0- Double-beta Decay, WIMPs, and Xenon: Will the Search Converge? - - PowerPoint PPT Presentation

0-ν Double-beta Decay, WIMPs, and Xenon: Will the Search Converge?

David Nygren LBNL / Stockholm U

Neutrino 2008 2

Conclusion

A high-pressure xenon gas TPC can offer superior performance to LXe for both direct WIMP search and 0-ν ββ search simultaneously How can this be?

Neutrino 2008 3

0-ν ββ ββ: TPC Performance Goals

• 1. Rejection at 100.0% level of charged backgrounds

from surfaces, based on 100.0% active, 100.0% closed, deadtime-free virtual fiducial surface

• 2. Rejection of γ backgrounds using topological

discriminants, to the highest possible level

• 3. Attractive scaling behavior to at least 1 ton
• 4. Energy resolution at 2480 keV:

δE/E <1% FWHM

Neutrino 2008 4

“Conventional” TPC Geometry

• HV plane

Readout plane B Readout plane A .

ions

Fiducial Surface Fiducial Surface

Slide: J J Gomez - NEXT project 5

Topology: spaghetti, with meatballs

ββ events: 2 γ events: 1 Gotthard TPC: ~ x30 rejection

Neutrino 2008 6

Scaling: 1000 kg Xe: ρ ~ 0.1 g/cm3 (~20 bars) ∅ = 225 cm, L =125 cm

A. Sensitive volume B. HV cathode plane C. Readout planes D. Flange for services E. Filler and neutron absorber, polyethylene, … F. Field cages and HV insulator, (rings are exaggerated here)

L

Slide: J J Gomez - NEXT project

Energy resolution: 1% FWHM, or better!

δE/E: 1% FWHM Exposure: 0.5 ton-year mββ = 60 meV

Neutrino 2008 8

WIMP: TPC Performance Goals

• 1. Recoil energy range: 0 < Erecoil < 50 keV

Lowest possible threshold: Emin < 3 keV ?

• 2. Maximum rejection of γ backgrounds:

Ionization/scintillation (S2 /S1) ratios are quite different for nuclear and electron recoils What do the data show?

Neutrino 2008 9

Slide: Elena Aprile

Xenon10

Neutrino 2008 10

Gamma events (e - R) Neutron events (N - R) Why do the γ events show large S2/S1 fluctuations at all energies, but nuclear recoil events do not?

Log10 S2/S1

Neutrino 2008 11

For electrons in LXe, fluctuations in energy partitioning are large

ionization ⇔ scintillation

• a. Landau fluctuations (δ-rays) lead to a highly

non-gaussian distribution of ionization +

• b. High atomic density in LXe, and the

existence of a conduction band ⇒ exaggerated fluctuations in recombination Slow-moving nuclear recoils do not create δ-rays

Slide: Giorgio Gratta - EXO 12

1 kV/cm

Strong anti-correlations in LXe are due to anomalously large fluctuations in energy partition between ionization and scintillation

~570 ~570 keV keV

Bi-207 source

EXO: predicted energy resolution: ~3.4 % FWHM @ Q(ββ) EXO data

Neutrino 2008 13

Big Impact for WIMP Search in LXe!

Scintillation (S1) & Ionization (S2) are the signals used to reject electron recoils: S2/S1 But, in LXe, since S2 and S1 fluctuations are anti-correlated and anomalously large

Bad news for discrimination power in LXe!

Neutrino 2008 14

Gamma events (e-R) Neutron events (N-R) The large fluctuations and anti-correlation

• f S2/S1 leads to a

reduced efficiency for nuclear recoil events

Log10 S2/S1

Neutrino 2008 15

Xenon: Strong dependence of energy resolution on density!

For ρ >0.55 g/cm3, energy resolution deteriorates rapidly

Ionization signal only

Neutrino 2008 16

Xenon: Strong dependence of energy resolution on density!

For ρ <0.55 g/cm3, ionization energy resolution is “intrinsic”

Ionization signal only

Here, the fluctuations are normal Bad!

Neutrino 2008 17

Density ρ < 0.55 g/cm3

• Anomalous fluctuations are absent
• S2/S1 has normal fluctuations, much better γ rejection
• Good news for direct recoil WIMP search !
• Energy resolution is “intrinsic”
• What is intrinsic resolution, and how do I get it?
• δE/E = σN /N x 2.35 (FWHM) σN =(FN)1/2
• F is the Fano factor, a constraint on fluctuations
• N is the number of electron-ion pairs: N = E/W
• Intrinsic resolution from ionization signal only
• I don’t need to measure primary scintillation precisely
• Good news for 0-ν ββ

ββ search !

Neutrino 2008 18

“Intrinsic” Energy Resolution for Ionization at 136Xe Q-value

Q-value (136Xe → 136Ba) = 2480 KeV W = ΔE per ion/electron pair in xenon gas = 21.9 eV, but W depends on Electric field strength, might be ~24.8 eV N = number of ion pairs = Q/W N = 2480 x 103 eV/24.8 eV = ~100,000 electron/ion pairs σN = (FN)1/2 F is the Fano factor - constraint on fluctuations F = 0.13 - 0.17 (measured for xenon gas) - take F = 0.15 σN = (FN)1/2 ~124 electrons rms @ 2480 keV

124: This is a rather small number!

Neutrino 2008 19

“Intrinsic” Energy Resolution for Ionization at 136Xe Q-value

δE/E = 2.35 x (FW/Q)1/2 δE/E ~2.8 x 10-3 FWHM @ 2480 keV

(xenon gas - ionization intrinsic fluctuations only)

Germanium diodes, in practice, δE/E = 1 - 2.4 x 10-3 FWHM Fano factor for LXe: ~20 ⇒ δE/E ≥ 33 x 10-3 FWHM

Neutrino 2008 20

High-pressure xenon gas TPC

• Fiducial volume surface:

– Single, continuous, fully active, variable,... – 100.000% rejection of charged particles (surfaces) – but: TPC needs t0 to place event in z coordinate

• Excellent tracking capability:

– Available in gas phase only! – Topological discrimination against single electrons (meatballs) – X-ray fluorescence can tag γ photo-conversion events

• Can HPXe TPC deliver the energy resolution?

Neutrino 2008 21

Generalization

• If fluctuations are uncorrelated, then

σN = ((F + L + G)N)1/2 F = Fano factor = 0.15 L = loss of primary ionization (set to 0) G = fluctuations & noise in gain process Goal: Make sure that G is smaller than F If F = 0.15, Is this possible ??

Neutrino 2008 22

Avalanche Charge Gain

• Avalanche: exponential amplification

Early fluctuations are amplified exponentially

• for wire (E ~1/r)

0.6 < G < 0.9 *

• σN = ((0.15 + 0.8)N) 1/2 = 328
• δE/E = ~7.0 x 10-3 FWHM

The benefit from a small Fano factor is lost! Micromegas should do better, but, in general… Avalanche devices can’t deliver G < F!

*Alkhazov G D Nucl. Inst. & Meth. 89 (1970) 155 (for cylindrical proportional counters)

Neutrino 2008 23

HPXe TPC: G < F ?

• Answer: Yes, with electroluminescence!

δE/E = 4 x 10-3 FWHM (in principle, @ 2480 keV) Why Electroluminescence? Amplification is linear with Voltage Fluctuations are very small

• With EL, it is possible to realize G = ~0.1
• Each electron is counted with ~10% precision

Neutrino 2008 24

• H. E. Palmer & L. A. Braby
• Nucl. Inst. & Meth. 116 (1974) 587-589



Electroluminescence (EL)

Neutrino 2008 25

55Fe Resolution: 8.4% FWHM

Neutrino 2008 26

55Fe Resolution: 8.4% FWHM

From this spectrum: G ~0.19

Neutrino 2008 27

Fluctuations in EL

G for EL contains three terms:

1. Fluctuations in nuv (UV photons per e): σuv = K/√nuv

– nuv ~ HV/Eγ = 6600/10 eV ~ 660 K < 1 ?

2. Fluctuations in npe (detected photons/e): σpe = 1/√npe

– npe ~ solid angle x QE x WLS x nuv = 0.1 x 0.25 x 0.5 x 660 ~ 8

3. Fluctuations in PMT single PE response: σpmt ~ 0.5 G = σ2 = 1/(nuv) + (1 + σ2

pmt)/ npe)

The more photo-electrons, the better!

Neutrino 2008 28

Fluctuations in EL

G for EL contains three terms:

1. Fluctuations in nuv (UV photons per e): σuv = K/√nuv

– nuv ~ HV/Eγ = 6600/10 eV ~ 660 K < 1 ?

2. Fluctuations in npe (detected photons/e): σpe = 1/√npe

– npe ~ solid angle x QE x WLS x nuv = 0.1 x 0.25 x 0.5 x 660 ~ 8

3. Fluctuations in PMT single PE response: σpmt ~ 0.5 G = σ2 = 1/(660) + (1 + σ2

pmt)/8) ~ 0.19

The more photo-electrons, the better!

Beppo-SAX satellite: a HPXe TPC in space!

Neutrino 2008 30

Neutrino 2008 31

Neutrino 2008 32

Electro-Luminescent Readout

• To keep G < F = 0.15, then: G = (1 + 0.5)/npe

npe > 10/electron For 2480 keV, Ne = 1 x 105 ⇒ Σnpe > 1,000,000

Neutrino 2008 33

Electro-Luminescent Readout

How to detect this much signal? Readout planes are PMT arrays Answer : Use both TPC readout planes

– If EL signal is generated in plane “A” – do “tracking” in Plane “A” – but: record “energy” in plane “B”

Neutrino 2008 34

Electro-Luminescent Readout

How to detect this much signal? Readout planes are PMT arrays Answer : Use both TPC readout planes

– If EL signal is generated in plane “B” – do “tracking” in Plane “B” – but: record “energy” in plane “A”

Neutrino 2008 35

TPC Signal

Transparent -HV plane Readout plane B Readout plane A .

ions

record energy signal here…

Signal: WIMP or ββ event

EL signal created here

Neutrino 2008 36

How to generate EL @ 20 bars?

• Answer: multi-wire plane with ~5 mm pitch

– Optical gain of 300, but no charge gain – Wire diameter: 0.2 - 0.3 mm - robust! – Luminous region: <1 mm – 1/r field: negligible degradation of resolution! – G ~0.08 may be possible…

δE/E < 4 x 10-3 FWHM @ 2480keV

Neutrino 2008 37

Electron drift in xenon gas

• Drift velocity is very slow: ~1 mm/µs
• ββ event is ~100 mm, only ~1 mm of track is

making light - the signal spans up to 100 µs

• The “energy” signal is spread out over the

entire readout plane “B”, ~1000 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

Neutrino 2008 38

By realizing that the key idea is counting electrons, we arrive at a new, attractive detector concept:  Superior S2/S1 discrimination for γ backgrounds  Near-intrinsic energy resolution in HP-Xe EL TPC  Ionization signal alone is sufficient to achieve this  Enriched/depleted xenon: SD/SI (WIMP search)  Enriched/depleted xenon: Signal/background (ββ)  WIMP + ββ search: superior performance!

Neutrino 2008 39

A novel technique emerges

 Electroluminescent readout is essential to achieve near intrinsic energy resolution in xenon  Separated-function readout planes provide tracking and integration of energy signal  All signals are detected as/converted to UV light: S1 , t0, S2 , tracking, & Eββ  Dynamic range OK: keV - MeV energy range

Neutrino 2008 40

Reality checks…

• Need to show that all this can work!

– What is S2/S1 in HPXe from 3 - 60 keV? – δE/E + tracking at MeV energies? – Radiopurity and background rejection?

• Milestone: a reasonably large system

– HPXe with 49 PMTs per readout plane – We (almost) know how to do this now! – “NEXT” & “DUET” efforts underway now

Slide: James White 41

7-PMT, 20 Bar Test Cell @ TAMU

Slide: James White 42

7-PMT,20 bar Test Cell

James White 43

241Am

Gammas

60 keV 30 keV (1st Look – PMT gains not yet calibrated)

Neutrino 2008 44

In conclusion…

WIMPs and 2-beta, no-ν We’ve experiments to do! High High risk risk is the name Of this quite serious game - So why do just one, but not two?

Neutrino 2008 45