Search for a dark photon in the $π^0 \to e^+e^-γ$ decay

TL;DR

The study searches for a light dark photon $U$ in the mass window $20$–$100\ \mathrm{MeV}$ via the Dalitz decay $\pi^0\to e^+e^-\gamma$ produced in $pp\to pp\pi^0$, using the WASA-at-COSY detector. It builds a robust event selection and background model, reconstructs the $e^+e^-$ invariant-mass spectrum, and performs a non-negative constrained fit to set 90% CL limits on $BR(\pi^0\to \gamma(U\to e^+e^-))$ across $M_U$, which are translated into limits on the mixing parameter $\epsilon^2$. The resulting limit $\epsilon^2<5\times 10^{-6}$ over $20$–$100$ MeV tightens constraints on the $(M_U, \epsilon^2)$ plane, particularly limiting the region favored by the muon $g-2$ anomaly and improving on prior measurements at low $M_U$. The analysis demonstrates the utility of large $\pi^0$ decay data samples for probing dark-sector scenarios and informs future high-statistics searches.

Abstract

The presently world largest data sample of pi0 --> gamma e+e- decays containing nearly 5E5 events was collected using the WASA detector at COSY. A search for a dark photon U produced in the pi0 --> gamma U --> gamma e+e- decay from the pp-->ppπ^0 reaction was carried out. An upper limit on the square of the U-gamma mixing strength parameter epsilon^2 of 5e-6 at 90% CL was obtained for the mass range 20 MeV <M_U< 100 MeV. This result together with other recent experimental limits significantly reduces the M_U vs. ε^2 parameter space preferred by the measured value of the muon anomalous magnetic moment.

Figures (8)

• Figure 1: Feynman diagrams for a) the lowest order electromagnetic $\pi^0\to e^+e^-\gamma$ decay and a possible contribution of $U$ vector boson to: b) $\pi^0\to e^+e^-\gamma$ and c) lepton $g-2$.
• Figure 2: Detector performance plots for a data sample with two reconstructed protons, an $e^+e^-$ pair and a photon. a) Distribution of the missing mass squared with respect to the two protons registered in the FD before electron identification. Experimental data (black points); simulations: $\pi^0\to e^+e^-\gamma$ and $\pi^0\to\gamma\gamma$ (broken line), random coincidences of two events (dotted line), and the sum (solid line). b) Distribution of $MM(pp)$ after electron identification: experimental data (black points) and sum of Monte Carlo simulations (solid line). c) The reconstructed invariant mass of the $e^+e^-\gamma$ system after particle identification cut.
• Figure 3: The reconstructed $e^+e^-$ invariant mass $q=IM(e^+e^-)$: a) before and b) after the cuts for reducing the conversion background. The experimental data are denoted by black points. Results of simulations for $\pi^0\to\gamma\gamma$ (blue line) and $\pi^0\to e^+e^-\gamma$ (green line) decays are normalized according to the known branching ratios. The normalization of random coincidences (dotted line) was fitted in order to reproduce the $IM(e^+e^-) > 150$ MeV range. The sum of all simulated contributions is given by the red line.
• Figure 4: Distribution of the distance $R$ between the COSY beam axis and the reconstructed point of closest approach of $e^+$ and $e^-$ tracks: experimental data (black crosses); simulations for $\pi^0\to\gamma\gamma$ (blue line), the $\pi^0\to e^+e^-\gamma$ decay (green line), and the sum of the two contributions (red line).
• Figure 5: Correlation between $R$ and $IM_b$ variables for the experimental data. The selection cut is shown by the diagonal line. The events below the line mainly come from photon conversions in the beam pipe.