Measurement of single charged pion production in charged-current $ν_μ$-Ar interactions with the MicroBooNE detector

We present flux-averaged charged-current $ν_μ$ cross-section measurements on argon for final states containing exactly one $π^\pm$ and no other hadrons except nucleons. The analysis uses data from the MicroBooNE experiment in the Booster Neutrino Beam, corresponding to $1.11 \times 10^{21}$ protons on target. Total and single-differential cross-section measurements are provided within a phase space restricted to muon momenta above 150 MeV, pion momenta above 100 MeV, and muon-pion opening angles smaller than 2.65 rad. Differential cross sections are reported with respect to the scattering angles of the muon and pion relative to the beam direction, their momenta, and their combined opening angle. The differential cross section with respect to muon momentum is based on a subset of selected events with the muon track fully contained in the detector, whereas the cross section with respect to pion momentum is based on a subset of selected events rich in pions that have not hadronically scattered on the argon before coming to rest. The latter has not been measured on argon before. The total cross section is measured as $(3.75~\pm0.07\textrm{(stat.)}\pm0.80~\textrm{(syst.)}) \times 10^{-38} , \text{cm}^2/\text{Ar}$ at a mean energy of approximately 0.8 GeV. Comparisons of the measured cross sections with predictions from multiple neutrino-nucleus interaction generators show good overall agreement, except at very forward muon angles.

💡 Research Summary

This paper presents the first high‑statistics measurement of charged‑current muon‑neutrino interactions on argon that produce exactly one charged pion (CC1π±) and any number of nucleons, using the MicroBooNE detector exposed to the Fermilab Booster Neutrino Beam (BNB). The data set corresponds to 1.11 × 10²¹ protons on target collected between 2015 and 2020, yielding a mean neutrino energy of about 0.8 GeV. The analysis defines a restricted phase space: muon momentum >150 MeV/c, pion momentum >100 MeV/c, and muon‑pion opening angle <2.65 rad. Within this region, 6 816 signal‑like events are selected after background subtraction.

The neutrino flux is simulated with a Geant4‑based model tuned to HARP and SciBooNE hadron‑production data; the interaction model uses GENIE v3.0.6 (tune G18_10a_02_11a) with T2K‑constrained CCQE and CCMEC components and the hA 2018 intranuclear cascade for final‑state interactions. Known deficiencies, such as an over‑prediction of charge‑exchange below 400 MeV, are examined with a dedicated modified sample, and the impact on selection efficiency is found to be negligible.

Event reconstruction employs the LArSoft framework and Pandora pattern‑recognition, followed by a support‑vector‑machine classifier to separate neutrino‑like slices from cosmic‑ray activity. Special care is taken to mitigate track fragmentation and Michel‑electron merging, which could otherwise mimic a back‑to‑back muon‑pion topology. Two exclusive subsamples are defined: one with fully contained muon tracks (used for dσ/dpμ) and another enriched in pions that stop without hadronic scattering (used for dσ/dpπ).

Systematic uncertainties are evaluated from ten sources, including flux (≈ 7 %), detector response (≈ 5 %), and interaction‑model variations (≈ 10 %). The total systematic error on the flux‑integrated cross section is 21 %. The measured total cross section is
σ(CC1π±) = (3.75 ± 0.07 (stat) ± 0.80 (syst)) × 10⁻³⁸ cm² per argon nucleus.

Differential cross sections are reported as functions of muon angle, pion angle, muon‑pion opening angle, muon momentum, and pion momentum. The pion‑momentum distribution is the first such measurement on argon, extending the limited statistics of the earlier ArgoNeuT result. Comparisons with several generators (GENIE, NuWro, GiBUU, NEUT) show overall agreement, but all models underestimate the rate at very forward muon angles (θμ < 20°), indicating possible shortcomings in forward pion production or final‑state interaction modeling.

These results provide essential input for the tuning of neutrino‑interaction generators used by the DUNE and Short‑Baseline Neutrino programs, both of which employ liquid‑argon time‑projection chambers. By reducing model uncertainties in CC1π± channels, the measurements help improve background predictions for oscillation analyses and contribute to a more reliable interpretation of future argon‑based neutrino experiments.