Backgroundandsensitivity predictionsforXENON1T › dm16 › talks › selvi.pdf · 2016-02-19 ·...
Transcript of Backgroundandsensitivity predictionsforXENON1T › dm16 › talks › selvi.pdf · 2016-02-19 ·...
Marco Selvi INFN -‐ Sezione di Bologna
(on behalf of the XENON collaboration)
Background and sensitivity predictions for XENON1T
1
Feb 19th 2016, UCLA Dark Matter 2016
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Outline
2
• Description of the detector;
• Prediction of the ER and NR backgrounds through a GEANT4 simulation of the experiment;�
• Conversion of the energy released in the TPC by signal and background events into the S1 and S2 signals seen in the detector;�
• Study the sensitivity to WIMP-nucleus spin-independent interactions.�
E. Aprile et al. (the XENON collaboration)“Physics reach of the XENON1T dark matter experiment”,arXiv:1512.07501, submitted to JCAP
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
The XENON1T experiment
3
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
The XENON1T experiment
4
96 cm
97 cm χχ
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Main Backgrounds
Nuclear Recoils (NR): • Radiogenic neutrons: spontaneous fission and (α, n) reaction from the U and Th
chains in the detector components.• Muon-induced neutrons.• Coherent scattering of neutrinos (mostly solar) off the Xe nuclei.
Electron recoils (ER): • low energy Compton
scatters from the radioactive contaminants in the detector components: U and Th chains, 40K, 60Co, 137Cs.
• Intrinsic contaminants: β decays of 222Rn daughters, 85Kr, 136Xe.
• Elastic scattering of solar neutrinos off electrons.
NR
ER
5
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
MC simulation in GEANT4
6
Diving bell
PTFE pillars
Supports
Field shaping
electrodes
PMTs
PMT bases
PTFE reflector
Copper plate
PTFE bottom plateCathode
Top PMT array
Bottom PMT array
A single PMT
The whole TPC
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
ER from the materials
Electronic Recoil Energy [keV]0 500 1000 1500 2000 2500
] -1
keV
)⋅
day
⋅
ER R
ate
[ ( k
g
8−10
7−10
6−10
5−10
4−10
3−10
2−10
1−10
Pb214
Bi214 Bi214
Bi214
Bi214
Bi214Bi214
Bi214
Bi214 Bi214
Bi214
Bi214
Tl208
Tl208 Tl208Bi212
Ac228 Ac228
Ac228 Tl208Cs137
Co60 Co60
K40
Total Cryostat Shells
Cryostat Flanges SS in the TPC
PMT and Bases PTFE and Cu in the TPC
]2 [mm2R0 20 40 60 80 100 120 140 160 180 200 220
310×
Z [m
m]
-400
-300
-200
-100
0
100
200
300
400 ] -1
keV
)⋅
day
⋅
Rate
[ ( k
g
-610
-510
-410
-310
-210
-110
ER background �from the materials�
in 1 t Fiducial Volume, �in [1, 12] keV:�
7.3 10-6 (kg day keV)-1
7
Extensive screening campaign using gamma (Ge ) and mass (ICP-MS) spectrometry.�Eur.Phys.J. C75 (2015) 546
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Total ER background
8
Assumptions on the intrinsic backgrounds:�0.2 ppt of natKr (already achieved in XENON1T distillation column tests),10 µBq/kg of 222Rn (conservative estimation based on Rn emanation meas.).
Fiducial Mass [kg]200 400 600 800 1000 1200 1400 1600 1800
] -1
keV
)⋅
day
⋅
ER R
ate
[ (kg
7−10
6−10
5−10
4−10
3−10Total Xe136 Total except materials
Rn222 Kr85 materials
νsolar
Electronic Recoil Energy [keV]0 500 1000 1500 2000 2500
] -1
keV
)⋅
day
⋅
ER R
ate
[ ( k
g
8−10
7−10
6−10
5−10
4−10
3−10
2−10
1−10
Total Rn222
materials Kr85
Xe136 νsolar
222Rn (mainly from 214Pb β-decay) is the most relevant source of ER background in most of the TPC.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Total ER background
9
• The numbers above are before any ER/NR discrimination.• In 1 t FV, the background from the material is at the same level as the
one from solar neutrinos and from Kr
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
NR from radiogenic neutrons
10]2 [mm2R
0 20 40 60 80 100 120 140 160 180 200 220310×
Z [m
m]
-400
-300
-200
-100
0
100
200
300
400 ]-1
keV
)⋅
day
⋅
NR
Rate
[ ( k
g
-910
-810
-710
-610
-510
Fiducial Mass [kg]200 400 600 800 1000 1200 1400 1600 1800
]-1
NR
Rate
[ y
-210
-110
1
[3, 50] keV[4, 50] keV[5, 50] keV
Neutron Energy [MeV]0 2 4 6 8 10
]-1
Neu
tron
Yiel
d pe
r Par
ent D
ecay
[MeV
-1610
-1510
-1410
-1310
-1210
-1110
-1010
-910
-810
-710
-610
-510Th230U - 238
Pb207U - 235
Pb206Ra - 226
Ac228Th - 232
Pb208Th - 228
In 1t FV, [4, 50] keV: �0.6 events / y �
(with no discrimination)
Neutron yield from �SOURCES-4C
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Nuclear Recoil Energy [keV]0 10 20 30 40 50 60 70 80 90 100
]-1
keV
)⋅
day
⋅
NR
Rate
[(kg
-1110
-1010
-910
-810
-710
-610
-510
-410
-310
-210
CNNSRadiogenic neutronsMuon-induced neutrons
Total NR background
11
Source Background (ev/y)
Radiogenic neutrons 0.6 ± 20%
Muon-‐induced neutrons
< 0.01 (muon veto ON)
CNNS 0.02 ± 20%
Total NR 0.62 ± 0.12
Single ScaBer, 1 t Fiducial Volume, [4, 50] keVr, 100% NR acceptance
Given the very steep spectrum of NR from CNNS, its contribuQon will become more relevant aSer the conversion into the S1, S2 signals,
considering the detector response and energy resoluQon.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Light & charge emission model
12
ER: NEST model� M. Szydagis et al., JINST 6 (2011) P10002, [arXiv:1106.1613]
NR: direct measurements of Leff,�Qy measured by XENON100�E. Aprile et al., Phys.Rev.D88 (2013) 012006, [arXiv:1304.1427]
NEST is based at low energies on the direct �measurement at Columbia and Zurich:�Phys. Rev. D 86, 112004 (2012) �Phys. Rev. D 87, 115015 (2013) ��
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
@122 keV
Light Collection EfGiciency
13
MeshMeshMeshWiresMesh
Ly at 500 V/cm | 4.6 PE/keV
]2 [mm2R0 20 40 60 80 100 120 140 160 180 200 220
310×
Z [m
m]
400−
300−
200−
100−
0
100
200
300
400
LCE
[%]
26
28
30
32
34
36
38
40
42
44
Absorption length [m]10 20 30 40 50 60 70 80 90 100
LCE
[%]
20
25
30
35
40
PTFE with 99% reflectivityPTFE with 95% reflectivityPTFE with 90% reflectivityOur assumption
Ly [P
E/ke
V]
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
We generate optical photons uniformly inside the TPC, and follow their propagation inside it with all the reflections, refractions, absorptions, etc.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Signal and backgrounds in S1
14
S1 [PE]0 10 20 30 40 50 60 70 80
]-1
PE)
⋅ d
ay
⋅Ev
ent R
ate
[ (k
g
-1110
-1010
-910
-810
-710
-610
-510
-410Total (ER+NR) backgroundER backgroundNR background from neutrons
νNR background from 2 cm-46 = 2 10σ, 2WIMP, m = 10 GeV/c
2 cm-47 = 2 10σ, 2WIMP, m = 100 GeV/c2 cm-46 = 2 10σ, 2WIMP, m = 1000 GeV/c
- Electric field in the TPC: 500 V/cm- Lower threshold of the region of interest: �
3 PE (lowest one used so far by XENON100)
- Higher end: 70 PE, corresponding on average to 50 keVr �(include >90% of NR spectrum for 100 GeV WIMP)
- S2 > 8 electrons
ER background is the most relevant one, �apart from the low S1 region where CNNS becomes dominant.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Statistical model
15
• 2D statistical model: �S1 and discrimination.�
• Profile Likelihood method, �exclusion test statistic, with CLs�(Eur. Phys. J. C71 (2011) 1554).�
• Systematic uncertainties are included in the model �as nuisance parameters, and profiled out.���
• The most relevant uncertainty is related to Leff, which rules the light output of nuclear recoils in LXe. Thus it affects, in a correlated way, both the WIMP signal and the NR backgrounds.�We consider also other systematics: Qy, and the overall uncertainty on ER and NR background (assumed 10% and 20%, respectively).�
• We study the sensitivity bands via generation of MC toy experiments.
§ ER Background events! NR Background events! (NR) Signal events!
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
XENON1T sensitivity
16
] 2WIMP mass [GeV/c6 10 20 30 100 200 1000 2000 10000
]2W
IMP-
nucl
eon
Cros
s Sec
tion
[cm
49−10
48−10
47−10
46−10
45−10
44−10
43−10
42−10
41−10
40−10
39−10 DAMA/Na
DAMA/ICDMS-Si (2013)
XENON10 (2013)
SuperCDMS (2014) PandaX (2014)DarkSide-50 (2015)
XENON100 (2012) LUX (2013)
y)⋅XENON1T (2 t
y)⋅XENONnT (20 t
Billard et al., Neutrino Discovery limit (2013)
y⋅XENON1T Sensitivity in 2 tExpected limit (90% CL)
expectedσ 2 ± expectedσ 1 ±
With a 2 t y exposure, with XENON1T we’ll reach a sensitivity to spin-independent WIMP-nucleon interactions of 1.6 10-47 cm2 for a 50 GeV/c2 WIMP.
Dotted blue line shows the XENON1T sensitivity assuming a cutoff below 3 keV
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Sensitivity VS time
17
Livetime [days]1 10 210 310
]2 [
cmσ
spin
-inde
pend
ent W
IMP-
nucl
eon
-4810
-4710
-4610
-4510
-4410
XENON100 (2012)
LUX (2013)
XENON1T (2 year exposure)
LUX (2013)
LUX (2015)
In less than 10 days we can reach the sensitivity �of the currently running experiments
mχ = 50 GeV/c2
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
LUX2015 emission model
18
Nuclear Recoil Energy [keV]1 10 210
Phot
on Y
ield
[ph/
keV
]
0
2
4
6
8
10
12
14
modelXENON100
modelLUX2015
Nuclear Recoil Energy [keV]1 10 210
/keV
]-
Elec
tron
Yie
ld [e
0
2
4
6
8
10
12 modelXENON100
modelLUX2015
With the recent LUX measurement (arXiv:1512.03506):• Larger photon and electron yield, in particular at low energy (below 3
keV),• Much smaller systematic uncertainties.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
XENON1T sensitivity
19
Assuming the LUX2015 emission model
Potential to detect CNNS �from solar neutrinos.
Significant improvement in sensitivity to WIMPs�
at low masses, below 10 GeV/c2.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Summary & Conclusions
20
• ER and NR backgrounds from the detector materials are reduced to a subdominant level in the inner 1 t FV.�
• The most relevant background comes from 222Rn in LXe, (assuming a uniform contamination of 10 µBq/kg). Total ER background: 1.8 10-4 (kg day keV)-1.
• With a 2 t y exposure, with XENON1T we’ll reach a sensitivity to spin-independent WIMP-nucleon interaction of 1.6 10-47 cm2 for a 50 GeV/c2 WIMP.�
• Good potential to detect for the first time coherent scattering of neutrinos from the Sun, in particular assuming the LUX2015 emission model. Improved sensitivity at low mass WIMPs too.
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Backups
21
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
MC simulation in GEANT4
Diving bell
PTFE pillars
Supports
Field shaping
electrodes
PMTs
PMT bases
PTFE reflector
Copper plate
PTFE bottom plateCathode
22
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
ER from the materials Extensive screening campaign using Ge gamma spectrometry
and mass spectrometry (ICP-MS)
Screening of PMTs: E. Aprile et al., Eur.Phys.J. C75 (2015) 546, [arXiv:1503.07698]Dedicated paper on screening of XENON1T materials in preparation
23
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
NR from µ-‐induced n
24
Negligible, thanks to the water shield and Cherenkov muon veto surrounding the detector
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
NR from CNNS
25
Nuclear Recoil Energy [keV]1 10
]-1
keV
)⋅
y
⋅Ra
te [(
t
5−10
4−10
3−10
2−10
1−10
1
10
210 TotalB8
hepDSNAtm
Total rate above 4 keV: 1.8 10-2 (t y)-1
Total rate above 1 keV: 90 (t y)-1
(Coherent neutrino-�nucleus scattering)
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
WIMP signal
26
Halo model: v0 = 220 km/s, vSun = 250 km/s, �vesc = 544 km/s, ρ=0.3 GeV/cm3
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
2d statistical model
27
ER recoils NR recoils
Total
§ 2 t∙y exposure !§ mχ =100 GeV/c2!§ μs=50 (enhanced)!§ μbER=1300"§ μbNR=2.3"
§ ER Background events! NR Background events! (NR) Signal events!
S1 distributionsRecoil energy spectra converted into S1 distributions(n° PE detected) through• Light Yield LY (n° γ produced/keV)• Charge Yield QY (n° e- produced/keV)• Detector performance (LCE, PMTs QE and CE)
Y distributionsY is an idealized version of the discrimination parameter log10(S2/S1)§ ER events Gauss(0,1) § NR events Gauss(-2.58,0.92)
These distributions reproduce the XENON100 ER/NR discrimination and NR acceptance performance
NO discrimination cut on ER events (full data set used)
S1 [PE]0 10 20 30 40 50 60 70 80
]-1
PE)
⋅ d
ay
⋅Ev
ent R
ate
[ (k
g
-1110
-1010
-910
-810
-710
-610
-510
-410Total (ER+NR) backgroundER backgroundNR background from neutrons
νNR background from 2 cm-46 = 2 10σ, 2WIMP, m = 10 GeV/c
2 cm-47 = 2 10σ, 2WIMP, m = 100 GeV/c2 cm-46 = 2 10σ, 2WIMP, m = 1000 GeV/c
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
Systematic uncertainty on Leff
28
The number of expected events from CNNS and low mass WIMPs is highly affected by Leff variations
μs= μs(t) and μbNR= μbNR(t)""
!
§ The RELATIVE SCINTILLATION EFFICIENCY Leff rules the light output of nuclear recoils in Xenon
§ Leff is the major systematic uncertainty for XENON1T; no direct experimental measurements below 3 keV.
Parameterization of the uncertainty on Leff Gaussian nuisance parameter t
"
S1 SPECTRAL SHAPES of signal and backgrounds change very slightly under different Leff ! We keep fixed the
shape of S1 spectra"!
mχ=100 GeV/c2"
IMPACT OF Leff VARIATIONS
NUMBER of EXPECTED EVENTS in [3,70] PE
CNNS neutrinos behave like 6 GeV/c2 WIMPNeutrons behave like 30 GeV/c2 WIMP
Conservatively extrapolated down
to zero at 1 keV
M. Selvi – UCLA DM 2016 – XENON1T backgrounds & sensitivity
ProGile likelihood method
29
μs =10"
Modified p-value CLs method!
μs =5"
Likelihood function(unbinned, extended)
Parameter of interest μs !
Exclusion test statisticProfile Likelihood ratio
Compute test statistic P.D.F.under signal hypothesis H!under bkg-only hypothesis H 0
Run hypotheses tests"§ Generated 104 MC toy experiments under H0 § Rejection test of each signal hypothesis H! (different !s) using med[qμ|Hμ] as observed test statistic qμ
obs"
§ The significance of each test is given by the p-value
The uncertainty on Leff , affecting both WIMP signal and NR background in a correlated way, is included in the likelihood as a nuisance parameter and profiled out.