betaFeedback

WorksPhysics and Astronomy

Testing the Pseudo-Dirac Nature of Atmospheric Neutrinos

Abstract

Pseudo-Dirac neutrinos are characterized by extremely small mass-squared differences between two nearly degenerate, strongly-mixed, active-sterile states. These differences are associated with new very-long oscillation lengths and provide a probe of the nature of neutrinos. Atmospheric neutrinos can probe these oscillations as they allow access to a broad range of baseline distances, from tens of kilometers to the Earth's diameter. We investigate the sensitivity of an experiment like the Deep Underground Neutrino Experiment (DUNE) to the pseudo-Dirac nature of neutrinos through measurements of the atmospheric neutrino flux. We concentrate on the modifications to the atmospheric neutrino energy and angular spectra induced by one new oscillation frequency associated with the ν3ν_3 state and determine the parameter space that can be probed with atmospheric neutrinos at a DUNE-like detector. We also comment on the sensitivity of these measurements to the sub-leading "solar" oscillations, which are also characterized by a relatively small mass-squared difference.

Introduction

The discovery of neutrino oscillations in solar, atmospheric, accelerator, and reactor experiments embodies one of the most important breakthroughs in particle physics over the past few decades. It constitutes the first direct evidence of physics beyond the Standard Model (SM), since neutrinos are massless in the original formulation of the SM.

Neutrino oscillations are quantum mechanical interference processes in which neutrinos change flavors after traversing large distances. The flavor-conversion probability depends on the propagation distance LL, the neutrino energy EE, the mass-squared differences Δmij2=mi2−mj2\Delta m_{ij}^{2}=m_i^2-m_j^2 among the neutrinos and mixing angles θij\theta_{ij} (i,j=1,2,3i,j=1,2,3, i≠ji\ne j) that characterize charged-current weak interactions [1]. The current data have established the three-flavor paradigm of oscillation on a rather firm footing, and the mixing angles and mass-squared differences are measured with impressive precision. The three main features of the standard picture that remain undetermined are the sign of Δm312\Delta m_{31}^{2}, i.e., the neutrino mass ordering, the octant of the mixing angle θ23\theta_{23}, and the Dirac CP phase δCP\delta_{\mathrm{CP}}.

Neutrino oscillation experiments remain insensitive to the absolute neutrino mass scale. Information about this parameter can come from experiments on beta decay and neutrinoless double beta decay as well as from cosmology. Among these, the combined analysis of DESI DR2 plus Planck CamSpec provides the best bound of 0.0642 eV on the sum of neutrino masses at the 95% level, disfavoring the inverted mass ordering [2]. The small mass-squared differences measured in oscillation experiments, together with the cosmological mass bound, imply that the neutrino masses are very small, safely below 0.1 eV.

Understanding the origin of such small neutrino masses remains one of the central open questions in particle physics. Unlike charged fermions, neutrinos are electrically neutral and can be Majorana or Dirac particles. A Dirac neutrino consists of four independent degrees of freedom corresponding to left-handed and right-handed neutrinos and antineutrinos. In this case, total lepton number is a conserved symmetry. Dirac neutrinos require the introduction of right-handed neutrino fields that are singlets under the SM gauge group. In contrast, Majorana neutrinos are their own antiparticles. A Majorana neutrino consists of two independent degrees of freedom: a left-handed and a right-handed one. Majorana neutrino masses violate the lepton number by two units. Such terms can arise through the dimension-five Weinberg operator [3] whose ultraviolet completion might come via one of the various seesaw mechanisms [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14] (for an overview, see, for example, [15]). These can be naturally implemented in grand unified theories and provide an elegant formulation for generating small neutrino masses in which the violation of lepton number is related to the small neutrino masses. Neutrino oscillations cannot distinguish between the Dirac and Majorana nature of the neutrinos. Evidence for this can come from lepton-number-violating processes such as neutrinoless double beta decay. Determining the Dirac or Majorana nature of neutrinos is among the primary objectives of particle physics.

An intriguing intermediate possibility is that neutrinos are pseudo-Dirac fermions [16, 17, 18, 19]. In this scenario, the dominant neutrino mass originates from a Dirac mass term, while small Majorana mass terms split each Dirac neutrino into two nearly degenerate Majorana states. Pseudo-Dirac neutrinos are Majorana fermions but behave, as far as most experimental probes are concerned, as Dirac neutrinos.

As far as oscillations are concerned, if the neutrinos are pseudo-Dirac, every active neutrino is accompanied by a sterile partner and a new, extremely small mass-squared difference that satisfies

ΔmPD2≪∣Δm212∣,∣Δm312∣,(1)\Delta m_{\mathrm{PD}}^{2}\ll\left|\Delta m_{21}^{2}\right|,\left|\Delta m_{31}^{2}\right|, \tag*{(1)}

such that conventional three-flavor oscillation phenomenology remains essentially unchanged at short-enough baselines. However, oscillations into sterile partners are expected to become observable for sufficiently large values of L/EL/E. Hence, the presence of pseudo-Dirac neutrinos modifies the standard oscillation framework by introducing additional oscillation frequencies associated with the tiny active–sterile mass splittings, which changes the flavor transition probabilities [20, 21]. Assuming the mixing between the active–sterile pair to be maximal, robust constraints have been obtained on the mass splitting associated with the first two mass eigenstates, Δm11′2,Δm22′2≲10−11 eV2\Delta m_{11'}^2,\Delta m_{22'}^2 \lesssim10^{-11}\ \mathrm{eV}^2 [22, 23, 24].

Existing constraints on the mass splitting associated with the third mass eigenstate from atmospheric, accelerator, and reactor data are considerably weaker: Δm33′2≲10−5 eV2\Delta m_{33'}^2 \lesssim10^{-5}\ \mathrm{eV}^2 [25]. Studies have also been done to probe the pseudo-Dirac nature of neutrinos in future reactor and accelerator facilities [26, 27, 28, 29] and a study to probe the pseudo-Dirac nature in JUNO using solar neutrinos can be found in [30]. Since these effects accumulate only over extremely large baselines, high-energy astrophysical neutrinos provide unique opportunities to probe such scenarios (see, for example, [31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42]).

Among all available neutrino sources, atmospheric neutrinos span energies from sub-GeV to multi-GeV while traversing baselines ranging from a few tens of kilometers to the Earth’s diameter. Consequently, atmospheric neutrinos probe an enormous range of L/EL/E, making them potentially sensitive to tiny mass-squared splittings. Furthermore, the propagation of atmospheric neutrinos through the Earth experiences matter effects that can significantly modify active–sterile oscillations.

In this work, we investigate pseudo-Dirac neutrinos at a DUNE-like detector. Motivated by existing experimental constraints, we concentrate on the pseudo-Dirac splitting associated with the third mass eigenstate, Δm33′2\Delta m_{33'}^2, which presently remains the least constrained among the pseudo-Dirac splittings and is most relevant for the oscillations of atmospheric neutrinos. Hence, we investigate pseudo-Dirac neutrinos within the framework of a 3+13+1 oscillation formalism. We study the modifications to neutrino oscillation probabilities, investigate their impact on atmospheric neutrinos, and assess the sensitivity of a DUNE-like detector to probe these effects via observations of atmospheric neutrinos.

The paper is organized as follows. Section II introduces the pseudo-Dirac neutrino framework together with the relevant theoretical motivation and the 3+13+1 oscillation formalism of interest. Section III describes the numerical methodology and the atmospheric neutrino analysis. The results and phenomenological implications are discussed in Section IV. Section V contains a discussion of how sensitive is the atmospheric neutrino data to ”solar” oscillations. Finally, Section VI summarizes our results.

Pseudo-Dirac Neutrinos

We start by postulating that the neutrino masses are a consequence of the existence of gauge-singlet “right-handed-neutrino” fields. In this case, the most general renormalizable Lagrangian consistent with SU(3)C×SU(2)L×U(1)Y\mathrm{SU}(3)_C \times\mathrm{SU}(2)_L \times\mathrm{U}(1)_Y gauge invariance is

LνSM=Lold−λαiLαHNi−∑i=1nMi2NiNi+H.c.,(2)\mathcal{L}_{\nu\mathrm{SM}}=\mathcal{L}_{\mathrm{old}}-\lambda_{\alpha i}L^\alpha HN^i-\sum_{i=1}^{n}\frac{M_i}{2}N^iN^i+\mathrm{H.c.}, \tag*{(2)}

where Lold\mathcal{L}_{\mathrm{old}} is the minimal SM Lagrangian, HH is the Higgs weak doublet, LαL^\alpha, α=e,μ,τ\alpha=e,\mu,\tau, are lepton weak doublets, λαi\lambda_{\alpha i} are neutrino Yukawa couplings, and MiM_i are Majorana masses for the right-handed neutrino fields NiN_i fields that have been added. The number of NiN^i fields is nn and, in order to fit the neutrino oscillation data, n≥2n\geq2. Henceforth, when relevant, we assume n=3n=3.

After electroweak symmetry breaking, the 6×66\times6 neutrino mass matrix in the (νLNc)\left(\begin{smallmatrix}\nu_L\\N^c\end{smallmatrix}\right) basis is

M=(0mDmDTM),(3)\mathcal{M}=\begin{pmatrix}0&m_D\\m_D^T&M\end{pmatrix}, \tag*{(3)}

where mD=λvm_D=\lambda v is the usual Dirac mass related to the Higgs vacuum expectation value vv. M=0M=0 yields the Dirac limit (i.e., pairs of mass-degenerate Majorana fermions form Dirac fermions). On the other hand, when M≫mDM\gg m_D, we can integrate out the right-handed neutrinos and end up with light active Majorana neutrinos (i.e., the seesaw mechanism).

The limit M≪mDM\ll m_D yields the pseudo-Dirac case. The neutrino masses are pairwise quasi-degenerate with magnitudes driven by the eigenvalues of mDm_D. The small mass splittings are proportional to MM. We refer to these as Δmii′2\Delta m_{ii'}^2, i=1,2,3i=1,2,3, where νi′\nu_i' and νi\nu_i are the six different pairwise-quasi-degenerate states.

Current constraints from solar neutrinos have greatly constrained two out of the three new oscillation frequencies Δmii′2\Delta m_{ii'}^2 associated with the ν1\nu_1 and ν2\nu_2 states [22, 23, 24, 25, 30]. Since solar neutrino experiments primarily involve electron neutrinos that have negligible mixing with ν3\nu_3, solar data provides comparatively weaker constraints on Δm33′2\Delta m^2_{33'}. Hence, bounds from Earth-bound experiments can compete with those from solar neutrinos in spite of the very long Earth-Sun distance.

Δm33′2\Delta m^2_{33'} is proportional to m32m_3^2 (see, for example, [22]), where m3m_3 is the average mass of ν3\nu_3 and ν3′\nu_{3'}. In turn, for the normal neutrino mass ordering and ignoring the possibility that all neutrino masses are quasi-degenerate, m32m_3^2 is of order Δm312\Delta m^2_{31}, the so-called atmospheric mass-squared difference. Hence, we define the relevant new oscillation frequency via

Δm33′2=2ϵΔm312,(4)\Delta m^2_{33'} = 2\epsilon\Delta m^2_{31}, \tag*{(4)}

where ϵ\epsilon is a dimensionless quantity. Throughout we use Δm33′2=5×10−3×ϵ eV2\Delta m^2_{33'} = 5 \times10^{-3} \times\epsilon\ \mathrm{eV}^2. We will further assume that the active–sterile mixing angle is maximal, as will become clearer in the next subsection.1 We refer to [22] for more details.

The pseudo-Dirac 3+1 oscillation formalism

In the limit Δm22′2=Δm11′2=0\Delta m^2_{22'} = \Delta m^2_{11'} = 0, the relevant 4×44 \times4 Hamiltonian that describes neutrino flavor evolution in the presence of matter is,

H3+1=12E[UM2U†+Hmat],(5)\mathcal{H}_{3+1} = \frac{1}{2E}\left[UM^2U^\dagger+ \mathcal{H}_{\mathrm{mat}}\right], \tag*{(5)}

where

U=(Ue1Ue2Ue32−Ue32Uμ1Uμ2Uμ32−Uμ32Uτ1Uτ2Uτ32−Uτ32001212).(6)U = \begin{pmatrix} U_{e1} & U_{e2} & \frac{U_{e3}}{\sqrt{2}} & -\frac{U_{e3}}{\sqrt{2}} \\ U_{\mu1} & U_{\mu2} & \frac{U_{\mu3}}{\sqrt{2}} & -\frac{U_{\mu3}}{\sqrt{2}} \\ U_{\tau1} & U_{\tau2} & \frac{U_{\tau3}}{\sqrt{2}} & -\frac{U_{\tau3}}{\sqrt{2}} \\ 0 & 0 & \frac{1}{\sqrt{2}} & \frac{1}{\sqrt{2}} \end{pmatrix}. \tag*{(6)}

UαiU_{\alpha i} are the elements of the standard PMNS mixing matrix, UPMNS=R23(θ23)R13(θ13)R12(θ12)U_{\mathrm{PMNS}} = R_{23}(\theta_{23})R_{13}(\theta_{13})R_{12}(\theta_{12}),

M2=(00000Δm2120000Δm312(1+ϵ)0000Δm312(1−ϵ)),(7)M^2 = \begin{pmatrix} 0 & 0 & 0 & 0 \\ 0 & \Delta m^2_{21} & 0 & 0 \\ 0 & 0 & \Delta m^2_{31}(1+\epsilon) & 0 \\ 0 & 0 & 0 & \Delta m^2_{31}(1-\epsilon) \end{pmatrix}, \tag*{(7)}

and

Hmat=(A00000000000000A2),(8)\mathcal{H}_{\mathrm{mat}} = \begin{pmatrix} A & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & \frac{A}{2} \end{pmatrix}, \tag*{(8)}

where A=2EVCCA = 2EV_{\mathrm{CC}} is the effective contribution in the Hamiltonian due to the matter potential VCC=2GFNeV_{\mathrm{CC}} = \sqrt{2}G_FN_e, with NeN_e being the ambient electron density. We briefly discuss approximate solutions to the neutrino oscillation probabilities in the presence of constant matter in Appendix § 0.2.1.

Atmospheric Neutrinos in a Liquid Argon Detector

In modeling our simulations of atmospheric neutrinos, we obtained the atmospheric neutrino flux from the publicly available Honda files using the solar minimum variant for the Homestake site [43]. Information regarding the total cross sections was obtained through GLoBES [44], adjusting for 40Ar{}^{40}\mathrm{Ar} as the target of a DUNE-like setup. To account for the matter effects that neutrinos experience along the different trajectories they take through the layers of the Earth from the atmosphere to the detector, we utilized the Preliminary Reference Earth Model (PREM) using an approximation with 5 layers: crust, upper mantle, lower mantle, outer core, and inner core [45].

We consider a setup similar to that proposed for DUNE [46]. DUNE is a long-baseline neutrino experiment under construction in South Dakota, with its far detector located approximately 1.5 km underground. The experiment will ultimately employ a 40 kt liquid argon time projection chamber (LArTPC) to detect neutrinos produced at Fermilab and traversing a baseline of 1300 km. In addition to its primary long-baseline neutrino program, the detector will also be sensitive to atmospheric, astrophysical, and solar neutrinos. The large detector mass, together with its excellent energy and angular resolution, makes this DUNE-like configuration well suited for the study presented here.

The analysis of atmospheric neutrino events is performed using an in-house code. The events per energy and cos⁡θ\cos\theta bin are calculated using

dNdEmdcos⁡θm=∬d(cos⁡θt) dEt ϕ(Et,cos⁡θt)RE(Et,Em)Rθ(θt,θm)Pαβ(Et)ϕ(Et,θt)Σ(Et).(9)\frac{dN}{dE_m d\cos\theta_m} = \iint d(\cos\theta_t)\,dE_t\,\phi(E_t,\cos\theta_t)R_E(E_t,E_m)R_\theta(\theta_t,\theta_m)P_{\alpha\beta}(E_t)\phi(E_t,\theta_t)\Sigma(E_t). \tag*{(9)}

Here ϕ(Et,cos⁡θt)\phi(E_t,\cos\theta_t) is the neutrino flux as a function of the energy and direction, and Σ(Et)\Sigma(E_t) is the cross section. We consider the energy range of Et∈[0.2,10.5]E_t \in[0.2, 10.5] GeV and cos⁡θt∈[−1,1]\cos\theta_t \in[-1,1] with 80 bins each. The energy resolutions used for detected electrons and muons are,

σE(μ)=0.1Et,σE(e)=0.2Et.(10)\sigma_E(\mu)=0.1\sqrt{E_t},\quad\sigma_E(e)=0.2\sqrt{E_t}. \tag*{(10)}

For the angular resolution, we consider

σθ=15π/180.0, for Et>1 GeV,(11)\sigma_\theta= 15\pi/180.0,\ \text{for } E_t > 1\ \text{GeV}, \tag*{(11)}
σθ=ϑπ/180.0, for Et<1 GeV,(12)\sigma_\theta= \vartheta\pi/180.0,\ \text{for } E_t < 1\ \text{GeV}, \tag*{(12)}

where we consider ϑ\vartheta (in degrees) values equal to 15, 30, 40, 50. The values for σE(μ),σE(e),σθ(Eν>1 GeV)\sigma_E(\mu),\sigma_E(e),\sigma_\theta(E_\nu> 1\ \text{GeV}) are consistent with those published by the DUNE collaboration in [47]. For σθ(Eν<1 GeV)\sigma_\theta(E_\nu< 1\ \text{GeV}), the resolution in [47] is most consistent with ϑ=50\vartheta=50. The Gaussian resolution functions are defined as

Rθ(θt,θm)=Nθexp⁡[−(cos⁡θt−cos⁡θm2σθsin⁡θt)2],(13)R_\theta(\theta_t,\theta_m)=N_\theta\exp\left[-\left(\frac{\cos\theta_t-\cos\theta_m}{\sqrt{2}\sigma_\theta\sin\theta_t}\right)^2\right], \tag*{(13)}
RE(Et,Em)=NEexp⁡[−(Et−Em2σE)2],(14)R_E(E_t,E_m)=N_E\exp\left[-\left(\frac{E_t-E_m}{\sqrt{2}\sigma_E}\right)^2\right], \tag*{(14)}

where the normalization constants are

Nθ=∫−1+1exp⁡[−(θt−θm2σθ)2]dcos⁡θm,(15)N_\theta=\int_{-1}^{+1}\exp\left[-\left(\frac{\theta_t-\theta_m}{\sqrt{2}\sigma_\theta}\right)^2\right]d\cos\theta_m, \tag*{(15)}
NE=∫Emin⁡Emax⁡exp⁡[−(Et−Em2σE)2]dEm.(16)N_E=\int_{E_{\min}}^{E_{\max}}\exp\left[-\left(\frac{E_t-E_m}{\sqrt{2}\sigma_E}\right)^2\right]dE_m. \tag*{(16)}

We have used 20 bins each for measured energy and cos⁡θ\cos\theta, with the corresponding ranges being the same as the true case.

We used the events per energy and angular bins for true (ex) and test (th) cases to calculate χ2\chi^2 defined as,

χ2=χstat2+χpull2=∑i∑α[Ni,αth+∑kCi,αkξk−Ni,αex]2Ni,αex+∑kξk2,(17)\chi^2=\chi^2_{\mathrm{stat}}+\chi^2_{\mathrm{pull}}=\sum_i\sum_\alpha\frac{\left[N^{\mathrm{th}}_{i,\alpha}+\sum_k C_{i,\alpha k}\xi_k-N^{\mathrm{ex}}_{i,\alpha}\right]^2}{N^{\mathrm{ex}}_{i,\alpha}}+\sum_k\xi_k^2, \tag*{(17)}

where ξk\xi_k are the pull variables and Ci,αkC_{i,\alpha k} are the systematic error coefficients corresponding to the pull variables for each flavor. The pull variables are flux error, zenith error, cross-section error, systematic error, and tilt error, proportional to the theoretical event rate Ni,αthN^{\mathrm{th}}_{i,\alpha}, and are listed in Table I. The zenith error is proportional to the measured angle, i.e., cos⁡θm\cos\theta_m. We minimize χ2\chi^2 by varying the oscillation parameters in the test (th) and obtain Δχ2\Delta\chi^2 relative to the value of χ2\chi^2 at the minimum.

We used the true and test values of oscillation parameters listed in Table II, following the global NuFit values [48]. We kept δCP=0∘\delta_{\mathrm{CP}}=0^\circ fixed in both true and test. The total exposure considered is 200 kt·yr, which is equivalent to 10 years of runtime with a 20 kton detector module, relevant in the case of a DUNE Phase I-like setup [46].

ParametersPercentage
Flux20%
Zenith Angle5%
Cross Section10%
Systematics5%
Tilt5%

Table 1. Different uncertainties used in our analysis.

ParametersTrue (NO)True (IO)Test Range
θ12\theta_{12}33.68∘33.68^\circ33.68∘33.68^\circ31.6∘31.6^\circ, 36.0∘36.0^\circ
θ13\theta_{13}8.56∘8.56^\circ8.56∘8.56^\circ8.15∘8.15^\circ, 8.9∘8.9^\circ
θ23\theta_{23}43.3∘43.3^\circ43.3∘43.3^\circ41.3∘41.3^\circ, 49.9∘49.9^\circ
Δm212\Delta m_{21}^{2}7.49×10−5 eV27.49 \times10^{-5}\ \mathrm{eV}^{2}7.49×10−5 eV27.49 \times10^{-5}\ \mathrm{eV}^{2}(6.9,8.0)×10−5 eV2(6.9, 8.0) \times10^{-5}\ \mathrm{eV}^{2}
Δm312\Delta m_{31}^{2}2.513×10−3 eV22.513 \times10^{-3}\ \mathrm{eV}^{2}−2.484×10−3 eV2-2.484 \times10^{-3}\ \mathrm{eV}^{2}±(2.45,2.59)×10−3 eV2\pm(2.45, 2.59) \times10^{-3}\ \mathrm{eV}^{2}

Table 2. The true and test values of the oscillation parameters

Sensitivity to the pseudo-Dirac Nature of ν3\nu_{3}

Before presenting the results for the sensitivity to the pseudo-Dirac nature of ν3\nu_{3}, we first illustrate in Figure 1 the differences in the oscillograms for the dominant muon-neutrino disappearance probability between the pseudo-Dirac hypothesis and the standard scenario for ϵ\epsilon equal to 0.001, 0.01, and 0.1. Non-zero values of the ϵ\epsilon parameter lead to distortions at sub-GeV energies and for negative values of cos⁡(θz)\cos(\theta_z). This is not surprising since a large-enough new oscillation phase is required for observable effects and negative cos⁡(θz)\cos(\theta_z) values correspond to longer baselines (and larger matter effects). The effect is seen to be very small – below a percent – for ϵ=0.001\epsilon= 0.001 but is order one for ϵ≳0.01\epsilon\gtrsim0.01.

Difference of the $\nu_\mu$ disappearance probability between the pseudo-Dirac hypothesis and the standard scenario

Figure 1. The difference of the νμ\nu_\mu disappearance probability between the pseudo-Dirac hypothesis and the standard scenario for ϵ=0.001\epsilon= 0.001 (top), ϵ=0.01\epsilon= 0.01 (middle), and ϵ=0.1\epsilon= 0.1 (bottom), as a function of the neutrino energy EνE_\nu and the neutrino direction cos⁡(θ)\cos(\theta).

Our estimates for the sensitivity to the pseudo-Dirac nature of ν3\nu_{3} are depicted in Figure 2. We highlight four options for the sub-GeV angular resolution: 15∘15^\circ, 30∘30^\circ, 40∘40^\circ, and 50∘50^\circ. The latter is the most conservative choice in our analysis and was inspired by results made public by the DUNE collaboration [47]. We assumed the neutrino mass ordering to be normal (NO); all other assumptions are listed within the panels of Figure 2. We present results for the inverted neutrino mass ordering (IO) in Appendix § 0.9. We include both the muon-neutrino and electron-neutrino “data” in our analyses; the sensitivity, however, is dominated by the muon-neutrino-associated samples.

Sensitivity curves for atmospheric neutrino data

Figure 2. Sensitivity to ϵ\epsilon of 200 kt·yr of atmospheric neutrino data at a DUNE-like detector, assuming NO. On the left is the sensitivity computed for fixed values of all other oscillation parameters for different values of the angular resolution below 1 GeV. On the right, we illustrate the effect of profiling over θ23\theta_{23} (green) and Δm312\Delta m^{2}_{31} (gold), relative to the sensitivity obtained for fixed values of all oscillation parameters (blue curve), for a fixed value of the angular resolution below 1 GeV.

In the left panel of Figure 2, all standard oscillation parameters are fixed to the values listed in Table II. In this case, the only free parameter is the value of the dimensionless parameter ϵ\epsilon that sets the hypothetical new oscillation frequency. These results indicate that the sub-GeV angular reconstruction significantly impacts the sensitivity. For the most conservative case presented here, a DUNE-like experiment is sensitive to ϵ≥4.0×10−2\epsilon\geq4.0 \times10^{-2} at the 3σ3\sigma level. This corresponds to Δm33′2≥2.0×10−4 eV2\Delta m_{33'}^{2} \geq2.0 \times10^{-4}\ \mathrm{eV}^{2}. For the more optimistic case we considered (15∘15^\circ angular smearing below 1 GeV), the 3σ3\sigma sensitivity limit is ϵ≥1.2×10−2\epsilon\geq1.2 \times10^{-2} which corresponds to Δm33′2≥6.0×10−5 eV2\Delta m_{33'}^{2} \geq6.0 \times10^{-5}\ \mathrm{eV}^{2}. For sub-GeV neutrino energies, Δm33′2∼5.0×10−5 eV2\Delta m_{33'}^{2} \sim5.0 \times10^{-5}\ \mathrm{eV}^{2} corresponds to an oscillation length of order the Earth’s diameter.

These results are competitive with similar studies in the literature, most notably a recent DUNE beam study [25]. This aligns with the fact that atmospheric and beam neutrinos probe, roughly, similar L/EL/E scales. We emphasize, however, that our result indicates a slightly worse relative sensitivity. Essentially, while atmospheric neutrino studies can take advantage of longer baselines and thus can potentially probe relatively small new mass-squared differences, most of the sensitivity lies where the detector angular resolution is the poorest (i.e., low neutrino energies). In other words, the access to longer baselines and relatively larger values of L/EL/E is offset by the experiment’s ability to determine L/EL/E precisely. We chose to report different instances of sub-GeV angular smearing for this reason.

To probe the robustness of our analysis and to account for the uncertainties present in the current neutrino oscillation global fit values, we also explored the impact of profiling over the individual oscillation parameters, taking the current uncertainties quoted in [48] into account. In the right-hand panel of Figure 2, we included the effects of profiling, one at a time, over the oscillation parameters that potentially most impact the sensitivity computed here: θ23\theta_{23} and Δm312\Delta m_{31}^{2}. Here, the angular resolution for low-energy neutrinos is fixed (50∘50^\circ angular resolution for neutrino energies below 1 GeV). We profile over the so-called atmospheric parameters because they contribute the most to the dominant oscillation amplitude and phase when it comes to the disappearance of atmospheric muon neutrinos.

Our results indicate that profiling over θ23\theta_{23} yields the strongest effect overall, followed by Δm312\Delta m_{31}^{2}. These results are consistent with expectations from a qualitative inspection of the leading-order terms in the oscillation probabilities

presented in Appendix § 0.2.1.

In a little more detail, terms like cos⁡2θ34msin⁡2θ23\cos^{2}\theta_{34}^{m}\sin2\theta_{23}, sin⁡2θ34msin⁡2θ23\sin^{2}\theta_{34}^{m}\sin2\theta_{23}, and sin⁡2θ34msin⁡4θ23\sin^{2}\theta_{34}^{m}\sin^{4}\theta_{23} in the probabilities indicate a correlation between ϵ\epsilon and θ23\theta_{23}. This is illustrated in Figure 3. In the left panel of Figure 3, we depict the correlation between ϵ\epsilon and θ23\theta_{23}. In particular, the red curve depicts the values of the test θ23\theta_{23} at which the minimum χ2\chi^{2} is obtained for a given value of ϵ\epsilon. These differ from the true θ23=43.3∘\theta_{23}=43.3^\circ (cyan line) for large-enough values of ϵ\epsilon; the correlation is clear. Hence, profiling over θ23\theta_{23} yields a relatively lower Δχ2\Delta\chi^{2} for higher values of ϵ\epsilon. Similarly, using Eqs. (27), (28) one can see that higher values of ϵ\epsilon can be partially compensated by shifting the vale of Δm312\Delta m^{2}_{31}. This is reflected in the red curve in the right-hand panel of Figure 3. It indicates that the minimum χ2\chi^{2} occurs at higher test-Δm312\Delta m^{2}_{31} values relative to the true value (blue line) as ϵ\epsilon increases. Quantitatively, for ϵ≲0.1\epsilon\lesssim0.1, the θ23\theta_{23}-related effects are more pronounced.

Sensitivity in the $\epsilon \times \sin^{2}\theta_{23}$ and $\epsilon \times \Delta m^{2}_{31}$ planes

Figure 3. Sensitivity in the ϵ×sin⁡2θ23\epsilon\times\sin^{2}\theta_{23} (left), and ϵ×Δm312\epsilon\times\Delta m^{2}_{31} plane (right) for NO. The red curve shows the position of the minimum value of χ2\chi^{2} as a function of ϵ\epsilon. The blue line indicates the true values on the relevant standard oscillation parameter.

Sensitivity to Solar Parameters

Since the sensitivity to the pseudo-Dirac mass splitting is close to that of the solar MSW mass-squared difference, it is worthwhile to investigate the sensitivity of the atmospheric neutrinos at a DUNE-like experiment to the solar parameters Δm212\Delta m^{2}_{21} and sin⁡2θ12\sin^{2}\theta_{12}. Similar studies were performed in [49] in the context of SuperKamiokande data.

We use as input for the standard 3ν3\nu-oscillation parameters the values listed in Table II. All other oscillation parameters, apart from the solar under investigation, have been kept fixed in the analysis.

For atmospheric neutrinos, the leading-order dependence on θ12\theta_{12} or Δm212\Delta m^{2}_{21} is subdominant, suggesting weak sensitivity to these parameters. Figure 4 depicts the sensitivity to Δm212\Delta m^{2}_{21} (left panel) and sin⁡2θ12\sin^{2}\theta_{12} (right panel). In the left (right) panel, the value of θ12\theta_{12} (Δm212\Delta m^{2}_{21}) is kept fixed at the value tabulated in Table II, and we assume NO. There are two curves in each panel indicating the different values of ϑ\vartheta that set the angular resolution for Eν<1E_{\nu}<1 GeV: 50∘50^\circ and 15∘15^\circ. As expected, the uncertainty in the direction of low-energy neutrinos has a measurable impact. Atmospheric neutrino data can only set a modest, resolution-dependent limit on Δm212\Delta m^{2}_{21}, even when Δm312\Delta m^{2}_{31} is assumed to be perfectly known. The hypothesis Δm212=0\Delta m^{2}_{21}=0 cannot be ruled out at more than 2σ2\sigma, but values of Δm212\Delta m^{2}_{21} larger than 4×10−44\times10^{-4} eV2^{2} can be ruled out at the 3σ3\sigma level even when we assume the more pessimistic uncertainty on the direction of low-energy neutrinos. On the other hand, the data can place virtually no bound on sin⁡2θ12\sin^{2}\theta_{12}.

Sensitivity to $\Delta m^{2}_{21}$ and $\sin^{2}\theta_{12}$

Figure 4. Sensitivity to Δm212\Delta m^{2}_{21} (left) and sin⁡2θ12\sin^{2}\theta_{12} (right) for standard oscillations, assuming NO and 200 kt⋅\cdotyr of data. The dotted and dashed lines correspond to different choices for the angular resolution when Eν<1E_{\nu}<1 GeV: 50∘50^\circ (most conservative) and 15∘15^\circ (most optimistic), respectively.

Figure 5 depicts the sensitivity to the solar oscillation parameters in the Δm122×sin⁡2θ12\Delta m^{2}_{12}\times\sin^{2}\theta_{12} plane for two different angular resolutions associated with low-energy atmospheric neutrinos: 15∘15^\circ on the left, 50∘50^\circ on the right. As before, we have assumed NO. The figure demonstrates strong Δm212−sin⁡2θ12\Delta m^{2}_{21}-\sin^{2}\theta_{12} correlation, as is expected when L/EL/E is small. The sensitivity, even with the most optimistic angular resolution, allows all values of sin⁡2θ12\sin^{2}\theta_{12} for Δm212≲15×10−5\Delta m^{2}_{21}\lesssim15\times10^{-5} eV2^{2}.

Sensitivity to the solar oscillation parameters in the $\Delta m^{2}_{21} \times \sin^{2}\theta_{12}$-plane

Figure 5. Sensitivity to the solar oscillation parameters in the Δm212×sin⁡2θ12\Delta m^{2}_{21} \times\sin^{2}\theta_{12}-plane for standard oscillations assuming 200 kt·yr of data. We assume NO. The left panel corresponds to an angular resolution for Eν<1E_{\nu}<1 GeV of 15∘15^\circ (most optimistic), and the right panel corresponds to 50∘50^\circ (most conservative). The star in each panel indicates the best-fit value from NuFit [48].

The simulated data is most sensitive to νμ\nu_{\mu} disappearance, so the information on the solar mixing angle comes, in part, from Uμ2U_{\mu2} (or Uμ1U_{\mu1}; they are related to Uμ3U_{\mu3} through the unitarity of the mixing matrix). Since we fixed δ=0\delta=0 throughout,

Uμ2=cos⁡θ23cos⁡θ12−sin⁡θ13sin⁡θ23sin⁡θ12,(18)U_{\mu2}=\cos\theta_{23}\cos\theta_{12}-\sin\theta_{13}\sin\theta_{23}\sin\theta_{12}, \tag*{(18)}

and, numerically,

∣Uμ2∣2≃0.5cos⁡2θ12−0.08sin⁡2θ12+0.01sin⁡2θ12.(19)\lvert U_{\mu2}\rvert^{2}\simeq0.5\cos^{2}\theta_{12}-0.08\sin2\theta_{12}+0.01\sin^{2}\theta_{12}. \tag*{(19)}

In particular, ∣Uμ2∣2\lvert U_{\mu2}\rvert^{2} decreases as sin⁡2θ12\sin^{2}\theta_{12} increases for sin⁡2θ12≲0.98\sin^{2}\theta_{12}\lesssim0.98.

Conclusions

We investigated the sensitivity of atmospheric neutrinos at a DUNE-like experiment to the pseudo-Dirac nature of the third neutrino mass eigenstate. We considered the scenario in which the active and sterile components are maximally mixed and the new physics is characterized by a dimensionless parameter, ϵ\epsilon, related to the additional mass-squared splitting through Δm33′2=2ϵΔm312\Delta m^{2}_{33'}=2\epsilon\Delta m^{2}_{31}. Atmospheric neutrinos are suitable for this search because they cover a broad range of energies (roughly 0.1−1000.1-100 GeV) and propagation distances (roughly 2020–1300013000 km) and hence provide access to several orders of L/EL/E that are not accessible to other conventional terrestrial neutrino sources.

Assuming an exposure of 200 kt·yr, we studied the sensitivity to ϵ\epsilon, concentrating on the normal neutrino mass ordering. Results for IO can be found in Appendix § 0.9. The sensitivity depends strongly on the angular reconstruction of sub-GeV atmospheric-neutrino events. For a sub-GeV angular resolution of 50∘50^\circ, the most conservative scenario considered here, if the data are consistent with the standard oscillation scenario, we expect it to translate into a 3σ3\sigma upper bound of

ϵ<4.0×10−2,orΔm33′2<2.0×10−4 eV2.\epsilon< 4.0 \times10^{-2}, \qquad\text{or} \qquad\Delta m^{2}_{33'} < 2.0 \times10^{-4}\,\mathrm{eV}^{2}.

Given a more optimistic angular resolution of 15∘15^\circ, future atmospheric data should be able to constrain

ϵ<1.2×10−2,orΔm33′2<6×10−5 eV2.\epsilon< 1.2 \times10^{-2}, \qquad\text{or} \qquad\Delta m^{2}_{33'} < 6 \times10^{-5}\,\mathrm{eV}^{2}.

The angular resolution at energies below 1 GeV plays a central role in the sensitivity of atmospheric neutrinos to new, long oscillation lengths.

We further examined the robustness of the sensitivity to the pseudo-Dirac nature of ν3\nu_{3} against uncertainties in the standard oscillation parameters. Profiling over θ23\theta_{23} and Δm312\Delta m^{2}_{31} within the current allowed ranges weakens the sensitivity relative to the case in which all standard parameters are fixed. The effect is more pronounced for θ23\theta_{23}. A more detailed determination of the pseudo-Dirac sensitivity must therefore account for correlations with the parameters governing the dominant atmospheric oscillation amplitude and frequency. Nevertheless, the atmospheric-neutrino reach remains comparable to that of related terrestrial searches and provides complementary information.

We also explored the sensitivity of atmospheric neutrinos in a DUNE-like detector to the solar parameters Δm212\Delta m^{2}_{21} and sin⁡2θ12\sin^{2}\theta_{12}. This question is similar to that of the pseudo-Dirac nature of ν3\nu_{3}, i.e., it is also related to the sensitivity to new long oscillation lengths. We find that the sensitivity to the solar oscillation parameters is limited. This is consistent with the bounds on Δm33′2\Delta m^{2}_{33'} summarized above, which not coincidentally, are of order Δm212\Delta m^{2}_{21}.

Our analysis highlights the potential and limitations of atmospheric neutrinos as complementary probes of the pseudo-Dirac nature of light neutrinos. It underscores the importance of the angular reconstruction of sub-GeV neutrinos, control of systematic uncertainties, and the treatment of correlations among the known oscillation parameters as central ingredients in determining the final experimental reach.

We anticipate that atmospheric neutrinos collected by the Hyper-Kamiokande experiment [50] should have similar sensitivity to the pseudo-Dirac nature of ν3\nu_3. Ultimately, the combination of atmospheric-neutrino observations with accelerator, reactor, solar, and astrophysical measurements will provide coverage across several orders of magnitude in pseudo-Dirac mass splittings.

Note added: While some of the authors are members of the DUNE collaboration, this work is not officially associated with DUNE; only publicly available detector specifications were used to simulate and analyze hypothetical data in a DUNE-like detector.

Acknowledgments

SG acknowledges Northwestern University, where the work was initiated, for hospitality and the Fulbright-Nehru Academic and Professional Excellence fellowship for funding her visit to the US. SP acknowledges the funding from ANID-Chile under FONDECYT Postdoctorado No. 3250376. The work of OMB and AdG was supported in part by the US Department of Energy (DOE) grants #de-sc0010143 and #de-sc0010813 and in part by the National Science Foundation under Grant Nos. PHY-1630782 and PHY-1748958.

Approximate Analytic Expressions for the Oscillation Probabilities

We can write the total Hamiltonian that governs the 3+1 system of interest as the sum of two Hamiltonians,

H3+1=12E(H(0)+H(1)).(20)\mathcal{H}_{3+1} = \frac{1}{2E}\left(\mathcal{H}^{(0)}+\mathcal{H}^{(1)}\right). \tag*{(20)}

H(0)\mathcal{H}^{(0)} is zeroth-order Hamiltonian proportional to Δm312\Delta m_{31}^{2} and H(1)\mathcal{H}^{(1)} is the perturbative Hamiltonian containing the small terms proportional to ϵ\epsilon and Δm212\Delta m_{21}^{2}. We use perturbation theory, considering Δm212\Delta m_{21}^{2} (in units of Δm312\Delta m_{31}^{2}) and ϵ\epsilon as small parameters, to calculate the eigenvalues of the Hamiltonian and the oscillation frequencies in matter. Throughout this appendix, we assume the matter potential AA is constant. The eigenvalues of H3+1/2E\mathcal{H}_{3+1}/2E are,

λ1m=12(Δm212+Ac132−Δ21m),(21)\lambda_{1}^{m} = \frac{1}{2}\left(\Delta m_{21}^{2}+Ac_{13}^{2}-\Delta_{21}^{m}\right), \tag*{(21)}
λ2m=12(Δm212+Ac132+Δ21m),(22)\lambda_{2}^{m} = \frac{1}{2}\left(\Delta m_{21}^{2}+Ac_{13}^{2}+\Delta_{21}^{m}\right), \tag*{(22)}
λ3m=Δm312+14(2A−Acos⁡2θ13)−12Δ43m,(23)\lambda_{3}^{m} = \Delta m_{31}^{2}+\frac{1}{4}\left(2A-A\cos2\theta_{13}\right)-\frac{1}{2}\Delta_{43}^{m}, \tag*{(23)}
λ4m=Δm312+14(2A−Acos⁡2θ13)+12Δ43m,(24)\lambda_{4}^{m} = \Delta m_{31}^{2}+\frac{1}{4}\left(2A-A\cos2\theta_{13}\right)+\frac{1}{2}\Delta_{43}^{m}, \tag*{(24)}

where the modified mass-squared differences are,

Δ21m=(Δm212cos⁡2θ12−Ac132)2+(Δm212sin⁡2θ12)2,(25)\Delta_{21}^{m} = \sqrt{\left(\Delta m_{21}^{2}\cos2\theta_{12}-Ac_{13}^{2}\right)^{2}+\left(\Delta m_{21}^{2}\sin2\theta_{12}\right)^{2}}, \tag*{(25)}
Δ43m=A2cos⁡22θ13+(4Δm312ϵ)2=A2cos⁡22θ13+(2Δm33′2)2,(26)\Delta_{43}^{m} = \sqrt{A^{2}\cos^{2}2\theta_{13}+\left(4\Delta m_{31}^{2}\epsilon\right)^{2}}=\sqrt{A^{2}\cos^{2}2\theta_{13}+\left(2\Delta m_{33'}^{2}\right)^{2}}, \tag*{(26)}
Δ31m=Δ31+14(2A−Acos⁡2θ13)−12Δ43m−12(Δ21+Ac132−Δ21m),(27)\Delta_{31}^{m} = \Delta_{31}+\frac{1}{4}\left(2A-A\cos2\theta_{13}\right)-\frac{1}{2}\Delta_{43}^{m}-\frac{1}{2}\left(\Delta_{21}+Ac_{13}^{2}-\Delta_{21}^{m}\right), \tag*{(27)}
Δ32m=Δ31+14(2A−Acos⁡2θ13)−12Δ43m−12(Δ21+Ac132+Δ21m).(28)\Delta_{32}^{m} = \Delta_{31}+\frac{1}{4}\left(2A-A\cos2\theta_{13}\right)-\frac{1}{2}\Delta_{43}^{m}-\frac{1}{2}\left(\Delta_{21}+Ac_{13}^{2}+\Delta_{21}^{m}\right). \tag*{(28)}

We can write the mixing matrix in matter, obtained from the eigenvectors of the Hamiltonian, as

Upd=R23⋅(c12mc13ms12mc13ms13mc34ms13ms34m−s12mc12m00−c12ms13m−s12ms13mc13mc34mc13ms34m00−s34mc34m)≡R23(θ23).R12m(θ12m).R13m(θ13m).R34m(θ34m).(29)U_{\mathrm{pd}}=R_{23}\cdot \begin{pmatrix} c_{12}^{m}c_{13}^{m} & s_{12}^{m}c_{13}^{m} & s_{13}^{m}c_{34}^{m} & s_{13}^{m}s_{34}^{m}\\ -s_{12}^{m} & c_{12}^{m} & 0 & 0\\ -c_{12}^{m}s_{13}^{m} & -s_{12}^{m}s_{13}^{m} & c_{13}^{m}c_{34}^{m} & c_{13}^{m}s_{34}^{m}\\ 0 & 0 & -s_{34}^{m} & c_{34}^{m} \end{pmatrix} \equiv R_{23}(\theta_{23}).R_{12}^{m}(\theta_{12}^{m}).R_{13}^{m}(\theta_{13}^{m}).R_{34}^{m}(\theta_{34}^{m}). \tag*{(29)}

Along with the “matter” mixing angles θ12m\theta^{m}_{12} and θ13m\theta^{m}_{13}, there is also an effective mixing angle θ34m\theta^{m}_{34}. It turns out (cf. Eq. (32)) that θ34m\theta^{m}_{34} is very small when Δm312ϵ≪Acos⁡2θ13\Delta m^{2}_{31}\epsilon\ll A\cos2\theta_{13}. Therefore, terms in the oscillation probabilities proportional to sin⁡2θ34m\sin^{2}\theta^{m}_{34} will be sub-leading. Using perturbation theory, the modified mixing angles satisfy

sin⁡θ13m=sin⁡θ13[1+Acos⁡2θ13Δm312],cos⁡θ13m=cos⁡θ13[1−Asin⁡2θ13Δm312].(30)\sin\theta^{m}_{13}=\sin\theta_{13}\left[1+\frac{A\cos^{2}\theta_{13}}{\Delta m^{2}_{31}}\right],\quad\cos\theta^{m}_{13}=\cos\theta_{13}\left[1-\frac{A\sin^{2}\theta_{13}}{\Delta m^{2}_{31}}\right]. \tag*{(30)}
sin⁡2θ12m=Δm212sin⁡2θ12Δm21m,cos⁡2θ12m=Δm212cos⁡2θ12−Acos⁡2θ13Δm21m.(31)\sin2\theta^{m}_{12}=\frac{\Delta m^{2}_{21}\sin2\theta_{12}}{\Delta m^{m}_{21}},\quad\cos2\theta^{m}_{12}=\frac{\Delta m^{2}_{21}\cos2\theta_{12}-A\cos^{2}\theta_{13}}{\Delta m^{m}_{21}}. \tag*{(31)}
sin⁡2θ34m=4Δm312ϵΔm43m=2Δm33′2Δm43m,cos⁡2θ34m=Acos⁡2θ13Δm43m.(32)\sin2\theta^{m}_{34}=\frac{4\Delta m^{2}_{31}\epsilon}{\Delta m^{m}_{43}}=\frac{2\Delta m^{2}_{33'}}{\Delta m^{m}_{43}},\quad\cos2\theta^{m}_{34}=\frac{A\cos2\theta_{13}}{\Delta m^{m}_{43}}. \tag*{(32)}

The matter equivalent version of Δm212\Delta m^{2}_{21} is modified relative to its vacuum value by a function of the matter potential AA and the mixing angles θ13\theta_{13} and θ12\theta_{12}. On the other hand, the matter equivalent to the atmospheric mass-squared difference is a function of both the matter potential and the size of the pseudo-Dirac splitting, proportional to ϵ\epsilon. A modified mass-squared difference Δm43m\Delta m^{m}_{43} also arises in matter and is a function of AA and ϵ\epsilon. However, it is a stronger function of AA, as seen in Eq. (26).

To express the oscillation probabilities in constant matter, we define

P0=s13mcos⁡(2θ12m)sin⁡(2θ23)+sin⁡(2θ12m)(c232−(s13m)2s232),(33)P^{0}=s^{m}_{13}\cos(2\theta^{m}_{12})\sin(2\theta_{23})+\sin(2\theta^{m}_{12})\left(c^{2}_{23}-(s^{m}_{13})^{2}s^{2}_{23}\right), \tag*{(33)}
Pμμ1=[s12mc13msin⁡(2θ23)+c12ms232sin⁡(2θ13m)]2,(34)P^{1}_{\mu\mu}=\left[s^{m}_{12}c^{m}_{13}\sin(2\theta_{23})+c^{m}_{12}s^{2}_{23}\sin(2\theta^{m}_{13})\right]^{2}, \tag*{(34)}
Pμμ2=(c13m)2sin⁡2(2θ23)+s234sin⁡2(2θ13m),(35)P^{2}_{\mu\mu}=(c^{m}_{13})^{2}\sin^{2}(2\theta_{23})+s^{4}_{23}\sin^{2}(2\theta^{m}_{13}), \tag*{(35)}
Pμe1=(c13m)2s13msin⁡(2θ23)sin⁡(2θ12m)+(c12m)2s232sin⁡2(2θ13m).(36)P^{1}_{\mu e}=(c^{m}_{13})^{2}s^{m}_{13}\sin(2\theta_{23})\sin(2\theta^{m}_{12})+(c^{m}_{12})^{2}s^{2}_{23}\sin^{2}(2\theta^{m}_{13}). \tag*{(36)}

The appearance and disappearance probabilities in matter in the presence of a pseudo-Dirac-related Δm33′2\Delta m^{2}_{33'} are

Pμe=(c13m)2sin⁡(2θ12m)P0sin⁡2Φ21m+(c34m)2[Pμe1(sin⁡2Φ31m−sin⁡2Φ32m)+s232sin⁡2(2θ13m)sin⁡2Φ32m]+(s34m)2[Pμe1(sin⁡2Φ41m−sin⁡2Φ42m)+sin⁡2(2θ13m)s232(sin⁡2Φ42m−(c34m)2sin⁡2Φ43m)].(37)\begin{aligned} P_{\mu e}={}&(c^{m}_{13})^{2}\sin(2\theta^{m}_{12})P^{0}\sin^{2}\Phi^{m}_{21}+(c^{m}_{34})^{2}\left[P^{1}_{\mu e}\left(\sin^{2}\Phi^{m}_{31}-\sin^{2}\Phi^{m}_{32}\right)+s^{2}_{23}\sin^{2}(2\theta^{m}_{13})\sin^{2}\Phi^{m}_{32}\right]\\ &+(s^{m}_{34})^{2}\left[P^{1}_{\mu e}\left(\sin^{2}\Phi^{m}_{41}-\sin^{2}\Phi^{m}_{42}\right)+\sin^{2}(2\theta^{m}_{13})s^{2}_{23}\left(\sin^{2}\Phi^{m}_{42}-(c^{m}_{34})^{2}\sin^{2}\Phi^{m}_{43}\right)\right]. \tag*{(37)} \end{aligned}
Pee=1−(c13m)4sin⁡2(2θ12m)sin⁡2Φ21m+(c34m)2[(c12m)2sin⁡2(2θ13m)sin⁡2Φ31m+(s12m)2sin⁡2(2θ13m)sin⁡2Φ32m]+(s34m)2[(c12m)2sin⁡2(2θ13m)sin⁡2Φ41m+(s12m)2sin⁡2(2θ13m)sin⁡2Φ42m]+(s13m)4sin⁡2(2θ34m)sin⁡2Φ43m.(38)\begin{aligned} P_{ee}=1-&(c^{m}_{13})^{4}\sin^{2}(2\theta^{m}_{12})\sin^{2}\Phi^{m}_{21}+(c^{m}_{34})^{2}\left[(c^{m}_{12})^{2}\sin^{2}(2\theta^{m}_{13})\sin^{2}\Phi^{m}_{31}+(s^{m}_{12})^{2}\sin^{2}(2\theta^{m}_{13})\sin^{2}\Phi^{m}_{32}\right]\\ &+(s^{m}_{34})^{2}\left[(c^{m}_{12})^{2}\sin^{2}(2\theta^{m}_{13})\sin^{2}\Phi^{m}_{41}+(s^{m}_{12})^{2}\sin^{2}(2\theta^{m}_{13})\sin^{2}\Phi^{m}_{42}\right]+(s^{m}_{13})^{4}\sin^{2}(2\theta^{m}_{34})\sin^{2}\Phi^{m}_{43}. \tag*{(38)} \end{aligned}
Pμμ=1−(P0)2sin⁡2Φ21m−(c34m)2[Pμμ1(sin⁡2Φ31m−sin⁡2Φ32m)+Pμμ2sin⁡2Φ32m]−(s34m)2[Pμμ1(sin⁡2Φ41m−sin⁡2Φ42m)+Pμμ2sin⁡2Φ42m]−(c13m)4s432sin⁡2(2θ34m)sin⁡2Φ43m.(39)\begin{aligned} P_{\mu\mu}=1-(P^{0})^{2}\sin^{2}\Phi^{m}_{21}-(c^{m}_{34})^{2}\left[P^{1}_{\mu\mu}\left(\sin^{2}\Phi^{m}_{31}-\sin^{2}\Phi^{m}_{32}\right)+P^{2}_{\mu\mu}\sin^{2}\Phi^{m}_{32}\right]\\ &-(s^{m}_{34})^{2}\left[P^{1}_{\mu\mu}\left(\sin^{2}\Phi^{m}_{41}-\sin^{2}\Phi^{m}_{42}\right)+P^{2}_{\mu\mu}\sin^{2}\Phi^{m}_{42}\right]-(c^{m}_{13})^{4}s^{2}_{43}\sin^{2}(2\theta^{m}_{34})\sin^{2}\Phi^{m}_{43}. \tag*{(39)} \end{aligned}

where Φijm≡ΔmijmL/4E\Phi^{m}_{ij}\equiv\Delta m^{m}_{ij}L/4E.

A few points related to the expressions above are worthy of note:

  • The new oscillation frequencies Φ41m\Phi^{m}_{41}, Φ42m\Phi^{m}_{42}, Φ43m\Phi^{m}_{43} terms are subleading due to the smallness of sin⁡2θ34m\sin^{2}\theta^{m}_{34}.

  • The presence of terms with both θ34m\theta^{m}_{34} and θ23\theta_{23} in the amplitude, e.g., sin⁡22θ34m\sin^{2}2\theta^{m}_{34} together with sin⁡4θ23\sin^{4}\theta_{23} in PμμP_{\mu\mu}, translate into a correlation between ϵ\epsilon and θ23\theta_{23}; an increase in ϵ\epsilon (and hence sin⁡22θ34m\sin^{2}2\theta^{m}_{34}) can be compensated by a decrease in θ23\theta_{23}.

  • There is a correlation between Δm312\Delta m^{2}_{31} and Δm33′2\Delta m^{2}_{33'}. These can be in the expressions for Δm31m\Delta m^{m}_{31} and Δm32m\Delta m^{m}_{32} (using Δm43m\Delta m^{m}_{43}); larger values of Δm312\Delta m^{2}_{31} can be compensated by an increase in ϵ\epsilon and hence of Δm43m\Delta m^{m}_{43}.

For completeness, we list the expressions for the oscillation probabilities in the standard oscillation scenario – three active flavors – under the same conditions:

Pμe=(c13m)2sin⁡(2θ12m)P0sin⁡2Φ21m+[Pμe1(sin⁡2Φ31−sin⁡2Φ32)+s232sin⁡2(2θ13m)sin⁡2Φ32],(40)P_{\mu e} = (c_{13}^{m})^2 \sin(2\theta_{12}^{m}) P^0 \sin^2 \Phi_{21}^{m} + \left[P_{\mu e}^{1}\left(\sin^2 \Phi_{31} - \sin^2 \Phi_{32}\right) + s_{23}^{2}\sin^2(2\theta_{13}^{m})\sin^2 \Phi_{32}\right], \tag*{(40)}
Pee=1−(c13m)4sin⁡2(2θ12m)sin⁡2Φ21m+sin⁡2(2θ13m)[(c13m)2sin⁡2Φ31+(s13m)2sin⁡2Φ32],(41)P_{ee} = 1 - (c_{13}^{m})^4 \sin^2(2\theta_{12}^{m})\sin^2 \Phi_{21}^{m} + \sin^2(2\theta_{13}^{m})\left[(c_{13}^{m})^2\sin^2\Phi_{31} + (s_{13}^{m})^2\sin^2\Phi_{32}\right], \tag*{(41)}
Pμμ=1−(P0)2sin⁡2Φ21m−[Pμμ1(sin⁡2Φ31−sin⁡2Φ32)+Pμμ2sin⁡2Φ32],(42)P_{\mu\mu} = 1 - (P^0)^2\sin^2\Phi_{21}^{m} - \left[P_{\mu\mu}^{1}\left(\sin^2\Phi_{31} - \sin^2\Phi_{32}\right) + P_{\mu\mu}^{2}\sin^2\Phi_{32}\right], \tag*{(42)}

where Φ31\Phi_{31} and Φ32\Phi_{32} are

Φ31=L4E(Δ31+14(2A−Acos⁡2θ13)−12(Δ21+Ac132−Δ21m)),(43)\Phi_{31} = \frac{L}{4E}\left(\Delta_{31} + \frac{1}{4}(2A - A\cos2\theta_{13}) - \frac{1}{2}\left(\Delta_{21} + Ac_{13}^{2} - \Delta_{21}^{m}\right)\right), \tag*{(43)}
Φ32=L4E(Δ31+14(2A−Acos⁡2θ13)−12(Δ21+Ac132+Δ21m)).(44)\Phi_{32} = \frac{L}{4E}\left(\Delta_{31} + \frac{1}{4}(2A - A\cos2\theta_{13}) - \frac{1}{2}\left(\Delta_{21} + Ac_{13}^{2} + \Delta_{21}^{m}\right)\right). \tag*{(44)}

Results for the Inverted Ordering

Here we collect the results obtained assuming that the neutrino mass ordering is inverted. Note that, in this case, the connection between Δm33′2\Delta m_{33'}^{2} and ∣Δm312∣\lvert\Delta m_{31}^{2}\rvert is not especially natural since, unless all neutrino masses are degenerate, m32≪∣Δm312∣m_{3}^{2} \ll\lvert\Delta m_{31}^{2}\rvert.

Sensitivity to $\epsilon$ of atmospheric neutrino data at a DUNE-like detector assuming IO

Figure 6. Sensitivity to ϵ\epsilon of 200 kt·yr of atmospheric neutrino data at a DUNE-like detector, assuming IO. On the left is the sensitivity computed for fixed values of all other oscillation parameters for different values of the angular resolution below 1 GeV. On the right, we illustrate the effect of profiling over θ23\theta_{23} (green) and Δm312\Delta m_{31}^{2} (gold), relative to the sensitivity obtained for fixed values of all oscillation parameters (blue curve), for a fixed value of the angular resolution below 1 GeV.

Sensitivity in the $\epsilon \times \sin^{2}\theta_{23}$ and $\epsilon \times \Delta m_{31}^{2}$ planes for IO

Figure 7. Sensitivity in the ϵ×sin⁡2θ23\epsilon\times\sin^{2}\theta_{23} (left), and ϵ×Δm312\epsilon\times\Delta m_{31}^{2} plane (right) for IO. The red curve shows the position of the minimum value of χ2\chi^{2} as a function of ϵ\epsilon. The blue line indicates the true values on the relevant standard oscillation parameter.

Sensitivity to $\Delta m_{21}^{2}$ and $\sin^{2}\theta_{12}$ for standard oscillations

Figure 8. Sensitivity to Δm212\Delta m_{21}^{2} (left) and sin⁡2θ12\sin^{2}\theta_{12} (right) for standard oscillations, assuming IO and 200 kt·yr of data. The dotted and dashed correspond to different choices for the angular resolution when Eν<1E_{\nu} < 1 GeV: 50° (most conservative) and 15° (most optimistic), respectively.

Sensitivity to the solar oscillation parameters

Figure 9. Sensitivity to the solar oscillation parameters in the Δm212×sin⁡2θ12\Delta m_{21}^{2} \times\sin^{2}\theta_{12}-plane for standard oscillations assuming 200 kt·yr of data. We assume IO. The left panel corresponds to an angular resolution for Eν<1 GeVE_{\nu} < 1\,\mathrm{GeV} of 15∘15^\circ (most optimistic), and the right panel corresponds to 50∘50^\circ (most conservative). The star in each panel indicates the best-fit value from NuFit [48].

References

  1. [1]F. Takahashi et al. (Particle Data Group), Review of Particle Physics, Int. J. Mod. Phys. A 41, 2630011 (2026).
  2. [2]R. Jimenez, C. P. Garay, F. Simpson, and L. Verde, From evidence to evident: Decisive cosmological evidence for the normal neutrino mass hierarchy (2026), arXiv:2606.18987 [astro-ph.CO].arxiv.org/abs/2606.18987
  3. [3]S. Weinberg, Baryon and Lepton Nonconserving Processes, Phys. Rev. Lett. 43, 1566 (1979).
  4. [4]P. Minkowski, μ → eγ at a Rate of One Out of 10⁹ Muon Decays?, Phys. Lett. B 67, 421 (1977).
  5. [5]T. Yanagida, Horizontal Gauge Symmetry and Masses of Neutrinos, in Proceedings of the Workshop on the Unified Theory and the Baryon Number in the Universe, edited by O. Sawada and A. Sugamoto (Tsukuba, Japan, 1979) p. 95.
  6. [6]M. Gell-Mann, P. Ramond, and R. Slansky, Complex Spinors and Unified Theories, in Supergravity, edited by P. van Nieuwenhuizen and D. Z. Freedman (North-Holland, Amsterdam, 1979) pp. 315–321, arXiv:1306.4609 [hep-th].
  7. [7]R. N. Mohapatra and G. Senjanovic, Neutrino Mass and Spontaneous Parity Nonconservation, Phys. Rev. Lett. 44, 912 (1980).DOI
  8. [8]W. Konetschny and W. Kummer, Nonconservation of Total Lepton Number with Scalar Bosons, Phys. Lett. B 70, 433 (1977).DOI
  9. [9]M. Magg and C. Wetterich, Neutrino Mass Problem and Gauge Hierarchy, Phys. Lett. B 94, 61 (1980).DOI
  10. [10]T. P. Cheng and L.-F. Li, Neutrino Masses, Mixings and Oscillations in SU(2) x U(1) Models of Electroweak Interactions, Phys. Rev. D 22, 2860 (1980).DOI
  11. [11]J. Schechter and J. W. F. Valle, Neutrino Masses in SU(2) x U(1) Theories, Phys. Rev. D 22, 2227 (1980).DOI
  12. [12]G. Lazarides, Q. Shafi, and C. Wetterich, Proton Lifetime and Fermion Masses in an SO(10) Model, Nucl. Phys. B 181, 287 (1981).DOI
  13. [13]R. N. Mohapatra and G. Senjanovic, Neutrino Masses and Mixings in Gauge Models with Spontaneous Parity Violation, Phys. Rev. D 23, 165 (1981).DOI
  14. [14]R. Foot, H. Lew, X. G. He, and G. C. Joshi, Seesaw Neutrino Masses Induced by a Triplet of Leptons, Z. Phys. C 44, 441 (1989).DOI
  15. [15]A. de Gouvêa, Neutrino Mass Models, Ann. Rev. Nucl. Part. Sci. 66, 197 (2016).DOI
  16. [16]L. Wolfenstein, Different Varieties of Massive Dirac Neutrinos, Nucl. Phys. B 186, 147 (1981).DOI
  17. [17]S. T. Petcov, On Pseudodirac Neutrinos, Neutrino Oscillations and Neutrinoless Double beta Decay, Phys. Lett. B 110, 245 (1982).
  18. [18]J. W. F. Valle and M. Singer, Lepton Number Violation With Quasi Dirac Neutrinos, Phys. Rev. D 28, 540 (1983).DOI
  19. [19]M. Doi, M. Kenmoku, T. Kotani, H. Nishiura, and E. Takasugi, PSEUDODIRAC NEUTRINO, Prog. Theor. Phys. 70, 1331 (1983).
  20. [20]A. S. Joshipura and S. D. Rindani, Phenomenology of pseudoDirac neutrinos, Phys. Lett. B 494, 114 (2000), arXiv:hep-ph/0007334.
  21. [21]M. Kobayashi and C. S. Lim, Pseudo Dirac scenario for neutrino oscillations, Phys. Rev. D 64, 013003 (2001), arXiv:hep-ph/0012266.DOI
  22. [22]A. de Gouvêa, W.-C. Huang, and J. Jenkins, Pseudo-dirac neutrinos in the new standard model, Physical Review D 80, 073007 (2009).DOI
  23. [23]A. Donini, P. Hernandez, J. Lopez-Pavon, and M. Maltoni, Minimal models with light sterile neutrinos, JHEP 07 (07), 105, arXiv:1106.0064 [hep-ph].
  24. [24]S. Asarifard and Y. Farzan, Revisiting pseudo-Dirac neutrino scenario after recent solar neutrino data, Phys. Rev. D 107, 075029 (2023), arXiv:2211.09105 [hep-ph].
  25. [25]G. Anamiati, R. M. Fonseca, and M. Hirsch, Quasi-dirac neutrino oscillations, Physical Review D 97, 095008 (2018).DOI
  26. [26]G. Amatiati, V. De Romeri, M. Hirsch, C. A. Ternes, and M. Tórtola, Quasi-Dirac neutrino oscillations at DUNE and JUNO, Phys. Rev. D 100, 035032 (2019), arXiv:1907.00980 [hep-ph].
  27. [27]N.-I. Boudjemaa, F. F. Deppisch, and S. S. Pattanaik, Probing Quasi-Dirac Neutrino Oscillations at Long Baseline Experiments (2026), arXiv:2607.06862 [hep-ph].
  28. [28]A. Abada, J. P. Pinheiro, and S. Urrea, Investigating dune oscillations sensitivity to sterile pseudo-dirac neutrinos (2025), arXiv:2506.16390 [hep-ph].arxiv.org/abs/2506.16390
  29. [29]M. Sen, Constraining pseudo-dirac neutrinos from a galactic core-collapse supernova (2023), arXiv:2205.13291 [hep-ph].arxiv.org/abs/2205.13291
  30. [30]J. Franklin, Y. F. Perez-Gonzalez, and J. Turner, JUNO as a probe of the pseudo-Dirac nature using solar neutrinos, Phys. Rev. D 108, 035010 (2023), arXiv:2304.05418 [hep-ph].arxiv.org/abs/2304.05418
  31. [31]J. F. Beacom, N. F. Bell, D. Hooper, J. G. Learned, S. Pakvasa, and T. J. Weiler, Pseudo-dirac neutrinos: A challenge for neutrino telescopes, Phys. Rev. Lett. 92, 011101 (2004).pubmed.ncbi.nlm.nih.gov/14753977
  32. [32]S. Pakvasa, A. Joshipura, and S. Mohanty, Explanation for the low flux of high energy astrophysical muon-neutrinos, Phys. Rev. Lett. 110, 171802 (2013), arXiv:1209.5630 [hep-ph].DOI
  33. [33]A. S. Joshipura, S. Mohanty, and S. Pakvasa, Pseudo-Dirac neutrinos via a mirror world and depletion of ultrahigh energy neutrinos, Phys. Rev. D 89, 033003 (2014), arXiv:1307.5712 [hep-ph].DOI
  34. [34]T. Rink and M. Sen, Constraints on pseudo-Dirac neutrinos using high-energy neutrinos from NGC 1068, Phys. Lett. B 851, 138558 (2024), arXiv:2211.16520 [hep-ph].arxiv.org/abs/2211.16520
  35. [35]K. Carloni, I. Martínez-Soler, C. A. Arguelles, K. S. Babu, and P. S. B. Dev, Probing pseudo-Dirac neutrinos with astrophysical sources at IceCube, Phys. Rev. D 109, L051702 (2024), arXiv:2212.00737 [astro-ph.HE].DOI
  36. [36]P. S. B. Dev, P. A. N. Machado, and I. Martínez-Soler, Pseudo-Dirac neutrinos and relic neutrino matter effect on the high-energy neutrino flavor composition, Phys. Lett. B 862, 139306 (2025), arXiv:2406.18507 [hep-ph].DOI
  37. [37]K. Dixit, L. S. Miranda, and S. Razzaque, Searching for pseudo-dirac neutrinos from astrophysical sources in IceCube data, Eur. Phys. J. C 85, 1481 (2025), arXiv:2406.06476 [astro-ph.HE].arxiv.org/abs/2406.06476
  38. [38]K. Carloni, Y. Porto, C. A. Arguelles, P. S. B. Dev, and S. Jana, Signatures of quasi-Dirac neutrinos in diffuse high-energy astrophysical neutrino data, Phys. Rev. D 113, 115004 (2026), arXiv:2503.19960 [hep-ph].arxiv.org/abs/2503.19960
  39. [39]L. P. S. Leal, D. R. Machado, and R. Z. Funchal, Cosmogenic neutrinos as probes of new physics, JHEP (08), 057, arXiv:2504.10576 [hep-ph].DOI
  40. [40]E. Cabrera, M. Jin, C. A. Arguelles, and A. Esmaili, Searching for sub-eV sterile neutrinos in neutrino telescopes, Phys. Rev. D 113, 075015 (2026), arXiv:2509.20142 [hep-ph].
  41. [41]C. S. Fong and Y. Porto, Constraining the pseudo-dirac nature of neutrinos using astrophysical neutrino flavor data, Phys. Rev. D 112, 063001 (2025).DOI
  42. [42]M. MacDonald, K. Carloni, C. A. Arguelles, I. Martínez-Soler, and R. Alves Batista, Exploring New Propagation Scales With Galactic Neutrinos, arXiv:2512.10744 [hep-ph] (2025), preprint.arxiv.org/abs/2512.10744
  43. [43]M. Honda, M. S. Athar, T. Kajita, K. Kasahara, and S. Midorikawa, Atmospheric neutrino flux calculation using the nrmsise-00 atmospheric model, Phys. Rev. D 92, 023004 (2015).DOI
  44. [44]P. Huber, J. Kopp, M. Lindner, M. Rolinec, and W. Winter, New features in the simulation of neutrino oscillation experiments with globes 3.0, Comput. Phys. Commun. 177, 432 (2007), arXiv:hep-ph/0701187.DOI
  45. [45]A. M. Dziewonski and D. L. Anderson, Preliminary reference earth model, Physics of the Earth and Planetary Interiors 25, 297 (1981).
  46. [46]B. Abi et al. (DUNE), Deep Underground Neutrino Experiment (DUNE), Far Detector Technical Design Report, Volume I Introduction to DUNE, JINST 15 (08), T08008, arXiv:2002.02967 [physics.ins-det].
  47. [47]S. Abdus et al. (DUNE), Reconstruction of atmospheric neutrinos in DUNE’s horizontal-drift far-detector module, Eur. Phys. J. C 86, 904 (2026), arXiv:2601.05697 [hep-ex].
  48. [48]I. Esteban, M. C. Gonzalez-Garcia, M. Maltoni, I. Martinez-Soler, J. P. Pinheiro, and T. Schwetz, Nufit-6.0: updated global analysis of three-flavor neutrino oscillations, Journal of High Energy Physics 2024, 216 (2024).
  49. [49]O. L. G. Peres and A. Y. Smirnov, Testing the solar neutrino conversion with atmospheric neutrinos, Phys. Lett. B 456, 204 (1999), arXiv:hep-ph/9902312.DOI
  50. [50]K. Abe et al. (Hyper-Kamiokande), Hyper-Kamiokande Design Report, arXiv:1805.04163 [physics.ins-det] (2018), design report.arxiv.org/abs/1805.04163

Paper details

vlnc:W_jew7awea
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 11

downloads · 1

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents