Physicists working with the Super‑Kamiokande (SK) detector searched for two unusual ways a proton might decay: into a positron plus two neutral pions (p → e+ π0 π0) and into an antimuon plus two neutral pions (p → μ+ π0 π0). The idea is to test whether protons, which are normally stable in everyday life, can break apart in these three‑body channels. Using 0.401 megaton‑years of SK data, the team found no clear signal and set new lower limits on how long a proton would live if it decays that way.

The search is motivated by theoretical work that extends the Standard Model of particle physics. Grand Unification Theories (GUTs) and related studies predict that protons might slowly decay if a quantity called baryon number is not strictly conserved. A recent theoretical study suggested that three‑body decays that produce two neutral pions together with a charged anti‑lepton could occur at rates comparable to the more commonly studied two‑body decays (for example p → e+ π0). These predictions prompted the first dedicated search for p → e+ π0 π0 and p → μ+ π0 π0 in SK.

Super‑Kamiokande is a large underground water‑Cherenkov detector. It holds about 50 kilotons of ultra‑pure water and is lined with thousands of light sensors called photomultiplier tubes (11,129 inward‑facing 20‑inch tubes in the inner detector). When a charged particle moves through the water faster than light does in water, it makes a flash of blue light (Cherenkov light) that the sensors record. The analysis used data from all of SK’s pure‑water running periods (phases I through V) and relied on simulated signal events and simulated backgrounds from atmospheric neutrinos to decide which observed events could be proton decays.

In the data the team found one candidate event for each decay mode. Those counts match the number expected from ordinary atmospheric neutrinos, which are a known background that can mimic proton decay signatures. Because the observations are compatible with background, the group set lower bounds on the proton lifetime divided by the branching ratio (τ/B): τ/B(p → e+ π0 π0) > 7.2 × 10^33 years and τ/B(p → μ+ π0 π0) > 4.5 × 10^33 years at 90% confidence level. These limits are more than an order of magnitude stronger than the earlier results from the IMB‑3 experiment, which reported limits of 1.47 × 10^32 years and 1.01 × 10^32 years for the two modes.