Transcript Dark Matter Search with Noble Liquids: the ArDM experiment

Dark Matter Search with Noble Liquids:
the ArDM experiment
Sergio Navas
University of Granada, Spain
Sixth International Heidelberg Conference
on Dark Matter in Astro and Particle Physics
(DARK MATTER 07)
University of Sydney, Australia, 23-28 September 2007
The ArDM collaboration
A. Badertscher, A. Baeztner, R. Chandrasekharan, L. Kaufmann, L. Knecht, M. Laffranchi,
A. Marchionni, G. Natterer, P. Otiougova, A. Rubbia (contact person), J. Ulbricht
ETH Zurich, Switzerland
C. Amsler, V. Boccone, S. Horikawa, C. Regenfus, J. Rochet
Zurich University, Switzerland
A. Bueno, M.C. Carmona-Benitez, J. Lozano, A.J. Melgarejo S. Navas, A.G. Ruiz
University of Granada, Spain
M. Daniel, M. Del Prado, L. Romero
CIEMAT, Spain
P. Mijakowski, P. Przewlocki, E. Rondio
Soltan Institute Warszawa, Poland
H. Chagani, P. Lightfoot, P. Majewski, K. Mavrokoridis, N. Spooner
University of Sheffield, England
http://neutrino.ethz.ch/ArDM
S. Navas (U. Granada), DARK07
2
Argon as target for WIMP detection
NIM A 327 (1993) 205 & NIM A 449 (2000) 147
WARP
NIM A 574 (2007) 83-88
XENON arXiv:0706.0039 [astro-ph]
WIMP
40
18 Argon
High event rate
due to high density
[1.4 g/cc at 87 K
(boiling point at 1atm)],
high atomic number
WIMP
( 10-3)
Tmax 2MAr c2 2
Argon Recoil
(in the range of tens of KeV)
e-


Scintillation from
excited Ar dimers
 40 000  / MeV
e-
e-
Escaping ionization
electrons
 42 000 e- / MeV
Our aim is to detect the ionization charge and
scintillation light independently
S. Navas (U. Granada), DARK07
3
M. Suzuki et al. NIM 192 (1982) 565
Processes induced by charged particles in Argon
 Columnar recombination decreases the
secondary electron yield at the favor of
scintillation photons. It is affected by an
external drift field Edrift
 Different ratio of scintillation to ionization
for faster electron and slow ion tracks
 Observed quenching of triplet (slow)
component in high density ionization core
Ar+
Ar*
Ar*
Ar2*
tS ≈ 6 ns
Ar2+
tT≈ 1.6 ms
UV Light
(two components)
recombination
l ≈ 128 nm
+
Charge
Ar**
S. Navas (U. Granada), DARK07
• Various physical processes leading to scintillation & ionization
• Yields are particle, energy and drift field dependent
• Simulation describes different response to WIMP and MIPs
4
Experimental strategy
 Ton-scale LAr detector providing self-shielding
 Direct detection of ionization charge and primary scintillation light
 LAr volume operated as TPC (3D event imaging)
• charge readout with fine spatial granularity (transverse coordinate)
• longitudinal coordinate from drift time (time difference between primary
scintillation light and charge collection time)
 Efficient rejection of external
 background
• Compton processes are source of low energy deposits within the fiducial volume
• however, often producing multiple scatters in the active volume
 Efficient rejection of neutron background
• irreducible genuine nuclear recoils are produced by fast neutrons elastically
scattering off target nuclei
• however, high probability of multiple scatters within active volume
S. Navas (U. Granada), DARK07
5
Prototype layout
•. Cylindrical volume, drift length ≈ 120 cm
• 850 kg LAr target
8 Polyethylene pillars as mechanical support.
2x LEM for the electron multiplication and readout (Gain ≈ 103 – 104)
Greinacher chain: supplies the right voltages to the field shapers
rings and the cathode up to 500kV  ≈ 4 kV/cm
LAr
WIMP



WLS + Light reflector
Cathode: semi-transparent
in order to let the scintillation
light pass trough …
e-
E-field
The aluminized foils reflect the
scintillation light (>95%)
WLS + Light reflector
120 cm
GAr
The field shapers are needed
to make an homogeneous E
PMTs below the cathode to
detect the scintillation light.
S. Navas (U. Granada), DARK07
6
Detector vessel
Cryogenics
LN2 out
Secondary LAr out
DEWAR
Liquid Argon should be kept
Primary LAr out
free from electronegative impurities
Insulation vacuum
connection top
(O2 contamination < 1 ppb)
Vacuum Insulation
Liquefying system
Purification
cartridge
Insulation vacuum
connection bottom
Cryogenic connections bottom
LN2 in / Secondary Argon in /
Primary Argon in
Secondary LAr
S. Navas (U. Granada), DARK07
Liquid N2 bath
Vacuum superinsulation
External wall
7
DEWAR at CERN
S. Navas (U. Granada), DARK07
8
High Voltage system for drift field generation
•
A cascade of rectifier cells (Greinacher/Cockroft-Walton circuit)
•
The total voltage we aim to reach is Vtot = 500 kV (≈ 4 kV/cm)
•
Tests in liquid nitrogen have been performed
Cathode
Polypropylene capacitors
82 nF
2.5 kV/stage
210 stages
S. Navas (U. Granada), DARK07
9
Layout of the charge readout system
GARFIELD simulations indicate an
expected single electron gain ≈ 104
Stable Measured Gain ≈ 104 measured
GAr
 Distance between stages: 3 mm
 Avalanche spreads into several holes at second stage
Gain 103
LAr
 Higher Gain reached as with one stage, with good stability
 Hole dimension: 500 mm diameter, 800 mm distance.
 Thickness of PCB: 1.6 mm
S. Navas (U. Granada), DARK07
V/d (kV/cm)
10
Segmented LEM
o Final LEM charge readout system will be
segmented
o Orthogonal strips readout
o Number of channels: 1024
o Strip width: 1.5mm
Prototype of a segmented LEM.
 Kapton flex-prints are used for signal transfers to the readout electronics
 The flex-prints, connected on one side to the LEM board, exit the dewar through a slot, sealed
with epoxy-resin, in a vacuum tight feed-through flange
Custom-made front-end charge preamp + shaper G ~15mV/fC
32 channels/cable
to ZIF connectors on the LEM board
S. Navas (U. Granada), DARK07
to front-end preamplifers
11
Layout of Light Readout system and PMT
Photomultiplier tube: potential PMTs
ETL 9357KFL (low background)
R5912-MOD and R5912-02MO from Hamamatsu
Wavelength shifter (WLS):
Tetra-Phenyl-Butadiene (TPB) evaporated on reflector
Reflectivity @430nm ~97%
Shifting eff. 128 nm  430nm >97%
Tetratex
reflecting foil
S. Navas (U. Granada), DARK07
12
GEANT4 simulation detector top view
 On average, 50% of the produced photons hit PMTs
Average incident angle of photons on PMTs: 40º
WLS on wall
WLS on PMT
S. Navas (U. Granada), DARK07
13
Light measurements in Liquid Argon (preliminary)
Event separation in liquid argon
Scintillation light from  in 1200 mbar liquid argon

Por un lado hay más luz de centelleo
DETECTADA de alphas que de betas
100
nVs
 events
L50/Ltot
radioactive source:  (5.4 MeV) +  (Q = 1.163 MeV)
PM Amplitude

210Po
10-4
0
6000
Time (ns)
  events separate well from ,e events
 Fast and slow light components distinguishable
,e events
IS

IT
~ 0.3 for  particles
~ 1.3 for  particles
[Phys.Rev.B27 (1983) 5279]
nVs
Ar2*
1u+ (singlet)
3u+(triplet)
t=6 ns (fast)
t=1.7 ms (slow)
#photoelectrons
Rejection
(@E=0V/cm)
>10
>
>20
> 5x103
>30
>
102
105
14
PMT tests in LAr
PMT after LAr immersion test
 from 241Am source
No  source
S. Navas (U. Granada), DARK07
15
Slow Control Devices
 A series of custom designed Slow Control devices
have been built, tested and installed to monitor temp.,
level, pressure …
PT10K resistors
Temperature
sensor
Levelmeters
10 K at 0ºC
Range: -200 to 400ºC
Capacity levelmeter
Electronic circuits
 0.7 pF/mm
precission of  0.03 mm
S. Navas (U. Granada), DARK07
16
Background rejection
Charge/Light:
Neutrons and WIMPs
interact with the argon
nucleus
interact with shell
electrons
Visible energy
Light shape:
Amplitude
e-/
e/-like
Background rejection tools:
– Different shape of the scintillation light
(ratio fast/slow components)
– Exploit 3D imaging capabilities of the detector
Amplitude
– Different light/charge ratios
Time
WIMP-like
Time
Neutron Background from detector components
Neutron sources:
 Uranium and Thorium contamination (spontaneous fission)
of the detector components and the surronding rock:
 Muon-induced neutrons from surrounding rock, shielding
and detector components
No. of events
• flux about 3.8  10-6 cm-2 s-1 (at 2450 m.w.e.)
• can be shielded, e.g. by a hydrocarbon shield
Geant4 simulation
0
Neutron energy (MeV)
10
Argon recoil energy (keV)
300
Event numbers per year
Component
Container
LEM (std. materials)
LEM (PEEK)
PMTs (std. materials)
PMTs (screened)
n per year
WIMP-like recoils
~ 400
~ 30
~ 10000
~ 900
< 18
<1
~ 12000
~ 1000
~ 600
~ 50
No. of recoils
High energy neutrons penetrate shielding, are thereby moderated and can cause WIMP-like events.
0
Nuclear recoils:
 70% scatter more than once within the
fiducial volume  advantage of large detectors
 10% produce a WIMP-like event
(single scattering, recoil energy  [30,100] keV)
Compared with ~ 3500 WIMP events at  = 10-43 cm-2
Low Background Materials are crucial
S. Navas (U. Granada), DARK07
18
Intrinsic background from Argon 39 isotope
Natural argon from liquefaction of air contains small fractions of 39Ar radioactive isotope
(well known to geophysicists)
•
•
Induced in atmospheric argon by cosmic rays
Concentration in natural Ar: 8.1x10–16
39Ar/Ar
[H.H. Loosli, Earth and Planetary Science Letters, 63 (1983) 51
and “Nachweis von 39Ar in atmosphärischem Argon” PhD thesis
University Bern 1968]
•
T1/2 = 269 years, Q=565 KeV , <E>= 218 keV
•
Integrated rate in 1 ton LAr  1kHz
[WARP Coll.] astro-ph/0603131
Energy(MeV)
To suppress 39Ar fraction we consider using Ar extracted from well gases (extracted from underground
natural gas). On the other hand, this source, evenly distributed in the target, provides precise calibration
and monitoring of the detector response.
S. Navas (U. Granada), DARK07
19
Assembly at CERN
Top flange
Access platform
Detector insertion
Reflector foil
S. Navas (U. Granada), DARK07
20
The ArDM schedule for near future
 Test of detector in vacuum, at CERN:
High voltage system, purity
Currently in preparation
 Test with gaseous argon, at CERN:
PMTs, high voltage system and small
version of LEM plates
Within a month
 Test in liquid argon, at CERN:
Recirculation and purification system
Before end of 2007
 Test underground at shallow depth
2008?
S. Navas (U. Granada), DARK07
21
Conclusions
 Construction and first tests of the ArDM detector are ongoing.
 Three technical key points:
 High drift field
 Charge readout with LEM
 Light readout with PMTs
 After tests at CERN and possibly at shallow depth, the detector
will be moved underground (presumably to the Canfranc underground
laboratory in Spain).
 Depending on the rejection power, the present ArDM detector can
reach sensitivities of the order of 10-8 pb.
 The technique of ArDM is scalable. Larger detectors of 10 tons or
more are a realistic perspective.
S. Navas (U. Granada), DARK07
22
S. Navas (U. Granada), DARK07
23
Estimated event rates on Argon
With true recoil energy threshold
≈ 30 keVr
CREST
Edelweiss I
ZEPLIN III
WARP
CDMS
XENON10
≈100 event/ton/day
≈1 event/ton/day
≈1 event/ton/100 day
Assumptions for simulation:
S. Navas (U. Granada), DARK07
• Cross-section normalized to nucleon
  = 10–42 cm2 =10–6 pb
 MWIMP = 100 GeV
• Halo Model
 WIMP Density = 0.5 GeV/cm3
 vesc = 600 km/s
• Interaction
 Spin independent
 Engel Form factor
24
WIMPs vs.
39Ar
background discrimination
This is MONTE CARLO, thisrelies heavily on
MC, there is no reason to belive this is OK,
this is exactly what the 1 ton test at CERN
should prove.
Full GEANT4 simulation
E = 5 kV/cm
LEM ampli= 103
PMTQE = 10 %
0.7 phe / keV
S. Navas (U. Granada), DARK07
CUTS:
True recoil energy > 30 keV
Q > 2000 electrons
light  phe
phe = 2 : ~91 WIMP evts/day
phe = 4 : ~85 WIMP evts/day
(If Quenching = 0.28)
phe = 2 : ~39 WIMP evts/day
phe = 4 : ~9 WIMP evts/day
25