XMASS experiment WIN 05 8 th June 2005

XMASS experiment WIN 05 8 th June 2005 A. Takeda for the XMASS collaboration Kamioka Observatory, ICRR, University of Tokyo 1. Introduction 2. R&D status using prototype detector 3. Summary 1

1. Introduction Ø What’s XMASS Multi purpose low-background experiment with liq. Xe l Xenon MASSive detector for solar neutrino (pp/7 Be) l Xenon neutrino MASS detector (bb decay) l Xenon detector for Weakly Interacting MASSive Particles (DM search) Solar neutrino Dark matter Double beta 2

Ø Why liquid xenon l Large Z (=54) Self-shielding effect l Large photon yield (~42 photons/ke. V ~ Na. I(Tl)) Low threshold l High density (~3 g/cm 3) Compact detector (10 ton: sphere with diameter of ~2 m) l Purification (distillation) l No long life radioactive isotope l Scintillation wavelength (175 nm, detected directly by PMT) l Relative high temperature (~165 K) 3

Ø Key idea: self-shielding effect for low energy events External g ray from U/Th-chain Volume for shielding Fiducial volume PMTs BG normalized by mass Single phase liquid Xe 23 ton all volume 20 cm wall cut 30 cm wall cut (10 ton FV) Large self-shield effect 0 1 Me. V 2 Me. V 3 Me. V 4

Ø Strategy of the scale-up 100 kg Prototype 10 ton detector 800 kg detector With light guide ～ 30 cm ～ 80 cm ～ 2. 5 m R&D Dark matter search We are now here Multipurpose detector (solar neutrino, bb …) 5

Trend of Dark matter (WIMPs) direct searches l Recoiled nuclei are mainly observed by 3 ways Scintillation Na. I, Xe, Ca. F 2, etc. Phonon Ge, Te. O 2, Al 2 O 3, Li. F, etc Ge, Si Ionization Ge l Taking two type of signals simultaneously is recent trend CDMS, EDELWEISS: phonon + ionization u g ray reduction owing to powerful particle ID u However, seems to be difficult to realize a large and uniform detector due to complicated technique 6

Strategy chosen by XMASS l Make large mass and uniform detector (with liq. Xe) Same style as successful experiments of Super-K, SNO, Kam. LAND, etc. l Reduce g ray BG by fiducial volume cut (self shielding) Super-K SNO Kam. LAND 7

Ø 800 kg detector Main purpose: Dark Matter search External g ray BG: 60 cm, 346 kg 40 cm, 100 kg Achieved pp & 7 Be solar n ~80 cm diameter l ~800 -2” PMTs immersed into liq. Xe l 70% photo-coverage ~5 ke. Vee threshold Expected dark matter signal (assuming 10 -42 cm 2, Q. F. =0. 2 50 Ge. V / 100 Ge. V, ) 8

Geometry design l A tentative design (not final one) 12 pentagons / pentakisdodecahedron l Total 840 hex PMTs immersed into liq. Xe l 70% photo-coverage l Radius to inner face ~43 cm This geometry has been coded in a Geant 4 based simulator 9

5. 8 cm (edge to edge) Hamamatsu R 8778 MOD(hex) l Hexagonal quartz window l Effective area: f 50 mm (min) l QE <~25 % (target) l Aiming for 1/10 lower background than R 8778 5. 4 cm 0. 3 cm (rim) 12 cm c. f. R 8778 U 1. 8± 0. 2 x 10 -2 Bq Th 6. 9± 1. 3 x 10 -3 Bq 40 K 1. 4± 0. 2 x 10 -1 Bq l Prototype has been manufactured already l Now, being tested 10

Expected sensitivities Cross section to nucleon [pb] 10 -4 10 -6 10 -8 10 -10 XMASS FV 0. 5 ton year Eth = 5 ke. Vee~25 p. e. , 3 s discovery w/o any pulse shape info. 106 104 102 1 Edelweiss Al 2 O 3 Tokyo Li. F Modane Na. I CRESST UKDMC Na. I XMASS(Ann. Mod. ) NAIAD 10 -2 XMASS(Sepc. ) 10 -4 l Large improvements will be expected SI ~ 10 -45 cm 2 = 10 -9 pb SD~ 10 -39 cm 2 = 10 -3 pb Plots except for XMASS: http: //dmtools. berkeley. edu Gaitskell & Mandic 11

2. R&D status using prototype detector 100 kg prototype Ø Main purpose l Confirmation of estimated 800 kg detector performance ~30 cm cube 3 kg fiducial With light guide version u Vertex and energy reconstruction by fitter u Miss fitting due to dead angle of the cubic detector (“wall effect”, will be explained later) can be removed with light guide u Self shielding power l BG study u Understanding of the source of BG u Measuring photon yield and its attenuation length 12

Ø 100 kg prototype detector In the Kamioka Mine (near the Super-K) 2, 700 m. w. e. OFHC cubic chamber 54 2 -inch low BG PMTs Hamamatsu R 8778 16% photocoverage Liq. Xe (31 cm)3 Gamma ray shield Mg. F 2 window 13

4 p shield with door material 1. 0 m 1. 9 m thickness Polyethylene 15 cm Boron 5 cm Lead 15 cm EVOH sheets 30μm OF Cupper 5 cm Rn free air (~3 m. Bq/m 3) 14

Ø Progress so far l 1 st run (Dec. 2003) u Confirmed performances of vertex & energy reconstruction u Confirmed self shielding power for external g rays u Measured the internal background concentration l 2 nd run (Aug. 2004) u Succeeded to reduce Kr from Xe by distillation u Photo electron yield is increased u Measured Rn concentration inside the shield l 3 rd run (Mar. 2005) with light guide u Confirmed the miss fitting (only for the prototype detector) was removed u Now, BG data is under analysis 15

Ø Vertex and energy reconstruction Reconstruction is performed by PMT charge pattern (not timing) Reconstructed here 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. exp(- m ) m n Log(L) = å Log( ) n ! PMT QADC L: likelihood F(x, y, z, i) x total p. e. m: S F(x, y, z, i) n: observed number of p. e. F(x, y, z, i): acceptance for i-th PMT (MC) VUV photon characteristics: Lemit=42 ph/ke. V tabs=100 cm tscat=30 cm FADC Hit timing === Background event sample === QADC, FADC, and hit timing 16 information are available for analysis

1. Performance of the vertex reconstruction Collimated g ray source run from 3 holes (137 Cs, 662 ke. V) hole C hole B hole A DATA MC → Vertex reconstruction works well + + + C BA 17

2. Performance of the energy reconstruction Collimated g ray source run from center hole 137 Cs, 662 ke. V All volume 20 cm FV 10 cm FV s=65 ke. V@peak (s/E ~ 10%) Similar peak position in each fiducial. No position bias → Energy reconstruction works well 18

Ø Demonstration of self shielding effect z position distribution of the collimated g ray source run → Data and MC agree well γ 19

Event rate (/kg/day/ke. V) Shelf shielding for real data and MC ~1. 6 Hz, 4 fold, triggered by ~0. 4 p. e. 3. 9 days livetime REAL DATA Aug. 04 run preliminary MC simulation All volume 20 cm FV 10 cm FV (3 kg) 10 -2/kg/day/ke. V Miss-reconstruction due to dead-angle region from PMTs. l Good agreement (< factor 2) l Self shielding effect can be seen clearly. l Very low background (10 -2 /kg/day/ke. V@100 -300 ke. V) 20

Ø Internal backgrounds in liq. Xe were measured Main sources in liq. Xe are Kr, U-chain and Th-chain l Kr = 3. 3± 1. 1 ppt (by mass spectrometer) → Achieved by distillation l U-chain = (33± 7)x 10 -14 g/g (by prototype detector) Delayed coincidence search (radiation equilibrium assumed) 214 Bi 214 Po 210 Pb a (7. 7 Me. V) b (Q=3. 3 Me. V) t 1/2=164 ms l Th-chain < 23 x 10 -14 g/g(90%CL) (by prototype detector) Delayed coincidence search (radiation equilibrium assumed) 208 Po 212 Bi 212 Po a (8. 8 Me. V) b (Q=2. 3 Me. V) t 1/2=299 ns 21

Kr concentration in Xe cpd/kg/ke. V l 85 Kr makes BG in low enegy region 102 Target = Xe Kr 0. 1 ppm 1 10 -2 DM signal 10 -4 (10 -6 pb, 50 Ge. V, 100 Ge. V) 10 -6 0 200 l Kr can easily mix with Xe because both Kr and Xe are rare gas 400 600 800 energy (ke. V) l Commercial Xe contains a few ppb Kr 22

Xe purification system l XMASS succeeds to reduce Kr concentration in Xe from ~3[ppb] to 3. 3(± 1. 1)[ppt] with one cycle (~1/1000) • Processing speed : 0. 6 kg / hour Boiling point (@2 atm) • Design factor : 1/1000 Kr / 1 pass • Purified Xe : Off gas = 99: 1 Raw Xe: ~3 ppb Kr Lower (178 K) ~3 m ~1% Xe 178. 1 K Kr 129. 4 K Off gas Xe: 330± 100 ppb Kr (measured) Purified Xe: Operation@2 atm Higher (180 K) ~99% 3. 3± 1. 1 ppt Kr (measured) (preliminary) 23

Summary of BG measurement Now (prototype detector) Goal (800 kg detector) 1/100 l g ray BG ~ 10 -2 cpd/kg/ke. V → Increase volume for self shielding → Decrease radioactive impurities in PMTs (~1/10) l 238 U l 232 Th = (33± 7)× 10 -14 g/g → Remove by filter < 23× 10 -14 g/g (90% C. L. ) → Remove by filter (Only upper limit) l Kr = 3. 3± 1. 1 ppt → Achieve by 2 purification pass 1/33 1/12 1/3 10 -4 cpd/kg/ke. V 1× 10 -14 g/g 2× 10 -14 g/g 1 ppt Very near to the target level! 24

Ø Remaining problem: wall effect (only for the prototype detector) HIT 1 ? Dead angle MC If true vertex is used for fiducial volume cut 10 -1 HIT HIT 10 -2 l Scintillation lights at the dead angle 0 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 1000 2000 3000 Energy (ke. V) No wall effect This effect does not occur with the sphere shape 800 kg detector 25

Ø Prototype detector with light guide Purpose: remove the wall effect and understand the source of BG in the DM region Active veto Fiducial 10 X 10 cm 3 (~3 kg Xe) 10 cm PTFE light guide (UV reflection) 6 pieces 10 cm 26

Light guide setup l Edging of PTFE surfaces 222 Rn decays (210 Pb b, 64 ke. V endpoint) implanted in PTFE surfaces might make the dominant BG α 222 Rn 218 Po air We edged the PTFE inside ~10μm PTFE α 214 Pb α N. J. T. Smith et al. , Phys. Lett. B 485 (2000) 9 Position distribution of 210 Pb (in Na. I) 210 Pb Implanted to ~0. 1μm 0 0. 05 0. 10 Z [mm] 0. 15 Recoil process implants 30% of the original surface Rn decays 27

Expected BG spectrum Efficiency curve 0. 48 p. e. /ke. V 1. 0 0. 8 0. 6 cpd/ke. V/kg Efficiency ・ MC simulation was done with GEANT 3 Signal window (10 -15 ke. V) 0. 4 1 BG spectrum PMT K PMT Th PMT U Fast neutron BG (90% C. L. upper limit) 10 -2 0. 2 0 0 20 40 60 80 100 energy(ke. V) Efficiency~30% @10 ke. V 10 -4 0 20 40 60 80 100 energy(ke. V) Expected BG~10 -2 cpd/ke. V/kg → Very low BG ~ 10 -2 cpd/ke. V/kg @ <100 ke. V 28

Result 1: comparing the data taken with and without light guide Collimated g ray source run from hole-B (137 Cs, 662 ke. V) l with light guide Counts l w/o light guide 10 cm fiducial Energy [ke. V] Hole-B fiducial volume Energy [ke. V] Reduce the events due to the wall effect 29

10 Outside the light guide(Data) 1 Live time (3. 3 days) 10 -1 10 -2 events/ke. V/kg/day Result 2: Obtained energy spectrum outside the light guide 10 1 10 -2 10 -3 10 -4 0 1000 2000 3000 energy(ke. V) Outside light guide(MC) 1000 2000 3000 energy(ke. V) l Good agreement (< factor 2) l Trigger rate is same as the measurement witout guide (Aug. 2004) 30

3. Summary l XMASS experiment: Multi purpose low-background experiment with large mass liq. Xe l 800 kg detector: Designed for dark matter shearch mainly, and 102 improvement of sensitivity above existing experiments is expected l R&D with the 100 kg prototype detector Most of the performances required for 800 kg detector are confirmed 31

32

XMASS collaboration • ICRR, Kamioka Y. Suzuki, M. Nakahata, Y. Itow, S. Moriyama, M. Shiozawa, Y. Takeuchi , M. Miura, Y. Koshio, K. Ishihara, K. Abe, A. Takeda, T. Namba, H. Ogawa, S. Fukuda, Y. Ashie, A. Minamino, R. Nambu, J. Hosaka, K. Taki, T. Iida, K. Ueshima • ICRR, RCNN T. Kajita, K. Kaneyuki • Saga Univ. H. Ohsumi, Y. Iimori, • Tokai Univ. K. Nishijima, T. Hashimoto, Y. Nakajima, Y. Sakurai • Gifu Univ. S. Tasaka • Waseda Univ. S. Suzuki, K. Kawasaki, J. Kikuchi, T. Doke, A. Ota • Yokohama National Univ. S. Nakamura, T. Fukuda, S. Oda, N. Kobayashi, A. Hashimoto • Miyagi Univ. of Education Y. Fukuda, T. Sato • Seoul National Univ. Soo-Bong Kim, In-Seok Kang • INR-Kiev Y. Zdesenko, O. Ponkratenko • UCI H. Sobel, M. Smy, M. Vagins, P. Cravens • Sejong univ. Y. Kim • Ewha Womans Univ. K. Lim 33 • Indiana Univ. M. Ishitsuka

Electronics of light guide run (Mar. 2005) Fan-out ADC delay Inner PMT× 6 PMT 400 ns Sum Amp PMT ID sum OD sum ID x 8 PMT Sum Amp Outer PMT× 48 FADC(8 ch/500 MHz) 1μs ID sum OD sum PMT Gain : 8. 25× 106 FADC(2 ch/250 MHz) 16μs Discri(VME) Discri(NIM)~1/4 p. e. thres. ID ≧ 2 hit Gate generator Trigger module x 8 OD ≧ 4 hit 34

measurement Data taking : 3/9 ~ 3/19, 2005 ➢Background run : 3. 7 days runtime, 3. 3 days livetime ➢Trigger rate 1. 5 Hz (inner 0. 2 Hz, outer 1. 4 Hz) 137 Cs / 60 Co / 57 Co / 133 Ba source run ➢Calibration run : ➢ ID hit event FADC ID sum OD hit event TDC 35

Ø Background study cpd/kg/ke. V Expected spectra in all volume 1 Outside of the shield 238 U in PMTs 232 Th in PMTs 10 -1 40 K 210 Pb 10 -2 0 1000 in PMTs in lead shield 2000 3000 l Outside of the shield 0. 71 cm-2 s-1 (>500 ke. V) l RI sources in PMTs 238 U : 1. 8× 10 -2 Bq/PMT 232 Th : 6. 9× 10 -3 Bq/PMT 40 K : 1. 4× 10 -1 Bq/PMT l 210 Pb in the lead shield 250 Bq/kg ke. V 36

• Low energy calibration source (1) Scintillation photons Liq. Xe X-ray D Attenuation length of 20 ke. V x-ray in liq. Xe is short ~ 50 μm The overall size of the source itself should be small not to block the scintillation photons • EC decaying nuclei preferable X-rays • Candidates : 71 Ge(463 d), 153 Gd(263 d), 103 Pd(17 d) Irradiate neutrons to natural Pd wire of 10 μm diameter 102 Pd(n, γ) 103 Pd EC decay of 103 Pd produce 20 ke. V x-ray 37

• Low energy calibration source (2) ☆ 125 I(X-ray source) : 27. 5 ke. V (59. 9 day) ☆ Temperature Range : -200 ~ +100 in Centigrade ☆ Overall source diameter < 20 mm ☆ Weak source ~ a few k. Bq Material A F= 10 mm Liq. Xe 125 I (1 k. Bq) Electrodeposition 5 mm Length 60 mm Source position Coating Material B Thick=3 mm 38

• Plan of prototype detector ☆ Introduce RI source(103 Pd, 125 I, …) inside the chamber → Source driving system is ready → Detailed study of the energy and vertex fitter Motor Wire Low energy g source (103 Pd, 125 I, …) Position accuracy is within 1 mm 39

40

41

42