A new measurement of the properties of the rare decay K -> pi+ e+ e-

Abstract

A large low-background sample of events (10300) has been collected for the rare decay of kaons in flight K+ -> pi+ e+ e- by experiment E865 at the Brookhaven AGS. The decay products were accepted by a broad band high-resolution charged particle spectrometer with particle identification. The branching ratio (2.94 +- 0.05(stat.) +- 0.13(syst.) +- 0.05(model))*10**{-7} was determined normalizing to events from the decay chain K+ -> pi+ pi0; pi0 -> e+ e- gamma. From the analysis of the decay distributions the vector nature of this decay is firmly established now, and limits on scalar and tensor contributions are deduced. From the (e+ e-) invariant mass distribution the decay form factor f(z)=f0(1+ delta*z) (z=M(ee)**2/m(K)**2) is determined with delta=2.14 +- 0.13 +- 0.15. Chiral QCD perturbation theory predictions for the form factor are also tested, and terms beyond leading order O(p**4) are found to be important.

Figures (5)

• Figure 1: Plan view of the E865 detector. P1 - P4: proportional chambers; C1, C2: Čerenkov counters; ABCD: scintillator hodoscopes. The beam passes through holes in the calorimeter and muon stack, dead regions in the proportional chambers, He bags and H$_2$ filled beam tubes in the Čerenkov counters, which are not shown here.
• Figure 2: $M_{ee}$ distribution for accepted $K_{dal}$ events with (lower curve) and without (upper curve) high-mass trigger required. The Monte Carlo simulation (histogram) includes events from other $K^+$ decays with a $\pi^0$ in the final state.
• Figure 3: Scatter plot $M_{ee}$ versus $M_{\pi ee}$ for $K_{\pi e e}$ candidates. Insert: $M_{\pi ee}$ mass for $K_{\pi e e}$ candidates with $M_{ee}>0.15$ GeV. The histogram shows the Monte Carlo simulation.
• Figure 4: Angular (left) and invariant mass (right) distributions for $K_{\pi e e}$ events (data points) compared to Monte Carlo simulations (histogram) assuming a pure vector interaction. A linear form factor parametrisation with $\delta=2.14$ is used. The dashed histogram (right) corresponds to a constant form factor ($\delta=0$).
• Figure 5: The measured form factor $|f_V(z)|$ versus $z$. The solid line shows the linear fit (Eq. \ref{['eq:lambda']} with $\delta^\prime=0$), the dashed line the next-to-leading-order ChPT fit (Eq. \ref{['eq:chpt']}).