Surface events suppression in the germanium bolometers EDELWEISS experiment Xavier-François Navick (CEA Dapnia) TAUP Sendai September 07

Ionization-heat detectors Biais amplifier Charge amplifier Thermal Sensor (NTD Ge) Al electrodes (charge collection ) HPGe single crystal WIMP neutron X or   rays D = 2 cm Al electrode V = 1-6 Volts T = 20 mK TAUP2007 X-F Navick September 2007

 Simultaneous measurement of ionization and heat => Evt per evt identification  Q = Eionisation / Erecul  Q = 1 for electronic recoil (ambiant radioactivity)  Q  0.3 for nuclear recoil (WIMP and neutron)  discrimination  /n > 99.9% pour Er> 15keV Rejection power TAUP2007 X-F Navick September 2007

Developed by CEA Saclay and Canberra Optimized NTD size in collaboration with LBNL for sub keV resolution New holder and connectors (Teflon and copper only) Edelweiss-II Ge/NTD detectors TAUP2007 X-F Navick September 2007

Incomplete charge collection Amorphous layers (Ge or Si) improve the charge collection At very low T and V : electric field is screened => diffusion phase Some carriers are trapped in the “wrong” electrode => incomplete collection TAUP2007 X-F Navick September 2007

Effect of the amorphous layers TAUP2007 X-F Navick September 2007

2003 I Edelweiss-I data with phonon trigger TAUP2007 X-F Navick September 2007

E = 5.3 MeV, Q = 0.3  ’s from 210 Po (E  =5.3 MeV) Q=0.3   decays near surfaces Rate ~ 400 /m²/d As expected, non-fiducial part more exposed 210 Pb on Cu covers or Ge surfaces Should see Pb recoils and  ’s No 206 Pb recoil peak at 100 keV observed as heat-only events: 210 Pb implanted in Cu, not Ge. Rate of 0.3 < Q < 1.0 events at low energy consistent with expected surface  ’s does not exclude contribution from 14 C By removing Cu between detectors, these events should disappear, or ID by coincidences Surface contamination TAUP2007 X-F Navick September 2007

Edelweiss-II first data taking TAUP2007 X-F Navick September 2007

Edelweiss-II first data taking TAUP2007 X-F Navick September keV recoils, no ionization 5.3 MeV alpha

 Initial effort : improve the cupper treatment (Rn contamination study) and reduce the surface exposure to radon  Passive rejection : improve the charge collection for surface event  Physics of the Ge and Si amorphous underlayer  Detectors with thick electrodes  Active rejection : identification of the surface events  Interleaved electrodes localization of the event  Pulse shape analysis of the charge signals  Detectors sensitive to athermal phonons  Ge/NbSi detectors Surface events suppression TAUP2007 X-F Navick September 2007

Passive protection Amorphous layers: Al / aGe:H / cGe / aGe:H / Al => GGA Thick electrodes (high entrance window) 20mK E Al aGe:H GeAl aGe:H TAUP2007 X-F Navick September 2007

electrodes NTD Ge thermometer guard ring Top view Cross-section 24 mm Interleaved electrodes design TAUP2007 X-F Navick September g Ge crystal - Hydrogenated a-Ge underlayer for improved charge collection - Annular aluminum electrodes, interconnected by ultrasonic bonding (strip width: 200 mm; pitch: 2 mm) - 7 measurement channels: 6 charge (7 MHz bandwidth) + heat (Ge NTD)

Mode of operation TAUP2007 X-F Navick September 2007 Bulk events: Q a = Q c = 0 Q b = - Q d Surface events (top surface): Q c = Q d = 0 Q a  0 & Q b  0 (bottom surface): Q a = Q b = 0 Q c  0 & Q d  0 * Voltage biases: V a = 1V, V b = 2V, V c = -1V, V d = -2V, V g = 0.5V, V h = -0.5V Type III: low-field area bulk event channel d (V d ) channel c (V c ) guard g (V g ) channel a (V a ) channel b (V b ) near- surface event low- field area guard h (V h ) NTD Ge thermometer

Rejection of the surface event TAUP2007 X-F Navick September 2007 Voltage biases: V a = -0.25V, V b = 2V, V c = 0.25V, V d = - 2V, V g = 0.5V, V h = - 0.5V Trigger threshold in ionization: 12 keV (e.e.) Baseline energy resolution of the ionization channels: 1 keV (e.e.) Heat channel resolution: 4 keV (FWHM) T = 17 mK 241Am  source Cuts: |Q a | > 2 keV & |Q b + Q d | | > 2 keV (e.e.)  event rejected Ionisation yield E recoil (keV) Ionisation yield before selection: 3596 evtsafter selection: 1134 evts Events of incomplete charge collection

Reduction of the fiducial volume TAUP2007 X-F Navick September Am  source Cf neutron source E recoil (keV) Ionisation yield 38 evts32 evts before selection: 3472 evtsafter selection: 1058 evts Loss in fiducial volume: (38-32) / 38 ~ 15 %

Near-surface, 122 keV  event: centre electrode signal in red guard electrode signal in blue best fit simulation in black Experimental setup (-1 V collection voltage) 1000 ns  Event location Due to electronic noise, surface event rejection by pulse-shape analysis is limited in practice to high-energy events only (  50 keV e.e.) Test at LSM on a 320g Edw-I detector TAUP2007 X-F Navick September 2007 Localization by ionization shape analysis

Surface event discrimination: Nb-Si thin films as athermal phonon sensors developed in CSNSM in Orsay Sensor A Sensor B Bulk event Surface event Signal amplitude (a.u.) Thermal component (energy-sensitive only) Athermal phonon (position-sensitive) component of signals sensor A’s response sensor B’s response Responses of sensors A and B Ge detectors with NbSi sensors Time (s) TAUP2007 X-F Navick September 2007 See Claudia Nones’s talk Surface event discrimination: Nb-Si thin films as athermal phonon sensors developed in CSNSM in Orsay Sensor B Sensor A

(a) Ionization yield vs. recoil energy (thermal component of heat signals only): note the large population of surface events of poor charge collection 57 Co  source (b) With the proper cuts in the athermal component of phonon signals to remove surface events Ionisation yield (a.u.) Recoil energy (keV) 200g Ge detector with NbSi sensors in the Edw-I set-up at LSM TAUP2007 X-F Navick September 2007

Summary TAUP2007 X-F Navick September 2007 The surface events appears to limit the sensitivity of ionization-heat germanium bolometers to WIMP. Different techniques are under development in Edelweiss to reduce or reject these events. The most promising ones lead to an active discrimination by athermal phonon measurement with NbSi thin film or by charge collection on interleaved electrodes. We are starting to apply these techniques in LSM with prototype detectors. In the next step of Edelweiss-II we will produce massive bolometers with active rejection.