Giving up S2/S1 discrimination in exchange for a lower energy threshold
Peter Sorensen
on behalf of the XENON10 Collaboration
Identification of Dark Matter 2010, Montpellier
Talk Overview
•the XENON10 direct detection experiment
•limitations of S1 (scintillation signal) threshold
•new analysis of S2 (electron signal) energy scale
•using S2 width to obtain approximate z coordinate
•S2-only (no discrimination) dark matter limits
Peter Sorensen, LLNL
1
IDM, Montpellier FR
26 July 2010
arXiv:1001.2834v1 [astro-ph.IM]
XENON10 detector
14 kg Xe
S2: secondary scintillation
S1: primary scintillation
S2 hit pattern gives (x,y)
tS2 - tS1 gives z
Peter Sorensen, LLNL
2
IDM, Montpellier FR
26 July 2010
Discrimination (old news)
neutrons
γ,β
γ,β
neutrons
Sorensen PhD Thesis (2008)
Peter Sorensen, LLNL
3
IDM, Montpellier FR
26 July 2010
a XENON10 event in slow-motion
Peter Sorensen, LLNL
4
IDM, Montpellier FR
26 July 2010
S1 determines threshold
PRD 80 115005 (2009)
“...incorporating the S2 threshold,
resulting in an energy-dependent
acceptance at low energies ... is
complicated.”
-- M Kuhlen, N Weiner et al. JCAP02 (2010) 030
TRUE... but for a clear
account see arxiv:
1007.3549
reminder: keVr = phe in this case, i.e. assumption
that Leff=0.19
Peter Sorensen, LLNL
5
IDM, Montpellier FR
26 July 2010
Nuclear recoils from AmBe neutrons
log10(S2 / S1)
5.4 kg fiducial target
3.3
3.1
2.9
2.7
2.5
2.3
2.1
1.9
1.7
1.5
1.3
1.1
0.9
0.7
0.5
0
typical events at threshold:
~ 30 S2 electrons / ~2 S1 photoelectrons
10
20
30
40
keVr
50
60
70
reminder: keVr = phe in this case, i.e. assumption that Leff=0.19
Peter Sorensen, LLNL
6
IDM, Montpellier FR
26 July 2010
S1 detection efficiency in XENON10
black (circles) curve from:
Nucl. Instr. Meth. A 601 339 (2009)
Compton
background (~flat)
/
&&!
•roughly 10% total detection efficiency for
scintillation photon => PMT photo-electron
•require n-fold PMT coincidence :
•XENON10: n>1
•ZEPLIN III: n>2
&!!
*!
%!
.295-6:01,
•we measure 7 eV photons (175 nm)
)!
• efficiency predicted by Poisson stats. in PMT
photo-electrons (simulation)
$!
(!
#!
'!
5!"
5!'
5!$
"!
&!
!/
!
"
#
$ % &! &" &# &$ &% "! "" "# "$
+,-,.-,+/012-2!,3,.-4256/////701,8
S1 detection limits threshold...
what about S2?
Peter Sorensen, LLNL
7
IDM, Montpellier FR
26 July 2010
Single electron pulses ~25 phe
,-.0/12,3.
'!
%!
1 e-
#!
"$+,-.
!
!
"!
#!
$!
%!)!+,-.
%!
&!
'!
(!
56627+8"
)!
µ/
*! "!! ""! "#! "$! "%! "&!
56627+8#
56627+/9:;3.!.3.<=>6:+8#
"4&
,-.0/12,3.
#&+,-.
#
"!
"
!4&
&
!
!
"4&
"
!4&
")4*
"*
(*
!
)!
)"
µ/
µ/
)#
)$
"!$
"!%
µ/
Peter Sorensen, LLNL
8
IDM, Montpellier FR
100
Figure 4.6: The summed event record for a typical ∼ 4 keVee background γ scatter, showing a 61 µs drift time between S1 and S2, and
also an isolated 25 phe single-electron S2 pulse at 103 µs. The primary S2 pulse (at 80 µs) triggered the DAQ in this event, however the S2
trigger was sensitive to single-electron S2 pulses as small as ∼ 10 phe (see Fig. ??). The PMT hit-patterns are indicated for each pulse; the
lower subplots show a zoom (and PMT hits on individual channels) on each pulse.
26 July 2010
Recent work on S2 calibration
S2 trigger efficiency
3
10
keVr:
1
2
4
8
16
32
64 256
from XENON10 neutron calibration (AmBe).
The calibration, analysis, etc already described
in (with focus on S1 signal):
counts / electron
Nucl. Instr. Meth. A 601 339 (2009)
2
10
require S1
S2 (electron) response to nuclear recoils,
rather than S1 (scintillation) response
1
10
S2 energy resolution assumed
Poisson in number of electrons
fit range
0
10
5
10
20
50
100
200
electrons extracted from liquid xenon
Peter Sorensen, LLNL
500
9
IDM, Montpellier FR
26 July 2010
Result of monte carlo best fit
E = 730 V/cm
PRC 81 025808 (2010)
energy scaling obtained from scattering angle
Nucl. Instr. Meth. A 601 339 (2009)
This work
energy scaling obtained from S1, via Leff
energy scaling obtained from S2
systematic uncertainty
The Qy curve allows us to assign a keVr value to every event, based on its S2
NOTE: conservative 1σ stat. uncertainties given by √Ne
Peter Sorensen, LLNL
10
IDM, Montpellier FR
26 July 2010
accidentals
0
10
trigger
threshold
sensitivity
maintained w/o
S1 requirement
~8 keVee
single-scatter Compton background in XENON10
ï1
10
same data scaled to keVr
ï2
green/red show +/1sigma uncertainty in
energy calibration
10
2
5
10
20
electrons
same data as previously reported in
PRD 80 115005 (2009)
PRL 100 021303 (2008)
except no S2/S1 discrimination
(hence Compton-dominated)
50
100
200
1
counts / keVr / kg / day
1
10
0
10
ï1
10
0.5
Peter Sorensen, LLNL
11
1
2
keVr
5
10
[from Qy]
IDM, Montpellier FR
20
26 July 2010
50
8
8
7
7
6
6
5
5
phe / sample
phe / sample
Typical nuclear recoil events at threshold
4
3
2
4
3
2
1
1
0
0
224 phe
7000
7500
8000
8500
ADC samples [105 MHz]
209 phe
9000
7000
7500
8000
8500
ADC samples [105 MHz]
These are 9 electron (about 1 keVr) S2 single scatter nuclear recoils. Clean events, but no S1 found.
Note width variation (from electron diffusion)
Peter Sorensen, LLNL
12
IDM, Montpellier FR
26 July 2010
9000
z fiducial cut from S2 width
neutron calibration data
S2 width cut selects events
in middle/bottom of target
(z cut with blurry edges)
7<S2<100 electrons
R<3 cm (~1.2 kg)
S2 width cut 0.240 < σS2 < 0.295 μs
(34% acceptance for NR)
3
0.7
10
events with S1=0 appear at z=15 cm
0.6
mS2 [µs]
0.5
2
10
0.4
0.3
1
10
0.2
0.1
(includes events with S1=0)
0
0
10
1
3
Peter Sorensen, LLNL
5
7
9
z [cm]
11
13
15
13
0
0.1
0.2
0.3
0.4
mS2 [µs]
IDM, Montpellier FR
0.5
0.6
26 July 2010
0.7
z fiducial cut from S2 width
12.5 live days of “WS2” dark matter search data
(Aug/Sept 2006, not blind, not published)
S2 width cut selects events
in middle/bottom of target
(z cut with blurry edges)
7<S2<100 electrons
R<3 cm (~1.2 kg)
S2 width cut 0.240 < σS2 < 0.295 μs
(34% acceptance for NR)
3
0.7
10
events with S1=0 appear at z=15 cm
0.6
events in gaseous xenon
mS2 [µs]
0.5
2
10
0.4
0.3
1
10
0.2
0.1
0
0
10
1
3
Peter Sorensen, LLNL
5
7
9
z [cm]
11
13
15
14
0
0.1
0.2
0.3
0.4
mS2 [µs]
IDM, Montpellier FR
0.5
0.6
26 July 2010
0.7
WS2 data
no explicit z cut because not all events have S1
R<8 cm
R<3 cm
R<3 cm + S2 width cut
2
3
10
counts / keVr / kg / day
1
10
0
10
ï1
10
9 e-
ï2
10
18 e-
29 e-
<1.7> counts/keVr after S2 width cut
2
10
1
10
0
10
ï1
10
ï2
5
10
20
S2 electrons
50
10
100
0.5
1
keVr
2
[from Qy]
5
10
XENON10 R-z plane
full active target
z: 15 cm
R<10 cm
Qy analysis
z: 15 cm
R<8 cm
height
counts / electron / kg / day
10
liquid
xenon
target
diameter
Peter Sorensen, LLNL
15
usual fiducial target
z: 9.4 cm
R<8 cm
this fiducial target
z: 15 cm
R<3 cm
(NOTE effective z reduced by S2 width cut)
IDM, Montpellier FR
26 July 2010
20
(preliminary) dark matter exclusion limits
Notice: this S2-only exclusion limit curve is preliminary, and has not been fully reviewed by
the XENON10 collaboration. Pending review it is subject to change.
arXiv:1001.2834
arXiv:1007.1005
•Max Gap 90% C.L. upper limit
between 1.6 keVr and 3.8 keVr
•12.5 live days
•1.2 kg target
•conservative -1σ Qy energy
calibration
•no account of resolution (this
would improve limits)
XENON10 S2-only
(no discrimination)
Peter Sorensen, LLNL
16
IDM, Montpellier FR
26 July 2010
Summary
•very quiet in the center of liquid xenon detectors (self shielding)
•S2 threshold is LOW, ~1 keVr
•need confirmation of S2 calibration at low energies
•more stringent S2-only limits from XENON100 ?
Peter Sorensen, LLNL
17
IDM, Montpellier FR
26 July 2010