Accurate Measurement of Reactor Neutrinos Close to Surface with STEREO

Accurate measurement of reactor neutrinos close to surface with STEREO

Photo:ILL Vladimir Savu IRFU,CEA, Paris-Saclay University on behalf of the STEREO collaboration

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 1 / 23

06.12.2019 – AAP 2019 Motivation Sterile neutrino search and reactorν ¯e measurements Motivation of STEREO

I Oscillation test

L/E ∼ 10 m/3 MeV → ∼ 1eV sterile neutrino 2 2 Two new parameters: sin (2θnew ) and∆ mnew Physical Review D 83, 073006 (2011), G. Mention et al.

∆ χ 2 4 1 4 9 5.5 Daya Bay Huber model w/ 68% C.L. I absolute ﬂux 5.0

/ fission] 4.5 normalization studies 2 cm 4.0 − 43 C.L

spectral shape studies [10 I 3.5 68%

239 95%

σ −43 σ238 =(10.1±1.0)×10 99.7% σ =(6.04±0.60)×10−43 3.0 241 5.2 5.6 6.0 6.4 6.8 7.2 −43 2 σ235 [10 cm / fission]

Phys. Rev. Lett. 118, 251801 (2017) Physics Letters B 773 (2017) STEREO will test precisely these 3 questions with a pure 235U spectrum

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 1 / 23 The STEREO experiment Experimental site ILL research facility, Grenoble, France

Research reactor core 58 MWth → 1019 ν¯ s−1 e Water channel 15 m.w.e overburden 235 X Highly U enriched X Compact core (40cm ∅) X Short baseline measurement: 9.4m < Lcore < 11.2m

93 tons moved on air cushions AutumnAutumn 2016 2016

I Surface-level experiment I γ and neutron background from neighboring experiments

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 2 / 23 The STEREO experiment Data taking Data taking

I Phase-I: 66 days reactor ON – 22 days reactor OFF I Phase-II: 119 days reactor ON – 211 days reactor OFF I Data taking eﬃciency: 98.5%, 14% dead-time+veto-time after oﬀ-line cuts 50 days reactor ON cycles with long periods of reactor OFF allowing for a good characterization of the background

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 3 / 23 The STEREO experiment The Stereo detector The STEREO detector

Acrylic buﬀers PMT Mineral oil I designed for a relative measurement X six identical Target cells ﬁlled with Gd doped LS Gamma-catcher ------External crown scintillator liquid (no Gd)

Target 6 cells scintillator liquid (Gd loaded) Counts Counts

L Without oscillation

Center cell with oscillation

Border cell with oscillation

1 2 3 4 5 6 7 Evisible [MeV] E

I reﬂectivity of VM2000 foils is above 98% in the 400-950 nm range in air

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 4 / 23 The STEREO experiment Detector response Light cross-talks

I The evolution of the light cross-talks between cells is monitored by looking at muon events I Phase I light cross-talks are evolving due to the liquid scintillator inﬁltrating into the acrylic walls. X Repaired during summer 2017 X Energy reconstruction formalism to account for the evolving light cross-talks I Phase II light cross-talks are reduced and stabilized

20 20 Cell2 → Cell3 Cell2 → Cell3 [%] [%] ij Cell6 → GC Back ij Cell6 → GC Back L Cell3 → GC D19 L → 15 15 Cell3 GC D19 → Cell5 Cell6 Cell5 → Cell6

10 10

5 5

0 0 10/31/16 12/31/16 03/02/17 10/31/17 12/31/17 03/02/18 Date [MM/DD/YY] Date [MM/DD/YY]

Phase I light cross-talks Phase II light cross-talks

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 5 / 23 The STEREO experiment Detector response Calibration

I Circulation of radioactive sources (68Ge,137 Cs,54 Mn,65 Zn,60 Co,42 K,24 Na, AmBe) along 3 diﬀerent calibration systems: I internal calibration tubes - 5 z positions I LED system, used to study the single I along the perimeter of the detector photoelectron and the PMT-DAQ I below the central long axis of the detector linearity in the detector Target

Non-linearity [%] 0

−1

−2

−3 0 200 400 600 800 1000 1200 1400 Charge [PE]

Top view of STEREO detector No sign of non-linearity up to 10 MeV for all PMTs

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 6 / 23 The STEREO experiment Detector response Simulation ﬁne tuning - raw charges

I Very good agreement between data and 20 simulation at the raw charge level: most 24Na Data points within ±1% 24Na MC I The energy reconstruction will correct to ﬁrst order the mean oﬀset between the TG cells

Preliminary

24Na Data 24Na MC Dilatation Factor

Preliminary Cell 1 Cell 2 Cell 4 Cell 5 Cell 6

10 45 80 30 60 10 45 80 30 60 10 45 80 Height [cm]

Measured (black) and simulated (red) collected Summary of all the dilatation factors ﬁtted from charges in cell 1 (top) and in cell 2 (bottom) when the 25 measured and simulated positions of the the 24Na source is located at the middle of cell 1. 54Mn source.

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 7 / 23 The STEREO experiment Detector response Simulation ﬁne tuning - Quenching

dE I Quenching for high dx leads to a non-linear Preliminary response at low energy deposits

dL dE ≈ dX dX dE 1 + KB dx Norm. Calib. Coefficient I The eﬀective Birks coeﬃcient KB is adjusted in the MC in order to match the experimental quenching curve I Agreement between experimental and simulated curves reaches sub-% accuracy

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 8 / 23 The STEREO experiment Detector response Energy reconstruction

−→ Q are the collected charges −1 M matrix constructed from regular monitoring: mij = Ci · Lji −−→ −1−→ Erec = M Q 54 I Ci calibration coeﬃcients( Mn radioactive source)

I Lji cross-talks between cells (cosmics)

Preliminary Cell 5

Data/MC agreement of the reconstructed Vertex distribtution for the 54Mn source deployed inside the energy distribution for a 54Mn calibration internal calibration tube

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 9 / 23 The STEREO experiment Detector response Energy reconstruction for all the calibration sources

I Residuals from all the radioactive sources I Energy resolution (δE/E) is well reproduced are contained in a ±1% band in the MC for all the radioactive sources I Energy resolution saturates at 5% due to volume eﬀects

Cell 1 Cell 2

Cell 3 Cell 4

Preliminary Preliminary Cell 5 Cell 6

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 10 / 23 The STEREO experiment Detector response Detector response - Y-axis

I Same quality and pattern of residuals as for I Neutrons from AmBe source are able to the X,Z axes tested with the calibration traverse the GC cells and reach the Target tubes cells I No further correction needs to be applied I We look at the n-H capture peak Target & Gamma-Catcher AmBe n-H (mid-height)

Preliminary

Cell ID

Data/MC Erec of neutron captures by hydrogen, for the average of 3 horizontal deployment Y-axis scan with AmBe source positions of an AmBe source at mid-height along the Y-axis of the detector

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 11 / 23 The STEREO experiment Detector response 12B β-decay spectrum

I Complementary data to the calibration sources I extended energy range Preliminary I continuous spectrum I extended vertex distribution

I β-decay spectrum of 12B → selection of cosmic muon captures on 12C:

− 12 12 µ + C → B + νµ

I Good agreement between the measured and simulated 12B spectra

Use the 12B spectrum and the calibration sources to perform a global ﬁt of the energy scale

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 12 / 23 The STEREO experiment Detector response 12B β-decay spectrum - global ﬁt

Preliminary

I Global ﬁt of sources + 12B spectrum I Supplementary constraints on the energy Cell 4 scale I Polynomial ﬁt function: Global Fit Data MC Erec = Pol(Erec )

I Distortions of the measured neutrino spectrum induced by the ﬁtted deviation of Data MC Erec to Erec are contained in an envelope obtained by varying the calibration coeﬃcients with ±1%

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 13 / 23 The STEREO experiment Detector response Time stability

Fit of the n-H capture peak in Cell 2 Relative deviation to the mean

n-H

Preliminary

n-H capture of neutrons from cosmic rays I Preliminary

I Distribution of vertices similar to νe

I Stable detector response at 0.3% level in each cell

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 14 / 23 The STEREO experiment Background and shielding Signal and background

Inverse beta decay (IBD) interaction: n thermalisation νe

diffusion p γ + e+ νe + p e + n energy deposit γ Gd(n,γ)γ)) → annihilation ~ 8MeV γ cascade e- e+ γ

Delayed event → neutron capture Prompt event → e+ P Gd: i Eγ,i ∼ 8MeV E + ≈ E − 0.8MeV vis,e νe H: Eγ ∼ 2.2MeV

Accidental background Correlated background from cosmic rays I natural radioactivity I stopping muons I reactor induced neutrons and gammas I spallation neutrons

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 15 / 23 The STEREO experiment Background and shielding Background and shielding

I Passive shielding for γ, neutrons and magnetic ﬁelds I Water channel and active veto for cosmic muons

Accidental background Correlated background from cosmic rays I natural radioactivity I stopping muons I reactor induced neutrons and gammas I spallation neutrons

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 16 / 23 The STEREO experiment ν¯e selection and eﬃciency studies

ν¯e signal selection and eﬃciency studies

n thermalisation νe τ ~ 20μs γ diffusion p + I Mean cut efficiency: e γ Gd(n,γ) energy deposit ~ 8MeV γ cascade annihilation 61.4±0.9% e- e+ γ I Dominated by neutron efficiency (delayed Clean after Clean before correlated pair t signal)

1.6 < Eprompt < 7.1 MeVE delayed > 4.5 MeV

AmBe source at z=45cm 3000 Data 2500 Simulation 2000

events 1500

1000

500

10 20 30 40 50 60 70 correlation time [µs] 1.5 1.0 0.5 data/MC 10 20 30 40 50 60 70 correlation time [µs]

Good agreement with Monte-Carlo Good agreement with Monte-Carlo for the neutron in correlation time detection eﬃciency at the % level

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 17 / 23 The STEREO experiment ν¯e selection and eﬃciency studies Improved Gd Gamma cascade simulation

I Delayed signal: gamma cascade from (n,γ)Gd → correct modelization of primary importance for small detectors (low gamma containment) I Improved simulation using FIFRELIN deexcitation of Gd isotopes using experimental data completed by nuclear models

AmBe source at mid-height

→ Eur.Phys.J.A (2019)55:183 – 107 cascades available on zenodo:2653786

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 18 / 23 e-recoils

The STEREO experiment Correlated background andν ¯e extraction

Correlated background andν ¯e extraction

e-recoil signal

p-recoil signal Pulse Shape Discrimination (PSD) for prompt signal

τe < τp I electron recoils (γ,ν ¯e ...) Qtail/Qtot (e) < Qtail/Qtot (p) I proton recoils (fast neutrons...)

Tail charge - Qtail Total charge - Qtot Cell 2 -- E = [2.125 - 2.625] MeV -2 ln λ| /ndf = 50.1/63 Correlated background 3.5 p-recoils min e-recoils prob = 0.88 (cosmics): 3 On : 119 days 2.5 a × Off : 211 days I rate sensitive to Rate (/day) × environment 2 On - a Off neutrinos On Acc. I stable in shape 1.5 Off Acc. 1 µ σ Fit model : G(A, ν, ν) Fitted → build model from 0.5 Nν = 10.37 ± 0.52 /day reactor-oﬀ data 0

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 PSD ν¯e signal extraction from reactor-on data, χ2 / ndf 8.12 / 9 with self-consistent2 background rescaling for each cell, energy20 binMean 0.004 ± 0.097 unit) Sigma 0.87 ± 0.10 σ 15 0 nb counts I 1) Gaussian shape for neutrino10 component Two hypotheses for theν ¯e signal extraction: 5 −2 I 2) Stability of the PSD spectrum shape of 0 Residuals ( 0 0.05 0.1correlated0.15 background0.2 0.25 −4 −2 0 2 4 σ Vladimir Savu (CEA) PSD 06.12.2019 – AAPResiduals 2019 ( unit) 19 / 23 The STEREO experiment Oscillation analysis Oscillation analysis

Oscillation test: look for relative distortions of theν ¯e -spectrum between cells X reduced systematics Counts X prediction independent

Without oscillation Ncells NE Center cell with oscillation X X Dl,i − φiMl,i(µ, σ, α) Border cell with oscillation χ2 = σ 1 2 3 4 5 6 7 l i l,i Evisible [MeV] Ncells !2 Ncells !2 !2 X αNormU X αEscaleU αEscaleC + l + l + 0 σNormU σEscaleU σEscaleC l l l l 0

Dl,i: measured spectra Ml,i: simulated spectra – taking into account cell diﬀerences, detection eﬃciencies etc. φi: free normalization parameter in energy bin i common for all cells

{α}: nuisance parameters taking account systematics (energy scale, uncorrelated norm)

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 20 / 23 The STEREO experiment Oscillation analysis Non oscillation hypothesis

Cell 1 Cell 2

10 10

8 8

6 6 Very good agreement

Rate / day I Rate / day 4 4

2 2 between data and

2 3 4 5 6 7 2 3 4 5 6 7 non-oscillated model Reconstructed Energy (MeV) Reconstructed Energy (MeV) Cell 3 Cell 4 10 8 I no sterile hypothesis 8 not rejected 6 6

1000 Phase-II prediction-independent shape-only no oscillation hypothesis 4 4 Rate / day Rate / day Counts 800 2 2 600 p-value = 0.40

2 3 4 5 6 7 2 3 4 5 6 7 400 Reconstructed Energy (MeV) Reconstructed Energy (MeV) 200

Cell 5 Cell 6 0 8 0 5 10 15 20 25 30 8 ∆χ2 6 6

4 4 minimized pull terms stay Rate / day I Preliminary Rate / day 2 2 within ±1σ 0 2 3 4 5 6 7 2 3 4 5 6 7 Reconstructed Energy (MeV) Reconstructed Energy (MeV)

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 21 / 23 The STEREO experiment Oscillation analysis Exclusion contours

101

I Phase-II results ) 2

eV 1 ( I Raster-scan method 2 41 ∆χ2 distributions estimated m

∆ by MC pseudo experiments

RAA 95% C.L. RAA 99% C.L. RAA: Best ﬁt I Best-ﬁt value of the RAA Stereo: (2011) Exclusion Sensitivity 90% C.L. Preliminary rejected at 99 % C.L. 1 10− Exclusion (119 days) 90% C.L.

1 10− 1 2 sin (2θee)

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 22 / 23 The STEREO experiment Oscillation analysis Conclusion and perspectives

STEREO detects ν¯e at 10 m from reactor core with high precision I 43.4 k¯νe detected in phase-II, 65.5k phase I + phase II I 179 days ON, 233 days OFF show a very high stability of the background I Initial RAA contours (2011) now mostly excluded, best-ﬁt point at 99% C.L., with no signs of cell-to-cell systematics I High stability of the liquid scintillator for a long period of time I Good background rejection due to shielding and pulse shape discrimination I Very good control of the detector’s response and of the neutron eﬃciency

Preliminary Stay tuned! Perspectives in the near future: I Upcoming oscillation analysis long paper I Absolute measurement of the pure 235U antineutrino ﬂux I New reference spectrum shape for 235U ﬁssion I Statistics to be doubled by mid-2020

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 23 / 23 Spokesperson: David Lhuillier (CEA) Contact: [email protected] Website: www.stereo-experiment.org

L L

Photo: S. Schoppmann I

: o t o The STEREO Collaboration h P Energy reconstruction

−→ Q are the collected charges −1 M matrix constructed from regular monitoring: mij = Ci · Lji −−→ −1−→ Erec = M Q 54 I Ci calibration coeﬃcients( Mn radioactive source)

I Lji cross-talks between cells (cosmics)

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 1 / 6 Background stability and spectrum(Preliminary)

Background measured during reactor-oﬀ periods. 233 days available → high-statistics for stability tests

hPSD_MuDevCorrected_Cell7_Ebin1.625_HighLevelPoolhPSD_MuDevCorrected_Cell7_Ebin1.625_HighLevelPoolhPSD_MuDevCorrected_Cell7_Ebin1.625_LowLevelPool 1.8 Entries 0 ref 2 OFF1,Mean =15Entries m : 47 days 120460348914 0.05097 χ Std Dev 0.319 / ndf 310 / 265 2 Mean 0.13470.1358 e-recoils 140 χ / ndf 43.3 / 43 Slope (hPa-1.R-1 ) -0.00598 ± 0.00049 p0 Std Dev −0.1231 0.037920.04644 ± 0.0772 1.6 ref p-recoils 120 a × OFF2, =7 m : 136 days

2 Rate (/day) Rate / R χ / ndf 374 / 265 100 × Slope (hPa-1.R-1 ) -0.00620 ± 0.00020 OFF1 - a OFF2 1.4 ref 80 60 1.2 40 20 1 0

0 0.05 χ2 / ndf0.1 0.15 0.2 0.25 0.3 0.35 0.4 1.2 26.99 / 32 Q / Q tail tot

0.8 OFF2 Ratio 1.001 ± 0.003 OFF1 × 1.1 a 1 960 970 980 990 1000 1010 Pressure [hPa] 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Qtail / Qtot Correlation of the IBD candidates rate with PSD distribution for two independent reactor-oﬀ atmospheric pressure, for e-recoils et p-recoils dataset with diﬀerent pool water level

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 2 / 6 Background shape and S/N ratio

e-recoil correlated background accidental background p-recoil correlated background /500keV ] /500keV 60 -1

< S/B) > = 0.6 tot < S/B) > = 0.8 corr

Rate [ day Rate 40 < S/B) > = 6.8 acc

0 2 3 4 5 6 7 8 Reconstructed energy [MeV] Prompt background energy spectrum, decomposed into e-recoil, p-recoil and accidental components

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 3 / 6 Residuals(Preliminary)

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 4 / 6 1/L2 law(Preliminary)

1.3

[a.u.] χ2 / ndf 2.55 / 5

det,l 1.2

∈ Prob 0.769 / ,l ν

N 1.1

0.9

0.8 9.5 10 10.5 11 Mean Cell Baseline [m]

ν¯e ﬂux as a function of the reactor distance

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 5 / 6 ● STEREO and PROSPECT are high-precision HEU reactor experiments

● To increase the physics reach of our respective experiments, we are joining forces to combine our spectral measurements

● We have begun the process of sharing data and information to lay the groundwork for this analysis

235 ● We aim to produce a joint result that better illuminates the shape of the U antineutrino spectrum and compare it to leading theoretical models

Vladimir Savu (CEA) 06.12.2019 – AAP 2019 6 / 6