Stringent constraints on non-standard neutrino interactions using high-purity $ν_μ$ CC events in IceCube DeepCore

Abstract

The neutral-current (NC) non-standard interactions (NSI) of neutrinos with fermions can modify the flavor oscillations of atmospheric neutrinos as they propagate through the Earth. We present constraints on the NC-NSI parameters $\varepsilon_{μτ}$ and $\varepsilon_{ττ}-\varepsilon_{μμ}$ (one at a time) using a high-purity sample of $ν_μ$ charged-current (CC) atmospheric neutrino events collected by IceCube DeepCore over 7.5 years of livetime. These two parameters significantly affect the $ν_μ$ disappearance channel for which this golden event sample is optimized by the IceCube Collaboration. The best fit to this dataset is consistent with no NSI hypothesis, and we place the most stringent constraints to date: $-\,0.0094 < \varepsilon_{μτ} < 0.0079$ and $-\,0.030 < \varepsilon_{ττ}-\varepsilon_{μμ} < 0.029$ at 90% confidence level.

Figures (8)

• Figure 1: Constraints on the NSI parameters $\varepsilon_{\mu\tau}$ and $\varepsilon_{\tau\tau}-\varepsilon_{\mu\mu}$ at 90% confidence level (C.L.) from this analysis using the 8-year golden event sample of IceCube DeepCore. Comparison with limits from other experiments, such as IceCube DeepCore (2021) IceCubeCollaboration:2021euf, KM3NeT/ORCA (2025) KM3NeT:2024pte, ANTARES (2022) ANTARES:2021crm, IceCube (2022) IceCube:2022ubv, and Super-Kamiokande (2011) Super-Kamiokande:2011dam is shown. The limits for $\varepsilon_{\mu\tau}$ are obtained by fixing the complex phase to $0$ ( i.e., positive $\varepsilon_{\mu\tau}$) and $\pi$ ( i.e., negative $\varepsilon_{\mu\tau}$). For consistency, bounds from Refs. KM3NeT:2024pteANTARES:2021crmIceCube:2022ubvSuper-Kamiokande:2011dam have been rescaled to the NSI convention used in our work.
• Figure 2: Difference in the three-flavor neutrino oscillation probabilities $P(\nu_\mu \rightarrow \nu_\mu)$ between SI + NSI and SI scenarios in the plane of neutrino energy and cosine of zenith angle. The top and bottom panels correspond to the NSI scenarios with $\varepsilon_{\mu\tau} = \pm\,0.05$ and $\varepsilon_{\tau\tau} - \varepsilon_{\mu\mu} = \pm\,0.1$, respectively, considering one NSI parameter at a time. Here, we take $\theta_{23} = 45.57^\circ$ and $\Delta m^2_{31} = 2.48 \times 10^{-3}~{\rm eV}^2$.
• Figure 3: Difference of the expected event distributions between SI + NSI and SI scenarios ( i.e., $N_{\text{SI+NSI}} - N_{\text{SI}}$) in the plane of the reconstructed energy $E_{\text{reco}}$ and the reconstructed cosine of zenith angle $\cos\theta_{\text{reco}}$, for two NSI parameter choices taken one at a time. The top and bottom panels correspond to $\varepsilon_{\mu\tau} = +\,0.05$ and $\varepsilon_{\tau\tau} - \varepsilon_{\mu\mu} = +\,0.1$, respectively. The left and right panels show distributions for mixed and track-like samples, respectively. Here, we use the nominal values of the nuisance parameters as given in Table \ref{['tab:systematic_params']}.
• Figure 4: The observed (black-solid curve) and the expected (black-dashed curve) constraints on the magnitude $|\varepsilon_{\mu\tau}|$ and complex phase $\phi_{\mu\tau}$ at the 90% C.L. (2 DOF) using the 8-year golden event sample of DeepCore. The red star marker represents the best-fit values for magnitude and phase. The top and the side sub-panels represent the 1D projections of the 2D contours for $|\varepsilon_{\mu\tau}|$ and $\phi_{\mu\tau}$, respectively, while the $\chi^2$ is minimized over the other parameter. The gray dotted lines in the sub-panels represent the 68.3%, 90%, and 99.7% confidence levels for 1 DOF. The solid blue curves represent the previous IceCube DeepCore result IceCubeCollaboration:2021euf.
• Figure 5: The observed (black-solid curve) and the expected (black-dashed curve) $\Delta\chi^2_{\rm mod}$ as a function of the $\varepsilon_{\mu\tau}$ parameter, assuming real values with the complex phase fixed to $0$ and $\pi$. We use the 8-year golden event sample of DeepCore. Shaded bands show the 68% and 90% ranges for the expected $\Delta\chi^2_{\rm mod}$ calculated from 500 statistically fluctuated pseudo-experiments. The horizontal dotted lines represent the 68% and 90% confidence levels for 1 DOF.