Transcript CoGeNT data

Searching for Dark Matter with the
CoGeNT and C4 Detectors
Mark Kos, PNNL
1
PNNL-SA-92945
Overview
Status of CoGeNT
Latest results
Current understanding of backgrounds
Steps to help reduce/eliminate the major backgrounds we
see in CoGeNT
Simulation of the C4 backgrounds
Expected WIMP (Weakly Interacting Massive Particle)
sensitivity of C4
Implications for future low-mass dark matter searches
2
CoGeNT results and low-mass WIMPs
Published CoGeNT
analysis shows an excess
of events at low energies
that is inconsistent with
known backgrounds, but
hint at low mass WIMPs
(Weakly Interacting
Massive Particles)
Also, a hint of annual
modulation consistent
with WIMP dark matter
The CoGeNT results
have sparked an interest
in low-mass WIMPs
Need multiple detectors
with lower backgrounds
and lower thresholds to
test the CoGeNT results:
C-4
3
Phys. Rev. Lett. 107
(2011) 141301
arXiv:1208.5737
CoGeNT shield design
CoGeNT: 1 Ge crystal
(440 g) at the Soudan
mine (data taking since
Dec 2009)
4
The background picture
68Ge
Background
sum
arXiv:1208.5737
65Zn
Resistor
gammas ~324
events, ~16%
of data
L-shell
contribution
68Ga
49V
54Mn
51Cr
55Fe
73,74As
Muon-induced
neutrons 339
events, 16% of data
Tritium b-decay
150 events, 7%
of data
Cavern neutrons (from radioactivity) 54
events, 3% of data
5
Other sources of background simulated:
U and Th chain backgrounds in surrounding material (copper)
Muon-induced neutrons from the cavern
U and Th chain backgrounds in lead shielding
Spontaneous fission neutrons from shielding material
(a,n) neutrons from shielding material
These
backgrounds
are tiny
Muon-induced neutrons (largest background)
1 cm panels do not allow
muon-gamma separation
Veto operated at single photoelectron sensitivity
Generate ~12% dead time from
spurious germanium detectorveto coincidences.
True coincidences are
however observable and rate
is in good agreement with
Monte-carlo (next slide)
6
Muon-induced neutron simulation
m-
Two independent MC simulations used to
assess neutron contributions
muon induced neutron
natural radioactivity in cavern
#1: GEANT
Soudan muon flux, E, angular distribution
to generate (m,n) in full shield.
Includes e- and g (8% of neutron contribution)
#2 MCNP-Polimi:
Neutron generation in lead shielding
(largest contributor)
Reasonable agreement between simulations
(they use different inputs)
339 +/- 68 events (GEANT)
GEANT
CoGeNT data
MCNP-Polimi
Mostly neutrons,
~8% e- and g’s
(simulation)
7
Less than 16% neutron
fraction in CoGeNT data
after L-shell subtractions
Backgrounds from the front end electronics
(2nd largest background)
RESISTORS ARE HOT!
ILIAS database
SNOLAB
8
Description
U-238
(Bq/kg)
Th-232
(Bq/kg)
K-40
(Bq/kg)
Events in
CoGeNT
Carbon film
resistor
4.3
12.7
21.9
972 +/- 120
Metal film
resistor 1
4.3
0.5
37.5
324 +/- 164
Metal film
resistor 2
5.1
16.1
24.7
1208 +/- 160
Ceramic core
resistor
5.9
4.6
34.3
644 +/- 131
Metal on
ceramic
resistor
28
40.7
25.7
4509 +/- 352
Ceramic
15.5
0.2
13.8
993 +/- 200
Most beta-spectra and gammas are a flat
background in the CoGeNT analysis region
This is expected from
Compton scattering of
high energy photons at
these low energies
A background that can
be reduced by having
tightly packed detectors
and rejecting multiples
9
Without an assay we cannot be sure the
flat background is from the resistors,
but typical resistor backgrounds can
plausibly explain most of the CoGeNT
flat background
Tritium production in germanium (3rd largest
background)
Years of surface exposure
Cosmogenic production of tritium in Ge
while detector at surface
Tritium b-decay endpoint at 18.6 keV
Half-life of 12.33 yrs
Tritium production rate:
27.7 /kg-day
Astroparticle phys, 31, 417 (2009)
Based on IGEX data
Phys Lett, B432, 8 (2002)
Assuming a surface exposure
of CoGeNT detector of 2 yrs:
150 events in 0.5 – 3.0 keVee
(Geant4 simulation of 3H in CoGeNT)
Tritium (simulation)
10
Tritium decays underground
1
299
2
583
3
850
4
1103
5
1342
6
1568
7
1782
8
1983
9
10
CoGeNT Data
2174
2355
Neutrons from radioactivity in the cavern:
(a,n) + fission
Use Mei-Hime neutron flux:
3.78 X 10-6 cm-2 s-1
(Phys Rev D 73 , 053004 (2006))
Use Monte-carlo neutron
energy spectrum from Gran
Sasso (worst case)
Simulated background for
CoGeNT:
54 events in the dataset
Deep underground this background
is higher than muon-induced
neutrons
Something that experiments
pushing the zero-background limit
need to address
11
Radioactivity in the CoGeNT shield
SNOLAB assay of similar materials as used in CoGeNT
Material
12
238U
(mBq/kg)
232Th
(mBq/kg)
Lead sample1
0.41 +/- 0.17
0.08+/-0.08
Plastic lumber
121 +/- 4
68 +/- 3
Plastic lumber
(recycled)
115 +/- 5
80 +/- 4
Plastic lumber
McMaster-Carr
15 +/- 1
1.3 +/- 0.8
Aluminum plate
7.1 +/- 2.4
986 +/- 12
Aluminum
framing pieces
42 +/- 8
1348 +/- 50
210Po
(Bq/kg)
93 +/- 19
210Pb
 210Bi  210Po  206Pb
Ultra-low background lead
around CoGeNT: 0.02 Bq/kg 210Pb
Source of (a,n)
neutrons
List of all backgrounds (we know about)
13
Source
Events in CoGeNT dataset
(0.5 – 3 keVee)
Resistor backgrounds
~324
Muon induced events in
shielding
339 +/- 68
Tritium b-decay
<150
Cavern neutrons from
radioactivity
<54
U and Th backgrounds in
copper
<9
External cavern neutrons
(muon-induced)
<1.4
Old lead (210Pb + daughters)
<0.6
Spontaneous fission neutrons
in lead
<0.5
SF neutrons in HDPE
<0.2
HDPE (a,n)
<0.03
8B
<0.014
solar neutrinos
Extensive
simulations done
at PNNL
Backgrounds that Modulate: Radon
Radon levels modulate
underground –
Measured
Modulation out of
phase!
Inner shield is
inside a sealed
nitrogen purged box
So far it doesn’t
look like radon
CoGeNT data: Dec 3 2009 - March 6 2011
MINOS data: Averaged 2007-2011
14
Backgrounds that Modulate: Muons
MINOS muon flux modulation
measured in Soudan
Approximately +/-1.5%
Peaks three months after best fit
to present CoGeNT data
A 1.5% modulation of the
estimated 339 +/- 68 muoninduced events in shielding
predicts a modulation of 5 events
in the 0.5-3 keVee energy range
The CoGeNT data set contains
2124 events in the 0.5-3 keVee
energy range. A 5 event
modulation of muon induced
events could only produce a
0.2% modulation effect in the
CoGeNT data set.
15
Courtesy Alec T. Habig
Surface events and slow pulses
Surface events have degraded energy and pile
up in the lowest energy bins (like WIMPs)
Surface events (background dominated) on
average have slower pulses than bulk events
Rejection between bulk (fast pulses) and surface
(slow pulses) gets worse at lower energies
We can estimate the contribution of slow pulses
in the data by fitting for the slow and fast pulse
distributions
Still looks like there is an excess of events above
the expected background
16
Juan Collar (UC)
The next generation of
CoGeNT, CoGeNT-4 (C4)
Four ~1 kg germanium detectors
(unfortunately 4 detectors won’t be funded)
2 inch thick veto panels
Soudan Underground Lab
New DAQ with full energy range
17
Reducing the Background Rate for C4
2 inch veto panels make the muoninduced neutron background negligible
Thicker water shielding reduces the cavern
neutron rate + reduces (a,n) from shielding
Electroformed copper used to ease
manufacture + has 10 X less background
than OFHC
Redesigned frontend will significantly
reduce resistor background
18
Expected Background Numbers
19
Background
CoGeNT
(cnts/kg/day)
C-4
(cnts/kg/day)
Resistors
6.2
<1.4
Cu cryostat
5.0 X 10-2
7.0 X 10-3
Muon-induced
neutrons in Pb shield
1.9
2.6 X 10-4
Cavern neutrons
3.0 X 10-1
1.3 X 10-3
(a,n) in shielding
2.2 X 10-4
1.3 X 10-4
Total
~10
~2
The background picture for C4
For C4 tritium
may be the
dominant
background—
but can be
reduced by
minimizing
surface
exposure of
crystals
arXiv:1210.6282
20
What about a single 1 kg crystal?
It we can use the same
shield and maintain the
predicted background
rate things don’t look so
bad:
C4 1 kg
C4 4 kg
21
Implications for a low-mass dark matter
search with C4
WIMP sensitivity prediction based on
likelihood fit to background + WIMP signal
Using conservative background
assumptions of some resistor background
remaining and 2 years of surface
exposure (tritium)
22
C4 WIMP sensitivity will be very
competitive in the low-mass region and
complement other experiments in
excluding WIMP parameter space
Even a modest lowering of the energy
threshold can give a large increase in
sensitivity at low masses
Future of low-mass dark matter searches
C4 will compliment other
low-mass dark matter
experiments such as DAMIC,
CDMSLite, MAJORANA in
excluding parameter space
at low masses
For all these experiments it is
crucial to have a low
threshold and minimize
backgrounds
23
Dark Matter analysis with C4
Use all possible information
to get the most out of the
data: PDFs for signal in
energy and time dependence,
PDFs for backgrounds in
energy and time, constrain
backgrounds with
measurements outside the
signal region, etc.
24
Summary
We have done an extensive simulation of the radioactive
and cosmogenic backgrounds in the CoGeNT detector
arXiv:1208.5737
So far no explanation for excess at low energies and no
candidate for the time dependence of the data
C4 will continue with this technology but increase target
mass and reduce backgrounds
The next generation, C4, will address many of the current
concerns…2” thick veto panels, improved low-noise
design (lower energy threshold), lower background
cryostat
C4 will be able to push the limit of sensitivity in the lowmass WIMP parameter space arXiv:1210.6282
25
Modest neutron rejection with multiple
scattering
With 4 detectors we can
remove ~40% of neutron
energy depositions (multiple
scattering)
Neutron deposited energy
distribution before coincidence cut
After coincidence cut
26