Nuclear Matrix Elements For Neutrinoless Double-Beta Decay and Double-Electron Capture
Nuclear Matrix Elements For Neutrinoless Double-Beta Decay and Double-Electron Capture
Nuclear Matrix Elements For Neutrinoless Double-Beta Decay and Double-Electron Capture
Search
Collections
Journals
About
Contact us
My IOPscience
Nuclear matrix elements for neutrinoless double-beta decay and double-electron capture
This article has been downloaded from IOPscience. Please scroll down to see the full text article.
2012 J. Phys. G: Nucl. Part. Phys. 39 124006
(http://iopscience.iop.org/0954-3899/39/12/124006)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 81.151.89.87
The article was downloaded on 21/11/2012 at 11:17
IOP PUBLISHING
doi:10.1088/0954-3899/39/12/124006
E-mail: faessler@uni-tuebingen.de
1. Introduction
The physics community faces a challenging problem, discovering whether neutrinos are indeed
Majorana particles (i.e. identical to their own antiparticles) as many particle models suggest or
Dirac particles (i.e. different from their antiparticles). The best sensitivity on small Majorana
neutrino masses can be reached in the investigation of neutrinoless double-beta decay (0decay) [1, 2],
(A, Z) (A, Z + 2) + e + e
0954-3899/12/124006+19$33.00 2012 IOP Publishing Ltd
(1)
Printed in the UK & the USA
A Faessler et al
e
b + eb + (A, Z) (A, Z 2) .
(2)
A double asterisk in equation (2) means that, in general, the final atom (A, Z 2) is excited
with respect to both the electron shell, due to the formation of two vacancies for the electrons,
and the nucleus. Observing the 0-decay and/or 0ECEC would tell us that the total lepton
number is not a conserved quantity and that neutrinos are massive Majorana fermions.
The search for the 0-decay represents the new frontier of neutrino physics, allowing
us, in principle, to fix the neutrino mass scale, the neutrino nature and possibly CP violation
effects. There are a few tens of nuclear systems [4], which offer the opportunity to study the
0-decay and the most favorable are those with a large Q -value.
Neutrinoless double-beta decay has not yet been found. The strongest limits on the half-life
0
of the 0-decay were set in HeidelbergMoscow (76 Ge, 1.9 1025 years) [5], NEMO3
T1/2
(100 Mo, 1.0 1024 years) [6], CUORICINO (130 Te, 3.0 1024 years) [7] and KamLANDZen (136 Xe, 5.7 1024 years) [8] experiments. However, there is an unconfirmed, but not
refuted, claim of evidence for neutrinoless double decay in 76 Ge by some participants of the
0
25
= 2.23+0.44
HeidelbergMoscow Collaboration [9] with a half-life of T1/2
0.31 10 years. It is
expected that the GERDA experiment [10] in its first phase will check this result quite soon.
The main aim of experiments searching for 0-decay is the measurement of the effective
Majorana neutrino mass m
m =
Ue2j m j ,
(3)
(4)
is planned to be reached. This is the region of the inverted hierarchy of neutrino masses. In
the case of the normal mass hierarchy |m | is too small, a few meV, to be probed in the
0-decay experiments of the next generation.
We note that it is reasonable to hope that the search for the 0ECEC atoms, which
are sufficiently long lived to conduct a practical experiment, may also establish the Majorana
nature of neutrinos. This possibility is considered as alternative and complementary to searches
for the 0-decay.
To interpret the data from the 0-decay and the 0ECEC (neutrinoless double-electron
capture) accurately a better understanding of the nuclear structure effects important for the
description of the nuclear matrix elements (NMEs) is needed. In this connection it is crucial to
develop and advance theoretical methods capable of reliably evaluating NMEs, and realistically
assessing their uncertainties.
2
A Faessler et al
MF0
0
+ MGT
MT0
2
(geff
)
A
+ + +
HF (rkl )
= 0i
k l eff 2 + HGT (rkl )kl HT (rkl )Skl 0+f .
(gA )
kl
M 0 =
(7)
Here
k r kl )(
l r kl ) kl ,
Skl = 3(
k
l .
kl =
j0,0,2 (qrkl )hF,GT,T (q2 )q
2
dq,
HF,GT,T (rkl ) = R
q+E
0
(8)
(9)
where R is the nuclear radius and E is the average energy of the virtual intermediate states used
in the closure approximation. The closure approximation is adopted in all the calculations of
the NMEs relevant for 0-decay with the exception of the QRPA. The functions hF,GT,T (q2 )
are given by [12]
hF (q2 ) = fV2 (q2 ),
( p n )2 q2
q4
2
2 q2
1
2 2
,
+
f
(q
)
1
+
hGT (q2 ) = fV2 (q2 )
A
2
3
4m2p
3 q2 + m2
3 (q2 + m2 )2
(geff
A )
q4
1 2 2 ( p n )2 q2
1 2 2
q2
2
hT (q ) = fV (q )
.
+ f (q ) 2 2
2
3
4m2p 3 A
(q + m2 ) (q2 + m2 )2
(geff
A )
(10)
For the vector normalized to unity and axial-vector form factors the usual dipole approximation
is adopted: fV (q2 ) = 1/(1 + q2 /MV2 )2 , fA (q2 ) = 1/(1 + q2 /MA2 )2 . MV = 850 MeV, and MA =
1086 MeV. The difference in magnetic moments of proton and neutron is ( p n ) = 4.71,
and gA = 1.254 is assumed.
3
A Faessler et al
The above definition of the M0 includes the contribution of the higher order
terms of the nucleon current, and the GoldbergerTreiman PCAC relation, gP (q2 ) =
2m p gA (q2 )/(q2 + m2 ) was employed for the induced pseudoscalar term.
The nuclear matrix elements M 0 for the 0 decay must be evaluated using tools of
nuclear structure theory. Unfortunately, there are no observables that could be simply and
directly linked to the magnitude of 0 nuclear matrix elements and that could be used to
determine them in an essentially model independent way.
Over many years two approaches have been used: the quasiparticle random
phase approximation (QRPA)[1113] and the interacting shell model (ISM)[14]. There are
substantial differences between these approaches. The QRPA treats a large, single particle
model space, but heavily truncates the included configurations. The ISM, in contrast, treats
a small fraction of this model space, but allows the nucleons to correlate in many different
ways. In the last few years several new approaches have been used for the calculation of
the 0-decay NMEs: the angular momentum projected HartreeFockBogoliubov method
(PHFB) [15], the interacting boson model (IBM) [16], and the energy density functional
method (EDF) [17].
The standard QRPA method consists of two steps. First, the mean field corresponding to
the minimum energy is determined and the like-particle pairing interaction is taken into account
by employing the quasiparticle representation. In the second step the linearized equations of
motion are solved in order to describe small amplitude vibrational-like modes around that
minimum. In the renormalized version of QRPA (RQRPA) the violation of the Pauli exclusion
principle is partially corrected.
The drawback of QRPA is the fact that, unlike in BCS, the particle number is not conserved
automatically, even on average. For realistic Hamiltonians the differences between averaged
particle numbers on the RPA ground state and the exact particle numbers could be of the order of
unity (an extra or missing neutron or proton). The self-consistent renormalized QRPA method
(SRQRPA) removes this drawback by treating the BCS and QRPA vacua simultaneously [18].
For the neutron-proton systems the method was proposed and tested on the exactly solvable
simplified models in [19]. It is a generalization of the procedure proposed earlier in [20].
In the QRPA the phonon operators are defined as
k
k
A (pn)J,M ,
(11)
A
Y(pn)J
Q(k) = pn X(pn)J
J,M
(pn)J,M
where
and
are the usual variational amplitudes, and A(pn)J,M is the angular
momentum coupled two-quasiparticle creation operator. p, n signify the quantum numbers of
the proton, respectively neutron, orbits. The X and Y amplitudes, as well as the corresponding
energy eigenvalues k are determined by solving the QRPA eigenvalue equations for each J
A
B
X
X
=
.
(12)
B A
Y
Y
k
X(pn)J
k
Y(pn)J
The matrices A and B above are determined by the Hamiltonian rewritten in terms of the
quasiparticle operators:
a )(JM) |O
(13)
AJpn,p n = O|(ap an )(JM) H(a
p n
p an )(JM) (1)M (a a )(JM) |O
BJpn,p n = O|H(a
p n
Here, |O is the BCS vacuum state.
In the RQRPA and SRQRPA instead of |O the correlated QRPA ground state |0+
QRPA
is considered. Then instead of the standard X and Y the renormalized amplitudes are used
everywhere including the QRPA equations of motion:
m
m
X (pn,J ) = D1/2
pn X(pn,J ) ,
m
Y (pn,J ) = D1/2
pn Y(pn,J ) ,
(14)
A Faessler et al
D pn = 0+
QRPA A(pn)J,M , A(pn)J,M 0QRPA = 1 p n
J,k 2
1
=1
n D pn J,k (2J + 1)Y pn
2 jp + 1
J,k 2
1
p D p n J,k (2J + 1)Y p n .
2 jn + 1
Here, n(p) is the expectation value of the number of quasiparticles in the orbit n(p),
+ +
0QRPA an(p) an(p) 00 0+
QRPA
.
n(p)
2 jn(p) + 1
(15)
(16)
a+jn(p) ,m , a jn(p) ,m are the creation and annihilitation operators for the quasiparticle with quantum
numbers n(p), m. The renormalization coefficients D pn and the quasiparticle occupation
numbers j can be obtained iteratively using the equations of motion of the (S)RQRPA.
In the correlated QRPA ground state the occupation numbers are no longer the pure BCS
quantities. Instead, they depend, in addition, on the solutions of the QRPA equations of motion
for all multipoles J, and can be evaluated using
+
+
+
nQRPA
n(p) = 0QRPA m cn(p),m cn(p),m 0QRPA
2
2
2
n(p) .
(17)
(2 jn(p) + 1) vn(p) + un(p) vn(p)
Here, c+j,m is the creation operator for a proton in the orbit j p or a neutron in the orbit jn and
c j,m is the corresponding annihilation operator. The amplitudes v j p and v jn are obtained by
solving the gap equations.
In SRQRPA the BCS equations are reformulated. This is achieved by recalculating the u
and v amplitudes from the minimum condition of the RQRPA ground-state energy. In SRQRPA
thus the state around which the vibrational modes occur is no longer the quasiparticle vacuum,
but instead the Bogoliubov transformation is chosen in such a way that provides the optimal
and consistent basis while preserving the form of the phonon operator, equation (11).
In practice, the SRQRPA equations are solved double iteratively. One begins with the
standard BCS u, v amplitudes, solves the RQRPA equations of motion and calculates the
factors D pn . The u, v amplitudes are recalculated and the procedure is repeated until selfconsistency is achieved. Numerically, the double iteration procedure represents a challenging
problem. It was resolved in [21] where instead of the G-matrix based interaction the pairing
part (and only that part) of the problem was replaced by a pairing interaction that uses a
constant matrix element whose value was adjusted to reproduce the experimental oddeven
mass differences.
In the QRPA, RQRPA and SRQRPA the M 0 is written as the sum over the virtual
intermediate states, labeled by their angular momentum and parity J and indices ki and k f :
j
jn J
(1) jn + j p +J+J 2J + 1 p
MK =
jn j p J
J ,ki ,k f ,J pnp n
0+f c
p cn J |J k f J k f |J ki J ki || c p cn J 0i .
(18)
The operators OK , K = Fermi (F), GamowTeller (GT), and Tensor (T), contain neutrino
k ,k
potentials, spin and isospin operators, and RPA energies EJ i f . Two separate multipole
decompositions are built into equation (18). One is in terms the J of the virtual states in
5
A Faessler et al
the intermediate nucleus, the good quantum numbers of the QRPA and RQRPA. The other
decomposition is based on the angular momenta and parities J of the pairs of neutrons that
are transformed into protons with the same J . The nucleon orbits are labeled in equation
(18) by p, p , n, n .
The QRPA-like approaches do not allow the introduction of short-range correlations
(SRCs) into the two-nucleon relative wavefunction. The traditional way is to introduce an
explicit Jastrow-type correlation function f (r12 ) into the involved two-body transition matrix
elements (see equation (18)). In the parametrization of Miller and Spencer [22] we have
f (r12 ) = 1 cear (1 br2 ),
2
a = 1.1 fm2 ,
b = 0.68 fm2 .
(19)
These two parameters (a and b) are correlated and chosen in the way that the norm of the
relative wavefunction |
J is conserved.
Recently, it was proposed [13] that instead of the Jastrow method the UCOM approach
for the description of the two-body correlated wavefunction [23] was adopted. The UCOM
method produces good results for the binding energies of nuclei already at the HartreeFock
level [24].
A self-consistent calculation of the 0-decay NMEs in the QRPA-like approaches was
performed in [25]. The pairing and residual interactions as well as the two-nucleon short-range
correlations were for the first time derived from the same modern realistic nucleonnucleon
potentials, namely from the charge-dependent Bonn potential (CD-Bonn) and the Argonne
V18 potential. A method of choice was the coupled cluster method (CCM) [26]. For purpose
of numerical calculation of the 0-decay NMEs the CCM short-range correlation functions
were presented in an analytic form of Jastrow-like function as [25]
fA,B (r12 ) = 1 cear (1 br2 ).
2
(20)
(21)
The calculated NMEs with these short-range correlation functions agree to within a few percent
with those obtained without this approximation. We note that the dependence of the SRC on
the value of oscillator length b is rather weak.
In table 1 the QRPA and RQRPA results are presented separately for the different types
of two-nucleon short-range correlations (SRC) considered: MilllerSpencer Jastrow SRCs
(Jastrow) [11]; Fermi hypernetted chain SRCc (FHCh); unitary correlation operator method
SRCs (UCOM) [12]; the coupled cluster method SRCs derived from the Argonne and CDBonn potentials [25] based on an extension of the Brueckner theory (coupled cluster method =
CCM). Two different values of the axial coupling constant, free nucleon geff
A = gA = 1.254 and
quenched geff
A = 1.0, are taken into account. The strength of the particle-particle interaction
is adjusted so the experimental value of the 2-decay nuclear matrix element is correctly
reproduced [11]. The NME calculated within this procedure, which includes three different
model spaces, is denoted as the averaged 0-decay NME M 0 . We note that the values of
NMEs become essentially independent of the size of the single-particle basis and rather stable
with respect to the possible quenching of the gA .
From table 1 it follows that the QRPA values are about 10-15% larger in comparison with
the RQRPA values. The largest NMEs are those calculated with the CCM CD-Bonn correlation
function. In comparison the NMEs obtained with the CCM CD-Argonne correlation function
and the UCOM SRCs are about 10% smaller. This is explained by the fact that the CCM
Argonne correlation function cuts out the small r12 part from the relative wavefunction of the
6
A Faessler et al
Table 1. Averaged 0 nuclear matrix elements M 0 and their variance (in parentheses)
calculated within the QRPA and the RQRPA. Different types of two-nucleon short-range
correlations (SRC) are considered: Miller-Spencer Jastrow SRCs (Jastrow) [11]; Fermi hypernetted
chain SRCc (FHCh); unitary correlation operator method SRCs (UCOM) [12]; the coupled cluster
method SRCs derived from the Argonne and CD-Bonn potentials [25]. Three sets of single particle
level schemes are used, ranging in size from 9 to 23 orbits. The strength of the particle-particle
interaction is adjusted so the experimental value of the 2-decay nuclear matrix element is
eff
correctly reproduced. Both free nucleon (geff
A = gA = 1.254) and quenched (gA = 1.0) values
of axial-vector coupling constant are considered. We note that unlike in [11, 12, 25] r0 = 1.2 fm
instead of r0 = 1.1 fm is assumed.
M 0
Nucleus
SRC
CCM SRC
transition
geff
A
Method
Jastrow
FHCh
UCOM
76
1.25
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
QRPA
RQRPA
4.92(0.19)
4.28(0.13)
4.18(0.15)
3.77(0.14)
4.39(0.16)
3.81(0.14)
3.59(0.13)
3.17(0.10)
1.22(0.03)
1.31(0.15)
1.32(0.08)
1.22(0.12)
3.64(0.21)
3.03(0.21)
2.96(0.15)
2.55(0.13)
2.99(0.21)
2.64(0.17)
2.38(0.17)
2.14(0.14)
3.97(0.14)
3.52(0.13)
3.11(0.09)
2.77(0.09)
3.56(0.13)
3.22(0.13)
2.55(0.08)
2.15(0.14)
2.16(0.13)
2.02(0.12)
1.70(0.09)
1.59(0.09)
5.15(0.17)
4.48(0.13)
4.36(0.15)
3.94(0.13)
4.57(0.16)
3.97(0.14)
3.74(0.13)
3.32(0.10)
1.23(0.04)
1.33(0.15)
1.34(0.07)
1.24(0.12)
3.73(0.21)
3.12(0.21)
3.02(0.15)
2.63(0.13)
3.11(0.21)
2.75(0.19)
2.47(0.17)
2.21(0.14)
4.15(0.15)
3.68(0.14)
3.23(0.10)
2.88(0.09)
3.72(0.14)
3.36(0.15)
2.93(0.08)
2.66(0.09)
2.25(0.12)
2.11(0.14)
1.77(0.09)
1.66(0.10)
5.98(0.27)
5.17(0.20)
4.97(0.23)
4.47(0.20)
5.32(0.23)
4.59(0.17)
4.29(0.19)
3.79(0.13)
1.77(0.02)
1.77(0.02)
1.73(0.10)
1.57(0.14)
4.71(0.28)
3.88(0.26)
3.74(0.21)
3.20(0.17)
3.74(0.12)
3.21(0.22)
2.88(0.17)
2.55(0.17)
5.04(0.15)
4.45(0.15)
3.88(0.11)
3.44(0.10)
4.53(0.12)
4.07(0.13)
3.52(0.07)
3.17(0.08)
2.73(0.13)
2.54(0.15)
2.12(0.11)
1.97(0.11)
Ge
1.00
82
Se
1.25
1.00
96
Zr
1.25
1.00
100
Mo
1.25
1.00
116
Cd
1.25
1.00
128
Te
1.25
1.00
130
Te
1.25
1.00
136
Xe
1.25
1.00
Argonne
6.34(0.29)
5.42(0.21)
5.20(0.22)
4.59(0.15)
5.66(0.26)
4.84(0.21)
4.57(0.20)
4.00(0.15)
2.07(0.10)
2.01(0.17)
1.90(0.12)
1.69(0.13)
5.18(0.36)
4.20(0.34)
4.03(0.27)
3.43(0.25)
3.86(0.29)
3.34(0.24)
2.99(0.23)
2.69(0.19)
5.38(0.17)
4.71(0.17)
4.11(0.13)
3.62(0.12)
4.77(0.15)
4.27(0.15)
3.69(0.11)
3.29(0.11)
2.88(0.14)
2.68(0.16)
2.21(0.10)
2.06(0.11)
CD-Bonn
6.89(0.35)
5.93(0.25)
5.63(0.27)
5.04(0.24)
6.16(0.29)
5.30(0.22)
4.89(0.22)
4.29(0.16)
2.28(0.03)
2.19(0.22)
2.11(0.12)
1.88(0.16)
5.73(0.34)
4.67(0.31)
4.44(0.24)
3.75(0.21)
4.35(0.16)
3.72(0.26)
3.31(0.21)
2.92(0.21)
5.99(0.17)
5.26(0.16)
4.54(0.13)
4.00(0.12)
5.37(0.13)
4.80(0.14)
4.11(0.08)
3.69(0.09)
3.23(0.14)
3.00(0.17)
2.47(0.09)
2.30(0.12)
two-nucleons more than the CCM CD-Bonn correlation function. The smallest in magnitude
are matrix elements for the 0 decay obtained with the traditional approach of using
the MillerSpencer Jastrow SRC and the Fermi hypernetted chain SRCc.
In table 2 we show the calculated ranges of the nuclear matrix element M 0 evaluated
within the QRPA, RQRPA [25] and SRQRPA [27] in a self-consistent way with the CCM CDBonn and Argonne SRC functions by assuming both the standard (gA = 1.254) and quenched
(gA = 1.0) axial-vector couplings. These ranges quantify the uncertainty in the calculated
7
QRPA
RQRPA
0
Nuclei
M 0
0
T1/2
[years]
76
(5.0, 7.2)
(4.4, 6.4)
(3.7, 6.1)
(3.6, 5.5)
(2.1, 3.4)
(4.5, 6.2)
(3.8, 5.6)
(3.2, 5.0)
(3.2, 4.7)
(2.0, 3.2)
Ge
Se
100
Mo
130
Te
136
Xe
82
8
Table 2. The calculated ranges of the nuclear matrix element M 0 evaluated within the QRPA (column 2), RQRPA (column 4) and SRQRPA (column 6), with standard (geff
A = gA = 1.254)
and quenched (geff
A = 1.0) axial-vector couplings and with the coupled cluster method (CCM) CD-Bonn and Argonne short-range correlation (SRC) functions. Columns 3, 5 and 7 give
0exp
the 0-decay half-life ranges corresponding to values of the matrix-elements in columns 2, 4 and 6 for |m | = 50 meV. T1/2
is the experimental lower bound on the 0-decay
half-life for a given isotope.
SRQRPA
0
0
T1/2
[years]
(4.3, 6.2)
(3.9, 6.1)
(4.0, 5.5)
(3.6, 5.1)
(2.4, 3.6)
0
T1/2
[years]
0exp
T1/2
[y]
1.9 1025
3.2 1023
1.0 1024
3.0 1024
5.7 1024
[5]
[6]
[6]
[7]
[8]
A Faessler et al
A Faessler et al
0-decay NMEs of a given QRPA-like approach. By comparing the SRQRPA with the
RQRPA results we conclude that the requirement of conserving the particle number have not
caused substantial changes in the value of the 0 matrix elements in that case.
Given the interest in the subject, in table 2 we show also the range of predicted 0-decay
half-lives of 76 Ge, 82 Se, 100 Mo, 130 Te and 136 Xe corresponding to the full range of M 0 for
|m | = 50 meV. This is a rather conservative range within the considered QRPA framework.
It represents roughly the required sensitivity of the 0-decay experiment in the case of
inverted hierarchy of neutrino masses, which can be compared with the current bound on the
0exp
.
0-decay half-life T1/2
3. decay of deformed nuclei within QRPA
One of the best candidates for searching 0 decay is 150 Nd since it has the second highest
endpoint, Q =3.37 MeV, and the largest phase space factor for the decay (about 33 times
larger than that for 76 Ge, see e.g. [28]). The SNO+ experiment at the Sudbury Neutrino
Observatory will use a Nd-loaded scintillator to search for neutrinoless double-beta decay by
looking for a distortion in the energy spectrum of decays at the endpoint [29].
However, 150 Nd is well known to be a rather strongly deformed nucleus. This strongly
hinders a reliable theoretical evaluation of the corresponding 0-decay NMEs (for instance,
it does not seem feasible in the near future to reliably treat this nucleus within the large-scale
nuclear shell model (ISM), see, e.g., [14]). Recently, more phenomenological approaches like
the pseudo-SU(3) model [30], the PHFB approach [15], the IBM [16], and the EDF [17]
have been employed to calculate M 0 for strongly deformed heavy nuclei (a comparative
analysis of different approximations involved in some of the models can be found in [31]).
The results of these models generally reveal a substantial suppression of M 0 for 150 Nd as
compared with the QRPA result of [11] where 150 Nd and 150 Sm were treated as spherical
nuclei. However, the calculated NMEs M 0 for 150 Nd show a rather significant spread.
One of the most up-to-date microscopic ways to describe the effect of nuclear deformation
on -decay NMEs M 2 and M 0 is provided by the QRPA. A QRPA approach for calculating
-decay amplitudes in deformed nuclei has been developed in a series of papers [3236].
M 2 were calculated in [32, 33] with schematic separable forces, and in [34]with realistic
residual interaction. It was demonstrated in [3234] that deformation introduces a mechanism
of suppression of the M 2 matrix element which gets stronger when deformations of the initial
and final nuclei differ from each other. A similar dependence of the suppression of both M 2
and M 0 matrix elements on the difference in deformations has been found in the PHFB [15]
and the ISM [14].
In [35, 36], the first QRPA calculations of M 0 with an account for nuclear deformation
were done. The calculations showed a suppression of M 0 for 150 Nd by about 40% as compared
with our previous spherical QRPA result [11]. In the next section we review the results of
[3236].
3.1. Formalism
The NMEs M 2 and M 0 , as the scalar measures of the decay rates, can be calculated in
any coordinate system. For strongly deformed, axially symmetric, nuclei the most convenient
choice is the intrinsic coordinate system associated with the rotating nucleus. This employs
the adiabatic BohrMottelson approximation that is well justified for 150 Nd, 160 Gd and 160 Dy,
which indeed reveal strong deformations. As for 150 Sm, the enhanced quadrupole moment of
this nucleus is an indication of its static deformation.
9
A Faessler et al
Though it is difficult to evaluate the effects beyond the adiabatic approximation employed
here, one might anticipate already without calculations that the smaller the deformation is, the
smaller should be the deviation of the calculated observables from the ones obtained in the
spherical limit. In this connection it is worth noting that spherical QRPA results can be exactly
reproduced in the present QRPA calculation by letting the deformation vanish, in spite of the
formal inapplicability of the adiabatic ansatz for the wavefunction in this limit.
We give here, for completeness, the formalism of the QRPA calculations of NMEs M 2
and M 0 in deformed nuclei as developed in [3236].
Nuclear excitations in the intrinsic system |K are characterized by the projection of
the total angular momentum onto the nuclear symmetry axis K (the only projection which is
conserved in strongly deformed nuclei) and the parity .
The intrinsic states |K , m are generated within the QRPA by a phonon creation operator
acting on the ground-state wavefunction:
m
m
A pn,K .
|K , m = Q 0+
=
Xpn,K
A Ypn,K
(22)
g.s. ; Q
m,K
m,K
pn,K
pn
Here, Apn,K = ap an and A pn,K = a p an are the two-quasiparticle creation and annihilation
operators, respectively, with the bar denoting the time-reversal operation. The quasiparticle
pairs pn are defined by the selection rules p n = K and p n = , where is the
single-particle (s.p.) parity and is the projection of the total s.p. angular momentum on
the nuclear symmetry axis ( = p, n). The s.p. states |p and |n of protons and neutrons
are calculated by solving the Schrodinger equation with the deformed axially symmetric
WoodsSaxon potential [34]. In the cylindrical coordinates the deformed WoodsSaxon s.p.
wavefunctions | with > 0 are decomposed over the deformed harmonic oscillator
s.p. wavefunctions (with the principal quantum numbers (Nnz )) and the spin wavefunctions
| = 12 :
| =
bNnz |Nnz = |,
(23)
Nnz
where N = n + nz (n = 2n + ||), nz and n are the number of nodes of the basis functions
in the z- and -directions, respectively; = and are the projections of the orbital
and spin angular momentum onto the symmetry axis z. For the s.p. states with the negative
projection = | |, which are degenerate in energy with = | |, the time-reversed
version of equation (23) is used as a definition (see also [34]). The states (, ) comprise the
whole single-particle model space.
The deformed harmonic oscillator wavefunctions |Nnz can be further decomposed over
the spherical harmonic oscillator ones |nr l by calculating the corresponding spatial overlap
rl
= nr l|Nnz (nr is the radial quantum number, l and are the orbital
integrals AnNn
z
angular momentum and its projection onto z-axes, respectively), see appendix of [34] for more
details. Thereby, the wavefunction (23) can be re-expressed as
| =
B | ,
(24)
where | =
j
Cl 1 |nr l
j
being the ClebschGordan coefficient.
with Cl
1
,
K,m
YKm
YKm
B(K) A(K)
10
(25)
A Faessler et al
m
with realistic residual interaction are solved to get the forward XiK
, backward YiKm amplitudes
m
and the excitation energies Kmi and K f of the mth K state in the intermediate nucleus. The
matrix A and B are defined by
A pn,p n (K) = pn,p n (E p + En ) + g pp (u p un u p un + v p vn v p vn )Vpnp
n
p n
p n
JK
jn n JK
where | p n , JK = m p mn CJK
C j p p jn n
j p m p jn mn | p m p |n mn , and Fp p nn = B p Bn (1)
jn n
is defined for the sake of simplicity ((1)
is the phase arising from the time-reversed
states |n).
The particleparticle Vpn,p
n and particle-hole Vpn ,p n interaction matrix elements
in the representation (26) for the QRPA matrices A, B (25) in the deformed WoodsSaxon
single-particle basis can then be given in terms of the spherical G matrix elements as follows:
JK
Vpn,p
=
2
Fp
F JK
G( p n p n , J),
(28)
n
p nn p p n n
p n p n
Vpn ,p n = 2
J
p n p n
JK
JK
Fp ppnn n Fp ppn n
n G( p n p n , J),
(29)
where Kpn
= p + n = p + n .
The structure of the intermediate |0+ and |1+ states is only needed within the QRPA to
calculate 2-decay NMEs M 2 [34], whereas all possible |K states are needed to construct
the NMEs M 0 .
2
is given within the QRPA in the intrinsic system by the following
The matrix element MGT
expression:
0+f |K |K + , m f K + , m f |K + , mi K + , mi |K |0+
i
2
MGT
=
.
(30)
K,mi m f
K=0,1 m m
i
Instead of the usual approximation of the energy denominator in equation (30) as K,mi m f =
(K,m f + K,mi )/2 (see, e.g., [32, 33]), here another prescription is used in which the whole
calculated QRPA energy spectrum is shifted in such a way as to have the first calculated 1+
state exactly at the corresponding experimental energy. In this case the energy denominator
in equation (30) acquires the form K,mi m f = (K,m f K,1 f + K,mi K,1i )/2 + 1+1 , with
1+1 being the experimental excitation energy of the first 1+ state measured from the mean g.s.
energy (E0i + E0 f )/2.
The two sets of intermediate nuclear states generated from the initial and final g.s. do not
come out identical within the QRPA. Therefore, the overlap factor of these states is introduced
in equation (30) [32] as follows:
m m
m
(31)
Xl f Kf Xli Ki Yl f Kf Ylmi Ki Rl f li BCS f |BCSi .
K + , m f |K + , mi =
li l f
11
A Faessler et al
The factor Rl f li , which includes the overlaps of single particle wavefunctions of the initial and
final nuclei is given by
(i) ( f )
(f)
(i) ( f )
(i) ( f )
(32)
Rll = p p |p p u(i)
p u p + v p v p nn |n n un un + vn vn ,
and the last term BCS f |BCSi in equation (31) corresponds to the overlap factor of the initial
and final BCS vacua in the form given in [32].
The matrix element M 0 is given within the QRPA in the intrinsic system by a sum of the
partial amplitudes of transitions via all the intermediate states K :
M 0 =
M 0 (K );
M 0 (K ) =
s(def) O (K ).
(33)
Here, we use the notation of Appendix B in [12], stands for the set of four single-particle
indices {p, p , n, n }, and O (K ) is a two-nucleon transition amplitude via the K states in
the intrinsic frame:
0+f cp cn |K m f K m f |K mi K mi |cp cn 0+
(34)
O (K ) =
i .
mi ,m f
The two sets of intermediate nuclear states generated from the initial and final g.s. (labeled by
mi and m f , respectively) do not come out identical within the QRPA. A standard way to tackle
this problem is to introduce in equation (34) the overlap factor of these states K m f |K mi ,
whose representation is given below, equation (37). Two-body matrix elements s(def) of the
neutrino potential in equation (33) in a deformed WoodsSaxon single-particle basis are
decomposed over the spherical harmonic oscillator ones according to equations (27) and (29):
JK
Fp
F JK
s(sph)
(J),
(35)
s(def)
pp nn =
p nn p p n n p p n n
J
p p
n n
s(sph)
pp nn (J)
(1)
jn + j p +J+J
jp
jn
jn
j p
J
p(1), p (2); J O (1, 2)n(1), n (2); J ,
J
(36)
mi
mi
K mi |cp cn |0+
i = u p vn X pn,K + v p unYpn,K .
(37)
li l f
Representations for Rl f li and the overlap factor BCS f |BCSi between the initial and final
BCS vacua are given in [32].
12
A Faessler et al
A Faessler et al
sph (2 =0)
Nuclear transition
geff
A
0
0
T1/2
def
[years]
0
0
T1/2
[years]
0-decay
76
76
Ge Se
150
Nd150 Sm
160
Gd160 Dy
0.94
1.25
0.94
1.25
0.94
4.10
5.30
4.52
6.12
9.4 1026
5.6 1026
2.3 1025
1.2 1025
4.00
4.69
2.55
3.34
3.76
9.8 1026
7.2 1026
7.1 1025
4.1 1025
2.3 1026
3.23
2.64
2.05
0ECEC
152
152
Gd Se (KL1 )
Er164 Dy (L1 L1 )
180
W180 Hf (KK)
164
1.269
1.269
1.269
7.50
7.20
6.22
Table 4. The matrix elements M 0 for the 0 decay 150 Nd150 Sm calculated in different
0 (in years) for an assumed effective Majorana neutrino
models. The corresponding half-lives T1/2
mass m = 50 meV are also shown.
Method
Pseudo-SU(3) [30]
PHFB [15]
IBM [16]
EDF [17]
M 0
0
T1/2
(1025 y)
2.95 0.4
5.6 1.5
1.57
18.7
3.24 0.44
4.6 1.2
2.32
8.54
1.71
16.5
A Faessler et al
50 meV seems to be short enough to hope that the SNO+ experiment will be able to approach
the inverse hierarchy of the neutrino mass spectrum.
4. The resonant neutrinoless double-electron capture
The resonant 0ECEC (neutrinoless double-electron capture) was already considered as
a process which might prove the Majorana nature of neutrinos and the violation of the
total lepton number, by Winter [40] in 1955. The possibility of a resonant enhancement
of the 0ECEC in the case of a mass degeneracy between the initial and final atoms was
pointed out by Bernabeu, De Rujula and Jarlskog as well as by Vergados about 30 years ago
[41, 42]. They estimated the half-life of the process by introducing different simplifications:
(i) non-relativistic atomic wavefunctions at the nuclear origin; (ii) qualitative evaluation of
was assumed to be
NME of the process; (iii) the degeneracy parameter = MA,Z MA,Z2
within the range (0,10) keV representing the accuracy of atomic mass measurement at that
are masses of the initial and final excited atoms, respectively. A list
time. MA,Z and MA,Z2
of promising isotopes based on the degeneracy requirement associated with arbitrary nuclear
excitation and on the natural abundance of daughter atom was presented.
In 2004 Sujkowski and Wycech [43] and Lukaszuk et al [44] analyzed the 0ECEC process
for nuclear 0+ 0+ transitions accompanied by a photon emission in the resonance and nonresonance modes. By assuming |m | = 1 eV and 1 error in the atomic mass determination
the resonant 0ECEC rates of six selected isotopes were calculated by considering the
perturbation theory approach.
In 2009 a new theoretical approach to the 0ECEC, a unified description of the oscillations
(38)
where J denotes angular momentum and parity of final nucleus. The degeneracy parameter
can be expressed as
= Q Bab E ,
= MA,Z MA,Z2
(39)
15
A Faessler et al
where Q stands for a difference between the initial and final atomic masses in ground states
and E is an excitation energy of the daughter nucleus. Bab = Ea + Eb + EC is the energy of
two electron holes, whose quantum numbers (n, j, l) are denoted by indices a and b and EC
is the interaction energy of the two holes. The binding energies of single electron holes Ea
are known with an accuracy of a few eV [46]. The width of the excited final atom with the
electron holes is given by
ab = a + b + .
(40)
Here, a,b is one-hole atomic width and is the de-excitation width of the daughter nucleus,
which can be neglected. Numerical values of ab are about up to a few tens of eV. By
factorizing the electron shell structure and nuclear matrix element for lepton numbers violating
the amplitude associated with nuclear transitions 0+ J = 01 , 11 one gets
(geff )2
1 2
G me A Fab M 0ECEC (J ).
(41)
4
R
Here, Fab is a combination of averaged upper and lower bispinor components of the atomic
electron wavefunctions [3] and M 0ECEC (J ) is the nuclear matrix element. We note that by
neglecting the lower bispinor components M 0ECEC (0+ ) takes the form of the 0-decay
NME for ground state to ground state transition after replacing isospin operators by + . R
is the nuclear radius and gA is the axial-vector coupling constant.
The probability of the 0ECEC is increased by many orders of magnitude provided the
resonance condition is satisfied within a few tens of electron-volts. For a long time there was
no way to identify promising isotopes for the experimental search for 0ECEC, because of
the poor experimental accuracy of measurement of Q-values of the order of 110 keV for
medium heavy nuclei. Progress in the precision measurement of atomic masses with Penning
traps [4749] has revived the interest in the old idea of the resonance 0ECEC. The accuracy
of Q-values at around 100 eV was achieved. The estimates of the 0ECEC half-lives were
recently improved by more accurate measurements of Q-values for 74 Se [50, 51], 96 Ru [59],
106
Cd [53, 52], 102 Pd [52], 112 Sn [54], 120 Te [55], 136 Ce [56], 144 Sm [52], 152 Gd [57], 156 Dy
[58], 162 Er [59], 164 Er [60], 168 Yb [59] and 180 W [61]. It allowed the exclusion of some isotopes
from the list of the most promising candidates (e.g., 112 Sn, 164 Er, 180 W) for searching for the
0ECEC.
Among the promising isotopes, 152 Gd has likely resonance transitions to the 0+ ground
states of the final nucleus as it follows from improved measurement of Q-value for this
transition with accuracy of about 100 eV [57]. A detailed calculation of the 0ECEC of 152 Gd
was performed in [62] (see table 3). The atomic electron wave functions were treated in the
relativistic DiracHartreeFock approximation [63]. The NME for ground state to ground
state transition 152 Gd 152 Sm was calculated within the proton-neutron deformed QRPA
with a realistic residual interaction [57]. For the favored capture of electrons from K and
L shells in the case of 152 Gd the 0ECEC half-life is in the range 4.7 1028 4.8 1029
years. This transition is still rather far from the resonant level. Currently, the 0ECEC halflife of 152 Gd is 23 orders of magnitude longer than the half-life of 0 decay of 76 Ge
corresponding to the same value of |m | and is the smallest known half-life among known
0ECEC.
The resonant 0ECEC has some important advantages with respect to experimental
signatures and background conditions. The ground state to ground state resonant 0ECEC
transitions can be detected by monitoring the X rays or Auger electrons emitted from the
excited electron shell of the atom. This can be achieved, e.g., by calorimetric measurements.
The de-excitation of the final excited nucleus proceeds in most cases through a cascade of
easy to detect rays. A coincidence setup can cut down any background rate right from the
Vab (J ) =
16
A Faessler et al
beginning, thereby requiring significantly less active or passive shielding. A clear detection
of these rays would already signal the resonant 0ECEC without any doubt, as there are
no background processes feeding those particular nuclear levels. We note that standard model
allowed double-electron capture with emission of two neutrinos,
e
b + eb + (A, Z) (A, Z 2) + e + e ,
(42)
A Faessler et al
of the nuclear matrix elements is a very important and challenging problem. We presented
improved calculation of the 0-decay and 0ECEC NMEs, which includes a consistent
treatment of the two-nucleon short-range correlations and deformation effects. In addition, the
possibility of measuring the 0-decay NME was addressed.
Acknowledgments
This work was supported in part by the Deutsche Forschungsgemeinschaft within the project
Nuclear matrix elements of Neutrino Physics and Cosmology FA67/40-1, the VEGA Grant
agency of the Slovak Republic under the contract number 1/0639/09 and by the grant of the
Ministry of Education and Science of the Russian Federation (contract 12.741.12.0150).
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
18
Krivoruchenko M I, Simkovic
F, Frekers D and Faessler A 2011 Nucl. Phys. A 859 140
Tretyak V I and Zdesenko Y G 2002 At. Data Nucl. Data Tables 80 83
Baudis L et al (The HeidelbergMoscow Collaboration) 1999 Phys. Rev. Lett. 83 41
Tretyak V I (The NEMOIII Collaboration) 2011 AIP Conf. Proc. 1417 125
Arnaboldi C et al (The CUORE Collaboration) 2004 Phys. Lett. B 584 260
Gando A et al (The KamLAND-Zen Collaboration) 2012 arXiv:1201.4664 [hep-ex]
Klapdor-Kleingrothaus H V and Krivosheina I V 2006 Mod. Phys. Lett. A 21 1547
Jochum J (GERDA Collaboration) 2010 Prog. Part. Nucl. Phys. 64 261
Schonert S (GERDA Collaboration) 2010 J. Phys.: Conf. Ser. 203 012014
Simkovic
F, Faessler A, Rodin V A, Vogel P and Engel J 2008 Phys. Rev. C 77 045503
Kortelainen M, Civitarese O, Suhonen J and Toivanen J 2007 Phys. Lett. B 647 128
Kortelainen M and Suhonen J 2007 Phys. Rev. C 75 051303
Kortelainen M and Suhonen J 2007 Phys. Rev. C 76 024315
Menendez J, Poves A, Caurier E and Nowacki F 2009 Nucl. Phys. A 818 139
Rath P K, Chandra R, Chaturvedi K, Raina P K and Hirsch J G 2010 Phys. Rev. C 82 064310
Barea J and Iachello F 2009 Phys. Rev. C 79 044301
Rodriguez T R and Martinez-Pinedo G 2010 Phys. Rev. Lett. 105 252503
Simkovic
F, Faessler A and Vogel P 2009 Phys. Rev. C 79 015502
Delion D S, Dukelsky J and Schuck P 1997 Phys. Rev. C 55 2340
Krmpotic F et al 1998 Nucl. Phys. A 637 295
Dukelsky J and Schuck P 1996 Phys. Lett. B 387 233
Simkovic
F, Faessler A, Muther H, Rodin V and Stauf M 2009 Phys. Rev. C 79 055501
Muther H and Polls A 1999 Phys. Rev. C 61 014304
Muther H and Polls A 2000 Prog. Part. Nucl. Phys. 45 243
A Faessler et al
[32] Simkovic
F, Pacearescu L and Faessler A 2004 Nucl. Phys. A 733 321
[33] Alvarez-Rodriguez R, Sarriguren P, Guerra E Moya de, Pacearescu L, Faessler A and Simkovic
F 2004 Phys.
Rev. C 70 064309
[45] Simkovic
F and Krivoruchenko M I 2009 Phys. Part. Nucl. Lett. 6 298
[46] Larkins F B 1977 At. Data Nucl. Data Tables 20 313
[47] Douysset G et al 2001 Phys. Rev. Lett. 86 4259
[48] Blaum K 2006 Phys. Rep. 425 1
[49] Blaum K, Novikov Y N and Werth G 2010 Contemp. Phys. 51 149
[50] Kolhinen V S et al 2010 Phys. Lett B 684 17
[51] Mount B J, Redshaw M and Myers E G 2010 Phys. Rev. C 81 032501
[52] Goncharov M et al 2011 Phys. Rev. C 84 028501
[53] Smorra C et al 2012 Phys. Rev. C 85 027601
[54] Rahaman S et al 2009 Phys. Rev. Lett. 103 042501
[55] Scielzo N D et al 2009 Phys. Rev. C 80 025501
[56] S V and Kolhinen et al 2011 Phys. Lett. B 697 116
[57] Eliseev S et al 2011 Phys. Rev. Lett. 106 052504
[58] Eliseev S et al 2011 Phys. Rev. C 84 012501
[59] Eliseev S et al 2011 Phys. Rev. C 83 038501
[60] Eliseev S et al 2011 Phys. Rev. Lett. 107 152501
[61] Droese Ch et al 2011 Nucl. Phys. A 875 1
19