Fei Gao* for the XENON collaboration
Search for Event Rate Modulation in XENON100 Electronic Recoil Data
1
University of California, San Diego
Oct 29th, 2015, Kashiwa, Japan
* affiliated with Shanghai Jiao Tong University
XENON Collaboration
2
Columbia Nikhef Mainz Muenster
MPIK
Purdue Coimbra Subatech Bologna Weizmann
UCLA
Rice
Bern
Zurich
LNGS Torino
RPI
NYUAD
Stockholm
Chicago
UCSD
~130 Scientists from 21 institutes
Annual modulation and DAMA/LIBRADark matter (DM) signal rate is expected to be annually modulating
A key feature to distinguish signals from overwhelming backgrounds
DAMA/LIBRA:
3
peak phase 152 days (June 1)
9.3 sigma significance
Phase (144 +- 7) days
only for single hit
No signal above 6 keV
Seems to be a convincing evidence, HOWEVER…
Freese et al., Rev. Mod. Phys. 85, 1561 (2013)
Bernabei et al., Eur. Phys. J. C 73, 12 (2013)
Nuclear Recoil Interpretation
4
Nuclear recoil interpretations of DAMA/LIBRA modulation have been challenged by several more sensitive experiments with background rejection power
How about Leptophilic DM?❖ DAMA/LIBRA annual modulation can be interpreted
as signals from Leptophilic DM models
❖ We tested three representative models in XENON100 using the electronic recoil data:
❖ 1, DM-electron scattering through axial-vector coupling
❖ 2, Mirror DM model
❖ 3, Luminous DM model
5
Light Response in XENON100
6
light response is determined with low energy measurements interpolated by NEST v0.98, uncertainties are from NEST and spread of measurements
DAMA/LIBRA 2-6 keV Electronic recoil (ER) corresponds to 3-14 PE in XENON100
Baudis et al., Phys. Rev. D 87, 115015 (2013) Aprile et al., Phys. Rev. D 90, 062009 (2014)
Szydagis et al., J. Instrum. 6, P10002 (2011)
❖ DAMA/LIBRA rate converted to XENON100 spectrum assuming leptophilic DM model, axial vector coupling
❖ Energy response, resolution and cut acceptance applied
❖ Compare XENON100 average rate with DAMA/LIBRA modulation amplitude
❖ Constraints on DM interpretation of DAMA/LIBRA (assuming 100% modulation):
❖ WIMPs-electron scattering 4.4-sigma
❖ Mirror dark matter model 3.6-sigma
❖ Luminous dark matter model 4.6-sigma
DAMA/LIBRA Comparison
S1 (PE)0 5 10 15 20 25 30
Rat
e (c
ount
s/PE
/tonn
e/da
y)
0
2
4
6
8
10
12
Mean electronic recoil energy (keV)1 2 3 4 5 6 7 8 9 10
DAMA/LIBRA annually modulated spectrum (2-6)keV interpreted asleptophilic DM, axial-vector couplingmirror DMXENON100 70 summer live days
7
S1 [PE]5 10 15 20 25 30
Cut A
ccep
tanc
e
00.10.20.30.40.50.60.70.80.9
1
2.0 - 5.8 keV 5.8 - 10.4 keV
Energy [keV]2 4 6 8 10
Science 349, 851 (2015)
Energy [keV]0 2 4 6 8 10 12
Rate
[cou
nts/k
eV/to
n/da
y]
10−
0
10
20
30
40
DAMA/LIBRA Modulation
Search for Modulations❖ The first LXe TPC with more than one year of stable
running conditions
❖ The first modulation search for DM at Gran Sasso Lab after DAMA/LIBRA
❖ Demonstration for future XENON modulation searches
❖ Search for leptophilic DM signals
❖ Require good understand the stability of detector and backgrounds
8
Stability of the Detector
Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
C]oT
[∆
-0.15-0.1
-0.050
0.050.1
0.150.2
0.250.3
C oT(Room) avg:20.7∆×-110C oT(Gas Xenon) avg:-86.5∆
C oT(Liquid Xenon) avg:-91.6∆
(b)
Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
Liqu
id L
evel
[mm
]
2.3
2.4
2.5
2.6
2.7
2.8Level Meter 1Level Meter 2
(c)
Time Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
Rate
[Cou
nts/D
ay]
0
0.5
1
1.5
2
2.5 (f)
(51.1%)σ1±
Background rate (SS) 2.0 - 5.8 keV
9
PRL 115, 091302 (2015)
❖ Detector pressor (2)
❖ Room pressor
❖ LXe temperature (4)
❖ PTR temperature
❖ Room temperature
❖ Purification flow rate
❖ LXe levels (2)
❖ PMT gain
❖ Radon level (2)
Very tiny absolute variations
No correlations with ER rate No significant impact on ER rate!
Aprile et al., Astropart. Phys., 35, 573-590 (2012)
❖ Stability of cut acceptance is derived from weekly ER calibrations sources
❖ The acceptance variation further accounts for the variation of the detector parameters like LXe level.
❖ The dips of acceptance are due to increment of noise level.
❖ The fluctuation of acceptance is taken into account for the event rate modulation analysis.
Stability of Cut Acceptance
Time Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
Rate
[Cou
nts/D
ay]
0
0.5
1
1.5
2
2.5 (f)
(51.1%)σ1±
Background rate (SS) 2.0 - 5.8 keV
Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
Cut A
ccep
tanc
e
0.60.65
0.70.75
0.80.85
0.90.95
1 (e)
(3.8%)σ1±
Average Cut Acceptance2.0 - 5.8 keV
10
PRL 115, 091302 (2015)
❖ Co60 (T1/2 = 5.3 year) gamma background is time dependent, but the absolute contribution is negligible.
❖ Radon and krypton background concentration are time dependent due to tiny air leak
❖ Radon contributes to the overall background by less than 20%. Hence the absolute contribution to fluctuation is negligible.
❖ No correlation between radon and ER rate.
❖ Krypton concentration varies in time due to air leak. The size of its variation is taken into account.
Stability of Backgrounds
Time Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
Rate
[Cou
nts/D
ay]
0
0.5
1
1.5
2
2.5 (f)
(51.1%)σ1±
Background rate (SS) 2.0 - 5.8 keV
Mar/11 May/11 Jul/11 Sep/11 Nov/11 Jan/12 Mar/12
Bq/k
g]µ
Rado
n Le
vel [
50
60
70
80
90
100 (d)
(8.6%)σ1±
Radon inside LXe
11
Profile Likelihood Analysis
f(t) = ✏(t)
✓C +Kt+A cos
✓2⇡
(t� �)
P
◆◆
L =
nY
i=1
˜f(ti)
!Poiss (n|N
exp
(E))L✏LKLE
12
Cut acceptance constant background
krypton variation due to air leak
modulation
observed number of events
energy response
Constraints
We performed an unbinned profile likelihood analysis to search for modulation signal
Event Rate:
Discovery Potential
Period [Days]8 10 20 30 40 100 200 300 10000
2468
101214161820
P = 100P = 365P = 500
(local)σ2
(local)σ3
0.0010.0020.010.020.1Frequency [1/Days]
-2log(L 0
/L1)
13
Simulated modulation signals
A=2.7 events/(keV · tonne · day) ~ best fit value for P=365.25 days
Average significance of 3 sigma assuming ~25% DAMA/LIBRA modulation
Characteristic plateau
no rise at large periodgood resolution on period
no resolution on period
PRL 115, 091302 (2015)
The data is only sensitive to modulation with period < 500 days
❖ No evident peak crossing the 1-sigma global significance threshold!
❖ SS in the Low-E (2.0-5.8 keV) range shows increasing significance at long period region. 2.8-sigma local significance at one year period
❖ MS background only control sample in Low-E range shows similar power spectrum as SS. This disfavors an WIMPs interpretation of the SS spectrum
❖ SS in high-E (5.8-10.4 keV) does not show high significance at long period region
Modulation Search Results
Period [Days]7 8 910 20 30 40 50 100 200 300 400
2468
10121416
(local)σ2
(local)σ3
(global)σ1
1 ye
ar
Frequency [1/Days]
SS, 2.0 - 5.8 keV
0.0020.010.020.030.1
Period [Days]
7 8 9 10 20 30 40 50 60 100 200 300 400
2468
10121416
(local)σ2
(local)σ3
(global)σ1MS, 2.0 - 5.8 keV0.0020.0030.010.020.030.040.1
Period [Days]7 8 910 20 30 40 50 100 200 300 400
2468
10121416
(local)σ2
(local)σ3
(global)σ1SS, 5.8 - 10.4 keV0.0020.010.020.030.1
-2log(L 0
/L1)
-2log(L 0
/L1)
-2log(L 0
/L1)
14
PRL 115, 091302 (2015)
Signal, Low-E
Background, Low-E
Side band, high-E
❖ The phase (112+-15) days (April 22) is not consistent with the standard halo model (June 2) at 2.6-sigma
❖ The amplitude of is also too small (only ~25%) compared with the expected DAMA/LIBRA modulation signal in XENON100.
❖ The DM interpretation of DAMA/LIBRA annual modulation as being due to WIMPs electron scattering through axial vector coupling is disfavored at 4.8-sigma from a PL analysis
DAMA/LIBRA Comparison (2D)
Phase [Days]20 40 60 80 100 120 140 160 180 200A
mpl
itude
[eve
nts/
(keV
·tonn
e·da
y)]
0
2
4
6
8
10
12 Expected
expectedphase fromDM halo
XENON10095% C.L.99.73% C.L.best fit
20 40 60 80 100 120 140 160 180 200
24681012
σ1
σ2
σ3
2 4 6 810120
2
4
6
8
10
12
σ1 σ2 σ3
-2lo
g(L 1/L
max
)
-2log(L1/Lmax)
DAMA/LIBRA
C = (5.5± 0.6) events/(keV · tonne · day),
A = (2.7± 0.8) events/(keV · tonne · day),
� = (112± 15) days
15
PRL 115, 091302 (2015)
DAMA/LIBRA(expected) = 11.5± 1.2stat± 0.7syst
Summary❖ The first stable LXe TPC sufficient for modulation searches.
❖ No significant modulation is found in the XENON100 electronic recoil data.
❖ The increasing significances at long period in both SS and MS samples does not favor a dark matter interpretation
❖ Leptophilic DM models to interpret DAMA/LIBRA modulation have been challenged by XENON100
❖ WIMPs-electron scattering 4.4-sigma
❖ Mirror dark matter model 3.6-sigma
❖ Luminous dark matter model 4.6-sigma
❖ More data is ready for modulation searches.
16
PRL 115, 091302 (2015)
Science 349, 851 (2015)
XENON100 ER background
ER background:Beta decay from krypton
Beta decay from radon
Gammas from materials
Very good data/MC absolute matching inside fiducial volume!!!
Radioactivity from screening values, no turning!
Need XENON1T to suppressPMT background
17
E. Aprile et al. (XENON100), Phys. Rev. D83, 082001 (2011)
WIMP-electron scattering Requirement: no loop-induced nuclear recoil - axial vector interaction
Advantage: natural model with coupling to electrons
Disadvantage: bad spectrum match with DAMA/LIBRA
Kopp et.al, PRD 80, 083502 (2009)18
Mirror electron scattering❖ Multi-component mirror models
❖ DM halos are composed of a multi-component plasma of mirror particles (same mass as their partners)
❖ Mirror electron scatters on electron through kinematically mixed coupling
❖ Scatter rate proportional to number of loosely bound electrons (binding energy < 1 eV)
❖ Constant scaling of 0.89 between XENON100 and DAMA/LIBRA
19
. R. Foot, Int.J.Mod.Phys. A29, 1430013 (2014)
Luminous Dark Matter
❖ Upper scattering inelastic dark matter scattering, the interaction rate is determined by dipole moment
❖ ~keV mass splitting produce X-rays
❖ Interaction in the earth besides the detector, and produce X-rays inside the detector
❖ 3.0 keV mass splitting fits well with the DAMA/LIBRA modulation
20
. B. Feldstein et.al, Phys.Rev. D82, 075019 (2010)
Radon Correlation Analysis
21
M. Weber, PhD Dissertation, University of Heidelberg (2013) www.ub.uni-heidelberg.de/archiv/15155
Calibration between radon and krypton
22
Leak rate is calculated fromthe correlation analysis between external and internal radon
three RGMS measurements of kryptonacross the run
Perfect linear correlation betweenkrypton levels and air leak