Dark Matter search with bolometric detectors PhD Student: Filippo Orio Dottorato in Fisica XXIII Ciclo Supervisor: prof. Fernando Ferroni Seminario di Dottorato - 4 giugno 29 1
Outline Introduction to Dark Matter Direct Dark Matter search The CUORE experiment CUORE background Trigger studies in CUORE Conclusions and perspectives 2
Evidences for Dark Matter 1933 Fritz Zwicky 197s Vera Rubin Observation Prediction Applying virial theorem to study of Coma cluster, he concluded that mass of galaxies in cluster was O(1 2 ) what inferred from luminosity Studying spiral galaxies rotational curves, she found inconsistencies between observations and newtonian predictions for orbital velocities 3
Evidences for Dark Matter 26 Bullet Cluster The stars of the galaxies, observable in visible light, were not greatly affected by the collision, and most passed right through, gravitationally slowed but not otherwise altered. The hot gas, seen in X-rays, represents most of the mass of the ordinary (baryonic) matter in the cluster pair. The gases interact electromagnetically, causing the gases of both clusters to slow much more than the stars. The third component, the dark matter, was detected indirectly by the gravitational lensing of background objects. The lensing is stronger in two separated regions near the visible galaxies. This provides support for the idea that most of the mass in the cluster pair is in the form of collisionless dark matter. 4
What is Dark Matter? Different astronomical and cosmological studies reveal that most of our Universe is dark Dark Matter represents the 22% of the whole energy and the 85% of the total mass but... What is Dark Matter made of? Many different candidates Data favor cold non-baryonic Dark Matter Non-relativistic velocities Gravity and weak interactions Stable Weak Interacting Massive Particles (WIMPs) are the most likely candidates 5
How to detect a WIMP Simple idea: detect matter recoil after WIMP elastic scattering χ χ Signal: χ+n χ+n n n Backgrounds: n+n n+n γ γ ν+n ν+n γ+e - γ+e - Very low energies (1-1 kev) Many backgrounds Many different techniques Many different target materials Many different experiments N N +α, e - 6
Present searches Cross-section [cm 2 ] (normalised to nucleon) 1-41 1-42 1-43 1-44 1-45 1-46 9526135 WIMP Mass [GeV/c 2 ] http://dmtools.brown.edu/ Gaitskell,Mandic,Filippini 1 1 1 2 1 3 Several assumptions and modeling required Theoretical and experimental uncertanties Exclusion plots can only exclude DM models x x x DATA listed top to bottom on plot Edelweiss I final limit, 62 kg-days Ge 2+22+23 limit WARP 2.3L, 96.5 kg-days 55 kev threshold CRESST 27 6 kg-day CaWO4 ZEPLIN III (Dec 28) result CDMS: 24+25 (reanalysis) +28 Ge XENON1 27 (Net 136 kg-d) Trotta et al 28, CMSSM Bayesian: 68% contour Trotta et al 28, CMSSM Bayesian: 95% contour Ellis et. al Theory region post-lep benchmark points Baltz and Gondolo 23 Baltz and Gondolo, 24, Markov Chain Monte Carlos 9526135 NO POTENTIALITY OF DISCOVERY 7
The annual modulation Since the Earth rotates around the Sun and the Sun itself moves into the galactic reference frame, we can express Earth's speed with respect to the Galaxy as the sum of the two motions γ= In this framework, the expected WIMP signal in the Earth reference system can be written (first order Taylor approximation) 8
The DAMA claim Highly radiopure NaI scintillators @LNGS MODEL INDEPENDENT 9
The CUORE experiment Proposed for neutrinoless Double Beta Decay search 988 TeO2 crystals (5x5x5 cm3) for a total mass of 741 kg Crystals operate as bolometers at 1 mk Well known technique, used for CUORICINO (stopped in June 28) 19 towers 52 detectors each CUORE- (the first tower of CUORE) will start in 21 Hosted at Laboratori Nazionali del Gran Sasso, a natural shield of 15 m of rock (35 meters of water equivalent) 1
The bolometric technique Heat bath Weak thermal coupling Thermometer: NTD Ge thermistor TeO2 Absorber C~1-9 J/K Energy release All the particle energy is converted into phonons Temperature variation: ΔT=E/C Resistance variation Voltage variation Detector response: ~.2 mk/mev ~ 3 MΩ/MeV ~.2 µv/mev FWHM Resolution ~ 5 kev 11
From V to kev Voltage (mv) 34 32 3 28 26 Amplified 24 signal 22 2 18 Estimate pulse amplitude 16 1 2 3 4 5 Time (s) Energy calibration is performed in specific runs in which a known radioactive source (usually thorium) is inserted in the cryostat. The amplitude spectrum becomes populated by well known monochromatic lines counts 5 Histogram3 Entries 7819 Identify Mean 43.6 4 RMS 359.4 known lines counts 5 4 Histogram3 Entries 7821 Mean 734.8 RMS 659.5 Calibrate 3 3 2 2 1 1 2 4 6 8 1 12 14 Amplitude [mv] 5 1 15 2 25 3 Energy [kev] 12
WIMP nuclear recoils in TeO2 Using cross section quoted by DAMA, we can calculate the expected rate of WIMP nuclear recoils and plot the annual modulation amplitude [cpd/kev/kg] max (!dr/de R ).3.25.2.15.1.5 3 " =.3 GeV/cm w M w = 4 GeV M w = 6 GeV M = 8 GeV = 1 GeV w TeO 2 M w The region of interest is under 1 kev -.5 1 2 3 4 5 6 7 8 9 1 [kev] E R 13
What we need Low background at energies of kevs High efficiency on signal identification High background rejection Low radioactive background Smart trigger 14
The test detector Chinese Crystal Validation Run 4 crystals CUORE-like 2 thermistors per crystal Operated in Hall C @LNGS Live time of 64.9 days cpd/kev/kg CCVR1 Low Energy Spectrum (ch 1, 3, 5, 6) 1 1 1 1 1 2 Many radioactive peaks 1 2 3 4 5 6 Energy [kev] 15
Tellurium isotopes contributions CCVR crystals contain large amount of Te long-life isotopes Nuclide Z N Decay mode Half Life (d) Ex (kev) 121 Te 52 69 e+b + 16.78±.35 121m Te 52 69 IT, e+b + 154±7 293.98±.3 123m Te 52 71 IT 119.7±1 247.55±.4 125m Te 52 73 IT 57.4±.15 144.795±.1 127m Te 52 75 IT, b - 19±2 88.26±.8 129m Te 52 77 IT, b - 33.6±.1 15.5±5 cpd/kev/kg 125m Event rate in 145 kev line of Te (ch 1,3,5,6) 8 7 6 5 4 3 2 1 2! / ndf 27.93 / 24 Base 1.84 ±. Ampl 4.51 ±.8 " 1/2 62.74 ± 2.59 1 2 3 4 5 6 7 days Fit on CCVR data (ch 1, 3, 5, 6) Using GEANT4, we can simulate N decays for each isotope. Then we normalize each contribution with the CCVR data counts 5 4 3 2 1 12 13 14 15 16 17 18 Energy [kev] 16
Simulations and clean spectrum 12x Te simulations (normalized with data) CCVR Bkg subtracted spectrum (ch 1,3, 5, 6) cpd/kev/kg 1 Sum of all the simulated 12x Te contributions cpd/kev/kg 1 1 1 1 1 2 1 2 3 4 5 6 Energy [kev] 1 1 Trigger threshold 1 2 3 4 5 6 Energy [kev] CCVR rate under 5 kev is almost 1cpd/keV/kg. It includes X-rays (cannot be removed) and some other fake signals which could be rejected by a pulse shape analysis At lower energies, however, trigger threshold effect is evident We need to lower the threshold and look at background again 17
The key role of trigger Typical noise values are of the order of few kev A signal of thousands of kev over a noise of few kev is easy to detect If the signal to detect has an amplitude of the same order of the noise, the detection is quite harder We are sensitive in 13 Te νdbd region (253 kev) We need a trigger able to detect very low energy signals, distinguishing them from noise 18
MonteCarlo Pulse Generation Voltage (mv) 161 168 166 164 162 5 kev Actually we are not able to distinguish very low energy signals from background Voltage (mv) 16 1 2 3 4 5 Time (s) 1635 163 1625 162 1615 161 165 4 kev 16 1 2 3 4 5 Time (s) In order to study trigger efficiencies, we use a Pulse Generator Starting from thermal model considerations, we can generate pulses of known amplitude and superimpose real noise 19
CCVR Standard trigger Voltage (mv) 1714 Derivative Trigger 1712 171 178 176 174 172 A rod with fixed x-length ( AVERAGE ), whose terminal points are samples of the acquisition. If the angular coefficient of the rod is above a fixed value ( THRESHOLD ) for a fixed time ( DEBOUNCE ), the trigger fires. If not, the rod is shifted forward. 17 1698 AVERAGE THRESHOLD 1696 1 2 3 4 5 Time (s) 2
Efficiencies and fake probability Fake probability (%) Fake (%) Trigger Threshold Derivative Energy Threshold in kev (95% signal efficiency) Real Threshold (sig.eff 95%) Trigger Threshold Derivative Ch. 4 8 1 12 14 16 18 2 22 Ch. 4 8 1 12 14 16 18 2 22 16 269 67 13 2 16 11 14 17 19 21 25 27 29 Trigger Debounce Time 24 32 4 48 87 9 13 1 Trigger Debounce Time 24 32 4 48 13 14 17 2 16 17 18 22 18 19 21 25 21 22 25 29 22 25 28 33 25 27 29 37 29 29 34 4 31 32 37 43 56 56 24 28 32 36 41 43 49 5 Varying Debounce and Threshold (Average is fixed at 4 ms) we can test Derivative Trigger on the same simulated low energy pulses. We evaluated the energy threshold (the lowest energy where signal efficiency is at least 95%) and the probability to have a fake trigger in a window of 5 s. Imposing null fake probability Channel Energy Threshold (kev) 1 23 2 32 3 9 4 17 5 13 6 16 7 13 21
What we have done We calculated the expected modulation signal in CUORE, according according to DAMA to DAMA measurement claim We studied CCVR radioactive background and estimated an event rate of about 1 cpd/kev/kg in 2-5 kev region We studied Derivative Trigger performances on simulations and found its lowest possible thresholds (still too high) 22
What we are doing Using external heaters (resistances glued to the crystals, used for offline temperature stabilization), we are studying Derivative Trigger efficiencies on real data We are studying heater behaviour at low energies (linearity, characterization, secondary effects due to positioning and gluing...) We are studying pulse shape parameters in order to apply selection criteria to reject background We are involved in assembling, commissioning and analyzing Chinese Crystal Validation Run 2 (4 crystals) and another prototype with 36 crystals 23
What we will do Build more performing trigger, based on signal online filtering Evaluate expected event rate under 1 kev Evaluate CUORE sensitivity to Dark Matter Annual Modulation 24