Development of a High Pressure Xenon Imager with Optimal Energy Resolution Azriel Goldschmidt on work with David Nygren Instrumentation Series Seminar LBNL, March 2009

Physics Motivations Neutrinoless double beta decay (bb0n): Tests Majorana nature of neutrino Helps determine absolute neutrino mass If observed, lepton number NOT conserved Current situation: controversial (one claim), may require new and richer approach WIMP dark matter: Direct detection This talk focuses on the bb0n

Rare nuclear transition between same mass nuclei Energetically allowed for even-even nuclei (Z,A) → (Z+2,A) + e-1 + n1 + e-2 + n2 (Z,A) → (Z+2,A) + e-1 + e-2 (Z,A) → (Z+2,A) + e-1 + e-2 + c

Double Beta Decay Spectra mention that a number of other exotic particles can also contribute to BB decay.

H-M Claim Inverted 50 meV Or ~ 1027 yr Normal

How to look for neutrino-less decay Measure the spectrum of the electrons

bb0n Experiments CANDLES 48Ca CaF2 scintillator crystals COBRA 116Cd CdZnTe crystals CUORE 128Te TeO2 Bolometers EXO 136Xe Liquid Xenon TPC GERDA 76Ge Enriched Ge diode MAJORANA 76Ge Enriched Ge diode SNO+ 150Nd Nd loaded liquid scintillator SuperNEMO 82Se Foils in track/calorimeter Indicate the variation in expected rate (modulo m^2) from the matrix element calculations

Past Results 48Ca >1.4x1022 y <(7.2-44.7) eV 76Ge >1.9x1025 y Elliott & Vogel Annu. Rev. Part. Sci. 2002 52:115 Past Results 48Ca >1.4x1022 y <(7.2-44.7) eV 76Ge >1.9x1025 y <0.35 eV >1.6x1025 y <(0.33-1.35) eV =1.2x1025 y = 0.44 eV 82Se >2.1x1023 y <(1.2-3.2) eV 100Mo >5.8x1023 y <(0.6-2.7) eV 116Cd >1.7x1023 y <1.7 eV 128Te >7.7x1024 y <(1.1-1.5) eV 130Te >3.0x1024 y <(0.41-0.98) eV 136Xe >4.5x1023 y <(1.8-5.2) eV 150Nd >1.2x1021 y <3.0 eV

H-M: Only claimed evidence of 0nbb detection with 11 kg of 86% enriched 76Ge for 13 years Klapdor-Kleingrothaus et al Phys.Lett.B586:198-212,2004. T1/2~1.19x1025y <m> ~ 0.44 eV NIM A522, 371 (2004)

CUORE Cryogenic “calorimeters” CUORICINO: 40.7kg TeO2 (34% abundant 130Te) T0n1/2 ≥ 3.0 × 1024 yr (90% C.L.) <mn> ≤ 0.19 – 0.68 eV Resolution DE/E = 2 x 10-3 FWHM at 2.5 MeV CUORE ~1000 crystals, 720 kg 60Co (cosmogenic) decays, 2 gammas summed

“Gotthard TPC” Pioneer TPC detector for 0-  decay search 5 bars, enriched 136Xe (3.3 kg) + 4% CH4 MWPC readout plane, wires ganged for energy No scintillation detection  no TPC start signal! No measurement of drift distance E/E ~ 80 x 10-3 FWHM (1592 keV) 66 x 10-3 FWHM (2480 keV) Reasons for this less-than-optimum resolution are not clear… Likely: uncorrectable losses to electronegative impurities Possible: Undetectable losses to quenching (4% CH4) But: ~30x topological rejection of  interactions!

Charge & scintillation EXO-200 200 kg Enriched 136Xe Charge & scintillation light readout

EXO-200 expected E resolution Anticorrelation between ionization and scintillation signals in liquid xenon can be used to improve the energy resolution 570 keV g Extrapolates to dE/E = 33 10-3 FWHM @ Q0nbb

Energy resolution in a Xe Dual Phase (XENON) Extrapolates to dE/E = 21 10-3 FWHM @ Q0nbb Aprile, Paris 2008

What’s needed… Long lifetimes (>1025 years) require: Large Mass of relevant isotope (>100 kg) Small or No background: Clean materials Underground, away from cosmic rays Background rejection methods: Energy resolution Event topology Particle identification Identification of daughter nucleus Years of data-taking

Why use Xe for bb0n search Only inert gas with a bb0n candidate Long bb2n lifetime ~1022-1023 y (not seen yet) No need to grow crystals Can be re-purified in place (recirculation) No long lived Xe isotopes Noble gas: easier to purify 136Xe enrichment easier (natural 8.9%): - noble gas (no chemistry involved) - centrifuge efficiency ~ Δm (136Xe vs. ~131Xe)

Energy partition in xenon When a particle deposits energy in xenon, where does the energy go? Ionization Scintillation: VUV ~170 nm (1, 2 …) Heat How is the energy partitioned? Complex responses, different for , , nuclei Dependence on xenon density , E-field Processes still not perfectly understood

LXe or HPXe? With high-pressure xenon (HPXe) A measurement of ionization alone is sufficient to obtain good energy resolution…

Xenon: Strong dependence of energy resolution on density! Ionization signal only For  >0.55 g/cm3, energy resolution deteriorates rapidly

DE/E ~ 1-2 10-3 FWHM DE/E ~ 35 10-3 FWHM

What is this factor “G”? In a very real sense: G is a measure of the precision with which a single electron (from an ionizing track) can be counted.

Electro-Luminescence (EL) (Gas Proportional Scintillation) Electrons drift in low electric field region Electrons then enter a high electric field region Electrons gain energy, excite xenon, lose energy Xenon generates UV Electron starts over, gaining energy again Linear growth of signal with voltage Photon generation up to ~1000/e, but no ionization Early history irrelevant,  fluctuations are small Maybe… G ~ F?

Electroluminescence in 4.5 bar of Xenon Corresponds to 5 x 10-3 FWHM When naively extrapolated to Qbb of 2.5 MeV (compare with the Fano limited 2.8 x 10-3 FWHM best case)

Fluctuations in Electroluminescence (EL) EL is a linear gain process G for EL contains three terms: Fluctuations in nuv (UV photons per e): Fluctuations in npe (detected photons/e): Fluctuations in photo-detector single PE response: G = 2 = 1/(nuv) + (1 + 2pmt)/ npe) For G = F = 0.15  npe ≥ 10 The more photo-electrons, the better! Equivalent noise: much less than 1 electron rms!

Virtues of an EL readout Immune to microphonics Absence of positive ion space charge Linearity of gain versus pressure, HV Isotropic signal dispersion in space Trigger, energy, and tracking functions accomplished with optical detectors

Detector Concept Use enriched High Pressure Xenon TPC to provide image of the decay particles Design to also get an energy measurement as close to the intrinsic resolution as possible

High-pressure xenon gas TPC Fiducial volume surface: Single, continuous, fully active, variable,... 100.00% rejection of charged particles (surfaces) TPC with t0 to place event in z coordinate Tracking: Available in gas phase only Topological rejection of single electron events

Separated Function TPC with Electroluminescence Readout Plane A - position Readout Plane B - energy Electroluminescent Layer

Electro-Luminescent Readout For optimal energy resolution, 105 e- * 10 pe/e- = 106 photoelectrons need to be detected! Energy readout plane is a PMT array electron (secondary) drift is very slow: ~1 mm/s This spreads out the arriving signal in time - up to 100 s for many  events The signal is spread out over the entire readout cathode-side, 100’s of PMTs These two factors greatly reduce the dynamic range needed for readout of the signals  No problem to read out <5 kev to >5000 keV

Single 2.49 MeV e- in 20 atm Xe (background) MC simulation ~5 cm Using tracking information, separate single electrons (like these) from 2-electron events that should have “blobs” at BOTH ends of the combined track

Backgrounds for the bb0n search NEXT Collaboration

Can one measure Ba++ Directly? Extract the ion from the high pressure into a vacuum Measure mass and charge directly A mass 136, ++ ion is a unique signature of Ba++. (Assumption is Xe++ cannot survive long enough to be a problem) This has been done for Ba++ in Ar gas Sinclair, TPC Workshop Paris 2008

Barium ions are guided towards the exit orifice and focused using an asymmetric field technique. The second chamber is maintained at a pressure of ~10-30 mb Using a cryopump and is lined with an RF carpet. An RF funnel guides the ions Towards the RF quadrupole which is at high vacuum. The ion is identified using TOF and magnetic rigidity Sinclair, TPC Workshop Paris 2008

Top EL/Scint Detector (Tracking) EL Grid Field Cage Cathode Grids Ba Channel Bottom EL/Scint Detector (Energy) Sinclair, TPC Workshop Paris 2008

HPXe and the Dark Matter search Liquid Xenon has the lead on this (since energy resolution is not critical), however HPXe offers better discrimination between nuclear recoils and electrons There are ideas that would enable a lower threshold in the gas phase (useful for testing the DAMA/LIBRA positive result) Challenge at low recoil energy for both LXe and HPXe is that the primary scintillation signal (for trigger ing and fiducialization) is Tiny

Big Impact for WIMP Search in LXe Scintillation (S1) & Ionization (S2) are the signals used to reject electron recoils: S2/S1 But, in LXe: S2/S1 fluctuations are anomalously large Bad news for discrimination power in LXe… (though may be not critical in the presence of self shielding)

7-PMT 20 Bar Test Cell cathode anode + fluorescence grid J. White, TPC08

7-PMT,20 bar Test Cell 1 inch R7378A J. White, TPC08

Going from concept to an R&D program at LBNL Build a test detector large enough to demonstrate ~ DE/E ~ 5 * 10-3 at ~2.5 MeV Gas system (to take out electronegative impurities) Energy side readout (PMTs most likely) Enough “tracking-side” sensors to achieve energy resolution 50 cm size chamber to house ~20 Atm of Xe Monte Carlo simulation for detector optimization ongoing (2 students and 0.25*yours-truly) Groups in Canada/US (part or EXO) and in Spain (NEXT collaboration) pursuing this line of research as well –both healthy competition and collaboration- Stay tuned…and thanks for listening!

Other possible uses of HPXe imagers with optimal resolution Nuclear safeguards Check fuel content of fuel rods Homeland security Directional information from “Compton camera” to identify U and Pu isotopes from a “source”

Summary

Backup Slides

Double beta decay Only 2-v decays Only 0-v decays Rate No backgrounds above Q-value Energy Q-value The ideal result is a spectrum of all  events, with a 0- signal present as a narrow peak, well-separated from 2-

 particles K. N. Pushkin et al, 2004 IEEE Nuclear Science Symposium proceedings A scary result: adding a tiny amount of simple molecules (CH4, N2, H2 ) to HPXe quenches both ionization and scintillation for ’s  particle: dE/dx is very high Gotthard TPC: 4% CH4 Loss(): factor of 6 For  particles, what was effect on energy resolution? Surely small but not known, and needs investigation (~25 bars)

Molecular Chemistry of Xenon Scintillation: Excimer formation: Xe*+ Xe  Xe2*  h + Xe Recombination: Xe+ + e–  Xe*  Density-dependent processes also exist: Xe*+ Xe*  Xe**  Xe++ e- + heat Two excimers are consumed! More likely for both high  + high ionization density Quenching of both ionization and scintillation can occur! Xe* + M  Xe + M*  Xe + M + heat (similarly for Xe2*, Xe**, Xe2*+… ) Xe+ + e–(hot) + M  Xe+ + e–(cold) + M*  Xe+ + e–(cold) + M + heat  e–(cold) + Xe+  Xe*

Energy Resolution Factors in Xenon Gas Detectors Intrinsic fluctuations Fano factor (partition of energy): small for  < 0.55 g/cm3 Loss of signal (primary): Recombination, quenching by molecular additives (heat) Loss of signal (secondary): Capture by grids or electronegative impurities Gain process fluctuations: Avalanche charge gain fluctuations are large Gain process stability: Positive ion effects, density and mix sensitivity,... Long tracks  extended signals Baseline shifts, electronic non-linearities, wall effect,...

 Sensitivity Issues Target (from oscillations): m ~0.050 eV = 50 meV Masses could be higher… ∑m < 0.61 eV There are ~109 relic neutrinos for each baryon  the total  mass could be  ∑all visible matter Goal: 100’s to 1000’s kg active mass likely to be necessary Rejection level of internal/external backgrounds: Less than one event per ~1027 atoms/year! Energy resolution needed: E/E <<10 x 10-3 FWHM, with gaussian behavior

Xenon: Strong dependence of energy resolution on density! Ionization signal only For  >0.55 g/cm3, energy resolution deteriorates rapidly

TPC:  Signal & Backgrounds -HV plane Readout plane B Readout plane A Fiducial volume surface . * electrons ions Signal:  event Backgrounds