1. CRESST Cryogenic Rare Event Search with Superconducting Thermometers
Max-Planck-Institut für Physik
University of Oxford
Technische Universität München
Laboratori Nazionali del Gran Sasso
Universität Tübingen
Cryogenic Dark Matter search
Located in Hall A of LNGS
Scintillating CaWO4 target crystals
Up to 33 crystals in modular structure (10 kg target mass)
Different materials possible

2. Which recoil nucleus is seen in CaWO4
assuming
σ ∝ A2
Threshold 10keV
For small WIMP masses <10 GeV only O recoils above threshold
Ca is important around 10 GeV
For large masses W dominates due to σ∝A2

3. CRESST set-up at LNGS shielding: underground laboratory
45 cm PE (12 tons)
muon-veto
radon box
20 cm lead (24 tons)
14 cm copper (10 tons)
use only radio-pure materials

4. Coldbox closed
Half Cu/Pb schield closed

5. CRESST type cryogenic Detectors
Resistance [m]
normal- conducting
super- conducting T R
heat bath
thermal link
thermometer (W-film)‏
absorber crystal
SQUID based read out circuit Width of transition: ~1mK, keV signals: few μK Longterm stablity: ~ μK
Temperature pulse
Advantages of technique:
- measures deposited energy independent of interaction type - Very low energy threshold and excellent energy resolution - Many different target materials E

6. 300g scintillaing CaWO4 crystal
CRESST-II Detectors
Discrimination of nuclear recoils from radioactive backgrounds by simultaneous measurement of phonons and scintillation light
Light detector
W sensor
300g scintillaing CaWO4 crystal
β+γ α O W
Light reflector
(scintillating)
W sensor
Identification of recoiling nucleus possible

7. 300 g CRESST-II Detector Module
The phonon detector: 300 g cylindrical CaWO4 crystal. Evaporated tungsten thermometer with attached heater.
Clamps not
scintillating
The light detector: Ø=40 mm silicon on sapphire wafer. Tungsten thermometer with attached aluminum phonon collectors and thermal link. Part of thermal link used as heater
CRESST-II: up to 33 detector modules

8. Spectral features at low Energies
Cu Kα 8.1 keV
keV
41Ca 3.61 keV keV
210Pb 46.5 keV
@ 46.5 keV
Very precise energy calibration
Lines down to 3.6 keV identified with excellent energy resolution of 300 eV.

9. Present run All results preliminary running since summer 2009
10 detectors running (1 ZnWO4)
Clamps not covered with scintillator
data analysis is still in progress
Data discussed are from 9 CaWO4 detectors ( about 400 kgd)

10. Stability

11. Excellent nuclear recoil discrimination
neutron calibration
γ + β band
O-recoil-band
Lines calculated from e-resolution of light detector and known quenching factor
Excellent nuclear recoil discrimination

12. Data α α γ + β band α band O-recoil-band W-recoil-band Pb Pb
clamp not scint.
210Po Pb (104 keV) + α (5.4 Me)‏ Pb α
crystal

13. What are these events in O-band ?
Detector
E0.1[keV]
events 5
12.35 20
11.85 2 29
11.65 4 33
15.55 43 45
19.15 47
17.35 51
9.65 6 55
22.25 3
total 32
Try to estimate background
Neutrons ?
α leakage ?
Low mass WIMPs ?
Check for coincidences

14. Neutrons ? No double coincidences
2 tripple coincidences : Orecoil + Orecoil + 1.8MeV gamma
Orecoil + 30 KeV gamma MeV gamma
O+O coincidences can only come from neutrons
When there are coincidences there are also single O recoils
Goal: Estimate single oxygen recoils from observed coincidences.
Ratio of coincidence/single depends on the source of neutrons.
Neutrons from sf or alpha-n
mostly n-n, few n-gamma
very few triple coincidences and none with MeV gammas
unlike observed 2 events
Neutrons induced by muons in Pb/Cu shield
shower like events with high multiplicity
and high energy gammas,
similar to observed 2 events.

15. Multiplicity of oxygen recoils in signal region
Multiplicity with neutron source
Multiplicity from muon induced neutrons
Only very view oxygen recoils seen in coincidence in 3 detectors, no high energy gammas involved
Oxygen recoils seen in coincidence in a large number of detector modules together with high energy gammas in others. Similar to the observed events
Coincident/single = 17/8 = 2.1 ±0.9
Single neutron background of 0.94±0.83 events in O band

16. Degraded alphas from external contamination in clamps
Discrete alpha lines from contamination in crystal are no problem
Degraded alphas with continuous energy distribution down to lowest energies from external contamination in clamps

17. Estimation of α background in oxygen band
Detector with highest external alpha background
Also highest 206Pb recoil background, with long tail extending from 100 keV down to 40 keV
Alpha band
Oxygen band
Signal region
206Pb recoils
Background of degraded external α's from contaminated clamps.
Oxygen and α band partially overlap and some α's may leak into signal band.
Estimate dN/dE in overlap free region of alpha band and then compute
expectation in oxygen band assuming constant dN/dE.

18. Estimation of Background of degraded alphas
Detector
Nfound
Nproj 5 4
1.05 20 2
0.51 29
1.02 33 3
0.76 43
1.06 45
0.55 47
0.53 51
0.98 55
0.46
total
6.93
Nfound: Number of alpha events found in reference region of alpha band
Nproj: The projected number of alpha events in signal region of the oxygen recoil band
Estimate of α background in signal energy range of oxygen band: 6.93 counts

19. Summary of Background Estimates
Neutron background
Alpha background:
Leakage from gamma band:
Sum of background estimates:
Observed oxygen recoil signals :
background estimate of 8.7 ± 1.4 not enough to explain observed 32 signals,
leaving space for a light WIMP
m about 15 GeV or less
σ about a few times 10-5pb

20. Inelastic Dark Matter
preliminary
preliminary
preliminary
preliminary

21. Conclusions
CRESST detectors are very powerful and able to perform precision measurements
Inelastic Dark Matter scenario becomes very unlikely to explain the DAMA result
Neutron background is negligible and can not explain our signals in oxygen band
Background from degraded alphas is less then observed signals in oxygen band, a precise estimate is difficult.
If the alpha estimate is roughly O.K, there is space for a light WIMP
A new run with strongly reduced alpha background is the next step. It should help to pin down the nature of the observed signals with high confidence.

22. Thank You

23. Detector Performance in wide energy range
Enormous dynamic range
Excellent linearity and energy resolution in whole energy range
Perfect discrimination of β and γ from α’s
Identification of alpha peaks from emitters in crystal
β+γ
α peaks
Nuclear recoils from neutrons or WIMPs

24. Identification of a WIMP-signal
The trouble starts if you see a signal
Annual modulation
you have to prove that your background is constant in time !
and your detector runs stable
Use different target nuclei in same detector
cross section for background depends differently on target nucleus (mass) than WIMP-scattering
unique feature of CRESST Detectors