Pragmatic view of short-baseline neutrino oscillations

C. Giunti, M. Laveder, Yufeng Li, H. W. LongPhysical 2013 · DOI · arXiv

Abstract

We present the results of global analyses of short-baseline neutrino oscillation data in $3+1$, $3+2$ and $3+1+1$ neutrino mixing schemes. We show that the data do not allow us to abandon the simplest $3+1$ scheme in favor of the more complex $3+2$ and $3+1+1$ schemes. We present the allowed region in the $3+1$ parameter space, which is located at $\Delta m_{41}^{2}$ between 0.82 and $2.19\text{ eV}^{2}$ at $3\sigma$. The case of no oscillations is disfavored by about $6\sigma$, which decreases dramatically to about $2\sigma$ if the Liquid Scintillating Neutrino Detector (LSND) data are not considered. Hence, new high-precision experiments are needed to check the LSND signal.

Introduction

The Standard Model describes three active neutrino flavors ($\nu_e, \nu_\mu, \nu_\tau$). However, persistent anomalies in short-baseline (SBL) experiments—most notably the LSND experiment's observed excess of $\bar{\nu}_e$ in a $\bar{\nu}_\mu$ beam—suggest the existence of "sterile" neutrinos. These hypothetical particles would not participate in weak interactions but could mix with active neutrinos, leading to oscillations over short distances that the standard three-neutrino framework cannot explain.

This paper provides a comprehensive global fit of available SBL data to determine the most pragmatic model for these anomalies. By comparing $3+1$, $3+2$, and $3+1+1$ mixing schemes, the authors evaluate whether the added complexity of multiple sterile neutrinos is statistically justified.

Global Analysis Framework

The study utilizes a least-squares fit across three primary schemes:

  • 3+1: Three active neutrinos plus one eV-scale sterile neutrino.
  • 3+2: Three active plus two eV-scale sterile neutrinos.
  • 3+1+1: Three active, one eV-scale sterile, and one much heavier sterile neutrino.

The analysis incorporates a vast dataset including $\nu_\mu \to \nu_e$ appearance (LSND, MiniBooNE, NOMAD, etc.), $\nu_e$ disappearance (Reactor and Gallium anomalies), and $\nu_\mu$ disappearance (CDHSW, MINOS, atmospheric data).

The MiniBooNE Low-Energy Excess

A significant challenge in SBL physics is the MiniBooNE low-energy anomaly—an unexplained excess of events at low energies. The authors argue that this excess likely has a non-oscillation origin. As shown in Figure 3, while the models can fit the higher-energy data reasonably well, they struggle to capture the sharp rise at low energies.

MiniBooNE neutrino and antineutrino appearance probabilities compared to 3+1, 3+2, and 3+1+1 models.

Excluding these anomalous low-energy bins (the "HIG" fit) significantly improves the overall goodness-of-fit (GoF) from 6% to 29%, suggesting that the $3+1$ model is a robust description of the remaining data.

Global Fit Results and Experimental Tension

The $3+1$ model emerges as the most pragmatic choice, offering the best balance of simplicity and fit quality. The preferred mass-squared difference is $\Delta m^2_{41} \approx 1.6 \text{ eV}^2$, with a $3\sigma$ allowed range of $0.82$ to $2.19 \text{ eV}^2$.

However, a persistent tension exists between appearance and disappearance datasets. Appearance experiments like LSND prefer mixing parameters that are largely excluded by disappearance experiments. Figure 5 illustrates this tension: the global fit (GLO) region represents a statistical compromise between the appearance (APP) signal and the disappearance (DIS) limits.

Global fit (GLO) allowed regions in the 3+1 parameter space compared to appearance (APP) and disappearance (DIS) constraints.

The statistical significance of the oscillation signal is heavily dependent on LSND. Including LSND rejects the no-oscillation hypothesis at $6.2\sigma$; without it, the significance drops to $2.1\sigma$.

Conclusion

While current data strongly suggests the existence of at least one sterile neutrino, the $3+1$ model remains sufficient. More complex $3+2$ or $3+1+1$ schemes do not significantly resolve the appearance-disappearance tension or improve the fit to the MiniBooNE low-energy excess. Given the heavy reliance on the LSND signal, new high-precision experiments are essential to provide the final word on whether sterile neutrinos truly exist.

Summary updated

Abstract

We present the results of global analyses of short-baseline neutrino oscillation data in $3+1$, $3+2$ and $3+1+1$ neutrino mixing schemes. We show that the data do not allow us to abandon the simplest $3+1$ scheme in favor of the more complex $3+2$ and $3+1+1$ schemes. We present the allowed region in the $3+1$ parameter space, which is located at $\ensuremath{\Delta}{m}_{41}^{2}$ between 0.82 and $2.19\text{ }\text{ }{\mathrm{eV}}^{2}$ at $3\ensuremath{\sigma}$. The case of no oscillations is disfavored by about $6\ensuremath{\sigma}$, which decreases dramatically to about $2\ensuremath{\sigma}$ if the Liquid Scintillating Neutrino Detector (LSND) data are not considered. Hence, new high-precision experiments are needed to check the LSND signal.
BibTeX
@inproceedings{giunti2013pragmatic,
  author = {C. Giunti and M. Laveder and Yufeng Li and H. W. Long},
  title = {{Pragmatic view of short-baseline neutrino oscillations}},
  booktitle = {Physical},
  year = {2013},
  doi = {10.1103/physrevd.88.073008},
}
Figure