Uncovering Majorana nature through a precision measurement of CP phase

J.C. Carrasco-Martınez,1, 2 F.N. Dıaz,1 and A.M. Gago1

1Seccion Fısica, Departamento de Ciencias, Pontificia Universidad Catolica del Peru, Apartado 1761, Lima, Peru2Department of Physics, University of California, Berkeley, CA 94720, USA

We show the possibility to discover the neutrino nature by measuring the Majorana CP phase atthe DUNE experiment. This phase is turned on by a decoherence environment, possibly originatedby physics at the Planck scale. A sizable distortion in the measurement of the Dirac CP violationphase δCP is observed at DUNE when compared with T2HK measurement due to decoherence andnon-null Majorana phase. Being that, when the measurement of the Majorana phase is performedat DUNE, it reaches a precision of 23 (21) % for a decoherence parameter Γ = 4.5(5.5)× 10−24GeVand a Majorana phase equal to 1.5π. The latter precision is similar to the one obtained at the T2Kexperiment at its current Dirac CP violation phase measurement.

I. INTRODUCTION

The origin of neutrino masses is one of the most rel-evant questions of modern elementary particle physics[1, 2]. The Standard Model (SM) Higgs mechanism couldgenerate the neutrino masses if they were Dirac particles.However, this SM alternative does not explain why theneutrino masses are less than one-millionth of the elec-tron’s mass, the smallest charged lepton. The generalbelief is that the latter is resolved through the seesawmechanism, being the most inexpensive case when neu-trinos are Majorana particles [3–15]. The Majorana neu-trinos, a fermion that is its own antiparticles, imply thetotal lepton charge violation, a conserved number withinthe SM processes [16, 17]. Therefore, the quest for elu-cidating the Majorana nature of neutrinos is one of ourbest chances to learn about what is beyond the SM. Theusual way to look for Majorana neutrinos is using theneutrinoless double beta decay, a lepton charge violatingprocess not allowed by the SM [18–21]. Until now, thereis no signal from the latter [22–25] or other proposes wayof searches for Majorana neutrinos [26–29].

The standard oscillation neutrino (SO) probabilitiestake the same form regardless of the neutrino nature.Only the Dirac CP phase is observable while the twoother CP violation phases, when the neutrino is Majo-rana, are absorbed [30]. However, if we suppose a newphysics(NP) phenomenon, subleading to the SO mecha-nism, exists, the latter situation could be no longer valid.Therefore, this NP could help the Majorana phases toemerge in the oscillation probabilities. The NP phenom-ena we focus on forecasts an interaction between the neu-trino system with the environment at the Planck scalelevel, which is characterized by a foamy space-time [31].The loss of quantum coherence (decoherence) in the neu-trino system is the interaction’s main signature. Thefoamy space-time is predicted in the context of stringsand branes [32–34], and quantum gravity [35]. This quan-tum decoherence phenomena in the neutrino system havebeen vastly studied in the literature [36–41] wherein mostof the cases containing only the damping effects. Never-theless, the phenomenology of the interaction betweenneutrino and a quantum decoherence environment goes

beyond the damping effects, being that this might addnew contributions to the violation of CP or CPT [42],[43] and [44]. Through these kinds of contributions inthe neutrino oscillation formula, we can reveal the neu-trino’s Majorana nature. The observability of the Majo-rana nature using neutrino oscillation is a rather novelapproach.

At this point it is worth mentioning that our work’scornerstone hypothesis is that the Dirac CP violationphase measured at T2K experiment represents its realvalue since it is unaffected by quantum decoherence inneutrino oscillations. Thus, considering that the effectsof a quantum decoherence environment, coupled with anon zero Majorana phase, are sizable in the DUNE data,our strategy is divided into two steps. The first step isto assess how much a measurement at DUNE experiment[45] of the Dirac CP violation phase can deviate from thetrue one. For this purpose, and taking the pure SO asa theoretical hypothesis in the fitting, the CP violationphase obtained at DUNE is compared with the projectedone at the T2HK experiment [46]. The T2HK projectedmeasurement comes to be an upgrade in the precision ofthe CP violation measurement at T2K. The second step,and final goal in this letter, is to go beyond testing theDUNE accuracy for determining a Majorana CP phasesimultaneously with the decoherence parameter.

II. GENERAL THEORETICAL FORMALISM

The treatment for a neutrino subsystem embodied inan infinity unknown reservoir or environment, with whichthe former interacts weakly, can be obtained by the Lind-blad Master equation [31]:

∂ρ(t)

∂t= −i[H, ρ(t)] +D[ρ(t)], (1)

where ρ(t) is the reduced (neutrino) density matrix, ob-tained after trace over the degrees of freedom of the envi-ronment, H is the Hamiltonian of the neutrino subsystemand D[ρ(t)] is the dissipative term which encloses the de-coherence phenomena. The aforementioned factor can be

written as D[ρ(t)] = 12

∑j

([Aj , ρ(t)A†j ] + [Ajρ(t), A†j ]

).

arX

iv:2

011.

0125

4v2

[he

p-ph

] 1

8 O

ct 2

021

2

Thus, if we work in a three-level system the operatorsρ, H and Aj can be written as follows: ρ =

∑ρµtµ,

H =∑hµtµ, and Aj =

∑ajµtµ where µ is running from

0 to 8, t0 is the identity matrix and tk the Gell-Mannmatrices (k = 1, ..., 8) that satisfy [ta, tb] = i

∑c fabctc,

where fabc are the structure constants of SU(3). The

hermiticity of Aj , which is secured demanding a time-increase Von Neumman entropy, allow us to write asymmetric D ≡ Dkj dissipative/decoherence matrix as:Dkj = 1

2

∑l,m,n(anl)fknmfmlj , where anl = ~an.~al with

components Dµ0 = D0µ = 0, and ~ar = {a1r, a

2r, ..., a

8r}.

The matrix A ≡ anl should be positive [31] in orderto fulfill the complete positivity condition which statesthat the eigenvalues of the mixing matrix ρ(t) should bepositive at any time. Besides, given their inner productstructure, the Dkj must satisfy the Cauchy-Schwartz in-equalities. Adding the conservation of the probability tothe aforementioned conditions we get the following evo-lution equation for ρ(t):

ρ0 = 0, ρk = (Hkj +Dkj)ρj = Mkjρj , (2)

where Hkj =∑i hifijk. The matrix form of the solution

of Eq. (B1) is:

%(t) = eMt%(0), (3)

where % is an eight column vector compose by the ρk andM≡ Mkj . Hence, the neutrino oscillation probabilityνα → νβ is given by:

Pνα→νβ =1

3+

1

2(%β(0))T %α(t), (4)

or written in terms of the coefficients ραj (0):

Pνα→νβ =1

3+

1

2

∑i,j

ρβi (0)ραj (0)[eMt]ij , (5)

where β, α = e, µ, τ and i, j = 1, ..., 8.

A. The ραj (0) coefficients and Majorana phases

The coefficients ραj (0) and ρβj (0) encloses the elementsof the neutrino mixing matrix, as shown in AppendixA. Thus, in order to include the Majorana phases, it isenough to make the replacement:

UMajorana = UPMNS.diag(1, e−iφ1 , e−iφ2) (6)

where φ1 and φ2 are the well-known Majorana phases.Therefore, the coefficients, when the Majorana phases

are included, takes the following form:

ρα1 → ρα1 cosφ1 − ρα2 sinφ1

ρα2 → ρα2 cosφ1 + ρα1 sinφ1

ρα3 → ρα3ρα4 → ρα4 cosφ2 − ρα5 sinφ2

ρα5 → ρα5 cosφ2 + ρα4 sinφ2

ρα6 → ρα6 cos ∆φ− ρα7 sin ∆φ

ρα7 → ρα7 cos ∆φ+ ρα6 sin ∆φ

ρα8 → ρα8 ,

(7)

where ραj are the coefficients considering only the UPMNS

mixing elements. For other Majorana neutrino mix-ing matrix parametrizations, the value of the Majoranaphases in the above equations can be reinterpreted, seeAppendix A.

III. TRANSITION PROBABILITY -PERTURBATIVE APPROACH

Our selected texture of decoherence matrix D in themass vacuum basis (MVB), can be seen as composed bytwo matrices: one is a diagonal one with its all elementequals Dd = −Γ × I. The other one, Dnd, is composedby its off-diagonal part, having as non-null only the given−Γij(= −Γji) elements (the diagonal is zero). As a con-sequence from the latter, the decoherence matrix in themass matter basis (MMB) Dm, obtained after rotatingthe aforementioned one, is defined by two different ma-trices, i.e. Dm = Dd

m+Dndm . The Dd

m = −Γ×I is purelydiagonal, the Dd is unaltered by the rotation, while Dnd

m

is the rotated matrix of the non-diagonal matrix Dnd inthe MVB.

Considering, that the transition probability when neu-trinos travel through matter is:

Pνα→νβ =1

3+

1

2(%βm(0))T %αm(t), (8)

where %αm(t) = eMt%αm(0), with M = Hm + Dm. BeingHm the Hamiltonian is written in the MMB. Due to Dd

m

commutes with any matrix we have:

%αm(t) = e−Γte(Hm+Dndm )t%αm(0) = e−Γt%′

αm(t) (9)

Given that %αm(0) = %′αm(0) we can rewrite the probabilityin the following way:

Pνα→νβ =1

3+

1

2e−Γt(%′

βm(0))T %′

αm(t). (10)

The solution of %′αm(t), which is based on a power seriessolution expanded in θ13, α∆ = ∆m2

12/∆m213 and a single

Γij = Γijt, is shown in the Appendix B. It follows theprocedure the given in [39].

After assessing the magnitude of the CP-odd termsin the transition, probability caused, individually, by

3

100

0.02

0.04

0.06

0.08

0.1

100

φ1/π ∈ [0 : 2]

φ1/π = 0.5

φ1/π = 1.5

φ1/π = 0SO

P(ν

µ→

νe)

(a) (b)

Eν/GeV

FIG. 1: Oscillation probability depending on the neutrinoenergy for DUNE experiment. The figures (a) and (b) rep-resent the νµ → νe and νµ → νe appearance channels re-spectively. The off-diagonal decoherence parameter is Γ28 =−Γ/√

3. We consider Γ28 = −Γ/√

3, δCP/π = 1.4, andΓ = 2.5× 10−24GeV.

each one of the off-diagonal elements Γij (i. e. thosewho activate the Majorana phases) we conclude that

Γ28 = −Γ/√

3 gives us the most significant deviationfrom the standard oscillation formulae. We achieve thelatter by fixing the off-diagonal elements at their maxi-mum allowed values and taking the diagonal ones equalto −Γ [44]. In the Appendix C is displayed the correspon-dence between each off-diagonal elements, Γij , its abilityto turn on CP-odd or CPT-odd terms and its connectionto either φ1, φ2, or ∆φ = φ1 − φ2.

Therefore, taking the off-diagonal part of the decoher-ence matrix Dnd formed only with non-null −Γ28 ,thefollowing semi-analytical perturbative νµ → νe transi-tion probability formula for SO plus decoherence (DE) isobtained:

P SO⊕

DEνµ→νe =

(1− e−Γ

)3

+ P SOνµ→νee

−Γ − Γ28√3

sin 2θ12 sin2 θ23

× sinφ1e−Γ + Γ28θ13

sin 2θ23

2√

3(A− 1)A∆

(((1−A2) cos δ

+A2 cos (δ −∆)− cos (δ −A∆))

cosφ1 +((1−A2) sin δ

+A2 sin (δ −∆)− sin (δ −A∆))

cos 2θ12 sinφ1

)e−Γ

+ Γ28α∆sin 2θ12√

3A2∆

(2 sin2

(A∆

2

)cos 2θ23 cosφ1 − cos 2θ12

× sin2 θ23

(sinA∆−A∆

)sinφ1

)e−Γ + ...

(11)

where P SOνµ→νe is the SO probability in matter (t → L),

∆ = (m23 − m2

1)L/(2E) and A =√

2GFneL/∆, whereGF is the Fermi constant, ne is the electron number den-sity and L is the source-detector distance. The valid-ity of the transition probability formula relies on havingthe parameter-perturbative expansion Γ28, Γ28θ13 andΓ28α∆ of order: 10−2, 10−3, and 10−4, respectively. Forinstance, the latter values can be attained for |Γ28| =

100

0

0.05

0.1

Eν/GeV

P(νµ→νe)−P

(νµ→νe)

φ1/π ∈ [0 : 2]

φ1/π = 0.5

φ1/π = 1.5

φ1/π = 0

SO

FIG. 2: CP asymmetry depending on the neutrino energy.The off-diagonal decoherence parameter is Γ28 = −Γ/

√3. We

consider δCP/π = 1.4, and Γ = 2.5× 10−24GeV.

2×10−24 GeV, and at DUNE baseline L = 1300 km. Fur-thermore, to get the antineutrino transition probabilityis enough to: φ1 → −φ1, δ → −δ and A → −A, mean-while, the νµ survival probability formula is not showndue to its negligible decoherence effect.

The νµ → νe transition probability displayed in Fig. 1is numerically calculated at DUNE baseline and for themaximum value of Γ28 = −Γ/

√3, with Γ = 2.5 × 10−24

GeV and the following values for the SO parameters,taken from [49]: θ12 = 33.82◦, θ13 = 8.61◦, θ23 = 48.3◦,∆m2

21 = 7.39 × 10−5eV2, and ∆m231 = 2.523 × 10−3eV2

(normal hierarchy), that are going to be fixed along thisletter. The Dirac CP phase is taken as: δtrueCP /π = 1.4inspired in the hint given by the T2K experiment [55]From this figure it is notorious the energy-independentincrease of the SO

⊕DE probability respect to standard

one, regardless the value of φ1, a feature that has beenalready pointed out in [39, 40], for other shape of thedecoherence matrix. However, the intensity of this incre-ment depends on φ1, for example, in case of φ1/π = 1.5(φ1/π = 0.5) the SO

⊕DE neutrino (antineutrino) prob-

ability grows much less than its antineutrino (neutrino)counterpart. For φ1/π = 0 the gain is proportionally thesame for both, neutrinos or antineutrinos.

In order to quantify the CP violating effects from theextra terms containing the Majorana phase given in ourperturbatives formulae, we use the CP violation asym-metry ∆P = Pνµ→νe − Pνµ→νe :

∆P SO⊕

DE ' ∆P SOe−Γ+2Γ

3sin 2θ12 sin2 θ23 sinφ1e

−Γ+...

(12)here it is displayed only the leading term Γ28 ∼ O (0.01)

taken Γ28 = −Γ/√

3, which is its maximum allowedvalue. The predictions from Eq. (12) are illustrated inFig. 2 where the νµ → νe(νµ → νe) transition prob-ability is numerically calculated at DUNE baseline forΓ28 = −Γ/

√3, with Γ = 2.5 × 10−24 GeV. In Fig. 2 we

4

see how the overall negative (positive) sign of the deco-herence contribution for φ1/π = 1.5 (φ1/π = 0.5) dimin-ish (increases) the ∆P amplitude, whilst for φ1/π = 0.0is, as expected, nearly equal to the SO case.

IV. SIMULATION AND RESULTS

The DUNE and T2HK simulated data samples are gen-erated with GLoBES [50, 51] and nuSQuIDS [52] intro-ducing the configuration and inputs from [45, 46, 53, 54],and selecting the optimized fluxes for neutrino and an-tineutrino with 5 years of exposure time per each modefor DUNE with a 40-kt detector. While for T2HK, with258-kt detector, we consider 3 and 9 years for neutrinoand antineutrino mode, respectively. These simulatedsamples are created for non-null values of Γtrue andφtrue1 and for a value of the Dirac CP violation phaseset on the measurement performed by the T2K exper-iment: δtrueCP /π = 1.4 [55]. At this point it is impor-tant to mention that due to the small statistics and thelarge size of the uncertainties, we disregard the measure-ment of the Dirac CP violation claimed by NOvA exper-iment, which is δCP/π ∼ 0.82 [56]. In this analysis, theT2K measurement is considered as the true value of theDirac CP violation phase since it should be unaltered byany quantum decoherence effects. This is because of thesmall size of the higher decoherence contributions thatwould be Γ ∼ O(0.001), a consequence of combining thesource-detector distance of the T2K experiment with theΓ elected for this study. It should be expected, that theT2HK experiment, with the same source-detector dis-tance, would be also unaffected by the quantum deco-herence effects. Within our analysis, the T2HK DiracCP violation phase simulated measurement, which is anupgrade in the precision of the one performed at T2K,will be used as a reference point with the expectationsat DUNE. The χ2 analysis for DUNE and T2HK relieson the comparison between the SO phenomena, adoptedas theoretical hypothesis, and simulated data that in-corporates the quantum decoherence effects, where theprescription given in [40, 57] is followed. The calculationof the ∆χ2 is described by:

∆χ2 =χ2(θtest13 , δtestCP ; θtrue13 , δtrueCP ,Γtrue, φtrue1 )

− χ2min(θfit13 , δ

fitCP; θtrue13 , δtrueCP ,Γtrue, φtrue1 )

(13)

where θfit13 and δfitCP are the best-fit points which mini-mizes the χ2, considering priors at 3σ for the rest of theoscillation parameters but δCP. The DUNE and T2HK∆χ2 contours, projected into sin2 θ13 vs δCP planes andobtained after marginalizing over the rest of SO param-eters, are presented in Fig. 3. As expected, for T2HK,

the sin2 θfit13 and δfitCP are similar to the true ones be-ing unmodified by the parameters chosen for decoher-ence. Meanwhile, for DUNE there is a slight increase

of sin2 θfit13 , respect to the sin2 θtrue13 (= 0.0224), explicitlyshown in Table. I. This increment is the consequence of

1.0 1.5

0.015

0.020

0.025

0.030

0.035φ1/π = 0.5

1σ

3σ

5σ

1.0 1.5

φ1/π = 0.5

1.0 1.5

0.015

0.020

0.025

0.030

0.035φ1/π = 1.0

1.0 1.5

φ1/π = 1.0

1.0 1.5

0.015

0.020

0.025

0.030

0.035φ1/π = 1.5

1.0 1.5

φ1/π = 1.5

sin

2θ 1

3

δCP/π

FIG. 3: ∆χ2 contours (2 dof) considering the effects ofdecoherence with Majorana phases on the standard oscilla-tion fits. The solid and dashed lines are decoherence withΓ28 = −Γ/

√3 for the DUNE and T2HK experiments, respec-

tively. The left column is Γ = 2.5× 10−24GeV and the rightcolumn is Γ = 3.5× 10−24GeV. We consider δtrueCP /π = 1.4.

trying to adjust the theoretical hypothesis, SO, with theenergy-independent increase of the SO

⊕DE probability

amplitude embodied in the data, and modulated by theintensity of Γ (see the third term of Eq. 11).

Γ = 2.5× 10−24 GeV φ1/π = 0.5 φ1/π = 1.0 φ1/π = 1.5

sin2 θfit13 0.0241 0.0242 0.0247

Nσ 0.31σ 0.34σ 0.55σ

δfitCP/π 1.43 1.33 1.13

Nσ 0.08σ 1.19σ 4.34σ

Γ = 3.5× 10−24 GeV φ1/π = 0.5 φ1/π = 1.0 φ1/π = 1.5

sin2 θfit13 0.0247 0.0250 0.0256

Nσ 0.54σ 0.61σ 0.87σ

δfitCP/π 1.44 1.28 1.06

Nσ 0.14σ 2.37σ 5.47σ

TABLE I: Fitted values for sin2 θ13, δCP and their respectiveshifts in terms of σ units. We consider δtrueCP /π = 1.4

5

The δfitCP for DUNE, when φ1/π = 1.5, is moving awayfrom δtrueCP /π(= 1.4) towards ∼ π, minimizing the magni-tude of the CP violation asymmetry. For φ1/π = 0.5 the

δfitCP takes almost exactly the value of the true one goingin the direction to maximize the CP violation asymme-try. Both features, expressed numerically in Table. I,can be explained from the need to accommodate the re-duction (increase) of ∆P , when φ1/π = 1.5(0.5), seenin Fig. 2. The quantified dislocation, in terms of σ,

from sin2 θfit13 and δfitCP (for DUNE) to the correspond-ing true ones (for T2HK), for Γ = {2.5, 3.5}× 10−24GeVand φ1/π = {0.0, 0.5, 1.5}, is depicted in Table I. ForΓ = 3.5(2.5) × 10−24GeV the most prominent shift isfound for φ1/π = 1.5 with 0.87(0.55)σ and 5.47(4.34)σ

for sin2 θfit13 and δfitCP, respectively. While, for Γ =

3.5 (2.5) × 10−24 GeV, the dislocation of δfitCP reaches3σ (2σ) and 5σ (3σ) when φ1/π takes values below

1.01 (1.03) and 1.30 (1.10), respectively, the sin2 θfit13

is clearly stable in front of changes along the φ1 inter-val. The less significant distortion is for φ1/π = 0.5 with

0.54(0.31)σ and 0.14(0.08)σ for sin2 θfit13 and δfitCP, respec-tively. A way to discriminate between different values ofφ1 is through the ratio (R) of the number of σ devia-

tion for sin2 θfit13 to the corresponding ones for δfitCP. Infact, a sort of discernment is achieved, for instance, forΓ = 3.5(2.5) × 10−24 GeV, R ∼ 0.16(0.13) − 0.26(0.29)for the interval φ1/π = 1.0, 1.5 reaching values up to∼ 3.86(3.88) for φ1/π = 0.5. A plus that reinforce theutility of R it is its low variations against changes of Γ.

The aforementioned analysis had the purpose of

searching for distortions in sin2 θfit13 and δfitCP, consideringpure SO as theoretical hypothesis. Now, the aim is togo one step further and to explore the capacity of DUNEfor measuring the Majorana phase, and also Γ, under the(SO) plus decoherence (DE) as theoretical hypothesis,for φ1/π = 0.5, 1, 1.5 and for Γ = 4.5, 5.5 × 10−24 GeV.The latter values cannot be compared with the limitsimposed by Ice Cube [58] since we are considering a non-diagonal scenario for the decoherence matrix. In Fig. 4the different allowed regions are displayed considering68 % and 90 % C.L for 2 dof. Here it is seen thatDUNE is able to measure φ1/π = 1.50 ± 0.35(0.32) andφ1/π = 0.50 ± 0.35(0.32) and Γ = 4.50 ± 1.38(5.50 ±1.42)× 10−24 GeV. While for φ1/π = 1.0± 0.19(0.15) aΓ = 4.50± 1.42(5.50± 1.46)× 10−24 GeV is obtained. Itis interesting to remark that either for φ1/π = 0.5 or 1.5the value of π is excluded at 90% C.L., independent thevalue of Γ. Thus, we are able to separate between thesedifferent values of φ1.

V. SUMMARY AND CONCLUSIONS

We demonstrated the possibility of uncovering the Ma-jorana nature of neutrinos in the DUNE experiment if adecoherence environment, with some determined char-acteristics, interacts with the neutrino system. Planck

0.0 0.5 1.0 1.5

2

3

4

5

6

7

8Γtrue/10−24GeV = 4.5φtrue1 /π = [0.5, 1.0, 1.5]

(a)

0.0 0.5 1.0 1.5

Γtrue/10−24GeV = 5.5 (b)

68%C.L.

90%C.L.

Best Fit

φ1/π

Γ/10−

24G

eV

FIG. 4: DUNE’s ability to constrain (2 dof) the decoherenceparameter and the Majorana phase. The red, green and bluelines represent φtrue/π = 0.5, 1.0, 1.5, respectively.

scale physics could be causing this decoherence environ-ment. Our approach is at first to show the strong dis-placement that it would be exhibited by the measuredvalue of δCP at DUNE, in comparison with the one mea-sured at T2HK, which would be unaffected by the de-coherence effects. This displacement can be as large as5.47 σ for a Majorana phase φ1/π = 1.5 and a deco-herence parameter Γ = 3.5 × 10−24 GeV. At next, weassessed the power of DUNE experiment in constrainingthe Majorana phase achieving, for instance, a precision of23% (21%) for φ1/π = 1.5 with Γ = 4.5(5.5)×10−24GeV.These values of precision are compatible with the currentresults on the Dirac CP phase reached by the T2K exper-iment [55]. Finally, we can conclude that, if decoherenceexists in the manner is predicted here, there would bean interesting chance for DUNE to perform a first timea measurement of the Majorana CP phase, with somereasonable uncertainties.

VI. ACKNOWLEDGEMENTS

A. M. Gago acknowledges funding by the Direccionde Gestion de la Investigacion at PUCP, through grantsDGI-2017-3-0019 and DGI 2019-3-0044. F. N. Dıaz ac-knowledges CONCYTEC for the graduate fellowship un-der Grant No. 236-2015-FONDECYT. The authors alsowant to thank F. de Zela, J. Jones-Perez, J. L. Bazo, C.A. Arguelles and Mario A. Acero for useful suggestionsand reading the manuscript.

Appendix A: Analysis of the ραj for differentparametrizations

The starting point of our analysis are the coefficientsραj , where j = 0, 1, ..., 8 and α = e, µ, τ , which are ex-

6

pressed in terms of the mixing matrix [47]:

ρα0 =√

2/3

ρα1 = 2Re(U∗α1Uα2)

ρα2 = −2Im(U∗α1Uα2)

ρα3 = |Uα1|2 − |Uα2|2ρα4 = 2Re(U∗α1Uα3)

ρα5 = −2Im(U∗α1Uα3)

ρα6 = 2Re(U∗α2Uα3)

ρα7 = −2Im(U∗α2Uα3)

ρα3 =1√3

(|Uα1|2 + |Uα2|2 − 2|Uα3|2)

(A1)

where the Uαj are the matrix elements of the PMNS ma-trix (UPMNS) [48] without taking into account the Majo-rana phases.

1. Symmetrical Parametrization of the mixingmatrix

The elements of symmetric parametrization of the mix-ing matrix, given in Eq. (5) in [59], assuming the relationδ = φ13 − φ12 − φ23, can be written as follows:

Ue1 → Ue1

Ue2 → Ue2e−iφ12

Ue3 → Ue3e−i(φ23+φ12)

Uµ1 → Uµ1eiφ12

Uµ2 → Uµ2

Uµ3 → Uµ3e−iφ23

Uτ1 → Uτ1ei(φ23+φ12)

Uτ2 → Uτ2eiφ23

Uτ3 → Uτ3

(A2)

where φ13, φ12 and φ23 are the CP phases used in [59].

The corresponding ραj are described by the followingrelations:

ρα1 → ρα1 cosφ12 − ρα2 sinφ12

ρα2 = ρα2 cosφ12 + ρα1 sinφ12

ρα3 → ρα3ρα4 → ρα4 cos (φ12 + φ23)− ρα5 sin (φ12 + φ23)

ρα5 → ρα5 cos (φ12 + φ23) + ρα4 sin (φ12 + φ23)

ρα6 → ρα6 cosφ23 − ρα7 sinφ23

ρα7 → ρα7 cosφ23 + ρα7 sinφ23

ρα8 → ρα8 ,

(A3)

where ραj are given in Eq. (A1)

2. Particle data group parametrization type I:PDG I

Here we analyze the mixing matrix parametrizationgiven in [60], which includes the Majorana phases φ1 andφ2:

UMajorana = UPMNS.diag(exp iφ1, exp iφ2, 1) (A4)

The corresponding ραj are described by the followingequations:

ρα1 → ρα1 cos ∆φ+ ρα2 sin ∆φ

ρα2 → ρα2 cos ∆φ− ρα1 sin ∆φ

ρα3 → ρα3ρα4 → ρα4 cosφ1 − ρα5 sinφ1

ρα5 → ρα5 cosφ1 + ρα4 sinφ1

ρα6 → ρα6 cosφ2 − ρα7 sinφ2

ρα7 → ρα7 cosφ2 + ρα6 sinφ2

ρα8 → ρα8 ,

(A5)

Below, we present in table II a summary of the equiv-alences between the different parameterizations.

Sym. ↔ PDG I Sym. ↔ Our Work PDG I ↔ Our Work

φ12 + φ23 ↔ φ1 φ12 ↔ φ1 φ1 ↔ ∆φ

φ23 ↔ φ2 φ12 + φ23 ↔ φ2 φ2 ↔ φ1

φ12 ↔ ∆φ φ23 ↔ −∆φ ∆φ↔ −φ2

TABLE II: Parameterizations comparison.

where ∆φ = φ1 − φ2.

Appendix B: Probability Calculation

For solving %′αm(t) we must start with the next differ-ential equation:

˙%′αm = (Hm + Dndm )%′

αm, (B1)

which is similar to Eq.(2) presented in our letter. Beforecontinue, we must point out that the following procedureis similar to the one given in [39]. The Eq.(B1) can besimplified using this change of variable:

%αm(t) = eHmt%α(t), (B2)

then the Eq. B1 :

eHmt ˙%α + HmeHmt%α = (Hm + Dnd

m )eHmt%α (B3)

thus we get:

˙%α = e−HmtDndm e−Hmt%α, (B4)

7

the matrix e−HmtDndm e−Hmt can be expanded perturba-

tively in power series of the small parameters θ13, andα∆ which turns out to be:

e−HmtDndm e−Hmt = Γij(D

(0) +θ13D(θ13) +α∆D

(α∆) + ...)(B5)

we can factor out the decoherence parameter Γij since itis a common factor of all the elements in the decoherencematrix Dnd

m in the MMB (a consequence of its definitionin the mass vacuum basis that is an off-diagonal matrixwith only non-null terms in a given −Γij element). Re-placing Eq.(B5) into Eq.(B4):

˙%α = Γij(D(0) + θ13D

(θ13) + α∆D(α∆) + ...)%α (B6)

the above equation can be solved perturbatively treating%α as a power series in θ13, α∆ and Γij :

%α = %(0) + θ13%(θ) + α∆%

(α∆) + α∆θ13%(α∆θ13) + ...

+ Γij %(Γij) + Γijθ13%

(Γijθ13) + Γijα∆%(Γijα∆) + ...

(B7)

Then substituing Eq.(B7) into Eq.(B6) we produce asequence of first order differential equations each ofthem collecting equal power terms. The Γij-independent

terms of the %α expansion: %(0) + θ13%(θ) + α∆%

(α∆) +α∆θ13%

(α∆θ13) + ... corresponds to the initial condition%α(0), which is constant in time and coincides with theinitial condition for the standard oscillation case, since atthat instant the environment is decoupled (not interact-ing) with the neutrino system. Considering all the laterplus the condition that %′αm(0) = %α(0), we can rewrite

Eq.(B2) as follows:

%′αm(t) = eHmt(%′

αm(0) + Γij(...))., (B8)

with Γij = Γijt. The second term at the right-hand sideof the equation above contains the explicit solution of thepower series of %α.

Appendix C: CP-odd, CPT-odd terms andMajorana phases

CPV CPTV Non-null Majorana phase

Γ13,Γ23,Γ18,Γ28,Γ12 Γ23,Γ28,Γ12 φ1

Γ34,Γ35,Γ48,Γ58,Γ45 Γ35,Γ58,Γ45 φ2

Γ37,Γ36,Γ68,Γ67,Γ78 Γ37,Γ67,Γ78 ∆φ

Γ14,Γ24,Γ15,Γ25 Γ24,Γ15 φ1, φ2

Γ16,Γ17,Γ26,Γ27 Γ17,Γ26 φ1,∆φ

Γ46,Γ47,Γ56,Γ57 Γ47,Γ56 φ2,∆φ

TABLE III: Violation of symmetries by non-diagonal decoher-ence elements and their dependence on the Majorana phases.

In table III we present a classification of the corre-spondence between each one of the off-diagonal elementsΓij and φ1, φ2, or ∆φ, also pointing out its connectionwith CP-odd, CPT-odd terms or both in the oscillationprobabilities which incorporates quantum decoherence.

[1] H. Murayama, Prog. Part. Nucl. Phys. 57, 3-21 (2006)[2] J. W. F. Valle, Adv. Ser. Direct. High Energy Phys. 25,

25-37 (2015)[3] T. Yanagida, Conf. Proc. C 7902131, 95-99 (1979)[4] M. Gell-Mann, P. Ramond and R. Slansky, Conf. Proc.

C 790927, 315-321 (1979)[5] S. L. Glashow, NATO Sci. Ser. B 61, 687 (1980).[6] P. Minkowski, Phys. Lett. B 67, 421-428 (1977)[7] R. N. Mohapatra and G. Senjanovic, Phys. Rev. Lett.

44, 912 (1980)[8] M. Magg and C. Wetterich, Phys. Lett. B 94, 61-64

(1980)[9] J. Schechter and J. W. F. Valle, Phys. Rev. D 22, 2227

(1980)[10] R. N. Mohapatra and G. Senjanovic, Phys. Rev. D 23,

165 (1981)[11] R. Foot, H. Lew, X. G. He and G. C. Joshi, Z. Phys. C

44, 441 (1989)[12] R. E. Shrock, Phys. Rev. D 24, 1232 (1981)[13] W. Konetschny and W. Kummer, Phys. Lett. B 70, 433-

435 (1977)[14] T. P. Cheng and L. F. Li, Phys. Rev. D 22, 2860 (1980)[15] G. Lazarides, Q. Shafi and C. Wetterich, Nucl. Phys. B

181, 287-300 (1981)

[16] A. Atre, T. Han, S. Pascoli and B. Zhang, JHEP 05, 030(2009)

[17] A. Zee, Nucl. Phys. B 264, 99-110 (1986)[18] S. M. Bilenky and C. Giunti, Mod. Phys. Lett. A 27,

1230015 (2012)[19] A. Faessler and F. Simkovic, J. Phys. G 24, 2139-2178

(1998)[20] Avignone F.T., Elliott S.R. and Engel J., Rev. Mod.

Phys. 80, 481-516 (2008)[21] J. D. Vergados, H. Ejiri and F. Simkovic, Rept. Prog.

Phys. 75, 106301 (2012)[22] M. Agostini et al. [GERDA], Science 365, 1445 (2019)[23] G. Anton et al. [EXO-200], Phys. Rev. Lett. 123, no.16,

161802 (2019)[24] D. Q. Adams et al. [CUORE], Phys. Rev. Lett. 124,

no.12, 122501 (2020)[25] Gando A., Gando Y., Hachiya T., Hayashi A.,

Hayashida S., Ikeda H. et al. [KamLAND-Zen], Phys.Rev. Lett. 117, no.8, 082503 (2016)

[26] A. M. Sirunyan et al. [CMS], JHEP 01, 122 (2019)[27] R. Aaij et al. [LHCb], Phys. Rev. Lett. 112, no.13,

131802 (2014)[28] A. B. Balantekin, A. de Gouvea and B. Kayser, Phys.

Lett. B 789, 488-495 (2019)

8

[29] K. Bora, D. Borah and D. Dutta, Phys. Rev. D 96, no.7,075006 (2017)

[30] C. Giunti, Phys. Lett. B 686, 41-43 (2010)[31] F. Benatti and R. Floreanini, JHEP 0002, 032 (2000);

Phys. Rev. D 64, 085015 (2001)[32] J. R. Ellis, N. E. Mavromatos and D. V. Nanopoulos,

Phys. Lett. B 293, 37 (1992)[33] J. R. Ellis, N. E. Mavromatos and D. V. Nanopoulos, Int.

J. Mod. Phys. A 11, 1489 (1996)[34] F. Benatti and R. Floreanini, Annals Phys. 273, 58-71

(1999)[35] S. Hawking, Commun. Math. Phys. 87, 395-415 (1982);

Phys. Rev. D 37, 904-910 (1988); Phys. Rev. D 53, 3099-3107 (1996); Hawking S.W. and Hunter C.J., Phys. Rev.D 59, 044025 (1999)

[36] E. Lisi, A. Marrone and D. Montanino, Phys. Rev. Lett.85, 1166 (2000)

[37] G. Barenboim, N. E. Mavromatos, S. Sarkar andA. Waldron-Lauda, Nucl. Phys. B 758, 90 (2006)

[38] P. Bakhti, Y. Farzan and T. Schwetz, JHEP 1505, 007(2015)

[39] J. A. Carpio, E. Massoni and A. M. Gago, Phys. Rev. D97, no. 11, 115017 (2018)

[40] J. A. Carpio, E. Massoni and A. M. Gago, Phys. Rev. D100, no. 1, 015035 (2019)

[41] A. L. G. Gomes, R. A. Gomes and O. L. G. Peres,arXiv:2001.09250 [hep-ph].

[42] R. L. N. Oliveira, M. M. Guzzo and P. C. de Holanda,Phys. Rev. D 89, no. 5, 053002 (2014)

[43] A. Capolupo, S. M. Giampaolo and G. Lambiase, Phys.Lett. B 792, 298 (2019)

[44] J. C. Carrasco, F. N. Dıaz and A. M. Gago, Phys. Rev.

D 99, no. 7, 075022 (2019)[45] R. Acciarri et al. [DUNE Collaboration],

arXiv:1512.06148 [physics.ins-det].[46] K. Abe et al. [Hyper-Kamiokande], arXiv:1805.04163

[physics.ins-det].[47] A. M. Gago, E. M. Santos, W. J. C. Teves and

R. Zukanovich Funchal, arXiv:hep-ph/0208166 [hep-ph].[48] Z. Maki, M. Nakagawa and S. Sakata, Prog. Theor. Phys.

28 (1962), 870-880[49] http://www.nu-fit.org[50] P. Huber, M. Lindner and W. Winter, Comput. Phys.

Commun. 167, 195 (2005)[51] P. Huber, J. Kopp, M. Lindner, M. Rolinec and W. Win-

ter, Comput. Phys. Commun. 177, 432 (2007)[52] C. A. Arguelles Delgado, J. Salvado and C. N. Weaver,

Comput. Phys. Commun. 196, 569 (2015)[53] T. Alion et al. [DUNE Collaboration], arXiv:1606.09550

[physics.ins-det].[54] P. Huber, M. Mezzetto and T. Schwetz, JHEP 03, 021

(2008)[55] K. Abe et al. [T2K], Nature 580, no.7803, 339-344 (2020)[56] A. Himmel, “New Oscillation Results from the NOvA

Experiment – presented at Neutrino2020,” (2020).[57] F. N. Dıaz, J. Hoefken and A. M. Gago, Phys. Rev. D

102 (2020) no.5, 055020[58] P. Coloma, J. Lopez-Pavon, I. Martinez-Soler and

H. Nunokawa, Eur. Phys. J. C 78 (2018) no.8, 614[arXiv:1803.04438 [hep-ph]].

[59] W. Rodejohann and J. W. F. Valle, Phys. Rev. D 84(2011), 073011

[60] M. Tanabashi et al. [Particle Data Group], Phys. Rev. D98, no.3, 030001 (2018)