationalization: If quadratic radicands ($\sqrt{ax^2+bx+c}$) obstruct integration, a specific substitution ($x \to \frac{b+2c\eta}{c\eta^2-a}$) rationalizes the square root.
   • Iterated Integration: Variables are shifted to the rightmost position in GPL arguments using in-house algorithms, allowing the integrals to be expressed as iterated integrals (GPLs).

Handling Mass Configurations:

• Multi-massive cases: Instead of increasing MB dimensionality, the authors use Partial Fraction (PF) decomposition. By exploiting the linearity of the angular integral in parameter space, integrals with multiple massive legs are reduced to sums of single-massive integrals.
• Higher Denominator Powers: Recursion relations (analogous to Integration-by-Parts identities) are derived by differentiating with respect to kinematic invariants. This reduces any integral with arbitrary powers $j_i > 1$ to a set of master integrals (where $j_i=1$).

3. Key Results

A. Three-Denominator Integrals ($n=3$)

• All-Massless Case:
  • Derived to $\mathcal{O}(\epsilon^2)$.
  • Result expressed entirely in GPLs (first such result in literature).
  • Pole structure: Single $1/\epsilon$ pole (collinear singularity), no double pole.
  • No square roots appear in the integration variables; rationalization is not needed.
• Single-Massive Case:
  • Derived to $\mathcal{O}(\epsilon)$.
  • Requires a 4-fold MB integral. One integral contains a quadratic radicand, which is rationalized before integration.
• Multi-Massive Cases:
  • Reduced to single-massive results via PF decomposition involving coefficients derived from a two-point splitting lemma.
• Recursion: A complete set of 6 linear recursion relations allows reduction of any $j_i > 1$ to the master integral $I_{1,1,1}$.

B. Four-Denominator Integrals ($n=4$)

• All-Massless Case:
  • Derived to $\mathcal{O}(\epsilon^0)$ (finite part).
  • Requires evaluating a six-fold MB integral analytically.
  • Pole Structure: The $1/\epsilon$ pole is $S_4$-symmetric and determined by collinear singularities.
  • Finite Part: Involves weight-2 GPLs with an alphabet of 17 letters. Notably, 8 letters involve square roots (Gram determinant analogues) that emerge only when recasting iterated integrals into the GPL basis.
• Single-Massive Case:
  • Derived to $\mathcal{O}(\epsilon^0)$.
  • Requires a seven-fold MB integral.
  • Pole structure is $S_3$-symmetric (broken $S_4$ due to the massive leg).
  • Finite part involves weight-2 GPLs with 11 letters.
• Multi-Massive Cases:
  • Reduced to single-massive results via extended PF decomposition.
  • Numerical evaluation of all mass configurations takes $<1$ second.

4. Significance and Impact

• First Analytic Results: To the authors' knowledge, these are the first analytic results for $n=4$ angular integrals in terms of GPLs, and the first for $n=3$ to $\mathcal{O}(\epsilon^2)$ in this basis.
• Algorithmic Scalability: The method is fully algorithmic. The complexity increases computationally (number of folds) but not conceptually as $n$ increases.
• Numerical Efficiency: The GPL representation offers a massive speedup for numerical evaluation. Using GiNaC, evaluation takes ~1 second compared to ~30 minutes for direct numerical MB integration (a $\sim 1800\times$ speedup).
• Radial Convolution: The primary advantage of the GPL basis is that GPLs close under iterated integration. This allows the angular result to be convolved with the process-dependent radial component:
  • In many cases, this final step can be performed analytically, yielding fully analytic phase-space integrals.
  • If analytic radial integration is impossible, the remaining 1D integral is trivial to compute numerically with high precision.
• Future Outlook: The method is ready for extension to $n \ge 5$ and higher perturbative orders, provided the computational resources for higher-fold MB integrals are available. This is crucial for Next-to-Next-to-Leading Order (NNLO) and N$^3$LO calculations in processes like Semi-Inclusive Deep-Inelastic Scattering (SIDIS).

5. Conclusion

The paper presents a robust, algorithmic framework for solving complex phase-space angular integrals using Mellin-Barnes representations. By converting these into Goncharov Polylogarithms, the authors provide results that are not only analytically precise but also computationally efficient and structurally suitable for the full integration required in high-order QCD phenomenology.