Direct Detection of Dark Matter and the CDMS II Experiment

1
Direct Detection of Dark Matter and the CDMS II
Experiment
Mike Attisha
2
Introduction

• Galactic rotation curves imply the presence of
  unseen matter
• Outside lum. core expect v r-1/2
• Self-gravitating ball of ideal gas at a uniform
  temperature would have the correct mass profile
• Analysis of galactic clusters (motion, x-ray,
  grav. lensing) give consistent value ?M
  0.2-0.3(i.e. 20-30 of the critical density)
• ?baryonic 0.045 from BBN

3
Introduction

• CMB is a snapshot of the baryonic matter
  distribution at 300,00 years after Big Bang
• Not enough CMB structure to explain the
  large-scale structure we see today
• Cold Dark Matter is required to reconcile CMB
  power spectrum with structure formation

4
Direct Detection Overview
5
DAMA

• Experiment running since 1996
• Located at Gran Sasso Lab, Italy, at a depth of
  4000 mwe
• 9 ? 9.7 kg low-activity NaI scintillator
  crystals, each viewed by 2 PMTs
• Known technology
• Low cost
• Large mass
• 107,000 kg-days exposure through July 2002
• Annual modulation signal 6.3s

6
ZEPLIN I

• Located in Boulby Mine, UK (2800 mwe)
• Xe, Xe- ions created by recoils within 3kg
  liquid Xe fiducial region
• Recombination of ions much quicker for nuclear
  recoils than electron recoils
• Uses mean charge arrival time as discriminant
• Pulse shape analysis gives relatively weak
  discrimination between recoil types

7
EDELWEISS

• Located in the Modane Laboratory, France-Italy
  border
• Located at depth of 4000 mwe
• Uses similar technology to CDMS

8
CDMS II Overview

• Located at the Soudan mine in sunny Minnesota
• CDMS II is 2341 feet below the surface (2090 mwe)

9
Sources of Background

• Gammas / X-Rays
• 15cm Pb reduces photon flux by factor gt 1000
• 25cm Poly reduces muon-induced neutron flux from
  rock and lead by factor gt100

muons
External neutron
Internal neutron

• Electrons
• Produced by contaminants located inside the
  shield rejected via analysis

• Neutrons
• Reject internal neutrons produced by muons
  within the shield by using scintillator veto

• Cosmic Ray Muons
• Depth (2090mwe) reduces muon flux by factor
  50,000
• Scintillator veto tags muon-related events

10
Sources of Background
Detectors must effectively discriminate between
Nuclear Recoils (Neutrons, WIMPs) Electron
Recoils (gammas, betas)
Use Ge and Si based detectors with two-fold
interaction signature - Ionization signal -
Athermal phonon signal
11
ZIP Detector Physics

• Detector event creates
• electron-hole pairs (Q prop. Er)
• THz athermal phonons
• e-h pairs drift under field to electrode
• Neganov-Luke effect creates an additional phonon
  population as charges drift

phonon sensors
3 Volts
charge sensors
Detector event

• Ionization signal for Nuclear Recoils is
  suppressed by factor 3
• Allows us to discriminate between recoil types
  using the ionization yield
• y Q / Er

12
ZIP Detector Setup

• 250 g Ge or 100 g Si crystals arranged into
  Tower of 6 ZIPs
• 1 cm thick x 7.5 cm diameter
• Phonon signal gives rise to quasiparticles in Al
• Lower band gap in W transition-edge sensors (TES)
  causes quasiparticles to become trapped
• Each W sensor fed by 8 Al fins

Ge
Ge
Ge
Si
Ge
Si
13
ZIP Signals - Phonons

• Amplitude of phonon signal given by integrated
  area under pulse
• Risetime (5-20?s) characteristic of event
  physics
• 4 phonon channels provide event position
  information

A
D
B
C
14
ZIP Signals - Charge

• Signal risetime 3-10 ?s
• Outer and inner charge channels

15
Pulse Analysis
Raw Pulses
Baseline Noise
PSD of Raw Pulses
Baseline Noise and Raw Pulse PSD
16
Pulse Analysis

• Optimal Filter method
• Pulse is filtered in frequency space using noise
  traces and a pulse template
• Filtered pulse is transformed back to time to
  obtain amplitude and time offset

17
Signal Parameters - Amplitude/Position
133Ba

• 4 phonon channels allow us to reconstruct event
  location in x and y
• Use radioactive source shining through
  collimator to calibrate position measurement

• Amplitude obtained from Optimal Filter algorithm
• Absolute scaling determined via calibration
  datasets with lines at known energies

18
Signal Parameters - Phonon Risetime

• Risetime defined as time taken to rise from 10
  -gt 40 of the maximum amplitude
• Algorithm finds peak and walks down pulse to
  find times along the leading edge
• Different risetime for each phonon channel
• Minimum risetime gives greatest discrimination

19
Signal Parameters - Phonon Delay

• Delay defined as duration from Q start time to P
  start time
• Different delay for each phonon channel
• Minimum delay gives greatest discrimination

20
Background Discrimination

• Electron recoils interacting in surface layer
  have incomplete charge collection
• Yield is reduced - may be identified as nuclear
  recoil
• Dead layer events generally have a lower phonon
  delay and risetime compared to genuine nuclear
  recoils

252Cf
133Ba
Phonon Delay (µs)
Yield
21
Background Discrimination

• Electron recoils interacting in surface layer
  have incomplete charge collection
• Yield is reduced - may be identified as nuclear
  recoil
• Dead layer events generally have a lower phonon
  delay and risetime compared to genuine nuclear
  recoils
• Use Risetime cut
• Reject gt99 betas keeping gt60 nuclear recoils
• Combination of minimum risetime minimum delay
  gives best rejection

252Cf RT cut
133Ba - RT cut
Phonon Delay (µs)
133Ba
Yield
22
Defining Bands

• Fit yield values from calibration data sets to
  gaussian distributions
• 2? NR and gamma bands shown above

23
Calculating Cuts - cQin
Q outer
Q inner
Q outer (keV)

• Events depositing energy in the Q outer region
  have poor charge collection
• Q inner cut defined to select events depositing
  their energy in the bulk of the crystal (radially)

Q inner (keV)
24
Calculating Cuts - cRT2
Best surface event rejection obtained by cutting
on two parameters - Minimum phonon risetime -
Minimum phonon delay
25
Estimating Leakage
Gammas Apply risetime cut to 105 events in
calibration set - leaves no events in the nuclear
recoil band In the entire 52.6 live day
background dataset, we expect 0.13 (0.56 UL)
gammas to leak into the NR band
Betas Apply RT cut to 2000 events between bands
in calibration set - leaves no single scatters in
NR band In the entire 52.6 live day background
dataset, we expect 0.59 (2.52 UL) betas to leak
into the NR band
26
Final Run118 Result
Z3 Background
cuts
252Cf

• WIMP-search events (blue) from 53 kg-days of Ge
  (Z1235)
• Phonon Cuts remove low y b/g but keep majority
  of nuclear recoils
• 252Cf events (yellow) showing calibrated
  response of detector for nuclear recoils - expect
  WIMP signal in this region

27
Final Run118 Result

• Raw Ge exposure of 52.6 kg-d above analysis
  threshold of 10kev (20kev, Z1)
• Blinded analysis performed using only
  calibration data to define bands and cuts
• Data quality cuts made
• Q inner
• Chisq
• Baseline std.
• Cuts to remove background
• Veto coincidence cut
• Risetime cut
• NR band cut
• 19.4 kg-d Ge net exposure after cuts

28
CDMS II Limit

• Factor of 4 below best previous limits set by
  EDELWEISS
• New analysis methods and increased exposure
  promises 20x improvement over current limit

• Whats Next?
• Currently operating 2 towers
• Adding 3 new towers over the next 6 months (4kg
  Ge, 1.4kg Si)