New Physics Results from DarkSide-50. Masayuki Wada Princeton University on behalf of the DarkSide-50 Collaboration Feb

Transcription

1 New Physics Results from DarkSide5 Masayuki Wada Princeton University on behalf of the DarkSide5 Collaboration Feb. 8 Lake Louise Winter Institute 8

2 DETECTOR DARKSIDE 5 Radonfree (Rn levels < 5 mbq/m ) Assembly Clean Room,tonne Water Cherenkov Cosmic Ray Veto tonne Liquid Scintillator Neutron and γ s Veto Veto efficiency > 99.% Inner detector TPC

3 r e DARKSIDE5 THE TIMEPROJECTION CHAMBER (TPC) e e Electron Recoil (ER) acquired charge acquired charge time S Nuclear Recoil (NR) S (S) (S) (S) (S) time S/S ratio and Pulse Shape Discrimination (PSD) WIMPs will generate nuclear recoils (NRs) DM S S n WIMPlike signal! E drift xy position from S light fraction E mult S e e e S z t drift liquid argon (L)

4 DARKSIDE5 PULSE SHAPE DISCRIMINATION M. G. Boulay and A. Hime, Astropart. Phys. 5 (6) 79 4 Electron and nuclear recoils produce different excitation densities in the argon, leading to different ratios of singlet and triplet excitation states τsinglet ~ 7 ns τtriplet ~ 5 ns f9 ~. Electron Recoil (ER) PSD parameter F9: Ratio of detected light in the first 9 ns, compared to the total signal ~ Fraction of singlet states NR band f9 ~.7 Nuclear Recoil (NR) WIMPlike signal! ER band AmBe calibration data

5 A Blind Analysis of 54 livedays of data f S [PE] Blinding box (red outline) shown with 7day data: PR D 9, 8 (6) Goal: design an analysis that will have <. event of background in the tobedesigned search box. (Final box chosen: dashed red) 5

6 Nuclear Recoil BG Events / [5 PE] Surface α decays Energy [MeV] Degraded: under TPB Data Po Rn 8 Po Full energy in L Background rejection: S < 46 Selfvetoing in DS5! Small or no S Long scintillation tail from TPB Neutrons S [PE] n + B α + 7 Li Background rejection: TPC: multiscatter LS Veto n + B α + 7 Li + γ 6 Measured n efficiency with AmC for TPC singlenr:.9964±.4 Cosmogenics: Water Cherenkov Veto Neutrons in data counted

7 Electron Recoil BG S+Cherenkov S + Cherenkov events are Dangerous!! f Energy [kev nr ] Li(p,n) in SCENE 4 AmBe in DarkSide5 Cherenkov light is all in prompt. Move regular scintillation events into NR band S [PE] Background rejection: Underground S fraction in max PMT PSD: f9 = S fraction in first 9 ns Design cut to reduce ER to <.8 event of background 7 Background Est. Survive Surface a decays All. Cuts Cosmogenic n <. Radiogenic n <.5 ER S+Cherenkov.8* Total.9±.4 Goal achieved: open the box!

8 Final Data Set Energy [kev nr ] f % 5% 99% S [PE] 8 arxiv: 8.798

9 9% C.L. Exclusion Limits [cm ] Dark MatterNucleon σ SI [TeV/c ] M χ DarkSide5 (8) PANDAXII (7) XENONT (7) DEAP6 (proj.) LZ (proj.) WARP (7) DarkSide5 (5) DEAP6 (7) DarkSidek (t yr LUX (7) XENONT (proj.) XENONnT (proj.) proj.) DarkSidek (t yr proj.) Future tonne GADMC Detector (ktyr proj.) Future tonne GADMC detector (ktyr proj.) 9 arxiv: 8.798

10 Low Mass Dark Matter Search Ionization Only Analyses arxiv: arxiv:

11 Ionization Measurements Scintillation signal (S): threshold at ~ kevee / 6 kevnr weak sensitivity to low mass WIMPs. Ionization signal (S): threshold <. kevee /.4 kevnr Sensitive to low mass WIMPs!! Use Ionization (S) Only. PMTs have zero dark rate at 88K Gain in the gas region (~7 e/e ) Sensitive to a single extracted electron Radioactivity rate in the detector is remarkably low No need of PSD The electron yield for nuclear recoils increases at low energy kg day] Acceptance Detection Efficiency MC Simulation Fiducialization Trigger efficiency S Identification (f <.5) 9 Ne 4 (Analysis Threshold for six Inner PMTs) S [PE] The efficiency is flat above the analysis threshold of number of electrons >4. S yield from Singleelectrons Center PMT DS5 DATA Getter Off Getter On Fit Ext. e Ext. e 's Detection efficiency: estimated from Data + MC S light distribution pattern among PMTs S waveform including longitudinal diffusion due to drifting Events /[.5 N e S=PE Fiducialization: use volume under 7 central PMTs In DS5, we can detect down to single electron. arxiv: Ne S/η, ηc = PE/e N e One electron creates ± PE at the center PMT Number of electrons

12 Energy scale for ER and NR 7 provides two xrays,.8 kev and.7 kev. Decayed out with 5 day halflife and not remain in the last 5days data set. Good agreement of BR with expected value. Confirmation of the trigger efficiency and acceptance AmBe and AmC neutron sources are used to extract ionization yield at ROI. The difference between other measured points is take as systematics arxiv: day] kg /kev nr ] Events / [.5 N e [e Q y Ionization Yield E [kev nr ] 9 E Xe [kev nr ] Electron Recoil Energy Scale N e Xe Data Xe Model First days Last 5 days Single S(5 d) S + S(5 d) LUX 6 7 ZEPLINIII XENON Manzur et al. XENON 9 Aprile et al. 6 Bezrukov et al. kg day] Events /[N e Data ARIS SCENE AmBe AmC ARIS SCENE Joshi et al. 4 Joshi et al. 4 Cross Calibrated N e ε L/K BR Ratio =. ±. 7.7 kev Nuclear Recoil Energy Scale Xenon 7.8 kev Ne=47.9 Number of electrons gon Reduced Energy NR ionization yield is obtained by fitting AmBe and AmC neutron calibration data

13 Background and WIMP Signals Fit the BG components at high energy in the same dataset (field V/ cm) Events/[ kev] High Energy Spectra 5 4 Th 4 K 8 U low 5 9 U Cryo 6 Co 8 U up 85 Kr DATA Put constraints on the rate of BG at ROI BG events in ROI are modeled with ER Ionization yield and detector response. The excess of events due to tail of the delayed electrons are not modeled as BG. WIMP recoil spectra are modeled with day] kg Energy [kev] WIMP Search ROI E [kevnr 45 ] 4 DM spectra σ χ = M χ =.5 GeV/c M χ =5. GeV/c M χ =. GeV/c cm Extrapolation at low energy, ok within few % E [kevee] Data G4DS MC All Cryostat γrays PMTs γrays Kr day] kg NR Energy Quenching Ionization Detector Response Events / [N e Excess of events due to tail of delayed electrons Those events were also observed in XENON N e Ana. Thres. Mχ 4 GeV arxiv: Ana. Thres. Mχ 4 GeV Number of electrons Events / [kevee

14 8 9% C.L. Exclusion Limits [cm ] 9 No Quenching Fluctuation Dark MatterNucleon σ SI Binomial Fluctuation DarkSide5 DarkSide5 Binomial DarkSide5 No Quenching Fluctuation NEWSG 8 LUX 7 XENONT 7 PICO6 7 PICASSO 7 CRESSTIII 7 CDMSLite 7 PandaXII 6 XENON 6 CDEX 6 DAMIC 6 CRESSTII 5 SuperCDMS 4 COGENT CDMSlite 4 CDMS CRESST DAMA/LIBRA 8 Neutrino Floor 5 M χ 4 [GeV/c ] Profile Likelihood Method is used Uncertainties from both WIMP signals (NR ionization yield, single electron yields) and BG spectrum (rates, ER ionization yield) Due to lack of knowledge about fluctuation at low recoil energy, two cases are considered. Binomial fluctuation for NR energy quenching, ionization, and recombination processes. No Fluctuation for NR energy quenching process. Corresponding to apply hard cut off in quenched energy ~.6 kevnr. 4 arxiv:

15 SubGeV Dark Matter Search arxiv: Light DM scatter off electrons DM signals are also ER. The same measured spectrum as the WIMP search is used. kg day] Events / [N e E [kevee] FDM = 6 DM spectra σ χ = M χ = MeV/c M χ = MeV/c M χ = MeV/c cm Data G4DS MC All Cryostat γrays PMTs γrays Kr N e Two extreme cases of Dark Matter formfactor are considered kg day] E [kevee] FDM /q DM spectra σ χ = M χ = MeV/c M χ = MeV/c M χ = MeV/c cm Data G4DS MC All Cryostat γrays PMTs γrays Kr FDM= heavy mediator FDM /q light mediator Events / [N e N e

16 SubGeV Dark Matter Exclusion Limits arxiv: [cm ] FDM = FDM /q [cm ] Dark MatterElectron σ e 7 8 F DM = DarkSide5 XENON XENON Dark MatterElectron σ e 4 5 F DM /q DarkSide5 XENON XENON 9 M χ [MeV] 6 M χ [MeV] Profile Likelihood Method is used Uncertainties from ER ionization yield and single electron yields are included both DM spectra and BG spectra. Rates uncertainties are included in BG spectra. In the case of a heavy mediator, FDM =, we improve the exclusion limit in the range from MeV/c to 7 MeV/c. 6

17 Summary Blind Analysis is successfully done with 54 livedays of data. Pulse Shape Discrimination (f9) is strong discriminator and necessary for BG free WIMP search at high mass. Liquid gon is also sensitive to low mass WIMPs and subgev DM. Next generation DarkSidek is coming! arxiv: arxiv: arxiv:

18 Thank You! 8

19 Additional Rejection S/S f Energy [kev nr ] % 5% 99%. S/S cut with 5% acceptance + R< cm radial cut S [PE]

20 Energy [kev nr ] SingleScatter Unresolved MultiScatter NR Acceptance Total cut acceptance f 9 NR Acceptance Overall NR Acceptance Events / [. f9] Scintillation + FusedSilica Cherenkov Scintillation + PTFE Cherenkov S [PE] f f Radiogenic Cosmogenic Fission candidate 5% NR acceptance S [PE] Drift time [µs] Reconstructed radius [cm] 8 6 4

21 NR Ionization Yields AmBe neutron source 9 AmC neutron source Events / N e AmBe Data G4DS Fit Single S S + S Events / N e Am C Data G4DS Fit Single S S + S 4 Am C γ N e number of electrons N e number of electrons MC + Ionization model [] fit to NR data from AmBe and AmC. The systematic discrepancy between the extracted and measured ionization yield is taken as systematic uncertainty. /kev nr ] [e Q y E [kev nr ] 9 E Xe [kev nr ] Xe Data Xe Model LUX 6 ZEPLINIII XENON Manzur et al. XENON 9 Aprile et al. 6 Bezrukov et al. Data Xenon ARIS SCENE AmBe AmC ARIS SCENE Joshi et al. 4 Joshi et al. 4 Cross Calibrated ε gon [] F. Bezrukov, F. Kahlhoefer, and M. Lindner, Astropart. Phys. 5, 9 ().

22 Quality +Trgtime +Ssat f 9.9 Entries Integral.9794e e S [PE] Trigtime: the first pulse is within expected trigger time window Ssat: S pulse is not saturated

23 +Npulses f Entries Integral 4.e S [PE] Npulses: number of pulse is or if there is S (echo of S). Most of surface events are gone.

24 +4µs fid f 9.9 Entries 4474 Integral.88e S [PE] 4us fid: remove 4 us from top and bottom in t_drift. Lots of γs from PMTs, unresolved S+S events, and surface close to top are removed. 4

25 +Spmf f 9.9 Entries 8895 Integral.7e S [PE] Spmf: fraction of prompt light in the maximum PMT is less than a threshold, which is a function of t_drift and S Remove scintillation + Cherenkov events from Fused Silica 5

26 f 9.9 +min Suncorr Entries 8664 Integral.695e S [PE] min Suncorr: S have PE This is more like quality cut, but remove surface events, which number of electrons are reduced by the surface effect. 6

27 +xyrecon f 9.9 Entries 8556 Integral.688e S [PE] xyrecon: reasonable xy reconstructed values (reconstruction does not failed) 7

28 f 9.9 +S F9 Entries 8556 Integral.688e S [PE] s F9: F9 of S pulse <. Remove S + S pileup events 8

29 f min S/S Entries Integral.68e S [PE] min S/S: S/S need to be above threshold, which is a function of S. Remove strangely small S events, like surface events. 9

30 f max S/S Entries Integral 5.4e S [PE] max S/S: S/S need to be below threshold, which is a function of S. Remove strangely large S events, which we don t expect, but applied as a safety net.

31 f S i9/i Entries Integral 5.8e S [PE] S i9/i: S have reasonable rise time. Remove events in which S is actually S+S pulses.

32 f S TBA Entries Integral 5.9e S [PE] S TBA: zposition from S TopBottom asymmetry agrees with t_drift. Remove random pileup S and S.

33 f TPB Tail Entries Integral 5.97e S [PE] TPB Tail: remove events, which have long tail of scintillation caused by TPB scintillation. Remove surface events, in which alpha goes through TPB layer.

34 f NLL Entries 5445 Integral 5.66e S [PE] NLL: Negative Log Likelihood cut, which compare event position from S light distribution among PMTs and event position from t_drift and S xy. Remove Cherenkov + scintillation events which deposit energy in separate locations. 4

35 +R f 9.9 Entries 6687 Integral.576e S [PE] R : Radial cut as a function of t_drift 5

36 +Veto f 9 Entries Integral.58e S [PE] Veto: all veto cuts Remove neutrons 6