Search for proton decay via $p\to e^{+}{\pi }^0$ and $p\to {\mu }^{+}{\pi }^0$ in $0.31\mathbf{\text{megaton}}·\mathbf{\text{years}}$ exposure of the Super-Kamiokande water Cherenkov detector

We have searched for proton decay via $p\to e^{+}{\pi }^0$ and $p\to {\mu }^{+}{\pi }^0$ using Super-Kamiokande data from April 1996 to March 2015, $0.306\mathrm{megaton}·\mathrm{years}$ exposure in total. The atmospheric neutrino background rate in Super-Kamiokande IV is reduced to almost half that of phase I-III by tagging neutrons associated with neutrino interactions. The reach of the proton lifetime is further enhanced by introducing new signal criteria that select the decay of a proton in a hydrogen atom. No candidates were seen in the $p\to e^{+}{\pi }^0$ search. Two candidates that passed all of the selection criteria for $p\to {\mu }^{+}{\pi }^0$ have been observed, but these are consistent with the expected number of background events of 0.87. Lower limits on the proton lifetime are set at $\tau /B(p\to e^{+}{\pi }^0)>1.6\times 10^{34}\mathrm{years}$ and $\tau /B(p\to {\mu }^{+}{\pi }^0)>7.7\times 10^{33}\mathrm{years}$ at 90% confidence level.

Figure 1

The left panel shows ${\pi }^{+}$ cross sections on ${}^{12}\mathrm{C}$ as a function of momentum. Open circles indicate existing data [26], solid and dashed lines show after and before the tune, respectively. The colors correspond to the total cross section and exclusive channels: total (black), elastic scattering (green), absorption (red), charge exchange (blue), and $\pi$ production (light blue). The right panel shows cumulative fractions of neutral pion final state interactions as a function of momentum generated by $p\to e^{+}{\pi }^0$ MC. The fractions of events that undergo charge exchange, multiple $\pi$ production, scattering, and absorption are shown by the various shades as labeled. Neutral pions that exit the nucleus without experiencing any final state interactions are indicated by the portion labeled “No interaction.”

Figure 2

Distribution of the number of tagged neutrons for $p\to e^{+}{\pi }^0$ (top) and $p\to {\mu }^{+}{\pi }^0$ (bottom) after applying (Cut-1) through (Cut-5), excluding the signal region defined by (Cut-6). Dots show SK-IV data with $133.5\mathrm{kton}·\mathrm{years}$ exposure, the solid histogram shows the multiplicity of tagged neutrons per event in the atmospheric $\nu$ MC, and the dashed histogram shows the true multiplicity of neutrons per event.

Figure 3

Reconstructed proton mass vs total momentum for $p\to e^{+}{\pi }^0$ (top) and $p\to {\mu }^{+}{\pi }^0$ (bottom) after all cuts except (Cut-6). The left panels show signal MC, where light blue corresponds to free protons and dark blue is bound protons. The middle panels show atmospheric $\nu$ MC corresponding to 500 years live time of SK, and the right panels show SK-I to SK-IV data which contain 3408 and 1180 events for $p\to e^{+}{\pi }^0$ and $p\to {\mu }^{+}{\pi }^0$, respectively. The dot size is enlarged in the signal box.

Figure 4

Distributions of reconstructed invariant mass (left) and total momentum (right) for $p\to e^{+}{\pi }^0$ in the top panels and for $p\to {\mu }^{+}{\pi }^0$ in the bottom panels, after all selection cuts except cuts on the plotted variable. The dark blue histograms correspond to 90% confidence level allowed signal and the histograms filled by light blue show the portion contributed by free proton decay. The red histograms show atmospheric $\nu$ MC, and the dots are data with $0.306\mathrm{Mton}·\mathrm{years}$ exposure. Vertical dashed lines indicate the signal regions. The peak around $150\mathrm{MeV}/c^2$ in the total mass distribution of atmospheric $\nu$ and data in the left top panel arises from ${\pi }^0$ decays.

Figure 5

Event display of the second candidate event, zoomed to the region of the rings. The blue solid line and the tan dashed line show the reconstructed $e$-like and $\mu$-like rings, respectively. The dark orange solid line shows an additional $e$-like ring that was identified in the initial ring counting process, but it is rejected by the ring correction because it is too close in angle to the other $e$-like ring (blue line). As a result, this event is judged as a two-ring event.