1. 1 Searching For Dark Matter in the Universe: Direct (indirect) methods for the detection of Weakly Interacting Massive Particles (WIMPs) Nader Mirabolfathi University of California, Berkeley PIC -2004

2. 2 Evidence of Dark Matter: At Galactic scales… halo bulge disk sun Rotation curve of spiral galaxies imply the presence of dark matter  Expect v 2  1/r  Velocity is measured using atomic lines from stars or the 21cm H line for the hydrogen clouds around the galaxy Bergstrom, Rept.Prog.Phys. 63 (2000) 793 E. Corbelli & P. Salucci astro-ph/9909252 M m  If WIMPs are the halo, detect them on earth via scattering on nuclei in targets

3. 3 Cosmic Microwave Background Clusters of Galaxies Supernovae SN1a Large-Scale structure formation Many different approaches:  All agree that matter makes up approx. 27 % of the Universe and… 2003 Ω matter ΩΩ Evidence of Dark Matter: At Cosmological scales…... Big Bang Nucleosynthesis, CMB, and Structure Formation require that approx. 85% of the matter is Non Baryonic Cold Dark Matter

4. 4 Many CDM candidates: SUSY neutralinos Axions Gravitinos Kaluza-Klein states... Standard Model of Cosmology

5. 5 Candidate: Weakly Interacting Massive Particles Production = Annihilation (T ≥ m  ) Production suppressed (T<m  ) Freeze out 1 1010 1001000 m  / T (time  ) ~exp(-m/T) WIMP  : produced when T >> m  via annihilation through Z (+other channels). If interaction rates high enough, comoving density drops as exp(- m  /T) as T drops below m  : Annihilation continues Production suppressed.  Freeze out when annihilation too slow to keep up with Hubble expansion  Leaves a relic abundance:    h 2  10 -27 cm 3 s -1  ann v  fr  For   ~ 0.3: M ~ 10-1000 GeV  A ~ electroweak

6. 6 Direct Detection of WIMPs If WIMPs are the halo, detect them via elastic scattering on nuclei in targets (nuclear recoils) Energy spectrum & rate depend on target nucleus masses and WIMP distribution in Dark Matter Halo: Standard assumptions: E recoil Log(rate) Energy spectrum of recoils ~ falling exponential with ~ 15 keV Rate (based on  n  and  ) is of the order of a fraction of 1 event /kg/day  Isothermal and spherical  Maxwell- Boltzmann velocity distribution V 0 =230 km/s = 270 km/s,  = 0.3 GeV / cm 3 WIMP detector Measure recoil energy

7. 7 Experimental Challenges  Low (keV) energy threshold  Large target mass  Suppression of backgrounds from radioactivity and cosmic rays ( , , , neutrons) Deep sites Passive/active shielding  Discrimination of residual background Use WIMPS signatures WIMPs: Extremely small scattering rate, small energy of the recoiling nucleus, and subtle signatures… WIMPs Signatures: Nuclear recoils, not electron recoils Absence of multiple scattering Annual modulation Directionality Requirements:

8. 8 WIMPS Detection Methods (strategies)  00, neutrons  Nuclear recoil  Electron recoil 1)Increase the mass of the absorber and keep the background as low as possible. But how to distinguish WIMPs? i.Cosmological signature for the WIMPs assuming standard halo model. ii.Statistically remove the background. 2) Discriminate WIMPs against dominant back ground ( , ,  ). EVENT BY EVENT How? i.WIMPs are interacting with nucleons whereas , ,  interact with electrons. ii.Increase the mass. Sensitivity improves by 1/(MT) 1/2 Sensitivity improves by 1/(MT)

9. 9 Current Direct Detection Experiments

10. 10 DAMA-NaI Experiment

11. 11 NaI PMT Copper Lead Polyethelene DAMA - 100 kg NaI Experimental Apparatus Very elegant experimental setup - in place >1996 Low Activity NaI scintillator 9  9.7 kg NaI crystals, each viewed by 2 PMTs Located at Gran Sasso Underground Lab (3.8 kmwe) + Photon and Neutron shielding Two modes of Background discrimination –Pulse shape –Annual modulation: ~2% modulation amplitude POSITIVE SIGNAL

12. 12 Annual Modulation of Rate & Spectrum galactic center v0v0 Sun 230 km/s Earth 30 km/s (15 km/s in galactic plane) log dN/dErecoil Erecoil June Dec ~5% effect June Dec. WIMP Isothermal Halo (assume no co-rotation) v 0 ~ 230 km/s

13. 13 Annual Modulation Not distinguish between WIMP signal and Background directly From the amplitude of the modulation, we can calculate the underlying WIMP interaction rate Background June Dec WIMP Signal ±2% June Dec

14. 14 Modulation Amplitude There is clearly a modulation (4  - compared to null hypothesis) mean over 2-6 keVee (22 – 66 keV recoil) DAMA 2000 paper Figure 2 DAMA 1 5,000 kg-day DAMA 2 15,000 kg-day DAMA 3 + 4 38,000 kg-day Best fit to Ann Mod data alone Best Fit DAMA NaI/1-4 Best-fit WIMP model’s expected annual modulation does not appear to fit data; lowest point of 3  contour is much worse. Why? Additional constraint applied during max likelihood analysis: DC WIMP signal implied by AC signal must not exceed observed DC count rate  best-fit cross-section is decreased Minimum DAMA NaI/1-4 (3  )

15. 15 DAMA → LIBRA 3 more annual cycles acquired –58,000 + 49,800 = 107,800 kg-d –7 cycles total Improved DAQ –Multiple scatters? LIBRA – Large sodium Iodide Bulk for RAre processes –250 kg with improved radiopurity –Taking data. Results have not been announced. Further R&D toward 1-ton –NaI(Tl) radiopurification started

16. 16 ZEPLIN Zoned Electroluminescence and Primary Light In Noble gasses Location: Boulby Mine UK: UKDMC ZonEd Proportional scintillation in LIquid Noble gasses Or

17. 17 Why Xe? Available in large quantities. High atomic number (A=131) gives a high rate due to  WIMP-Nucleon  A 2 (if E is low). High density (~ 3g/cm 3 liquid). High light (175 nm) and ionization yield. Can be highly purified.  long light attenuation (m).  long free electron life time (~5ms). Easy to scale up to large volume. No long lived radioisotopes.

18. 18 Principle Of Detection Excitation Production and Decay of excited Xe 2 * states: 1)Through singlet (3ns) 2)Through triplets (27 nS) dE/dx determines the proportion of different channels=> Nuclear more dense give more singlets or faster Ionization Ionized state Xe 2 +, recombine with e - => Xe 2 * =>Above relaxation dE/dx determines the recombination time channels=> Nuclear recoils (ps scale) electrons (40 nS) Nuclear recoils are faster

19. 19 Recombination allowed. Only scintillation signal measured. Discrimination is based on the pulse shape. Discrimination is statistical. ZEPLIN I

20. 20 ZEPLIN I Results 30 keV 122 keV&136 keV 90 keV Linear response 1.5 p.e/keV  (E)=1.24E 1/2

21. 21 ZEPLINI Results (continued) Fiducial mass =3.2 Kg Mean event rate 2Hz. Trigger three fold coincidences at 1pe. 2keV threshold. Light yield 1.5-2.5pe/keV. Statistics 293 kg.day in Three runs. 2003 2002 2003 (SUF) 2002 2004 No in situ neutron calibration

22. 22 LTDs, phonon sensors and beyond! Who ? CDMS ( Cyogenic Dark Matter Search ) EDELWEISS ( Expérience pour DEtecter Les WIMPS En Sites Sousterrain ) CRESST (Cryogenic Rare Event Search with Superconducting Thermometers) Low Temperature Detectors LTD-1 1987LTD-9 2001

23. 23 LTDs, phonon sensors and beyond! Why? Advantages? After an interaction (event), all the excitations transform to heat.  Good resolution Phonon excitation~10 -6 eV compare to ~1or few eV for conventional semiconductor detectors.  Low threshold How to measure: Two methodes  T=E/C C  (T/  D ) 3   T could be big even with keV interaction. Using thermometers (Mott-Anderson or Superconducting thermometers) to measure  T. Low T  Density of thermally excited phonons (noise) is very low. But we need to collect phonons before they reach the Equilibrium in the absorber. At low T Electron-phonon interaction is more effective than ph-ph interaction  evaporated thermometers (electron bath). Temperature: EquilibriumLattice excitations (phonons) Advantages: Detecting the overall  T  No position dependence. Best resolution obtained with this kind of detectors:~100 eV at 5 MeV ? Weak points: C  Mass  Hard to increase the detector mass. Unable to reconstruct the history of evts. Advantages: Could reconstruct the history of an event. Thermometer collects constant fraction of phonons  independent of the absorber Mass. Weak points: Dispersion or position dependence of E. Homogeneity of thermometers.

24. 24 d3d3 d3>d3> d4d4 d4>d4> d2d2 d2>d2>d1d1 d1d1

25. 25 Comparison between the two types of signals  T=E/C total  T=  E/C film  T=E/C total To cold bath A) Temperature measurement B) Phonon measurement Absorber Thermometer Absorber

26. 26 Heat is not enough! Need another measurement to achieve event by event discrimination. The amount of charge created in a Semiconductor after an event depends on the type of interaction: Quenching factor (Q). Quenching factor for an electron recoil event (Most of the radioactive background) is bigger than for nuclear recoil events (WIMPs). By simultaneously measuring the charge and heat, one can discriminate - event by event WIMPs from the background. Charge? This defines the principle of detectors for CDMS and EDELWEISS experiments. Scintillation? CRESST: The same principle but scintillation instead of charge.

27. 27 Electron recoil Nuclear recoil Dead layer What is different between CDMS and EDELWEISS Collection E field needs to be very low ~3Volts/cm.  Dead layer (~50  m) > than traditional SM detectors (~1  m). limits discrimination! Most of the  bkgnd falls into DL region.  very important to deal with. Solutions Avoid surface event by: 1) Carefully dealing with surface contamination. 2) Introducing a blocking layer against the charge back diffusion  Introducing an amorphous Si layer below charge electrodes. Decrease DL to < 10  m Identify near surface events: 1) Using phonon signal. Only possible if athermal phonons measured. (CDMS current, EDELWEISS R&D) 2) Using charge signal rise time. Needs a large bandwidth electronics. (EDELWEISS R&D )

28. 28 Use of Ge NTD thermistors : FET readout The guard electrode ~50% volume

29. 29 ZIP 1 (Si) ZIP 2 (Si) ZIP 3 (Ge) ZIP 4 (Si) ZIP 5 (Ge) ZIP 6 (Si) SQUID cards FET cards 4 K 0.6 K 0.06 K 0.02 K CDMS Soudan first result with towerI Tower I: 4 Ge (250 g) and 2 Si (100 g) CDMS now running two towers 6 Ge and 6 Si Si and Ge combination helps to better understand the neutron bkgnd. Edelweiss 2002 1 Ge (320g) detectors No Si Edelweiss 2003 3 Ge (320g)

30. 30 Shielding Layered shielding (reduce , , neutrons) ~1 cm Cu walls of cold volume (cleanest material) Thin “mu-metal” magnetic shield (for SQUIDs) 10 cm polyethylene (further neutron moderation) 22.5 cm Pb, inner 5 cm is “ancient” (low in 210 Pb) 40 cm polyethylene (main neutron moderator) Active Veto (reject events associated with cosmics) Hermetic, 2” thick plastic scintillator veto wrapped around shield Reject residual cosmic-ray induced events Information stored as time history before detector triggers Expect > 99.99% efficiency for all , > 99% for interacting  MC indicates > 40% efficiency for  -induced showers from rock 30 cm parafin, 20cm Pb,1 cm Cu No active veto Dilution fridge : 17mK base.

31. 31       CDMS 2004 Results (Calibration  cuts) Neutron calibration after the run and Systematically check for gamma (e-recoil) calibration. Phonon position dependence removed. Nuclear and electron recoil bands defined (+/- 2  ) Phonon timing cuts defined with calibration data. Guard charge electrode defined. Veto coincident events defined (window 50  s). 4 Ge (850g) 2 Si (170 g) * 52 live days during 92 calendar days Selection criteria and nuclear recoil efficiency Veto coincidence (50 us window) - 97% Baseline stable (pileup, noise,…) - 95% Nuclear Recoil band (2 sigma) - 95% Phonon Timing cuts - 80% Charge outer electrode cut - 75% TOTAL - 53% Electron recoil Nuclear recoil Charge spectrum Phonon Spectrum

32. 32 WIMP search data with Ge detectors Recoil energy (keV) Charge yield Exposure –92 days (October 11, 2003 to January 11, 2004) –52.6 live days –20 kg-d net (after cuts) Data: Yield vs Energy –Timing cut off –Timing cut on –Yellow points from neutron calibration

33. 33 WIMP search data with Ge detectors Recoil energy (keV) Charge yield Exposure –92 days (October 11, 2003 to January 11, 2004) –52.6 live days –20 kg-d net (after cuts) Data: Yield vs Energy –Timing cut off –Timing cut on –Yellow points from neutron calibration

34. 34 WIMP search data with Ge detectors Recoil energy (keV) Charge yield Exposure –92 days (October 11, 2003 to January 11, 2004) –52.6 live days –20 kg-d net (after cuts) Data: Yield vs Energy –Timing cut off –Timing cut on –Yellow points from neutron calibration No nuclear-recoil candidates

35. 35 Comparing Cross section-WIMP Mass plots

36. 36 Future The presented results are from one tower CDMS II is now running two towers (6 *Ge (250 g) 6 *Si (100 g) Background of the second tower is very similar to tower I. Run stops mid July of this year New three towers of detectors will be installed October this year CDMS II ends by the end of 2005. March 2004 end EDELWEISS I Install EDEWEISS II with 21*320 g Ge NTD+ Install 7*400 g Nb x Si 1-x athermal phonon detectors (Dead layer rejection) The 100 liter dilution fridge has been successfully tested. Capacity for 120 detectors or 35 Kg Ge

37. 37 CRESST : Scintillation/Heat instead of Charge/Heat Gran Sasso Background discrimination by simultaneously measuring light/heat. Uses a cryogenic detector (the same as phonon detector) for light measurement. Works with different absorber materials: CaWO 4 (mainly), PbWO 4, BaF,..Advantage to change the absorber Phonon channel:320 g CaWO 4 (  =40mm,h=40mm), W-SPT (4*6 mm 2 ). Light channel:30*30*.4 mm 3 W-SPT. Reflector: Polymer foil, Teflon. Need 33 Modules to complete CRESST II goal

38. 38 CRESST Sensitivity and rejection High rejection: 99.7% E > 15 keV 99.9% E>20 keV 9.7 kg.day data Only half of the data analyzed. Data without neutron shield. Sensitivity limited by n.

39. 39 Future direct detection experiments

40. 40 DRIFT experiment Directional Recoil Identification From Tracks Standard halo model for WIMPs in our galaxy suggests that the axis of recoils changes in the 24 hours (earth). Axis of recoil is a cosmological signature for WIMPs. Ionization track in a low pressure gas (CS 2 ) depends on the type of interaction (Discrimination). Multi wire proportional chamber ?

41. 41 e-e- C +,S +  WIMPS E  S i  Time of flight  z Principle of DRIFT

42. 42 Low Prsure CS 2 (40 Torr) 1 m 3, 0.167 kg, 20 micron diameter wires 2 mm pitch. 1 mm track for nuclear recoils Many calibration runs with 55 Fe (5.9 keV X-rays) Neutron Calibration with 252 Cf. Polypropylene shielding (~ 50 cm). Dark matter run started. Energy threshold 15 keV. gammas C recoils S recoils DRIFT setup

43. 43 Discrimination in ZeplinII and III,IV,… Double phase Xe : Ionization Calibration of the prototype with gamma and neutron sources showed very good gamma/neutron discrimination (Cline et al. Astroparticle Phys. 12(2000) 373)

44. 44 ZEPLIN projected ~3Kg ZEPLIN I ~30kg ZEPLIN II ~1000 kg ZEPLIN IV

45. 45 Xenon: Perspective Dual phase Xe experimnent Light/Ionization Very-low BG PMT Prototype 1 cm drift 10 kg prototype underway 100 kg phase : 1 TPC Modular: each module 100 kg Self protected by outer Xe 1 Ton scale 99.5 % discrimination eff 16 keV threshold Reach:  ~10 -46 cm 2

46. 46 WIMPS indirect detection experiments AMANDA, ICECUBE (Southe po;e) ANTARES NESTOR Superkamiokande, Hyperkamiokande  -ray telescopes: CANGAROO, MAGIC, HESS Satellite experiments: AMS-02, GLAST

47. 47 WIMP indirect detection WIMP elastic scattering. But in average it will lose energy:  V<V escape  accumulates in the center of large massive objects like the sun earth or galaxy. Neutralino : Majorana particle  its own anti particle. If massive annihilates. Annihilation  ;b,c,t quarks;gauge and Higgs Bosons ,, ,e +, p -. Signature: search for excess of up-going muons Form direction from center of sun galaxy or Earth. Search for annihilation lines (galactic center, cosmological…)

48. 48 Neutrinos from the center of the earth, sun, galaxy. Assumptions: Dark matter in the galaxy due to  Density~ 0.3 GeV/cm 3 AMANDA, Super K…

49. 49 Amundsen-Scott South Pole station South Pole Dome Summer camp AMANDA road to work 1500 m 2000 m [not to scale] AMANDA

50. 50 PMT noise: ~1 kHz Optical Module AMANDA-II 19 strings 677 OMs Trigger rate: 80 Hz Data years: 2000- AMANDA

51. 51 Sensitivity to muon flux from neutralino annihilations in the center of the Earth: WIMP annihilations in the center of Earth E μ > 1 GeV Muon flux limits PRELIMINARY Look for vertically upgoing tracks NN optimized (on 20% data) to - remove misreconstructed atm. μ - suppress atmospheric ν - maximize sensitivity to WIMP signal Combine 3 years: 1997-99 Total livetime (80%): 422 days No WIMP signal found Disfavored by direct search (CDMS II) Limit for “hardest” channel:

52. 52 CDMS 2004

53. 53 Summary Direct detection Experiments (CDMS,EDELWEISS, CRESST..) have already explored the regions of the most optimistic SUSY models. Despite the lower amount of exposure (~20 kg.day compare to 110,000 kg.day), the event-by-event discrimination methods are giving the best sensitivities. Extremely high discrimination + large mass seems to be the only solution for the next generation of direct detection experiments. The current and next experiments (CDMSII, EDELWEISSII, ZEPLIN IV, XENON …) will explore the core of many SUSY models in few years. Indirect detection will be complementary but hardly competitive to low  scalar WIMPs detection. The accelerator (LHC) results + the direct detection experiments will soon (not in the cosmological sense!) let to discover the nature of the dark matter.

54. 54

55. 55 Lowest energy bin is most important when setting dark matter limit DAMA - Energy Spectra Bernabei et al. PLB450(1999)448 Data re-plotted by Gaitskell, CfPA Rate corresponding to best fit to Ann Mod data alone (  = 14.0 10 -42 cm 2 ) is shown as cyan - exceeds signal in 2-3 keV bin WIMP component assuming: (50 GeV,7.2x10 -42 cm 2 ) (50 GeV,14.0x10 -42 cm 2 ) Best fit to Ann Mod data alone CDMS 2000 limit (90% CL) Low Background in Det #8 constrains max  All Histograms already corrected for trigger / cut efficiencies

56. 56 Modulation Animation in NaI 50 GeV WIMP 000904.4 rjg Background Sun moving through WIMP Halo Threshold

57. 57 Depth (mwe) Log 10 (Muon Flux) (m -2 s -1 ) Depth (mwe) Muon flux: 4/m 2 /day Neutron flux:1.5e-6/cm 2 /s Muon flux: 70/m 2 /day Neutron:

58. 58 WIMP search data with Ge detectors Recoil energy (keV) Charge yield Exposure –92 days (October 11, 2003 to January 11, 2004) –52.6 live days –20 kg-d net (after cuts) Data: Yield vs Energy –Timing cut off –Timing cut on –Yellow points from neutron calibration Well, maybe 1….