Dark Matter Direct Search:

the XENON Experiment

José A. Matias Lopes

Universidade de Coimbra, 4 de Dezembro, 2008

What is our Universe made of?

Dark Energy

Atoms

Dark

Matter

Dark Matter from Gravitational Lensing

Galaxy Cluster Abell 2218

SUSY and Dark Matter

• Super Symmetry theory (SUSY) predicts:

• the existence of Weakly Interacting Massive Particles (WIMP) of dark matter

• interaction rate as low as 1 event /ton/year for exposure of tens of millions per sec.

The Via Lactea WIMP Halo

WIMP Detection: Scattering off Atoms

WIMP M. Attisha

WIMP

Nuclear

recoil

nuclear recoil spectrum

WIMP velocity distribution

f (v)dv

4v 2

v03

ev2 /v0

2

d3v

WIMP Detection: Elastic Scattering off Atoms

<V> = 220 km/s

Nuclei recoil energy:

Event rate:

ER

rq2

2mN

2v 2

mN

(1 cos) 50 keV

mmN

m mN

Q

Predicted Rates in Different WIMP Targets

Recoil energy [keVr]

MWIMP = 100 GeV/c2

σWN=4×10-43 cm2

Eth = 2 keVr

Differential rates Cross sections (normalized to nucleons)

CMSSM

Challenges of Direct Detection Experiments

n produced by fission and (α,n)

n produced by µ

muonsG. Heusser, 1995

hadrons

• Low event rates ⇒ ton-scale detectors

• Small deposited energies ⇒ low (~ few keV) energy thresholds

• Low backgrounds

• shield against cosmic rays (deep underground laboratories → µ-spallation reactions)

• low intrinsic radio-activity (ultra-pure materials → (α,n)-reactions)

• shield radio-activity from

surroundings (Pb, PE, H2O, etc)

• Good background rejection

• Particle identification

• nuclear vs. electron recoils

• Position sensitivity/fiducialisation

• Self-shielding

World Wide Direct WIMP Searches

Direct WIMP Detection Experiments

ER

LightCharge

Phonons

ZEPLIN, XENON,

LUX, WARP, ArDM

CRESST

ROSEBUDCDMS

EDELWEISS

DAMA, LIBRA,

XMASS, CLEAN,

ANAIS, KIMSHDMS, DRIFT,

GERDA, MAJORANA

COUPP,

PICASSO

NR WIMP

WIMP

Why Noble Liquids?

• Good Nuclear versus Electron Recoil discrimination

➡pulse shape of scintillation signal

➡ratio of ionization to scintillation signals

• High Scintillation Light Yields; transparent to their own light

➡low energy thresholds can be achieved

• Large Detector Masses are feasible

➡self-shielding => low-activity of inner fiducial volumes

➡good position-resolution in TPC operation mode (use ionization signal)

• Ionization Drift >> 1 m achieved

➡ corresponding to << ppm electronegative impurities

• Competitive Costs and Practicality of large instruments

Noble Liquids as Detector Media

• Liquid noble gases yield both charge and light

• Scintillation is decreased (∼ factor 2) when drift field to extract charge is applied

Z (A)BP (Tb) at

1 atm [K]

liquid density at

Tb [g/cc]

ionization [e-

/keV]

scintillation

[photon/keV]

He 2 (4) 4,2 0,13 39 15

Ne 10 (20) 27,1 1,21 46 7

Ar 18 (40) 87,3 1,40 42 40

Kr 36 (84) 119,8 2,41 49 25

Xe 54 (131) 165,0 3,06 64 46

Noble Liquids as Dark Matter Detectors

Differential rates

Integrated rates

MWIMP = 100 GeV/c2

σWIMP-N=4×10-43 cm2

Dense, homogeneous targets/detectors

High scintillation/ionization yields

Commercially easy to obtain and purify

Scintillation

Light

Intrinsic

Backgrounds

Ne (A=20)

$60/kg

100% even-even

nucleus

85 nm

requires

wavelength

shifter

Low BP (20 K), all

impurities frozen out

No radioactive isotopes

Ar (A=40)

$2/kg

100% even-even

nucleus

128 nm

requires

wavelength

shifter

Natural Ar contains 39Ar

at 1Bq/kg, corresp. to

~150 events/kg/day/keV

at low energies

Xe (A=131)

$3500/kg

50% odd nuclei

(129Xe, 131Xe)

175 nm

UV quartz PMT

window

No long lived isotopes85Kr can be removed by

distillation to ppt level

Charge and Light in Noble Liquids

triplet singlet Excitation/Ionization depends on dE/dx!

(recombination is stronger for NR)

=> discrimination of signal (WIMPs=>NR)

and (most of the) background (gammas=>ER)!

hνhν

2Xe 2Xe

ER Ionization

Excitation

Xe+

+Xe

Xe2+

+e-

Xe**+XeXe*

+Xe

Xe2*

VUV light: λ depends on gas

(85 nm Ne, 128 nm Ar, 175 nm Xe)

time constants depend on gas

( few ns/15.4μs Ne, 10ns/1.5μs Ar, 3/27 ns Xe)

excited molecular states

Noble Liquid Detectors: Existing and Proposed

Experiments

Single Phase

(liquid only)

PSD

Double Phase

(liquid and gas)

PSD and Charge/Light

Neon (A=20) miniCLEAN (100 kg)

CLEAN (10-100 t)

SIGN

Argon (A=40)DEAP-I (7 kg)

miniCLEAN (100 kg)

CLEAN (10-100 t)

ArDM (1 ton)

WARP (3.2 kg)

WARP (140 kg)

Xenon (A=131)

ZEPLIN I

XMASS (100 kg)

XMASS (800 kg)

XMASS (23 t)

ZEPLIN II + III (31 kg, 8 kg)

XENON10, XENON100,

XENON1t

LUX (300 kg)

XENON detection principle: Dual Phase Detector

• detects prompt (S1) light signal after a particle interacts in the active xenon

• drifts the charge, extracts into the gas phase and detects as proportional light (S2)

hν

e- Ed

Eg

Liquid Xe

Gas Xe

PMT array

PMT array

ER

hν

hν

hν

PMT array

PMT array

drift time

WIMP

S1 S2

drift time

Gamma

S1 S2

WIMP

gamma

electron recoil (ER)

nuclear

recoil (NR)

XENON at the Gran Sasso Laboratory

• since March 2006, in ex-LUNA box

• ~3100 m.w.e; muon flux ≈ 1 m-2 h-1

Muon Inte

nsity [m

-2y

-1]

Depth [meters water equivalent]

The first phase: XENON10

• March 06: detector first installed/tested outside the shield

• July 06: inserted into shield (20 cm Pb, 20 cm PE, Rn purge)

• August 24, 2006 – February 14, 2007: WIMP search run

The XENON10 Detector

•22 kg of liquid xenon

➡15 kg active volume

➡20 cm diameter, 15 cm drift

•Hamamatsu R8520 1’’×3.5 cm PMTs

•bialkali-photocathode Rb-Cs-Sb,

•Quartz window; ok at -100ºC and 5 bar

•Quantum efficiency > 20% @ 175 nm

•48 PMTs top, 41 PMTs bottom array

➡x-y position from PMT hit pattern; σx-y≈ 1 mm

➡z-position from ∆tdrift (vd,e- ≈ 2mm/µs), σZ≈0.3 mm

•Cooling: Pulse Tube Refrigerator (PTR),

➡90W, coupled via cold finger (LN2 for emergency)

Typical XENON10 Low-Energy Event

• 4 keVee event; S1: 8 p.e => 2 p.e./keV

S1

S2

S1 S2

Hit pattern of top PMTs

8 p.e. 3000 p.e.

drift time: 65 µs

XENON10 Performance at LNGS

• Stable pressure, temperature, PMT gain, liquid level, cryostat vacuum, HV...

➡ over many months (continuously monitored with ‘slow control system’)

PMT gain < ±2%

December 8, 06

Pressure: ∆P < ±0.006 atm

Temperature: ∆T < ±0.005 ºC

Start of WIMP Search Run

XENON10 Backgrounds :

Data and MC Simulations

x-y position of WS data

r < 8 cm fiducial volume cut

Data

MC total

Steel

PMTs

Ceramic FTs

Single hits in fiducial mass (8.9 kg)

• Gamma BG: dominated by steel (inner vessel and cryostat) and ceramic FTs

• Neutron BG: subdominant for XENON10 sensitivity goal (MC: < 1 event/year from

(α,n) in materials and < 5 events/year from µ-induced n’s)

164 keV 236 keV

XENON10 Calibrations: Gammas and Neutrons

energy scale

(S1 in p.e)

Nuclear recoil spectrum

164 keV

236 keVn-activated Xe:

Combined

energy

spectrum

AmBe source

n-activated Xe:

Anti-correlation of

charge/light

signals

57Co source

Data and Monte Carlo agree well:⇒ NR response at low energies well understood

XENON10 Blind WIMP Analysis Cuts

dt [μs] ↔ z

r < 8 cm fiducial volume cut

Wimp

search

region

3s

XENON10 WIMP Search Data

• WIMP search run Aug. 24. 2006 - Feb. 14, 2007: ~ 60 (blind) live days

• 136 kg-days exposure = 58.6 live days × 5.4 kg × 0.86 (ε) × 0.50 (50% NR acceptance)

WIMP ‘Box’ defined at

50% acceptance of NRs

(blue lines): [Mean,-3σ]

10 events in ‘box’ after all cuts

7.0 (+1.4 -1.0) statistical leakage

expected from the gamma (ER) band

50% NR

acceptance

~ 1800 events

2 6 8 10

1

34

57

9

Spatial Distribution of Events

‘Gaussian events’: nr. 3, 4, 5, 7,9

‘Non-Gaussian events’: nr: 1, 2, 6, 8, 10

Ev. nr. 1: S1 due to noise glitch (a posteriori)

Ev. 2, 6, 8, 10 -> not WIMPs!

Likely explanation: reduced S2/S1-events due to

double scatters with one scatter in a ‘dead’ LXe

region => no S2 for 2nd scatter

e-e-

e-

e-e-e-

-12 kV

ground

S1

S1

15 cm

1.2 cm

e-e-e-

-12 kV

S1

S1

15 cm

1.2 cm

15 c

m

XENON10 WIMP-Nucleon Cross Sections upper

limits for SI Interactions (90% CL)

Upper limits in WIMP-nucleon cross

section derived with Yellin Maximal

Gap Method [PRD 66 (2002)]

No background subtraction:

8.8 × 10-44 cm2 at 100 GeV/c2

(4.5 × 10-44 cm2 at 30 GeV/c2)

5.5 × 10-44 cm2 at 100 GeV/c2

(known background subtracted, not

shown)

factor 6 below previous best limit

WIMP mass [GeV/c2]

WIM

P-n

uc

leo

n S

I c

ros

s s

ec

tio

n [

cm

2]

CDMS-II

XENON10

Phys. Rev. Lett. , 100 (2008) 21303

XENON10 WIMP Search Results for SD

Interactions

Phys. Rev. Lett., 101 (2008) 91301

The XENON10 Collaboration

• Columbia University Elena Aprile, Bin Choi, Karl-Ludwig Giboni, Sharmila Kamat, Yun Lin, Maria Elena Monzani,

Guillaume Plante, Roberto Santorelli and Masaki Yamashita

• University of Zürich Laura Baudis, Jesse Angle, Ali Askin, Martin Bissok, Alfredo Ferella, Marijke Haffke, Alexander Kish,

Aaron Manalaysay, Stephan Schulte, Eirini Tziaferi

• Brown University Richard Gaitskell, Simon Fiorucci, Peter Sorensen and Luiz DeViveiros

• Lawrence Livermore National Laboratory Adam Bernstein, Chris Hagmann, Norm Madden and Celeste Winant

• Case Western Reserve University Tom Shutt, Peter Brusov, Eric Dahl, John Kwong and Alexander Bolozdynya

• Rice University Uwe Oberlack, Roman Gomez, Christopher Olsen and Peter Shagin

• Yale University Daniel McKinsey, Louis Kastens, Angel Manzur and Kaixuan Ni

• LNGS Francesco Arneodo and Serena Fattori

• Coimbra University Jose Matias Lopes, Luis Coelho, Luis Fernandes and Joaquim Santos

XENON100 (2008-2009)

• US/EU new collaboration to build 170 kg (70 kg fiducial) LXe detector in

(conventional: Pb, PE) XENON10 shield at LNGS

• Start of WIMP search data: early 2009

XENON10 XENON100

+ UCLA recently joined (now >30)

• x10 fiducial mass:170 kg LXe

(vs. 20 kg in XE10); 50 kg

fiducial mass (vs. 5 kg in

XE10).

• Improved shield (additional

inner copper layer)

• 242 R8520 PMTs with lower

background and higher QE (vs.

89 in XE10)

• a factor of 100 lower

background: lower background

stainless steel cryostat and

inner chamber

Bell

Top PMTs

Grids

Structure

Active LXe Shield

Teflon

Panels with

HV

Racetracks

Shield PMTs

Vacuum Cryostat

Bottom PMTs

Active LXe Target

Cathode Mesh

XENON100: a new detector

Status of XENON100:

Commission Phase at LNGS

• Cryostat, cryogenics, TPC with active veto installed underground and tested

• high-purity, double walled stainless steel cryostat assembled and tested at LNGS

• 242 (R8520) PMTs (98 top, 80 bottom, 64 in active LXe shield) installed, tested

170 kg

3

• Low activity PMTs: Hamamatsu R8520-06-Al

• 98 on top - 80 on bottom - 64 in active LXe shield

• QE~23% for top array; QE~33% for bottom array

XENON100: PMTs

Bottom PMT ArrayTop PMT Array

New XENON100 material screening facility

• Ultra-low background, 100 % efficient (2 kg) HPGe-spectrometer

• Shield: 5 cm of OFRP Cu (Norddeutsche Affinerie); 20 cm Plombum Pb (inner 5 cm: 3 Bq/kg 210Pb), air-lock system and Nitrogen purge against Rn

• First background spectrum: < 1 event/kg/day/keV above 40 keV

• Goal: screen all XENON100 detector/shield components for a complete BG model

Gator LNGS background

30 kg days (October 2007)

Gator Soudan background

44 kg days

XENON100 Materials Screened at LNGS Facility

Material 238U 232Th 40K 60Co

Stainless Steel

1.5mm (316Ti,

Nironit)

<2 mBq/kg <2 mBq/kg 10.5 mBq/kg 8.5 mBq/kg

Stainless Steel

25mm (316Ti, Nironit)<1.3 mBq/kg <0.9 mBq/kg <7.1 mBq/kg 1.4 mBq/kg

PMTs (R8520-AL) 0.25 mBq/PMT 0.19 mBq/PMT 7.0 mBq/PMT 0.67 mBq/PMT

PMT Bases 0.16 mBq/pc 0.10 mBq/pc <0.16 mBq/pc <0.01 mBq/pc

XENON100: expected Neutron Background

• internal neutron BG from detector materials + shield, from (α,n) and fission reactions

• numbers based on detailed MC, taking into account the measured U/Th activities of

all the materials

MaterialTotal

n-rate [yr-1]

Single NR rate

[μdru]

Stainless steel 17,7 0,29

PMTs 7,0 0,32

Teflon 16,8 0,97

Copper 1,6 0,01

Poly shield 416,3 0,49

Total 2,08 background from

muon-induced

neutrons is currently

being modeled

• background from U/Th/K/Co in stainless steel cryostat & PMTs

• active LXe veto shield on sides, and top/bottom of target mass

• <1 gamma; <0.2 neutron events for 2 month operation

XENON100: expected Gamma Background

<10 mdru

rate of single scatters (0-100 keV)

before S2/S1 discrimination

XENON100

XENON10

XENON100: projected results (2009)

XENON100 purification system

Kr distillation column

(down to ppt level)

Storage rack

Purification unit

(<0.1 ppm volatile/organics)

The XENON100 Slow Control

Set of instruments that locally monitors, records and controls most of the physical

parameters of the experiment.

Slow Control

Instrumentation

Xenon100 Detector

Pressures

Temperatures

Vacuum

Gas flow

Xe liquid levels

Photomultipliers

Anode

Cathode

Grids

High Voltages

Slow Control Server(Underground Computer)

Measures, controls,

records to disk and

makes them available to

the clients.

Runs the alarms through

emails and SMS

The XENON100 Slow Control

The Slow Control Structure

Slow Control Server

Internet

Alarms

SMS and Email

Slow Control Clients

Alarm Redundancy Server

Gets data from server and sends

status reports and email alarms.

Laboratori Nationali del Gran Sasso

Underground Facilities

0,0

0,5

1,0

1,5

2,0

2,5

3,0

0 3 6 9 12 15 18 21 24

E/N (10-17

V cm2 atom

-1)

Y/N

(

x 1

0-1

7 p

ho

ton

s e

lec

tro

n-1

cm

2 a

tom

-1)

this work

Santos et al., Monte Carlo [7]

Fraga et al., Boltzmann [11]

Fonseca et al. [17]

Fonseca et al., -90ºC [17]

Parsons et al. [13]

Bolozdynia et al. [14]

Akimov et al. [15]

Aprile et al., -95ºC [16]

Ngoc et al. [12]

Xenon

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0.0 5.0 10.0 15.0 20.0 25.0 30.0

E/N (10-17

V cm2 atom

-1)

Y/N

(x

10

-17 p

ho

ton

s e

lec

tro

n-1

cm

2 a

tom

-1)

Ar experimental

Ar Monte Carlo

Ar WARP

Xe experimental

JINST 2 (2007) P05001 Phys. Lett. B 668 (2008) 167-170

Absolute Electroluminescence Yield Measurements

Argon

R&D on Photosensors: Avalanche Photodiodes and

Gaseous Photomultipliers

XENON1T: the next step

• Possible new design with full coverage of PMTs.

• 3T Xe, 1T fiducial mass

• A new PMT (QUPID) designed by UCLA with Hamamatsu is under development . Alternative photosensors also under study (Coimbra & Columbia).

From XENON10 to XENON1T

DetectorsXENON10

(2006-2007)

XENON100

(2007-2009)

XENON1T

(2010-2012)

Fiducial Mass (kg) 5 50 1000

Gamma background

(dru:

events/kg/keV/day)

1 10-2 goal: <10-4

Neutron background < 1 per year < 1 per year goal: < 1/two years

Exposure 2 months 2 months 2 year

Photodetectors R8520lower bkg, higher

QE R8520QUPID, APD, GPMs

WIMP Sensitivity

at 100 GeV9 x 10-44 cm2 2 x 10-45 cm2

(projected)

3 x 10-47 cm2

(projected)

Future projected results

Summary

95%68% 1 event/kg/yr

1 event/t/yr

CDMS-II, XENON100, COUPP,

CRESST-II, EDELWEISS-II, ZEPLIN-

III,...

SuperCDMS1t, WARP1t, ArDM

XENON1t, EURECA, XMASS, ...

Many different techniques/targets are being employed to search for dark matter particles

Sensitivities are now approaching the theoretically interesting regions!

Next generation projects: should reach the ≲ 10-10 pb level => WIMP (astro)-physics

Nature 448, July 19, 2007

Theory example: CMSSM (Roszkowski, Ruiz, Trotta)