MiniBooNE and LSND data: non-standard neutrino interactions in a (3+1) scheme versus (3+2) oscillations

TL;DR

This work investigates whether charged-current non-standard neutrino interactions (NSI) in a (3+1) sterile-neutrino framework can reconcile the LSND and MiniBooNE appearance signals, including CP violation through interference between NSI and sterile oscillations with Δm^2_{41} ≃ 1 eV^2. By parameterizing production and detection NSI with ε_{ ext{αβ}}^X and exploiting a one-mass-scale-dominant limit, the authors derive compact expressions for appearance and disappearance probabilities and perform global fits to short-baseline data in two NSI realizations: a constrained NSI^c (5 parameters) and a general NSI^g (8 parameters). The NSI^c fit yields a moderate improvement over pure (3+1) oscillations, while NSI^g provides a much larger improvement and effectively decouples LSND from disappearance constraints, offering the best overall fit among the considered scenarios; compared with (3+2) oscillations, NSI models yield larger improvements per added parameter, though none address the MiniBooNE low-energy excess. The results imply testable predictions such as zero-distance appearance effects and motivate exploration of TeV-scale mediators at colliders, while acknowledging that the MiniBooNE low-energy anomaly remains unresolved within these frameworks.

Abstract

The recently observed event excess in MiniBooNE anti-neutrino data is in agreement with the LSND evidence for electron anti-neutrino appearance. We propose an explanation of these data in terms of a (3+1) scheme with a sterile neutrino including non-standard neutrino interactions (NSI) at neutrino production and detection. The interference between oscillations and NSI provides a source for CP violation which we use to reconcile different results from neutrino and anti-neutrino data. Our best fit results imply NSI at the level of a few percent relative to the standard weak interaction, in agreement with current bounds. We compare the quality of the NSI fit to the one obtained within the (3+1) and (3+2) pure oscillation frameworks. We also briefly comment on using NSI (in an effective two-flavour framework) to address a possible difference in neutrino and anti-neutrino results from the MINOS experiment.

Figures (9)

• Figure 1: Constraint from no-evidence data (NEV) compared to the combined allowed regions from LSND and MiniBooNE $\bar{\nu}$ data (shaded) at 90% and 99% CL for (3+1) oscillations. We show also the individual regions from LSND and MiniBooNE $\bar{\nu}$ data.
• Figure 2: Predicted event spectra at the best fit point to appearance data. Shown are the predictions for MiniBooNE neutrino (left), MiniBooNE anti-neutrino (middle), and LSND (right) compared to data. Dashed histograms refer to (3+1) oscillations, solid histograms to (3+1) NSI. The dashed vertical lines indicate the 475 MeV threshold used for MiniBooNE data. For LSND we show the spectrum with free normalisation and the value for the total transition probability (inset). The parameter and $\chi^2$ values are given in tab. \ref{['tab:fit-params']}.
• Figure 3: Predicted event spectra at the best fit point to global data. Shown are predictions for MiniBooNE neutrino (left), MiniBooNE anti-neutrino (middle), and LSND (right) compared to data. Dashed histograms refer to (3+1) oscillations, solid histograms to (3+1) NSI$^c$. The dashed vertical lines indicate the 475 MeV threshold used for MiniBooNE data. For LSND we show the spectrum with free normalisation and the value for the total transition probability (inset). The parameter and $\chi^2$ values are given in tab. \ref{['tab:fit-params']}.
• Figure 4: Global fit in the (3+1) NSI$^c$ framework. Left: allowed regions projected onto the plane of $|\alpha_{\mu e}|$ and $|\beta_{\mu e}|$ at 90% and 99% CL (2 dof). Right: $\Delta\chi^2$ as a function of $|\beta_{\mu e}|$. We minimise over all undisplayed parameters.
• Figure 5: Constraint from no-evidence data (NEV) compared to the combined allowed regions from LSND and MiniBooNE $\bar{\nu}$ data (shaded) at 90% and 99% CL for the (3+1) NSI$^c$ case. The regions correspond to sections of the 4-dimensional allowed regions at fixed values of $|\beta_{\mu e}|$ and $\delta$ as given in the legend. These values correspond to the ones where the NEV constraint and the LSND+MB$\bar{\nu}$ allowed region touch each other (at $\Delta\chi^2 = 11.7$).