First-hand information for everyone
This paper reports the first exact calculation, at leading colour, of the two-loop quantum corrections for producing a top–antitop pair together with a W boson. Using that amplitude, the author produces predictions for the inclusive production rate at next-to-next-to-leading order in quantum chromodynamics (NNLO QCD). The result helps test previous approximate methods and is a step toward more precise predictions for measurements that currently show some tension with theory.
The two-loop amplitude is a higher-order quantum correction needed at NNLO QCD. The author works in the leading‑colour approximation, which keeps the dominant terms in the expansion in the number of colours and light quark flavours and drops smaller colour-suppressed terms. At this approximation only 210 of the 722 two-loop Feynman diagrams contribute. The amplitude is decomposed into 24 tensor structures. Those form factors are written as combinations of 8,959 scalar loop integrals, which are reduced to 330 master integrals using integration‑by‑parts identities and modern algebraic tools (NeatIBP and FiniteFlow).
Evaluating the master integrals is the technical core of the work. The integrals satisfy systems of differential equations in the kinematic variables. For many integrals the equations are of a well‑understood type that leads to so‑called polylogarithmic functions. But some two‑loop integrals involve elliptic curves, which bring in more complicated functions. The author chooses a basis where the elliptic integrals are finite and the polylogarithmic part satisfies canonical differential equations. The whole amplitude is then expressed in terms of a set of special functions that separate the polylogarithmic and elliptic pieces. The finite remainder (the part left after removing ultraviolet and infrared singularities) is written in this basis; the paper reports a few hundred special functions appearing in the result.
To get numbers, the special functions are evaluated by solving the differential equations numerically. The boundary values for the integrals come from the public tool AMFlow, and the DEs are solved along linear paths using series expansions. This step is costly: evaluating the special functions takes about an hour per phase‑space point. The rational coefficients that multiply the special functions are reconstructed exactly from numerical evaluations modulo prime numbers (finite‑field techniques) using FiniteFlow; this takes on the order of ten minutes per point and can require hundreds of primes. Because evaluating the two‑loop contribution at each point is expensive, the author builds and interpolates a five‑dimensional grid for the double‑virtual part when computing the inclusive cross section.
Why this matters: having the exact leading‑colour two‑loop amplitude provides a solid check of previous dynamic approximations (soft approximation and massification) and is needed for reliable differential predictions. Important caveats remain. The calculation keeps only the leading‑colour terms, so subleading colour effects are neglected and could be enhanced near production thresholds. The presence of elliptic integrals prevents a simple, fully analytic form for all parts of the amplitude and makes numerical evaluation slow. Finally, the excerpted TeX source may be incomplete, so further technical details and full numerical results are given in the referenced papers rather than in this short summary.