Understanding neutrino pion production with the GiBUU model

TL;DR

This work evaluates the GiBUU model's ability to describe neutrino-induced pion production across MINERvA and MicroBooNE data, emphasizing baseline performance, 2p2h contributions, and in-medium effects such as Δ broadening and density-dependent final-state interactions. It decouples background scaling into $T_1$ and $T_2$ to disentangle one-body and MEC contributions, finding that data prefer different settings depending on the observable and target, with MINERvA favoring minimal in-medium modification while MicroBooNE often requires stronger in-medium effects. The analysis shows that a unified description across experiments is difficult, revealing tensions in angular and momentum distributions for CC$\pi^0$ and CC0$\pi$ channels and highlighting the critical role of in-medium dynamics and FSI in shaping final-state observables. The results underscore the need for more precise data and cross-experiment calibration, particularly in the non-dispersive region of transverse kinematic imbalance, to inform improvements in nuclear modeling for neutrino oscillation analyses.

Abstract

Pion production is a major source of systematic uncertainty in neutrino oscillation measurements. We report a systematic investigation of neutrino-induced pion production using MINERvA and MicroBooNE data within the GiBUU theoretical framework. The analysis begins by establishing baseline model parameters using inclusive and pionless data from MINERvA, MicroBooNE, and T2K experiments. We then examine the role of in-medium effects, including resonance broadening and nucleon-nucleon final-state interactions. While agreement with individual datasets can be achieved through specific model configurations, we demonstrate the difficulty of a unified description across all experiments: MINERvA measurements prefer minimum in-medium modifications, whereas MicroBooNE data require the maximum in-medium enhancement, revealing the complexity and richness of the underlying nuclear dynamics.

Figures (26)

• Figure 1: GiBUU$\nu_\mu$ CC interactions on $^{12}$C at time after interaction: $t=0, ~150,~\text{and}~500$ a.u. (from lower to upper), where $1~\text{a.u.}=0.2~\text{fm/}c\simeq 6.7\times10^{-25}~\textrm{s}$. The neutrino energy is 5 GeV. The population of the final-state protons ($N$ in an arbitrary unit) is shown as a function of the proton location, $R$, and its mass, $M_\textrm{p}$. The centre of the nucleus is at $R=0~\textrm{fm}$, and the boundary of the mean-field potential is seen at $R\simeq 5~\textrm{fm}$. The mass spread at large $R$ is a residual numerical artifact introduced during elastic scattering inside the nucleus.
• Figure 2: Similar to Fig. \ref{['fig:movie_m']} but as a function of the proton momentum $p_\textrm{p}$.
• Figure 3: TKI schematic from Ref. Lu:2015tcr.
• Figure 4: Shape-only comparison of $\delta\alpha_\textrm{T}$ from CC0$\pi$MINERvA:2018hba and CC$\pi^0$ measurements MINERvA:2020anu by MINERvA, where the final-state N is a proton and a $\pi+$proton system, respectively. Figure from Ref. MINERvA:2020anu. These data sets will be analyzed in Secs. \ref{['sec:cc0pi']} and \ref{['subsec:ccpi0']} where details will be given.
• Figure 5: CC inclusive cross section measured by (a) MINERvA on $^{12}\text{C}$MINERvA:2016ing and (b) MicroBooNE on $^{40}\text{Ar}$MicroBooNE:2021sfa, compared to GiBUU predictions with different $\mathcal{T}$ configurations. Note that the MicroBooNE prediction in (b) has been forward-folded with the detector response matrix provided in the corresponding publication. A similar procedure applies to all MicroBooNE predictions discussed in this paper. See also Appendix \ref{['app:xsec_decompose']} for a channel decomposition of MINERvA predictions.