Latest results of the STEREO Experiment a Search for a Sterile - - PowerPoint PPT Presentation

https://lpsc.in2p3.fr/ DocDB/ 0006/000673/001/logo- ILL.png

Supported by :

Latest results of the STEREO Experiment

a Search for a Sterile Neutrino ~1eV at Short Baseline

The 11th of September

Adrien Blanchet

(CEA Paris-Saclay)

• n behalf of the STEREO collaboration

TAUP 2019

2

Worldwide Experimental Program

STEREO PROSPECT Solid Neutrino 4 DANSS NEOS

Research reactor (Pure U235), Li-loaded Research reactor (Pure U235) Gd-loaded Power reactor (Mixed Pu9/U5), Gd-loaded

The STEREO Experiment

STEREO : Experimental Site

4

ILL research facility, Grenoble, France

Research reactor core ~ 58MWth

• Little damping of the oscillation pattern
• Compact core (∅40cm × 80cm)
• Short baseline
• Pure U235 neutrino spectrum : Highly enriched fuel

9.4 m < Lcore < 11.2 m → 1019¯ νe/s Challenging Mitigation of the Background

• Gamma & Neutron Background from neighboring

experiments --> heavy passive + active shielding

• Surface-level experiment (15 m.w.e overburden

thanks to water channel)

Autumn 2016

93 tons moved

• n air cushions

Autumn 2016

STEREO : The Detector

5

Oscillation analysis based on the relative comparison of the cell spectra. No dependence on the prediction of the reactor spectrum.

• Little damping of the oscillation pattern

also thanks to good energy resolution

STEREO : Detection Principle

6

Interaction Channel : Inverse Beta Decay (IBD)

Prompt Delayed

IBD tagging via Prompt- Delayed correlated pairs

Detection Efficiency ~ 60%

ΔT

Correlation distance :

STEREO : Data Taking

7

Phase-I ON = 66 days OFF = 22 days Phase-II ON = 119 days OFF = 211 days

• Data taking efficiency: 98.5%
• Inefficiencies for neutrino data :
• Calibrations: 3.3%
• 14% dead-time from selection cuts.

Data Analysis

0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4

Reconstructed Energy (MeV)

0.0 0.2 0.4 0.6 0.8 1.0

Counts (a.u.)

Data MC Data MC

2 4 6 8 10

Nominal Energy (MeV)

0.96 0.97 0.98 0.99 1.00 1.01 1.02 1.03 1.04

Data/MC Erec detector

Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6 Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6 Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6 Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6 Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6 Cell 1 Cell 2 Cell 3 Cell 4 Cell 5 Cell 6 Cs Mn Zn K AmBe n-H Co Na AmBe AmBe n-Gd

Prompt Signal : E Reconstruction

9

Energy calibration

• Weekly calibration with 54Mn sources
• Monitoring light leaks between cells using

cosmics data

• Dedicated algorithm to go from deposited

charge (Q) to reconstructed energy (E) parametrized by calibration coefficients (C) and light leaks coefficients (L) :

Testing detector's response with several gamma sources

Qj = ∑

i

MijEi = ∑

i

CiLijEi

Ei = ∑

j

(M−1)ijQj

Anchoring energy scale with Mn54 gammas

Remaining Background : Cosmic Induced

10

PSD properties of the LS allow to discriminate neutrinos from dominant remaining cosmic background

• PSD distributions are corrected from temperature

& evolution of optical properties

• ON data fitted using OFF model
• Background model given by OFF data
• Accidentals accurately measured and

included in the model

Neutrino rates are extracted for each cell and Energy bin Extraction relies on two hypothesis only

• Gaussian shape of the neutrino component
• Stability of the PSD spectrum shape of correlated

background

Using Pulse Shape Discrimination (= Qtail/Qtot)

T

• tal charge - Qtot

T ail charge - Qtail p-recoil signal τe e-recoil signal Qtail/Qtot (e) Qtail/Qtot (p) τp < <

Background Stability

11

Major Background Contribution : Cosmic Induced

• Clear correlation between atmospheric pressure

and counting rate

PSD background shapes independent

• f environmental conditions
• Splitting data in 2 groups : Low and High pressure
• Normalization factor = (93.3 ± 0.25)%
• Expected for a 10 hPa difference = (93.8 ± 0.3)%

High stability of the shape of the background PSD distribution is demontrated for all cells

Involved systematic uncertainties :

(Absolute-Predictions-less) Oscillation Analysis

12 Fully free parameters

• nly constraint : common to all cells

Δi = Δc=cell ; b=Ebin ≐ Dcb − ϕbMcb(Δm2

14, θ14)

Data MC

Source Uncertainty Cell-to-Cell Correlation Study

Cell-to-cell correlated Energy-Scale ± 1.1 % 100 %

• Time stability of the energy response with cosmic-

induced events

• Anchoring of the MC

Cell-to-cell uncorrelated Energy-Scale ± 0.5 % 0 %

• Cell-to-Cell deviations

Normalisation Factor ± 1.2 % 0 %

• Neutron capture efficiencies

(Absolute-Predictions)-less Oscillation Analysis

13

• Raster-scan method (best fit search in a given ∆m²

slice)

• Due to detector maintenance phase I & II can be

considered as independent measurements.

• Preliminary combination is done by summing the two χ²

and assuming a standard χ² law for the C.L.

• Best-fit value of the RAA rejected at C.L. ~ 99.8%
• Sensitivity still dominated by the statistical errors

Δχ2 = χ2(Δm2

14, θ14) − χ2 best fit χ2 = ∑

i ∑ j

Δi (V−1

cov)ij Δj

Contours drawn using a Δχ² :

Statistical inference cross-validated by 2 independent studies : χ2 = ∑

i (

Δi σstat.

i

)

2

+

syst.

∑

j (

αj σsyst.

j

)

2

with Covariance Matrix : with Nuisance Parameters :

10−1 1

sin2(2θee)

10−1 1 101

∆m2

41(eV2) RAA 95% C.L. RAA 99% C.L. RAA: Best fit Stereo: Grenoble Contour PII Saclay contour PII

Preliminary Preliminary

Pull Terms Contour (PII) Covariance Contour (PII)

Perspectives

• Neutrino detection efficiency
• Mostly depends on neutron event
• Topology cut efficiency depends on the spectrum

shape of the Gd cascade : Improved description in MC with FIFRELIN (cf. arXiv:1905.11967 + 10M cascades available on zenodo:2653787)

• MC efficiency corrected for border effects of the

Target

• Thermal Power estimation
• Well controlled uncertainties thanks to the ILL

redundant and regularly calibrated instrumentation

• Current estimate :

Testing RAA with Pure U5 Reactor Fuel

15

Uncertainty dominated by two components

Data/MC neutron detection efficiency

δP/P ≃ 1.4 %

Stereo experiment has a good potential to check the RAA thanks to the control of the two dominant systematics Current estimate of experimental uncertainty : 2.4 %

• Predicted Spectrum
• Pure U235 Huber Spectrum
• + % level corrections in the first 2 energy bins

(n-Al capture, off-equilibrium effect, spent fuel)

• Quantified comparison with a spectrum

prediction requires high precision in the treatment of the energy scale

• Residual distortions between

measured and simulated reconstructed energies are described by the most general function:

• New stringent constraints,

complementary to the calibration sources, can be added to the delta function by considering the boron data.

Testing the Spectrum Shape

16

STEREO Phase-II data only

EData = EMC (1 + δ(EMC))

δ(EMC)

4 6 8 10 12 14

Reconstructed energy (MeV)

2 −

10

1 −

10

events/day/100 keV

Data MC

B Spectrum

12

Capture of stopping muon at rest : Coincidence signal :

• Prompt (stopping muon)
• Delayed (beta from Boron 12 decay)

Boron 12 extracted :

• Dominant accidental background is measured on line.
• ~ 700 events/day
• S/B ratio ~ 0.1 to 1

Extracting Boron12 Spectrum

17

μ− + 12C → 12B

Boron12 in Data Boron12 in MC

Simulating the spectrum :

• Beta spectrum prediction using 3 main branches
• Vertices distribution from cosmic muon simulation
• Nitrogen12 background fitted (% level contribution)

Systematic Uncertainties :

• Weak magnetism corrections
• Distribution of µ-capture vertices
• Radiative corrections

Preliminary

1 2 3 4 5 6 7 8 9

(MeV)

rec MC

E

0.02 − 0.01 − 0.00 0.01 0.02

• 1

rec MC

/ E

rec Data

) = E

rec MC

(E δ

• Calib. Points

2 4 6 8 10 12 14

(MeV)

rec MC

E

0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 1.20

MC B

12

/ S

Data B

12

S

B Ratio

12

• Syst. only

B Spectrum Ratio

12

STEREO Neutrino spectrum range

Global Fit of the Energy Scale

18

EData EMC − 1 = δ(EMC)

SData SMC = 1 + δ(EMC) + EMC (δ′(EMC) + δ(EMC) S′MC SMC )

Point-like energy constraints (calibration sources) Spectrum constraints (Boron 12)

Common fit of

δ(EMC) can be chosen as a polynomial parametrization, or with a non-parametric expression (KDE)

δ(EMC)

Preliminary Preliminary

At the end, propagate the function on the neutrino spectrum and its uncertainty to infer on the Huber spectrum prediction

δ(EMC)

Stay tuned..!

Conclusion and Outlook

STEREO demonstrates its high precision capability:

• Detection – 43.4 k neutrinos detected in phase-II, 65.5 k total
• Backgrounds – Extensive measurements (233 days OFF and 185 ON) show a very high

stability of the background

• Sterile ν exclusion contours – Major fraction of the initial RAA contour is now rejected with

no sign of cell-to-cell systematics

• Fine-tuned MC allowing an accurate description of all relevant observables. Improved

control of the neutron detection efficiency thanks the the Fifrelin model. 1e7 Gd cascades available to the community on Zenodo.

Perspectives in the (very) near future:

Conclusion & Outlook

20

• Long paper on the oscillation analysis being written.
• Absolute flux measurement: potential of accurate measurement of the pure 235U neutrino

flux to test the RAA.

• Improved energy scale in development to provide a new reference of the 235U fission

neutrino spectum

• Statistic to be double by mid-2020.

Thanks for your Attention !

Conclusion & Outlook

Photo: ILL

Spokesperson: David Lhuillier (CEA) Contact: david.lhuillier@cea.fr Website: http://stereo-experiment.org

The STEREO Collaboration

Photo: S. Schoppmann

21

Backups

• 2011 : Re-evaluation of the neutrino reactor flux prediction
• Reactor Antineutrino Anomaly (RAA)
• All reactor short-baseline experiments are observing a deficit (~6%)
• Confirmed by recent accurate measurements from Daya Bay, RENO & Double Chooz
• Adding a fourth "sterile" neutrino may explain the deficit :

Motivation : Testing the Sterile Hypothesis

23

Δm2

14 = 2.3 eV2 ; sin2 (2θ14) ≃ 12 %

6% deficit

• Phys. Rev. D 83, 073006 (2011)

➡ STEREO will be able to measure an oscillation pattern without prediction

10.1007/JHEP06(2017)135

Motivation : Nuclear Physics Biases

24

Shape Anomalies : Several experiments revealed a "bump" around 5 MeV w.r.t. predicted spectrum

• Could be linked to underestimation of some

isotopes of uranium/plutonium

• Cannot explain the total deficit
• Everyone does not see the same spectral distortion

Rate Anomalies : Studies of decorrelation

• f isotopes contributions at commercial

reactors

→ Claim that the deficit is carried by U235

➡ STEREO will measure the pure U235 neutrino spectrum

10.1016/j.physletb.2017.08.035 10.1103/PhysRevLett.122.232501

Corrections on Huber Spectrum Prediction

25

antineutrino energy [MeV] 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4

• Relat. correction to pred. spectrum

0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

Al

28

Off-equilibrium Spent fuel

0.003 ± 0.103 0.002 ± 0.006 0.001 ± 0.004 bin

st

1 bin

nd

2 bin

rd

3

SH(Eν) → SH(Eν) × (1+δOE(Eν) + δSF(Eν) + δAl(Eν))

Off-Equilibrium effets Spent Fuel contributions Neutron captures on 27Al

Schreckenbach Beta spectrum have been measured after irradiating a sample 235U sample during ~12h. Long life FP contribution needs to be added a posteriori. STEREO is located under the water channel where the spent fuel is stored. Full history of spent fuel storage is taken into account. 27Al(n,γ)28Al induce neutrino emission from 28Al beta decay (2.86MeV end- point). Reactor core vessel essentially constituted of Al : important correction in the first Energy bin of the measured neutrino spectrum. Al vessel of the ILL reactor core

Preliminary

FIFRELIN's Gd Gamma Cascade

26

• Delayed signal: gamma cascade from (n,γ)Gd

→ correct modelization of primary importance for small detectors (low gamma containment)

• Improved simulation using FIFRELIN deexcitation
• f Gd isotopes using experimental data

completed by nuclear models

PSD Correction

27

Neutrino Extraction Method

28

• Merging of all periods is necessary to

minimize the bias on likelihood estimator

• Bias << statistical uncertainty for all

energy bin

• Constrains our binning choice

Extensive Biases Analysis

29

Note: Here expressed under χ² for better understanding, but using binned Log Likelihood for minimization

Likelihood Estimator can be Biased

∼ A N + B N2 + . . .

χ2 = ∑ bins ON−(a × OFF + Gν + acc) σ + . . .

• Measured neutrino rates corrected for

detection efficiency of each cell found in perfect agreement with the expected 1/L² law:

Results : 1/L² Law

30

Nν ϵdet, cell i ∝ 1 L2

STEREO Phase-II data only

Results : Oscillation Analysis

31

Statistical Inference done with a χ² formalism Oscillation scenarios (H) rejected using Δχ²

Δχ2 = χ2

H − χ2 best fit

χ2 = ∑

i ∑ j

Δi (V−1

cov)ij Δj

Vcov = Vstat + ∑

s=syst

Vs

Systematics

• Cell-Correlated Escale
• Cell-Uncorrelated Escale
• Cell-Uncorrelated Norm

Very good agreement between Data and Model non-oscillated × φi

Results : Oscillation Analysis

32

no-sterile hypothesis not rejected

2 −

10

1 −

10 1

)

14

θ (2

2

sin

1 −

10 1 10

)

2

(eV

2 14

m Δ

RAA contour @ 95 % CL RAA contour @ 99 % CL Best Fit Pseudo-Exp. PDFs @ 90% C.L. Law @ 90% C.L.

2

χ 1 d.o.f.

Preliminary

Validity of Normal χ² Laws for Rejection Contours

33

Sensitivity contours

In principle for relative oscillation analysis, ∆χ²(raster-scan) to p-value inference should be done by generating pseudo-experiences to generate ∆χ² PDFs. However the 1 d.o.f. χ² laws are sufficient in the region of interest

Δi=c,b = Dcb − ϕbMcb(Δm2

14, θ14)

Δχ2 = χ2(Δm2

14, θ14) − χ2 best fit

χ2 = ∑

i ∑ j

Δi (V−1

cov)ij Δj

Relative oscillation analysis used in STEREO :

P-Value (C.L.)

(Phase-II data only)

Signal over Noise Ratio

34 2 3 4 5 6 7

Reconstructed Energy (MeV)

10 20 30 40 50 60 70 80

Counts (/day/500keV)

Electron-recoil type background spectrum

n-H C

12

) γ C(n,n'

12

from γ n + 2 3 4 5 6 7

Reconstructed Energy (MeV)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

BG

/R

ν

R

Signal over Noise Ratio across the spectrum

Spectrum Average

Systematics on the Absolute Neutrino Flux Measurement

35

• Regular and accurate absolute calibration of the T and P
• The dominant uncertainty is from the measurement of the heavy

water flow in the primary circuit :

• The alpha parameter was calibrated with 0.9% precision using a

scale 1 mock-up of the primary circuit. Aging has been neglected.