the slides used in the Jodi Cooley presentation

KIPAC, Dec. 17, 2009Jodi Cooley, SMU, CDMS Collaboration

New Results from the Final Runs of the CDMS II Experiment

Jodi CooleySouthern Methodist University

Analysis Coordinatorfor the CDMS Collaboration

1

SLAC, Dec. 17, 2009Jodi Cooley, SMU, CDMS Collaboration

Overview

What we know and what we dont know about dark matter

CDMS-II experiment detection principle first results from the final CDMS II data runs

The future SuperCDMS

2


Introduction to Dark Matter

3


The Evidence for Dark Matter

4

Motion of Galaxies in Clusters

1933

Rotation Curves

1970


The Bullet Cluster Observations of the Bullet

Cluster in the optical and x-ray fields combined with gravitational lensing provide compelling evidence that the dark matter is particles.

blue = lensingred = x-rays

5

Clowe et al., ApJ, 648, 109

Gravitational lensing tells us mass location

No dark matter = lensing strongest near gas

Dark matter = lensing strongest near stars


The Cosmic Pie

Measurements from CMB + supernovae + LSS indicate that ~23% of our Universe is composed of dark matter.

6


What Could Dark Matter Be?

Warm or Cold? ordinary s can not

make up LSS of universe

7
http://nedwww.ipac.caltech.edu/level5/March02/Plionis/Plionis3_2.htmlhttp://nedwww.ipac.caltech.edu/level5/March02/Plionis/Plionis3_2.html


What Could Dark Matter Be?

Warm or Cold? ordinary s can not

make up LSS of universe

Baryonic or Non-Baryonic? to avoid skewing

formation of light elements in BBN

7


A Candidate is Born!

8

WMAP 0.095 < h2 < 0.129

Weakly Interacting Massive Particles

Particles in thermal equilibrium

Decoupling when non-relativistic

Freeze out when annihilation rate

! expansion rate

Relic abundance:

!" h2#$10

-27 cm

3s

-1 %$&'

ann v$(

freeze out

if m and ann determined by electroweak physics, then ~ 1

freeze out

annihilationproduction

relic abundance

h2

3 1027

< v >

1037

cm2


New stable, massive particle produced thermally in early universe

Weak-scale cross-section gives observed relic density


Motivated by Particle Physics Too!

New TeV physics is required to explain radiative stability of weak scale.

SuperSymmetry Extra Dimensions ...

These theories give rise to convenient dark matter candidates.

LSP, LKPBaltz et al., PRD 74, 103521 (2006)

stable

9



Particles in thermal equilibrium

Decoupling when non-relativistic

Freeze out when annihilation rate

! expansion rate

Relic abundance:

!" h2#$10

-27 cm

3s

-1 %$&'

ann v$(

freeze out

if m and ann determined by electroweak physics, then ~ 1

freeze out

annihilationproduction

relic abundance

h2

3 1027

< v >

Happy Coincidence!

Baltz et al., PRD 74, 103521 (2006)

stable

1037

cm2

10


How Do We Detect WIMPs?

IceCube

FGST

CDMS

WIMP scattering on earth WIMP production on earth

WIMP annihilation in the cosmos

11

CERN


The Spherical Cow

12


Direct Detection Event Rates

D. Cline, Scientific American 2003

Spherical Cow Halo Modellocal density (o) = 0.3 GeV/cm3, Maxwellian distrubution, rms velocity (vo) = 220 km/s, vesc = 650 km/s

13



D. Cline, Scientific American 2003

Spherical Cow Halo Modellocal density (o) = 0.3 GeV/cm3, Maxwellian distrubution, rms velocity (vo) = 220 km/s, vesc = 650 km/s

Interaction Detailsspin-independent, coherent scattering

A2

13



GeSiXe

0 50 100101

102

103

104

Recoil [keV]

Coun

ts [#

106

/kg/

keV/

day]

WIMP Elastic Scattering Differential RateWIMP Differential Event RateM = 100 GeV/c2-N = 10-45 cm2

Elastic scattering of a WIMP deposits small amounts of energy into recoiling nucleus (~ few 10s of keV)

Featureless exponential spectrum

Expected rate: < 0.01/kg-d

Radioactive background of most materials higher than this rate.

14

SiGe

Xe


Detection Challenges

Low energy thresholds (~10 keV)Rigid background controls

Clean materials shielding discrimination power

Substantial Depth neutrons look like WIMPS

Long exposures large masses, long term stablility

15


CDMS II

16


The CDMS Collaboration

17

California Institute of TechnologyZ. Ahmed, J. Filippini, S.R. Golwala, D. Moore, R.W. Ogburn

Case Western Reserve UniversityD. Akerib, C.N. Bailey, M.R. Dragowsky, D.R. Grant, R. Hennings-Yeomans

Fermi National Accelerator LaboratoryD. A. Bauer, F. DeJongh, J. Hall, D. Holmgren, L. Hsu, E. Ramberg, R.L. Schmitt, J. Yoo

Massachusetts Institute of TechnologyE. Figueroa-Feliciano, S. Hertel, S.W. Leman, K.A. McCarthy, P. Wikus

NIST *K. Irwin

Queens UniversityP. Di Stefano *, N. Fatemighomi *, J. Fox *, S. Liu *, P. Nadeau *, W. Rau

Santa Clara UniversityB. A. Young

Southern Methodist UniversityJ. Cooley

SLAC/KIPAC *E. do Couto e Silva, G.G. Godrey, J. Hasi, C. J. Kenney, P. C. Kim, R. Resch, J.G. Weisend

Stanford UniversityP.L. Brink, B. Cabrera, M. Cherry *, L. Novak, M. Pyle, A. Tomada, S. Yellin

Syracuse UniversityM. Kos, M. Kiveni, R. W. Schnee

Texas A&MJ. Erikson *, R. Mahapatra, M. Platt *

University of California, BerkeleyM. Daal, N. Mirabolfathi, A. Phipps, B. Sadoulet, D. Seitz, B. Serfass, K.M. Sundqvist

University of California, Santa BarbaraR. Bunker, D.O. Caldwell, H. Nelson, J. Sander

University of Colorado DenverB.A. Hines, M.E. Huber

University of FloridaT. Saab, D. Balakishiyeva, B. Welliver *

University of MinnesotaJ. Beaty, P. Cushman, S. Fallows, M. Fritts, O. Kamaev, V. Mandic, X. Qiu, A. Reisetter, J. Zhang

University of ZurichS. Arrenberg, T. Bruch, L. Baudis, M. Tarka


CDMS-II: The Big Picture

Discrimination from measurements of ionization and phonon energy.

ER b

ackg

roun

d

NR signa

l

Ephonon

Ech

arg

e

Keep backgrounds low as possible through shielding and material selection.

18

Use a combination of discrimination and shielding to maintain a


CDMS-II ZIP Detectors

Z-sensitive Ionization and Phonon mediated

230 g Ge or 100 g Si crystals (1 cm thick, 7.5 cm diameter)

Photolithographically patterned to collect athermal phonons and ionization signals

xy-position imaging Surface (z) event rejection

from pulse shapes and timing

30 detectors stacked into 5 towers of 6 detectors

19

3 (7.6 cm)

1 cm

1 tungsten 380 x 60 aluminum fins


ZIP Detectors: Charge

h+

-3V

h+ h+

e-e-e-

-3V

!

Vet

oed

by

gu

ard

rin

g

Vetoed

by g

uard

ring

~85%

~15%

Inner Channel: ionization measurementOuter Channel: fiducial volume

20


ZIP Detectors: PhononsAl Collector

W Transition-Edge Sensor

Si or Ge

quasiparticlediffusion

phonons

21

~

RTES ()

4

3

2

1

T (mK)Tc ~ 80mK

~ 10mK

Tungsten Transition Edge Sensor (TES)

4 SQUID readout channels, each reads out 1036 TESs in parallel


Most backgrounds (e, ) produce electron recoils

WIMPS and neutrons produce nuclear recoils.

Background Rejection

Ionization yield (ionization energy per unit phonon energy) strongly depends on particle type.

0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n yi

eld

22


Most backgrounds (e, ) produce electron recoils

WIMPS and neutrons produce nuclear recoils.

Background Rejection

Ionization yield (ionization energy per unit phonon energy) strongly depends on particle type.

0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n yi

eld

Particles that interact in the surface dead layer result in reduced ionization yield.

0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n yi

eld

22


Reduced Ionization Yield

Reduced charge yield is due to carrier back diffusion in surface events.

Dead layer is within ~10m of the surface.

~10 mdead layer -3Vcarrier back diffusion

h+ h+ h+

h+ h+

e-e-e-

e- e-

e-h+

rapid phonondown-conversion

23


Surface Event Rejection

Phonons near surface travel faster, resulting in shorter risetimes of phonon pulse.

Selection criteria set to accept ~0.5 background events.

24

BulkSurface

Delay + RiseTime [s]

Co

un

ts


Another View of Discrimination

25

3

! "! #! $! %! &!!!

!'(

&

&'(

)*+,-./01*234/56*789,1-:;

/

/

!"! !&! ! &! "!!&!

!

&!

"!

?!

@,2A;.-:*>/B-A-13/C;2;A*/=-*.>

/

/

&??E;/EF.6/G;AA;D

&??E;/HF2I;+*/0J*1

SLAC, Dec. 17, 2009Jodi Cooley, SMU, CDMS Collaboration26


Log 1

0(Muo

n Flux

) (m-

2 s-1

)

Depth (meters water equivalent)

SUF 17 mwe 0.5 n/d/kg (182.5 n/y/kg)Soudan 2090 mwe 0.05 n/y/kg

SNOLab 6060 mwe 0.2 n/y/ton (0.0002 n/y/kg)

26


Peeling the Shielding Onion

Active Muon Veto: rejects events from cosmic rays

27




27

Polyethyene: moderate neutrons produced from fission decays and from (,n) interactions resulting from U/Th decays

Pb: shielding from gammas resulting from radioactivity

Low Activity Lead Polyethylene

-metal (with copper inside)

Ancient lead

40 cm

22.5 cm

10 cm




27





Ancient lead

40 cm

22.5 cm

10 cm

Cu: shielding from gammas




27





Ancient lead

40 cm

22.5 cm

10 cm

Cu: shielding from gammas@ 40 mK!!

Phonon Sensors


Seven Total Data Runs: R123 - R124:- taken: (10/06 - 3/07) (4/07 - 7/07)- exposure: ~400 kg-d (Ge raw)- PRL 102, 011301 (2009)

R125 - R128- taken: (7/07 - 1/08) (1/08 - 4/08)

(5/08 - 8/08) (8/08 - 9/08)- exposure: ~ 600 kg-d (Ge raw)

R129:- taken: (11/08 - 3/09)

CDMS II Experiment

28

T1 T2

T3T5T4

30 detectors installed and operating in Soudan since June 2006.- 4.75 kg of Ge, 1.1 kg of Si


Results from Final Data Blind Analysis: Event selection and efficiencies were calculated without looking at the signal region of the WIMP-search data.

Event Selection:Veto-anticoincidence cutSingle-scatter cutQinner (fiducial volume) cutIonization yield cutPhonon timing cut

29

We unblinded the signal region November 5, 2009


Method 2Use singles and

multiples just outside NR band

Surface Event Background

30

Expected Surface leakage = N fail cutdata

*NSidebandfail cut

NSidebandpass cut

Method 1Use multiple-scatters in

NR band

Method 3Use singles and multiples from Ba calibration in

wide region 133Ba 252Cf

Correct for systematic effects due to different distributions in energy and face

Combined Estimate =


Neutron Background

31

Cosmogenic:

Radiogenic:

Nvetoed, SS, NRMC

Nunvetoed, SS, NRMC

Nvetoed, SS, NRdata

* =

3 vetoed, single scatter events

0.03 - 0.06 events

Materials measured using conventional HPGe detector @ 77 KSpectra confirmed by Monte Carlo.

Contamination levels used as inputs to Geant4 simulation.


Projected Sensitivity

32

612 raw kg-days194.1 kg-d WIMP equiv.

@ 60 GeV/c^2 (10 -100 keV analysis

energy range)

Surface Background

Neutron BackgroundCosmogenic

Radiogenic0.03 - 0.06


Opening the Box

33

Box opened November 5 , 2009 for 14 Ge ZIP detectors


Opening the Box

33


3 region maskedHide unvetoed singles

0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y

ield


Opening the Box

33



0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y

ieldLift mask, see 150

singles failing timing cut

0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y

ield


Opening the Box

33


Apply the timing cut ...


0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y



0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y

ield


Opening the Box

33


Apply the timing cut ...


0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y



0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y

ield

2 EVENTS OBSERVED!

0 10 20 30 40 50 60 70 80 90 1000

0.5

1

1.5

Recoil Energy (keV)

Ioni

zatio

n Y

ield

T1Z1 10/27/07

T3Z4 08/05/07


CDMS II Results

Upper limit at the 90% C.L. on the WIMP-nucleon cross-section is 3.8 x 10-44 cm2 for a WIMP of mass 70 GeV/c2

Note: An improved estimate of our detector masses (~9% decrease) was used in calculating these limits.

34

WIMP mass [GeV/c2]

WIM

P!nu

cleo

n !

SI [c

m2 ]

101 102 10310!44

10!43

10!42

10!41Ellis 2005 LEESTRoszkowski 2007 (95%)ZEPLIN III 2008XENON10 2007CDMS Soudan 2008CDMS 2009 GeCDMS Soudan (All)Expected Sensitivity

arXiv: 0912.3320


Inelastic Scattering

Disfavor all DAMA/LIBRA allowed region except for WIMPs of mass ~100 GeV with mass-splittings ~80-140 keV

Shown are only regions for which CDMS II and XENON10 are not compatible with DAMA/LIBRA at the 90% C.L.

35

!"#$%&'((%)*+,-./0

!"#$!&'((%(12344356%)7+,0

%

%

898 89/ 89:9

/9

;9


Closer Examination of Observed Events

36


Event Yield, Timing and Energy

37

4

0 20 40 60 80 1000

0.3

0.6

0.9

1.2

Ion

izati

on

Yie

ld

Recoil Energy (keV)

0

0.3

0.6

0.9

1.2

1.5

Ion

izati

on

Yie

ld

Fail Timing Cut

Pass Timing Cut

Fail Timing Cut

Pass Timing Cut

FIG. 2: Ionization yield versus recoil energy for events pass-ing all cuts, excluding yield and timing. The top (bottom)plot shows events for detector T1Z5(T3Z4). The solid redlines indicate the 2 electron and nuclear recoil bands. Thevertical dashed line represents the recoil energy threshold andthe sloping magenta dashed line is the ionization threshold.Events that pass the timing cut are shown with round mark-ers. The candidate events are the round markers inside thenuclear-recoil bands. (Color online.)

ate the pre-blinding misidentified surface event estimate.213Therefore, a refined calculation, which accounts for this214effect, produced a revised surface event leakage estimate215of 0.8 0.1(stat) 0.2(syst) events. Based on this re-216vised estimate, the probability to have observed two or217more surface events in this exposure is 20.4%. Inclusion218of the neutron background estimate increases the prob-219ability to have observed two or more background events220to 23.3%. These values indicate that the results of this221analysis cannot be interpreted as significant evidence for222WIMP interactions. We nonetheless note that we lack223sufficient additional information to definitively reject ei-224ther event as a signal event.225

To better quantify the consistency of the candidate226events with the nuclear recoil and surface event hypothe-227ses, we performed a likelihood ratio analysis using dis-228tributions for yield and timing of these two event classes229from calibration and WIMP-search multiple-scatter data230to calculate the likelihoods. We found that, in the case231of T1Z5 (T3Z4), 2.5% (0.01%) of surface events have a232likelihood ratio less consistent with the ionization-side233surface event hypothesis and 0.24% (0.02%) of surface234

!!" !# " # !" !#!!"

"

!"

$"

%"

&'()*+,-./01,.+/

&'()*+,-./02,),3405*(*).6.(0789

0

0!!"

"

!"

$"

%"

:"

&'()*+,-./01,.+/

0

0;*,+02,),340


What More Can We Say?

38

The two events occur during a time of nearly ideal detector performance.

They are separated in time by several months and occur on detectors in different towers (T1Z5 and T3Z4).

They occur on inner detectors where we have a stronger handle on our background estimate.

SLAC, Dec. 17, 2009Jodi Cooley, SMU, CDMS Collaboration39

Data Quality Item Resultmuon veto performance good

neutralization good

KS tests normal

noise levels typical

pre-pulse baseline rms typical

background electron-recoil rate typical

surface event rate typical

radial position well-contained

single-scatter identification good

special running conditions no

operator recorded issues no


W&C Dec. 18, 2009 41

Reconstruction Checks

ionization and phonon energies lookgood, phonon timing looks good

Could there be a problemwith the start time of the

charge pulse?

phonon chan A

phonon chan B

phonon chan C

phonon chan D

inner ionization

outer inonization

Candidate 2 (on det T3Z4)

ADC bin

fittedstarttime

What is thetrue start

time?

!2 o

f th

e f

itADC bin

Closeup of template fitto ionization pulse

40

Reconstruction Checksionization and phonon energies look good, phonon timing looks good, ...

... but could be problem with charge start time

T3Z4 Candidate

W&C Dec. 18, 2009 41

Reconstruction Checks

ionization and phonon energies lookgood, phonon timing looks good

Could there be a problemwith the start time of the

charge pulse?

phonon chan A

phonon chan B

phonon chan C

phonon chan D

inner ionization

outer inonization

Candidate 2 (on det T3Z4)

ADC bin

fittedstarttime

What is thetrue start

time?

!2 o

f th

e f

it

ADC bin

Closeup of template fitto ionization pulse

Note: This effects some events with ionization energy < ~6 keV. It does not effect candidate event on T1Z5.

Raw Unfiltered

Data.


Reconstruction (Cont.)

41

A refined calculation of the surface background taking into account larger errors in the timing estimate a low energy produced a post-unblinding leakage estimate of

Based on this revised estimate the probability of observing 2 or more events is 23% (includes neutron + surface event background).

With an improved reconstruction algorithm which includes this 2-fit, this pulse may fail the timing cut, but other events may be let into the signal region.


What Would it Take to Exclude these Events?

Reducing the surface event estimate by ~1/2 would remove both candidates while reducing our exposure by 28%

Additional events would not enter the signal region until we increased the surface event estimate by a factor of ~2.

42

W&C Dec. 17, 2009 40

0 e

vent

s

1 eve

nt

2 e

vent

s

3 o

r m

ore

eve

nts

0.1 1.0 10.0estimated surface event leakage from 133Ba

(no systematic correction applied)

Cut Varying StudyTightening the cut toyield ~1/2 the expectedsurface events, removesboth events from thesignal region and reducesthe exposure by ~28%

Additional events appearin the signal region afterloosening the cut to ~2Xthe expected leakage

The calculated limit doesntdepend strongly on chosen

surface-event rejection cutvalue

chosenleakage


Final Comments on this Analysis

Our results cannot be interpreted as significant evidence for WIMP interactions.

However, we cannot reject either event as signal.

43


The Future

44


Next Step: SuperCDMS Last CDMS II data taken on March 18, 2009 March 19, 2009: Warm up to begin the installation and

commissioning of the first SuperCDMS detectors. Commissioning runs of the first SuperCDMS tower is underway.

Fabrication of remaining detectors for the SuperCDMS Soudan project (15 kg Ge deployed in existing Soudan setup) underway. Installation and commissioning summer 2010.

Eventual goal: SuperCDMS SNOLAB (100 kg Ge deployed at SNOLAB)


Sensitivity of Future Detectors

46


Conclusions

47

We observe 2 events in the first analysis of the final data taken by CDMS II between July 07 and Sept. 08. This yields a cross section limit of < 3.8 x 10-44cm2 (90% CL) for a WIMP of mass 70 GeV/c2 when combining this result with previous analyses.

The results of this analysis cannot be interpreted as significant evidence for WIMP interactions, but we can not reject either event as a signal.

The first SuperTower of detectors has been installed and is operating in the Soudan Underground Laboratory. Remaining SuperTowers of detectors are planned to be installed in Summer 2010.

Stay tuned for this coming year. Several other promising technologies (liquid nobles, bubble chambers, ...) will have exciting results.

the slides used in the Jodi Cooley presentation

Documents

Transcript of the slides used in the Jodi Cooley presentation