1

1. Introduction2. 800kg detector design3. Summary

XMASS experiment

15th Sep. 2006

A. Takeda for the XMASS collaborationKamioka Observatory, ICRR,

University of Tokyo

IDM2006, Rhodes, Greece

2

1. Introduction

Dark matter

Double beta

Solar neutrino

What’s XMASS

Xenon MASSive detector for solar neutrino (pp/7Be) Xenon neutrino MASS detector ( decay) Xenon detector for Weakly Interacting MASSive Particles (DM search)

Multi purpose low-background experiment with liq. Xe

3

Why liquid xenon

Large Z (=54) Self-shielding effect Large photon yield (~42 photons/keV ~ NaI(Tl)) Low threshold High density (~3 g/cm3) Compact detector (10 ton: sphere with diameter of ~2m) Purification (distillation)

No long life radioactive isotope Scintillation wavelength (175 nm, detected directly by PMT) Relative high temperature (~165 K)

4

All volume20cm wall cut30cm wall cut (10ton FV)

BG

nor

ma

lize

d b

y m

ass

1MeV0 2MeV 3MeV

Large self-shield effect

External ray from U/Th-chain

Key idea: self-shielding effect for low energy events

tracking MC from external to Xenon

Blue : γ trackingPink : whole liquid xenonDeep pink : fiducial volume

Background are widely reduced in < 500keV low energy region

U-chain gamma rays

5

100kg Prototype 800kg detector

10 ton detector

～ 30cm

～ 80cm

～ 2.5mR&D

Dark matter search

Multipurpose detector

(solar neutrino, …)

Strategy of the scale-up

With light guide

We are here !

Vertex & energy reconstruction Self shielding power BG level

6

Trend of Dark matter (WIMPs) direct searches

Scintillation

Phonon Ionization

Ge, TeO2, Al2O3, LiF, etc

NaI, Xe, CaF2, etc.

Ge

Recoiled nuclei are mainly observed by 3 ways

CDMS, EDELWEISS: phonon + ionization

Taking two type of signals simultaneously is recent trend

ray reduction owing to powerful particle ID

However, seems to be difficult to realize a large and uniform detector due to complicated technique

Ge, Si

7

Strategy chosen by XMASS

Make large mass and uniform detector (with liq. Xe)

Reduce ray BG by fiducial volume cut (self shielding)

Same style as successful experiments of Super-K, SNO, KamLAND, etc.

Super-K KamLANDSNO

8

Main purpose: Dark Matter search

812-2” PMTs immersed into liq. Xe 70% photo-coverage

~80cm diameter

~4 p.e./ keV

External ray BG: 60cm, 346kg 40cm, 100kg

Expected dark matter signal(assuming 10-6 pb, Q.F.=0.2 50GeV / 100GeV,)

pp & 7Be solar

Achieved

2. 800kg detector design

0 100 200 300Energy [keV]

9

Expected sensitivities

Large improvements will be expected SI ~ 10-45 cm2 = 10-9 pb SD~ 10-39 cm2 = 10-3 pb

XMASS(Ann. Mod.)

XMASS(Sepc.)

Edelweiss Al2O3

Tokyo LiF

Modane NaI

CRESST

UKDMC NaI

NAIAD

XMASS FV 0.5 ton yearEth = 5 keVee~25 p.e., 3 discoveryw/o any pulse shape info.

10-6

10-4

10-8

Cro

ss s

ectio

n to

nuc

leon

[p

b]

10-4

10-2

1

102

104

106

Plots except for XMASS: http://dmtools.berkeley.eduGaitskell & Mandic

10-10

10

Basic performances have been already confirmed using 100 kg prototype detector

Status of 800 kg detector

Vertex and energy reconstruction by fitter Self shielding power BG level

Detector design is going using MC Structure and PMT arrangement (812 PMTs) Event reconstruction BG estimation

New excavation will be done soon

Necessary size of shielding around the chamber

11

Hex

1PMT 12 pentagons / pentakisdodecahedron

Structure of 800 kg detector

5 triangles make pentagon

34cm

31cm

31cm

10 PMTs per 1 triangle

1

2

3

4

5

6

7

8

9

10

agonal quarts window

HamamatsuR8778MOD

12

Total 812 hex PMTs (10PMTs/triangle×60 + 212 @gap) immersed into liq. Xe ~70% photo-coverage Radius to inner face ~44cm

Each rim of a PMT overlaps to maximize coverage

13

Event reconstruction

60

50

40

30

20

10

030 3426 3822

GeneratedR = 31cmE = 10keV

= 2.3 cm

Reconstructed position [cm]

Eve

nts

5 keV

10 keV

50 keV100 keV500 keV1 MeV

Fiducial volume

2010 300 40

8

10

6

4

2

0

12

Distance from the center [cm]

(r

econ

stru

cted

) [c

m]

Position resolution @Boundary of fiducial volume

10 keV ~ 3.2 cm 5 keV ~ 5.3 cm

14

5keV ~ 1MeV

0 5 10 15 20 25 30 35 40 45

50

45

40

35

30

25

20

15

10

5

50Distance from the center [cm]

R_reconstructed(cm) Generated VS reconstructed

• No wall effect

• Out of 45cm, some events occurring behind the PMT are miss reconstructed

• In the 40cm~44cm region, reconstructed events are concentrated around 42cm,

but they are not mistaken for those occurred in the center

• Up to <~40cm, events are well reconstructed with position resolution of ~2~5cm

• Out of 42cm, grid whose most similar distribution is selected because of no grid data

15

PMT PMT

Light leak events

Some scintillation lights generated behind the PMTenter the inner region

It is not problemif light shield is installed

PMT hit map

16

Achieved (measured by prototype detector) Goal (800kg detector)

ray from PMTs ~ 10-2 cpd/kg/keV 10-4 cpd/kg/keV → Increase volume for self shielding → Decrease radioactive impurities in PMTs (~1/10)

238U = (33±7)×10-14 g/g 1×10-14 g/g → Remove by filter

232Th < 23×10-14 g/g (90% C.L.) 2×10-14 g/g → Remove by filter (Only upper limit)

Kr = 3.3±1.1 ppt 1 ppt → Achieve by 2 purification pass

1/100

1/33

1/12

1/3

800kg BG study

17

Estimation of ray BG from PMTs

Energy [keV]

Cou

nts/

keV

/day

/kg

All volume

R<39.5cm

R<34.5cm

R<24.5cm

All volume

R<39.5cm

R<34.5cm

R<24.5cm

Statistics: 2.1 days

• U-chain• 1/10 lower BG PMT than R8778

No event is found below100keV after fiducial cut (R<24.5cm)

(Now, more statisticsis accumulating)

< 1×10-4 cpd/kg/keVcan be achieved

18

Water shield for ambient and fast neutron

wa

water

Liq. Xe

Generation point of or neutron

Necessary shielding was estimated for the estimation of the size of the new excavation

MC geometry

Configuration of the estimation

Put 80cm diameter liquid Xe ball Assume several size of water shield 50, 100, 150, and 200cm thickness Assume copper vessel (2cm thickness) for liquid Xe

19

attenuation

PMT BG level

attenuation by water shield

0 100 200 300Distance from LXe [cm]

10

104

De

tect

ed/g

ene

rate

d*s

urf

ace

[cm

2]

103

102

1

10-1

10-2 More than 200cm water is needed to reduce the BG to the PMT BG level

Initial energy spectrum from the rock

Deposit energy spectrum (200cm)

20

water: 200cm, n: 10MeV

Liq. Xe

water

No event is found fromthe generated neutron

of 105

< 2×10-2 counts/day/kg

fast neutron attenuation

• Fast n flux @Kamioka mine: (1.15±0.12) ×10-5 /cm2/sec

• Assuming all the energies are 10 MeV very conservatively

200cm water is enoughto reduce the BG to the PMT BG level

BG caused by thermal neutronis now under estimation

21

New excavation @Kamioka mine

New excavation for XMASS and other underground experiment will be made soon

~5 m

~15 m

~20 m

22

3. Summary XMASS experiment: Multi purpose low-background experiment with large mass liq. Xe

800 kg detector: Designed for dark matter search mainly, and 102 improvement of sensitivity above existing experiments is expected

Detector design of 800 kg detector is going

BG estimation Shielding New excavation

23

Backup

24

800kg detector: Main purpose: Dark Matter search

~800-2” PMTs immersed into liq. Xe 70% photo-coverage

~80cm diam

eter

~4 p.e./keV

External ray BG: 60cm, 346kg 40cm, 100kg

Expected dark matter signal(assuming 10-42 cm2, Q.F.=0.2 50GeV / 100GeV,)

pp & 7Be solar

Achieved

Photoelectrons (p.e.)

5 keV

10 keV

25

XMASS collaboration• ICRR, Kamioka Y. Suzuki, M. Nakahata, S. Moriyama, M. Shiozawa,

Y. Takeuchi , M. Miura, Y. Koshio, K. Abe, H. Sekiya, A. Takeda, H. Ogawa, A. Minamino, T. Iida, K. Ueshima• ICRR, RCNN T. Kajita, K. Kaneyuki• Saga Univ. H. Ohsumi• Tokai Univ. K. Nishijima, T. Maruyama, Y. Sakurai• Gifu Univ. S. Tasaka• Waseda Univ. S. Suzuki, J. Kikuchi, T. Doke, A. Ota, Y. Ebizuka• Yokohama National Univ. S. Nakamura, Y. Uchida, M, Kikuchi, K. Tomita, Y. Ozaki, T. Nagase, T. Kamei, M. Shibasaki, T. Ogiwara• Miyagi Univ. of Education Y. Fukuda, T. Sato• Nagoya ST Y. Itow• Seoul National Univ. Soo-Bong Kim• INR-Kiev O. Ponkratenko• Sejong univ. Y.D. Kim, J.I. Lee, S.H. Moon

26

Hamamatsu R8778MOD(hex)

12cm

5.4c

m5.

8cm

(ed

ge

to e

dg

e)

0.3cm(rim)

c.f. R8778 　 U 1.8±0.2x10-2 Bq Th 6.9±1.3x10-3 Bq40K 1.4±0.2x10-1 Bq

Hexagonal quartz window Effective area: 50mm (min) QE <~25 % (target) Aiming for 1/10 lower background than R8778

Prototype has been manufactured already Now, being tested

27

12

Ceramic dielectric parts to support dynodes

Glass parts for feed through & containment

12

For R8778mod 　 using quartz

For R8778mod 　 Reduce glass material

c.f. R8778 (used for 100kg chamber)　 U 1.8±0.2x10-2 Bq Th 6.9±1.3x10-3 Bq40K 1.4±0.2x10-1 BqU Th 40K

※measured byHPGe detectorin Kamioka

Improvement result will be coming soon!

28

Super-K

KamLAND (>0.8MeV)

Heidelberg Moscow

Kamioka Ge

ZEPLIN before PSD cut

DM signal for LXe100GeV 10-6pb

CDMS IIAfter PID

DAMANaI

Current XMASS (new improvement!)

XMASS800kg

Eve

nts/

kg/k

eV/d

ay

BG levels

29

100kg prototype

With light guide version

R&D status using prototype detector

Main purpose

Confirmation of estimated 800 kg detector performance

BG study

Vertex and energy reconstruction by fitter Miss fitting due to dead angle of the cubic detector (“wall effect”, will be explained later) can be removed with light guide Self shielding power

Understanding of the source of BG Measuring photon yield and its attenuation length

~30 cm cube3 kg fiducial

30

In theKamioka Mine

(near the Super-K)

Gamma ray shield

OFHC cubic chamber

Liq. Xe (31cm)3

MgF2 window

54 2-inch low BG PMTs

16% photo- coverage

Hamamatsu R8778

2,700 m.w.e.

100 kg prototype detector

31

1.0m

1.9mRn free air (~3mBq/m3)

material thickness

　　 Polyethylene 15cm

　　 Boron 5cm

　　 Lead 15cm

　　 EVOH sheets 30μm

　　 OF Cupper 5cm

4 shield with door

32

100 kg Run summary

1st run (Dec. 2003)

2nd run (Aug. 2004)

3rd run (Mar. 2005) with light guide

Confirmed performances of vertex & energy reconstruction Confirmed self shielding power for external rays Measured the internal background concentration

Succeeded to reduce Kr from Xe by distillation Photo electron yield is increased Measured Rn concentration inside the shield

Confirmed the miss fitting (only for the prototype detector) was removed Now, BG data is under analysis

33

PMT

n

nL )

!

)exp(Log()Log(

L: likelihood

: F(x,y,z,i)

x total p.e. F(x,y,z,i)

n: observed number of p.e.

=== Background event sample === QADC, FADC, and hit timing information are available for analysis

F(x,y,z,i): acceptance for i-th PMT (MC)VUV photon characteristics:

Lemit=42ph/keVabs=100cm

scat=30cm

Calculate PMT acceptances from various vertices by Monte Carlo. Vtx.: compare acceptance map F(x,y,z,i) Ene.: calc. from obs. p.e. & total accept.

Reconstructedhere

FADC Hit timing

QADC

Reconstruction is performed by PMT charge pattern (not timing)

Vertex and energy reconstruction

34→ Vertex reconstruction works well

Collimated ray source run from 3 holes (137Cs, 662keV)

DATA

MC

hole Ahole Bhole C

ABC+++

Performance of the vertex reconstruction

35

65keV@peak(/E ~ 10%)

Similar peak position in each fiducial.No position bias

Collimated ray source run from center hole (137Cs, 662keV)

→ Energy reconstruction

works well

All volume20cm FV10cm FV

Performance of the energy reconstruction

36

z position distribution of the collimated ray source run

→ Data and MC agree wellγ

Demonstration of self shielding effect

37

Shelf shielding for real data and MC

Good agreement (< factor 2) Self shielding effect can be seen clearly. Very low background (10-2 /kg/day/keV@100-300 keV)

REAL DATA MC simulation

All volume20cm FV

10cm FV(3kg)

All volume20cm FV

10cm FV(3kg)

Miss-reconstruction due to dead-angle region from PMTs.

Eve

nt r

ate

(/kg

/day

/keV

)

10-2/kg/day/keV

Aug. 04 runpreliminary

3.9dayslivetime

~1.6Hz, 4 fold, triggered by ~0.4p.e.

38

Kr = 3.3±1.1 ppt (by mass spectrometer)

→ Achieved by distillation

U-chain = (33±7)x10-14 g/g (by prototype detector)

Th-chain < 23x10-14 g/g(90%CL) (by prototype detector)

1/2=164s (Q=3.3MeV) (7.7MeV)

214Bi 214Po 210Pb

1/2=299ns (Q=2.3MeV) (8.8MeV)

212Bi 212Po 208Po

Delayed coincidence search

Delayed coincidence search

(radiation equilibrium assumed)

(radiation equilibrium assumed)

Internal backgrounds in liq. Xe were measured

Main sources in liq. Xe are Kr, U-chain and Th-chain

39　　　　　　　　

　　　　　　　　　　　　　

Kr 0.1ppm

DM signal(10-6 pb, 50GeV,100 GeV)

0 200 400 600 800

1

102

10-2

10-4

10-6

cpd/

kg/k

eV

energy (keV)

Target = Xe

Kr concentration in Xe

85Kr makes BG in low enegy region

Kr can easily mix with Xe because both Kr and Xe are rare gas

Commercial Xe contains a few ppb Kr

40

XMASS succeeds to reduce Kr concentration in Xe

from ~3[ppb] to 3.3(±1.1)[ppt] with one cycle (~1/1000)

~3mLower

Higher

~1%

~99%

Purified Xe: 3.3±1.1 ppt Kr (measured)

Off gas Xe:330±100 ppb Kr(measured)

Raw Xe: ~3 ppb Kr

(preliminary)

(178K)

(180K)Operation@2atm

• Processing speed : 0.6 kg / hour• Design factor : 1/1000 Kr / 1 pass• Purified Xe : Off gas = 99:1

Xe purification system

Boiling point

(@2 atm)

Xe 178.1K

Kr 129.4K

41

HIT

HIT

HIT

HIT

HIT

?

MCIf true vertex is usedfor fiducial volume cut

10-2

1

10-1

Energy (keV)0 1000 2000 3000

No wall effect

Remaining problem: wall effect (only for the prototype detector)

Scintillation lights at the dead angle from PMTs give quite uniform 1 p.e. signal for PMTs, and this cause miss reconstruction as if the vertex is around the center of detector

Deadangle

This effect does not occur with the sphere shape 800 kg detector

42

800 kg detector