Transcript First Results from the South Pole Acoustic Test Setup

The South Pole
Acoustic Test Setup
(SPATS)
Justin Vandenbroucke
for the IceCube Acoustic Neutrino Detection group
3rd International Workshop on Acoustic and Radio EeV Neutrino detection Activities
Rome
June 26, 2008
The IceCube Acoustic Neutrino Detection group
UC Berkeley
Buford Price
Justin Vandenbroucke
University of Stockholm
Christian Bohm
University of Gent
Yasser Abdou
Freija Descamps
University of Uppsala
Allan Hallgren
DESY Zeuthen
Martin Bothe
Rolf Nahnhauer
Delia Tosi
Lausanne
Mathieu Ribordy
University of Aachen
Karim Laihem
Matthias Schunk
Christopher Wiebusch
ARENA
University of Wuppertal
Klaus Helbing
Timo Karg
Benjamin Semburg
Justin Vandenbroucke
June 26, 2008
Outline
1)
2)
3)
4)
5)
ARENA
Motivation
Design
Deployment and performance
Results
Conclusion and what’s next
Justin Vandenbroucke
June 26, 2008
Part 1: SPATS motivation
ARENA
Justin Vandenbroucke
June 26, 2008
Observation of UHECR spectrum steepening at ~1019.7
simplest interpretation is GZK cutoff
plot
HiRes
5 sigma steepening
Auger
6 sigma steepening
Conclusion: GZK neutrinos exist!
ARENA
Justin Vandenbroucke
June 26, 2008
Particle physics with GZK :
Measure N @ ECM ~100 TeV
[A. Connolly]
Goal: Detect ~100 events
• sky maps
• spectra
100 events: measure Lint = 400 km ±
33%
ARENA
Justin
Vandenbroucke
tests e.g. models of extra
dimensions
June 26, 2008
GZK  event rates (Engel, Seckel, Stanev model)
zmax = 8, n = 3
We use  = 0.7
 = 0
Log(Ethr/eV)
~Veff for 1 evt/yr (km3)
16
4
17
5
18
9
19
50
Need ~100 km3 effective volume to get ~10 evts/yr
ARENA
Justin Vandenbroucke
June 26, 2008
Ice may be best medium for acoustic (hybrid) 
detection
acoustic signal strength ~ Gruneisen parameter 
ocean
salt
South Pole ice
c (m/s)
1530
4560
3920
 (/K)
25.5e-5
11.6e-5
12.5e-5
CP (J/kg/K)
3900
839
1720
fmax (kHz)
7.7
42
20
 = c2/CP
0.153
2.87
1.12
refraction
moderate
small?
small?
attenuation
few km
few km if pure?
few km?
noise
highly variable
small?
small?
ice only medium where optical, radio, and acoustic can be used
conclusion:
ocean noisy;
hybrid difficult
ARENA
salt impure + expensive
Justin Vandenbroucke
ice best ?
June 26, 2008
Predicted absorption length at South Pole: several km
absorption increases with depth (temperature):
theoretical model based on lab data
(P. B. Price GRL 2006)
instrument shallow ice
ARENA
Justin Vandenbroucke
June 26, 2008
Calculated acoustic radiation pattern in ice
(Ehad)
(10 EeV)
(1 EeV)
(100 EeV)
threshold = 9 mPa
ARENA
Justin Vandenbroucke
June 26, 2008
South Pole good for all methods (optical, radio, acoustic)
Build a hybrid array!
Goal: detect ~100 GZK  in a few years
LHC
- cross-calibration
- confidence in signals
- background rejection
- event reconstruction
Inexpensive:
- shallow, narrow holes
- simple electronics
ARENA
Justin Vandenbroucke
June 26, 2008
Simulated sensitivity of hybrid detector
(GZK events/yr)
~20 events/year
~40% hybrid
astro-ph/0512604
ARENA
Justin Vandenbroucke
June 26, 2008
Part 2: SPATS design
ARENA
Justin Vandenbroucke
June 26, 2008
Our host: IceCube
IceCube : 4800 DOMs on 80 strings
IceTop
currently
instrumented
IceTop : 160 Ice Cherenkov tanks on surface
AMANDA: 677 OMs surrounded by IceCube
AMANDA
40 strings installed
as of January 2008!
See A. Karle talk
Eiffel Tower
Digital Optical Module
ARENA
Justin Vandenbroucke
Svenska Dagbladet
June 26, 2008
The South Pole Acoustic Test Setup (SPATS)
• First step toward acoustic/hybrid detector at South Pole
• Purpose: measure ice properties in situ
• Measurement goals:
- Attenuation
- Noise floor
- Sound speed vs. depth
- Transients
- background for us
- interesting for glaciologists?
stick/slip glacier movement or bulk ice cracking?
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS array design
• IceCube holes, separate strings
• surface digitization
• IceCube surface cables
• 4 strings
• 7 stages per string
• stage = sensor + transmitter
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS geometry
Vertical
80 m
(for Strings ABC;
D slightly different)
100
140
190
250
Horizontal
1 sensor + 1 transmitter
at each depth
320
400
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS in-ice
hardware
transmitter module
(electronics)
transmitter piezo-ceramic
sensor module
3 piezo-ceramics inside
for full azimuthal coverage
Sebastian
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS sensor module
• Steel pressure housing
• 3 independent channels per
module
• Separated 120° in azimuth
• Few s acoustic propagation
delay between channels
• Transducer = one
piezoelectric disk per channel
• One custom amplifier board
per channel
• Differential analog output to
surface
ARENA
Justin Vandenbroucke
~10 cm
June 26, 2008
SPATS transmitter
module
• HV pulser:
• ~ 30 s pulse
up to 1.5 kV
gaussian shape
• ring shaped
piezo-ceramic
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS data
acquisition
1x master-pc
in IceCube Lab
4x embedded PC
buried in snow
above each string
power
digital comm.
GPS timing
triggered data
4x in-ice string
power
transmitter control
analog waveforms
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS String D
Designed to address new questions raised by
first year of data with 3 strings
• 100 m longer: deepest stage at 500 m
SPATS
HADES
• Improved sensor and transmitter design
• With Strings A, B, C: longer baselines
SPATS
• 2 “HADES” sensors:
SPATS
- Complementary dynamic range to SPATS sensors
- Housing impedance matched to ice (polyurethane)
SPATS
HADES
See B. Semburg talk
+ EPFL emitter
SPATS
ARENA
Deployed Dec. 24 2007
Justin Vandenbroucke
June 26, 2008
Retrievable pinger
• 1 Hz pulsed transmitter
• lowered to ~500 m and back up
• in water-filled holes before IceCube string
• operated in 6 holes
ARENA
Justin Vandenbroucke
June 26, 2008
Part 3: SPATS deployment
and performance
ARENA
Justin Vandenbroucke
June 26, 2008
Deployment
preparation
ARENA
Justin Vandenbroucke
June 26, 2008
Deployment
ARENA
Justin Vandenbroucke
June 26, 2008
4 strings successfully deployed at Pole Jan. + Dec. 2007
• SPATS B, hole 72, Jan. 11
• SPATS A, hole 78, Jan. 14
• SPATS C, hole 47, Jan. 22
• SPATS D, hole 68, Dec. 24
ARENA
first hybrid
optical/radio/acoustic
string
Justin Vandenbroucke
June 26, 2008
SPATS performance
- All 28 transmitters working
- 73 of 80 sensor channels working normally
- Continuous running except a few power outages; recover fine
- 4 string PC’s
running smoothly 6 ft
under -50° C snow
ARENA
Justin Vandenbroucke
June 26, 2008
Part 4: Results from SPATS
ARENA
Justin Vandenbroucke
June 26, 2008
counts
Noise
• Very Gaussian
• Very stable in time
• Decreases with depth
(refraction shadowing?)
one example channel
-60
See T. Karg talk
sample amplitude (V)
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS hears the IceCube drill!
Noise vs. time for each String B channel
• SPATS running
throughout 07/08
drilling season
• Drill in each hole
heard
• On way down and
up
• Heard at maximum
SPATS distance =
660 m
time
ARENA
Justin Vandenbroucke
June 26, 2008
We hear shear waves (emitter in frozen or water hole!)
transmitter in water
31 ms
• Mode conversion at interfaces: large
incident angle gives large S and small
P amplitudes
• If neutrinos produce S waves: vertex
distance and shower energy with one
sensor!
TS
TP
RP
water
TP
RP
TS
ice
water
side view
ARENA
ice
Justin Vandenbroucke
top view
June 26, 2008
Pressure and shear wave speed vs. depth measured
with SPATS
See F. Descamps talk
ARENA
Justin Vandenbroucke
June 26, 2008
Inter-string transmitter pulses
recorded through ≥125 m of South Pole ice
19 pulses from A6 to B6
1 pulse
t = 55 ms
signals detected from every string to
every other string
ARENA
Justin Vandenbroucke
1 ms
June 26, 2008
Inter-string transmitter pulses
Can hear longest baseline:
D to C = 543m
one pulse average of 89 pulses
ARENA
Justin Vandenbroucke
June 26, 2008
Attenuation analysis (1) inter-string data:
Ln(amplitude*distance) vs. distance
• 3 string data; some data points buried in noise
• significant improvements in run optimization underway: retrieve missing points
• 4 string analysis in progress
ARENA
Justin Vandenbroucke
June 26, 2008
Attenuation analysis (2) pinger data:
systematic effects
Waveform shape not consistent from one geometry to the other
Systematic effect
Effect on λ estimate
1. Interference with reflections from hole back wall
varying
2. Strong polar / azimuthal sensitivity dependence
decreases
3. Shear waves
decreases
4. Clock drift of ADC sample clocks
neutral
5. Saturation, i.e. limited dynamic range
increases
6. Ice quality: possible hole ice inhomogeneities & cracks
may affect pinger transmittivity and/or sensor sensitivity
varying
7. Noise: if not subtracted
increases
8. Transmission coefficient (combination of position of
pinger in hole, polar angle etc.)
decreases
ARENA
Justin Vandenbroucke
June 26, 2008
Many systematic effects in pinger data: an example
continuous recording of 9 pulses
P
1
pulse number
9
• Significant pulse to pulse variation
• Shear waves = “missing energy” in pressure-only analyses
• P, S wave amplitudes anticorrelated!
• Variation in transmission coefficients due to swinging/bouncing pinger?
ARENA
Justin Vandenbroucke
June 26, 2008
Attenuation analysis with pinger data:
summary of results
Affected by
systematic
effects
Analysis
Data set
Attenuation length
range of best fit (m)
1st oscillation amplitude (first min,
first max, or whichever comes first)
multi-hole,
single-level
150 to 
(1), (3), (5), (6),
(7), (8)
1st oscillation peak-to-peak pressure
+ peak-to-peak shear, noise
subtracted
multi-hole,
single-level
300 to 
(1), (6), (8)
FFT: variation of coefficients in
signal frequency range
singlehole, multilevel
80 to 200
(1?), (2), (6), (8)
FFT: integral over signal frequency
range
multi-hole,
single-level
300 to 500
(1?), (3), (6), (7),
(8)
(pressure energy) + (shear energy)
– (noise energy)
multi-hole,
single-level
100 to 600
(1), (2), (6), (8)
HADES (single pinger hole)
singlehole, multilevel

(1), (3), (6), (8)
Conclusion: attenuation length still poorly constrained due to systematics
ARENA
Justin Vandenbroucke
June 26, 2008
Part 5: What’s next?
ARENA
Justin Vandenbroucke
June 26, 2008
What’s next for attenuation measurement
Same systematics at longer baseline gives better constraints:
an illustration
(not real data!)
Conclusion: more pinger data in 08/09 season at longer distances
Also: significantly optimized 4 string data taking underway
ARENA
Justin Vandenbroucke
June 26, 2008
High energy outer strings: a possible next step for radio
and acoustic R&D with IceCube
• Final strings of IceCube likely
spaced to optimize ≥PeV
response
radio
• Radio and acoustic
instrumentation could improve
quantitatively and qualitatively
acoustic
• An opportunity for both R&D and
event detection
optical
• Simulations underway
See D. Tosi talk
• Seeking interested collaborators!
ARENA
Justin Vandenbroucke
June 26, 2008
Conclusions
SPATS first results:
arXiv:0708.2089
- Goal: ~100 GZK neutrinos in a few years
- South Pole ice best suited for acoustic (hybrid) detection?
- SPATS installed in 2007 to measure acoustic ice properties
- Running very smoothly
- Noise Gaussian, stable, decreases with depth; transients rare
- Pressure and shear wave speed vs. depth measured
- Attenuation analysis currently systematics limited
- New inter-string and pinger data will constrain attenuation better
- Toward a hybrid optical / radio / acoustic neutrino observatory at South Pole
See next 3 talks (T. Karg, F. Descamps, D. Tosi)
+ 1 tomorrow (B. Semburg) for more details
ARENA
Justin Vandenbroucke
June 26, 2008
extra slides
ARENA
Justin Vandenbroucke
June 26, 2008
Previous sound speed work in literature
P speed in firn measured
P speed beneath firn, and all S speed, modeled
D. G. Albert GRL 25 (1998)
ARENA
Justin Vandenbroucke
June 26, 2008
Overcoming module to module variation with ratios
Model:
ATA SB
1
 d AB

S BTA e
d AB
TA
SA
Assumes S, T independent of angle
TB
2 T’s + 2 S’s  couplings divide out:
dAC
dAD
ARENA
dBC
dBD
ATA SC
ATA SD
ATB SC
ATB SD
e
SB
 ((d AC d BC )(d AD d BD ))
e
x / 
One remaining unknown: 
Justin Vandenbroucke
June 26, 2008
Ratio method
(Monte Carlo)
Assume  measure  = 100 m
Treats angular variation (~40%) as uncorrelated variation in amplitudes
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS tested in frozen lake, Northern Sweden
100 km above Arctic Circle
April 2006
ARENA
Justin Vandenbroucke
June 26, 2008
Close up of pressure wave speed results
consistent with
previous result in firn
ARENA
constant between 250
and 500 m depth
Justin Vandenbroucke
June 26, 2008
Close up of shear wave speed results
• first measurement of shear wave speed
• constant between 250 and 500 m depth
ARENA
Justin Vandenbroucke
June 26, 2008
GZK  point to UHECR sources
GZK neutrinos from cosmological
distances point to UHECR sources
[D. Saltzberg]

D ~ 1 Gpc
 ~ R/D ~ 3°
ARENA
R ~ 50 Mpc
“GZK sphere”
of arbitrary B deflection/diffusion
Justin Vandenbroucke
June 26, 2008
Attenuation (1)
Scattering:
Rayleigh off ice
grains
10-2
1 km
10-4
10-6
10-8
10-10
103
ARENA
103
km
scat > 103 km
neutrino signal
104
frequency [Hz]
105
Justin Vandenbroucke
June 26, 2008
[]
Attenuation (2)
Absorption:
by thermal phonons
theoretical model
fits lab data
(P. B. Price GRL 2006)
f 2eU / kT

1 ( f / f 0 ) 2 e 2U / kT
abs ~ few km

ARENA
Justin Vandenbroucke
to be measured in situ
June 26, 2008
Refracted ray paths
surface noise shielded
deep signals from neutrinos ~unaffected
ARENA
Justin Vandenbroucke
June 26, 2008
String PC Connected in Acoustic
Junction Box
ARENA
Justin Vandenbroucke
June 26, 2008
air
Neutrino-induced cascades produce 3
detectable signals

dense medium
(3) Askaryan acoustic pancake
~few km?
(2) Askaryan radio cone
~1 km
(1) optical Cherenkov cone
~100 m
interaction
 particle shower
radio and acoustic (?) travel farther than optical in ice
ARENA
Justin Vandenbroucke
June 26, 2008
Confirmation of acoustic technique
Bipolar pulse characteristics confirmed
with Brookhaven proton beam
(Sulak et al NIM 1979)
Pressure proportional to shower energy:
coherence, calorimetry
Pressure proportional to (T):
thermo-acoustic origin
ARENA
Justin Vandenbroucke
June 26, 2008
The sound of one
neutrino clapping
time domain
1021 eV shower
1.05 km distant
with ocean absorption
SAUND-I noise
frequency domain
ARENA
Justin Vandenbroucke
June 26, 2008
Good neutrino pointing resolution
(benefit of flat acoustic radiation pattern)
Events
acoustic-only
point spread function
(assumes no noise hits!)
alternative:
combine O/R/A hits
overflow bin
Pointing error (deg)
ARENA
Justin Vandenbroucke
June 26, 2008
String PC
• A rugged computer in the IceTop
trench, in a waterproof Acoustic
Junction box next to the Surface
Junction Box
• Use a PC104(+) modular
embedded computing system
from RealTime Devices
• Developed for military applications
in extreme temperature and
vibration conditions
• All components rated to -40 C and
tested to -55 C
• All components take 5 VDC, total
usage ~25 W
ARENA
Justin Vandenbroucke
June 26, 2008
String PC: Components
CPU
3 Fast ADC’s
Relay board
Slow ADC
ARENA
Justin Vandenbroucke
June 26, 2008
Timing: Distribute IRIG-B
width:
decoding:
8
8
5
5
5
2
2
2
m
m
1
1
1
0
0
[m = marker]
0
5
1
2
0
8
(ms)
m
• From GPS clock in Master PC to all 4 strings:
- absolute time specified in 1-second frames of digital pulses
- 1 pulse every 10 ms, <1 s risetime
- pulse lengths encode bits to specify absolute time
• ADC boards can sample “marker” digital lines simultaneously with the
analog
• We use IRIG-B as a marker line  match it with the acoustic signals to
give 1-sample (~s) timing resolution
ARENA
Justin Vandenbroucke
June 26, 2008
Junction box
buried 2 m in snow
at -50 C
housing for embedded
PC, power supplies,
modems
down into ice
to indoor
computer
ARENA
Justin Vandenbroucke
June 26, 2008
SPATS Hub Service Board
1 per string in Master PC: distribute PWR + timing + comm’s
DSL in
PWR+
IRIG/Serial
out
PWR in
Serial in
PWR+DSL
out
IRIG
in
PCI
ARENA
Justin Vandenbroucke
June 26, 2008
Retrievable pinger: technical design
In hole:
- Emitter: ITC1001 (isotropic)
- HV pulser (30 s, fixed amplitude)
- Pressure loggers
On surface (“Acoustic Pinger Box”)
- Power (batteries)
- Trigger (PPS from GPS)
ARENA
Justin Vandenbroucke
June 26, 2008
Background transients
Rate: ~1/minute/channel
Loud examples:
5V
12 ms
ARENA
Justin Vandenbroucke
June 26, 2008
Attenuation analysis with pinger data: techniques
Variables
• Amplitude:
- max amplitude (9 pulses or averaged)
- first peak analysis
- peak-to-peak in the first period of the waveform
• Energy:
Calculate energies of pressure wave and shear wave and subtract the scaled
noise
Fit the energy to get attenuation
• FFT analysis:
If the λ does not depend on the frequency the amplitude attenuation can be
obtained evaluating or integrating the spectrum over the signal frequency range
Geometries
• Single hole, multi-level data analysis  minimize azimuthal variation
• Multi-hole, single-level data analysis; only aligned holes  minimize polar
variation
ARENA
Justin Vandenbroucke
June 26, 2008
Pinger double pulse: direct + reflection off rear hole wall
~0.8 ms
ARENA
Justin Vandenbroucke
June 26, 2008
Refracted neutrino pancake (SAUND velocity profile)
E = 3 x 1021 eV
ARENA
Justin Vandenbroucke
June 26, 2008