Higher partial wave contamination in finite-volume 1-to-2 transitions

Abstract

In their seminal work, Lellouch and Lüscher derived a conversion factor relating a finite-volume matrix element, calculable using numerical lattice QCD, with the infinite-volume decay amplitude for $K\to \pi \pi$. The conversion factor depends on the $\pi \pi \to \pi \pi$ scattering amplitude with the same total isospin (either zero or two) as the $\pi \pi$ decay channel. Although an infinite tower of $\pi \pi \to \pi \pi$ partial-wave components affect the conversion factor, the $S$-wave ($\ell =0$) component is expected to dominate, and only this contribution is included in the well-known Lellouch-Lüscher factor, with other $\pi \pi \to \pi \pi$ partial-wave amplitudes formally set to zero. However, as the precision of lattice calculations increases, it may become important to assess the systematic uncertainty arising from this approximation. With this motivation, we compare the $S$-wave-only results with those truncated at the next contaminating partial wave: the $G$-wave ($\ell =4$) for zero total momentum in the finite-volume frame and the $D$-wave ($\ell =2$) otherwise. Using the general framework for $1\stackrel{\mathcal{J}}{\to }2$ transitions, we quantify the effect of higher partial waves for systems with zero and non-zero total momentum as well as with anti-periodic boundary conditions, presenting both generic numerical examples and results for realistic $\pi \pi$ amplitudes taken from chiral perturbation theory and dispersive analysis. We also consider the accidental degeneracy occurring in the 8${}^{\text{th}}$ excited state of the zero-momentum system. This exhibits qualitatively new features at $\ell =4$, not seen in the $\ell =0$ truncation.

Comments: 31 pages, 6 figures, 2 tables