Direct dark matter search by annual modulation in XMASS-I

Contents lists available atScienceDirect

Physics

Letters

B

www.elsevier.com/locate/physletb

Direct

dark

matter

search

by

annual

modulation

in

XMASS-I

XMASS

Collaboration



K. Abe

a

,

e

_,

_{K. Hiraide}

a

,

e

_,

_{K. Ichimura}

a

,

e

_,

_{Y. Kishimoto}

a

,

e

_,

_{K. Kobayashi}

a

,

e

_,

_{M. Kobayashi}

a

,

e

_,

S. Moriyama

a

,

e

,

M. Nakahata

a

,

e

,

T. Norita

a

,

H. Ogawa

a

,

e

,

H. Sekiya

a

,

e

,

O. Takachio

a

,

A. Takeda

a

,

e

,

M. Yamashita

a

,

e

,

B.S. Yang

a

,

e

,

N.Y. Kim

b

,

Y.D. Kim

b

,

S. Tasaka

c

,

1

,

K. Fushimi

d

,

J. Liu

e

,

2

,

K. Martens

e

,

Y. Suzuki

e

,

B.D. Xu

e

,

R. Fujita

g

,

K. Hosokawa

g

,

K. Miuchi

g

,

Y. Onishi

g

,

N. Oka

g

,

Y. Takeuchi

g

,

e

,

Y.H. Kim

h

,

b

,

J.S. Lee

h

,

K.B. Lee

h

,

M.K. Lee

h

,

Y. Fukuda

i

,

Y. Itow

j

,

f

,

R. Kegasa

j

,

K. Kobayashi

j

,

K. Masuda

j

,

H. Takiya

j

,

K. Nishijima

k

,

S. Nakamura

l

a_{KamiokaObservatory,InstituteforCosmicRayResearch,theUniversityofTokyo,Higashi-Mozumi,Kamioka,Hida,Gifu,506-1205,Japan} b_{CenterofUndergroundPhysics,InstituteforBasicScience,70Yuseong-daero1689-gil,Yuseong-gu,Daejeon,305-811,SouthKorea} c_{InformationandMultimediaCenter,GifuUniversity,Gifu501-1193,Japan}

d_{Institute ofSocio-ArtsandSciences,TheUniversityofTokushima,1-1MinamijosanjimachoTokushimacity,Tokushima,770-8502,Japan} e_{KavliInstituteforthePhysicsandMathematicsoftheUniverse(WPI),theUniversityofTokyo,Kashiwa,Chiba,277-8582,Japan}

f_{Kobayashi-MaskawaInstitutefortheOriginofParticlesandtheUniverse,NagoyaUniversity,Furo-cho,Chikusa-ku,Nagoya,Aichi,464-8602,Japan} g_{DepartmentofPhysics,KobeUniversity,Kobe,Hyogo657-8501,Japan}

h_{KoreaResearchInstituteofStandardsandScience,Daejeon305-340,SouthKorea} i_{DepartmentofPhysics,MiyagiUniversityofEducation,Sendai,Miyagi980-0845,Japan} j_{SolarTerrestrialEnvironmentLaboratory,NagoyaUniversity,Nagoya,Aichi464-8602,Japan} k_{DepartmentofPhysics,TokaiUniversity,Hiratsuka,Kanagawa259-1292,Japan}

l_{DepartmentofPhysics,FacultyofEngineering,YokohamaNationalUniversity,Yokohama,Kanagawa240-8501,Japan}

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory:

Received3March2016

Receivedinrevisedform11April2016 Accepted25May2016

Availableonline30May2016 Editor:S.Dodelson

Keywords:

Darkmatter Annualmodulation Liquidxenon

AsearchfordarkmatterwasconductedbylookingforanannualmodulationsignalduetotheEarth’s rotationaroundtheSunusingXMASS,asinglephaseliquidxenondetector.Thedatausedforthisanalysis was 359.2livedaystimes832kgofexposureaccumulatedbetweenNovember2013andMarch 2015. WhenweassumeWeaklyInteractingMassiveParticle(WIMP)darkmatterelastically scatteringonthe targetnuclei,theexclusionupperlimitoftheWIMP–nucleoncrosssection4.3×10−41_cm2_{at8 GeV/c}2

was obtainedandweexcludealmostallthe DAMA/LIBRAallowedregioninthe6to16GeV/c2range at∼10−40_cm2_{.Theresultofasimplemodulationanalysis,withoutassuminganyspecificdarkmatter}

modelbutincludingelectron/

γ

events,showedaslightnegativeamplitude.Thep-valuesobtainedwith two independentanalysesare 0.014 and0.068 for nullhypothesis,respectively. We obtained 90%C.L. upper bounds that can be used totest various models.Thisis thefirst extensiveannualmodulation searchprobingthisregionwithanexposurecomparabletoDAMA/LIBRA.

©2016TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP3_.

 _{E-mailaddress:xmass.publications2@km.icrr.u-tokyo.ac.jp.}

1 _{NowatKamiokaObservatory,InstituteforCosmicRayResearch,theUniversity}

ofTokyo,Higashi-Mozumi,Kamioka,Hida,Gifu,506-1205,Japan.

2 _{NowatDepartmentofPhysics,theUniversityofSouthDakota,Vermillion,SD}

57069,USA.

1. Introduction

There isstrongevidencethat about5timesmoredarkmatter existsintheuniversethanordinarymatter.Despiteitsprominence, we do not yet knowwhat darkmatter is [1].Among many can-didates for darkmatter particles, WIMPs are well motivatedand havereceivedthemostattentiontodate.However,collider experi-mentsattheLHCdonotshowanyindicationforsuchparticlesso far [1].And noexperimentalindication forastandard WIMPwas

http://dx.doi.org/10.1016/j.physletb.2016.05.081

0370-2693/©2016TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP3_.

foundinhighsensitivitydirectsearchexperimentssuchasLUX

[2]

, XENON100

[3]

andSuperCDMS

[4]

either.Ontheother hand,that appears tocontradict experiments that report signals interpreted as

∼

10 GeV/c2 _{light WIMP darkmatter [8–10] for manyyears.}

In this situation, light mass WIMPs or other dark matter candi-datesaregettingmoreattention.Infact,XMASS,ahighlightyield andlow backgrounddetector, probed thispossibility and looked forsignals not only from nuclear recoils butalso from electrons andgamma raysemanatingfrominteractions ofother candidates suchasaxion-likeparticles,Super-WIMPsandsoon

[5–7]

.

Themost significantresult isthat of theDAMA/LIBRA experi-ment atthe Gran Sasso National Laboratory in Italy which indi-cated an annual modulation signature [11]. The Earth’s velocity relativetothedarkmatterdistributionchangesastheEarthmoves aroundtheSunandproducessuchamodulationinthedarkmatter signalrate.Thismodulationcanbeobservedwithterrestrial detec-tors

[12]

.The amplitude ofthe modulationcan be changedfrom positive(i.e.higherrateinJune thaninDecember)tonegativeat cross-overenergy [13] andit is possible to observethis effectif thedetectorthreshold islower than that energy.For100 GeV/c2

WIMP massand a Xe target, this isabout 20 keVnuclear recoil energyandit dependson theWIMPmassandthetarget materi-als.

TheDAMA/LIBRAexperimentreportedan observationofevent rateannualmodulationwitha9

σ

significancein1.33 ton

·

yearof datatakenover14annualcycleswith100to250 kgofNaI(Tl) de-tectors.TheirsignalmaybecausedbylightWIMPs,orothertypes ofdarkmatter producingelectronsorgammarays. Insuchcases, the signal is not observable to direct search experiments if they removeelectron events.In thissituation,darkmattermodels,for instance,withinteraction via darkmatter–electron scattering be-comewellmotivatedwhichproducekeVenergydepositioninthe detectorbecausethey provide aexplanation forthe DAMA/LIBRA result while avoiding other direct detection constraints [14–16]. Recently, in addition to the WIMP search result [3], an annual

modulation search was carried out by the XENON group using

onlyelectronic recoilevents intheir two phase Xedetector with the 34 kg fiducial volume in 224.6 live days data [17]. The re-sult disfavored the interpretation of the DAMA/LIBRA as WIMP– electron scattering through axial-vector coupling. XMASS uses a singlephasetechnologytoobserveonlyscintillationlightby look-ing forboth types ofsignals withoutany electricfield. Although

XMASS has a modest background ratelike that of DAMA/LIBRA,

XMASShasalargermassof832 kgofliquidxenonand,therefore, isabletoreachtheDAMA/LIBRAexposureinshorttime.Whilethe backgroundinthisrecentmodulationstudybytheXENON exper-imentislower, XMASShasa larger target mass andsignificantly longer exposure time. We will discussthe sensitivity later. Note thatXMASSteststhismodulationhypothesiswithalmosthalfthe energythreshold(

∼

1 keV)thantheirsinadifferentenvironment andundergroundsite.

2. TheXMASSexperiment

The XMASS detector is located at the Kamioka Observatory

(overburden2700m.w.e)inJapan.Thedetaileddesignand perfor-mancearedescribed in

[18]

.The detectorisimmersedinawater tank,10 m indiameterand10.5 m inheight, whichisequipped with72HamamatsuH3600photomultipliertubes(PMTs),andacts asanactivemuonvetoandapassiveradiationshieldagainst neu-tronsandgammaraysfromthesurroundingrock.642high

quan-tumefficiency(28–40% at175 nm)HamamatsuR10789PMTs are

mountedintheliquidxenondetector,anapproximatespherewith anaverageradius of40 cm.Thegain ofthePMTswas monitored weeklywithablueLED embeddedintheinnersurfaceofthe

de-tector.Thescintillationlightyieldresponsewastracedbyinserting a57Cosource

[19]

intothedetectorevery oneortwoweeks.The numberofeventsforeachsourcepositionwasabout20,000.

In November 2013, after refurbishing the detector to reduce the radioactivebackgroundfromthealuminum seal ofthePMTs’ window that was identified inthe commissioning run [18], data takingwasresumedwithaboutoneorderofmagnitudeimproved

background by covering these seal parts with plates made of

pure copper.The dataaccumulatedbetweenNovember 2013and

March 2015 were used for this analysis and we selected peri-ods with stabletemperature (172.6–173.0 K)and pressure of Xe (0.162–0.164 MPa absolute). After removing periods of operation withexcessive PMT noise ordata acquisition problems, thetotal livetimebecame359.2 days.

Inthispaper, twodifferentenergy scaleswere used: 1) keVee

represents an electron equivalent energy incorporating all the gamma-raycalibrations intheenergyrangebetween5.9 keVand 122 keV from 55Fe, 109Cd, 241Am and57Co sources by inserting sourcesintothesensitivevolumeofthedetector.Thenon-linearity ofenergyscalewas takenintoaccountwiththosecalibrations us-ing a non-linearity model from Doke et al. [20]. Below 5.9 keV,

we extrapolated based on this model. We found about 15%

en-ergyscaledifferencefromtheNobleElementSimulationTechnique (NEST) [21] at the thresholdenergy of 1.1 keVee (

∼

8

photoelec-trons)inthisanalysis. 2) keVnr denotesthe nuclearrecoilenergy

whichisestimatedfromthelight yieldat122 keVby using non-linearity response measurementatzeroelectricfield in [22].The energythreshold,inthiscase,correspondsto4.8 keVnr.

3. Dataanalysis

Events with4ormorePMThitsina 200 nscoincidence tim-ing windowwithoutamuon vetowereinitially selected.This re-sulted in3

.

3

×

107 events inthe energy regionbetween 1.1and 15 keVee.Inordertoavoideventscausedby afterpulsesofbright

events induced by, for example, high energy gamma-rays or al-pha particles, we rejected events occurring within 10 ms from the previous event andhaving a variance in their hit timings of greater than100 ns(this selectionreducesthe numberofevents to 2

.

8

×

107_{). A ‘Cherenkov cut’ removedevents which produce}

light predominantly fromCherenkov emission, inparticular from thebetadecaysof40_{KinthePMTphotocathode.Eventsforwhich}

more than 60% of their PMT hits arrive in the first 20 ns were classifiedasCherenkov-like events [5](thisselection reduces the number of events to 1

.

9

×

106). Finally, to remove background eventsthatoccurredinfrontofPMTwindow,wegive upper lim-itsonthevaluesof‘Max-photoelectron/Total-photoelectron’where Max-photoelectronandTotal-photoelectronare thelargest photo-electroncountsinonePMTamongallPMTsandthetotalnumber ofphotoelectronsintheevent,respectively(thisselectionreduces thenumber ofeventsto 3

.

6

×

105).These cutvalues variedasa function of photoelectron from about 0.2 at8 photoelectrons to about0.07at50photoelectrons.Thecount rateforthedataafter allthecutsis1.17(0.028)events/day/kg/keVeeat1.1(5.0) keVee.

The 57_{Co calibration data were taken atfrom z}

_{= −}

_{40 cm to}

+

40 cmalongthecenterverticalaxisofthedetectortotrack pho-toelectron yield and optical properties of the liquid xenon [18]. A difference of about 10% was observed as the position depen-denceforthisphotoelectronyield.Thephotoelectronyieldduring the data taking varied about10%. The absorption and scattering lengthforthescintillationlightaswellastheintrinsiclightyield oftheliquidxenonscintillatorareextractedfromthe57_Co

calibra-tiondatatheMonteCarlosimulation

[18]

.Withthatwefoundthat wecantracetheobservedphotoelectronchangeinthecalibration data as a change as the absorption length, while the scattering

Fig. 1. Lightyieldstabilitywasmonitoredwitha57_{Co122keVgammaraysource.}

Therelativeintrinsicscintillationlightyield(Ryield)wasobtainedbycomparingto

calibrationdatawiththeMonteCarlosimulationbyconsideringopticalparameters suchasabsorptionandscatteringlength.

Fig. 2. (Coloronline.)Observedcountrateasafunctionoftimeinthe1.1–1.6 keVee

(=4.8–6.8 keVnr)energyrange.Theblackerrorbarsshowthestatistical

uncer-taintyofthecountrate.Squarebracketsindicatethe1σ systematicerrorforeach timebin.Thesolidanddashedcurvesindicatetheexpectedcountratesassuming7 and8GeV/c2_{WIMPsrespectivelywithacrosssectionof2×10}−40_cm2_wherethe

WIMPsearchsensitivityclosedtoDAMA/LIBRA.

lengthremainsstableat52 cmwithastandarddeviationof

±

0

.

6%. Wethenre-evaluatetheabsorptionlengthandtherelative intrin-siclightyieldtoseethestabilityofthescintillationlightresponse by fixingthe scatteringlengthat52 cm. Theabsoluteabsorption length varied fromabout4 m to 11 m, butthe relative intrinsic lightyield(Ryield)stayedwithin

±

0

.

6% overtheentiredatataking

period (see

Fig. 1

).

Thetimedependenceofthephotoelectronyieldaffectsthe ef-ficiencyofthe cuts.Therefore, weevaluate theabsorption length dependence of the relative cut efficiencies through Monte Carlo simulation. If we normalize the overall efficiency at an

absorp-tion length of 8 m, this efficiency changes from

−

4% to

+

2%

over the relevant absorption range. The position dependence of the efficiency was taken into account as a correlated systematic error (

∼ ±

2

.

5%). This is the dominant systematic uncertainty in thepresentanalysis.Thesecondlargestcontributioncomesfroma gaininstability ofthewaveformdigitizers (CAENV1751)between April 2014 and September 2014 due to a different calibration methodofthedigitizersusedinthatperiod.Thiseffectcontributes anuncertaintyof0.3%totheenergyscale.OthereffectsfromLED calibration,triggerthresholdstability,timingcalibrationwere neg-ligible.Theobservedcountrateaftercutsasafunctionoftime in the energyregion between 1.1 and1.6keVee is shownin Fig. 2.

The systematic errors caused by the relative cut efficiencies are alsoshown.

Toretrievetheannualmodulationamplitudefromthedata,the leastsquaresmethodforthetime-binneddatawasused.Thedata set was divided into 40 time-bins (tbins) with roughly 10 days

of live time each. The data in each time-bin were then further dividedintoenergy-bins(Ebins)withawidthof0.5keVee.Two

fit-tingmethodswere performedindependently.Both ofthem fitall energy- andtime-binssimultaneously.Method1useda‘pullterm’

α

with

χ

2_definedas:

χ

2

=

Ebins



i tbins



j



(

Rdata i,j

−

Rexi,j

−

α

Ki,j

)

2

σ

(

stat

)

2_i,j

+

σ

(

sys

)

2_i,j



+

α

2

,

(1)

where Rdata_i,j ,Rex_i,j,

σ

(

stat

)

i,jand

σ

(

sys

)

i,j aredata,expectedevent

rate, statistical and systematic error, respectively, of the (i-th energy- and j-thtime-) bin.The time isdenoted asthe number of days fromJanuary 1, 2014. Ki,j represents the 1

σ

correlated

systematicerrorontheexpectedeventratebasedon therelative cut efficiencyin that bin.Method 2 used a covariancematrix to propagate the effects ofthe systematicerror. Its

χ

2 _{was defined}

as:

χ

2

=

N



bins

k,l

(

Rdata_k

−

Rex_k

)(

Vstat

+

Vsys

)

−_kl1

(

Rdata_l

−

Rex_l

),

(2)

where Nbins

(

=

Ebins

×

tbins) was the total number of bins and Rdata_k (ex) isthe eventratewherek

=

i

·

tbins

+

j.The matrix Vstat

containsthestatisticaluncertaintiesofthebins,andVsysisthe

co-variancematrixofthesystematicuncertaintiesasderivedfromthe relativecutefficiency.

4. Resultsanddiscussion

We performedtwo analyses, one assuming WIMPinteractions

and the other independent of any specific dark matter model.

HereafterwecalltheformercasetheWIMPanalysisandthelatter amodelindependentanalysis.

In the case of the WIMP analysis, the expected modulation

amplitudes become a function ofthe WIMP mass Ai(mχ

)

asthe WIMP mass mχ determines the recoilenergy spectrum. The ex-pectedrateinabinthenbecomes:

Rex_i,j

=

tj+



12tj tj−12tj



Ci

+

σ

χn

·

Ai

(

mχ

)

cos 2π

(

t

−

t0

)

T



dt

,

(3)

where

σ

χn is the WIMP–nucleon cross section. To obtain the

WIMP–nucleon cross section the data was fitted in the energy

range of 1.1–15 keVee. We assume a standard spherical

isother-mal galactic halo model with the most probable speed of v0

=

220 km

/

s, the Earth’s velocity relative to the dark matter dis-tribution of vE

=

232

+

15 sin2

π

(

t

−

t0

)/

T km

/

s, and a galactic

escapevelocity ofvesc

=

650 km

/

s,alocaldarkmatter densityof

0.3 GeV/cm3_,following

_[13]

_{.Intheanalysis, thesignalefficiencies}

for each WIMP massare estimated fromMonte Carlosimulation ofuniformlydistributed nuclearrecoileventsin theliquidxenon volume. The systematic error of the efficiencies comes from the uncertaintyofliquidxenonscintillationdecaytimeof25

±

1 ns[5]

andisestimatedasabout5%inthisanalysis.The expectedcount ratefor WIMPmassesof 7and8 GeV/c2 _{witha crosssection of}

2

×

10−40cm2 forthespin independentcaseareshownin

Fig. 2

as a function of time after all cuts. This demonstrates the high sensitivity of the XMASS detector to modulation. As both meth-ods found no significant signal, the 90% C.L. upper limit by the ‘pull term’ method onthe WIMP–nucleoncross section isshown in

Fig. 3

.Theexclusionupperlimitof4

.

3

×

10−41_cm2_at8GeV/c2

wasobtained.The

−

1

σ

scintillationefficiencyof

[22]

wasusedto obtain a conservative limit. Toevaluate the sensitivityof WIMP– nucleoncrosssection,wecarriedoutastatisticaltestbyapplying the same analysisto 10,000 dummysampleswith thesame sta-tistical andsystematic errors asdata butwithout modulationby the following procedure.At first, the time-averagedenergy spec-trumwasobtainedfromtheobserveddata.Then,weperformeda toyMonteCarlosimulationtosimulatetimevariationofeventrate

Fig. 3. (Coloronline.) Limitsonthespin-independentelasticWIMP–nucleoncross sectionasafunctionofWIMPmass.ThesolidlineshowstheXMASS90%C.L. ex-clusionfromtheannualmodulationanalysis.The±1σ and±2σ bandsrepresent theexpected90%exclusiondistributions. Limitsaswellasallowedregionsfrom othersearchesbasedoncountingmethodarealsoshown[2,3,23,8–10,5].

ofbackgroundateach energybinassumingthesamelivetimeas data and including systematic uncertainties. The

±

1

σ

and

±

2

σ

bands in Fig. 3 outline the expected 90% C.L. upper limit band forthe no-modulationhypothesis usingthedummysamples.The resultexcludestheDAMA/LIBRA allowed regionasinterpreted in

[8] for the WIMP masses higher than 8 GeV/c2. The difference between two fitting methods is less than 10%. The upper limit of 5

.

4

×

10−41_cm2 _{is obtained under different astrophysical}

as-sumptions of vesc

=

544 km

/

s [24]. The best fit parameters in

a mass range between 6 and 1000 GeV/c2 is a cross section of 3

.

2

×

10−42cm2 for a WIMP mass of 140 GeV/c2. This yields a statisticalsignificanceof2.7

σ

,however,inthiscase, theexpected unmodulatedeventrateexceedsthetotalobservedeventratebya factorof2,thereforetheseparametersweredeemedunphysical.

For the model independent analysis, the expected event rate wasestimatedas:

Rex_i,j

=

tj+



12tj tj−12tj



Ci

+

Aicos 2π

(

t

−

t0

)

T

dt

,

(4)

wherethefreeparametersCiandAi weretheunmodulatedevent

rate andthe modulation amplitude, respectively. t0 and T were

thephaseandperiodofthemodulation,andtj and



tj was the

time-bin’scenterandwidth,respectively. Inthefittingprocedure, the1.1–7.6keVee energyrangewas usedandthemodulation

pe-riod T was fixed to one year and the phase t0 to 152.5 days

(

∼

2ndofJune)whentheEarth’svelocityrelativetothedark mat-terdistributionis expectedtobe maximal. Fig. 4 showsthebest fitamplitudesasafunctionofenergyfor‘pullterm’after

correct-ing the efficiency. The efficiency was evaluated from gamma ray

MonteCarlosimulationwithaflatenergyspectrumuniformly dis-tributedin the sensitive volume (Fig. 4 inset). Both methods are

in good agreement and find a slight negative amplitude below

4 keVee. The

±

1

σ

and

±

2

σ

bands in Fig. 4 represent expected

amplitude coverage derived fromsame dummysample above by

the ‘pull term’ method. This test gave a p-value of0.014 (2.5

σ

) forthe ‘pullterm’ methodandof0.068 (1.8

σ

)forthe covariance matrixmethod.To be ableto test anymodel ofdark matter, we evaluated theconstraints onthe positive andnegative amplitude separately in Fig. 4. The upperlimits on the amplitudes in each energybinwerecalculatedbyconsideringonlyregionsofpositive ornegative amplitude.Theywere calculatedby integrating Gaus-sian distributionsbased on themean andsigmaofdata (

=

G

(

a

)

)

Fig. 4. (Coloronline.) Modulationamplitudeasafunctionofenergyforthemodel independentanalysesusingthe ‘pullterm’method(solidcircle).Solidlines rep-resent90%positive(negative)upperlimitsontheamplitude.The±1σ and±2σ

bandsrepresenttheexpectedamplituderegion(seedetailinthetext). DAMA/LI-BRAresult(square)isalsoshown[11].

fromzero.Thepositiveornegativeupperlimitsaresatisfiedwith 0.9for

aup 0 G

(

a

)

da

/

_∞ 0 G

(

a

)

da or

0 aupG

(

a

)

da

/

0 −∞G

(

a

)

da,wherea andaup aretheamplitudeandits90%C.L.upperlimit,respectively.

The‘pullterm’methodobtainedpositive(negative)upperlimitof 2.1(

−

2.1)

×

10−2 _{events/day/kg/keV}

ee between 1.1 and 1.6 keVee

andthelimitsbecomestricterathigherenergy.The energy reso-lution(

σ

/

E)at1.0(5.0) keVeeisestimatedtobe36% (19%)

com-paringgammaraycalibrationsanditsMonteCarlosimulation.Asa guideline,wemakedirectcomparisonswithotherexperimentsnot by consideringa specific darkmattermodelbutamplitude count rate.Themodulationamplitudeof

∼

2

×

10−2events/day/kg/keVee

between 2.0 and 3.5 keVee was obtained by DAMA/LIBRA [11]

and we estimate a 90% C.L.upper limit for XENON100as 3.7

×

10−3 events/day/kg/keVee (2.0–5.8 keVee) basedon

[17]

asitwas

not claimedasa signal.XMASS obtainedpositiveupper limitsof (1.7–3.7)

×

10−3 events/day/kg/keVee in same energy region and

gives the more stringent constraint. This fact is importantwhen wetestthedarkmattermodel.

5. Conclusions

Inconclusion, XMASSwithits largeexposure andhigh photo-electron yield (lowenergy threshold) conductedan annual mod-ulation search. For the WIMP analysis, the exclusion upperlimit of 4

.

3

×

10−41cm2 at8 GeV/c2 was obtainedandthe result

ex-cludes the DAMA/LIBRA allowed region for WIMP masses higher

than that. Inthe caseofthe modelindependent case,the analy-sis wascarried outfromtheenergythresholdof1.1 keVee which

islowerthanDAMA/LIBRAandXENON100.Thepositive(negative) upperlimitamplitudeof2.1 (

−

2.1)

×

10−2_{events/day/kg/keV}

ee

be-tween1.1and1.6 keVeeand(1.7–3.7)

×

10−3 counts/day/kg/keVee

between2 and6 keVee were obtained. As this analysisdoesnot

consideronlynuclear recoils,asimpleelectronorgammaray in-terpretationoftheDAMA/LIBRAsignalcanalsoobeythislimit.

Acknowledgements

WegratefullyacknowledgethecooperationofKamiokaMining andSmeltingCompany.Thisworkwas supportedbytheJapanese Ministry of Education, Culture, Sports, Science and Technology, Grant-in-Aid for Scientific Research, JSPS KAKENHI Grant Num-ber, 19GS0204, 26104004, andpartially by theNational Research

Foundation of Korea Grant funded by the Korean Government

References

[1]K.A.Olive,etal.,ParticleDataGroup,Chin.Phys.C38(2014)090001.

[2]D.S.Akerib,etal.,LUXcollaboration,Phys.Rev.Lett.112(2014)091303.

[3]E.Aprile,etal.,XENON100collaboration,Phys.Rev.Lett.109(2012)181301.

[4]R.Agnese,etal.,SuperCDMScollaboration,Phys.Rev.Lett.112(2014)241302.

[5]K.Abe,etal.,XMASScollaboration,Phys.Lett.B719(2013)78.

[6]K.Abe,etal.,XMASScollaboration,Phys.Rev.Lett.113(2014)121301.

[7]H.Uchida,etal.,XMASScollaboration,Prog.Theor.Exp.Phys.(2014)063C01.

[8]C.Savage,etal.,J.Cosmol.Astropart.Phys.04(2009)010.

[9]C.E.Aalseth,etal.,CoGeNTcollaboration,Phys.Rev.D88(2013)012002.

[10]R.Agnese,etal.,CDMScollaboration,Phys.Rev.Lett.111(2013)251301.

[11]R.Bernabei,etal.,Eur.Phys.J.C73(2013)2648.

[12]A.K.Drukier,K.Freese,D.N.Spergel,Phys.Rev.D33(1986)3495.

[13]J.D.Lewin,P.F.Smith,Astropart.Phys.6(1996)87.

[14]J.Kopp,etal.,Phys.Rev.D80(2009)083502.

[15]B.Feldstein,P.W.Graham,S.Rajendran,Phys.Rev.D82(2010)075019.

[16]R.Foot,Phys.Rev.D90(2014)121302(R); R.Foot,arXiv:1508.07402,2015.

[17]E.Aprile,etal.,XENON100collaboration,Phys.Rev.Lett.115(2015)091302; E.Aprile,etal.,XENON100collaboration,Science349 (6250)(2015)851.

[18]K. Abe, et al., XMASScollaboration, Nucl. Instrum. MethodsA 716(2013) 78.

[19]N.Y.Kim,etal.,Nucl.Instrum.MethodsA784(2015)499.

[20]T.Doke,R.Sawada,H.Tawara,in:TheInternationalWorkshoponTechnique andApplicationofXenonDetectors(Xenon01),WorldScientific,2001,p. 17.

[21]M.Szydagis,etal.,J.Instrum.8(2013)C10003.

[22]E.Aprile,etal.,Phys.Rev.Lett.107(2011)131302.

[23]J.Angle,etal.,XENON10collaboration,Phys.Rev.Lett.110(2013)249901.