Targeting ultra-high energy neutrinos with the ARIANNA experiment
A. Anker, S. W. Barwick, H. Bernhoff, D. Z. Besson, Nils Bingefors, G. Gaswint, C. Glaser, A. Hallgren, J. C. Hanson, R. Lahmann, U. Latif, J. Nam, A. Novikov, S. R. Klein, S. A. Kleinfelder, A. Nelles, M. P. Paul, C. Persichilli, S. R. Shively, J. Tatar, E. Unger, S.-H. Wang, G. Yodh
TTargeting ultra-high energy neutrinos with the ARIANNA experiment
A. Anker a , S. W. Barwick a , H. Bernhoff b , D. Z. Besson c,d , Nils Bingefors e , G. Gaswint a , C. Glaser a, ∗ , A.Hallgren e , J. C. Hanson f , R. Lahmann a,g , U. Latif c , J. Nam h , A. Novikov c,d , S. R. Klein i , S. A.Kleinfelder j , A. Nelles k,l , M. P. Paul a , C. Persichilli a , S. R. Shively a , J. Tatar a,m , E. Unger e , S.-H. Wang h ,G. Yodh a a Department of Physics and Astronomy, University of California, Irvine, CA 92697, USA b Uppsala University Department of Engineering Sciences, Division of Electricity, Uppsala, SE-752 37 Sweden c Department of Physics and Astronomy, University of Kansas, Lawrence, KS 66045, USA d National Research Nuclear University MEPhI (Moscow Engineering Physics Institute), Moscow 115409, Russia e Uppsala University Department of Physics and Astronomy, Uppsala, SE-752 37, Sweden f Whittier College Department of Physics, Whittier, CA 90602, USA g ECAP, Friedrich-Alexander Universit¨at Erlangen-N¨urnberg, 91058 Erlangen, Germany h Department of Physics and Leung Center for Cosmology and Particle Astrophysics, National Taiwan University, Taipei10617, Taiwan i Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA j Department of Electrical Engineering and Computer Science, University of California, Irvine, CA 92697, USA k DESY, 15738 Zeuthen, Germany l Humbolt-Universit¨at zu Berlin, Institut f¨ur Physik, 12489 Berlin, Germany m Research Cyberinfrastructure Center, University of California, Irvine, CA 92697 USA
Abstract
The measurement of ultra-high energy (UHE) neutrinos (E > eV) opens a new field of astronomy withthe potential to reveal the sources of ultra-high energy cosmic rays especially if combined with observations inthe electromagnetic spectrum and gravitational waves. The ARIANNA pilot detector explores the detectionof UHE neutrinos with a surface array of independent radio detector stations in Antarctica which allows fora cost-effective instrumentation of large volumes. Twelve stations are currently operating successfully at theMoore’s Bay site (Ross Ice Shelf) in Antarctica and at the South Pole. We will review the current state ofARIANNA and its main results. We report on a newly developed wind generator that successfully operates inthe harsh Antarctic conditions and powers the station for a substantial time during the dark winter months.The robust ARIANNA surface architecture, combined with environmentally friendly solar and wind powergenerators, can be installed at any deep ice location on the planet and operated autonomously. We report onthe detector capabilities to determine the neutrino direction by reconstructing the signal arrival direction ofa 800 m deep calibration pulser, and the reconstruction of the signal polarization using the more abundantcosmic-ray air showers. Finally, we describe a large-scale design – ARIA – that capitalizes on the successfulexperience of the ARIANNA operation and is designed sensitive enough to discover the first UHE neutrino. Keywords: neutrino, radio, Antarctica, cosmic ray, Askaryan radiation, ARIANNA
1. Introduction
The origin of ultra-high energy cosmic rays (UHECRs) is one of the biggest mysteries in astroparticlephysics. The deflection of cosmic rays in galactic and extra-galactic magnetic fields makes the determinationof their origin extremely challenging, and despite decades of research, no UHECR source could be identified.A more promising way to track down the sources of cosmic rays is the measurement of ultra-high energy ∗ Corresponding author
Email address: [email protected] (C. Glaser)
Preprint submitted to Elsevier June 25, 2019 a r X i v : . [ a s t r o - ph . I M ] J un UHE) neutrinos as neutrinos traverse the universe unimpeded and point back to their source. Neutrinos areeither produced directly at an astrophysical object through the acceleration of cosmic rays and succeedinginteractions with matter surrounding the source (called astrophysical neutrinos), or during the propagationof cosmic rays through the universe via interactions with photons from the cosmic microwave background(called cosmogenic neutrinos).Recently, the IceCube neutrino detector at the South Pole reported a detection of a 3 × eV astro-physical neutrino in coincidence with a flaring state of a blazar that was observed with gamma-ray telescopes(Aartsen et al., 2018b). With a significance of 3 σ , it constitutes the first evidence of a cosmic-ray source.This discovery shows the strength of neutrino astronomy and the multi-messenger approach as one of thebest opportunities to discover the sources of cosmic rays also at even higher energies.As the interaction cross section of UHE neutrinos is small, huge volumes need to be instrumented toobserve neutrinos with sufficient statistics. The ice sheets of Antarctica and Greenland provide a suitabledetection medium, and a sparse instrumentation of the ice with sensors allows for the construction ofterraton detectors. Neutrinos with energies above 10 eV are measured best with the radio technique(Aartsen et al., 2016) because of the small attenuation of radio signals in ice ( L a ≈ ◦ inice and a width of a few degrees, as depicted in Fig. 2. Hence, the observable Askaryan radio pulse dependsstrongly on the geometry as shown in Fig. 1. If the shower is observed directly on the Cherenkov cone, thesignal is strongest and extends to frequencies of ∼ ∼
50 MHz to ∼ (cid:31) ≈
15 cm) that are drilled from the ice surface. This comes with the advantages of a larger sensitivityper station, and being less impacted by firn effects: In the upper part of the ice layer (at the South Pole theupper ∼
200 m), the index of refraction changes from 1.78 of clear deep ice to 1.35 at the surface (Kravchenkoet al., 2004; Barwick et al., 2018) because of a change in the ice density. Therefore, the radio signals donot propagate on straight lines but undergo continuous Fresnel refraction. Still, the drilling is a substantialdeployment effort and only antennas that fit into the narrow hole can be used which limits the broadbandresponse and the sensitivity to the horizontal signal polarization.The ARIANNA detector, in turn, consists of autonomous detector stations located just slightly belowthe ice surface. This allows a quick deployment and the installation of large high-gain broadband LPDAantennas with different orientations, which enables a precise measurement of the signal polarization toreconstruct the neutrino direction, and the measurement of the frequency spectrum which is required todetermine the neutrino energy. Being close to the surface comes with additional advantages: In dipoleantennas a couple of meters below the surface (which is a straight-forward extension of the current stationlayout and part oft the ARIA concept), the Askaryan radio pulse can be observed two times, one pulse thatpropagates directly to the antenna and one that is reflected off the ice surface, for almost all events. Thisallows for an unambiguous neutrino identification, and the reconstruction of the distance to the interactionvertex. Furthermore, an ARIANNA station is also sensitive to the radio signals of cosmic-ray air showerswhich allows for an in-situ calibration, and continuous test and monitoring of the detector under realisticconditions. However, the smaller exposure to neutrinos per station requires the installation of more stations2 igure 1: Askaryan radio pulse for different viewing angles with respect to the Cherenkov cone = 0 ◦ for a hadronic showeraccording to the model presented in Alvarez-Mu˜niz et al. (2000) and Alvarez-Mu˜niz and Zas (1998). (left) Time domainrepresentation. The pulse start time is shifted for better visibility. (right) Frequency spectrum. to reach the same sensitivity as an array of deep detectors, and firn effects need to be taken into account inthe data analysis.In this paper we report on the accomplishments of the ARIANNA pilot array and present a improveddesign of a large-scale radio neutrino detector – called ARIA – that has emerged from the ARIANNAexperience. We will review relevant previously published ARIANNA results for completeness and presentseveral new achievements relevant for the design of ARIA, and other Askaryan based high-energy neutrinodetectors with a significant surface component of LPDA antennas. We present a novel wind generatorspecifically designed to work at the harsh Antarctic conditions capable of powering the station through thedark winter months. The robust ARIANNA surface architecture, combined with environmentally friendlysolar and wind power generators, can be installed at any deep ice location on the planet and operatedautonomously. This design avoids the necessity to lay power and communication cables from a centralizedpower generator, which is expensive and geographically restrictive.We report on an in-situ calibration measurement at the South Pole where the signal arrival direction isreconstructed from pulses of a 800 m deep calibration pulser and discuss its implications on the understandingand modelling of signal propagation through the ice. We then discuss how neutrino radio signals can beidentified and demonstrate this technique using the more abundant cosmic-ray air showers. Furthermore,air-shower radio signals are used to show the capabilities to reconstruct the signal polarization, an importantquantity to determine the neutrino direction. Finally, we will present the array and detector layout of ARIA,briefly discuss how neutrino signals are identified and how the neutrino properties can be reconstructed, andpresent its expected sensitivity.
2. The ARIANNA detector
The ARIANNA detector consists of autonomous and independent stations, i.e., the information of onestation is sufficient to measure a neutrino and multi-station coincidences are not required. The stationlayout is depicted in Fig. 2. Each station comprises two pairs of downward facing LPDA antennas withorthogonal orientation and are spatially separated by 6 m. In the second generation of ARIANNA stations,the downward facing LPDAs for neutrino detection are complemented by two pairs of upward pointingLPDAs for cosmic-ray detection and vetoing.ARIANNA stations are solar powered and communication takes place via the Iridium satellite networkor a high-speed long-range wifi connection. The ARIANNA pilot array was initially deployed at Moore’sBay on the Ross ice shelf near the coast of Antarctica (Gerhardt et al., 2010) but the autonomous natureof the ARIANNA design allows a deployment at any suitable site around the world, so that two ARIANNAstations were deployed in 2017 and 2018 at the South Pole and have been operating successfully since then.3 igure 2: Sketch of the ARIANNA detector at the Ross Ice Shelf.
The main part of the ARIANNA pilot array is the hexagonal radio array (HRA) (Barwick et al., 2015b).It consists of seven 4-channel stations and has been installed at the Moore’s Bay site. All stations have beenoperating successfully since their deployment (the first stations were deployed in 2012) demonstrating theenormous stability of the ARIANNA hardware in harsh Antarctic conditions.The ARIANNA hardware is based on the SST chip design (Kleinfelder et al., 2014). It currently supportsup to 8 input channels on a single board. The input is sampled at 1 GSPS or 2 GSPS and continuously storedin a switched capacitor array. The SST chip has a precise time synchronization between samples and acrosschannels of less than 5 ps (Chiem, 2017) which allows for a precise reconstruction of the signal directiondespite the small spatial extent of an ARIANNA station. Following a trigger, 256 samples are digitizedwith 12 bit ADCs and read into an FPGA and thereafter into an Mbed micro processor (arm Mbed) for thecalculation of a second trigger stage and data storage.Interesting events are triggered using a high and low threshold crossing requirement: The input signalneeds to cross a remotely adjustable positive and negative threshold within 5 ns. This requirement substan-tially reduces the trigger rate on thermal noise fluctuations compared to a single threshold trigger whileretaining the same trigger efficiency on Askaryan pulses. To further reduce the trigger rate on thermal noisefluctuations, coincidences between multiple channels are required. A typical setting is to require 2 triggersout of the 4 channels within 30 ns which corresponds to the maximal physical allowed propagation timebetween the antennas. The thresholds are set to 4 times the RMS noise and results in a typical triggerrate of 10 mHz. The threshold are typically set only once at the beginning of the season because the noiseconditions are very stable and a continuous adjustment of thresholds is not required. We note that weobserve short periods of elevated rates caused by non-thermal noise, i.e., the elevated rates are not causedby a change in the average RMS noise. Most of these occurrences are caused by high-wind speeds.We occasionally observe continuous narrowband signals from, e.g., air plane communication or weatherballoons. We therefore compute an additional L1 trigger that rejects events with a strong signal in only onefrequency bin. More advanced trigger schemes are not necessary at this stage of the experiment given theextremely radio quiet environment at Moore’s Bay and at the South Pole.
In addition to the 7 HRA stations with 4 downward facing LDPAs that serve as a pilot array of alarge scale neutrino detector, several other station designs have been installed for different science cases (cf.4 igure 3: Overview of ARIANNA stations. (upper left) dedicated cosmic-ray station 32 at Moore’s Bay, all antennas are inthe ice (upper right) 8 channel neutrino/cosmic ray combo station 52 at Moore’s Bay, all antennas are in the ice (lower left)horizontal cosmic-ray station 50, all antennas are above the ice (lower right) first ARIANNA station 51 at the South Pole, thedownward facing antennas and the dipole are in the ice whereas the upward facing antennas are above the ice.
Fig. 3).The capabilities of cosmic-ray measurement were first demonstrated using a dedicated cosmic-ray station(station 32) consisting of four upward facing LPDAs (Barwick et al., 2017) before the ARIANNA hardwarewas extended to support 8 channels. A year later in the 2017/2018 season, the first 8 channel station (station52) was deployed combining 4 downward and 4 upward facing LPDAs.In the same season, the first ARIANNA station was installed at the South Pole. Similar to station 52,it combines downward and upward facing LPDAs complemented by one dipole for a direct access of thevertical polarization. The upward facing LPDAs are placed above the surface to better study the RF noisesituation at the South Pole. The station is connected to the power grid of the ARA detector so it canrun all year. Since its deployment, the station has been running continuously proving that the ARIANNAhardware works reliably at the even colder temperatures at the South Pole which has average temperaturesof − ◦ C during the winter compared to − ◦ C at Moore’s Bay. The station was deployed relatively closeto the South Pole station and in direct vicinity to a large wind turbine that turned out to be a significantsource of RFI noise. Although these events could be identified in an offline analysis it complicates datataking, limits the uptime in certain periods and results in larger data volumes.In the 2018/2019 season, another ARIANNA station was installed at the South Pole further away fromany man made infrastructure. The station is now running completely autonomous with solar power and abattery to buffer periods of insufficient sunlight which results in a 100% uptime during the austral summer.Also the second ARIANNA station has been running reliably since it was turned on. Being further awayfrom South Pole station and the wind turbine has had the desired effect and we observe stable trigger rates. This wind turbine is a kW scale power system from a different experiment and completely different from the low-powersystem that is discussed in Sec. 2.3.
Another promising technique of a flavor sensitive neutrino detector (Feng et al., 2002; Hou and Huang,2002; Nam and Liu, 2017) is explored at the ARIANNA site similar to the approach of (Alvarez-Muniz et al.,2018). A tau neutrino interaction produces a tau lepton which has just the right combination of interactionand decay length that it can escape a solid medium, then decay in air and produce an air shower that isin-turn observable with a radio detector. A solid mountain range provides an excellent target material. TheMoore’s Bay site provides an optimal location as it is surrounded by the massive Transantarctic Mountainsand is extremely radio quiet.We prototype this technique with a horizontal cosmic-ray (HCR) station that consists of an array ofantennas pointed towards the mountain range. Cosmic-ray induced air showers constitute the main physicalbackground of such a tau neutrino detector. Hence, a precise measurement of air showers, in particular theangular reconstruction to distinguish air showers coming from slightly above the mountain from tau inducedair showers coming from below the mountain ridge, is crucial to study the feasibility of this technique.In the 2017/2018 season, an earlier HCR prototype station (Wang, 2017) was extended to the layoutdepicted in Fig. 3 bottom right. This station consists of eight LPDAs placed above the snow pointing at themountain range that surrounds Moores Bay. An initial analysis of data from this improved station designindicates an improvement of the angular resolution in elevation from 0 . ◦ to 0 . ◦ for signals originatingfrom close to the horizon. However, we observe a systematic offset of a couple of degrees which is mostly theresult of interference with signals reflected off the surface and due to uncertainties in the station geometry.A full analysis of the HCR station will be presented in a forthcoming publication. A wind power system would allow an autonomous station to run during the antarctic winter when thestation is in darkness for almost 6 months. This has a huge potential as it directly doubles the uptime of thedetector and its multi-messenger sensitivity. A wind generator capable of both surviving extreme Antarcticweather conditions and providing power at low wind-speeds has been in development since several years.The turbine has a novel geometry based on the traditional twisted Savonius turbine. Its main elementsare two twisted displaced half circle arc cross-sections, in the horizontal plane, that are swept vertically withan increasing azimuth around the axis of rotation (cf. Fig. 4 left). This twist helps to reduce vibrations,noise and the torque fluctuations. Above all it renders the turbine aerodynamically self starting. Integratedwith the turbine is the permanent magnet rotor of an air gap winding synchronous generator. This allowsfor a full electrical control of the turbine’s rotational speed through the loading of the generator.The first important milestone was achieved in 2018 when the first wind generator survived the wintermonths and powered an ARIANNA station for 24% of the winter time (cf. Fig. 4 right). In Nov 2018, alarger version was deployed at Moores Bay with several changes to improve performance at low temperatureswhich is shown in Fig. 4 left. Its dimensions are 0 . . . . . igure 4: (left) Photo of newest version of the wind generator installed in Nov 2018 at Moore’s Bay. (right) Shows operation ofARIANNA station throughout the dark winter (April-Sept) in 2018. During this part of the year, power was supplied entirelyby Wind Generator.
3. Reconstruction of signal direction
In this section, we present a measurement campaign performed at the South Pole during the australsummer in 2017/2018. This campaign measured the propagation of radio signals from deep in the ice to thesurface and constitutes an important proof-of-concept of a surface array. Furthermore, it provides importantinsights into the understanding and modelling of signal propagation through the ice and in particular thefirn. In the upper ice layers (the firn), the index of refraction changes from n = 1 .
78 of deep ice to n = 1 . .
75 km deep which allowed to place the pulser deep enough to be outside of the shadow zone . Signaltrajectories are bent downwards due to the changing index of refraction, such that signals emitted from ashallow depth are not able to reach the ARIANNA station. We note that exceptions from this classicalpicture have been observed (see e.g. Barwick et al. (2018)) and indeed also signals emitted from within theshadow zone are visible in this measurement. However, an analysis of this data is beyond the scope of thisarticle and will be addressed in an upcoming publication. Here, we focus on the ’classical’ region that isimportant for the default operation of a radio neutrino detector.This data is used to reconstruct the signal arrival direction which is then compared to the predicteddirections calculated from the known geometry and ray tracing the signal though the ice. We analyzed51 pulses emitted from a depth between 761 m and 841 m. The signal arrival direction was reconstructedfrom the time differences of the two pairs of parallel channels of the downward pointing LPDAs. The timedifferences were determined using a cross-correlation method. The advantage of this method is that it isindependent of the description of the antenna response as only the time differences of parallel channelsare considered where the antenna response is the same and thus cancels out as systematic uncertainty.Please refer to (Persichilli, 2018) for more details on the analysis and (Glaser et al., 2019c) for details of thereconstruction algorithm.For each pulse, we determine the expected arrival direction via ray tracing the signal through the ice,and compare it with the reconstructed direction. The result is presented in Fig. 5 and we find a scatter ofless than 1 ◦ . This has been achieved despite the small lever arm of only 6 m because of the excellent timesynchronization across channels of the ARIANNA hardware of better than 5 ps.7 [deg] e n t r i e s mean = 2.2STD = 0.5 [deg] e n t r i e s mean = -2.6STD = 0.6 Figure 5: Experimentally determined angular resolution from an in-situ measurement at the South Pole. The histograms showthe difference between the expected and reconstructed direction of the zenith angle (left) and azimuth angle (right).
We observe an offset of 2 . ◦ in the azimuth angle and 2 . ◦ in zenith angle. We thoroughly evaluatedall experimental uncertainties and can’t exclude that the observed offsets are due to uncertainties in thedetector calibration and position of the emitter. However, the offset in the zenith angle might also bedue to uncertainties in the calculation of the expected signal arrival direction, e.g., from uncertainties inthe assumed index of refraction profile. From the experience gained in this calibration campaign, we canimprove the detector calibration in the future such that it won’t limit the uncertainties in the directionalreconstruction.The main conclusions with relevance to a radio neutrino detector from this measurement are: First, radiosignals propagate from deep in the ice to the surface, i.e., neutrinos can be observed via an array of antennasplaced just slightly below the ice surface. Second, the index-of-refraction profile is understood well enoughand our mathematical modelling of the signal propagation via ray optics is accurate enough to predict theincoming signal direction within two to three degrees and might even be much better as some of the offsetcan originate from systematic uncertainties in the detector calibration or position of the emitter.This result underlines the importance to precisely measure the ice properties and to understand thepropagation of radio signals through the ice. During the last austral summer of 2018/19, an improvedmeasurement was performed that built up on the experiences gained with the measurement presented here.This new measurement will allow us to study propagation effects on the pulse form and the study ofbirefringence effects. We will report on this new measurement in an upcoming publication.We note that the signal arrival direction is different from the neutrino direction because the radio signalis emitted on a cone around the neutrino direction with an opening angle of about 56 ◦ . To determinethe neutrino direction, we also need to know the polarization of the radio signal, which is addressed inSec. 5. Given the angular direction of the arriving signal and the polarization, the neutrino direction canbe determined. It is limited by the few degree width of the Cherenkov cone, though it may be possible toconstrain the angular offset from the axis of the Cherenkov cone by examining the frequency dependence ofAskaryan pulse.
4. Search for neutrino signals
The unambiguous identification of radio pulses originating from a neutrino interaction in the ice is thefirst purpose of the ARIANNA detector. Already the distinctive detector response to an impulsive signalallows for an efficient discrimination of neutrino candidates against the thermal and anthropogenic noisebackground. In particular, the dispersion of the LPDA antenna, which comes as an unavoidable side effectof the LPDA’s high gain, produces a waveform that is clearly distinguishable from thermal noise fluctuationsand from most anthropogenic radio pulses. The ARIANNA detector response to an Askaryan pulse, whichwe refer to as neutrino template in the following, is presented in Fig. 6.8 igure 6: ARIANNA detector response to a on-cone neutrino radio pulse (cf. Fig. 1). The LPDA antenna response wasevaluated 30 ◦ off the boresight direction which corresponds to the most likely signal arrival direction. The high frequencycomponents are at the beginning of the waveform whereas lower frequency components are delayed longer and are shiftedtowards the end of the waveform. Figure from (Persichilli, 2018). The similarity of a triggered event with the neutrino template is determined by means of the Pearsoncorrelation coefficient χ which can take values from 0 to 1, where 1 represents perfect correlation and valuesclose to 0 represent totally un-correlated waveforms.The similarity estimator χ is combined with an estimator of the signal strength, because high amplitudesignals are less effected by noise and are therefore expected to have a higher χ coefficient than a neutrinopulse that hardly sticks out of the noise. Hence, for signals with a high signal-to-noise ratio, we can requirea higher χ value. The optimal cut value is determined in a separate Monte-Carlo study in which a detectorsimulation is performed for a large library of simulated Askaryan signals. The resulting distribution of χ vs.the peak-to-peak amplitude (P2P) is shown in Fig. 7 left.This neutrino signal distribution is compared to the distribution of all triggered events. We analyzed alldata from Dec. 2015 through April 2017. The corresponding distribution from all HRA stations equippedwith a series 100 amplifier is shown in Fig. 7 right. The distribution of the stations equipped with a series200 amplifier look qualitatively the same. The bulk of the events is clearly separated and far away fromthe expected neutrino signal space, and none of the recorded events reaches the approx. 85% neutrinoefficiency line. This analysis demonstrates that already a simple template matching technique is a powerfuldiscriminator that leads to a good neutrino efficiency and purity.However, an important physical background to the template matching technique is the radio emissionof cosmic-ray air showers that is picked up by the in ice antennas. Their radio pulse is very similar to theexpected Askaryan signal and cosmic rays are several orders of magnitude more abundant than neutrinoswith a typical event rate of one per day per ARIANNA station. The initial cosmic-ray signal always comesfrom above and enters the (downward facing) LPDA antennas through their backlobe. This normally distortsthe signal strongly enough to be distinguishable from Askaryan signals that come from below and enter theantenna through its sensitive direction. However, for some very specific air-shower directions, the cosmic-raypulse might be confused with a neutrino signal, and a preliminary analysis actually observed one event thatis in the neutrino signal space. Therefore, the newer generation of ARIANNA stations is equipped with fouradditional upward facing LPDAs which allows a clear tagging of cosmic-ray events. A simple cut on theamplitude ratio between upward and downward facing antennas reduces the cosmic-ray background to 0.3neutrino candidates in 3 calendar years for an array with more than 1000 detector stations while preserving99.7% of the neutrino triggers (Barwick et al., 2015a). The ARIANNA HRA stations have two different types of amplifiers with a slightly different gain which are referred to asseries 100 and 200 amplifiers. Newer stations were deployed with the series 200 amplifiers. igure 7: (left) Distribution of expected neutrino Askaryan signals in the similarity parameter χ and peak-to-peak amplitude.The color shows the number of entries per bin. The area above the red curve contains 85% of the simulated neutrino signals.(right) Same as the left plot but for triggered events. Figure from (Persichilli, 2018).
5. Cosmic ray test beam The measurement of cosmic rays comes with many advantages apart from rejecting cosmic-ray signalsfrom the neutrino search. Cosmic rays are not only a background that we need to get rid of but also aperfect calibration source for a radio-neutrino detector because their radio pulses are very similar to theAskaryan pulses that we expect from neutrinos. Both are very short bipolar pulses of just a few nanosecondslength which are difficult to generate articially. Therefore, measuring cosmic rays is the only way to fullytest the neutrino detector under realistic conditions. Furthermore, the radio emission of air showers is wellunderstood so that the reconstructed signal properties can be verified by theoretical predictions.First, we demonstrate the performance of the template matching technique by using this techniqueto identify cosmic-ray signals out of the large sample of all triggered events (Barwick et al., 2017). A twodimensional cut in the correlation parameter χ and the signal amplitude leads to a clear separation of cosmicrays from the background as shown in Fig. 8 left. The measured event rate together with a simulation ofthe ARIANNA acceptance and the uptime of the detector was converted to a cosmic-ray flux and shownin Fig. 8 right. The measured flux agrees with the more precise measurements of other experiments withinuncertainties which is an indirect test that the cosmic-ray identification works successfully.Second, we demonstrate the ARIANNA sensitivity to the signal polarization by reconstructing the po-larization of the cosmic-ray events and comparing it with the theoretical expectation (Glaser, 2018). Thepolarization of cosmic-ray radio signals was measured extensively by dedicated radio cosmic-ray detectorssuch as AERA (Aab et al., 2014) and LOFAR (Schellart et al., 2014) and is very well understood theoretically(de Vries et al., 2010; Glaser et al., 2016, 2019b) which allows for a precise prediction.We use all data collected by the dedicated cosmic-ray station 32 during the 2017/2018 season and selectcosmic rays using the template-matching method described above. In total we find 265 cosmic-ray eventsfrom which 135 pass the more stringent quality cut of having a signal-to-noise ratio (SNR) larger than 4 inall four channels. (We note that a SNR of 4 is a relatively weak cut, the average SNR of a pure noise traceis 3.) This cut allows to reconstruct the cosmic-ray direction using the same method as described in Sec. 3. For completeness of this article, this section partly reviews previously published results on the detection and selection ofcosmic rays of (Barwick et al., 2017). The signal-to-noise ratio is defined as the maximum amplitude divided by the RMS noise. igure 8: (left) Correlation of measured events with a cosmic-ray template as a function of signal amplitude. The backgrounddensity map shows the probability distributions for simulated air showers given an amplitude. The markers show the averagecorrelation value χ of measured events. The line indicates a cut separating the cosmic ray signals from the background. Alldiamonds are background events, while the signals are indicated by circles. See (Barwick et al., 2017) for more details. (right)Cosmic-ray flux measured by ARIANNA in comparison with other experiments. Figures and captions from (Barwick et al.,2017). The direction is required to predict the expected signal polarization, and to evaluate the antenna responsefor the correct direction during the polarization reconstruction.We reconstruct the signal polarization using a novel forward folding method that is described in detailin (Glaser et al., 2019c) and briefly summarized here: The incident electric field, i.e., the cosmic-ray radiopulse, is related to the measured voltage in the antennas in Fourier space via V ( f ) V ( f ) ... V n ( f ) = H θ ( f ) H φ ( f ) H θ ( f ) H φ ( f ) ... H θn ( f ) H φn ( f ) (cid:18) E θ ( f ) E φ ( f ) (cid:19) , (1)where V i is the Fourier transform of the measured voltage trace of antenna i , H θ,φi represents the responseof antenna i to the φ and θ polarization of the electric field E θ,φ from the direction ( ϕ, ϑ ).Traditionally, this system-of-equation is solved for the electric field for each frequency bin. However,this unfolding comes with several downsides and often leads to an amplification of noise, in particular,for horizontal air showers and higher frequency bands where the cosmic-ray signal is small (see extendeddiscussion in (Glaser et al., 2019c) for more details). Instead, we use a novel forward folding technique:We describe the cosmic-ray radio pulse analytically with just four free parameters, and determine theseparameters in a chi-square minimization directly on the measured voltage traces. This is, the analyticelectric-field is folded with the antenna response to obtain the expected signal in all antennas. These voltagetraces are then directly compared to the measured voltage traces. The parameters of the electric-field pulseare optimized to obtain the smallest quadratic difference (the χ ) between the two traces for all channels.The polarization angle is then just given by the ratio of the two electric-field components. Applyingthis method on the cosmic-ray data set of station 32 and comparing it to the theoretical expectation,we find a resolution of the signal polarization of 14 ◦ . The expected polarization is calculated accordingto the dominant geomagnetic emission process and is given by the vector product between the showeraxis and the geomagnetic field axis. We expect the resolution to improve significantly with an improveddetector calibration. To study this, we performed an end-to-end MC simulation using a representative setof CoREAS simulations (Huege et al., 2013) and including signal distortion due to noise interference. For a11ell-calibrated detector station we find that we can achieve a polarization resolution of ∼ ◦ which is limitedby the signal-to-noise ratio.Although not the primary objective of ARIANNA, a direct contribution to ultra-high-energy cosmic-ray(UHECR) physics is foreseen. A large-scale ARIANNA detector with hundreds of stations will provide asubstantial exposure to measure cosmic rays with reasonable statistics up to energies of 10 eV (Barwicket al., 2017). We will be able to measure the energy spectrum with competitive and independent systematicuncertainties (Gottowik et al., 2018). Furthermore, a small subset of cosmic rays are expected to be observedin coincidence with IceTop and IceCube if the detector is deployed at the South Pole. These events willprovide an exceptional amount of information and will be important for cross-calibration purposes. Inaddition, the radio signal is only sensitive to the electromagnetic air-shower component whereas IceTop andespecially IceCube provides a complementary measurement of the hadronic shower component which allowsfor a measurement of the cosmic-ray mass (Holt et al., 2019). This is facilitated by a newly developedtechnique to determine the air-shower energy from a single radio detector station (Welling et al., 2019).Also, the radio signal of inclined air showers, where this technique works best, extends over a large area(Aab et al., 2018). Thus, a sparse spacing of detector stations of O (1 km) is sufficient to perform thismeasurement.
6. ARIA: Optimized surface radio neutrino detector
The versatile experiences of the ARIANNA pilot array in terms of hardware stability, deployment anddetector operation in the harsh Antarctic conditions, neutrino and cosmic-ray identification, and advanceddata reconstruction led to the design of the Askaryan Radio In-ice Array (ARIA). We determined the SouthPole as the optimal location because of the good infrastructure and the cold, deep and clear ice whichattenuates the radio signal less than the ice at the Moore’s Bay site and more than compensates for thehigher logistical effort and lack of reflective layer from the water-ice interface beneath the ice shelf at Moore’sBay.ARIA combines the advantages of the ARIANNA design (autonomous stations, ease of deployment, re-liable hardware, deployable at any ice sheet on the planet) with an optimized station layout that achievesa high sensitivity to neutrinos, an unambiguous identification of the neutrino signals with multiple comple-mentary channels, and the ability to reconstruct the neutrino direction and energy for almost all events.Because no high-energy neutrino has been detected yet with the radio technique, especially the second pointis of utmost importance. All these goals can be achieved with the station design presented in Fig. 9.The ARIA station comprises several LPDA antenna with different orientations near the surface and onestring of four dipoles deployed at a depth of 15 m. The vertically oriented LPDAs provide a large sensitivityto the expected neutrino signals as the antenna is most sensitive towards the expected signal direction andpolarization. The four downward pointing LPDAs measure the two horizontal polarization componentsand provide the necessary addition information to reconstruct the polarization. The four upward pointingLPDAs take care of cosmic-ray rejection. The station is then completed with four deep dipoles to improvethe neutrino identification and provide information on the distance to the neutrino interaction vertex viadetecting both the direct and reflected pulse (see below).The autonomous nature of the stations and the vicinity of the antennas to the snow surface allow fora quick and easy deployment with essentially no requirements for supporting infrastructure. The use of anewly developed portable cylindrical hole melter allows to drill the borehole for the dipoles within a fewhours and very little monitoring (Heinen et al., 2017; RWTH Innovation, 2018). It was already successfullyused in 2018 at the Moore’s Bay site. This concept is currently adapted to also melt the slots for the LPDAsto avoid hand digging of trenches. A complete ARIA station can be deployed within one working day withabout four to five people.Neutrinos can be identified using several complementary techniques: • Using a template matching technique as described in Sec. 4 where cosmic-ray signals can be separatelyrejected by the upward pointing LPDAs. 12 igure 9: Sketch of the ARIA station design. • The additional information of the signal arrival times from the deep dipoles allow to determine if thesignal originated from above, i.e., cosmic rays or anthropogenic background, or from below where theneutrino signals come from. • The deep dipoles measure two pulses, a direct and reflect pulse, which is unique signature of a neutrinosignal (D’n’R technique, see below).
The direct and reflected (D’n’R) technique provides a unique signature of a neutrino signal and providesimportant information on the distance to the neutrino interaction vertex: For most geometries, signalsoriginating from deep in the ice have two paths to an antenna. One direct path and one reflected path,i.e., the signal gets reflected off the surface and reaches the antenna from above. We studied this techniquein detail using the novel Monte Carlo code NuRadioMC (Glaser et al., 2019a) that models the signalpropagation through the firn via ray tracing and includes a realistic treatment of the reflection at thesurface using the complex Fresnel reflection coefficients. We find that for most neutrino events, the incidentangle at the surface is such that is results in total internal reflection. Hence, the ice-air interface acts as aperfect mirror.The optimal depth of the dipoles is a trade off between two competing effects. On the one hand, the timedifference between both pulses increases with depth which allows for a better differentiation and improvesthe resolution of the vertex distance. On the other hand, the efficiency to detect both pulses reduces withincreasing depth. This is because of the narrowness of the Cherenkov cone. The optimal depth is around15 m where the ARIA dipoles are located. At this depth we get time differences between both pulses of afew tens of nanoseconds which allows for a clear separation. Even at low neutrino energies of 10 eV, theD’n’R detection efficiency is still above 80%. At 10 eV the detection efficiency of both pulses is 95% andeven larger at higher neutrino energies. 13 igure 10: D’n’R pulser measurement at Moore’s Bay. (left) Geometry of the setup. The direct path (solid line) and reflectedpath (dashed line) are calculated using the ray-tracing technique. (right) The received waveform in the 8 . Experimental evidence exists that reflected pulses can be observed in addition to direct pulses. TheARA collaboration broadcasted radio signals from a 1400 m deep emitter to their receiving antennas at200 m depth and observed both pulses (Allison et al., 2019). The measured time difference is compatiblewith the propagation times of the two paths that were determined via ray tracing. We note that thismeasurement implies that the signal propagated twice through the complete firn, and that the bendingof signal trajectories in the firn is not hampering this technique. More recently, we performed a similarmeasurement at the Moore’s Bay site where a short broadband pulse was emitted at a depth of 37 . . MB exponentialindex-of-refraction profile that describes the available index-of-refraction data well (Barwick et al., 2017).Technically, we used the analytic ray-tracing code of NuRadioMC (Glaser et al., 2019a). We find that alreadyat 8 . To determine the neutrino direction, both the signal arrival direction as well as the polarization needto be measured. The signal arrival direction is reconstructed from the signal arrival times of all antennas(cf. Sec. 3). The polarization is reconstructed from the LPDA measurements using the same method as forcosmic rays (cf. Sec. 5). It is important to point out that the systematic uncertainty of this measurementwill be small because the same antenna type is used to cover all three polarizations. The LPDA antennasare just oriented into different directions to be sensitive to different signal polarizations. We also note thatthe forward folding technique can naturally use all available antennas. Hence, we expect an improvementof the polarization resolution due to the additional antennas of an ARIA station compared to results ofthe previous section obtained with only four antennas. However, the method still needs to be adapted toneutrino signals that have a different pulse form. In addition, the pulse forms are slightly different betweenantennas due to the narrowness of the Cherenkov cone. This will be addressed in an upcoming publication.The determination of the neutrino energy is the most challenging part and requires the measurement ofthe distance to the neutrino interaction vertex, the viewing angle, i.e., which part of the Cherenkov coneis observed, and the signal polarization. The vertex distance is determined from the time difference of thedirect and reflected pulse observed in the dipoles. The viewing angle is reconstructed from the frequencyspectrum of the Askaryan signal. Here, the broadband sensitivity of the LPDA antennas is beneficial. Hence,the ARIA design has a good sensitivity to the neutrino energy. The energy resolution is physically limited by the random fluctuations of how much energy is transferred from theUHE neutrino into the particle shower that produces the Askaryan signal. This limits the energy resolution to about σ [log ( E rec /E true )] ∼ .
35. A preliminary analysis indicates that the contributions to the energy resolution from the un-certainty in vertex distance, viewing angle, and polarization are smaller than the aforementioned limit. igure 11: Layout of the proposed ARIA detector. The black dots represents the locations of the ARIA stations. The filledred circles show the positions of the ARA stations. The blue hexagon show the position of the IceCube detector. It is important to point out that ARIA is sensitive to the neutrino properties of essentially all triggeredevents. Each neutrino signal will be seen in three orthogonal LPDA orientations, which allows for a mea-surement of the polarization and frequency spectrum, and most events will have a direct and reflect pulsein the dipoles, which allows for the reconstruction of the vertex distance.
ARIA consists of an array of 130 identical and independent neutrino stations, arranged in a hexagonalpattern and separated from each other by ∼ <
10% of the neutrino events with E = 10 eV are observed by 2 or more stations. Thoughthe observation of the same event by two independent stations provides a cross-check, it also reduces thesensitivity for a fixed number of stations. Since ARIA is a discovery instrument, it is useful to place ARIAstations as far apart as possible, keeping in mind deployment constraints. It requires less time to deploystations over a smaller area. One possible geometric arrangement of the array is shown in Fig. 11. The sensitivity of the ARIA detector is calculated using the state-of-the art NuRadioMC simulationcode (Glaser et al., 2019a) using the prediction of the Askaryan signal of (Alvarez-Mu˜niz et al., 2000). Theexpected sensitivity to both an isotropic neutrino flux as well as to transient sources is presented in Fig. 12.In 5 years of full-time operation, ARIA could limit the fraction of protons in the cosmic rays at the highestenergies to 10% or less for a standard choice of source evolution (van Vliet et al., 2019), an importantmilestone . If the diffuse flux of high energy neutrinos discovered by IceCube continues to higher energieswith a hard power law spectrum (dN/dE proportional to E − . ), ARIA will observe 12 events in 3 calendaryears of full-time operation. It could be the first detection of a neutrino with energy > × GeV.One potentially transformative method to understand the ultra-high-energy universe involves neutrinoemission in coincidence with gravitational wave and/or electromagnetic emission. ARIA contributes totransient neutrino astronomy by virtue of it large instantaneous aperture, and broad field of view of Ω =2 . > GeV. This is calculated as follows: We use the model of (van Vliet et al., 2019) of the cosmogenic neutrino flux for a sourceevolution parameter of m = 3 .
4, a spectral index of the injection spectrum of α = 2 .
5, a cut-off rigidity of R = 100 EeV,and a proton fraction of 10% at E = 10 . eV. We then calculate after how much lifetime, 2.44 neutrinos will be observedwith the full ARIA detector. Assuming a non-observation and a zero background contribution, we can turn this into a 90%Feldman-Cousins confidence upper limit on the proton fraction of cosmic rays. igure 12: Expected sensitivity of the proposed ARIA detector in one-decade energy bins calculated using NuRadioMC. (left)ARIA sensitivity to an isotropic flux for 3, 5 and 10 years of operation assuming a uptime of 100%. The shaded band of the3 years expected limit represents uncertainties in the analysis efficiency. Also shown is the measured astrophysical neutrinoflux from IceCube using the high-energy starting event (HESE) selection (Kopper, 2017) and using a muon neutrino sample(C.Haack and Wiebusch, 2017), limits from existing experiments (IceCube (Aartsen et al., 2018a), Auger (Aab et al., 2015)and Anita (Gorham et al., 2018)), the expected sensitivity of the proposed GRAND10k detector (Alvarez-Muniz et al., 2018),and several theoretical models. (right) ARIA sensitivity to transient sources if in field-of-view (solid blue line). Also shownis the sensitivity of current experiments (Albert et al., 2017) and a theoretical model of a high-energy neutrino flux of aneutron-neutron star merger (Fang and Metzger, 2017).
7. Conclusions
The ARIANNA pilot array paves the way for a large-scale radio detector to discover cosmogenic neutrinos.The robust ARIANNA hardware has proven to work reliably in harsh Antarctic conditions. The dataacquisition system with its precise time synchronization will enable the reconstruction of the telltale radiosignatures of a high-energy neutrino interaction.The ARIANNA stations run completely autonomous with solar power through the summer and witha newly developed wind generator through most of the dark winter months. The development of a wind-power system that works at extreme cold temperatures is a great achievement as commercial systems failquickly. In 2018, the first prototype survived the winter and was able to power an ARIANNA station for asubstantial amount of time. This design is now being further improved to reach a higher wind yield. Hence,fully autonomous stations are a viable option for a large scale Askaryan detector at the South Pole. Theautonomous nature of the design even allows for an installation at any deep ice location on the planet, animportant goal to reach full sky coverage in the future. Furthermore, the real-time transmission of datavia the Iridium satellite network and corresponding ultra-high energy neutrino alerts will contribute to theexiting multi-messenger effort to identify point sources.We demonstrated that radio signals originating from deep in the ice can be measured using a surfacestation via an in-situ calibration measurement, an important proof-of-concept. An ARIANNA station is ableto reconstruct the predicted signal direction within a few degrees. This also shows that the ice propertiesand the signal propagation through the ice is well understood to correct for the bending of signal trajectoriesin the firn.Neutrino signals can be identified with high efficiency and purity using a template matching techniquethat exploits the distinctive detector response to an impulsive signal. We used the more abundant cosmic-ray radio signals for an in-situ calibration and demonstration of the detector capabilities. We applied the16emplate matching technique to obtain a pure sample of cosmic-ray events and found that the reconstructedpolarization is in good agreement with the theoretical expectation.Finally, guided by five years of successful operational experience with the ARIANNA pilot array, alarge-scale high energy neutrino detector – called ARIA – was designed and proposed (Barwick, 2018).The station design was optimized for ice conditions at the South Pole. Neutrinos are distinguished fromhigh rate background processes by several complementary channels, which is of utmost importance for adiscovery instrument. ARIA will have unprecedented sensitivity to high-energy neutrinos. It it sensitive toa proton fraction of as low as 10% in the ultra-high energy cosmic-ray composition and will measure severalastrophysical neutrinos per year if the flux observed by the IceCube detector continues to higher energies.Furthermore, the ARIA station is designed to have a good sensitivity to the neutrino direction and energywhich enables multi-messenger astronomy at the highest neutrino energies.
Acknowledgements
We are grateful to the U.S. National Science Foundation-Office of Polar Programs, the U.S. NationalScience Foundation-Physics Division (grant NSF-1607719) and the U.S. Department of Energy. We thankgenerous support from the German Research Foundation (DFG), grant NE 2031/2-1 and GL 914/1-1, theTaiwan Ministry of Science and Technology. H. Bernhoff acknowledge support from the Swedish Govern-ment strategic program Stand Up for Energy. E. Unger acknowledge support from the Uppsala universityVice-Chancellor’s travel grant (sponsored by the Knut and Alice Wallenberg Foundation) and the C.F.Liljewalch travel scholarships. D. Besson and A. Novikov acknowledge support from the MEPhI AcademicExcellence Project (Contract No. 02.a03.21.0005) and the Megagrant 2013 program of Russia, via agreement14.12.31.0006 from 24.06.2013
ReferencesReferences
Aab, A., Abreu, P., Aglietta, M., et al. (The Pierre Auger Collaboration), 2014. Probing the radio emission from air showerswith polarization measurements. Phys. Rev. D 89, 52002. doi: .Aab, A., Abreu, P., Aglietta, M., et al. (The Pierre Auger Collaboration), 2015. An improved limit to the diffuse flux of ultra-high energy neutrinos from the Pierre Auger Observatory. Phys. Rev. D 91, 092008. doi: , arXiv:1504.05397 .Aab, A., Abreu, P., Aglietta, M., et al. (Pierre Auger Collaboration), 2018. Observation of inclined EeV air showers with theradio detector of the Pierre Auger Observatory. J. Cosmol. Astropart. Phys. 10(2018)026 doi: , arXiv:1806.05386 .Aartsen, M.G., Abraham, K., Ackermann, M., et al. (IceCube Collaboration), 2016. Constraints on ultrahigh-energy cosmic-raysources from a search for neutrinos above 10 PeV with IceCube. Phys. Rev. Lett. 117, 241101. doi: .Aartsen, M.G., Ackermann, M., Adams, J., et al. (IceCube Collaboration), 2018a. Differential limit on the extremely-high-energy cosmic neutrino flux in the presence of astrophysical background from nine years of IceCube data. Phys. Rev. D 98,062003. doi: , arXiv:1807.01820 .Aartsen, M.G., Ackermann, M., Adams, J., et al. (The IceCube Collaboration), 2018b. Multimessenger observations ofa flaring blazar coincident with high-energy neutrino IceCube-170922A. Science 361, issue 6398, eaat1378 doi: , arXiv:1807.08816 .Abbott, B.P., Abbott, R., Abbott, T.D., et al., 2017. GW170817: Observation of Gravitational Waves from a Binary NeutronStar Inspiral. Phys. Rev. Lett. 119, 161101. doi: , arXiv:1710.05832 .Albert, A., Andre, M., Anghinolfi, M., et al., 2017. Search for High-energy Neutrinos from Binary Neutron Star MergerGW170817 with ANTARES, IceCube, and the Pierre Auger Observatory. Astrophys. J. 850, L35. doi: , arXiv:1710.05839 .Allison, P., Archambault, S., Auffenberg, J., et al. (ARA collaboration), 2019. Measurement of the real dielectric permittivity (cid:15) r of glacial ice. Astroparticle Physics 108, 63–73. doi: , arXiv:1712.03301 .Allison, P., Bard, R., Beatty, J.J., et al., 2016. Performance of two Askaryan Radio Array stations and first results in thesearch for ultrahigh energy neutrinos. Phys. Rev. D 93, 082003. doi: , arXiv:1507.08991 .Alvarez-Muniz, J., Batista, R.A., Balagopal, A., et al., 2018. The giant radio array for neutrino detection (GRAND): Scienceand design arXiv:1810.09994 .Alvarez-Mu˜niz, J., V´azquez, R.A., Zas, E., 2000. Calculation methods for radio pulses from high energy showers. Phys. Rev.D 62. doi: . lvarez-Mu˜niz, J., Zas, E., 1998. The LPM effect for EeV hadronic showers in ice: Implications for radio detection ofneutrinos. Phys. Lett. Sect. B Nucl. Elem. Part. High-Energy Phys. 434, 396–406. doi: , arXiv:astro-ph/9806098 .Askaryan, G.A., 1962. Excess negative charge of an electron-photon shower. Sov. Phys. JETP 14, 441.Barwick, S.W., 2018. Antarctic Radio Ice Array (ARIA). submitted to the National Science Foundation .Barwick, S.W., Berg, E.C., Besson, D.Z., et al. (ARIANNA collaboration), 2015a. A first search for cosmogenic neutrinos withthe ARIANNA Hexagonal Radio Array. Astropart. Phys. 70, 12–26. doi: .Barwick, S.W., Berg, E.C., Besson, D.Z., et al. (ARIANNA collaboration), 2015b. Design and performance of the ARIANNAHRA-3 neutrino detector systems. IEEE Transactions on Nuclear Science 62, 2202–2215. doi: , arXiv:http://arxiv.org/abs/1410.7369 .Barwick, S.W., Berg, E.C., Besson, D.Z., et al., 2018. Observation of classically ‘forbidden’ electromagnetic wave propagationand implications for neutrino detection. J. Cosmol. Astropart. Phys. 07(2018)055 doi: , arXiv:1804.10430 .Barwick, S.W., Besson, D.Z., Burgman, A., et al. (ARIANNA collaboration), 2017. Radio detection of air showers with theARIANNA experiment on the Ross Ice Shelf. Astropart. Phys. 90, 50–68. doi: .Casey, K., Fudge, T., Neumann, T., Steig, E., Cavitte, M., Blankenship, D., 2014. The 1500 m south pole ice core: recovering a40 ka environmental record. Annals of Glaciology 55, 137–146. URL: http://spicecore.org/ , doi: .C.Haack, Wiebusch, C., 2017. A measurement of the diffuse astrophysical muon neutrino flux using eight years of IceCubedata., in: Proceedings of 35th International Cosmic Ray Conference — PoS(ICRC2017), Busan, South Korea, Sissa Medialab.doi: .Chiem, E.Y.W., 2017. Multi-Gigahertz Synchronous Sampling and Triggering (SST) Circuit with Picosecond Timing Resolu-tion. Ph.D. thesis. University of California, Irvine, USA.Fang, K., Metzger, B.D., 2017. High-Energy Neutrinos from Millisecond Magnetars formed from the Merger of Binary NeutronStars. Astrophys. J. 849, 153. doi: , arXiv:1707.04263 .Feng, J.L., Fisher, P., Wilczek, F., Yu, T.M., 2002. Observability of earth-skimming ultrahigh energy neutrinos. Phys. Rev.Lett. 88, 161102. doi: .Gerhardt, L., Klein, S., Stezelberger, T., Barwick, S., Dookayka, K., Hanson, J., Nichol, R., 2010. A prototype station forARIANNA: A detector for cosmic neutrinos. Nuclear Instruments and Methods in Physics Research Section A: Accelerators,Spectrometers, Detectors and Associated Equipment 624, 85–91. doi: .Glaser, C., 2018. ARIANNA: Measurement of cosmic rays with a radio neutrino detector in Antarctica, in: Proc. of 8thARENA 2018, Catania, Italy. arXiv:1811.10661 .Glaser, C., Erdmann, M., H¨orandel, J.R., Huege, T., Schulz, J., 2016. Simulation of Radiation Energy Release in Air Showers.J. Cosmol. Astropart. Phys. 09(2016)24 doi: .Glaser, C., Garc´ıa-Fern´andez, D., Nelles, A., et al., 2019a. NuRadioMC: simulating the radio emission of neutrinos frominteraction to detector. submitted to The European Physical Journal C URL: https://github.com/nu-radio/NuRadioMC/ , arXiv:1906.01670 .Glaser, C., de Jong, S., Erdmann, M., H¨orandel, J.R., 2019b. An analytic description of the radio emission of air show-ers based on its emission mechanisms. Astroparticle Physics 104, 64–77. doi: , arXiv:1806.03620 .Glaser, C., Nelles, A., Plaisier, I., Welling, C., Barwick, S.W., Garca-Fernndez, D., Gaswint, G., Lahmann, R., Persichilli, C.,2019c. Nuradioreco: A reconstruction framework for radio neutrino detectors. The European Physical Journal C 79, 464.doi: , arXiv:1903.07023 .Gorham, P.W., Allison, P., Banerjee, O., et al., 2018. Constraints on the diffuse high-energy neutrino flux from the third flightof ANITA. Phys. Rev. D 98, 022001. doi: , arXiv:1803.02719 .Gottowik, M., Glaser, C., Huege, T., Rautenberg, J., 2018. Determination of the absolute energy scale of extensive air showersvia radio emission: systematic uncertainty of underlying first-principle calculations. Astroparticle Physics 103, 87–93.doi: , arXiv:1712.07442 .Heinen, D., Eliseev, D., Henke, C., Jeschke, S., Linder, P., Reuter, S., Schnitz, S., Scholz, F., Weinstock, L.S., Wickmann,S., Wiebusch, C., Zierke, S., 2017. EnEx-RANGE - Robust autonomous Acoustic Navigation in Glacial icE. Proc. of 7thARENA 2016, EPJ Web of Conferences 135, 06007. doi: .Holt, E.M., Schr¨oder, F.G., Haungs, A., 2019. Enhancing the cosmic-ray mass sensitivity of air-shower arrays by combiningradio and muon detectors. The European Physics Journal C 79, 371. doi: .Hou, G.W., Huang, M., 2002. Expected performance of a neutrino telescope for seeing AGN/GC behind a mountain, in: Proc.of first NCTS workshop Astroparticle Physics, WORLD SCIENTIFIC. pp. 105–116. doi: .Huege, T., Ludwig, M., James, C.W., 2013. Simulating radio emission from air showers with CoREAS, in: Proc. of 5th ARENA2012, AIP Conference Series 1535, pp. 128–132. doi: .Kleinfelder, S.A., Chiem, E., Prakash, T., 2014. The SST fully-synchronous multi-ghz analog waveform recorder with nyquist-rate bandwidth and flexible trigger capabilities, in: Proc. of 2014 IEEE Nuclear Science Symposium and Medical ImagingConference (NSS/MIC), Seattle, WA, 2014, pp. 1-3, IEEE. doi: , arXiv:1505.07085 .Kopper, C., 2017. Observation of astrophysical neutrinos in six years of icecube data, in: Proc. of 35th International CosmicRay Conference, Busan, South Korea — PoS(ICRC2017)981. doi: .Kravchenko, I., Besson, D., Meyers, J., 2004. In situ index-of-refraction measurements of the south polar firn with the RICEdetector. Journal of Glaciology 50, 522–532. doi: .arm Mbed, . Mbed platform and operating system. URL: .Nam, J., Liu, T.C., 2017. Feasibility of antenna array experiment for earth skimming tau-neutrino detection in antarctica, in: roc of 35th International Cosmic Ray Conference, Busan, South Korea — PoS(ICRC2017)944. doi: .Persichilli, C.R., 2018. Performance and Simulation of the ARIANNA Pilot Array, with Implications for Future Ultra-highEnergy Neutrino Astronomy. Ph.D. thesis. University of California, Irvine. URL: https://escholarship.org/uc/item/1340083p .RWTH Innovation, 2018. Patent: High efficiency head for fast melting probes. URL: .Schellart, P., Buitink, S., Corstanje, A., et al., 2014. Polarized radio emission from extensive air showers measured with LOFAR.J. Cosmol. Astropart. Phys. 10(2014)014 doi: , arXiv:1406.1355 .van Vliet, A., Batista, R.A., H¨orandel, J.R., 2019. Determining the fraction of cosmic-ray protons at ultra-high energies withcosmogenic neutrinos arXiv:1901.01899 .de Vries, K.D., Scholten, O., Werner, K., 2010. Macroscopic Geo-Magnetic Radiation Model; Polarization effects and finitevolume calculations. Proc. of ARENA 2010, Nucl. Instrum. Meth. A 662, S175–S178.Wang, S.H., 2017. Calibration, performance, and cosmic ray detection of arianna-hcr prototype station, in: Proc. of 35thInternational Cosmic Ray Conference, Busan, South Korea — PoS(ICRC2017)558. doi: .Welling, C., Glaser, C., Nelles, A., 2019. Reconstructing the cosmic-ray energy from the radio signal measured in one singlestation. submitted to J. Cosmol. Astropart. Phys. arXiv:1905.11185 ..