1. Dan Akerib Case Western Reserve University 7 July 2001 Snowmass, Colorado E6.2 Working Group The CDMS I & II Experiments: Challenges Met, Challenges Faced

2. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 2 The Cryogenic Dark Matter Search Collaboration Case Western Reserve University D.S. Akerib,D. Driscoll, S. Kamat, T.A. Perera, R.W. Schnee, G.Wang Fermi National Accelerator Laboratory M.B. Crisler, R. Dixon, D. Holmgren Lawrence Berkeley National Lab R.J. McDonald, R.R. Ross A. Smith Nat’l Institute of Standards & Tech. J. Martinis Princeton University T. Shutt Santa Clara University B.A. Young Stanford University D. Abrams, L. Baudis, P.L. Brink, B. Cabrera, C. Chang, T. Saab University of California, Berkeley S. Armel, V. Mandic, P. Meunier, M. Perillo-Isaac, W. Rau, B. Sadoulet, A.L. Spadafora University of California, Santa Barbara D.A. Bauer, R. Bunker, D.O. Caldwell, C. Maloney, H. Nelson, J. Sander, S. Yellin University of Colorado at Denver M. E. Huber University College London/Brown Univ. R.J. Gaitskell

3. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 3 WIMPs and Dark Matter Non-Baryonic dark matter  Dynamical measurements of clusters  m = 0.3  0.1  Corroborated by CMB + SNe Ia:  m ~ 0.3   ~ 0.7  BBN baryon density  b = 0.05  0.005  Structure formation requires Cold dark matter WIMPs: EW-scale couplings and 10 – 1000 GeV mass range  Thermally produced  Non-relativistic freeze-out  SUSY/LSP a natural candidate

4. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 4 Direct Detection in the Galactic Halo Galactic halo ~20% Machos  8 – 50% @ 95%C.L.  Basic paradigm intact Direct detection scattering experiment  Few keV recoil energy  < 1 event/kg/d Background suppression/rejection Low energy threshold Signal modulation WIMP detector ~10 keV energy nuclear recoil WIMP-Nucleus Scattering Importance of threshold and high quenching factor ¨ I/Xe a 50 keV true nuclear recoil threshold is equivalent to about 5 keV electron equivalent recoil

5. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 5 Selected results & goals CDMS I – best limit to date and first example of cryogenic detectors to surpass sensitivity of conventional detectors (HPGe, NaI) CDMS II – at Soudan to be 100x more sensitive DAMA 100kg NaI CDMS CDMS Stanford CDMS Soudan CRESST Genius Ge 100kg 12 m tank

6. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 6 CDMS Strategy Lines of defense  Underground site: hadrons,   Muon veto: cosmogenic , , n  Pb shield: ,   Poly shield: n  Recoil type: ,   Multiple-scatters: n  Position sensitive polyethylene outer moderator detectorsinner Pb shield dilution refrigerator Icebox outer Pb shield scintillator veto E thermal E charge Background Signal

7. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 7 Two Signals: Reject the Background Photon and electron backgrounds give more-ionizing electron recoils WIMPs and neutrons give less-ionizing nuclear recoils Plot as ratio: “Charge Yield” E recoil = E thermal – E thermal Y = E charge /E recoil E thermal E charge Background Signal (Y = Charge Yield) external gamma source external neutron source (blip detector) > 99.8% gamma rejection

8. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 8 Four 165 g Ge detectors, for total mass of 0.66 kg during 1999 Run Calorimetric measurement of total energy Energy resolution: sub-keV FWHM in phonons and ionization Inner Ionization Electrode Outer Ionization Electrode Passive Ge shielding (NTD-Ge thermistors on underside) Tower Wiring heat sinking holds cold FETs for amplifiers Berkeley Large Ionization- and Phonon-mediated Detectors Germanium BLIP Detectors

9. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 9 ZIP Ionization & Phonon Detectors ZIP: At end of fabrication steps involving µm photolithography at Stanford Nanofabrication Facility Fast athermal phonon technology  Superconducting thin films of W/Al  Stable Electrothermal Feedback configuration  Aluminum Quasiparticle Traps give area coverage

10. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 10 collimator X y Time delay Internal backgrounds  Tends to surfaces or edges Wimps  Uniform throughout bulk (zip detector) Position Sensitivity: fast phonon sensors

11. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 11 Basic simultaneous charge/ionization 1992 ~90%  -rejection  Suspected charge trapping at edges limits effectiveness Evolution from segmented electrode to “edgeless design” 1993- 1994 gives 99%  -rejection Early Stanford runs (1995-1997): reveals low-energy electrons  Electrons 10 - 100 keV stop in surface layer = “dead layer”  Reduced charge yield due to trapping defeats rejection of electron recoils  Sources: Tritium background traced to NTDs and eliminated in bakeout procedure Surface contamination – especially in earlier prototypes (too much handling)  Limits rejection to ~50% @ 10 – 20 keV Need ~factor 10 reduction to equal gammas/neutrons 4-part strategy (also applies to new ZIP detectors for CDMS II)  Cleanliness  Close-pack array Rejection History  Improve electrode structure  Fast phonon signal risetime

12. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 12 Electron Backgrounds Continuum beta contamination, problematic up to ~ 100 keV on thermal phonon-mediated Ge detectors Tritium contamination below 20 keV in Ge  Eliminated through bakeout procedure electron events Post muon veto

13. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 13 Basic simultaneous charge/ionization 1992 ~90%  -rejection  Suspected charge trapping at edges limits effectiveness Evolution from segmented electrode to “edgeless design” 1993- 1994 gives 99%  -rejection Early Stanford runs (1995-1997): reveals low-energy electrons  Electrons 10 - 100 keV stop in surface layer = “dead layer”  Reduced charge yield due to trapping defeats rejection of electron recoils  Sources: Tritium background traced to NTDs and eliminated in bakeout procedure Surface contamination – especially in earlier prototypes (too much handling)  Limits rejection to ~50% @ 10 – 20 keV Need ~factor 10 reduction to equal gammas/neutrons 4-part strategy (also applies to new ZIP detectors for CDMS II)  Cleanliness  Close-pack array Rejection History  Improve electrode structure  Fast phonon signal risetime

14. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 14 Electron Source (14C) probes charge collection at surface directly Conventional p-type implanted contact shows ~30% collection Significant improvement with new blocking contact Improved Charge Collection for Surface Events

15. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 15 Beta contamination in top detector in stack of four  Serendipitous population of tagged electron events  New electrodes of 1999 BLIP minimize “dead layer” and amount of charge lost during ionization measurement  >95% event-by-event rejection of surface electron- recoil backgrounds Surface-Event Discrimination in BLIPs 616 Neutrons (external source) 1334 Photons (external source) Ionization Threshold 233 Electrons (tagged contamination) 1999 SUF run

16. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 16 Basic simultaneous charge/ionization 1992 ~90%  -rejection  Suspected charge trapping at edges limits effectiveness Evolution from segmented electrode to “edgeless design” 1993- 1994 gives 99%  -rejection Early Stanford runs (1995-1997): reveals low-energy electrons  Electrons 10 - 100 keV stop in surface layer = “dead layer”  Reduced charge yield due to trapping defeats rejection of electron recoils  Sources: Tritium background traced to NTDs and eliminated in bakeout procedure Surface contamination – especially in earlier prototypes (too much handling)  Limits rejection to ~50% @ 10 – 20 keV Need ~factor 10 reduction to equal gammas/neutrons 4-part strategy (also applies to new ZIP detectors for CDMS II)  Cleanliness  Close-pack array Rejection History  Improve electrode structure  Fast phonon signal risetime

17. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 17 Surface-Event Discrimination in ZIPs: Risetime gammas neutrons Neutrons (low y, slow t r ) electrons surface bulk Rise time Bulk events well separated in charge yield… Charge yield, y …surface events not.

18. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 18 Summary of gamma/beta rejection history Steady improvement of rejection factors  Can we continue trend to next generation? Goals for CryoArray, see R.Gaitskell’s talk in E6, 9 July (Background fraction that leaks through)

19. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 19 Basic simultaneous charge/ionization 1992 ~90%  -rejection  Suspected charge trapping at edges limits effectiveness Evolution from segmented electrode to “edgeless design” 1993- 1994 gives 99%  -rejection Early Stanford runs (1995-1997): reveals low-energy electrons  Electrons 10 - 100 keV stop in surface layer = “dead layer”  Reduced charge yield due to trapping defeats rejection of electron recoils  Sources: Tritium background traced to NTDs and eliminated in bakeout procedure Surface contamination – especially in earlier prototypes (too much handling)  Limits rejection to ~50% @ 10 – 20 keV Need ~factor 10 reduction to equal gammas/neutrons 4-part strategy (also applies to new ZIP detectors for CDMS II)  Cleanliness  Close-pack array Rejection History  Improve electrode structure  Fast phonon signal risetime Succeeded with 1999 Data Set… see below

20. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 20 1999 CDMS Ge Data (BLIP) Combined data set from 3 BLIPs Muon anti-coincident 45 Live days – 10.6 kg-d exposure Well-separated , , nuclear recoils above 10 keV threshold 13 single-scatters consistent with residual neutron background  4 nuclear-recoil multiple-scatter events  Singles to multiples ratio established by MC  4 nuclear recoils in silicon Standard halo assumptions used to set limit Single scatters Nuclear recoils

21. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 21 Neutron Multiple Scatters Observe 4 neutron multiple scatters in 10-100 keV multiple events  3 neighbors, 1 non-neighbor  Calibration indicates negligible contamination by electron multiples Ionization Yield B6 Ionization Yield B4 photons neutron neutrons Ionization Yield B5,6 Ionization Yield B4,5 surface electrons photons Neighbor interaction B4 B3 B5 B6 Non-Neighbor interaction Neighbors Non-Neighbors

22. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 22 mostly neutrons Si ZIP measured external neutron background  For neutrons 50 keV - 10 MeV, Si has ~2x higher interaction rate per kg than Ge  Not WIMPs: Si cross-section too low (~6x lower rate per kg than Ge)  Electron-recoil leakage into nuclear recoil (NR) band small upper limit on electron-recoil leakage determined by electron, photon calibrations in 1998 Run data set:< 0.26 events in 20-100 keV range at 90% CL bulk events NR candidates 1998 CDMS Si Data (ZIP)

23. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 23 Dark Matter Limit from CDMS I CDMS 1999 DAMA 3  DAMA 2  DAMA 1996 Ge ionization Gondolo et al Bottino et al Excludes new parameter space Better than expected based on Ge singles  1 mulitple expected, 4 observed  Worse agreement 6% of the time  Likely to improve in new analysis with increased fiducial volume Bottom of DAMA NaI/1-2 2-  contour excluded at 89% Bottom of DAMA NaI/1-4 3-  contour excluded at 75% Simultaneous fit ruled out at > 99.8% CL PRL 84, 19 June 2000 astro-ph/0002471 Detailed PRD in preparation with increased fiducial mass (2x)

24. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 24 Compatibility of CDMS and DAMA Estimate DAMA Likelihood function based on “Figure 2” data (left) Simultatneous best fit to CDMS + DAMA  “standard” halo  A 2 scaling Ruled out at > 99.8% CL Accommodation?  Halo parameters?  Direct test with NaIAD CDMS bkg subtracted Best simultaneous fit to CDMS and DAMA predicts too little annual modulation in DAMA, too many events in CDMS DAMA residual spectrum

25. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 25 CDMS II – 100x improvement over present limits  Larger array & longer exposure  Second generation detectors with event positions  Ge (WIMP + n) and Si (WIMP/10 + n) — (per unit volume)  Deeper site for further reduction in cosmic-ray background Soudan Mine, Northern Minnesota 2300’ depth CDMS II Soudan II MINOS Genius Ge 100kg 12 m tank CDMS Soudan CDMS Stanford DAMA 100kg NaI CDMS (Latest) CRESST CDMS II

26. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 26 Already demonstrated discrimination to < 10 event / kg / year  >99.9% rejection of photons >10 keV (~0.5 events/keV/kg/day)  >99% rejection of surface-electrons >15 keV (~0.05 events/keV/kg/day) Identical Icebox, but no internal lead/poly, so fits seven Towers each with three Ge & three Si ZIP detectors  Total mass of Ge = 7 X 3 X 0.25 kg > 5 kg  Total mass of Si = 7 X 3 X 0.10 kg > 2 kg CDMS II Detector Deployment

27. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 27 2000-2005: CDMS II at Soudan Reduce neutron background from ~1 / kg / day to ~1 / kg / year  Soudan: Depth 713 m (2000 mwe)  First detectors in Jan 2001  Use layered polyethylene - lead - polyethylene shield (moderate the neutrons trapped inside the lead) Fridge Outer polyethylene Active Muon Veto Inner polyethylen e lead detector s Top View

28. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 28 CDMSII Deployment/Exposure Schedule Scenario 1-2-4-7 tower deployments  Factor of ~10 improvement in ~1.5 years  Factor of ~2 improvement each subsequent year T1 S T1 SUF T1-4 S 2000 2001 2002 200320042005 Full Science Running T1-7 S T1-2 S Soudan ready 1 tower in Soudan 2 tower2 in Soudan 30% 4 tower2 in Soudan 60% Begin Science ENGR

29. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 29 CDMS II goals @ Soudan (2070 mwe depth) Goal: 0.01 evt/kg/day= 0.0003 evt/kg/keV/day Units: /kg/keV/day at 15 keV (5kg Ge, 2kg Si - 2500 kg-days in Ge) ~1 per 0.25-kg detector per year 0.01 /kg /day 99.5%  rejection 95%  rejection

30. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 30 Sensitivity: CDMS II projections Based on exposure versus time and expected backgrounds  90% CL event-rate upper limit S 90  WIMP-nucleon cross section upper limit  Wn (90) at M = 40 GeV

31. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 31 Selected results & goals CDMS I – best limit to date and first example of cryogenic detectors to surpass sensitivity of conventional detectors (HPGe, NaI) CDMS II – at Soudan to be 100x more sensitive DAMA 100kg NaI CDMS CDMS Stanford CDMS Soudan CRESST Genius Ge 100kg 12 m tank

32. Snowmass 2001, E6.2 Dark Matter Detection: CDMS D. Akerib, CWRU 32 Conclusion Challenges met: technology is in hand Challenges ahead  Fabrication/yield: control of tungsten Tc understood  More of the same re cleanliness & screening Radon reduction/minimization Activation of materials  Operating complex cryogenic experiment at remote deep site If that weren’t hard enough… CryoArray: See R. Gaitskell’s talk in E6 on Mon 9 July  Description and goals for a 1000-kg experiment based on CDMS detectors  Goal of 100 event sample at 10 -46 cm 2, with <100 background events