Sensor development and calibration for acoustic neutrino detection in ice
Timo Karg, Martin Bissok, Karim Laihem, Benjamin Semburg, Delia Tosi
aa r X i v : . [ a s t r o - ph . I M ] J u l PROCEEDINGS OF THE 31 st ICRC, Ł ´OD ´Z 2009 1
Sensor development and calibration for acousticneutrino detection in ice
Timo Karg ∗ , Martin Bissok † , Karim Laihem † , Benjamin Semburg ∗ , and Delia Tosi ‡ for the IceCube collaboration §∗ Dept. of Physics, University of Wuppertal, D-42119 Wuppertal, Germany † III Physikalisches Institut, RWTH Aachen University, D-52056 Aachen, Germany ‡ DESY, D-15735 Zeuthen, Germany § Abstract . A promising approach to measure theexpected low flux of cosmic neutrinos at the highestenergies (E > ) and water (11 m )for the development, testing, and calibration of acous-tic sensors. Furthermore, these facilities allow forverification of the thermoacoustic model of soundgeneration through energy deposition in the ice by apulsed laser. Results from laboratory measurementsto disentangle the effects of the different environmen-tal influences and to test the thermoacoustic modelare presented. Keywords : acoustic neutrino detection, thermo-acoustic model, sensor calibration
I. I
NTRODUCTION
The detection and spectroscopy of extra-terrestrialultra high energy neutrinos would allow us to gain newinsights in the fields of astroparticle and particle physics.Apart from the possibility to study particle accelerationin cosmic sources, the measurement of the guaranteedflux of cosmogenic neutrinos [1] opens a new windowto study cosmic source evolution and particle physicsat unprecedented center of mass energies. However, thefluxes predicted for those neutrinos are very low [2], sodetectors with large target masses are required for theirdetection. One possibility to instrument volumes of iceof the order of
100 km with a reasonable number of sensor channels is to detect the acoustic signal emittedfrom the particle cascade at a neutrino interaction vertex[3].To study the properties of Antarctic ice relevant foracoustic neutrino detection the South Pole Acoustic TestSetup (SPATS) [4] has been frozen into the upper part ofIceCube [5] boreholes. SPATS consists of four verticalstrings reaching a depth of 500 m below the surface.The horizontal distances between strings cover the rangefrom 125 m to 543 m. Each string is instrumented withseven acoustic sensors and seven transmitters. The iceparameters to be measured are the sound speed profile,the acoustic attenuation length, the background noiselevel, and transient noise events in the frequency rangefrom 1 kHz to 100 kHz.For the design of a large scale acoustic neutrinodetector it is crucial to fully understand the in-situ response of the sensors as well as the thermoacousticsound generation mechanism.II. S ENSOR CALIBRATION
To study the acoustic properties of the Antarcticice, like the absolute background noise level, and todeduce the arrival direction and energy of a neutrinoin a future acoustic neutrino telescope it is essential tomeasure the sensitivity and directionality of the sensorsused, i.e. the output voltage as function of the incidentpressure, and its variation with the arrival direction ofthe incident acoustic wave relative to the sensor. Thesemeasurements can be carried out relatively easily in thelaboratory in liquid water. The two calibration methodsmost commonly used are • the comparison method, where an acoustic signalsent by a transmitter (with negligible angular vari-ation) is simultaneously recorded at equal distancewith a pre-calibrated receiver and the sensor to becalibrated. A comparison of the signal amplitudesin the two receivers allows for the derivation of thedesired sensitivity from the sensitivity of the pre-calibrated sensor. • the reciprocity method, which makes use of theelectroacoustic reciprocity principle to determinethe sensitivity of an acoustic receiver without hav-ing to use a pre-calibrated receiver (see e.g. [7]). KARG et al.
ACOUSTIC SENSOR CALIBRATION
All SPATS sensors have been calibrated in ◦ C waterwith the comparison method [8]. However, both calibra-tion methods are not suitable for in-situ calibration ofsensors in South Pole ice. There are no pre-calibratedsensors for ice available, and reciprocity calibrationrequires large setups which are not feasible for de-ployment in IceCube boreholes. Further, directionalitystudies require a change of relative positioning betweenemitter and receiver which is difficult to achieve in afrozen-in setup.It is not clear how results obtained in the laboratoryin liquid water can be transferred to an in-situ situationwhere the sensors are frozen into Antarctic ice. Weare studying the influence of the following three envi-ronmental parameters on the sensitivity separately: lowtemperature, increased ambient pressure, and differentacoustic coupling to the sensor. We will assume that sen-sitivity variations due to these factors obtained separatelycan be combined to a total sensitivity change for frozenin transmitters. This assumption can then be checkedfurther using the two different sensor types deployedwith the SPATS setup. Apart from the standard SPATSsensors with steel housing two HADES type sensors[9] have been deployed with the fourth SPATS string.These contain a piezoceramic sensor cast in resin andare believed to have different systematics.
A. Low temperatures
The ice temperature in the upper few hundred metersof South Pole ice is − ◦ C [10]. It is not feasible toproduce laboratory ice at this temperature in a largeenough volume to carry out calibration studies. Westudy the dependence of the sensitivity on temperaturein air. A signal sent by an emitter is recorded with asensor at different temperatures. To prevent changes inthe emissivity of the transmitter, the transmitter is keptat constant temperature outside the freezer, and onlythe sensor is cooled down. The recorded peak-to-peakamplitude is used as a measure of sensitivity. First resultsindicate a linear increase of sensitivity with decreasingtemperature (cf. Fig. 1). The sensitivity of a SPATSsensor is increased by a factor of . ± . when thetemperature is lowered from ◦ C to − ◦ C (averagedover all three sensor channels).
B. Static pressure
Acoustic sensors in deep polar ice are exposed toincreased ambient static pressure. During deploymentthis pressure is exerted by the water column in the bore-hole (max. 50 bar at 500 m depth). During re-freezing itincreases since the hole freezes from the top, developinga confined water volume. The pressure is believed todecrease slowly as strain in the hole ice equilibrates tothe bulk ice volume. The final static pressure on thesensor is unknown.A . cm inner diameter pressure vessel is availableat Uppsala university that allows for studies of sensor −70 −60 −50 −40 −30 −20 −10 0 1011.522.533.544.55 preliminary temperature [C] a m p li t ude pea k t o pea k [ V ] SignalVsTemp−Ch0fit: y=a x+b a = −0.021 +/ 0.001 T/ o C b = 2.32 +/ 0.03
Fig. 1: Measured peak-to-peak amplitude of one SPATSsensor channel in air at different temperatures and linearfit to the data. a m p li t ude V pp Feb09−POS3−20kHz−4V−Pscan−allCh 012
Fig. 2: Measured peak-to-peak amplitude of a SPATSsensor excited by a transmitter coupled from the outsideto the pressure vessel. All three channels of the sensorare shown.sensitivity as function of ambient pressure. Static pres-sures between 0 and 800 bar can be reached. In thisstudy the pressure is increased up to 100 bar. Acousticemitters for calibration purposes can be placed insidethe vessel or, free of pressure, outside of it. Cable feedsallow one to operate up to two sensors or transmittersinside the vessel. A sensor is placed in the center ofthe water filled vessel. The transmitter is coupled fromthe outside to the vessel. The recorded peak-to-peakamplitude is used as a measure of sensitivity while thepressure is increased. The sensor sensitivity is measuredby transmitting single cycle gated sine wave signals withdifferent central frequencies from 5 kHz to 100 kHz.Figure 2 shows the received signal amplitudes for the
ROCEEDINGS OF THE 31 st ICRC, Ł ´OD ´Z 2009 3 three sensor channels of a SPATS sensor as a functionof ambient pressure. No systematic variation of thesensitivity with ambient pressure is observed. Combin-ing all available data we conclude that the variationof sensitivity with static pressure is less than 30% forpressures below 100 bar.
C. Sensor-ice acoustic coupling
The acoustic coupling, i.e. the fractions of signal en-ergy transmitted and reflected at the interface of mediumand sensor, differs significantly between water and ice.It can be determined using the characteristic acousticimpedance of the medium and sensor, which is theproduct of density and sound velocity and is equivalentto the index of refraction in optics. Due to the differentsound speeds the characteristic acoustic impedance ofice is about . times higher than in water.Its influence will be studied in the Aachen AcousticLaboratory (Sec. III), where it will be possible to carryout reciprocal sensor calibrations in both water and ice,and also to use laser induced thermoacoustic signals asa calibrated sound source.III. N EW LABORATORY FACILITIES
Two new laboratories have been made available to theIceCube Acoustic Neutrino Detection working group forsignal generation studies and sensor development andcalibration. a) Wuppertal Water Tank Test Facility:
For rapidprototyping of sensors and calibration studies in water,the Wuppertal Water Tank Test Facility offers a cylin-drical water tank with a diameter of . m and depthof . m (
11 m ). The tank is built up from stackedconcrete rings and has a walkable platform on top. Itis equipped with a positioning system for sensors andtransmitters and a 16-channel PC based DAQ system(National Instruments USB-6251 BNC).The size of the water volume allows for the clean sep-aration of emitted acoustic signals and their reflectionsfrom the walls and surface. This makes it possible toinstall triangular reciprocity calibration setups with sidelengths of up to 1 m. Further, installations to measure thepolar and azimuthal sensitivity of sensors are possible. b) Aachen Acoustic Laboratory: The AachenAcoustic Laboratory is dedicated to the study of thermo-acoustic sound generation in ice. A schematic overviewof the setup can be seen in Fig. 3. The main part is acommercial cooling container ( × . × . ), whichcan reach temperatures down to − ◦ C. An IceToptank, an open cylindrical plastic tank with a diameterof 190 cm and a height of 100 cm [6], is located insidethe container. The IceTop tank has a freeze control unitby means of which the production of bubble-free ice ispossible. The freeze control unit mainly consists of acylindrical semipermeable membrane at the bottom ofthe container, which is connected to a vacuum reservoirand a pressure regulation system. The membrane allowsfor degassing of the water. A total volume of ≈ of Fig. 3: Overview of the AAL setup with zoom onthe mirror holder (top), the sensor positioning system(bottom, left) and a sensor (bottom, right).bubble-free ice can be produced. A full freezing cycletakes approximately sixty days with the freezing goingfrom top to bottom.On top of the container, a Nd:YAG Laser is installed ina light-tight box with an interlock connected to the lasercontrol unit. The laser has a pulse repetition rate of up to20 Hz and a peak energy per pulse of 55 mJ at 1064 nm,30 mJ at 532 nm, and 7 mJ at 355 nm wavelength. Thelaser beam is guided into the container and deposited invariable positions on the ice surface by a set of mirrorswith coatings for the above mentioned frequencies. Theoptical feed-through consists of a tilted quartz windowto avoid damage of the laser cavity by reflected laserlight. For the detection of thermoacoustic signals, 18sensors are mounted on a sensor positioning system.The positioning system has three levels, on each level 6sensors are placed in a hexagonal geometry. Along withthe sensors, 18 sound emitters are deployed for calibra-tion and test purposes. The sensors will be calibratedreciprocally. The positioning system will also includea reciprocal calibration setup for HADES sensors andthe ability to install a SPATS sensor for calibration pur-poses. In addition, two temperature sensors are deployedat each level. The acoustic sensors are pre-amplifiedlow-cost piezo based ultrasound sensors, usually usedfor distance measurement. The sensors show a strongvariation of signal strength with incident angle. Thisdirectionality has to be studied but is rather useful for KARG et al.
ACOUSTIC SENSOR CALIBRATION the suppression of reflected signals. The sensors are readout continuously by a LabVIEW-based DAQ frameworkwith a NI PCIe-6259M DAQ card. The frameworkincludes a temperature and acoustic noise monitoringsystem.IV. S
TUDIES OF THERMOACOUSTIC SIGNALGENERATION
A detailed understanding of thermoacoustic soundgeneration in ice is crucial for designing an acousticextension to the IceCube detector. The dependence of thesignal strength on the deposited energy as well as on thedistance to the sensor is of great interest. Also the pulseshape and the frequency content have to be studied sys-tematically with respect to various cascade parameters.The spatial distribution of the acoustic signal has to beinvestigated, i.e. the acoustic disk and its dependence onthe spatial and temporal energy deposition distribution.In addition, the AAL setup will be able to study thethermoacoustic effect in a wide temperature range from ◦ C to − ◦ C and possible differences of the effectin ice and water.In the Aachen Acoustic Laboratory, the thermo-acoustic signal is generated by a Nd:YAG laser. A laser-induced thermoacoustic signal differs from a signal pro-duced in a hadronic cascade. While a cascade’s energydeposition profile can be described by a Gaisser-Hillasfunction, the laser intensity drops of exponentially. Alsothe lateral profile of a cascade follows a NKG function,where, assuming a TEM mode in the far field region,the typical laser-beam profile is Gaussian. Knowing this,a recalculation of the signal properties from a laser-induced pulse to a cascade-generated pulse is possible.The frequency content of the signal is expected to varywith the beam diameter, while a too short penetrationdepth will result in an acoustic point source rather thana line source. The absorption coefficient of light inwater or ice varies strongly with wavelength. The firstwavelength of the laser (1064 nm) is absorbed after fewcentimeters, while the second harmonic (532 nm) hasan absorption length of ≈ m. The third harmonic at355 nm with an absorption length of ≈ m is expectedto be the most suitable wavelength to emulate a hadroniccascade with a typical length of 10 m and a diameter of10 cm. The diameter of the heated ice volume has to becontrolled by optics inside the container that will widenthe beam.With an array of 18 sensors, the Aachen Acoustic Lab-oratory allows the study of the spatial distribution of thegenerated sound field, as well as the frequency contentwith varying beam parameters. The first thermoacousticsignal has been generated and detected in a test setupwith preliminary sensor electronics in order to determinea reasonable gain for the pre-amplifier while avoidingsaturation. A small volume of bubble-free ice has beenproduced, containing a sensor and an emitter. Laserpulses (wavelength 1064 nm, 55 mJ per pulse) have beenshot at the ice block. A zoom on the first waveform is t [ms]0 0.5 1 1.5 2 2.5 3 3.5 4 U [ V ] −0.1−0.0500.050.1 Fig. 4: Thermoacoustic pulse in ice, generated with alaser at 1064 nm wavelength and a beam diameter of ≈ µ m .presented in Fig. 4. The distance between laser spot andsensor is approximately
15 cm and the sensor gain factoris 22. A Fourier transform of the pulse implies a pulsecentral frequency of ≈
100 kHz , which is expected forsuch a small beam diameter. In order to see the expectedbipolar pulse, further studies have to be performed todetermine the transfer function of the sensors.V. C
ONCLUSIONS
Detailed understanding of the thermoacoustic soundgeneration mechanism and the response of acoustic sen-sors in Antarctic ice is necessary to design an acousticextension for the IceCube neutrino telescope. While in-situ calibrations in deep South Pole ice are inherentlydifficult, different environmental influences can be stud-ied separately in the laboratory. No change in sensorresponse with increasing ambient pressure was found; alinear increase in sensitivity with decreasing temperaturewas observed. Intense pulsed laser beams can be usedto generate thermoacoustic signals in ice which can alsobe used as an in-ice calibration source.A
CKNOWLEDGMENTS
We are grateful for the support of the U.S. NationalScience Foundation and the hospitality of the NSFAmundsen-Scott South Pole Station.This work was supported by the German Ministry forEducation and Research.R
EFERENCES[1] V. S. Berezinsky and G. T. Zatsepin, Phys. Lett. B (1969)423.[2] R. Engel, D. Seckel and T. Stanev, Phys. Rev. D (2001)093010, arXiv:astro-ph/0101216 .[3] G. A. Askaryan, At. Energ. (1957) 152.[4] J. Vandenbroucke et al., Nucl. Instr. and Meth. A (2009),doi:10.1016/j.nima.2009.03.064, arXiv:0811.1087[astro-ph] .[5] http://icecube.wisc.edu/[6] T. K. Gaisser, in Proceedings of the 29th International CosmicRay Conference (ICRC 2005), arXiv:astro-ph/0509330 .[7] R. J. Urick, Principles of underwater sound , 3rd ed., PeninsulaPublishing (1983).
ROCEEDINGS OF THE 31 st ICRC, Ł ´OD ´Z 2009 5 [8] J.-H. Fischer, Diploma Thesis, Humboldt-Universit¨at zu Berlin(2006).[9] B. Semburg et al., Nucl. Instr. and Meth. A (2009),doi:10.1016/j.nima.2009.03.069, arXiv:0811.1114[astro-ph] .[10] P. B. Price et al., Proc. Nat. Acad. Sci. U.S.A.99