Background Free Search for 0 Decay of 76Ge with GERDA

Transcription

1 Background Free Search for 0 Decay of 76Ge with GERDA Victoria Wagner for the GERDA collaboration Max-Planck-Institut für Kernphysik Rencontres de Moriond, Electro Weak La Thuile, March

2 The GERDA Collaboration: searching for neutrinoless double beta decay of 76Ge 2

3 Signature & Experimental Challenges Measure sum energy of electrons T Observable: (T 2ν 1/ 2 0 ν 1 1/ 2 zero background regime ) N0 ν T = yr M t background, i.e. statistical fluctuation limited scenario T 21 0ν 1/ 2 0ν 1/ 2 M t Δ E BI M t: exposure [kg yr], ΔE: energy resolution, BI: background index [counts/(kev kg yr)] Need to achieve < 1 bck event in ROI excellent energy resolution 3

4 Germanium Detectors Measure sum energy of electrons -ν -ν 2νββ e---- e- 2ν e h+ 21 T 1/ 2= yr High Purity Germanium (HPGe) Detectors 3-4 kev FWHM at Q = 2039 kev (0.2%) HPGe detectors isotopically enriched in Ge (~87%) 76 h e e- e- 0νββ high detection efficiency of ββ: source = detector no intrinsic background [Astropart.Phys. 91 (2017) 15-21] discrimination of signal- from background like events using pulse shape analysis 4

5 1.4 km of rock 3500 m w.e. LNGS cosmic muon flux reduced by a factor ~ μ/m2/h 5

6 The Germanium Detector Array concept: operate bare HPGe detectors in LAr which serves as coolant & (active) shielding GERDA Phase I (Nov May 2013) 17.8 kg enriched semi-coaxial kg enriched BEGe exposure 21.6 kg y BI ~ 10-2 cts/(kev kg yr) T1/2 > yr (90% C.L.) 0ν m3 pure water muon veto & neutron absorber HPGe detectors PRL 111, (2013) (Dec ) 30 enriched BEGe (= 20.0 kg) + 7 enriched semi-coaxial (= 15.6 kg) LAr instrumentation goal: BI ~ 10-3 cts/(kev kg yr) Eur. Phys. J. C (2013)73:2330 6

7 Array wire bonding for contacting ~8cm low radioactivity electronics ~8cm new low mass holders with reduced mass and Cu Si 7

8 Discriminating Signal from Background Events ββ event ν ν local energy deposition (SSE) in single detector 2νββ background event γ energy deposition in multiple locations (MSE) in single detector γ γ or on detector surface ( ) pulse shape discrimination coincident energy deposition in β/ α more than one detector detector anti-coincindence additional energy deposition in LAr γ γ 0νββ LAr veto 8

9 LAr Instrumentation Hybrid Design 810 wavelength shifting fibers coupled to SiPMs 100cm 60cm 16 photomultiplier tubes (PMTs) 60cm Cu cylinder with wavelength shifting reflector foil 49cm 9

10 Start of Full Integration of Phase II Array finished in December 2015 all Ge and LAr detector channels working 10

11 Start of Full Integration of Phase II Array finished in December 2015 all Ge and LAr detector channels working First data release in June 2016 blinded region: Q ± 25 kev quality cuts (phys. acc. > 99.9%) 35 out of 37 detectors used for analysis events in coincidence with muon veto ( phys. acc.~ 99.9 %) 11

12 enriched Coax enriched BEGe GERDA Dec '15 June '16 FWHM [kev] Energy Scale and Resolution calibration data 3.5 physics data counts / (0.1 kev) counts / (1.0 kev) energy [kev] energy [kev] 2500 energy [kev] Q : BEGe's: 3.0(2) kev Coax: 4.0(2) kev 12

13 Performance of the LAr Veto counts / 25 kev enriched Coax 5.0 kg yr prior liquid argon (LAr) veto after LAr veto Monte Carlo 2 (T1/2= yr) 2νββ 3 GERDA :bck = 96:4 ( MeV) counts / 25 kev Ar enriched BEGe 5.8 kg yr K energy [kev] 210 Po energy [kev] 2νββ MC with T1/2 = yr from Phase I EPJC 75 (2015)

14 Pulse Shape Analysis: A/E for BEGe detector current time profile used to discriminate signal from background events A/E 1 E (A/E 1)/σA/E A α-region Signal region γ-lines 14

15 A/E Analysis for BEGe's 80% of background events rejected at Q and keep high signal efficiency = 87(2) % A/E 1 E (A/E 1)/σA/E A 2νββ acceptance % 15

16 Background Suppression LAr + PSD BEGe, 5.8 kg yr BI BI BI AC, MV = AC, MV, LAr +2.2 counts kev kg yr 3 = AC, MV, LAr, PSD +1.1 counts kev kg yr 3 = counts kev kg yr 16

17 Spectrum at Q counts / ( kev kg yr ) counts / ( kev kg yr ) counts / ( kev kg yr ) prior cuts after LAr veto (Phase II) 1 Phase I after all cuts limit (90% C.L.) 23.6 kg yr Q Extended unbinned profile likehood: Phase II enriched Coax kg yr energy [kev] flat background in kev signal = Gaussian with mean at Q and standard deviation E 7 parameters: 6 BI + common T1/ best fit for N = 0 lower limit T1/2 > yr Phase II enriched BEGe energy [kev] 5.8 kg yr with T1/2 sensitivity yr (90 % C.L.) Frequentist approach after Cowan et al., EPJC 71 (2011) energy [kev] 17

18 Current Status of GERDA Pre l since summer 2016 additional (blinded) data BEGe after LAr veto and PSD: BI = counts kev kg yr imi Coax BI = nar y counts kev kg yr 18

19 Conclusion GERDA sets new limit on the half-life of 0 decay of 76Ge 0ν 25 T 1/ 2> C. L. mββ <( )meV best energy resolution: FWHM = 3.0 kev (4.0 kev) BEGe (Coax) at Q flat background in ROI lowest background at Q : 10-3 counts/ (kev kg yr) will stay background-free important ingredients for discovery 19

20 Beyond GERDA LEGEND (Large Enriched Germanium Experiment for Neutrinoless Double Beta Decay) new collaboration formed in Oct 2016 (=GERDA+Majorana+new groups) Goals: - 1 t enriched Ge - first phase: 200 kg in existing LNGS - reduce background with respect to GERDA remain background-free best discovery potential Eur.Phys.J.C76 (2016)

21 Bonus Slides 21

22 Duty Cycle 22

23 Performance of the LAr Veto random coincidences: 2.3% 42 K line suppressed by factor K 42Ca + γ (1.5 MeV) in Ge + β- ( 2 MeV) in LAr K 40Ar (EC) no energy in LAr counts / 1 kev K 40 K energy [kev] 23

24 K Background 42 solution: transparent nylon cylinder coated with wave length shifter tested in test cryostat LArGe nylon from BOREXINO 24

25 Signals of BEGe's final drift paths of holes nearly independent of interaction point high gradient of weighting potential single site events (SSE) have similar pulse shape current signal = q v q: charge, v: velocity current [a.u.] figure taken from JINST 6 P03005, SSE time [ns] time [ns] MSE = superposition of SSE

26 PSD with Coaxial HPGe more detail in Eur.Phys.J C73 (2013) 2583 To identify signal like events artificial neural network algorithm TMlpANN from TMVA is used Input variables: times when charge pulse reach 1%, 3%,, 99% of maximum amplitude DEP events of at 1503 kev serve as signal sample FEP events at 1621 kev as multi site event sample 45% of background events rejected at Qbb with a 0 efficiency of % current signal = q v q: charge, v: velocity 2 efficiency 85 ± 2 % 26

27 Coax PSD 27

28 Background Suppression LAr + PSD Coax, 5.0 kg yr BI BI BI AC, MV = AC, MV, LAr +3.5 counts kev kg yr 3 = AC, MV, LAr, PSD +2.1 = counts kev kg yr 3 counts kev kg yr 28

29 Background Model Background Model BEGe Prelim befor inary res u e PS D & L lts Ar ve to 29

30 Background Composition at Q Monte Carlo Simulation of BEGe Background Prelim befor inary res u e PS D & L lts Ar ve to expect flat background in ROI enr Coax enr ~ 1/3 ~ 1/3 ~ 1/3 ~ 1/3 ~ 1/3 ~ 1/3 BI counts/(kev kg yr) Bi and 208Tl K LAr 42 BEGe 30

31 Effective Majorana Neutrino Mass Assuming light Majorana neutrino exchange (T 0 ν 1 1/ 2 2 ) mee observable effective Majorana mass: mee i U 2ei mi Access to absolute neutrino mass scale mass hierarchy Modified figure from Nucl. Part. Phys. Proc., 260: ,

32 Phase I + II Data Sets (T NA: Avogadro s constant, m76: molar mass of 76Ge M t: exposure [kg yr], T1/2: half-life of 0 decay, LAr LAr efficiency, PSD PSD efficiency, exposure averaged efficiency incl. active volume, enrichement, FEP ln 2 N A M t ) N 0 ν= ϵ ϵpsd ϵ LAr 0ν m76 T 1/ 2 0 ν 1 1/ 2 data set exposure [kg yr] Phase I gold 17.9 Phase I silver Energy resolution (kev, FWHM) Background index cnts/(kev kg yr) 0.57 (3) 4.3 (1) 11 ± (3) 4.3 (1) 30 ± 10 Phase I BEGe (2) 2.7 (2) Phase I extra (4) 4.2 (2) Phase II coax (4) 4.0 (2) Phase II BEGe (1) 3.0 (2) signal eff 32

33 Results from GERDA Phase I 21.6 kg y exposure blind analysis: events in ROI not available for analysis background index (BI) after pulse shape discrimination counts kev kg yr 10 times better BI than previous experiments BI=1.0 (1) 10 2 Q = 2039 kev number of events in Qββ±2σE after cuts (gray): 2.0 ± 0.3 expected from background 3 observed no signal observed at Qββ profile likelihood: best fit for N0 = 0 limit on the half-life 0ν 25 T 1/ 2 > yr (90% C.L.) claim rejected with 99% probability 0ν GERDA: 90% lower limit (T1/2 ) [Phys. Rev. Lett. 111 (2013) ] 0ν Claim: T1/2 = yr [Phys. Lett. B (2004)] 33