IFIC/08-19

Tri-bimaximal neutrino mixing and neutrinoless double beta decay

M. Hirsch,1, ∗ S. Morisi,1, † and J. W. F. Valle1, ‡

1
AHEP Group, Institut de Fı́sica Corpuscular – C.S.I.C./Universitat de València
Edificio Institutos de Paterna, Apt 22085, E–46071 Valencia, Spain
(Dated: November 22, 2018)

We present a tri-bimaximal lepton mixing scheme where the neutrinoless double beta decay rate
has a lower bound which correlates with the ratio α ≡ ∆m2sol /∆m2atm well determined by current
data, as well as with the unknown Majorana CP phase φ12 characterizing the solar neutrino sub-
system. For the special value φ12 = π2 (opposite CP-sign neutrinos) the ββ0ν rate vanishes at tree
level when ∆m2sol /∆m2atm = 3/80, only allowed at 3σ. For all other cases the rate is nonzero, and
lies within current and projected experimental sensitivities close to φ12 = 0. We suggest two model
arXiv:0804.1521v1 [hep-ph] 9 Apr 2008

realizations of this scheme in terms of an A4 × Z2 and A4 × Z4 flavour symmetries.

PACS numbers: 11.30.Hv 14.60.-z 14.60.Pq 14.80.Cp

Current neutrino oscillation data [1, 2, 3, 4, 5, 6, 7] in- and the neutrino mass matrix has the form
dicate a peculiar pattern [8] of neutrino masses and mix-
 
A 0 0
ings quite at variance with the structure of the Cabibbo- Mν ∼  0 B C  ,
 
Kobayashi-Maskawa quark mixing matrix [9]. However 0 C B
they do not yet fully determine the absolute scale of neu-
trino masses nor shed any light on the issue of leptonic the resulting lepton mixing matrix has exactly the tri-
CP violation, two demanding challenges left for future bimaximal structure given in Eq. (1).
experiments. Here we consider schemes where neutrinos get mass a
la seesaw, defined by the following mass matrices,
Lacking a basic theory for the origin of mass one needs  
theoretical models restricting the pattern of fermion α β γ
masses and mixings and providing guidance for future M l ∼  γ α β  = Uω Mdiag l
Uω† ;
 
experimental searches. An attractive phenomenological β γ α
ansatz for leptons is the Harrison-Perkins-Scott (HPS)
   
mixing [10] a 0 0 1 0 0
mD ∼  0 a b  ; MR ∼  0 1 0 
   
p √ 
2/3 1/ 3 0 0 b a 0 0 1
√ √ √ 
UHPS = −1/ 6 1/ 3 −1/ 2 (1)

√ √ √ This “texture” constitutes a new ansatz for the lepton
−1/ 6 1/ 3 1/ 2
sector that can be realized (see below) in the frame-
which predicts the following values for the lepton mixing work of A4 -based flavour symmetry models. The as-
angles: tan2 θatm = 1, sin2 θChooz = 0 and tan2 θsol = 0.5, sumed symmetry of the Dirac mass term holds in SO(10)
providing a good first approximation to the values [8] models where it comes from a 16 16 10 Yukawa coupling.
indicated by neutrino oscillation experiments [1, 2, 3, 4, In contrast with other existing tri-bimaximal A4 based
5]. schemes, the gauge singlet seesaw mass term character-
izing the heavy right-handed neutrinos is also a flavour
As noted earlier [11], when the charged lepton mass
singlet, instead of the neutrino Dirac mass term. This
matrix Ml obeys
makes the scheme extremely predictice, as it involves as
free parameters only the two modulii and the relative
M l M l† = Uω Mdiag
l2
Uω† ;
phase between a and b.
After the seesaw mechanism, one obtains the effective
where Uω is the “magic” unitary matrix
light neutrino mass matrix Mν given as
   
1 1 1 a2 0 0
1  1
Uω = √  1 ω ω 2  , M ν = mD mT ∼  0 a2 + b2 2ab  . (2)
  
3 MR D
1 ω2 ω 0 2ab a2 + b2
 2

Rewriting the effective light neutrino mass matrix in the 10 -2

basis where charged leptons are diagonal one finds

hmν i ≡ |mee| [eV]

2b2
 
a2 + 4ab3 + 3 − 31 b(2a + b) − 13 b(2a + b) 10 -3
Mν ≡  − 31 b(2a + b) 1 2b2  .
3 b(4a − b) a2 − 2ab3 + 3 

2
− 31 b(2a + b) a2 − 2ab 3 + 3
2b 1
3 b(4a − b)
10 -4
This matrix is fully determined by two complex parame-
ters a and b, which imply three physical real parameters,
namely two moduli and a relative phase, which is the 10 -5

only source of leptonic CP violation in the scheme.

We note that Mν is µ ↔ τ invariant, so it gives θ13 = 0
10 -6
and sin2 θ23 = 1/2 as predictions. The state (1, 1, 1)t is
0.027 0.031 0.034 0.037 0.04
an eigenstate of Mν with eigenvalue a2 , so the neutrino α
mass matrix Mν is diagonalized by the tri-bimaximal
mixing matrix, leading then to tan2 θ12 = 0.5. The three FIG. 1: Lower bound on the ββ0ν amplitude parameter mee
neutrino mass eigenvalues are as function of α ≡ ∆m2sol /∆m2atm for different values of the
Majorana phase φ12 = −π/2 + t where t = 0 (dark brown),
{m1 , m2 , m3 } = {(a + b)2 , a2 , −(a − b)2 }. 0.001 (brown), 0.004 (red), 0.011 (dark orange), 0.029 (or-
ange), 0.089 (yellow). The 1, 2 and 3σ ranges for α are also
Data from neutrino oscillation experiments [1, 2, 3, 4, 5] shown.
determine pretty well two of the three parameters on
the left-hand side [8], namely the solar and atmospheric
hmν i ≡ |mee| [eV]
mass-square splittings. The remaining observable is pre- 10 0
cisely the neutrino-exchange amplitude for neutrinoles
double beta decay, given by 10 -1

4ab 2b2
hmν i ≡ |mee | = |a2 + + |.
3 3 10 -2

This parameter can be given as a function of the three

10 -3
independent model parameters, which we choose to ex-
press in terms of the observables ∆m2atm , α and the rela-
10 -4
tive phase between a and b. The latter is directly related
to the Majorana CP phase [12, 13, 14, 15] characteriz- -Π 3Π Π
- €€€€€€€€€ - €€€€
Π
€ - €€€€€ 0 Π
€€€€€
Π
€€€€€ 3Π
€€€€€€€€€
Π
4 2 4 4 2 4
ing the solar neutrino sub-system, φ12 , in a symmetric
φ12
parametrization of the lepton mixing matrix where all
phases appear attached to the corresponding mixing an- FIG. 2: Lower bound on the neutrinoless double beta decay
gle [12, 13]. amplitude parameter mee as function of the Majorana CP
First we note that our scheme is compatible with neg- phase φ12 for α within the 1σ (yellow) and 2σ (blue) ranges.
ligible neutrinoless double beta decay, vanishing at the
tree level i. e. mee = 0. This happens only when CP is
ββ0ν amplitudes, which correspond to almost full de-
conserved with opposite CP parities [16, 17] between ν1
structive interference between opposite CP sign neutrinos
and ν2 and for
ν1 and ν2 .
∆m2sol 3
α= = = 0.0375, (3)
∆m2atm 80
as seen in Fig. 1, which is currently allowed at 3 σ. For
all other values of the CP phase the model gives a lower Notice that in the central region around the other CP
bound on the neutrinoless double beta decay which we conserving point φ12 = 0 the ββ0ν amplitude is sizeable,
display in Fig. 2, which we call the “Niemeyer” plot [32]. and depends very sensitively on the Majorana phase φ12 ,
This plot exhibits two dips characterized by very small as displayed in Fig. 3.
 3

hmν i ≡ |mee| [eV]

10 1 fields Li lic νic h Hi ϕ Φ ξ
SU (2)L 2 1 1 2 2 2 2 1
A4 3 3 3 1 3 1 3 1
Z2 + + − + + − − +

10 0 TABLE I: Lepton multiplet structure of model I

λijk with components β and γ, and similarly for the next

10 -1 two terms. Note that in this case there are additional
SU(2) ⊗ U(1) doublet Higgs scalar bosons Hi , Φi , ϕ
transforming non-trivially under the flavour symmetry.
10 -2 10 -1 We assume that these develop non-zero vacuum expecta-
|φ12 | tion values (vevs), with the structure

FIG. 3: Zoom of the region giving the maximum value for the hHi i ∼ (1, 1, 1); hΦi i ∼ (0, 0, 1)
lower bound on mee in Fig. 2.
Similar vev alignment condition has been used in
Ref. [27]. Note that the two zeros in mD follow from
the alignment condition hΦ1 i = hΦ2 i = 0.
It is a non-trivial task to produce a consistent flavour In contrast the second model contains only one SU(2)
symmetry leading to a structure of the effective neutrino ⊗ U(1) doublet Higgs boson and its A4 × Z4 symmetric
mass matrix Mν that has, at least as a first approxima- leading-order Lagrangian is written as
tion, the desired predictive pattern.
L = λ0 (Llc )hξ1 + λ(Llc φ)h
Here we suggest two possible realizations based on an
+λ′0 (Lν c φ′ )h̃ + λ′ (Lν c )h ξ2 + λR (ν c ν c )ξ3
A4 flavour symmetry for the neutrino mass matrix. The
discrete group A4 is a relatively small and simple flavour where λR is dimensionless while the others scale as in-
group consisting of the 12 even permutations among four verse mass. Note the appearance of gauge singlet scalars
objects. It has a three-dimensional irreducible represen- φ, φ′ and ξi , transforming non-trivially under the flavour
tation appropriate to describe the three generations ob- symmetry and coupling non-renormalizably to the lepton
served. Originally, A4 was proposed [18, 19] for under- doublets. Their only renormalizable is the one giving rise
standing degenerate neutrino spectrum with nearly max- to the large Majorana mass term. We assume that these
imal atmospheric neutrino mixing angle. More recently, “flavon” fields develop non-zero vacuum expectation val-
predictions for the solar neutrino mixing angle have ues (vevs), with the structure
also been incorporated within the so-called tri-bimaximal
neutrino mixing schemes [20, 21, 22, 23, 24, 25, 26]. hφi ∼ (1, 1, 1); hφ′ i ∼ (0, 0, 1)
In our phenomenological A4 -based flavor symmetry
Note that either way we obtain the desired predic-
schemes the neutrino mass comes from type-I seesaw
tive charged lepton and neutrino mass matrices discussed
mechanism with right-handed Majorana mass matrix
above.
proportional to the identity matrix. In both models lep-
tons transform as A4 -triplets, while the standard Higgs fields Li lic νic h φ φ′ ξ1 ξ2 ξ3
is a flavour singlet [33]. The lepton and scalar content of SU (2)L 2 1 1 2 1 1 1 1 1
the models are specified in Tables I and II. A4 3 3 3 1 3 3 1 1 1
The A4 × Z2 invariant Lagrangian characterizing the Z4 1 ω3 ω 1 ω ω3 ω ω3 ω2
first model is renormalizable, and given by
TABLE II: Lepton multiplet structure of model II
c c
L = λ0 (Ll )h + λ(Ll H)
+λ′0 (Lν c )ϕ + λ′ (Lν c Φ) + λR (ν c ν c )ξ.
In summary, here we have proposed two A4 -based
where the first term involves an A4 -invariant coupling λ0 flavour symmetries leading to tri-bimaximal lepton mix-
that provides α in M l , while the second involves a tensor ing, namely tan2 θatm = 1, sin2 θChooz = 0 and tan2 θsol =
 4

0.5. Although this implies a boring scenario for upcom- [3] KamLAND collaboration, T. Araki et al., Phys. Rev.
ing long baseline oscillation experiments [28, 29] aiming Lett. 94, 081801 (2005).
[4] T. Kajita, New J. Phys. 6, 194 (2004).
to probe θ13 and leptonic CP violation in oscillations, [5] K2K collaboration, M. H. Ahn et al., Phys. Rev. Lett.
we have analysed its implications for neutrinoless dou- 90, 041801 (2003), [hep-ex/0212007].
ble beta decay. We have seen how the ββ0ν amplitude [6] MINOS collaboration, arXiv:0708.1495 [hep-ex].
[7] I. Shimizu, Talk at the 10th International Conference on
parameter mee has a lower bound which correlates with
Topics in Astroparticle and Underground Physics, TAUP
the ratio α ≡ ∆m2sol /∆m2atm well determined by current 2007.
neutrino oscillation data, as well as with the Majorana [8] M. Maltoni, T. Schwetz, M. A. Tortola and J. W. F.
CP violating phase φ12 . Accelerator neutrino oscillation Valle, New J. Phys. 6, 122 (2004), arXiv ver-
sion [hep-ph/0405172] provides updated results; previous
experiments like MINOS, T2K and NOvA are expected works by other groups as well as the relevant experimen-
to improve the determination of α in the not-too-distant tal references are given therein.
future. [9] M. Kobayashi and T. Maskawa, Prog. Theor. Phys. 49,
652 (1973).
For the special value φ12 = π2 (opposite CP-sign neu- [10] P. F. Harrison, D. H. Perkins and W. G. Scott, Phys.
trinos) one finds that ββ0ν vanishes at tree level when Lett. B530, 167 (2002), [hep-ph/0202074].
∆m2sol /∆m2atm = 3/80. However this is only allowed at [11] E. Ma, Phys. Rev. D70, 031901 (2004),
[hep-ph/0404199].
3σ, as seen from Fig. 1, at 1 σ we currently have a lower [12] J. Schechter and J. W. F. Valle, Phys. Rev. D22, 2227
bound |mee | > −4
∼ few ×10 eV. For all other cases one (1980).
has a nonzero ββ0ν decay rate, with CP conservation [13] J. Schechter and J. W. F. Valle, Phys. Rev. D23, 1666
(1981).
with same CP-sign neutrinos already excluded. We have
[14] S. M. Bilenky, J. Hosek and S. T. Petcov, Phys. Lett.
also presented in Fig. 3 the lower bound in the region B94, 49 (1980).
close to φ12 = 0, corresponding to the case of same CP- [15] M. Doi, T. Kotani, H. Nishiura, K. Okuda and E. Taka-
parity neutrinos, where neutrinoless double beta decay sugi, Phys. Lett. B102, 323 (1981).
[16] J. Schechter and J. W. F. Valle, Phys. Rev. D24, 1883
could soon be observed. (1981), Err. D25, 283 (1982).
All our considerations refer to an effective low-energy [17] L. Wolfenstein, Phys. Lett. B107, 77 (1981).
model which assumes the vev alignment conditions, and [18] E. Ma and G. Rajasekaran, Phys. Rev. D64, 113012
(2001), [hep-ph/0106291].
the symmetry of the neutrino Dirac mass matrix, a re- [19] K. S. Babu, E. Ma and J. W. F. Valle, Phys. Lett. B552,
lation which holds in the framework of SO(10) unifica- 207 (2003), [hep-ph/0206292].
tion [30, 31]. A more complete picture formulated at [20] G. Altarelli and F. Feruglio, Nucl. Phys. B741, 215
(2006), [hep-ph/0512103].
the unified level is outside the scope of this letter. In
[21] M. Hirsch, A. Villanova del Moral, J. W. F. Valle
principle the structure presented here can be lifted to and E. Ma, Phys. Rev. D72, 091301 (2005),
the SO(10) level, though fitting the flavour structure of [hep-ph/0507148].
quarks will require additional fields and/or symmetries. [22] S.-L. Chen, M. Frigerio and E. Ma, Nucl. Phys. B724,
423 (2005), [hep-ph/0504181].
In such more complete scenario exact tri-bimaximality [23] A. Zee, Phys. Lett. B630, 58 (2005), [hep-ph/0508278].
would be just a first approximation, corrections lead- [24] E. Ma, Phys. Rev. D73, 057304 (2006).
ing to calculable deviations from the predictions reported [25] X.-G. He, Y.-Y. Keum and R. R. Volkas, JHEP 04, 039
(2006), [hep-ph/0601001].
here. These issues will be taken up elsewhere. [26] E. Ma, Mod. Phys. Lett. A22, 101 (2007),
[hep-ph/0610342].
Acknowledgments
[27] G. Altarelli and F. Feruglio, Nucl. Phys. B720, 64
Work supported by MEC grant FPA2005-01269, (2005), [hep-ph/0504165].
[28] ISS Physics Working Group, A. Bandyopadhyay et al.,
by European Commission Contracts MRTN-CT-2004- arXiv:0710.4947 [hep-ph].
503369 and ILIAS/N6 RII3-CT-2004-506222. [29] H. Nunokawa, S. J. Parke and J. W. F. Valle, Prog. Part.
Nucl. Phys. 60, 338 (2008), [arXiv:0710.0554 [hep-ph]].
[30] S. Morisi, M. Picariello and E. Torrente-Lujan,
hep-ph/0702034.
[31] W. Grimus and H. Kuhbock, Phys. Rev. D77, 055008
∗
Electronic address: mahirsch@ific.uv.es (2008), [0710.1585].
†
Electronic address: Stefano.Morisi@ific.uv.es [32] Due to its similarity to the columns of “Palacio da Alvo-
‡
Electronic address: valle@ific.uv.es rada” designed by this brilliant Brazilian arquitect.
[1] Super-Kamiokande collaboration, S. Fukuda et al., Phys. [33] Quarks are also flavor singlets, hence the quark sector is
Lett. B539, 179 (2002), [hep-ex/0205075]. completely standard and its flavor structure unpredicted.
[2] SNO collaboration, Q. R. Ahmad et al., Phys. Rev. Lett.
89, 011301 (2002), [nucl-ex/0204008].