The Cryogenic Dark Matter Search: First 5Tower Data and Improved Understanding of Ionization Collect

1
The Cryogenic Dark Matter Search First 5-Tower
Data and Improved Understanding of Ionization
Collection

• Catherine Bailey
• PhD Thesis Defense
• July 14, 2009

2
Outline

• Dark Matter and WIMPs
• CDMS
• 5-Tower Analysis and Results
• Ionization Collection Studies

3
Standard Cosmology
Supernova Cosmology Project
NASA/WMAP Science Team 2006
4
Evidence for Dark Matter
Gravitational Lensing
Spiral galaxy rotation curves
5
Big Bang Nucleosynthesis non-baryonic dark
matter

• Big Bang Nucleosynthesis
• Constrain baryon density based on relative
  abundance of light elements from hot big bang
• One-parameter model baryon density
• Best constraint D/H in primordial gas clouds
  (Burles Tytler) ?Baryons 0.05 0.005
• 4 of atoms (in pie chart)

theory
measurements
concordance
theory
6
Weakly Interacting Massive Particles (WIMPs)

• Annihilation stops when number density drops to
  the point that annihilation too slow to keep up
  with Hubble expansion
• Leaves a relic abundance
• ?ch2 ??10-27 cm3 s-1 ????A v??fr

exp(-m/T)
if mc and ?A determined by physics at electroweak
scale, then ?c? 0.3
7
WIMPs in the Galactic Halo
WIMPs the source of Mass in the Rotation Curves?
Ge
Scatter from a Nucleus in a Terrestrial Particle
Detector
8
WIMP Direct Detection Challenges

• Weakly interacting
• Low recoil energies 10 keV
• Low event energy thresholds
• Rates less than 0.01 events per day in a kilogram
  detector
• Large masses
• Long term stability
• Backgrounds much higher than event rate
• Background control (cleanliness, shielding )
• Underground site (reduce muon induced neutrons)
• Electron / Nuclear recoil event discrimination

9
WIMP Direct Detection Techniques
10
CDMS in a nutshell
Use a combination of discrimination and shielding
to maintain a zero background experiment with
low temperature semiconductor detectors

• Discrimination
• Phonons
• energy measure
• pulse shape
• Ionization
• dE/dx discrimination
• Shielding
• Passive (Pb, poly, depth)
• Active (muon veto shield)

11
Soudan Underground Lab
12
Soudan Installation
vacuum-sealed electronic connector box (E-box)
Plastic scintillator
detector cold volume (icebox)
lead
RF shielded class 10,000 clean room
Oxford Instruments 400µW dilution refrigerator
1 ft3 _at_40 mK!
outer polyethylene
0.05 unvetoed neutrons per kg-y (Monte Carlo)
ancient lead
inner polyethylene
13
ZIP Detectors
(Z-sensitive Ionization and Phonon)
Phonon side 4 quadrants of athermal phonon
sensors gt energy measurement
3 (7.6 cm)
Charge side 2 concentric electrodes
1 cm
Ge 250 g
Operated at 40 milliKelvin for good phonon
signal-to-noise
14
Anatomy of an event
Hot charge carriers (3eV/pair)
-3V
h
e-
Quasi-diffusive THz phonons
0V
Ballistic Neganov-Luke phonons
Ballistic low-frequency phonons
15
ZIP Detectors Ionization
Q inner Q outer
Essentially complete collection at 3V/cm (after
trap neutralization) Low-noise JFET amp at 140
K Zero-energy resolution 250 eV(3 keV _at_ 511
keV)
15
85
Fiducial volume cut from divided electrode
(guard ring)
16
ZIP Detectors Phonons
380x60µm Al fins
250x1µm W TES
Tungsten Transition Edge Sensor (TES)
4 SQUID readout channels, each 1036 W TESs in
parallel Zero-energy resolution 100 eV in each
channel, total 5 at higher energies (after
position correction)

Fast response 5 µs risetime
17
Ionization Yield
ER background
Echarge / Ephonon
NR signal
Ephonon
18
Ionization Yield
Density Plot with 50,000 events
Photons from external source
Ionization Yield
Neutrons from external source
Recoil Energy keV

• Reject bulk electron recoils using ionization
  yield
• Better than 10,0001 rejection

Ionization Yield Ionization / recoil energy
19
Near-Surface Events
Reduced charge yield from surface events (e.g.
K-40, Rn chain) from carrier back-diffusion can
mimic signal Greatly improved by aSi contact
(Shutt et al.), still dominant background for CDMS
10µm dead layer
Data from UC Berkeley calibration of T2Z5, née
G31 V. Mandic et al., NIM A 520, 171 (2004)
20
ZIP Detectors Z-sensitivity
10µm dead layer

• Primary risetime(time from 10 - 40 in phonon
  amplitude for largest pulse)
• Primary delay(time from 20 charge amplitude to
  20 phonon amplitude for largest pulse)

Surface Bulk
21
Discrimination Ionization Yield Phonon Timing

• Ionization Yield
• Ionization / recoil energy
• Reject bulk electron recoils
• Better than 10,0001 rejection
• Phonon Timing
• Faster phonon signals from surface events (more
  ballistic phonons)
• Rejects surface electron recoils gt2001

Calibration Data
Bulk Electron Recoils (background)
Surface Electron Recoils (background)
Nuclear Recoils (signal)
22
ZIP Tower
FET
SQUID
Detectors
23
First Five Tower Runs (2006-7)

• 30 ZIPs (5 Towers) in Soudan icebox
• 4.75 kg Ge, 1.1 kg Si
• First two runs analyzed
• Run 123 (21Oct06 21Mar07)
• 108 live days (430 kg-d Ge raw)
• Run 124 (20Apr07 16Jul07)
• 56 live days (224 kg-d Ge raw)

24
Analysis Cuts
muon

• Blind analysis
• Data Quality
• Physics
• Veto-anticoincidence
• Single-scatter
• Qinner (fiducial volume)
• Ionization yield
• Phonon timing

Q outer
Q inner
25
The WIMP Search DataBlind Ge Analysis
Prior to Phonon Timing Cut
All cuts set and frozen! Predict 77 ? 15 single
scatters in NR
26
The WIMP Search Data
Midnight PST, 4th of Feb, 2008
Prior to Phonon Timing Cut
97 Singles in Signal region rejected by Surface
Event Cut
27
Open The Box
Midnight PST, 4th of Feb, 2008
0 Observed Events
After Phonon Timing Cut
Expected Background 0.6 ? 0.5 surface events and
lt 0.2 neutrons
28
Results

• Strongest SI limits above 44 GeV/c2
• 4x data in hand
• Ongoing detector characterization and analysis
  optimization
• New detectors currently being installed
  commissioned

Phys. Rev. Lett. 102, 011301 (2009)
29
SuperCDMS

• New detectors
• Improved background rejection (due to larger
  volume to surface ratio)
• 2.5 times more mass per detector readout channel
  (1-inch thick instead of 1-cm thick)
• Improved discrimination through phonon timing
  (due to new phonon collection system)

30
Poor Ionization Collection
Density Plot with 50,000 events

• Surface charge trapping
• Phonon timing
• Bulk charge trapping
• Neutralization
• CDMS increasing sensitivity to WIMP interactions
  by increasing detector mass
• To remain background free, need to improve
  background rejection

Ionization Yield
Recoil Energy keV
Individual surface events (many have low yield
due to poor ionization collection)
31
Surface Trapping
Ionization Yield Ionization / recoil energy
Ionization Side 109Cd Source
Phonon Side 109Cd Source
gammas
ER
ER
Ionization Yield
Ionization Yield
109Cd surface events leak into signal region
109Cd surface events separate from signal region
betas
NR
NR
neutrons
Phonon Energy (keV)
Phonon Energy (keV)

• Ionization Yield discrimination is worse for
  phonon side events than ionization side events
• Phonon timing provides rejection of phonon side
  events
• Goal improve phonon side yield discrimination
  for SuperCDMS

32
Decrease Dead LayerHydrogenation of
amorphous-Si contacts

• Previous work showed promising results
• Germanium test devices w/ different surface
  treatments studied
• Lowest number of low yield events / smallest
  surface layer seen in device w/ nominal surfaces
  (no H-aSi)

Si / Ge substrate
aSi
Al
?Esubstrate
?EaSi
33
Decrease Dead LayerHydrogenation of
amorphous-Si contacts
Dead layer calculated from 241Am 60-keV events
Dead layer (mm)
Better Ionization Yield Rejection
G3D
G28B
34
Dead Layer Differencesfinding the culprit?
Dead layer calculated from 241Am 60-keV events
Ionization side sources
Ionization Side - Iron Ion Implanted (test)
Dead layer (mm)
Better Ionization Yield Rejection
Ionization Side -Normal Processing (control)
Ionization Side Iron Ion Implanted test
Normal Ionization Side processing control
-3V Charge Bias 3V

• Iron Ion Implantation used on phonon side of
  detector to create uniform W TES Tc across entire
  detector
• Induced lattice defects source of yield
  discrimination asymmetry? Possibly.

35
Reducing Ionization Yield Asymmetry Charge
Phonon Side Surface Events

• Selective Iron Ion Implantation across phonon
  side of detector to reduce lattice defect
  destruction
• New Tungsten Target in sputtering deposition
  system may allow reduced implantation to achieve
  target Tc

36
Bulk Trapping
Test facility data taken with an internal,
collimated 241Am source
60 keV gamma rays
Charge collected
Charge collected
time
time
Bulk charge trapping by lattice impurities
37
Solution to Bulk TrappingNeutralization

• Neutralize these impurities to maintain full
  charge collection (Neutral impurity sites have
  lower trapping cross sections than ionized sites)
• LED photons
• Energetic particles (either from radioactive
  sources or surface backgrounds)

Charge collected
time
38
Bulk TrappingEffects on Data Analysis
Ba calibration for comparison

• Soudan 2-tower data analysis resulted in one
  candidate event
• However, in data series exhibiting poor
  neutralization

Ionization Yield
Automatic LED flash to clear charge traps in
detectors
Run 119 candidate
time (hours)
39
Quantitative Monitoring of Detector
Neutralization State

• For Soudan 5-tower data analysis, quantitative
  figure-of-merit of neutralization
• Using this fraction of low yield events
• Developed neutralization cut
• Real-time monitoring
• Optimized neutralization process

Neutralization loss
40
Conclusions

• Worlds best limits on WIMP dark matter
  properties above 44 GeV/c2
• Data already taken and currently undergoing
  analysis to improve limits by 4x
• Surface trapping studies leading to modifications
  in future detector fabrication
• Bulk trapping studies led to analysis cuts,
  operational optimizations, improved data quality

41
(No Transcript)