skirmishes on the ldm frontier territorial disputes...

Skirmishes on the LDM frontier

territorial disputes over the low-recoil region

in direct detection

Josef Pradler

Johns Hopkins University

“Light Dark Matter”University of Michigan

April 15, 20131

1. New feeble signals at the direct DM detection threshold “Claiming territory over the low recoil region”

2. A critical look at DAMA’s Dark Matter claim

- Dark Photons => Haipeng An’s talk this morning

- New neutrino signals in direct detection

Outline

2

1210.5501 (PLB)

1210.7548

with Balraj Singh and Itay Yavin

with Itay Yavin

with Maxim Pospelov 1203.0545 (PRD)

with Haipeng An and Maxim Pospelov 1304.3461 (PLB)

1302.3884

http://arxiv.org/abs/1203.0545

http://arxiv.org/abs/1203.0545

A vision of a true neutrino observatory

• superconducting grains in filler material in magnetic field

• at low temperatures specific heat ~

=> single scatter of neutrino can make grain conducting

=> magnetic field collapses, induces electric signal in detector

T 3

• coherent enhancement for MeV-scale neutrinos from

=> spallation sources, supernovae, reactors, sun, earth

• cross section grows quadratically with neutrino energy

• helicity conservation forbids back-scattering

coherent neutrino-nucleus scattering

N2

d⇤

d cos �=

1

8⇥G2

FE2�

⇥Z(4 sin

2 �W � 1) +N⇤2

(1 + cos �)

• coherent enhancement for MeV-scale neutrinos from

=> spallation sources, supernovae, reactors, sun, earth

• cross section grows quadratically with neutrino energy

• helicity conservation forbids back-scattering

coherent neutrino-nucleus scattering

N2

d⇤

d cos �=

1

8⇥G2

FE2�

⇥Z(4 sin

2 �W � 1) +N⇤2

(1 + cos �)

(this process has not yet been observed)

=> direct DM detection

Witten 1985~ keV

=> direct DM detection

• nuclear recoil can be picked up in various channels:

heat ionizationscintillation

WIMPs vs. solar neutrinos• flux

• cross section

• recoil

�DM =�0v

mDM⇠ 105 cm�2s�1

✓100GeV

mDM

◆�pp = 6⇥ 1010 cm�2s�1

�8B = 6⇥ 106 cm�2s�1

� ' 10�44 cm2 ⇥N2

✓E⌫

1MeV

◆2

� = 10�44 cm2 ⇥ �44A2

✓µN

µn

◆2

Emax

R =(2µNv)2

2mN⇠

8>>>>><

>>>>>:

20 keV�A20

�

(mN ⌧ mDM )

4 keV�

mDM20GeV

�2

�100

A

�

(mDM ⌧ mN )

Emax

R =(2E�)2

2mN

⇠ 0.1 keV

✓20

A

◆✓E�

1MeV

◆2

“baryonic” neutrinosM. Pospelov PRD 2011

⌫b

• introduce new left-handed neutrino speciestogether with gauged

• couples to quarks, but not to leptons

• breaking of gives new gauge field mass

U(1)b

⌫b

U(1)b Vµ

⌫bsterile under SM-gauge group

active under U(1)b

LB = ⇥b�µ(i⇤µ � glqbVµ)⇥b �

1

3gb

X

q

q̄�µqVµ � 1

4Vµ�V

µ� +1

2m2

V VµVµ + Lm.

⌫b

8

• for effective Lagrangian reads

• measure interaction strength in units of :

jµNCB = ⇥b�µ⇥b

Q2 ⌧ m2V

GF

Le� = �GBjµNCB

X

N=n,p

N�µN, GB = qbgbglm2

V

N =|GB |GF

⇥ 100�✓3GeV

mV

◆2 ✓ glgb10�2

◆


⌫b

9

• crucial insight:

• For solar flux, deuteron breakup in SNO does not constrain scenario

��bN (elastic)

��bN (inelastic)� A2

E4�R

4N

� O(108)

this ratio makes direct detection experiments competitive with large scale neutrino experiments


⌫b

10

direct detection of solar

dR(t)

dER= NT

L0

L(t)

�2 X

i

�i

Z

Emin⌫

dE�dfidE�

d�

dERPb(t, E�)

overall flux modulation

average overneutrino spectrum i

L(t) = L0

⇢1� � cos

2⇥(t� t0)

1 yr

�� L0 = 1AU

t0 ' 3 Jan (perihelion)

� = 0.0167 (eccentricity)

⌫b

like SM-neutrinos with G2F (N/2)2 ! G2

BA2

appearance probability

11

dR(t)

dER= NT

L0

L(t)

�2 X

i

�i

Z

Emin⌫

dE�dfidE�

d�

dERPb(t, E�)

more modulation here

Losc

L0

' 0.5⇥✓10�10 eV

�m2

◆✓E⌫

10MeV

◆oscillation-length on the order sun-earth distance

=> flip phase for high energy part of the neutrino spectrum? explain DAMA?

direct detection of ⌫b

Appearanceprobability

• considering small values instandard solar story unfolds

�m2b

Pb(earth) ' sin2(2�b) sin2

�m2

bL(t)

4E

�

N 2e� ⇥ N 2

2� sin2 2✓b

=> for fast oscillations PbG2B � N 2

e�G2F

(from a tribimaximal ansatz assuming mixing to ) ⌫2

[see also arXiv:1103.3261]

13

17F

15O

13Npphep

8BGermanium

ER [keV]

reco

ilsp

ectrum

[even

ts/keV

/kg-y]

1010.10.01

108

107

106

105

104

103

102

101

1

17F

15O

13Npphep

8B

E! [MeV]

solar!flux[cm

!2s!

1M

eV!1]

1010.1

1012

1010

108

106

104

102

1

AGe = 70� 76

8B Ne� = 100

direct detection of ⌫b

14

!bDAMA 2010

"2/d.o.f. = 9.3/8

!m2 = 2.52 ! 10!10 eV2

Ne! = 102

.

Ev [keVee]

modulation

amplitu

de(cpd/kg/keV

ee)

2015105

0.03

0.025

0.02

0.015

0.01

0.005

0

-0.005

-0.01

fit only first 10 bins

DAMAmodulation amplitude

15

CRESST-IIsignal

• 8 CaWO4 crystals, measure scintillation light and phonons from nuclear recoil

• in a nutshell: 67 events in acceptance region - half of which are attributed to backgrounds

WCaO

Ne! = 100

!m2 = 2.5! 10!10 eV2

dotted: hep neutrinossolid: 8B neutrinos

CaWO4 target

Er [keV]

dR/dE

r(cpd/kg/keV

)

2018161412108642

102

101

1

10!1

10!2

10!3

10!4

10!5

10!6

Angloher et al EPJC 2012

16

• we follow CRESST in their modeling of backgrounds

CRESST-IIfits

=> e/gamma events known => other bkg. essentially flat

CRESST-II 730 kg!days

Ev (keV)

counts/keV

40353025201510

10

8

6

4

2

0

observedn!

Pbe!/"

#b


Ev (keV)

counts/keV

40353025201510

10

8

6

4

2

0

• we follow CRESST in their modeling of backgrounds

CRESST-IIfits

=> e/gamma events known => other bkg. essentially flat

• use Poisson log-Likelihoodto fit ⌫b

�2P = 2

X

i

yi � ni + ni ln

✓ni

yi

◆�

• best fit yields

�2P /d.o.f. = 27.8/28 (recoil spectrum only)


Ev (keV)

counts/keV

40353025201510

10

8

6

4

2

0

observedn!

Pbe!/"

#b


Ev (keV)

counts/keV

40353025201510

10

8

6

4

2

0

direct detection experiments as neutrino observatories

99% C.L.

!m2 [eV2]

Ne!

10!910!10

100

10

1

Xenon10 low th.Xenon100CDMS-II low th.CRESST-II excl.CoGeNT hullDAMACoGeNTCRESST-II

99% C.L.

!m2 [eV2]

Ne!

10!910!10

100

10

1

b

18

reg CDMS-II Si as of today

19

2

technologies [6].The Cryogenic Dark Matter Search (CDMS) collabora-

tion identifies nuclear recoils (including those that wouldoccur in WIMP interactions) using semiconductor detec-tors operated at 40 mK. These detectors use simulta-neous measurements of ionization and non-equilibriumphonons to identify such events among the far more nu-merous background of electron recoils. During 2003-2008the collaboration operated CDMS II, an array of Ge andSi detectors located at the Soudan Underground Labo-ratory [7]. Previous results from the CDMS II instal-lation [8–11] have set world-leading upper limits on theWIMP-nucleon scattering cross section and constrainedsome non-WIMP dark matter candidates [12].

The low atomic mass of Si generally makes it a lesssensitive target for spin-independent WIMP interactionsrelative to the larger coherent enhancement of the scat-tering cross section for heavy nuclei. On the other hand,the lower atomic mass of Si is advantageous in searchesfor WIMPs of relatively low mass due to more favorablescattering kinematics. A WIMP of mass . 40 GeV/c2

will transfer more recoil energy to a Si nucleus than aGe nucleus on average. For a recoil energy threshold of⇠ 10 keV, WIMPs of su�ciently low mass (. 10 GeV/c2)will generate more detectable recoils in a Si detector sincethe more numerous Ge recoils would fall below the detec-tor threshold. New particles at such masses are generallydisfavored in fits of models to precision electroweak data(e.g. [13]), but viable models in this regime do exist (e.g.[14, 15]). Renewed interest in this mass range has beenmotivated by results from the DAMA/LIBRA [16], Co-GeNT [17], and CRESST [18] experiments, which can beinterpreted as evidence of low-mass WIMP scattering.

In its final configuration, the CDMS II array consistedof 30 Z-sensitive ionization and phonon (ZIP) detectors:19 Ge (⇠239 g each) and 11 Si (⇠106 g each), for a to-tal of ⇠4.6 kg of Ge and ⇠1.2 kg of Si. Each CDMS IIdetector is a semiconductor disk, 7.6 cm in diameter and1 cm thick, instrumented to detect the phonons and ion-ization generated by particle interaction within the crys-tal [7]. We discriminate nuclear recoils from backgroundelectron recoils using the ratio of ionization to phononrecoil energy (ionization “yield”). Electron recoils thatoccur within ⇠10 µm of a detector surface can exhibitreduced ionization collection. These events are identi-fied by phonon pulse-shape discrimination. Our overallmisidentification rate of electron recoils is less than 1 in106.

We consider data from the Si detectors using the finalfour run periods of the full CDMS II detector installa-tion, acquired between July 2007 and September 2008.The Ge results from this data set have been described inprevious publications [11]. Of the 11 Si detectors, threewere excluded from the WIMP-search analysis: two dueto wiring failures that led to incomplete collection of theionization signal and one due to unstable response onone of its four phonon channels. Periods of poor perfor-mance, as identified by a series of Kolmogorov-Smirnov

0 20 40 60 80 1000

20

40

60

80

100

120

140

:,03�([

SRVXUH��NJï

GD\V�

5HFRLO�(QHUJ\��NH9�

0 20 40 60 80 1000

0.25

0.5

0.75

1

:,03�HIILFLHQF\

5DZ�([SRVXUH�*RRG�1XFOHDU�5HFRLO�)LGXFLDO�9ROXPH�3KRQRQ�7LPLng

FIG. 1. Exposure (left y axis) and e�ciency (right y axis)as functions of recoil energy after application of each WIMP-selection criterion shown. Each curve from top to bottomshows the cumulative e↵ect of successive cuts on the data,such that the bold solid curve shows the overall e�ciency ofthis analysis. The abrupt drops in acceptance at low recoilenergies reflect the elevated energy thresholds chosen for somedetectors.

tests, were also excluded from analysis. After all suchexclusions, the data collected by the 8 Si detectors con-sidered in this analysis represent a total exposure of 140.2kg-days prior to the application of the WIMP candidateselection criteria.The responses of these detectors to electron and nu-

clear recoils were calibrated using events from extensiveexposures to 133Ba and 252Cf sources in situ at Soudan.Electron recoils from the former were used to empiricallycharacterize and correct for the dependence of phononpulse shape on event position and energy. The 356 keVgamma ray from the 133Ba source has a ⇠ 4.2 cm atten-uation length in Si, and thus the Si detectors generallydo not show a clear line at 356 keV. Their energy scaleswere calibrated using 356 keV events with total energiesshared between the Si detector and a neighboring detec-tor.WIMP-candidate events were identified by a series of

selection criteria. As with the Ge data analysis, eventsin and near the WIMP-candidate region were automat-ically removed from the data set during the analysis,and all WIMP-selection criteria were defined blindly us-ing calibration and remaining WIMP-search data. Thus,WIMP candidates had no impact on the definition ofthe selection criteria. A WIMP candidate was requiredto have phonon and ionization signals inconsistent withnoise alone in exactly one ZIP detector and to exhibit nocoincident energy in the scintillating veto shield or in anyof the other 29 ZIP detectors. Events in coincidence withthe NuMI beam [19] were also vetoed. We demanded thatany candidate event occur within the detector’s fiducialvolume, defined by requiring signal consistent with noisein the outer ionization electrode. Candidate events werealso required to have ionization yield and phonon pulsetiming consistent with a nuclear recoil. The recoil energy

Agnese et al. 2013

Ne! = 49

!m2 = 4! 10!9 eV2

CDMS-II Si 140 kg!days

Er [keV]

dR/dE

r(even

ts/keV

)

14131211109876

1

10!1

10!2

10!3

phase off by one month! :(

(2–6 keV)

days since Jan 1, 1995

5000450040003500

0.040.02

0-0.02-0.04

(2–5 keV)

residuals

(cpd/kg/keV

)

0.040.02

0-0.02-0.04

DAMA/LIBRA 0.87 ton!yr

(2–4 keV)

0.040.02

0-0.02-0.04

DAMAtime series

20

Outlookdirect detection

Kr

threshold

S2software

8.4keV

Ne! = 100

!m2 = 2.5 ! 10!10 eV2

hep

8B

XENON100, 100.9 live days

S2 (PE)

even

ts

700600500400300200

10000

1000

100

10

1

0.1

prediction for Xenon100 low-threshold analysis

prediction for COUPPbubble chamber (CF3I)

COUPP, 60 kg!1 year

Ne! = 60

!m2 = 3! 10!10 eV2

DM

hep

8B

Ethr (keV)

even

ts

30252015105

10000

1000

100

10

1

0.1

21

Outlookdirect detection

• even lighter targets (He) are even better! Neutrinos give more recoil than WIMPs ( )

• directional detection

�

dR

d cos �Nhas strongest “AFB”

in direction vavg(�)

in direction of sun

(WIMPs)

(neutrinos)

Emax

R (�)

Emax

R (⇥)=

E2

�

v2µ2

⇥

' 25⇥✓

E�

10MeV

◆2

✓4GeV

µ⇥

◆2

Outlooksolar neutrino experiments

• elastic scattering off scintillating mineral oil with ultra-pure setups

14C

hep

8BR14 = 10!23

C9H12

(quenched) energy (keV)

even

ts/day

/M

eV/to

n

20018016014012010080604020

103

102

101

1

10!1

10!2

10!3

10!4

10!5 mainly protonrecoils

Borexino has C14 contamination of

10�18

• supernovae production of => signal in direct detection possible?

• may affect dynamics of explosions

• sensitivity to truly tiny mass splittings over cosmological distances

• Neff = ?

⌫b

⇡

n

n

n

n

⌫b⌫b

freezes out quickly=> high initial T spectrum!

insulating scattering sphere

⌫b

24

Outlookastrophysical signatures

with Liang Dai (JHU)

Part II

apropos DAMA

25

observed modulation amplitude

the higher the background, the stronger the signal must be modulated

=> take a closer look at the DAMA backgrounds to see what is needed

26

smax

m � sobsm

✓1 +

B0

S0

◆⇡ 2%⇥

✓1 +

B0

S0

◆smax

m = Sm/S0

sobsm =Sm

R=

Sm

B + S0

' 2%

Sm ' 0.02 cpd/kg/keV

S = S0 + Sm cos!(t� t0)

DAMA signal interpretationin the presence of backgrounds

• “Single-hit” spectrum(all other detectors act as a veto)

Ê

Ê

Ê ÊÊÊÊÊ Ê

ÊÊÊÊÊ Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê Ê

Ê Ê Ê Ê ÊÊ Ê Ê

Ê Ê

2 4 6 8 100.0

0.5

1.0

1.5

2.0

2.5

3.0

Energy HkeVL

RateHcpdêkgê

keVL

R ⇠ 1 cpd/kg/keV

stat. error bar shown

noise

signalregion

increased backgrounds in the signal region?27

DAMA signal interpretationin the presence of backgrounds

!!

!+

4!

2+

0+

0+

EC

EC"

40K

40Ar

40Ca

89.25(17) %

Q = 1311.07(11) keV10.55(11) %

1460 keV

0.2(1) %

0.00100(12) %

Q = 1504.69(19) keV

!1.2504(30) · 109 a

"

(nuclear recoil )⌧ 1 keV

X-raysAuger e

⇠ 3 keV

Ar

needs MC to find rate at 3 keV => we employ the results from

Kudryavtsev et al 2010

� (1.4MeV)

28

Potassium background in DAMA at 3 keV

!!

!+

4!

2+

0+

0+

EC

EC"

40K

40Ar

40Ca

89.25(17) %

Q = 1311.07(11) keV10.55(11) %

1460 keV

0.2(1) %

0.00100(12) %

Q = 1504.69(19) keV

!1.2504(30) · 109 a

"

29


!!

!+

4!

2+

0+

0+

EC

EC"

40K

40Ar

40Ca

89.25(17) %

Q = 1311.07(11) keV10.55(11) %

1460 keV

0.2(1) %

0.00100(12) %

Q = 1504.69(19) keV

!1.2504(30) · 109 a

"

⌫̄e


X-raysAuger e

⇠ 3 keV

Ar29


!!

!+

4!

2+

0+

0+

EC

EC"

40K

40Ar

40Ca

89.25(17) %

Q = 1311.07(11) keV10.55(11) %

1460 keV

0.2(1) %

0.00100(12) %

Q = 1504.69(19) keV

!1.2504(30) · 109 a

"

⌫̄e


X-raysAuger e

⇠ 3 keV

Ar

angular momentum change by 4 units,“3rd forbidden unique weak decay”

=> the ONLY such EC realized in nature

29


40K

238U+232Th+129IDAMA reported

!! (ii)

EC" + "

EC", " escapes (iii)

EC, g.s. (i)

.

E (keV)

cpd/k

g/keV

100001000100101

101

1

10!1

10!2

10!3

10!4

10!5

shape = energy resolution

30

Simulated DAMA spectrum using reported contaminations

40K

238U+232Th+129IDAMA reported

!! (ii)

EC" + "

EC", " escapes (iii)

EC, g.s. (i)

.

E (keV)

cpd/k

g/keV

100001000100101

101

1

10!1

10!2

10!3

10!4

10!5

• strong indication of a flat background component

• and Compton background at low energies are flat!

=> work out implication for modulation fraction

Bflat ' 0.85 cpd/kg/keV

��

shape = energy resolution

30

Simulated DAMA spectrum using reported contaminations

7%

10%

30%

7%

10%

30%

required modulation fraction

Bflat = 0.85 cpd/kg/keV

excluded

claimed

DAMA

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

• challenges standard WIMP scenario with Maxwellian halo:

• for 13 ppb potassium contamination

required!

sm . 10%

sm & 20%

31

required modulation fractionif a flat background is present

7%

10%

30%

7%

10%

30%



excluded

claimed

DAMA

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

• critique 1: potassium is measured at

• critique 2: EC to g.s. is only 10% => our discussion is “captious”

• critique 3: upper limit on signal claimed => allows for 6-10% modulation!

natK = 13ppb

S0 0.25 cpd/kg/keV

results triggered responses by DAMAwhich only raised more questions

Bernabei et al.

1210.6199

1211.6346

7%

10%

30%

7%

10%

30%



excluded

claimed

DAMA

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1




natK = 13ppb

S0 0.25 cpd/kg/keV

Great. A new number in print.


Bernabei et al.

1210.6199

1211.6346

7%

10%

30%

7%

10%

30%



excluded

claimed

DAMA

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1




natK = 13ppb

S0 0.25 cpd/kg/keV


Confirms our findings.


Bernabei et al.

1210.6199

1211.6346

7%

10%

30%

7%

10%

30%



excluded

claimed

DAMA

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1

BREC (%)

natK

(ppb)

10.80.60.40.20

102

101

1




natK = 13ppb

S0 0.25 cpd/kg/keV


Confirms our findings.

=> let’s check. Requires “slide-forensic”.

Number not in print.


Bernabei et al.

1210.6199

1211.6346

DAMA/LIBRA 0.53 tonuyr

Example of a correct approach: S0

vs background

background

around 2-6 keV: • straight line extrapolation from higher energy

• <natK> (|13 ppb)

background

Residuals: upper limit of S0

In the energy region 2-4 keV:

S0

<~ 0.25 cpd/kg/keV

single-hit counting rate

multi-detectors set-up each one in anticoincidence with all the others

data under the energy threshold of the experiment: wait for the new higher Q.E. PMTs

Counting-rate/S0

(upper limit) ratio is very much higher in other experiments !

DM annual modulation signature offers a powerful tool for background rejection (see in Freese et al. )

Software energy threshold of the

experiment

talk by Nozzoli for DAMA, TAUP 2009

referenced by DAMA’s reply to our paper:

none of our questions/concerns have been addressed.

instead our assumption of a flat background was criticized as being ad hoc

=> their own model is not supported by data

34


Ê

Ê Ê

Ê

Ê

Ê

Ê

Ê

Ê

Ê

Ê

Ê

ÊÊ Ê

ÊÊ

Ê ÊÊ

ÊÊ

ÊÊ

ÊÊ

ÊÊ

ÊÊ

ÊÊ

Ê

ÊÊ

2 4 6 8 100.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

1.4

Energy HkeVL

RateHcpdêkgê

keVL

DAMA backgroundreported in Nozzoli @7D

Gaussian + Flat + Linear Fitwith 13 ppb

with 14 ppb

Flat background 0.85 cpdêkeVêkgused in arXiv:1210.5501

interpretation of the signal in terms of DM is seen to be very sensitive to assumptions on the background....

Bernabei et al.

1210.6199

1211.6346

This is how it’s done

35

copper encapsulation the upper limits on the activities shownin Table 4 have been obtained at the HPGe test bench at theLSC and are given at 95% C.L. Upper limits for 238U are worsethan those obtained for other isotopes because of the low inten-sity of its gamma emissions; in this case, equilibrium in the nat-ural chains has not been imposed.! For archaeological lead, upper limits on 210Pb, 232Th and 238U

activities quoted in [23] have been used.! For radon content in the air filling the inner volume of the

shielding, there is no real estimate. Radon content in the labora-tory air is being continuously monitored, and the inner volumeof the shielding is flushed with boil-off nitrogen, to guaranteeits radon-free quality. An arbitrary value for the radon contentin the inner volume air of about one hundredth of the externalair radon content has been assumed in our background model(0.6 Bq/m3), compatible with the absence of lines coming fromradon daughter isotopes in the measured background. Never-theless, this contribution should be considered as an upperlimit.

4. Geant4 simulation

A complete simulation (using the Geant4 package [24]) of theANAIS-0 module with shielding in different experimental configu-rations has been carried out. Geant4 is widely used for this purposeand its reliability has been established in a large number of validat-ing studies, changing experimental conditions and detection tech-niques. Some recent works using Geant4 with similar context andgoals can be found in the literature [25–28]. Accurate modeling ofthe detector and shielding components and physical processes par-ticipating in the interaction mechanisms for every backgroundsource is required in order to reproduce qualitatively and quantita-tively the background event rate of the experiment. VersionGeant4.9.4.p01 and its corresponding data libraries have been usedfor all the simulations presented here. Physical processes and mod-els commonly used in the context of underground experimentshave been implemented for interactions of alpha, beta and gammaparticles; for example, the low energy models based on the Liver-more data libraries for the electromagnetic interactions. Range cutvalues of 10 (100) lm (that are converted into the correspondingenergy cuts for every material) have been typically used for elec-trons (photons) in the simulation of bulk emissions; these cut val-ues are (recommended) production thresholds for secondaryparticles. The Geant4 Radioactive Decay Module has been usedfor simulating radioactive contaminations, after checking carefullythe energy conservation in the decay of all the considered isotopes;problems encountered in previous versions of Geant4 code seem tobe surpassed. In particular, a very relevant (for this work) modifi-cation, included in the used Geant4 version, is related with theshape of the 40K beta spectrum3, because the proper shape factorfor the third unique forbidden beta decay has been used [29] (seeFig. 3).

The simulated geometry (see Fig. 4) includes the NaI crystal,Teflon and reflector lining, quartz windows, light guides (optional),photomultipliers, copper encapsulation, Mylar window, HV dividercircuit and its mechanical support, and the shield made of botharchaeological and standard low activity lead.

Decays of radioactive impurities in the most relevant materialsof the experimental layout, including mainly 238U and 232Th chainsand 40K, have been simulated assuming (unless otherwise stated) a

Table 3Activities considered for the simulation of the different PMT models used in theANAIS-0 module. In the case of the Ham LB and Ham ULB PMTs, the values measuredfor each available unit are given separately.

Detector component Isotope Activity (mBq/PMT)

ET LB PMT 40K 420 ± 50232Th 24 ± 4238U 220 ± 12

Ham LB PMT 40K 678 ± 42 647 ± 56232Th 68 ± 3 75 ± 4238U 100 ± 3 109 ± 5

Ham ULB PMT 40K 12 ± 7 24 ± 9232Th 3.6 ± 1.2 1.8 ± 1.0238U 47 ± 28 59 ± 28226Ra 8.0 ± 1.2 6.1 ± 1.260Co 4.1 ± 0.7 5.1 ± 0.8

Table 4Activities considered in the simulation for the different components of the ANAIS-0module layout. Upper limits are given at 95% C.L.

Simulated component Isotope Activity

Copper encapsulation 40K < 11 mBq232Th < 4.1 mBq238U < 140 mBq226Ra < 2 mBq60Co < 0.94 mBq

Quartz optical window 40K < 12 mBq/kg232Th < 2.2 mBq/kg238U < 100 mBq/kg226Ra < 1.9 mBq/kg

Light guides 40K < 21 mBq/guide232Th < 4.1 mBq/guide238U < 120 mBq/guide226Ra < 4.7 mBq/guide

Optical coupling grease 40K < 200 mBq/kg232Th < 200 mBq/kg238U < 2000 mBq/kg226Ra < 30 mBq/kg

Archaeological lead 210Pb < 20 mBq/kg232Th < 0.3 mBq/kg238U < 0.2 mBq/kg

Inner volume air 222Rn < 0.6 Bq/m3

3 Problem number 1068 at Geant4 Problem Tracking System: http://bugzilla-geant4.kek.jp/

0 200 400 600 800 1000 1200 1400 1600 1800 20000.00.20.40.60.81.01.21.41.61.82.0

cpd/k

eV/kg

Energy (keV)

Fig. 3. Simulated spectra for bulk 40K contamination in the ANAIS-0 NaI crystal areshown both for an allowed beta spectrum shape (Geant4.9.3 version) in blue andtaking into account the proper shape factor for the third unique forbidden betadecay (Geant4.9.4 version) in red. It can be seen a much better accordance with themeasured background spectrum (in black) shape at 1 MeV for the latter: in bothcases 12.7 ± 0.5 mBq/kg of 40K have been considered, as deduced from the analysisof coincidences (see text) (For interpretation of the references to colour in thisfigure legend, the reader is referred to the web version of this article.).

S. Cebrián et al. / Astroparticle Physics 37 (2012) 60–69 63

Background model for a NaI (Tl) detector devoted to dark matter searches

S. Cebrián a,b, C. Cuesta a,b, J. Amaré a,b, S. Borjabad b, D. Fortuño a, E. García a,b, C. Ginestra a,b, H. Gómez a,b,M. Martínez a,b, M.A. Oliván a,b, Y. Ortigoza a,b, A. Ortiz de Solórzano a,b, C. Pobes a,b, J. Puimedón a,b,M.L. Sarsa a,b,!, J.A. Villar a,b

a Laboratorio de Física Nuclear y Astropartículas, Universidad de Zaragoza, C/ Pedro Cerbuna 12, 50009 Zaragoza, Spainb Laboratorio Subterráneo de Canfranc, Paseo de los Ayerbe s.n., 22880 Canfranc Estación, Huesca, Spain

a r t i c l e i n f o

Article history:Received 29 May 2012Received in revised form 9 July 2012Accepted 21 July 2012Available online 7 August 2012

Keywords:Sodium iodide scintillatorsDark matterANAISLow backgroundBackground simulation

a b s t r a c t

NaI (Tl) is a well known high light yield scintillator. Very large crystals can be grown to be used in a widerange of applications. In particular, such large crystals are very good-performing detectors in the searchfor dark matter, where they have been used for a long time and reported first evidence of the presence ofan annual modulation in the detection rate, compatible with that expected for a dark matter signal. In theframe of the ANAIS (Annual modulation with NaI Scintillators) dark matter search project, a large andlong effort has been carried out in order to characterize the background of sodium iodide crystals. In thispaper we present in detail our background model for a 9.6 kg NaI (Tl) detector taking data at the CanfrancUnderground Laboratory (LSC): most of the contaminations contributing to the background have beenprecisely identified and quantified by different complementary techniques such as HPGe spectrometry,discrimination of alpha particles vs. beta/gamma background by Pulse Shape Analysis (PSA) and coinci-dence techniques; then, Monte Carlo (MC) simulations using Geant4 package have been carried out forthe different contributions. Only a few assumptions are required in order to explain most of the measuredbackground at high energy, supporting the goodness of the proposed model for the present ANAIS pro-totype whose background is dominated by 40K bulk contamination. At low energy, some non-explainedbackground components are still present and additional work is required to improve background under-standing, but some plausible background sources contributing in this range have been studied in thiswork. Prospects of achievable backgrounds, at low and high energy, for the ANAIS-upgraded detectors,relying on the proposed background model conveniently scaled, are also presented.

! 2012 Elsevier B.V. All rights reserved.

1. Introduction

Sodium iodide scintillators have been widely used for radiationdetection [1,2]. In particular, they have been applied since thenineties in the direct search for dark matter, profiting from the fea-sibility of growing large mass crystals, the high light output, thesensitivity to spin-dependent interacting WIMP candidates, thelow and high mass isotope combination, and the particle discrim-ination capability (although limited at low energy). On the otherhand, they suffer from poor energy resolution and low RelativeScintillation Efficiency factor for nuclear vs. electron recoil events.Several experimental efforts with NaI (Tl) detectors in the searchfor dark matter can be found in the literature [3–12]. Among theseexperiments, DAMA/LIBRA results have produced a large impact inthe field by analyzing the annual modulation in the rates: a mod-ulation compatible with that expected for galactic halo WIMPs was

reported after seven cycles of measurement at the Laboratori Nazi-onali del Gran Sasso, Italy [5] and has been confirmed with muchmore statistical significance by LIBRA data, accumulating six morecycles at this moment [11]. Complete understanding of the DAMA/LIBRA background at low energy has not yet been achieved andsome open questions remain [13,14]. Results obtained with othertarget materials and detection techniques have been ruling outfor years the most plausible compatibility scenarios (some of themost recent and significant negative results can be found in[15,16]), but recently CoGeNT and CRESST experiments have re-ported excess of events attributable to WIMPs [17,18], and even,in the former experiment, the presence of an annual modulationin the rate has been observed [19].

Confirming the DAMA/LIBRA modulation observation in a mod-el independent way is the main goal of the ANAIS experiment (An-nual modulation with NaI Scintillators), projected to be carried outat the Canfranc Underground Laboratory (LSC), Spain, with up to250 kg of NaI (Tl) [20]. In the context of ANAIS, a long effort tounderstand and reduce the radioactive background in sodium io-dide detectors has been done. In this paper we report on the esti-

0927-6505/$ - see front matter ! 2012 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.astropartphys.2012.07.009

! Corresponding author at. Laboratorio de Física Nuclear y Astropartículas,Universidad de Zaragoza, C/ Pedro Cerbuna 12, 50009 Zaragoza, Spain.

E-mail address: [email protected] (M.L. Sarsa).

Astroparticle Physics 37 (2012) 60–69

Contents lists available at SciVerse ScienceDirect

Astroparticle Physics

journal homepage: www.elsevier .com/ locate/ast ropart

ANAIS collaboration

DAMA shouldshow us the K40shoulder

A count rate in DAMA much greater than 0.04 cpd/kg/keV will severely undermine a DM interpretation of the signal.

Conclusions

• DAMA data speaks against a “vanilla” Dark Matter interpretation

=> a minimum of 20% modulation in any putative signal may at least be required

=> after a decade of modulation maybe it is time to take a more global look at the data set

• new neutrinos with enhanced baryonic currents can be tested in

direct detection experiments

=> “DM-like” signals from new neutrino physics can explain DM anomalies CoGeNT and CRESST-II, unchallenged by other searches

=> upcoming experimental results will conclusively probe the most interesting parameter space

skirmishes on the ldm frontier territorial disputes...

Documents