Dark matter search with EDELWEISS Recent results and prospects

Gilles Gerbier- CEA Saclay/IRFU on behalf of Edelweiss (and CDMS) collaboration(s) and also some personal views

PPC2011 CERN -june 15th 2011

1 Expérience pour DEtecter Les Wimps En SIte Souterrainn Outline

 Edelweiss : “ID technology”, results of 1 year run

 Combination of data with CDMS : results

 Prospects of Edelweiss for 2012-2013

 Low mass WIMP’s : new result, comments wrt CoGeNT

 Further future : EURECA @ ULISSE Principle of WIMP direct detection WIMP Galactic WIMP Interaction in a terrestrial detector velocity v ~ 200 km/s θr local density ρ0 Energy deposition Nuclear recoil Er

• Relevant parameters:

- mass mχ ~ 10 GeV to 10 TeV for usual extensions of the Standard Model - WIMP-nucleon cross-section σ, weakly 1 evt / kg / yr contrained but of the order of EW scale

• Non-relativistic diffusion: 1 evt / ton / yr ~ 1 - 100 keV

• Interaction rate:

0 v 100 100 GeV 0 0 v0 1 1 R ~ ~ 0.04 8 3 1 kg day m mN A m 10 pb 0.3GeV cm 230 km s

3

Principle of WIMP direct detection WIMP Galactic WIMP Interaction in a terrestrial detector velocity v ~ 200 km/s θr local density ρ0 Energy deposition Nuclear recoil Er

• Low-threshold detectors • Relevant parameters: • Ultra-low-background detectors : - mass mχ ~ 10 GeV to 10 TeV for usual extensions of the Standard Model- « Passive » bckgd reduction (shields, - WIMP-nucleon cross-section σ, weakly 1 evt / kg / yr contrained but of the order ofradiopurity, EW scale external vetos..) - « Active » bckgd reduction • Non-relativistic diffusion: (discrimination of electron recoils, multiple scatters..) 1 evt / ton / yr ~ 1 - 100 keV • We consider only the « spin- • Interaction rate: independent » channel here → single

0 v 100 100 GeV WIMP-nucleon0 0 cross-sectionv0 ∀ target1 1 R ~ ~ 0.04 8 3 1 kg day m mN A m 10 pb 0.3GeV cm 230 km s

4

The EDELWEISS collaboration

Oxford - sep 10 Karlsruhe - oct 09 FRANCE ITALIE

• CEA Saclay (IRFU and IRAMIS) 4800 mwe LSM (Fréjus) • CSNSM Orsay (CNRS/IN2P3 + Univ. Paris Sud) Altitudes 1228 m 1263 m 1298 m • IPNLyon (CNRS/IN2P3 + Univ. Lyon 1) Distances 0 m 6210 m 12 868 m • Institut Néel Grenoble (CNRS/INP) • Karlsruhe Institute of Technology • JINR Dubna 131 km • Oxford University (joined in 2009) from here • Sheffield University (joined in 2010) The EDELWEISS-II infrastructure

 Cryogenic installation (18 mK) :  Reversed geometry cryostat, pulse tubes

 Remotely controlled  Can host up to 40kg of detectors  Shieldings :

 Clean room + deradonized air  Active muon veto (>98% coverage)

 50-cm PE shield

 20-cm lead shield ⇒ γ background reduced by ~3 wrt EDW1  Other items:  Remotely controlled sources for gamma calibrations + regenerations  AmBe sources of neutron calibrations

 Detector storage & repair within the clean room 3  Radon detector down to few mBq/m  He3 neutron detector (thermal neutron monitoring inside shields) sensitivity ~10-9 n/cm2/s  Liquid scintillator neutron counter (study of muon induced neutrons)

6 The EDELWEISS-II infrastructure

 Cryogenic installation (18 mK) :  Reversed geometry cryostat, pulse tubes

 Remotely controlled  Can host up to 40kg of detectors  Shieldings :

 Clean room + deradonized air  Active muon veto (>98% coverage)

 50-cm PE shield

 20-cm lead shield ⇒ γ background reduced by ~3 wrt EDW1  Other items:  Remotely controlled sources for gamma calibrations + regenerations  AmBe sources of neutron calibrations

 Detector storage & repair within the clean room 3  Radon detector down to few mBq/m  He3 neutron detector (thermal neutron monitoring inside shields) sensitivity ~10-9 n/cm2/s  Liquid scintillator neutron counter (study of muon induced neutrons)

7 « EDELWEISS-I-like » detectors

 Germanium bolometers  Ionization measurement @ few V/cm  Heat measurement (NTD sensor) @ 20 mK:

 heat signal = k x recoil energy + Luke effect

 Discriminating variable between electronic and nuclear recoils : « Q » = ionization/recoil energy

⇒ Full separation between ER and NR electron recoils

inelastic neutron diffusion

Neutron calibration (S. Scorza PhD thesis) nuclear recoils

8 « EDELWEISS-I-like » detectors

 Germanium bolometers  Ionization measurement @ few V/cm  Heat measurement (NTD sensor) @ 20 mK:

 heat signal = k x recoil energy + Luke effect

 Discriminating variable between electronic and nuclear recoils : « Q » = ionization/recoil energy

⇒ Full separation between ER and NR

Surface interactions

 Population identified as originating from 210Pb at surface of detector or Cu cover

9 Rejecting surface events with interleaved electrodes ‘a’ electrodes (+4V) the « ID » (InterDigit) detector collecting ‘b’ electrodes (-1.5V) field shapping

fiducial volume First 200g ID

• Keep the EDW-I NTD phonon ‘c’ (-4V) ‘d’ (+1.5V) detector • Modify the E field near the - First detector built 2007 surfaces with interleaved - 1x200g + 3x400g tested in 2008 electrodes - 10x400g running 1 year 2009-2010 • Use ‘b’ and ‘d’ signals as vetos - 800g detectors tested and running 2010-2011 against surface events (NB : an alternative solution, NbSi films, studied for 10 years was abandoned in 2007)

10 Ionization yield for surface events

210Pb calibration (2008) Background data

6x104 210Bi 6x104 210 6x104 Po 210Pb

1 evt

PLB 681 (2009) 305-309 [arXiv:0905.0753]

 Much better than anticipated !

11 WIMP search with ID-400 detectors

1) ~ 20 kg.d in 2008 during validation runs of 2 ID detectors 2) Physics run Apr 2009 - May 2010 (10 detectors) : ~ 380 kg.d  Published : 2008 + 6 months 2009 = 160 kg.d - PLB 687 (2010) 294–298 [arXiv: 0912.0805]  Final results of the complete run presented here - submited [arXiv:1103.4070]

 All heat sensors working  55/60 electronics channels working  calibrations show we can use all detectors for WIMP search thanks to redundancy

Fiducial zone ~170g

12 ID-400 : gamma calibration

• Regular calibrations between background runs with two motorized 133Ba sources (356 keV)

• all IDs stacked • same analysis/cuts as for bg data • more than 350 000 fiducial evts

• anomalous events observed: - for 20

13 WIMP search : first six month result

160 kg.d x 90% NR band = 144 kg.d

X 15

« WIMP candidate » Er = 21 keV coincidences bolo-bolo+veto => muon-induced neutrons in fiducial volume

PL B 687 (2010) 294–298 [arXiv:0912.0805]

14 WIMP search : final results

fiducial volume events, 427 kg.d

Five WIMP candidates: - four 20.8 < E < 23.2 keV - one @ 172 keV

« alpha tail »

15 Elastic WIMP scattering limit

 Fiducial exposure 384 kg.d

 4.4 x 10-8 pb at Mχ=85 GeV

 (x2.7 better than 2009 result, 35 x better than 2003)

 Sensitivity limited at low mass due to background

 Backgrounds

arXiv:1103.4070

16 Elastic WIMP scattering limit

 Fiducial exposure 384 kg.d

 4.4 x 10-8 pb at Mχ=85 GeV

 (x2.7 better than 2009 result, 35 x better than 2003)

 Sensitivity limited at low mass due to background

 Backgrounds

arXiv:1103.4070

17 Inelastic dark matter

δ=0  Dark matter modulated signal claimed by DAMA/LIBRA vs. null detection in all the other direct detection experiments

 χ + m => χ* + m (δ~100 keV)

δ=60 1 1 mE v R min c 2 2mE _ _ _ _ R _ _ _ v el min  Signal globally reduced and suppressed at low recoil energies

δ=120  Heavier targets preferred

 Modulation is enhanced

A. Torrento / CEA-Saclay 18 Constraints on inelastic dark matter

arXiv:1103.4070

Ana Sofia Torrento analyis

 same data & analysis as in the elastic case

 use vesc = 544 km/s (RAVE survey, arXiv:0611671, 2007)  DAMA allowed region excluded for Mχ > 90 GeV (90%CL)

19 Elastic WIMP scattering limits : similar sensitivities from CDMS and Edelweiss

 Limits very close with identical nucleus : why not combine data ourselves before others do ?

arXiv:1103.4070

20 Combining Germanium data CDMS+EDELWEISS : the data

 “Simple merger method”

 More sophisticated ones also used in paper

 Quite good spirit and efficiency to complete analysis in few months CDMS+EDELWEISS : the limits

1 TeV: x2 behind Xe100

100 GeV: x3 behind Xe100

Below 70 GeV: >x5 behind Xe100 Edelweiss III : new detectors FID800 (2010)

Fiducial mass ~640g

• All fiducial volume: more Preliminary statistics than stacked ID-400 statistics • No event in NR • Expected to be and indeed better than IDs !

ID (350000 γ)

24 Towards 5x10-9 pb : EDELWEISS-III

 Program under way, funded

 Infrastructure : Upgrades of cabling, cryogenics, acquisition and shielding within the current EDW-II setup  Special care with neutrons : additionnal inner PE shield

 Detectors : ~ 40 FID800 bolometers installed beginning 2012 : 26 kg fiducial ⇒ 3000 kg.d by end 2012

25 Low mass ? Low mass WIMPs ?

 Need very low threshold

 Obtained with a « heat only » detector = naked Ge 330 g crystal with one NTD sensor

 NB : sensitive to full recoil energy (electron/ nuclear)

 Very high sensitivity obtained

 Trigger lower than 1.7 keV 90 % of time

 Select good pulse shapes : risetime/duration 2-8 keV Calibration Data Physical evts Physical evts

Typical raw 2 keV pulses

Jocelyn Domange thesis Noise 27 Data with 1 detector

 Care was taken on control of low energy part of spectrum : no electronic noise kept, correct efficiency for physics pulses

 Quality cuts => 200 days data

 Software cut @ 2keV

 Rising spectrum at low energy but :

 10 keV comogenics Backgrounds from Zn65 and Ge68  Compton : flat Surface β’s  Surface beta’s from 210Pb : from 210Pb rising towards low energy Compton  Releases of stress  WIMPs ? Yes, possible to fit, but can be as well background

28 Data with 1 detector

 Care was taken on control of low energy part of spectrum : no electronic noise kept, correct efficiency for physics pulses

 Quality cuts => 200 days data

 Software cut @ 2keV

 Rising spectrum at low energy but :

 10 keV comogenics Backgrounds from Zn65 and Ge68  Compton : flat Surface β’s  Surface beta’s from 210Pb : from 210Pb rising towards low energy Compton  Releases of stress  WIMPs ? Yes, possible to fit, but can be as well background

29

Start of personal comments Annual modulation : CoGeNT 1106.0650v1

 400 g detector with 0.5 keVee threshold, 425 days, 145 kg.d

 => full spectrum

1106.0650v1 (CoGeNT) 1106.1066v1 (Hooper, Kelso)

1.2 10-40 pb WIMP 7 GeV, v0=250 km/s Vesc= 550 km/s

 Nb : fraction of signal above 0.5 keVee ≈ 10 % Modulation ?

 Modulation of energy integrated rate vs time bins of 30 days shown

 fitted modulation amplitude of 8% (1) and 16 % (2) wrt total rate

 predicted modulation amplitude of 8 % (1) and 6 % (2) wrt total rate

 (Hooper, microMegas*…)

Fitted (1)

 But printed curve for 0.5-3 keV corresponds to 12 % not 6% (2)  Mail exchange on going 2.7σ with J Collar on this

* http://pisrv0.pit.physik.uni-tuebingen.de/darkmatter/micromegas_gexp/index.php Modulation ?

 Modulation of energy integrated rate vs time bins of 30 days shown

 fitted modulation amplitude of 8% (1) and 16+-5 % (2) wrt total rate

 predicted modulation amplitude of 8 % (1) and 6 % (2) wrt total rate

Plain curve (WIMP as before) 0.64 /kg.d

Fitted (1) 0.4

(2) 2.7σ 0.5 0.9 3

 Good curve is the blue

 Most of modulation is in a region with little expected signal Allowed σ vs M from full spectrum and anticipated modulation fractions in both E windows

 D.Hooper 1106.1066v1

8% 6%

0.5-0.9 keV 0.5-3.0 keV Predicted modulations (% of total rate) Pulling ropes 1

 Changing v0 and vesc to adapt to data modul fractions 8 and 16 %

=> 0.5-3.0 kev 10 %  but 16 %

0.5-3.0 kev 0.5-0.9 kev Pulling ropes 2

 Compatibility with DAMA => choose other v0/vesc + increase QNa. Who cares about QNa ? => Mod frac ≈ 5-6 %

0.5-3.0 kev => Mod frac ≈ 7-8 %

0.5-0.9 kev Pulling ropes 3 : Na “quenching factor” in NaI(Tl)

0.45

0.40

NaI Saclay paper (1999) Astr Phys, 11 (1999) 282 0.25

Chagani et al. (2006) Proceedings IDM 2006 arXiv : 0611156v1

 0.4-0.45 looks incompatible with measurements

End of personal comments Back to Edelweiss : going to lower thresholds

 Goal : have 2 parameters to separate nuclear/electron recoils and/or surface/volume events at 2 keVr

 R&D on going at CSNSM-Orsay to reach few 100 ev thresholds on ionisation and heat (30µe/√Hz)

 Ionisation : working on cooper pair box read out, qubit device

 Heat sensors : supraconductor resistive “meander”

 “Quick” test : large polarisation of ionisation electrodes to benefit from Luke effect => “ionisation” only detector Towards 10-10 pb : EURECA

 Dark matter experiment, to search for WIMP interactions to σ~10-10pb

 (~1 event/tonne/year)

 CRESST and EDELWEISS, ROSEBUD additional groups

 Cryogenic (<100mK) calorimeters

 Multiple target materials: Ge, CaWO4, ZnWO4….

 2 phases Phase 1 :150 kg : 2015 Phase 2 : 1 t : 2018 ULISSE : LSM extension Timeline: ULISSE 2010/2011: Design Study  TDR 30 000 m3 2011/12: Digging out of LSM. Tunneling machine starts sep11@ 20m/d. Begin construction of EURECA components away from LSM. Aim for Actual lab ~150kg stage (10−9 pb).

2014: LSM extension ready to receive EURECA. Excavated safety galery 2015: Begin data taking and in parallel improve and upgrade. Actual road tunnel 2018: One tonne target installed.

 Common work with Super CDMS/GeoDM ongoing

 New lab available for new projects

 Funding decision of ULISSE by end july

13 m

15 m Summary & outlook

 Important progress with cryogenic Ge detectors in Edelweiss

 Confortable phase space for improvements and larger mass

 Versatility for adaptation to lower thresholds

 Hopefully new data in coming year to contribute clarifying actual situation at low mass

 Ton scale experiment EURECA in sight

 Enlarging “cryogenic” collaboration ongoing, towards larger integration ?