Next Steps In Applied Antineutrino Physics at LLNL

1. Next Steps In Applied Antineutrino Physics at LLNL Adam Bernstein, Group Leader, Advanced Detectors Group, Lawrence Livermore National LaboratoryDec 14 2007 neutrinos.llnl.gov nuclear.llnl.gov

2. Outline • LLNL/SNL Work to Date • Thoughts about Practical Near-Field Monitoring • Next Steps at LLNL • Range Extension: Studies of Argon Coherent Scatter Detection • Understanding Backgrounds and Operating at the Surface of the Earth • White Paper Status

3. “Standard” Applied Antineutrino Physics at LLNL/SNL Determine on/off status within 5 hours with 99.9% C.L. Track Pu content to~50 kg - with known power and initial fuel content Measure thermal power to 3% in one week burnup model with one free parameter 0 Time in hours Detector is stable to ~ 1%; burnup is ~ 10% 130 J. App. Phys. publication pending Relative count rate 1.5 tons 235U consumed 250 kg 239Pu produced Continuous, non-intrusive, self-calibrated, unattended, low cost and channel count, operable for months to years with rare maintenance NIM A 572 (2007)

4. What is Needed for Near-Field Cooperative Monitoring and Safeguards ? • 3x3x3 meter deployment at SONGS is already demonstrably non-intrusive for reactor operators • Acceptance depends on ability to meet diversion detection goals, cost, ease of use and operator/IAEA acceptance - not primarily on the physical footprint • there is plenty of physical space with overburden in many safeguarded reactors worldwide • New designs will be non-toxic, have negligible flammability, no cryogenics, be self-calibrated and easy to deploy • For Near-Field Monitoring at 10-100 meters, Inverse Beta Detectors May Suffice • What is the Use of Coherent Scatter Detectors ? • Above-Ground Detection Would Expand Deployment Opportunities

5. Antineutrino ν + Ar → ν + Ar Basic Principles of Coherent Scattering Neutrino-nucleus scatter coherent for En < 50 MeV (in Argon) supernova, solar, reactor neutrinos Recoil energies among noble elements Argon (Z=18) gives the greatest number of detectable ionizations per unit mass AtomicNumber Cross-section Neutron Number

6. Beyond Cross-Section: Detectable Coherent Scatter Rates For Reactor Monitoring • Discovery - Ge and Ar both have potential • Exploitation – Argon enjoys scalability and (possibly) cost advantages • Full CNS detection efficiency required for any significant reduction in footprint • Near-field deployment needs already well met with well engineeredinverse beta detectors • There may be promise in scaling to 0.1-10 kilometer ranges

7. 1-10 primaryscintillationphotons in liquid –very difficult to see these 1-10 primaryionization electrons(after quenching >~25 photoelectronsper primary electron Herein lies the signal This signal strength has already been measured in existing ten kg noble detectors

8. Predicted Signal and Background in a 10 kg detector Ar-39 the dominantbackground: what canbe done about it ? The neutrino signal including nuclear quenching Modest (few cm) passive shieldssuffice to screenexternal backgrounds In this simulation: external neutrons and gammas, internal Ar-39 Not yet in this simulation: PMTs ~10000 emitted gammas per day (20 mBq/tube)

9. Are PMT Backgrounds Manageable ? Simulated internalbackgrounds In 100 kg xenon detector (5 keV threshold) 104 suppression • ~10,000 decays per day total from PMTs - must incorporate in model Fiducial and energy cuts shouldsuppress these: most PMT gammaswill be above energy thresholdor multiply scatter Real backgrounds 10 kg xenon detector (4.5 keV threshold) 59 days livetime

10. The Ar-39 beta background • 565 keV endpoint – 0.9 Bq/kg in Natural Argon • An important background for coherent scatter • Gram quantities of depleted Ar created by recovery from underground natural gas reserves (Princeton, Calaprice, Galbiati et. al.) • Kilogram quantities manufactured by Russian group Activity limit : at least a factor of 20 lower than natural Argon • This would eliminate Ar-39 as a concern for coherent scatter in Ar • cost could be an issue • discovery can be done even with Natural Ar

11. Is The Signal Within Reach of Existing Dual-Phase Detectors ? Lossless drift of electrons over 10 cm distances amply demonstrated in many LAr/LXe experiments – Argon purification techniques are well understood Sensitivity to single primary electrons – accomplished in 10 kg Xe detectors (ZEPLIN, XENON10) Quench factor: gas-phase quench measurement, consistent with predictions, has been measured at LLNL – this must be repeated in liquid

12. Field cage mounted inside Argon-filled chamber 12 in. Gas Phase Studies of Very Low Energy Nuclear Recoils Calibration & Noise-floor estimation By studying nuclear recoils in the gas phase, we learn about: ionization, gas phase quenching, light collection, scintillation properties Calibration55Fe 5.9 keV X-rays Noise wall Single- photoelectron response of PMT Energy (integral units) • Only 1 PMT in this detector • ~20 in full scale detector • 1% Ni for wavelngth shifting

13. 60 keV Neutron Source: Neutrons Recoils at 8 keV and below Borated plastic Neutron shield Lead Gamma shield LLNL LINAC Li-target ~60 keV neutron generator Argon detector 7Li (p,n) 7Be 100 Hz rep. rate ~105 neutrons / spill Neutron beam Gamma Background 478 keV from 7Li(p,p’)

14. The Predicted Recoil Spectrum Actual detector response including PMT coll. eff. Predicted effect of quenching Deposited energy(before quenching) keV 2-4) Neutron kinematics, quenching optical collection efficiency 1) Incident neutrons selectedby resonance

15. 200-μsec time trace during neutron beam measurement - Gate width Extended event ~6 s Single p.e. ~20 ns X-ray or neutron ~2 s A Menagerie of Raw Events ←Neutron beam on→

16. Gamma signal only above neutron recoil threshold Energy threshold for neutron recoils Extraction of a Quench Factor –the Lowest Ever Measured in Ar ? Derived Quench factor: (preliminary) 0.22 Predicted: 0.2 Residual signalattributed to neutrons 8 keV neutron recoilgenerates 1.8 keV electronequivalent energy deposition

17. For comparison: Liquid n-recoil Results from McKinsey Group, Yale

18. A 10 kg Liquid Argon Coherent Neutrino Detector HV Design by W. Stoeffl Coherent ScatterGroup: Chris Hagmann Celeste Winant Kareem Kazkaz Igor Jovanovic Michael Foxe Wolfgang Stoeffl Turbo-pump Pulse tube fridge PMTs Valves Support Bellows Level gauge Gain region Super Insulation Drift region Liquid Nitrogen transport reservoir Insulation Vacuum Liquid Argon 87K

19. Background Considerations for Antineutrino Detectors at the Surface of the Earth • Veto trigger rates increase by 5-10 relative to ‘SONGS1’ - what about deadtime ? • Correlated backgrounds gammas neutrons, pions, protons - are an additional concern, beyond the usual problem of time-correlated events from muons We are justbeginning to studythis problem http://abyss.uoregon.edu/~js/glossary/cosmic_rays.html

20. First Consideration: Shrink Deadtime By Shrinking Detector SONGS (3 meter)3 veto - ~30% dead at sea level (~5% at 10 m.w.e.) (1.5 meter)3 detector/veto – ~5% dead time at sea level - but more elaborate vetoing strategies may be needed * “Standard” veto (100 microsecond following any cosmic) Example: (water detector now deployed at SONGS, below ground) Target (1.5 m)3 Current (3 m)3

21. Second Consideration: Studying Above Ground Time Correlated Backgrounds • Characterize with Monte Carlo • Measure in meter2 detector arrays (Muon, Liquid Scint., Plastic, 3He) • Deploy existing prototypes at SONGS and measure signal and background empirically in antineutrino detectors • Explore alternative means to reject backgrounds • Water Cerenkov detectors • Segmentation (Jim’s talk) • Others..

22. Monte Carlo Generation and Detection of Sea Level Backgrounds A) Public Code package CRY: nuclear.llnl.gov Due to strong natl. lab interest in surface detection of plutonium and uranium, codes exist to study time correlated backgrounds at sea level – like antineutrinos, fission chains are highly time correlated All secondaries propagated through 42 layers of atmospheretime correlated energy spectra recorded with up to 300 m horiz separation B) GEANT and MCNP models of detectors

23. Benchmark examples from sea-level flux (CRY) code - Muon, Pion, Neutron energy spectra match h.e. data

24. A First Comparison (For Us) Of Sea Level Showers In Meter2 Detector Arrays Cumulative number of counts Simulation Time until Next count 2 minutes Cumulative number of counts Data Time until Next count 2 seconds Mixed Array of 3He, NaI, PSD and plastic– 100 detectors, here near 1 ton of lead

25. B) Initial Background Modeling For Water Cerenkov Detectors • Fast neutrons should not be problem since they are below the Cerenkov threshold up to high energies • But: energy scale is smeared by low light collection 250 kg detectorNow deployed below-ground At San Onofre above-ground test in ’08-09

26. Conclusions • Dual Phase Detectors appear to have the sensitivity needed forcoherent scatter discovery • Significant Infrastructure for background measurement and modeling at the Earth’s surface will help guide and small surface antineutrino detector designs

27. A Range of Applications

28. White Paper: A Review of the State of the Art in Antineutrino Detection as Applied to Nonproliferation of Nuclear Weapons Summary for policy and physics community to understand state of the art and R&D program • 1 Introduction 1 • 2 Current Safeguards and Cooperative Monitoring Practice for Light Water Reactors 3 • 3 Current Nuclear Explosion Detection Technology 10 • 4 Production and Detection of Antineutrinos From Nuclear Reactors and Nuclear Explosions 11 • 5 Near-Field Detection 13 • 6 Mid-Field Detection 13 • 7 Far-Field Detection 14 • 8 Overview of Fundamental Physics Using Reactor Antineutrinos 17 • Research and Development Needs for Basic and Applied Antineutrino • Detection 25 Inputs received for all but two chapters editing in progress..