Ionization Imaging:
a better way to search for
0-v  decay?
David Nygren
LBNL Physics Division
SCIPP - November 14, 2006
1
Two Types of Double Beta Decay
A known background process and
an important calibration tool
T1  1019 yrs.
2
2 
This process not yet observed
particle = antiparticle
1
2
2
 G    m
T1
2
Z
Neutrinoless double
beta decay lifetime
0 
SCIPP - November 14, 2006
Neutrino
effective
mass
2
0-v  Decay
• If 0-v decays occur, then:
– Neutrino mass ≠0
(now we know this!)
– Decay rate measures effective mass mv 
– Neutrinos are Majorana particles
– Lepton number is not conserved
• Because the physics impact is so great,
the experimental result must be robust.
SCIPP - November 14, 2006
3
Uncertainties…
• Hierarchy uncertain
o Determines needed sensitivity
• Matrix element calculations uncertain
o Order of magnitude in rate
• Effective mass uncertain
o Phases enter: mv  = |∑ |Uei|2 imi |
• Direct tests by 3H kinematics uncertain
o For mv  << 1 eV, technically very difficult!
• Best experimental approach: uncertain!
SCIPP - November 14, 2006
4
A “Robust” Experiment:
Only
Only
2-v decays!
0-v decays!
Rate
No backgrounds
above Q-value!
0
Energy
Q-value
The experimental result is a spectrum of all  events,
with very small or negligible backgrounds.
SCIPP - November 14, 2006
5
A robust experiment:
• Has negligible overlap of 0- and 2- events
– excellent energy resolution is essential!
• Selects 0-v  and 2-v  events identically
– does not depend solely on end-point energy!
• Scales to large active mass
– M ~ 1/ mv2  >1000 kg needed?
• Rejects all backgrounds effectively
– no insensitive surfaces!
SCIPP - November 14, 2006
6
Energy Resolution
The Gold Standard:
Energy resolution with Germanium detector:
E/E ~ 1.25 x 10-3 FWHM at 2.6 MeV
Germanium Detectors:
– Excellent electron and hole mobilities
• Complete charge collection
– Small level of recombination
• Charge collection independent of track topology
– Small energy per ion/electron pair
• Fluctuations small
SCIPP - November 14, 2006
7
Energy Resolution
“Extra margin in energy resolution is very
desirable because non-gaussian
characteristics are often present in the
tails of the experimental distribution.”
“To realize an energy resolution near the
limit imposed by physical processes, the
detector and target must be the same.”
SCIPP - November 14, 2006
8
Using Energy to Detect 
1.
2.
3.
4.
5.
Get a large quantity of candidate nuclei
Put them in an electron detector
Shield and purify
Acquire data for a few years (“plug and pray”)
Cut on energy to select out the neutrinoless events
100’s of kg target Condensed matter
strongly preferred
Theory
Spectra from Ludwig DeBraekeleer
Esum of 2 final state electrons
Practice
positive signal claim is
disputed!
Spectra from Klapdor Kleingrothaus et. al.
SCIPP - November 14, 2006
9
 Background rejection is essential - energy resolution may not be enough !
Present Status
• Heidelberg-Moscow (H-M) claim:
mv = 0.44 +0.14-0.20 eV (best value) disputed!
0v1/2 = (8
— 18.3) x 1024 y (95% c.l.)
Scale: ~11 kg of 76Ge, for ~7 years
No other claim for a positive result exists
SCIPP - November 14, 2006
10
Worldwide Activity
•
•
•
•
•
•
•
•
•
•
•
•
•
•
CAMEO.........
CANDLES.....
COBRA.........
CUORE.........
DCBA............
EXO..............
GERDA.........
GSO..............
Majorana.......
MOON...........
Nano-crystals
Super-NEMO
Xe.................
XMASS.........
CdWO4 crystals in liquid scintillator
CaF2 crystals in liquid scintillator
CdTe semiconductors
TeO2 bolometers, Cuoricino now
Nd foils and tracking chambers
Xe TPC; liquid now, maybe gas later
Germanium crystals in LN
Gd2SiO5 crystals in liquid scintillator
Segmented Ge crystals
Mo foils and plastic scintillators
Suspended nanoparticles
Foils with tracking
Xe dissolved in liquid scintillator
Liquid xenon
SCIPP - November 14, 2006
11
Present Perspective…
• Cuoricino (130Te): background-limited
– E/E only  Cuore needs factor of ~20
• Majorana (76Ge): pre-construction stage
– E/E + multi-site rejection (x10), but
– a factor of several 100 needed beyond HM
• Common to both:
– Multi-detector coincidences can reject many
backgrounds, but:
– Large rejection factor needed for success
SCIPP - November 14, 2006
12
Xenon
• Qv = 2.49 MeV (136Xe)
• Previous experiments inconclusive
– HPXe TPC in Gotthard tunnel (5 bar, no start time)
– Russian experiments with various MWPC
• EXO
–
–
–
–
EXO-200 underway with LXe @ WIPP
No laser tagging of barium daughter: R&D stage
Strong anti-correlation of ionization/scintillation
Results eagerly awaited
SCIPP - November 14, 2006
13
NUSAG Recommendations:
• “…support research in two or more 0-v 
experiments to explore the region of
degenerate neutrino masses (mv  > 100
meV)…”
• “The knowledge gained and the technology
developed in the first phase should then be
used in a second phase to extend exploration
into the inverted hierarchy region of (mv  > 10
- 20 meV) with a single experiment.”
But: no explicit encouragement for new ideas!
SCIPP - November 14, 2006
14
Experimental Approach
“We believe that an
Imaging Ionization Chamber
is most likely to meet all criteria imposed
for a robust experiment.”
An Imaging Ionization Chamber (IIC) is
a TPC without gain at the readout plane
SCIPP - November 14, 2006
15
Imaging Ionization Chamber
-HV plane
Pixel Readout plane
Pixel Readout plane
~99%Xe +
~1% CH4
@ 20 bars
.
electrons
ions
SCIPP - November 14, 2006
16
Proportional Gain:
good results for low-energy x-rays
MWPC, GEM,
micromegas
all work well...
but:
why are there
so many events
below the peak?
SCIPP - November 14, 2006
17
Proportional Gain:
poor resolution for MeV energies
Typical E/E: 4 - 6.6% FWHM @ 2.5 MeV
– Gain variations?
• gas density, composition
• mechanics, calibration maps
– Extended tracks?
• ballistic deficit in signal processing
• impact of space charge on gain
– Intrinsic physical phenomena?
• sensitivity to dE/dx density variations
• large scintillation/ionization fluctuations
SCIPP - November 14, 2006
18
Ionization Chamber Mode
• Reason # 1:
– “Best energy resolution can only be
obtained through direct charge integration”
• There is a lot to learn here...
• Reason # 2:
– “Gain may be needed at HV plane…”
• This is a new, but very speculative aspect
SCIPP - November 14, 2006
19
Imaging Ionization Chamber
is filled with 136Xe gas
• Xe is relatively safe and easy to enrich
• EXO has 200 kg highly enriched in 136Xe
• high density is desirable to contain event
• But there is an upper limit:  < 0.55 g/cm3
• working density:  ~ 0.1 g/cm3 (LXe = 2.95 g/cm3)
• ~20 bars; critical point P = 58 bars, T = 290° K
• 1000 kg in ~10 m3 :  = 200 cm, L =300 cm
• dn/dx ~ 1 fC/cm = 6000 electron/ion pairs/cm
SCIPP - November 14, 2006
20
UV Scintillation in HPXe
• N ~ Nelectron/ion pairs ~ 118,000 at Q = 2.49 MeV
– can use for event start time
– ratio depends on E field
– spectrum peaks at ~170 nm ~ 7 eV
• However, small amount of “CH4” is necessary
– to cool electron drift and keep diffusion low for tracking
• Does methane absorb 170 nm light? - No
• Does methane quench scintillation? - don’t know, but
– 2% added to LXe without loss!
– Does HPXe behave similarly? maybe...
SCIPP - November 14, 2006
21
Imaging Ionization Chamber…
• “provides adequate S/N for good tracking”
– detailed imaging of event topologies
• “provides fully closed, active fiducial surface”
– ex post facto variable definition with mm resolution
• “provides energy resolution of 1% FWHM”
– avoids scintillation/ionization anti-correlation
• “may permit detection of birth of Ba daughter”
– automatic process tags both space and time
SCIPP - November 14, 2006
22
Event Characteristics in IIC
– High density of xenon constrains  event:
• Total track length ~10 - 20 cm max
– Multiple scattering will be prominent in xenon
• Unclear if B-field would help identification
– True  events will have two “blobby” ends
• Shown to reject background by ~30 in Gotthard TPC
– Bremsstrahlung and fluorescence ’s
• Distinct satellite “blobs” may be visible
SCIPP - November 14, 2006
23
Imaging Ionization Chamber
has a fully “decorated” pixel readout plane
– no grids or wires: eliminates microphonics
– pixel size is 5 mm x 5 mm (4 x 104 /m2)
• ~ 40 - 80 contiguous “hit” pixels for E = Q
• dn/dx = ~3000 electrons/(5mm)
– ultra-low noise readout electronics - BNL ASIC
• <n> = ~30 e– rms for 4 s shaping time, with pixel!
• Other noise terms must be included  <n>= 60 e– rms?
– “waveform capture” essential for extended tracks
SCIPP - November 14, 2006
24
Pixel geometry
A low capacitance solution: a 7-pixel hexagonal sub-module:
Or, a 16 channel 4x4
rectangular array…
SCIPP - November 14, 2006
25
Imaging Ionization Chamber
collects electrons on pixel readout plane
–
–
–
–
–
–
–
all energy information is derived from q =  Idt
current is very small until electrons approach pixel
pixels with no net charge have bipolar currents
drift velocity is small, 0.1< Vd <0.5 cm/s
diffusion after 1.5 m drift is ~ 2 mm rms
event is reconstructed from contiguous hit pixels
noise adds only from hit pixels + some neighbors
SCIPP - November 14, 2006
26
Energy Resolution…
Q-value of 136Xe = 2.48 MeV
W = E per ion/electron pair = 21 eV
N = number of ion pairs = Q/W
N  2.49 x 106 eV/21 eV = 118,350
N2 = FN (0.05 < F < 0.17)
F = 0.17 for pure noble gases (theory)

N = (FN)1/2 ~ 140 electrons rms
SCIPP - November 14, 2006
28
Energy Resolution…
• If ionization were the only issue:
E/E = 2.9 x 10-3 FWHM
• Other contributions:
– electronic noise from N = 49 pixels in event
• N1/2 x <n> if noise is gaussian ~ 7 x 60 = 430 e–
– ballistic deficit in signal processing
– “locked” charge caused by slow-moving ions
E/E < 10.0 x 10-3 FWHM
SCIPP - November 14, 2006
29
IIC and Imaging Power
• The 3-D imaging of the IIC provides:
– Topology reconstruction
– Energy resolution independent of scale
– Active and continuous fiducial surfaces
– Variable fiducial surfaces ex post facto
Rejection of ionizing backgrounds from
surfaces can be essentially 100%
SCIPP - November 14, 2006
31
Perspective
• The basic IIC concept offers:
– Stable operation: ionization mode
– Excellent energy resolution: ~1% FWHM
– Good scaling: active mass ~1000 kg
– No dead surfaces: 3-D event placement
– Active, adjustable fiducial boundaries
– Topological rejection of backgrounds
– Possibility to evolve further…
SCIPP - November 14, 2006
32
Barium Daughter Atom
– In a volume of ~1027 xenon atoms, a 
event creates one barium atomic ion.
– The Ba ion drifts out to the HV plane,
and in ~ 1 second, the ion will be lost!
– Is this a hopeless situation?
SCIPP - November 14, 2006
33
Barium Daughter Atom
– In xenon/CH4, the Ba++ ion will survive:
• IP(Xe) = 12.13 eV, IP(CH4) = 13.0 eV
• First IP(Ba+) = 5.212 eV
• Second IP(Ba++) = 10.004 eV
– if impurities exist with IP less than 10 eV:
• Ba++ becomes Ba+ through charge exchange
SCIPP - November 14, 2006
34
Ion Mobilities
Is there a straightforward way to detect
and identify the barium daughter?
• Ba and Xe ion masses are ~identical…
• Ba+ and Xe+ ion charges are identical…
• Ion mobilities should be the ~same,
Right??
SCIPP - November 14, 2006
35
Ion Mobilities…
• But: Ion mobilities are quite different!
– The cause is resonant charge exchange
– RCE is macroscopic quantum mechanics
•
•
•
•
occurs only for ions in their parent gases
no energy barrier exists for Xe+ in xenon
energy barrier exists for Ba ions in xenon
resonant charge exchange is a long-range
process; glancing collisions = back-scatter
– RCE increases viscosity of ions
SCIPP - November 14, 2006
36
Ion Mobilities in Xenon
– Mobility differences have been measured at low
pressures, where clustering effects are small:
• (Xe+) = 0.6 cm2/V-sec
• (Cs+) = 0.88 cm2/V-sec (Cs is between Ba and Xe)
– So, the barium ion should move faster by ~50%!
(maybe even faster if Ba++ is stable)
RCE can provide a way to detect Ba daughter!
SCIPP - November 14, 2006
37
Ion mobility in dense gases?
• Ion mobility data at high pressure does
not apparently exist in the literature.
– Clustering may be prominent at 20 bars.
– Clustering phenomena are complex, and
may introduce very different behavior
• Not clear whether this will help or hurt!
• Low pressure measurements not adaptable to
high pressures like 20 bars - need new method
SCIPP - November 14, 2006
38
Ba Daughter Detection
• If we assume that barium ion mobility is
not identical to xenon ion mobility, then:
• A barium ion will arrive at the HV plane at a
different time than the Xe+ ion track image.
• If event time origin and mobilities of the barium
and xenon ions are known, an arrival time for
the barium daughter at HV plane is predicted.
• The unique t between Xe+ and Ba+ ions is a
robust signature for a true  event.
SCIPP - November 14, 2006
39
Arrival Time Separation
• Assume low-pressure data…
– Assume drift distance: L= T = 250 mm
– Thermal transport diffusion:  ~ .25 mm
/L = 0.25/250 = 1/1000
/(L) = 21/2/(Ba - Xe)T ~ 1/235
– Arrival times are very precisely determined
SCIPP - November 14, 2006
40
Detection of Ion Arrival
• Detection of ion arrival may be possible:
– Ions drift at thermal energies to HV plane…
Then:
– Ions are attracted to “high field pore (HFP)”
•
•
•
•
•
Very high electric field inside HFP
Ions can enter, but electrons are blocked
High energy tail of M-B distribution relevant
Ba++ may be critical for desired outcome
Hoped-for outcome: ≥1 electron appears
SCIPP - November 14, 2006
41
Blind GEM or “Microwell”
HV plane
Drift region:
Low E-field
Resistive
back side
blocked to
electrons
ions
Very High E-field inside pore;
low work-function surface?
SCIPP - November 14, 2006
42
The barium daughter “Echo”
– If a single electron appears, high E-field in
blind GEM causes electron avalanche.
– Electron avalanche will saturate, producing
a large pulse of electrons, more than 103.
– Electron pulse returns to pixel plane, at a
spot on the projected event track.
– This spot on the projected track is very
close to origin of the barium daughter.
SCIPP - November 14, 2006
43
“Birth Detection”
Because the echo tagging is so precise
in space and time
(if it can be done at all)
I refer to this process as
“Birth detection”
SCIPP - November 14, 2006
44
The Return Image Echo
• The Xe+ ions will also enter HFP, producing
an “echo” of the track.
• The track echo time will be distinct from the
pulse due to the barium daughter.
• Maybe: transfer charge to C2H4: IP =11.6 eV
– Complex organic molecule may be much less
likely to liberate electrons than Ba++ or Ba+
– Will a mobility difference still exist?
SCIPP - November 14, 2006
45
Event Quality
• strong primary UV scintillation gives to
– Enough intensity, even at low visible energy
• electron track image provides topology
– Energy resolution limited by electronic noise
• ion track echo also places event in space
– Don’t need all 115,000 echoes from HV plane
• barium daughter echo is elegant tag method
– Can some detection scheme be found?
• an over-constrained reconstruction possible.
SCIPP - November 14, 2006
46
Imaging Ionization Chamber
-HV plane
Pixel Readout plane
Pixel Readout plane
.
electrons
ions
SCIPP - November 14, 2006
47
What to do?
• An R&D (and library) effort is needed to:
– Develop good simulation tools
– Optimize S/N with real electronics
– Measure E/E in HPXe IIC with  rays
– Investigate benefits of organic additives
– Determine ion mobilities in HPXe.
– Explore ion-induced avalanche processes
SCIPP - November 14, 2006
48
What to do?
• An R&D (and library) effort is needed to:
– Develop good simulation tools
– Optimize S/N with real electronics
– Measure E/E in HPXe IIC with  rays
– Investigate benefits of organic additives
– Determine ion mobilities in HPXe.
– Explore ion-induced avalanche processes
• A proposal has been rejected by DOE!
SCIPP - November 14, 2006
49
SCIPP - November 14, 2006
50
SCIPP - November 14, 2006
51
SCIPP - November 14, 2006
52
Summary
• A novel concept for a robust 0- 
decay search has been developed:
– E/E ~ 1% FWHM
– Detailed & constrained 3-D event topology
– Active, variable fiducial boundaries
– Identification of Ba daughter possible, in
principle, by exploitation of macroscopic
quantum mechanical phenomenon, RCE
SCIPP - November 14, 2006
53
Acknowledgments
•
•
•
•
•
Azriel Goldschmidt - LBNL
Adam Bernstein - LLNL
Mike Heffner - LLNL
Jacques Millaud - LLNL/LBNL
Leslie Rosenberg - UW
SCIPP - November 14, 2006
54