CDMS. Vuk Mandic. 13. July 2010

Transcription

1 Dark Matter Direct Detection: CDMS Vuk Mandic University of Minnesota 13. July 2010

2 CDMS II Collaboration California Institute of Technology Z. Ahmed, J. Filippini, S.R. Golwala, D. Moore, R.W. Ogburn Case Western Reserve University D. Akerib, C.N. Bailey, M.R. Dragowsky, D.R. Grant, R. Hennings-Yeomans Fermi National Accelerator Laboratory D. A. Bauer, F. DeJongh, J. Hall, D. Holmgren, L. Hsu, E. Ramberg, R.L. Schmitt, J. Yoo Massachusetts Institute of Technology E. Figueroa-Feliciano, S. Hertel, S.W. Leman, K.A. McCarthy, P. Wikus NIST * K. Irwin Queen s University it P. Di Stefano *, N. Fatemighomi *, J. Fox *, S. Liu *, P. Nadeau *, W. Rau Santa Clara University B. A. Young Southern Methodist University J. Cooley SLAC/KIPAC * E. do Couto e Silva, G.G. Godrey, J. Hasi, C. J. Kenney, P. C. Kim, R. Resch, JG J.G. Weisend Stanford University P.L. Brink, B. Cabrera, M. Cherry *, L. Novak, M. Pyle, A. Tomada, S. Yellin Syracuse University M. Kos, M. Kiveni, i R. W. Schnee Texas A&M J. Erikson *, R. Mahapatra, M. Platt * University of California, Berkeley M. Daal, N. Mirabolfathi, A. Phipps, B. Sadoulet, D. Seitz, B. Serfass, K.M. Sundqvist University of California, Santa Barbara R. Bunker, D.O. Caldwell, H. Nelson, J. Sander University of Colorado Denver B.A. Hines, M.E. Huber University of Florida T. Saab, D. Balakishiyeva, B. Welliver * University of Minnesota J. Beaty, H. Chagani *, P. Cushman, S. Fallows, M. Fritts, O. Kamaev, V. Mandic, X. Qiu, A. Reisetter, J. Zhang University of Zurich S. Arrenberg, T. Bruch, L. Baudis, M. Tarka * new collaborators or new institutions in SuperCDMS 2

3 Cryogenic Dark Matter Search Search for Dark Matter in the form of WIMPs. Identify/suppress all known forms of particle interactions. ti Cosmogenic: Deep underground: Soudan mine, Minnesota, 713m below surface. Muon scintillator veto: Reject muon-coincident events. Ambient neutron and electromagnetic background: Passive shielding: Pb and polyethylene. Residual gamma and beta backgrounds: Ge/Si-based detectors cooled to ~40 mk Phonon and ionization signals WIMP Terrestrial Particle Event-by-event identification Detector Remaining neutron background: Ge vs Si event rates Single vs multiple event rates WIMP-Nucleus Scattering energy transferred appears in wake of recoiling nucleus 3

4 CDMS II Detector: ZIP Z-sensitive Ionization and Phonon Detector 3 diameter, 1cm thick, 250 g (Ge) R Sensors deposited and I bias photolithographicaly patterned on the surface. Two charge/ionization electrodes on one surface: Inner disk and outer ring. Four phonon sensors, each covering one quadrant, on the opposite surface. SQUID array R bias D C A B Phonon D R feedback Q outer Q inner V qbias 4

5 Ionization and Phonon Signals Ionization: - Fast: 1 μs srisetime rise-time, 40 μs sfalltime fall-time. - Good measure of the Event Time. A B D C Phonons: - Pulse shape (start-time, rise-time, energy distribution among 4 quadrants) depends on event position. Ionization and Phonon signal amplitudes reveal the recoil energy. Timing and amplitude of the phonon signals can be used to reconstruct event position. Allows position correction of any nonuniformities (T c gradient). 5

6 Ionization Yield Ionization yield: ionization signal divided by recoil energy. 133 Ba γ-source used to define the electron-recoil recoil band. Calibration Data 13x our WIMP-search background 252 Cf n-source used to define the nuclear-recoil band. The bands are well separated down to below 10 kev! 6

7 Ionization Yield Ionization yield: ionization signal divided by recoil energy. 133 Ba γ-source used to define the electron-recoil recoil band. Calibration Data 13x our WIMP-search background 252 Cf n-source used to define the nuclear-recoil band. The bands are well separated down to below 10 kev! Small fraction of electron-recoils trickles down to the nuclear recoil band surface events! 7

8 Surface-Event Rejection Calibration Data Reject Keep Phonon pulse-shape p contains information on interaction depth: 8

9 WIMP Search Exposure Total raw exposure is 612 kg-days recorded data some detectors not analyzed for WIMP scatters periods of poor data quality removed this work 2008 result PRL102, (2009) Data taken from 9/08-3/09: primarily an engineering run 9

10 Runs : Blind Analysis All WIMP search data We unblinded the signal region November 5,

11 Events Failing Timing Cut All WIMP search data failing the timing cut 150 events in the NR band fail the timing cut, consistency checks deemed ok 11

12 Events Passing Timing Cut All WIMP search data passing the timing cut Event 1: Tower 1, ZIP 5 (T1Z5) Sat. Oct. 27, :48pm CDT Event 2: Tower 3, ZIP 4 (T3Z4) Sun. Aug. 5, :41 pm CDT 2 events in the NR band pass the timing cut! 12

13 WIMP Signal? 2 candidate events: Periods of nearly ideal experimental performance. Different months, different detectors. So, maybe Expected backgrounds: Surface event leakage (based on calibration studies): 0.8 ± 0.2 events. Neutron background: <0.1 events. Cosmogenic and radiogenic. Data, Simulations, Counting. 23% probability of observing 2 or more events, given these backgrounds. Results of this analysis cannot be interpreted as significant evidence for WIMP interactions, but we cannot reject either event as signal. 13

14 90%CL Spin-Independent Limit In the presence of 2 events (no bg subtraction): CDMS Combined Soudan mass 70 GeV σ = 3.8 x cm 2 (90% C.L.) Science 327, 1619 (2010) Sensitivity curve assuming: 0.8 ±0.1(stat.) ( ) ±0.2(sys.) ( y ) surface events cosmogenic neutrons radiogenic neutrons 14

15 Runs : Low Energy Electromagnetic signatures in CDMS detectors possibly new physics. Similar to our standard analysis: Use electron-recoil events. Do not impose timing cut. Low recoil energies particularly interesting. Understand the backgrounds well. Several lines due to cosmogenic activation. Line widths (energy resolution) well understood. Use kev window. Observed Electron-Recoil Spectrum Feature at 6.54 kev likely due to de-excitation of 55 Mn PRD81, (2010) (cosmogenic activation). 15

16 Solar Axion Background Axion-photon coupling: In Coulomb field of the nucleus, a γ. Recoil energy = incident energy. Standard solar model gives the axion flux: Time and energy dependence of solar axion conversion rate for g aγγ =10-8 GeV -1 Coherent Bragg diffraction: momentum transfer equal to reciprocal lattice vector. For a given direction (sky location) there are preferred recoil energies. Complex modulation o pattern, PRL (2009) dependent on incident/recoil energy. Sets the relevant energy scale PRL103, (2009) 16

17 Solar Axion Background Similar studies were done in the past: SOLEX, COSME, DAMA New feature: angular orientation is well understood: Uncertainty of 3 dominated by the relative tower-cryostat orientation. Place a new 95% CL on the axion- photon coupling: g aγγ < GeV -1 Applies to axion mass below 0.1 kev. Larger masses suppressed in the solar axion flux. Expect ~10x improvement with SuperCDMS-100 kg. PRL103, (2009) 17

18 Galactic Axion Background Repeat the analysis for galactic axions. Non-relativistic, axio-electric coupling. Signal appears at the axion rest mass. Place an upper limit on g aee at each axion mass. g aee < for a 2.5 kev axion Incompatible with galactic axion interpretation of DAMA signal. 55 Mn feature at 6.54 kev not subtracted (no direct constraint on this contribution). PRL103, (2009) 18

19 Low-Energy Spectra Can attempt a comparison with DAMA/LIBRA signal, in the interpretation of electromagnetic energy deposition by WIMPs. Big uncertainty how does cross section change between Ge and NaI? Assume Z 2 dependence. Scale CDMS (Ge) ratetoto estimate total rate in NaI. Compare with total rate observed by DAMA at the 3.15 kev peak. Observe large discrepancy. Could be reduced if the 40 K contamination is understood (leading to a 3.2 kev line). 90% CL upper limits PRD81, (2010) 19

20 Outlook: CDMS II SuperCDMS GEODM CDMS II 3 x 1cm ~ 0.25 kg/det 16 detectors = 4 kg ~2 yrs operation SuperCDMS 3 x 1 ~ 0.64 kg/det Soudan SNOlab 25 detectors t = 15 kg 150 detectors t = 100 kg 2 yrs ~ 8000 kg-d 3 yrs ~ kg-d SuperCDMS SNOlab and Ge-Observatory for Dark Matter (GEODM) 6 x 2 ~ 5.1 kg/det SNOlab DUSEL 20 detectors t = 100 kg 300 detectors t = 1.5 ton 3 yrs ~ 100,000 kg-d 4 yrs ~ 1.5 Mkg-d 20

21 New Technology: izip Detector Improved yield=q/p performance: Surface rejection: 1:3000 (currently 1:350) Less Al, tangential E-field. Ionization side asymmetry: Surface rejection: 1:1000 Phonon timing and symmetry between two sides: Surface rejection: 1:3000 Likely correlations with other two parameters. Meets requirements for the ton-scale experiment. Could not measure overall rejection efficiency in a surface facility (cosmogenic neutrons dominate over the surface leakage). 21

22 Spare Slides 22

23 Fit DAMA/LIBRA Spectrum 23

24 Maximum Likelihood Analysis (Axion models) Solar axion flux Expected rate (solar) Rate model (solar) Background rate model Likelihood definition Expected rate (galactic) 24

25 mzip Minor design changes. 1 thick, 2.5x suppression of surface events. Stadium phonon sensor design: Covers more surface area Improves phonon collection and SNR. Mercedes -like phonon sensor layout. Better phonon signal at the outer edge. Breaks degeneracies in position reconstruction. Improves phonon timing information. 25

26 mzip at Soudan Meets requirements for 15-kg stage (possibly even beyond). SuperTower 1 installed at Soudan. Five 1 -thick detectors + 2 end- cap veto detectors. t Started to look at the data. 26

27 Phonon Sensors Al Collector Al quasiparticle trap quasiparticle W diffusion Transition-Edge Sensor phonons Si or Ge Measurement of athermal phonon signals maximizes information Fast pulse, excellent energy and timing resolution R TES (Ω Ω) superconducting W Transition-Edge Sensor: a really good thermometer ~ 10mK normal T c ~ 80mK T (mk) 27

28 Phonon Sensors 60 μm wide 380 μm Al fins 4 quadrants 37 cells per quadrant 6x4 array of W transition-edge sensors per cell Each W sensor fed by 8 Al fins ~1000 TES per quadrant, wired in parallel! 28

29 Phonon Signal Readout 10 pa/ Hz First-stage amplification: Low-noise SQUID-based amplifier. SQUIDs operated at 600 mk. Current through the sensor ~1μA. Bandwidth ~ 70 khz Amplified 10x by the SQUID. Further amplification at room temperature. 29

30 Ionization Electrodes & Readout Phonon sensors also serve as ground for ionization readout. Opposite side has two electrodes: Inner disk, outer ring. Defines fiducial volume: reject events close to the edge. First-stage amplification: FET-based amplifier. High-impedance: Susceptible to microphonics vacuum coax. Keep gate wire short FET must be in the cryostat heat load. These are major drivers of the cold hardware design. Sub-microsecond rise-time R F C F 40μs fall-time 30

31 Ionization Channel Noise Amplifier bandwidth ~160 khz. Noise consistent with 0.5 nv/ Hz of the FET. Equivalent to ~1 kev energy threshold. h 31

32 Position Reconstruction Exposed one detector to a largesurface 109 Cd source, behind a Pb collimator. Event-position reconstructed using phonon start-times. 32