Search for WIMPs with background evaluation in a fiducial volume by - - PowerPoint PPT Presentation

Search for WIMPs with background evaluation in a fiducial volume by XMASS-I.

35th International Cosmic Ray Conference (ICRC2017) July 12-20, 2017 BEXCO, BUSAN, KOREA Institute for Cosmic Ray Research, Univ. of Tokyo Hiroshi Ogawa for XMASS collaboration

Introduction

• WIMPs elastically scatter off nuclei in

targets, producing nuclear recoils. Xenon is a popular target for WIMP searches.

• In this presentation, high precision WIMP

search in XMASS is processed using fiducial volume cut.

• XMASS concept : FV cut and “Self shielding” in

LXe reduces the external background.

• Background in fiducial volume is evaluated

with systematic error.

• Then, expected background and WIMPs

signal are compared with the reduced data. WIMPs cross section limit is estimated for each WIMP mass.

XMASS concept

Dark Matter （WIMP）

Deposit Energy

The XMASS detector

• Single phase (scintillation only) liquid Xenon detector.
• 630 hexagonal & 12 round PMTs with 28-39% Q.E.
• photocathode coverage: ~ 62%
• Each PMT signal is recorded by 10-bit 1GS/s waveform

digitizers

• The inner detector surface is covered by copper plate to

reduce gaps in this surface.

• 832 kg Lxe sensitive volume (0.8 mf). Self-shielding +

reconstruction can reduce BG from external and detector surface. 1.2m diameter

Copper ring covers the aluminum seal (→huge background origin!) of PMT.

Background evaluation in XMASS-I

• RI screening of detector materials was

implemented using HPGe detector.

• RI activity are estimated by spectrum

fitting for > 400pe (~30keV) with constraints from related assays screening results.

Background evaluation in XMASS-I (Cont.)

• Pb210 in copper surface & bulk.
• Alpha-ray events are identified using

scintillation decay time.

• Copper surface/bulk 210Pb concentration are

estimated by alpha-ray from 210Po decay.

• ※Alpha from bulk makes continues spectrum by

energy loss in copper

• Contamination with 210Pb (~20 mBq/kg)

in the bulk of the oxygen-free copper was measured by a low background alpha-particle counter (XIA Ultra-Lo-1800)

• RI in liquid xenon.
• Coincidence analysis is used for Rn222 as Bi214-

Po214 (164μs) decay, Kr85 as beta-gamma (1.015μs, 0.434%).

• Concentration of C14 and Ar39 in Lxe are

estimated from spectrum fits.

Po210 Copper bulk p.e.

0 30 60 90 120 x 103

Po210 Copper surface

Event reduction

• (1) pre-selection : remove electronic noise

events, Cherenkov events and so on.

• To remove surface events, two kind of

vertex reconstruction are used.

• (2) Timing based vertex reconstruction

R(T) : First hit timing of each PMT is used. Position is fitted by likelihood. Events are kept if R(T)<38cm.

• (3) NPE based vertex reconstruction R(PE) :

an acceptance map is made for both volume and surface events with the XMASS MC. position is calculated by

• likelihood. Also energy is estimated. Events

are selected if R(PE) <20cm yielding a ~100kg fiducial mass.

• Dataset : 2013Nov 20 - 2016Mar 29
• >2yr of stable operation.
• Livetime :705.879days.
• In pre-selection, Reduction of

Cherenkov event is effective.

• PMT hit timing is used.
• Main origin of Cherenkov : K40 beta in

PMT photo-cathode.

• After pre-selection applying R(T) +

R(PE) gives another O(10-3) reduction.

• After applying all cuts (pre-selection

+ R(T) cut + R(PE) cut), event rate becomes ~4×10-3 /day/kg/keVee @ 5-5.5keV.

7

Event/day/kg/keVee Reconstruction energy [keVee]

Pre-selection + R(T)<38cm+R(PE)<20cm

Background MC

• Background MC is generated

using XMASS MC for each RI’s decay mode and its activity.

• Optical parameters of LXe are

traced with our Co57 and Co60 regular (weekly) calibration.

• Same event reduction is applied

to background MC, which has the same livetime as the dataset.

• ~90% of remaining BG is of

detector surface origin (not internal BG). => mis- reconstructed events.

Reconstruction energy [keVee] Event/day/kg/keVee

Blue histogram : all MC with statistical error Pb210 Plate bulk N.R. Pb210 Ring bulk Internal PMT Al seal Pb210 Surface PMT gamma

Pre-selection + R(T)<38cm+R(PE)<20cm

Mis-reconstructed events :

• Remaining events most likely
• ccurred in places with no

direct path for light to hit a nearby photocathode and thus get reconstructed in fiducial

• volume. Possibilities are :
• Surface events directly on

the inner surface of copper plates.

• Light leakage from a gap. :

An event which is generated inside a gap often has a special pattern. But reconstruction put it into the fiducial volume.

“Butt” of plate Cu ring ⇔～100μm

PMT holder

Data：may happen in gap reconR=16.4cm Pb210 surface MC reconstE=2-5keV Plate surface Window surface Surface of detector Fiducial volume region R=20cm

Systematic error for BG

• The systematic error evaluation

for mis-reconstruction requires an understanding of surface condition uncertainly.

• It can be estimated considering

the specific condition of each relevant detector part.

Reconstruction energy [keVee] Event/day/kg/keVee

Contents Systematic error (each event position and RI) 2-15keVee 15-30keVee (1) Plate gap dependence +6.2/-22.8% +1.9/-6.9% (2) Ring roughness +6.6/-7.0% +2.0/-2.1% (3) Cu ref dependence +5.2/-0.0% +2.5/-0.0% (4) Plate floating +0.0/-4.6% +0.0/-1.4% (5) Al seal dependence +0.7/-0.7% +0.0/-0.0% (6) Reconstruction +3.0/-6.2% +0.0/-0.0% (7) Timing (decaytime, TTS) +4.6/-2.9% +0.4/-5.3% (8) Timing (response in detector surface) +0.0/-8.0% +0.0/-0.0% (9) Absorption & scattering +0.7/-6.7% +1.5/-1.1% (10) Dead tube origin +10.3/-0.0% +45.2/-0.0% (11) N.R. +0.7/-0.7% +0.0/-0.0%

(1) Plate gap geometry (4)Plate floating (3) Cu reflection dependence

Microscope obs. for ring side

(2)Ring roughness Sum of remaining BGMC Surface condition validity

Systematic error for BG (cont)

• The validity of reconstruction,

dependence on optical properties

• f Lxe, detector response at

detector surface and so on are treated as systematic errors.

Contents Systematic error (each event position and RI) 2-15keVee 15-30keVee (1) Plate gap dependence +6.2/-22.8% +1.9/-6.9% (2) Ring roughness +6.6/-7.0% +2.0/-2.1% (3) Cu ref dependence +5.2/-0.0% +2.5/-0.0% (4) Plate floating +0.0/-4.6% +0.0/-1.4% (5) Al seal dependence +0.7/-0.7% +0.0/-0.0% (6) Reconstruction +3.0/-6.2% +0.0/-0.0% (7) Timing (decaytime, TTS) +4.6/-2.9% +0.4/-5.3% (8) Timing (response in detector surface) +0.0/-8.0% +0.0/-0.0% (9) Absorption & scattering +0.7/-6.7% +1.5/-1.1% (10) Dead tube origin +10.3/-0.0% +45.2/-0.0% (11) N.R. +0.7/-0.7% +0.0/-0.0%

Reconstruction energy [keVee] Event/day/kg/keVee ※histogram color Data, BGMC ※bar color Systematic error RI activity error MC stat error

WIMP search region (2-15keVee)

Search for WIMPs with background evaluation in the fiducial volume.

• We performed a WIMP dark matter

search with events of the final reduction sample with consideration for background events predicted by background MC as preliminar result.

• The energy spectrum of the data was

fitted with background MC plus WIMP MC in the energy range of 2-15keVee. The systematic error is included both background MC and WIMP MC.

• A 90% C.L. upper limit on the WIMP-

nucleon cross section was derived.

• Our preliminary exclusion limit is

2.2x10-44cm2 for 60GeV WIMPs mass.

+ Data ・ BGMC + WIMP(90% C.L.) ■ 1s error for best fit WIMP 60 GeV, 2.2e-44 cm2(90%CL)

Conclusion

• A high sensitivity WIMP search are conducted using 706 days of data.
• Fiducial volume cuts are applied to both data and background MC.
• The systematic error is evaluated for background MC.
• A WIMP dark matter search in the events of the final reduction sample

from the data considering known backgrounds simulated with our background MC is derived as preliminar result. A preliminary exclusion limit is 2.2x10-44cm2 for 60GeV WIMPs.

backup

Light leakage from Alseal

• Low energy spectrum has large

dependence of Alseal geometry.

• We observed the Alseal

geometry by PMT cutting.

• We evaluated 3 kind of model

and estimate the spectrum shape.

• These model dependence is

evaluated as systematic error of background.

（１） Observed geometry in room temperature （２）(1) + Low temp shrink. (center model) Data (w/o FV cut） BGMC(Al seal） (1)Observed geometry (room temp) (2) (1) + Low temp shrink.=> center model (3) (2) + shrink more BGMC(others) (3) (2) + shrink more Event/day/kg/keV NPE scale (Co57 122keV) energy [keV] w/o FV cut

To improve the sensitivity :

• Increase the fiducial volume
• Select ultra low BG detector

material ➜ material screening for RI measurement.

• Discriminate against BG events,

especially surface BG. ➜ Developed a new PMT.

• Reduce inner detector RI ➜

filtering(222Rn), Distillation (85Kr) and screening for Rn emanation.

D.C.Mailing Ph.D (2014) Fig 1.5

XMASS-I

Internal RI in LXe Leakage from detector material/surface RI

New PMTs for future XMASS

• Surface events can be identified and rejected very effectively by new dome-

shaped PMTs, which have high and uniform collection efficiency for whole area.

• Performance test was carried out using the first batch of the new PMTs.
• Reduction of radioactivity in PMT parts was done.

2inch hex shape current PMT R10789 (for XMASS-I)

3inch dome shape new PMT R13111

17