Improved Standard-Model predictions for $\eta^{(\prime)}\to \ell^+ \ell^-$

The rare decays $\eta^{(\prime)}\to\ell^+\ell^-$, $\ell\in\{e,\mu\}$, are highly suppressed in the Standard Model, both by their chirality structure and the required loop attaching the lepton line to the $\eta^{(\prime)}\to\gamma^*\gamma^*$ matrix element. The latter is described by a single scalar function, the transition form factor, which has recently been studied in great detail for $\eta^{(\prime)}$ in the context of the pseudoscalar-pole contributions to hadronic light-by-light scattering in the anomalous magnetic moment of the muon. Based on these results, we evaluate the corresponding prediction for the $\eta^{(\prime)}$ dilepton decays, supplemented by an improved evaluation of the asymptotic contributions including pseudoscalar mass effects. In particular, the dispersive representation for the $\eta^{(\prime)}$ transition form factors allows us, for the first time, to perform a robust evaluation of the imaginary parts due to subleading channels besides the dominant two-photon cut. Our final results are $\text{Br}[\eta\to e^+e^-]=5.37(4)(2)[4]\times 10^{-9}$, $\text{Br}[\eta\to \mu^+\mu^-]=4.54(4)(2)[4]\times 10^{-6}$, $\text{Br}[\eta'\to e^+e^-]=1.80(2)(3)[3]\times 10^{-10}$, and $\text{Br}[\eta'\to \mu^+\mu^-]=1.22(2)(2)[3]\times 10^{-7}$, where the errors refer to the uncertainty in the normalized branching fraction, the one propagated from $\text{Br}[\eta^{(\prime)}\to\gamma\gamma]$, and the total uncertainty, respectively. The branching fraction for $\eta\to\mu^+\mu^-$ exhibits a mild $1.6\sigma$ tension with experiment, and we explore the bounds that can be derived on physics beyond the Standard Model.