Search for Majorana Neutrinos Near the Inverted Mass Hierarchy
Region with KamLAND-Zen
A. Gando,1 Y. Gando,1 T. Hachiya,1 A. Hayashi,1 S. Hayashida,1 H. Ikeda,1 K. Inoue,1,2 K. Ishidoshiro,1 Y. Karino,1
M. Koga,1,2 S. Matsuda,1 T. Mitsui,1 K. Nakamura,1,2 S. Obara,1 T. Oura,1 H. Ozaki,1 I. Shimizu,1 Y. Shirahata,1 J. Shirai,1
A. Suzuki,1 T. Takai,1 K. Tamae,1 Y. Teraoka,1 K. Ueshima,1 H. Watanabe,1 A. Kozlov,2 Y. Takemoto,2 S. Yoshida,3
K. Fushimi,4 T. I. Banks,5 B. E. Berger,2,5 B. K. Fujikawa,2,5 T. O’Donnell,5 L. A. Winslow,6 Y. Efremenko,2,7
H. J. Karwowski,8 D. M. Markoff,8 W. Tornow,2,8 J. A. Detwiler,2,9 S. Enomoto,2,9 and M. P. Decowski2,10
(KamLAND-Zen Collaboration)
1Research Center for Neutrino Science, Tohoku University, Sendai 980-8578, Japan
2Kavli Institute for the Physics and Mathematics of the Universe (WPI), The University of Tokyo Institutes for Advanced Study,
The University of Tokyo, Kashiwa, Chiba 277-8583, Japan
3Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan
4Faculty of Integrated Arts and Science, University of Tokushima, Tokushima 770-8502, Japan
5Physics Department, University of California, Berkeley, and Lawrence Berkeley National Laboratory,
Berkeley, California 94720, USA
6Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA
7Department of Physics and Astronomy, University of Tennessee, Knoxville, Tennessee 37996, USA
8Triangle Universities Nuclear Laboratory, Durham, North Carolina 27708, USA; Physics Departments at Duke University, Durham,
North Carolina 27705, USA; and North Carolina Central University, Durham, North Carolina 27701, USA
9Center for Experimental Nuclear Physics and Astrophysics, University of Washington, Seattle, Washington 98195, USA
10Nikhef and the University of Amsterdam, Science Park, Amsterdam 1098XG, The Netherlands
(Received 10 May 2016; revised manuscript received 21 June 2016; published 16 August 2016)
We present an improved search for neutrinoless double-beta (0νββ) decay of 136Xe in the KamLAND-
Zen experiment. Owing to purification of the xenon-loaded liquid scintillator, we achieved a significant
reduction of the 110mAg contaminant identified in previous searches. Combining the results from the
first and second phase, we obtain a lower limit for the 0νββ decay half-life of T0ν1=2 > 1.07 × 10
26 yr at
90% C.L., an almost sixfold improvement over previous limits. Using commonly adopted nuclear matrix
element calculations, the corresponding upper limits on the effective Majorana neutrino mass are in the
range 61–165 meV. For the most optimistic nuclear matrix elements, this limit reaches the bottom of the
quasidegenerate neutrino mass region.
DOI: 10.1103/PhysRevLett.117.082503
Neutrinoless double-beta (0νββ) decay is an exotic
nuclear process predicted by extensions of the Standard
Model of particle physics. Observation of this decay
demonstrates the nonconservation of lepton number, and
proves that neutrinos have a Majorana mass component.
In the framework of light Majorana neutrino exchange, its
decay rate is proportional to the square of the effective
Majorana neutrino mass hmββi≡ jPiU2eimνi j. Recent 0νββ
searches [1] involving 76Ge (GERDA [2]) and 136Xe
(KamLAND-Zen [3] and EXO-200 [4]) provide upper
limits on hmββi of ∼0.2–0.4 eV using available nuclear
matrix element (NME) values from the literature. The
sensitivities of these searches correspond to mass scales
in the so-called quasidegenerate mass region.
KamLAND-Zen is a double-beta decay experiment that
exploits the existing detection infrastructure and radio-
purity of KamLAND [5,6]. The KamLAND-Zen detector
consists of 13 tons of Xe-loaded liquid scintillator (Xe-LS)
contained in a 3.08-m-diameter spherical inner balloon (IB)
located at the center of the KamLAND detector. The IB
is constructed from 25-μm-thick transparent nylon film
and is surrounded by 1 kton of liquid scintillator (LS)
contained in a 13-m-diameter spherical outer balloon. The
outer LS acts as an active shield. The scintillation photons
are viewed by 1879 photomultiplier tubes (PMTs) mounted
on the inner surface of the containment vessel. The Xe-LS
consists of 80.7% decane and 19.3% pseudocumene
(1,2,4-trimethylbenzene) by volume, 2.29 g=liter of the
fluor PPO (2,5-diphenyloxazole), and ð2.91  0.04Þ% by
weight of isotopically enriched xenon gas. The isotopic
abundances in the enriched xenon were measured by a
residual gas analyzer to be ð90.77  0.08Þ% 136Xe, ð8.96
0.02Þ% 134Xe. Other xenon isotopes have negligible
presence. The two electrons emitted from 136Xe ββ decay
PRL 117, 082503 (2016) P HY S I CA L R EV I EW LE T T ER S
week ending
19 AUGUST 2016
0031-9007=16=117(8)=082503(6) 082503-1 © 2016 American Physical Society
produce scintillation light and their summed energy is
observed. Hypothetical 0νββ decays would produce a peak
at the Q value of the decay, distinguishable from 2νββ
decays that have a continuous spectrum.
In the first phase of KamLAND-Zen (phase I) [3], we
obtained a lower limit of T0ν1=2 > 1.9 × 10
25 yr (90% C.L.)
on the 136Xe 0νββ decay half-life. The sensitivity of the
phase-I search was limited by the presence of an unex-
pected background peak, consistent with 110mAg β− decay
(τ ¼ 360 day, Q ¼ 3.01 MeV), just above the 2.458 MeV
Q value of 136Xe ββ decay. After completing phase I, we
embarked on a Xe-LS purification campaign that continued
for 18 months. First, we extracted the Xe-LS in small
batches from the IB through a teflon tube whose intake was
near the bottom of the IB volume. We then isolated and
stored the Xe before placing the Xe-depleted LS back in the
top of the IB where it was later replaced by a new LS. This
new LS was initially purified by water extraction followed
by vacuum distillation. The replacement of the Xe-depleted
LS was performed in three cycles equivalent to one IB
volume exchange for each cycle. The LS was purified by
vacuum distillation during each cycle. We also purified a
mix of recovered and new Xe through distillation and
refining with a heated zirconium getter. Finally, the Xe was
dissolved into the purified LS. In December of 2013, we
started the second science run (phase II), and found a
reduction of 110mAg by more than a factor of 10. We report
on the analysis of the complete phase-II data set, collected
between December 11, 2013 and October 27, 2015. The
total live time is 534.5 days after muon spallation cuts,
discussed later. This corresponds to an exposure of 504 kg
yr of 136Xe with the whole Xe-LS volume.
Following the end of phase II, we performed a detector
calibration campaign using radioactive sources deployed at
various positions along the central axis of the IB. The event
position reconstruction—determined from the scintillation
photon arrival times—reproduces the known source posi-
tions to within 2.0 cm; the reconstruction performance is
better than 1.0 cm for events occurring within 1 m of
the IB center. The energy scale was studied using γ rays
from 60Co, 68Ge, and 137Cs radioactive sources, γ rays from
the capture of spallation neutrons on protons and 12C, and
β þ γ-ray emissions from 214Bi, a daughter of 222Rn
(τ ¼ 5.5 day) that was introduced during the Xe-LS puri-
fication. The calibration data indicate that the reconstructed
energy varies by less than 1.0% throughout the Xe-LS
volume, and the time variation of the energy scale is
less than 1.0%. Uncertainties from the nonlinear energy
response due to scintillator quenching and Cherenkov light
production are constrained by the calibrations. The light
yield of the Xe-LS is 7% lower than that of the outer LS,
which is corrected in the detector simulation, while the
nonlinearities for both the LS regions are consistent.
The observed energy resolution is σ ∼ 7.3%=
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
EðMeVÞp ,
slightly worse relative to phase I due to an increased
number of dead PMTs.
We apply the following series of cuts to select ββ decay
events. (i) The reconstructed vertex must be within 2.0 m of
the detector center. (ii) Muons and events within 2 ms after
muons are rejected. (iii) 214Bi-214Po (τ ¼ 237 μs) decays
are eliminated by a delayed coincidence tag, requiring the
time and distance between the prompt 214Bi and delayed
214Po decay events to be less than 1.9 ms and 1.7 m,
respectively. The cut removes ð99.95  0.01Þ% of
214Bi-214Po decays, where the inefficiency is dominated
by the timing cut, and the uncertainty is estimated from
analysis of periods with high Rn levels. The same cut is not
effective for 212Bi-212Po (τ ¼ 0.4 μs) decays that occur
within a single ∼200-ns-long data acquisition event win-
dow. Therefore, the cut is augmented with a double-pulse
identification in the photon arrival time distribution after
subtracting the time of flight from the vertex to each PMT.
The 212Bi-212Po rejection efficiency is ð95  3Þ%, con-
firmed with high-Rn data. (iv) Reactor ν¯e’s identified by a
delayed coincidence of positrons and neutron-capture γ’s
[6] are discarded. (v) Poorly reconstructed events are
rejected. These events are tagged using a vertex-time-
charge discriminator that measures how well the observed
PMT time-charge distributions agree with those expected
based on the reconstructed vertex [7]. The total cut
inefficiency for ββ events is less than 0.1%.
Background sources external to the Xe-LS are domi-
nated by radioactive impurities on the IB film. Based on a
spectral fit to events reconstructed around the IB, we find
that the dominant background sources are 134Cs (β þ γ,
τ ¼ 2.97 yr) in the energy region 1.2 < E < 2.0 MeV
(2νββ window), and 214Bi in the region 2.3 < E <
2.7 MeV (0νββ window). The observed activity ratio of
134Cs to 137Cs (0.662 MeV γ, τ ¼ 43.4 yr) indicates that the
IB film was contaminated by fallout from the Fukushima-I
reactor accident in 2011 [8]. 214Bi is a daughter of 238U, a
naturally occurring contaminant. The observed rate of 214Bi
decays indicates that the 238U concentration in the nylon
film is 0.16 ppb assuming secular equilibrium, while
the ex situ measurement by ICP-MS yielded 2 ppt. The
nonuniform 214Bi event distribution observed on the IB
indicates that this discrepancy is caused by dust contami-
nation rather than decay chain nonequilibrium. Figure 1(a)
shows the vertex distribution of candidate events, and the
predicted 214Bi background events from a Monte Carlo
(MC) simulation in the 0νββ window. The z distribution
of 214Bi decays on the IB is evaluated from the data, and
used for the radioactive decay model in the MC simulation.
For the 214Bi background, the vertex dispersion model was
constructed from a full MC simulation based on Geant4
[9,10] including decay-particle tracking, scintillation
photon processes, and finite PMT timing resolution. This
MC simulation reproduces the observed vertex distance
PRL 117, 082503 (2016) P HY S I CA L R EV I EW LE T T ER S
week ending
19 AUGUST 2016
082503-2
between 214Bi and 214Po sequential decay events from the
initial 222Rn contamination within the Xe-LS.
An enlarged 3.5-m-radius spherical volume was used to
study a high statistics sample of muon spallation products
and better constrain their background contributions. This
included a region outside the IB. We assess a 22%
systematic uncertainty on the calculated spallation yields
in the Xe-LS, taking account of the observed ð20  2Þ%
increase in the spallation-neutron flux in the Xe-LS relative
to the outer LS. In the 0νββ window, events from 10C
decays (βþ, τ ¼ 27.8 s, Q ¼ 3.65 MeV) dominate the
contribution from muon spallation. A triple-coincidence
tag of a muon, a neutron identified by neutron-capture γ
rays, and the 10C decay [11], reduces the 10C background
with an efficiency of ð64  4Þ%. Post-muon spallation-
neutron events are recorded by newly introduced
dead-time free electronics. We apply spherical volume cuts
(ΔR < 1.6 m) around the reconstructed neutron vertices
for 180 s after the preceding muon. We estimate that the
remaining 10C background after cuts is ð1.01  0.26Þ×
10−2 ðton dayÞ−1, where ton is a unit of Xe-LS mass. Other
shorter-lived products, e.g., 6He and 12B, are also reduced
by the triple-coincidence tag and have a minor contribution
to the background. The dead time introduced by all the
spallation cuts is 7%. In the Xe-LS, long-lived 137Xe (β−,
τ ¼ 5.5 min, Q ¼ 4.17 MeV) is a background source
produced by neutron capture on 136Xe. Based on the
spallation-neutron rate and the 136Xe capture cross section
[12], the production yield of 137Xe is estimated to be
ð3.9  2.0Þ × 10−3 ðton dayÞ−1, which is consistent with
the simulation study in FLUKA [13,14].
We perform the 0νββ decay analysis using a 2-m-radius
fiducial volume (FV) as described above to utilize the
deployed 136Xe mass. However, the sensitivity is domi-
nated by the innermost 1-m-radius spherical volume due to
the background from the IB. The region outside this radius
serves to strongly constrain the tails of the IB background
extending into the innermost region. Further, anticipating
the decay of the 110mAg background identified in phase I,
we divide the phase-II data set into two equal time periods
(period 1 and period 2), each roughly equal to one average
lifetime of the 110mAg decay rate. Table I lists the number of
observed events, and the estimated and best-fit background
contributions in the 0νββ window within a 1-m-radius
spherical volume for each of the two time periods. The fit is
described in detail below. We find a precipitous decrease in
the event rate in the 0νββ window in period 2. The period-2
background components are well constrained near the
values listed in Table I with the exception of 110mAg.
The hypothesis of standard radioactive decay of the 0νββ
window background with the decay rate of 110mAg is
disfavored relative to the hypothesis of a faster reduction
at 96% C.L. The origin of this apparent reduction of 110mAg
is unknown, but we speculate that much of it settled to the
bottom of the IB where only a small fraction of 110mAg
decays is reconstructed in the inner Xe-LS volume. In order
to allow the 0νββ window background the greatest freedom
in the fit, the 0νββ decay analyses are performed
independently for period 1 and period 2.
The 2νββ decay rate can, in principle, be estimated from
the same analysis used to derive the 0νββ decay limits, but
the region outside of 1-m radius contributes negligibly to
the 2νββ decay rate estimate and is dominated by system-
atic uncertainty arising from the IB background. To obtain a
)2 (m2+Y2X
0 1 2 3 4
Z 
(m
)
2−
1−
0
1
2
B
i E
ve
nt
 R
at
e 
(E
ve
nts
/B
in)
21
4
Si
m
ul
at
ed
 
1
10
210
310
410
510
610
(a) 2.3 < E < 2.7 MeV
Visible Energy (MeV)
1 2 3 4
Ev
en
ts
/0
.0
5M
eV
210
310
410
Data
Xe 2νββ136
Spallation
IB/External
Total
Th232U+238
Po210Bi+210+
K40Kr+85+
(b)
3(R/1.54m)
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Ev
en
ts
/B
in
1
10
210
310 Data
νββ136
IB
Total
Spallation
Ag110m
(c)
FIG. 1. (a) Vertex distribution of candidate events (black points)
and reproduced 214Bi background events in a Monte Carlo (MC)
simulation (color histogram) for 2.3 < E < 2.7 MeV (the 0νββ
window). The normalization of the MC event histogram is
arbitrary. The solid and thick dashed lines indicate the shape
of the IB and the 1-m-radius spherical volume, respectively.
The thin dashed lines illustrate the shape of the equal-volume
spherical half shells that compose the 2-m-radius spherical
fiducial volume for the 0νββ analysis. (b) An example of the
energy spectrum in a volume bin with high 214Bi background
events around the lower part of the IB film [shaded region in
(a) at 1.47 < R < 1.53 m, z < 0]. (c) R3 vertex distribution
of candidate events in the 0νββ window. The curves show the
best-fit background model components.
PRL 117, 082503 (2016) P HY S I CA L R EV I EW LE T T ER S
week ending
19 AUGUST 2016
082503-3
2νββ decay rate free of such systematic uncertainty, we
perform a separate estimate using a likelihood fit to the
binned energy spectrum of the selected candidates between
0.5 and 4.8 MeV, limited to a volume within the 1-m-radius
spherical fiducial volume (FV2ν). The corresponding fidu-
cial exposure of 136Xe is 126 kg yr. The contributions from
major backgrounds in the Xe-LS, such as 85Kr, 40K, 210Bi,
and the 228Th-208Pb subchain of the 232Th series are free
parameters and are left unconstrained in the spectral fit.
We confirmed that the 134Cs contribution in the Xe-LS is
negligible from a fit. The contributions from the
222Rn-210Pb subchain of the 238U series, muon spallation
products, and detector energy response model parameters
are allowed to vary but are constrained by their independent
estimations. The 2νββ decay rates for period 1 and period 2
are 100.1þ1.1−1.8 ðtondayÞ−1 and 100.1þ1.0−0.9 ðtondayÞ−1, respec-
tively, and are in agreement within the statistical uncer-
tainties. The resolution tail in 2νββ decays is an important
background in the 0νββ analysis. Such tail events are
reproduced in 214Bi decays with high-Rn data assuming the
Gaussian resolution, indicating that a contribution from
energy reconstruction failures is negligible.
We assess the systematic uncertainty of the FV2ν cut
based on the study of uniformly distributed 214Bi events
from initial 222Rn contamination throughout the Xe-LS. We
obtain a 3.0% systematic error on FV2ν, consistent with the
1.0 cm radial vertex bias in the source calibration data.
Other sources of systematic uncertainty, such as xenon
mass (0.8%), detector energy scale (0.3%) and efficiency
(0.2%), and 136Xe enrichment (0.09%), only have a small
contribution; the overall uncertainty is 3.1%. The measured
2νββ decay half-life of 136Xe is T2ν1=2 ¼ 2.21  0.02ðstatÞ
0.07ðsystÞ × 1021 yr. This result is consistent with our
previous result based on phase-I data, T2ν1=2 ¼ 2.30 
0.02ðstatÞ  0.12ðsystÞ × 1021 yr [15], and with the result
obtained by EXO-200, T2ν1=2 ¼ 2.165  0.016ðstatÞ 
0.059ðsystÞ × 1021 yr [16].
For the 0νββ analysis, using the larger 2-m-radius FV, the
dominant 214Bi background on the IB is radially attenuated
but larger in the lower hemisphere. Sowe divide the FV into
20 equal-volume bins for each of the upper and lower
hemispheres [see Fig. 1(a)]. We perform a simultaneous fit
to the energy spectra for all volume bins. The z dependence
of 214Bi on the IB film is extracted from a fixed energy
window dominated by these events. The 214Bi background
contribution is then broken into two independent distribu-
tions in the upper and lower hemispheres whose normal-
izations are floated as free parameters. The fit reproduces the
energy spectra for each volume bin; Fig. 1(b) shows an
example of the energy spectrum in a volume bin with high
214Bi background events around the IB film. The radial
dependences of candidate events and best-fit background
contributions in the 0νββwindow are illustrated in Fig. 1(c).
The possible background contributions from 110mAg are
free parameters in the fit. We consider three independent
components: 110mAg uniformly dispersed in the Xe-LS
volume, and on the surfaces of each the lower and upper
IB films. We also examined nonuniform 110mAg sources,
with different assumed radial dependences, in the Xe-LS but
determined that this has little impact on the 0νββ limit.
As described above, the fits are performed independently
for period 1 and period 2 in the region 0.8 < E < 4.8 MeV.
We found no event excess over the background expec-
tation for both data sets. The 90% C.L. upper limits
on the 136Xe 0νββ decay rate are < 5.5 ðkton dayÞ−1 and
< 3.4 ðkton dayÞ−1 for period 1 and period 2, respectively.
TABLE I. Summary of the number of observed events, and the estimated and best-fit background contributions in the energy region
2.3 < E < 2.7 MeV (0νββ window) within the 1-m-radius spherical volume for each of the two time periods.
Period-1 Period-2
(270.7 days) (263.8 days)
Observed events 22 11
Background Estimated Best-fit Estimated Best-fit
136Xe 2νββ    5.48    5.29
Residual radioactivity in Xe-LS
214Bi (238U series) 0.23  0.04 0.25 0.028  0.005 0.03
208Tl (232Th series)    0.001    0.001
110mAg    8.5    0.0
External (Radioactivity in IB)
214Bi (238U series)    2.56    2.45
208Tl (232Th series)    0.02    0.03
110mAg    0.003    0.002
Spallation products
10C 2.7  0.7 3.3 2.6  0.7 2.8
6He 0.07  0.18 0.08 0.07  0.18 0.08
12B 0.15  0.04 0.16 0.14  0.04 0.15
137Xe 0.5  0.2 0.5 0.5  0.2 0.4
PRL 117, 082503 (2016) P HY S I CA L R EV I EW LE T T ER S
week ending
19 AUGUST 2016
082503-4
To demonstrate the low background levels achieved in the
0νββ region, Fig. 2 shows the energy spectra within a 1-m
radius, together with the best-fit background composition
and the 90% C.L. upper limit for 0νββ decays. Combining
the results, we obtain a 90% C.L. upper limit of
< 2.4 ðkton dayÞ−1, or T0ν1=2 > 9.2 × 1025 yr (90% C.L.).
We find that a fit including potential backgrounds from
88Y, 208Bi, and 60Co [3] does not change the obtained limit.
A MC of an ensemble of experiments assuming the best-fit
background spectrum without a 0νββ signal indicates a
sensitivity of 5.6 × 1025 yr, and the probability of obtaining
a limit stronger than the presented result is 12%. For
comparison, the sensitivity of an analysis in which
the 110mAg background rates in period 1 and period 2 are
constrained to the 110mAg half-life is 4.5 × 1025 yr.
Combining the phase-I and phase-II results, we
obtain T0ν1=2 > 1.07 × 10
26 yr (90% C.L.). This corresponds
to an almost sixfold improvement over the previous
KamLAND-Zen limit using only the phase-I data, owing
to a significant reduction of the 110mAg contaminant and the
increase in the exposure of 136Xe.
From the limit on the 136Xe 0νββ decay half-life, we
obtain a 90% C.L. upper limit of hmββi < ð61 − 165Þ meV
using an improved phase space factor calculation [17,18]
and commonly used NME calculations [19–25] assuming
the axial coupling constant gA ≃ 1.27. Figure 3 illustrates
the allowed range of hmββi as a function of the lightest
neutrino mass mlightest under the assumption that the decay
mechanism is dominated by exchange of a pure-Majorana
Standard Model neutrino. The shaded regions include the
uncertainties inUei and the neutrino mass splitting, for each
hierarchy. Also drawn are the experimental limits from
the 0νββ decay searches for each nucleus [2,26–28]. The
upper limit on < mββ > from KamLAND-Zen is the most
stringent, and it also provides the strongest constraint on
mlightest considering extreme cases of the combination of
CP phases and the uncertainties from neutrino oscillation
parameters [29,30]. We obtain a 90% C.L. upper limit
of mlightest < ð180–480Þ meV.
In conclusion, we have demonstrated effective back-
ground reduction in the Xe-loaded liquid scintillator by
purification, and enhanced the 0νββ decay search sensi-
tivity in KamLAND-Zen. Our search constrains the mass
1 2 3 4
Ev
en
ts
/0
.0
5M
eV
1−10
1
10
210
310
410 Data
Total
Total
νββ(0
νββ136
νββ136
Ag110m
Bi210Th+232U+238
K40Kr+85Po+210+
IB/External
Spallation
Ev
en
ts
/0
.0
5M
eV
1
2
3
4
2.4 2.6 2.8 3
Ev
en
ts
/0
.0
5M
eV
0
1
2
3
4
FIG. 2. (a) Energy spectrum of selected ββ candidates within a
1-m-radius spherical volume in period 2 drawn together with
best-fit backgrounds, the 2νββ decay spectrum, and the 90% C.L.
upper limit for 0νββ decay. [(b) and (c)] Close-up energy spectra
for 2.3 < E < 3.0 MeV in period 1 and period 2, respectively.
 (eV)lightestm
4−10 3−10 2−10 1−10
3−10
2−10
1−10
1
IH
NH
Xe)136KamLAND-Zen (
A
50 100 150
Ca
Ge
Se
Zr
Mo
Cd
Te
Te
Xe
Nd
 
(eV
)
m
FIG. 3. EffectiveMajorana neutrino mass hmββi as a function of
the lightest neutrino mass mlightest. The dark shaded regions are
the predictions based on best-fit values of neutrino oscillation
parameters for the normal hierarchy (NH) and the inverted
hierarchy (IH), and the light shaded regions indicate the 3σ
ranges calculated from the oscillation parameter uncertainties
[29,30]. The horizontal bands indicate 90% C.L. upper limits on
hmββi with 136Xe from KamLAND-Zen (this work), and with
other nuclei from Refs. [2,26–28], considering an improved
phase space factor calculation [17,18] and commonly used NME
calculations [19–25]. The side panel shows the corresponding
limits for each nucleus as a function of the mass number.
PRL 117, 082503 (2016) P HY S I CA L R EV I EW LE T T ER S
week ending
19 AUGUST 2016
082503-5
scale to lie below ∼100 meV, and the most advantageous
nuclear matrix element calculations indicate an effective
Majorana neutrino mass limit near the bottom of the
quasidegenerate neutrino mass region. The current
KamLAND-Zen search is limited by backgrounds from
214Bi, 110mAg, and muon spallation, and partially by the tail
of 2νββ decays. In order to improve the search sensitivity,
we plan to upgrade the KamLAND-Zen experiment with a
larger Xe-LS volume loaded with 800 kg of enriched Xe,
corresponding to a twofold increase in 136Xe, contained in a
larger balloon with lower radioactive background contam-
inants. If further radioactive background reduction is
achieved, the background will be dominated by muon
spallation, which can be further reduced by optimization
of the spallation cut criteria. Such an improved search
will allow hmββi to be probed below 50 meV, starting to
constrain the inverted mass hierarchy region under the
assumption that neutrinos are Majorana particles. The
sensitivity of the experiment can be pushed further by
improving the energy resolution to minimize the leakage of
the 2νββ tail into the 0νββ analysis window. Such improve-
ment is the target of a future detector upgrade.
The authors wish to acknowledge Professor A. Piepke
for providing radioactive sources for KamLAND. The
KamLAND-Zen experiment is supported by JSPS
KAKENHI Grants No. 21000001 and No. 26104002;
the World Premier International Research Center
Initiative (WPI Initiative), MEXT, Japan; Stichting FOM
in the Netherlands; and under the U.S. Department of
Energy (DOE) Contract No. DE-AC02-05CH11231, as
well as other DOE and NSF grants to individual institu-
tions. The Kamioka Mining and Smelting Company has
provided service for activities in the mine. We acknowledge
the support of NII for SINET4.
[1] S. Dell’Oro, S. Marcocci, M. Viel, and F. Vissani, Adv. High
Energy Phys. 2016, 2162659 (2016).
[2] M. Agostini et al. (GERDA Collaboration), Phys. Rev. Lett.
111, 122503 (2013).
[3] A. Gando et al. (KamLAND-Zen Collaboration), Phys. Rev.
Lett. 110, 062502 (2013).
[4] J. Albert et al. (EXO Collaboration), Nature (London) 510,
229 (2014).
[5] A. Gando et al. (KamLAND Collaboration), Phys. Rev. C
92, 055808 (2015).
[6] A. Gando et al. (KamLAND Collaboration), Phys. Rev. D
88, 033001 (2013).
[7] S. Abe et al. (KamLAND Collaboration), Phys. Rev. C 84,
035804 (2011).
[8] A. Gando et al. (KamLAND-Zen Collaboration), Phys. Rev.
C 85, 045504 (2012).
[9] S. Agostinelli et al., Nucl. Instrum. Methods Phys. Res.,
Sect. A 506, 250 (2003).
[10] J. Allison et al., IEEE Trans. Nucl. Sci. 53, 270 (2006).
[11] S. Abe et al. (KamLAND Collaboration), Phys. Rev. C 81,
025807 (2010).
[12] J. B. Albert et al., arXiv:1605.05794v1.
[13] G. Battistoni, F. Cerutti, A. Fassò, A. Ferrari, S. Muraro,
J. Ranft, S. Roesler, and P. R. Sala, AIP Conf. Proc. 896, 31
(2007).
[14] A. Ferrari, P. R. Sala, A. Fassó, and J. Ranft, FLUKA:
A Multiparticle Transport Code (CERN, Geneva, 2005).
[15] A. Gando et al. (KamLAND-Zen Collaboration), Phys. Rev.
C 86, 021601 (2012).
[16] J. B. Albert et al. (EXO Collaboration), Phys. Rev. C 89,
015502 (2014).
[17] J. Kotila and F. Iachello, Phys. Rev. C 85, 034316 (2012).
[18] S. Stoica and M. Mirea, Phys. Rev. C 88, 037303 (2013);
updated in arXiv:1411.5506v3.
[19] T. R. Rodríguez and G. Martínez-Pinedo, Phys. Rev. Lett.
105, 252503 (2010).
[20] J. Menéndez, A. Poves, E. Caurier, and F. Nowacki, Nucl.
Phys. A818, 139 (2009).
[21] J. Barea, J. Kotila, and F. Iachello, Phys. Rev. C 91, 034304
(2015).
[22] J. Hyvärinen and J. Suhonen, Phys. Rev. C 91, 024613
(2015).
[23] A. Meroni, S. T. Petcov, and F. Šimkovic, J. High Energy
Phys. 02 (2013) 025.
[24] F. Šimkovic, V. Rodin, A. Faessler, and P. Vogel, Phys. Rev.
C 87, 045501 (2013).
[25] M. T. Mustonen and J. Engel, Phys. Rev. C 87, 064302
(2013).
[26] K. Alfonso et al. (CUORE Collaboration), Phys. Rev. Lett.
115, 102502 (2015).
[27] R. Arnold et al. (NEMO-3 Collaboration), Phys. Rev. D 92,
072011 (2015).
[28] A. S. Barabash, Phys. At. Nucl. 74, 603 (2011).
[29] S. Dell’Oro, S. Marcocci, and F. Vissani, Phys. Rev. D 90,
033005 (2014).
[30] F. Capozzi, G. L. Fogli, E. Lisi, A. Marrone, D. Montanino,
and A. Palazzo, Phys. Rev. D 89, 093018 (2014).
PRL 117, 082503 (2016) P HY S I CA L R EV I EW LE T T ER S
week ending
19 AUGUST 2016
082503-6