Toward Hybrid OpticalRadioAcoustic Detection of EeV Neutrinos

1
Toward Hybrid Optical/Radio/Acoustic Detection of
EeV Neutrinos

• Justin Vandenbroucke
• (UC Berkeley, justinav_at_berkeley.edu)
• with
• Dave Besson
• Sebastian Böser
• Rolf Nahnhauer
• Rodín Porrata
• Buford Price
• 2nd Workshop on TeV Particle Astrophysics,
  Madison, August 30 2006

2
The goal GZK physics with an IceCube extension
at South Pole

• 100 GZK events (e.g. 10 yrs _at_ 10/yr) would give
  a quantitative measurement including energy,
  angular, and temporal distributions
• Non-optical techniques must be used at these
  energies and their systematics are not well
  understood
• ? Use a hybrid technique same advantages of
  Auger and accelerator detectors

3
Goal 1 Identify UHECR sources

• - Neutrinos generally point to sources
• - However, GZK neutrinos are not produced in the
  source or even in its radiation field but 50 Mpc
  away
• - But its still true

D. Saltzberg
2 Gpc
?
GZK sphere of arbitrary B deflection/diffusion
? (50 Mpc) / (2 Gpc) 1.4
4
Goal 2 Measure ??N _at_ ECM 100 TeV
A. Connolly
100 events measure Lint 400 km 33
5
The Engel, Seckel, Stanev (ESS) GZK flux model
zmax 8, n 3
We use ?? 0.7
?? 0
6
A simple hybrid optical/radio/acoustic detector
Monte Carlo

• 1016 - 1020 eV ? 2? down-going neutrinos
• All flavor, all interaction (first bang only)
• Optical only muons for now (no light from
  showers)
• Radio acoustic hadronic shower for all
  channels (LPM washes out EM component), Esh
  0.2E?
• Vertices uniformly in fiducial cylinder
• AMANDA, RICE, and SAUND code

7
An example hybrid array
Optical 80 IceCube 13 IceCube-Plus (Halzen
Hooper astro-ph/0310152) holes at 1 km radius
(2.5 km deep) Radio/Acoustic 91 holes, 1 km
spacing, 1.5 km deep
shift real array to avoid clean air sector
8
Acoustic simulation

• Based on SAUND tools
• Differences from water
• - signal 10x higher
• - noise 10x lower, limited by sensors (not
  ambient)?
• different refraction (opposite and smaller)
• shear waves?
• - Unknown ice properties to be measured by SPATS
• - For now we use a model for absorption length,
  extrapolated from lab measurements (P. B. Price
  astro-ph/0506648)

9
Sound velocity profile in South Pole ice
Sound channel ridge
measured in firn (J. Weihaupt)
Firn (uncompactified snow) in top 200 m Vsound
increasing with density? refraction. Rcurvature
200 m!
predicted in bulk (using IceCube-measured
temperature profile and A. Gow temperature
coefficient) - measure with SPATS?
10
Strong refraction in firn
Acoustic upward
Radio downward
Signals always bend toward minimum propagation
speed, but Radio adores vacuum c 3e8
m/s Sound abhors vacuum c 0
11
Signals from bulk ice (neutrinos) somewhat
refracted
source in bulk
(emit a ray every 5)
12
Predicted depth (temperature)-dependent acoustic
absorption at 10 kHz
P. B. Price model absorption frequency-independen
t but temperature (depth)-dependent
In simulation, integrate over absorption from
source to receiver
instrumented
13
Acoustic detection contours in ice
Contours for Pthr 9 mPa raw discriminator, no
filter
14
Coincident effective volumes event ratesfor
IceCube (I), an optical extension (O), and
combinations with surrounding A R arrays
(GZK events/yr)
astro-ph/0512604
Curves with I/O will improve when light from
cascades included
15
Event reconstruction

• For physics we need E? and/or (??, ??), perhaps
  from (x, y, z)cascade
• A, R can get good pointing from cascades (O gets
  30 in ice)
• Multiple constraints
• O, R, A x timing, radiation pattern, hit
  amplitudes, up/down going, polarization
• How best to use and combine information?
• 1) timing most powerful (esp. for R, A)
• 2) radiation pattern (R cone, A pancake, O
  candies) also useful
• 3) hit amplitude most uncertain (except for O)
• Hybrid reconstruction?
• When possible with sub-arrays but improved with
  hybrid array
• When impossible with sub-arrays but possible with
  hybrid array
• ? lower multiplicity threshold (maximize
  physics/)

16
Mono or hybrid reconstruction from timing alone

• - For unscattered signals, Ni hits in sub-array i
  constrain source
• to Ni -1 hyperboloids

• NRNA hits determine (NR -1) (NA -1)
  hyperboloids

- Alternative exploiting cacoustic ltlt cradio,
we get (NR - 1) hyperboloids and (NA) spheres,
because t(emission) t(first radio hit) compared
to acoustic hit time

• Also true for OA, even with scattering tO tR
  few ?s ltlt tA s)

? Reconstruction possible with 1 fewer total hits

• Linear analytical solution exists for most
  (NO,NR,NA) with at least 4 hits

• Acoustic shear waves? Another velocity

17
Proof-of-principle Monte Carlo

• Demonstrate we get a single solution with
  reasonable precision
• Choose source and module locations randomly for
  each event (array and radiation pattern
  independence)
• Time resolution smear by 5 ns (R) and 10 ?s
  (A)
• No refraction (will worsen resolution)
• No noise hits (will require higher multiplicity)
• No receiver location error (will add absolute
  resolution floor)

18
Cascade location reconstruction results
5 acoustic hits 2.0 m
5 radio hits 48.8 m
19
Instead of using timing only, we could use
radiation pattern geometry only (no amplitudes)

• Radio beamed in thin cone, acoustic in thin
  pancake
• Bad for event rate, good for reconstruction
• Acoustic even with pancake thickness and
  refraction,very flat ? fit a plane through the
  hit modules, upward normal points to the GZK
  source
• Only requires 3 hits on 3 strings
• What about E?? Need vertex not just direction
• But now a 2D problem transform to the plane and
  intersect hyperbola within it (need 3-4 hits)
• Similar for radio 5 parameters determine a cone
  (known opening angle) ? need 5 hits

20
Another demo MC pointing resolutionusing
acoustic radiation pattern only (no timing)
determine hits, fit plane, compare neutrino
direction
actual radiation pattern no refraction no noise
hits 0.5 km hole spacing isotropic 1019 eV ?s
overflow bin
21
Conclusions

• Optical high energy neutrino detection proven by
  AMANDA with thousands of atmospheric neutrinos
• GZK physics will require new techniques with
  large uncertainties
• Bootstrap them using coincidence with IceCube and
  with each other
• Join efforts with a large hybrid array with
  hybrid advantages
• R/A shallower narrower cheaper holes
• 10 GZK events per year are possible
• Hybrid reconstruction techniques are promising
• South Pole possibly best place on Earth for all 3
  techniques
• Such a detector could discover UHECR sources and
  measure a cross section at 100 TeV ECM

22
Extra slides
23
O(91) radio/acoustic strings for a fraction of
the IceCube cost?

• Holes 3 times smaller in diameter (20 cm) and
  1.5 km deep
• Don LeBar (Ice Coring and Drilling Services)
  drilling estimate 33k per km hole length after
  400k drill upgrade to make it weatherproof and
  portable (cf. SalSA 600k/hole)
• Sensors simpler than PMTs
• Cables and DAQ Only 5 radio channels per string
  (optical fiber). 300 acoustic modules per
  string, but
• Cable channel reduction Send acoustic signals to
  local in-ice DAQ module (eg 16 sensor modules per
  DAQ module) which builds triggers and sends to
  surface
• Acoustic bandwidth and timing requirements are
  easy (csound 10-5 clight!)
• Acoustic data bandwidth per string 0.1-1 Gbit,
  could fit on a single ethernet cable per string

24
Acoustic event rate depends on threshold (noise
level) and hole spacing
Trigger 3 strings hit ESS GZK events per year
Need low-noise sensors (DESY) and low-noise ice
(South Pole?) Frequency filtering may lower
effective noise level For hybrid MC, set
threshold at 9 mPa a few sigma
25
Optical simulation

• Check Halzen Hoopers rate estimate with
  standard simulation tools run a common event set
  through optical, radio, and acoustic simulations
• For now, only simulate the muon channel (cascades
  in progress)
• Use standard AMANDA simulation tools muon
  propagation, ice properties, detector response
• Define a coincidence to be hits at 2 of 5
  neighboring modules on one string within 1000 ns
• Require 10 coincidences in the entire array
  within 2.5 ?s
• For optical-only events, require gt 182 channels
  hit (a muon energy cut proxy) to reject
  atmospheric background
• Do not apply Nch requirement when seeking
  coincidence with radio or acoustic

26
Radio simulationUsing RICE Monte Carlo

• Dipole antennas in pairs to resolve up-down
  ambiguity
• 30 bandwidth, center frequency 300 MHz in air
• Effective height length/?
• Radio absorption model based on measurements by
  Besson, Barwick, Gorham (accepted by J. Glac.)
• Trigger require 3 pairs in coincidence
• Use full radio MC

27
Resolution results one sub-array alone, 6 hits
acoustic
radio
28
Resolution results 1 radio 4 acoustic hits
intersect 4 spheres without the radio hit we
would not know the sphere radii, or would have
too few hyperboloids