A new limit on 0nbb-decay of Mo from the CUPID-Mo … · 2020-06-30 · Benjamin Schmidt, Lawrence...

Benjamin Schmidt (CUORE, CUPID/CUPID-Mo)

Lawrence Berkeley National Laboratory - 06/20/2020

A new limit on 0nbb-decay of 100Mo from the CUPID-Mo demonstrator for CUPID

CUPID-Mo Neutrino 2020 results, LBNL seminar 06/29/2020 Benjamin Schmidt, Lawrence Berkeley National Laboratory

Single beta decay and the neutrino

2

Wolfgang Pauli, “Letter to the radioactive ladies and gentlemen”, (1930)

Phys. Lett. B 590 (2004) 35-38

2 - body expected e-Energy (18.6 keV)


Single beta decay and the neutrino

3

Phys. Lett. B 590 (2004) 35-38

Wolfgang Pauli, “Letter to the radioactive ladies and gentlemen”, (1930)


Neutrinoless double beta decay Light Majorana neutrino exchange

4

Effective Majorana mass:

hm��i2 =��⌃i=1,2,3U

2e,imi

��

(T 0⌫��1/2 )�1 = G0⌫

��M0⌫��2 hm��i2

m2e

Energy (keV)0 500 1000 1500 2000 2500 3000 3500 4000

Frac

tion

of c

ount

s / k

eV

6−10

5−10

4−10

3−10

2−10

2nbb

0nbb (hypothetical)

100Mo

Reconstructed Energy (keV)


Neutrinoless double beta decay Light Majorana neutrino exchange

5

100Mo

76Ge

130Te

82Se

136Xe

Phase space

Nuclear Matrix Element (NME)

Effective Majorana mass hm��i2 =

��⌃i=1,2,3U2e,imi

��

Rep. Prog. Phys. 80 (2017) 046301

(T 0⌫��1/2 )�1 = G0⌫

��M0⌫��2 hm��i2

m2e


(meV)lightestm1−10 1 10 210 310

(meV

)ββ

m

1−10

1

10

210

310

Normal hierarchy

Inverted hierarchy

(meV)lightestm1−10 1 10 210 310

(meV

)ββ

m

1−10

1

10

210

310

Normal ordering

Inverted ordering

0 0.2 0.4 0.6 0.8 1

1−10

1

10

210

310

Other isotopes

MoSe

Ge

Xe

6

Inverted Ordering

Normal Ordering

mbb Vector addition/Cancellation in complex space

Phys. Rev. D64 :053010,2001

Im

Real

<mbb>2


The experimental challenge

Probe a process with a half-life larger > 1025 yr - 1026 yr

7

x x 105 yr

Next generation: Need to find single events in a ton of isotope x year(s) of exposure! 3 x 10-14 Bq/g We go to extreme length to limit ubiquitous radioactivity

15 Bq / banana


Experimentally considered 0nbb isotopes

8

11 / 35 experimentally considered candidate isotopes

Isotopic Abundance [atomic %]0 5 10 15 20 25 30 35

Q-v

alue

[keV

]

1000

1500

2000

2500

3000

3500

4000

4500Ca48

Ge76

Se82

Zr96

Mo100

Cd116

Sn124

Te128

Te130Xe136

Nd150

208Tl

214BiIsotope choice considerations:

high Q-value (3034 keV) -> large phase space, typically low natural radioactivity backgrounds Backgrounds -> improve signal/background through good energy resolution -> dedicated Background suppression/particle ID

CUPID-Mo


Scintillating cryogenic Li2MoO4 calorimeters

9

Cop

per:

Ther

mal

Bat

h

NTD-Ge thermistor

Teflon: weak thermal linkC(T) a T3

Thermal bath @ 20 mK

R(T ) = R0ep

T0/T

NTD-Ge thermistoras sensor

Source 100Mo = Detector Li2MoO4 high efficiency

Thermometers)

Light)

Energy)release)

Scin4lla4ng)bolometer)

Light)Detector)

Figure 3: Operating principle of a scintillating bolometer. The release of energy inside ascintillating crystal follows two channels: light production and thermal excitation.

A scintillating bolometer functions by operating a scintillating crystal as a281

cryogenic bolometer (as described above) and coupling it to a light detector, as282

shown in Fig. 3. As it is for other large mass bolometers, the device works only283

at extremely low temperatures (⇠10 mK).284

When a particle traverses the scintillating crystal and interacts with the285

lattice, a large fraction of the energy is transferred into the crystal as heat,286

raising the internal energy, thus inducing the already mentioned temperature287

rise. A small fraction of the deposited energy produces scintillation light that288

propagates as photons outside the crystal. These are then detected by a separate289

light detector facing the crystal. The light detectors used so far for scintillating290

bolometers are bolometers themselves and consist of germanium wafers, kept291

at the same temperature as the main bolometer. Scintillation photons deposit292

heat into the wafer and induce a temperature rise, which is then measured by293

a second thermistor.294

The signals registered by the two thermistors are conventionally named heat295

(the one generated in the main bolometer) and light (the one induced in the296

light detector). Although they have the same nature (temperature rises), they297

originate by di↵erent processes.298

An interesting feature of scintillating bolometers is that the ratio between the299

light and heat signals depends on the particle mass and charge. Indeed, while300

the thermal response of a bolometer has only a slight dependence on the particle301

10

Bolometer

Ge light detector

NTD-Ge thermistor

Cop

per:

Ther

mal

Bat

h


Scintillating cryogenic Li2MoO4 calorimeters

10

Thermal bath @ 20 mK

Thermometers)

Light)

Energy)release)


Light)Detector)























10

Bolometer

Time (s)0 0.5 1 1.5 2 2.5 3

Vol

tage

(mV

)

120−

100−

80−

60−

40−

20−

Time (s)0 0.5 1 1.5 2 2.5 3

Vol

tage

(mV

)

25−

24−

23−

22−

21−

20−

C(T) a T3

LMO example pulse @ 2615 keV LD example pulse @ 2615 keV (LMO)


CUPID-Mo at Laboratoire Souterrain de Modane France (2018 - 2020)

• 4800 m.w.e. rock overburden • shared EDELWEISS cryogenic

infrastructure operated at @ 20 - 22 mK

• 20 Li2100MoO4 detectors of ~210 g, ~97% enriched (2.26 kg 100Mo)

• Ge light detectors • Ge-NTD based sensor readout

• All Li2100MoO4, 19 light detectors operational

• physics data taking March 2019 - June 2020

11


CUPID-Mo design

• As of April 2020: Accumulated > 2.3 kg*yr of physics data (Blinded)

12


Laboratoire Souterrain de Modane


13


Laboratoire Souterrain de Modane


14


The EDELWEISS/CUPID-Mo cryogenic infrastructure


15

Active and passive shielding designed for the EDELWEISS-II dark matter search (Final results in 2010)

• 100 m2 plastic scintillator muon-veto system

• 50 cm PE shielding • 20 cm lead shield

innermost 2 cm is roman lead • Radon free air circulation in

between lead and Cu cryostat • Inversed geometry wet dilution

refrigerator with GM cryocoolers for 100K screen and He liquefier

• 10 days between LHe refill • In-house front end electronics

(Grenoble)


The CUPID-Mo design


16

Crystal growth and 100Mo enrichment

NIIC, Novosibirsk, Russia• purification of enriched Mo (from the

NEMO-3 experiment) to MoO3 • low radioactivity Li2CO3 • double crystallization (low thermal

gradient Czochralski technique) • surface polish with radio-pure SiO2

oil based slurry • Storage in dry N2 atmosphere

(Li2MoO4 is slightly hygroscopic)

4.158 kg Li2MoO4

2.264 kg 100Mo


CUPID-Mo design

17

Modular tower design:

• Compatible with existing EDELWEISS cryostat design

• Detector mounting in CSNSM & LAL clean-rooms (Orsay)

• Decoupling of LMO and light detectors from vibrations

• NOSV-Cu for radio-purity


CUPID-Mo tower suspension

Suspended tower design

Particularly important for the LD operation in (dry + wet) cryostat with vibrations from thermal machines

18


CUPID-Mo calibration

19

• LMO detectors have relatively low mass ~210 g and low density 3.07 g/cm3

• Significant amount of time dedicated to calibration (2 days / LHe refill) 20-25% of data taking

Energy LD (keV)5 10 15 20 25 30

Cou

nts /

(0.1

keV

)

0

100

200

300

400

500

600

700

800

αCu K

αMo K

βMo K

Energy (keV)0 500 1000 1500 2000 2500 3000

Cou

nts /

keV

1

10

210

310

410

510CUPID-Mo, Neutrino 2020U/Th calibration, Preliminary

Tl208Bi214

Bi214Bi214

Bi214Pb214

Ac228

DE

SE

Bi214

• Low energy calibration sources are potentially dangerous for the EDELWEISS dark matter search

• Use the Mo x-ray escape peak from high intensity irradiation of the crystals (60Co)

19 LD sum spectrum 0.3 keV s at 17.4 keV

19/20 LMO sum spectrum


CUPID-Mo performance

• Li2100MoO4 scintillates at 600 nm • Typical measured light yield of

~0.6/0.7/0.9 (keV/MeV) for b,g

• Difference in light yield expected from tower design • a scintillation light yield of 20% - compared to b,g • > 99.9% alpha separation extrapolated for all detectors

• Good uniformity/performance suitable for larger arrays!CUPID-Mo commissioning results EPJ-C 80:44 (2020)

20

LMO detector2 4 6 8 10 12 14 16 18 20

Rela

tive

Ligh

t Yie

ld (k

eV/M

eV)

0

0.2

0.4

0.6

0.8

1

1.2

Top LDBottom LDSingle LD

Example Heat/Light separation - CUPID-Mo

commissioninga

b,g


CUPID-Mo: The Neutrino 2020 data

21

Days (tot) Days (sel) Exposure kg x yr RetainedPhysics 240 200 (224) 2.17 (2.4) 94%

(93%)Calibration 73 59 (65) 0.6 (0.7) 88% (89%)Special 21 — — —

Downtime 46 — — —

• March 2019 - April 2020 (380 days) • 7 Long datasets, 1-2 month scale • 3 Short datasets (single calibration periods)

Not used in the Neutrino analysis - extra work needed on energy-scale uncertainty

• Rejection of periods of temperature instabilities


CUPID-Mo: Data production and cuts

22

Base cuts

Single trigger BaselineSlope

Light Yield

Sum of LD Consistency of LD

Pulse Shape analysis

Principal component analysis (PCA)

Multiplicity

Pulser rejection M1 - single crystal

Muon veto anti-coincidence

Thermometers)

Light)

Energy)release)


Light)Detector)























10

Bolometer

Trigger efficiency CUORE based data production chain at NERSC


CUPID-Mo: Trigger efficiency

23

1−10 1 10 210 310 energy [keV]

0

0.2

0.4

0.6

0.8

1

effi

cien

cy

Injected pulses trigger efficiency ch 19

1−10 1 10 210 energy [keV]

0

0.2

0.4

0.6

0.8

1

effi

cien

cy

Injected pulses trigger efficiency ch 40

LMO

LD

Optimum trigger Derivative trigger

Optimum trigger Derivative trigger

• Use of Optimum Filter to obtain lower thresholds • Surplus coincidence information • Choose conservative 10 sigma trigger threshold

• Evaluation of trigger efficiency: Inject avg. pulse template into noise • —> Typical LMO threshold ~ 11 keV (90% efficiency),

fully efficient above analysis threshold (45 keV) —> Typical LD threshold ~ 0.5 keV (90% efficiency)

thresholdEntries 1449

Mean 0.8392

Std Dev 1.089

1 2 3 4 5 6 7 8 Energy [keV]

0

20

40

60

80

100

120

140

160

180

Cha

nnel

s


Mean 0.8392

Std Dev 1.089

trigger∈thres. at 90% Mean 0.85Median 0.52

Thresholds distribution at 90% of trigger efficiency


Mean 15.96

Std Dev 18.59

0 20 40 60 80 100 120 140 160 180 200 Energy [keV]

0

5

10

15

20

25

30

35

40

45

Cha

nnel

s


Mean 15.96

Std Dev 18.59

trigger∈thres. at 90% analysis thresholdMean 15.9Median 10.7

Thresholds distribution at 90% of trigger efficiency

2019 data 2019 dataCUPID-Mo Preliminary CUPID-Mo Preliminary


CUPID-Mo: Data quality cuts

• No energy dependence of these cuts • No line of sight between LMO detectors • Appreciable 210Pb, 210Po presence • 210Po serves as a clean sample of events (no pulser, heat-only

other event contamination) for the efficiency estimate

24

Base cuts

Single trigger BaselineSlope

Multiplicity

Pulser rejection M1 - single crystal

(Muon veto anti-coincidence)

Light detector

LMO

LMO

LMO

LMO

Energy (keV)5395 5400 5405 5410 5415 5420

Cou

nts/

(0.6

keV

)

50

100

150

200

250 Base Cuts

Base+M1 Cuts e ~ 98 % CUPID-Mo Preliminary


CUPID-Mo - Pulse Shape Analysis

25



Energy (keV)0 500 1000 1500 2000 2500

Rec

onst

ruct

ion

Erro

r (AU

)

0

50

100

150

200

250

300

350

PCA Reconstruction Errors (Background)PCA Reconstruction Errors (Background) Normalized PCA Reconstruction Errors (Background)

Energy (keV)500 1000 1500 2000 2500 3000

Nor

mal

ized

Rec

onst

ruct

ion

Erro

r

10−

5−

0

5

10

15

20

25

30

35

40

Normalized PCA Reconstruction Errors (Background)

• Perform a Principal Component Analysis (PCA) • Train on 1 MeV - 2 MeV 2nbb events in physics data • 1st component - contains main amplitude information - similar to average pulse

• Define the (Normalized) Reconstruction Error E with respect to 1st (1st plus 2nd) PCA component as pulse shape analysis variable

E =X

i

(xi,rec. � xi)2

<latexit sha1_base64="zTInzU20aeAXmc+kBujCYO4BGoI=">AAACDnicbVDLSgMxFM3UV62vUZdugqVQQYeZUlAXQlEElxXsA9pxyKSZNjTzIMmIZZgvcOOvuHGhiFvX7vwbM+0stHoh5OSce7k5x40YFdI0v7TCwuLS8kpxtbS2vrG5pW/vtEUYc0xaOGQh77pIEEYD0pJUMtKNOEG+y0jHHV9keueOcEHD4EZOImL7aBhQj2IkFeXolUt4Bvsi9p2EprB6r65DTrCRwiOYPdKD26SWOnrZNMxpwb/AykEZ5NV09M/+IMSxTwKJGRKiZ5mRtBPEJcWMpKV+LEiE8BgNSU/BAPlE2MnUTgorihlAL+TqBBJO2Z8TCfKFmPiu6vSRHIl5LSP/03qx9E7shAZRLEmAZ4u8mEEZwiwbOKDKumQTBRDmVP0V4hHiCEuVYEmFYM1b/gvaNcOqG6fX9XLjPI+jCPbAPqgCCxyDBrgCTdACGDyAJ/ACXrVH7Vl7095nrQUtn9kFv0r7+AbfzJrG</latexit>

Linear energy dependence Predictive extrapolation to 3 MeV

1st PCA component

CUPID-Mo Preliminary


CUPID-Mo - Pulse Shape Analysis

26



Energy (keV)0 500 1000 1500 2000 2500 3000 3500 4000

Effic

ienc

y0

0.2

0.4

0.6

0.8

1

• Optimize PCA cut based on the goal of mitigating pile-up on calibration data

• Maximize signal sensitivity using Cowan’s metric

• S = NEMO-3 exclusion limit • B = Bg from 2615 keV tails (scaled)

• Under further study to improve pile-up rejection

• Efficiency evaluation (physics data) • Evaluate PCA cut with LY cuts applied • Fit of 1st order polynomial in high stat. 2nbb region • Evaluation of fit uncertainty at Qbb ; 1% Dataset level

Qbb (3034 keV)


e ~ 97 % r2(S +B) · ln(1 + S

B)� S

<latexit sha1_base64="+wHs4t1+UkFr2A6PQ9AEIGirmtU=">AAACE3icbVA9SwNBEN2LXzF+RS1tFoMQDYY7CahdiI1lJCYRciHsbfaSxb29c3dOCMf9Bxv/io2FIrY2dv4bNx+FRh8MPN6bYWaeFwmuwba/rMzC4tLySnY1t7a+sbmV395p6TBWlDVpKEJ14xHNBJesCRwEu4kUI4EnWNu7vRj77XumNA/lNYwi1g3IQHKfUwJG6uWPXH2nIDkpNkq1Q+zSfghYyKJTcn1FaNJIk1p6iI9xI8318gW7bE+A/xJnRgpohnov/+n2QxoHTAIVROuOY0fQTYgCTgVLc26sWUToLRmwjqGSBEx3k8lPKT4wSh/7oTIlAU/UnxMJCbQeBZ7pDAgM9bw3Fv/zOjH4Z92EyygGJul0kR8LDCEeB4T7XDEKYmQIoYqbWzEdEhMGmBjHITjzL/8lrZOyUymfX1UK1dosjizaQ/uoiBx0iqroEtVRE1H0gJ7QC3q1Hq1n6816n7ZmrNnMLvoF6+MbI0ub0A==</latexit>


Energy LMO (keV)0 1000 2000 3000 4000 5000 6000

Ener

gy L

D (k

eV)

0

500

1000

1500

2000

2500

3000

1

10

210

ββQ

CUPID-Mo - The light yield cuts

27

Example Heat/Light separation (PCA cut applied)

200 days of physics data

a

b,g

Light Yield

Weighted sum of LD Consistency of LD


Energy LMO (keV)0 1000 2000 3000 4000 5000 6000

Ener

gy L

D (k

eV)

0

500

1000

1500

2000

2500

3000

1

10

210

ββQ


28

Example Heat/Light separation (PCA cut applied)

200 days of physics data

a

b,g

Light Yield


LY cut defined with 3 sigma acceptance on calibration data



• Efficiency evaluation (physics data) • Evaluate the LD cut efficiency from the 2nbb spectrum

after applying the PSA cut

• Systematics from energy extrapolation • Excess broadening with respect to expected

LD width, sqrt(sB2 + nScint * 2.07 eV) observed • Consider two models

Excess broadening a E — excess broadening a sqrt(E) • systematic quantification with Toy MC incl. fit

uncertainties

29

Light Yield


+0.9�0.2%

<latexit sha1_base64="ObU4EY3BZuyeOh3QMgzaNnxlAns=">AAACEHicbVDLSsNAFJ3UV62vqEs3g6UoqCEpBe2u6MZlBfuAJpTJdNIOnUzCzEQooZ/gxl9x40IRty7d+TdO2iy09cCFwzn3cu89fsyoVLb9bRRWVtfWN4qbpa3tnd09c/+gLaNEYNLCEYtE10eSMMpJS1HFSDcWBIU+Ix1/fJP5nQciJI34vZrExAvRkNOAYqS01DdPXJn4UiE8Ts9sqw5dF17YVnUKoXvuhkiNRJi6lWmpb5Zty54BLhMnJ2WQo9k3v9xBhJOQcIUZkrLn2LHyUiQUxYxMS24iSazXoiHpacpRSKSXzh6awopWBjCIhC6u4Ez9PZGiUMpJ6OvO7Ea56GXif14vUcGVl1IeJ4pwPF8UJAyqCGbpwAEVBCs20QRhQfWtEI+QQFjpDLMQnMWXl0m7ajk1q35XKzeu8ziK4Agcg1PggEvQALegCVoAg0fwDF7Bm/FkvBjvxse8tWDkM4fgD4zPH9rGmpg=</latexit>

LD excess width LD expected width - photon stat.


CUPID-Mo - Analysis cut efficiencies

30



Cut \ DATASET 1 2 3 4 5 6 7Base 99.0(3) 100 99.8(2) 99.9(1) 99.9(1) 99.7(3) 99.7(2)Slope 98.2(1) 98.6(2) 98.4(1) 98.8(1) 98.9(1) 98.9(1) 99.0(1)

M1 97.3(6) 98.0(6) 97.2(6) 97.6(5) 98.3(5) 98.8(6) 97.8(5)Light Det 96.8(2) 96.7(2) 95.5(2) 97.5(1) 97.5(2) 97.2(2) 97.2(2)

PCA 96.6(9) 96.8(11) 97.2(7) 96.2(7) 96.6(10) 97.6(11) 98.9(10)Total 88.4(11) 90.5(12) 88.6(9) 90.4(9) 91.4(11) 92.3(13) 92.9(13)

✏ = (90.5± 0.4 (stat.) +0.9�0.2 (syst.))%

<latexit sha1_base64="eMwdOI6d0ksbhDaOPS0kwjDeUSI=">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</latexit>

Total efficiency (exposure weighted avg.)CUPID-Mo Preliminary


Energy (keV)0 500 1000 1500 2000 2500 3000 3500 4000

Cou

nts /

keV

1−10

1

10

210

310

410

CUPID-Mo, Neutrino 20202.17 kg x yr, Preliminary

Tl208ββν2

Physics dataNormalized U/Th data

CUPID-Mo the blinded data

31

• 19/20 detectors with good performance • Analysis efficiency 90.5 %

• 200 days of physics data, ~7 keV FWHM @ 2615 keV (calibration)

Blinded 0nbb dataset


CUPID-Mo - ROI definition for counting analysis

32

2540 2560 2580 2600 2620 2640 2660 2680 2700Energy (keV)

1

10

210

Even

ts /

keV

Tl - 7.1 keV FWHM effective208


ROI

+ +

Energy [keV]2000 2200 2400 2600 2800 3000 3200

cts

1

10

210

310

410

510

610

Smeared MC - DS 8237 Channel 3

Detector resolution

BG Index

0nbb containmentBremsstrahlung escape

ROI definition




CUPID-Mo - Energy scale

33

Energy (keV)0 500 1000 1500 2000 2500

Cou

nts /

(2 k

eV)

1

10

210

310

410CUPID-Mo, Neutrino 20202.17 kg x yr, PreliminaryPb212

Co60Co60

Tl208

K40

Mo99

Bi214

0 500 1000 1500 2000 2500 3000 3500 4000

Fit E

nerg

y (k

eV)

500

1000

1500

2000

2500

3000

3500

4000

Mo100��

Q

/NDF = 0.2529 / 22�

0.4016 keV±) = -0.1802 ��

(QbiasE

Literature value - Energy (keV)0 500 1000 1500 2000 2500 3000 3500 4000-1

-0.50

0.51

1.5

Res

idua

ls (k

eV) 0


• Energy scale is set with pol2 in calibration data • Check consistency in time in calibration data

• Estimate possible energy bias based on physics data, EB = (-0.2 ± 0.4) keV

208Tl40K

208Tl214Pb

214Bi


CUPID-Mo - ROI (Ch,DS) based resolution

34

2540 2560 2580 2600 2620 2640 2660 2680 2700Energy (keV)

1

10

210

Even

ts /

keV

Tl - 7.1 keV FWHM effective208 • Simultaneous unbinned extended maximum likelihood (UEML) fit to extract the Ch,DS - based resolutions • Fit model:

smeared step function (multi-compton) Gauss (photopeak) Linear (multi-photon + 2nbb)

• —> Extract the gaussian width on a Ch,DS basis

1 DatasetCUPID-Mo Preliminary


Energy (keV)0 500 1000 1500 2000 2500 3000 3500

FWH

M (k

eV)

0

2

4

6

8

10

Mo100��

Q

/NDF = 2.2237 / 42�0.3860 keV±FWHM at 3034 keV: 7.6097 keV

1.15028e-03±= 1.47590e+00 0

p

1.95813e-03±= 1.83680e-02 1

p

• Obtain a global scaling factor Calibration @2615 keV <-> Physics @3034 keV • Test several hypothesis:

• linear, sqrt, pol2 fit -> linear is ruled out by calibration data -> take remaining more conservative estimate (pol2)

CUPID-Mo ROI resolution scaling

35

Energy (keV)0 500 1000 1500 2000 2500 3000 3500

FWH

M (k

eV)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Mo100��

Q

/NDF = 0.9601 / 32�0.1024±Ratio Extrap at 3034 keV: 1.0593

0.7468 keV±FWHM at 3034 keV is: 7.6757 keV


208Tl40K

208Tl

214Pb214Bi


CUPID-Mo ROI definition Bg index from Geant4 MC model

36

• Detailed Geant4 MC model

500 1000 1500 2000 2500 3000Energy [keV]

2−10

1−10

1

10

210

Cou

nts/

keV Experimental

JAGS reconstruction

Spectrum M1

Preliminary

500 1000 1500 2000 2500 3000Energy [keV]

0.8

0.9

1

1.1

1.2

1.3

Dat

a/M

odel

ratio Data/Model

σ1 σ2 σ3

M1


Room temperatureelectronics

Cryostat shields

Internal lead

Internal polyethylene

Pump lines

• Two fits: RooFit and JAGS (MCMC) • M1 - Gamma analysis: BI expectation for 0nbb ROI

4 ± 2 counts /keV/kg/yr


Energy (keV)2500 2600 2700 2800 2900 3000 3100 3200 3300 3400

Cou

nts

/ ( 5

keV

)

3−10

2−10

1−10

1

10

210

CUPID-Mo ROI definition Bg index from sideband data

37

• Perform unbinned extended maximum likelihood fit on Bg data excluding [3010, 3060] keV • Phenomenological Bg model:

Exponential - approximates both 2nbb spectral shape and U/Th calibration tail Flat component - conservative estimate of 2nbb pile-up and remaining muon-induced events

~ [4-5] 10-3 counts/keV/kg/yr strongly dependent on fit range




CUPID-Mo - ROI

38

• Optimize signal ROI for Poisson counting analysis in Signal, Background likelihood space

• Maximize mean limiting setting sensitivity for a poisson counting analysis with • an expected final CUPID-Mo exposure of

2.8 kg x yr • a background index of

0.005 counts /keV/kg/yr

—> Large central ROI ~18 keV average width


CUPID-Mo - Limit setting

• Two analyses: • Bayesian counting analysis in

central ROI + sidebands (3 bin fit) • Central bin/ROI:

75% Signal & Bg • Sideband:

1% Signal & Bg • Bg: Exponential + flat • Use Gaussian priors on

exponential from fit in [2650,2980] keV region

• Poisson counting analysis as cross-check

39

0 0.005 0.01 0.015 0.02 0.025Input BI [counts/keV/kg/yr]

0.4

0.6

0.8

1

1.2

1.4

1.6

2410×

90%

c.i.

lim

it [y

r]�0 1/2

T

0

20

40

60

80

100

120

140

Prob

abili

ty d

ensi

ty -

# to

ys

0

20

40

60

80

100

120

140

Prob

abili

ty d

ensi

ty -

# to

ys

0 counts in ROI - Poisson limit

CUPID-Mo, Neutrino 2020yr, Preliminary×2.17 kgToy-MC study

0 cts ROI

1 cts ROI

2 cts ROI

Toy study


New world leading limit on 0nbb of 100Mo

perfectly consistent with Poisson analysis

CUPID-Mo - New 0nbb result

40

Energy (keV)2500 2600 2700 2800 2900 3000 3100 3200

Cou

nts

/ keV

1

10

ROI evtsSideband evtsAnalysis regionMean ROITl208


T 0⌫1/2 > 1.4 · 1024 yr, 90% c.i. (stat.+ syst.)

<latexit sha1_base64="HdycxliPgXWt0vD8nIwUEb/m7Uo=">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</latexit>

BAT signal posteriorCUPID-Mo Preliminary


CUPID-Mo systematics

41

Energy (keV)2500 2550 2600 2650 2700 2750

Cou

nts /

keV

1

10

210

ROI evts

Sideband evtsMean ROI

Tl208CUPID-Mo, Neutrino 20202.17 kg x yr, Preliminary

CUPID-Mo ROI - Unblinded

• Isotope concentration (96.6 ± 0.2) %

• Containment (~75% on average in peak) • Geant4 modeling & density/volume uncertainty: 1.1 % • Energy scale and resolution uncertainties

included with MC sampling in containment of chosen ROI

• Analysis efficiency • All cuts stat. & PCA extrapolation - Gaussian prior • LD resolution model - Uniform, asymmetric

• Potential non gaussianity - Containment reduction • Use 2615 keV peak in calibration data

- (2.5 ± 2.5)%

✏ = (90.5± 0.4 (stat.) +0.9�0.2 (syst.))%


Energy [keV]2000 2200 2400 2600 2800 3000 3200

cts

1

10

210

310

410

510

610

Smeared MC - DS 8237 Channel 3

MC simulation of 0nbbCUPID-Mo Preliminary


CUPID-Mo - Effective Majorana Neutrino mass

42

With only 1 year of data and ~2 kg of 100Mo CUPID-Mo is able to set a limit of mbb < (0.31-0.54) eV 90% c.i.

considering gA = 1.27 and the following NME calculations:

F. Šimkovic, V. Rodin, A. Faessler, P. Vogel, Phys. Rev. C 87, 045501 (2013). https://doi.org/10.1103/PhysRevC.87.045501N.L. Vaquero, T.R. Rodríguez, J.L. Egido, Phys. Rev. Lett. 111, 142501 (2013). https://doi.org/10.1103/PhysRevLett.111.142501J. Barea, J. Kotila, F. Iachello, Phys. Rev. C 91, 034304 (2015). https://doi.org/10.1103/PhysRevC.91.034304J. Hyvärinen, J. Suhonen, Phys. Rev. C 91, 024613 (2015). https://doi.org/10.1103/PhysRevC.91.024613L.S. Song, J.M. Yao, P. Ring, J. Meng, Phys. Rev. C 95, 024305 (2017). https://doi.org/10.1103/PhysRevC.95.024305P.K. Rath et al., Phys.Rev.C88, 064322 (2013). https://doi.org/10.1103/PhysRevC.88.064322F. Šimkovic, A. Smetana, and P. Vogel, Phys. Rev. C 98, 064325 (2018). https://doi.org/10.1103/PhysRevC.98.064325P.K. Rath, Ramesh Chandra, K. Chaturvedi and P. K. Raina, Front. Phys. 64, 1 (2019). https://doi.org/10.3389/fphy.2019.00064

(meV)lightestm1−10 1 10 210 310

(meV

)ββ

m

1−10

1

10

210

310

Mo), 2.17 kg x yr100CUPID-Mo (Neutrino 2020 - Preliminary

Normal hierarchy

Inverted hierarchy

0 0.2 0.4 0.6 0.8 11−10

1

10

210

310

Other isotopes

Mo Se

TeGe

Xe


https://doi.org/10.1103/PhysRevC.87.045501

https://doi.org/10.1103/PhysRevLett.111.142501






https://doi.org/10.3389/fphy.2019.00064


New world leading limit on 0nbb of 100Mo, Neutrino 2020 poster #419

• 2nbb result, CUPID-Mo technology (arXiv: 1912.07272), Neutrino 2020 poster #525

• Preliminary status/results from Bg model analysis Neutrino 2020 poster #418

• Bg index of [4 - 5] x 10-3 counts/keV/kg/yr in 0nbb ROI with non-optimized setup for dark matter search, (lower than expected)

• Further CUPID-Mo updates at NEUTRINO 2020: • 0nbb/2nbb decay to excited states & low energy

prospects, 56Co high energy calibration Neutrino 2020 posters #374, #382, #448

CUPID-Mo at Neutrino 2020

43

Energy (keV)2500 2600 2700 2800 2900 3000 3100 3200

Cou

nts

/ keV

1

10

ROI evtsSideband evtsAnalysis regionMean ROITl208


T 0⌫1/2 > 1.4 · 1024 yr, 90% c.i. (stat.+ syst.)

<latexit sha1_base64="HdycxliPgXWt0vD8nIwUEb/m7Uo=">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</latexit>

T 2⌫1/2 =

⇥7.12+0.18

�0.14 (stat.)± 0.10 (syst.)⇤· 1018 yr

<latexit sha1_base64="PbBYCXopw5NyZx2gRkgcsFkr/Rs=">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</latexit>


CUPID-Mo - What’s next ?

44

Physics data taking finished as of June 20th 2.8 kg x yr total (before cuts) - 20-30 % more data

Currently ongoing high energy calibration campaign Fe wire irradiated at 88-inch (LBNL) in collaboration by Andew Voyles and Rick Normann -> ~660 Bq activity in 56Co Currently deployed at Laboratoire Souterrain de Modane

500 1000 1500 2000 2500 3000Energy [keV]

2−10

1−10

1

10

210

Cou

nts/

keV Experimental

JAGS reconstruction

Spectrum M1

Preliminary

500 1000 1500 2000 2500 3000Energy [keV]

0.8

0.9

1

1.1

1.2

1.3

Dat

a/M

odel

ratio Data/Model

σ1 σ2 σ3

M1

Next up: Focus on Bg model and 2nbb precision analyses, 0nbb/2nbb to excited states and low mass dark matter search



CUPID-Mo results for CUPID

45

• Excellent crystal radiopurity (Neutrino 2020 poster #404)• [0.3 - 1] µBq/kg for U/Th

100 µBq/kg 210Pb

• Li2MoO4 bolometric performance in non-optimal environment • Efficient alpha rejection over 1 year of data taking

• LD performance hit due to AC-biasing/demodulation sampling limitation (0.5 kHz)

• High analysis efficiency • ~7 keV calibration resolution @ 2615 keV

~8 keV physics resolution @ 3034 keV • ~20 mK instead of ~10-15 mK • Sub-optimal/(No) heater based gain-stabilization

Cou

nts /

10

keV

1

10

10 2

10 3

4000 6000 8000 10000

210Po (surface / bulk)

190Pt 238U

234U,226Ra 222Rn

218Po

Bi-Po events

Energy (keV)

✏ = (90.5± 0.4 (stat.) +0.9�0.2 (syst.))%


2540 2560 2580 2600 2620 2640 2660 2680 2700Energy (keV)

1

10

210

Even

ts /

keV

Tl - 7.1 keV FWHM effective208




The CUPID-Mo collaboration

46

Zoom Unblinding

UCB and LBNL led analysis team: Giovanni Benato, Toby Dixon, Roger Huang, Laura Marini, Benjamin Schmidt, Vivek Singh and Bradford Welliver


Backup

Neutrino 2020 Poster links: CUPID-Mo 0nbb analysis https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-419.pdf CUPID-Mo performance https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-404.pdf CUPID-Mo 56Co calibration campaign https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-374.pdf CUPID-Mo background model https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-418.pdf CUPID-Mo low energy analysis prospects https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-448.pdf CUPID-Mo sensitivity for 0nbb/2nbb decay to excited states https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-382.pdf 2nbb analysis with CUPID-Mo technology https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-525.pdf

47

https://nusoft.fnal.gov/nova/nu2020postersession/pdf/posterPDF-419.pdf







A new limit on 0nbb-decay of Mo from the CUPID-Mo … · 2020-06-30 · Benjamin Schmidt, Lawrence...

Documents

Transcript of A new limit on 0nbb-decay of Mo from the CUPID-Mo … · 2020-06-30 · Benjamin Schmidt, Lawrence...