Searching for Neutrino Radio Flashes from the Moon with LOFAR

0.27 MB·English

by  Stijn Buitink

Searching for Neutrino Radio Flashes from the Moon with
LOFAR
Stijn Buitink∗,†, Arthur Corstanje†, Emilio Enriquez†, Heino Falcke†,∗∗, Wilfred
Frieswijk∗∗, Jörg Hörandel†,‡, Maaijke Mevius∗∗, Anna Nelles†,‡, Satyendra
Thoudam†, Pim Schellart†, Olaf Scholten∗, Sander ter Veen†, Martin van den
Akker† and the LOFAR collaboration∗∗
3
1
0 ∗KernfysischVersnellerInstituut,9747AAGroningen,TheNetherlands
2 †Dept.ofAstrophysics/IMAPP,RadboudUniversityNijmegen,6500GLNijmegen,TheNetherlands
n ∗∗NetherlandsInstituteforRadioAstronomy(ASTRON),7990AADwingeloo,TheNetherlands
a ‡Nikhef,ScienceParkAmsterdam,1098XGAmsterdam,TheNetherlands
J

2 Abstract. Ultra-high-energyneutrinosandcosmicraysproduceshortradioflashesthroughtheAskaryaneffectwhenthey
2 impactontheMoon.EarthboundradiotelescopescansearchtheLunarsurfaceforthesesignals.Anewgenerationoflow-
frequency,digitalradioarrays,spearheadedbyLOFAR,willallowforsearcheswithunprecedentedsensitivity.Inthefirst

] stageoftheNuMoonproject,low-frequencyobservationswerecarriedoutwiththeWesterborkSynthesisRadioTelescope,
M
leading to the most stringent limit on the cosmic neutrino flux above 1023 eV. With LOFAR we will be able to reach a
I sensitivityofoveranorderofmagnitudebetterandtodecreasethethresholdenergy.
.
h Keywords: neutrinos,cosmicrays,radiotelescopes,Moon,radioemission
p PACS: 95.85.Ry,95.55.Jz,95.55.Vj,96.20.-n
-
o
r 1. INTRODUCTION directly. The detection of UHE neutrinos that are pro-
t
s ducedinGZKinteractionsordirectlyinthesourceisan
a
A century after their discovery, cosmic rays (CRs) still attractivealternativetofindthesesources.Neutrinosare
[
representoneoftheunsolvedquestionsinastrophysics. notaffectedbymagneticfields,andtheycantravelover

1 WhilemodernexperimentslikethePierreAugerObser- cosmicdistancesalmostunattenuated.
v vatory (Auger) [1], HiRes [2] and the Telescope Array It was first suggested by Dagkesamanskii and
5
(TA) [3] can efficiently detect ultra-high-energy (UHE) Zheleznykh [9] to use the Moon as a detector for
8
CRs,twocomplicationsremainthatmakeithardtoiden- UHEneutrinosandCRs.AtenergiesaboveseveralPeV
1
5 tifytheirsources.First,thetrajectoriesofCRsarebentin the Moon is opaque to neutrinos. Neutrinos impinging
. the (inter-)Galactic magnetic fields, so the arrival direc- on the Moon will interact below the lunar surface. In
1
tionsoftheCRsatEarthdonotalignwiththepositions a charged current interaction ∼ 20% of the energy is
0
3 oftheirsources.Thisdeflectiondependsonthechargeof converted into a hadronic cascade. The leptons that are
1 the CRs, and therefore on their composition. An Auger createdareinprincipledetectable,thoughinthecaseof
: studyofpossiblecorrelationwithAGNs[1]disfavorsCR muons and taus the relevant cross-sections at UHE are
v
isotropy above 6·1019 eV. However, composition stud- highly uncertain. Electrons initiate an electromagnetic
i
X ies by Auger show a trend towards a heavier compo- showerwhichishardtodetectwiththelunarCherenkov
r sition at the highest energies [4], which suggests large techniquebecauseitiselongatedduetotheLPMeffect
a
separations between CR arrival directions and sources. [10].However,atsufficientlyhighenergy,thelengthand
Incontrast,HiResdatadoesnotrejectisotropyandsug- depth of the cascade is reduced because of the increase
gestsalightCRcomposition[5].Asecondcomplication of the photonuclear and electronuclear cross sections
is that CRs above an energy of 4·1019 eV will interact [11]. Particle cascades beneath the lunar surface can be
withthecosmicmicrowavebackground(CMB).Inthese detectedbytheradiationtheyemitthroughtheAskaryan
Greisen-Zatsepin-Kuzmin (GZK) interactions pions are effect[12].
created,whichproduceneutrinoswhentheydecay[6,7]. GLUE[13]wasoneofthepioneeringexperimentsat
AsteepeningoftheCRspectrumatthisenergyhasbeen highfrequencies(>1GHz).Presently,LUNASKA[14]
found experimentally [8, 2]. The attenuation length of performshighfrequencymeasurementswithATCAand
protonsduetotheGZKeffectis∼50Mpc.Sourcesout- theParkesradiotelescope.AttheExpandedVeryLarge
side this radius cannot be found by studying UHE CRs Array, RESUN [15] performs 1.4 GHz lunar observa-
tions.
TheNuMoonprojectobservesinalowfrequencywin-
dowwhichoffersanoptimaldetectionprobabilityatthe
ν
highest energies. In the first phase, observations were
performed with the Westerbork Synthesis Radio Tele-
scope(WSRT)leadingtothemostconstraininglimiton
θ
theUHEneutrinofluxabove1023eV[16,17].Presently, ν
we are preparing observations with the Low Frequency
Array(LOFAR)[24].
b
2. RADIOEMISSIONFROMLUNAR
PARTICLECASCADES
Neutrinos and CRs impacting on the Moon initiate cas-
cades below the surface, which develop a time-varying
negativechargeexcesspropagatingatthespeedoflight FIGURE 1. Geometry: θν is defined as the angle between
thedirectionoftheneutrinoandtheobserver;bistherelative
invacuum,andtherebyproduceradioemission[12].Al-
distance of the point where the pulse emerges from the lunar
thoughtheradiationismostintenseattheCherenkovan- surfacetothecenterofthefaceoftheMoon.
gle, it can be misleading to think of it in terms of pure
Cherenkov radiation. The rise and decay of a moving
chargeitselfwouldalsoproduceradiationiftheindexof
refraction of the medium were unity [18, 19]. Actually, b
at low frequencies, when the wavelength is larger than
0.8
thecascadelength,theradiationiscoherentatanglesfar
10-3
fromtheCherenkovangle[20].
The spread of the radiation around the Cherenkov 0.6
angle is of large importance for the detectability of the
radio pulse. To observe the pulse on Earth, it needs to 0.4 10-4
beabletoescapetheMoon.SincetheCherenkovangle
is equal to the angle of total internal reflection on the 0.2
surface, the high frequency component can only leave 10-5
the Moon if the shower points upwards to the surface. 0
21 22 23 24
log(E)
This is only the case when a particle enters the Moon
from behind, near the lunar rim. The low frequency
FIGURE2. Distributionofradiopulseswithpowergreater
component,ontheotherhand,isspreadoutoveramuch
than500Jyasafunctionofenergyandb.
largeropeningangle,andpartoftheradiationcanescape
theMoonevenifthecascadeisdirecteddownwardsinto
theMoon.At∼150MHzthewholevisiblesurfaceofthe
Mooncontributestotheeffectivedetectorvolume[21]. 180 10-2
n
q
We simulated the distribution of observable pulses 160
overthelunarsurfaceforanisotropicneutrinofluxwith 140
anenergyspectrumE−1intherange1021−1024eV.We
120 10-3
assumethat20%oftheneutrinoenergyisdepositedina
100
hadroniccascadeforallflavors.Theneutrinocrosssec-
80
tions [22] are extrapolated to high energies for a 500 m
deep layer of regolith, and the radio pulse parametriza- 60 10-4
tionofAlvarez-Muñizetal.[23]isusedforthehadronic 40
showers. As a mean value for the attenuation length for 20
the radiated power we have taken λ =(9/ν[GHz]) m. 0
Surfaceroughnessisnotincludedsinceitdoesnotplay 0 0.2 0.4 0.6 0.8 b
animportantroleatlowfrequencies[21].
FIGURE3. Distributionofradiopulseswithpowergreater
The geometry definitions used here are sketched in
Fig.1:θ isdefinedastheanglebetweenthedirectionof than500Jyasafunctionofθν andb.
ν
theneutrinoandtheobserver;bistherelativedistanceof
thepointwherethepulseemergesfromthelunarsurface
tothecenterofthefaceoftheMoon(b=1corresponds
totherimoftheMoon).
In Fig. 2 the distribution of radio pulses above a
thresholdof500Jyinthe110−190MHzbandisplotted
asafunctionofneutrinoenergyandb.Theenergyspec-
trumhasbeenreweightedtoE−2 andthetotalobserved
flux has been normalized to unity. The color scale thus
representstherelativeprobabilitiesofobservinganevent
in different parts of the parameter space. The threshold
energy for detection is ∼1021 eV and at one decade in
energyhigherthewholesurfaceoftheMooncontributes
FIGURE 4. Arial view of the LOFAR core. The island,
to the detector volume. Even at higher energies events
called the superterp contains 6 of the 24 core stations. It is
aremorelikelytooccurclosetotherim.Thisismainly dividedinlargerandsmallerpatcheswhichhousetheLBAand
a projection effect: the high-b region represents a very HBAantennasrespectively.
largepartofthevisiblehalfofthelunarsphere.
In Fig. 3 the same set of events is now plotted as a
function of θ and b. There is a strong clustering of nal is transported to the central processor where it is
ν
eventstowardstherimoftheMoonwithanarrivalangle firstcorrectedforionosphericdispersion.AtLOFARfre-
θ in the range 110◦−150◦. Neutrinos from 50◦−90◦ quencies, a typical ionospheric electron content of 10
ν
can produce observable pulses on the complete lunar TEC-units (TECU=1016 electron/m2) causes a disper-
surface,whileforevensmallervaluesofθ ,observable sionwhichspreadsabandwidth-limitedsignalpulseover
ν
events are rare. Hence, localization of a low-frequency hundredsofnanoseconds[25].Inordertobeabletoper-
radio pulse on the Moon provides constraints on the form a pulse search the signal needs to be de-dispersed
arrivaldirectionoftheneutrino. withanaccuracyof1TECU.Aninterestingapproachis
measuring the absolute TEC value via the Faraday ro-
tation of polarized light due to ionospheric plasma and
3. NUMOONWITHLOFAR the Earth magnetic field. We are currently investigating
thepossibilitytousethepolarizedlightoftherimofthe
LOFAR[24]isanewkindofradiotelescopeconsisting Moonforthispurpose.
of thousands of simple omni-directional antennas that In the next step, the de-dispersed station beams are
takesadvantageofrecentdevelopmentsinfastelectron- combined to form array beams of ∼0.1 degree width.
ics.Thereceivedradiosignalisdigitizedateachantenna Upto50suchbeamscanbesynthesizedsimultaneously
andsenttoacentralprocessorwheredirectedbeamsare tocoverthecompletesurfaceoftheMoon.Eachofthese
synthesizedbyapplyingtheappropriatetimedelays.Be- beams will be searched in real-time for short pulses.
cause the antennas themselves contain no moving com- When a candidate pulse is found in one of the beams a
ponentstheyarerelativelycheapandcanbeconstructed coincidencecheckwillbeperformed.Aradiopulseorig-
inlargeamounts.Anothergreatadvantageofsynthesiz- inating from a certain location on the Moon is strong
ingbeamsthroughsoftwareisthatmultiplebeamsindif- in only one of the array beams (although neighboring
ferentdirectionscanbeinoperationatthesametime. beams may also receive a signal, depending on beam
Each LOFAR station consists of 96 Low Band An- shape and sidelobes). This is a powerful vetoing tech-
tennas (LBA), which operate in the 10−80 MHz range, niquethatremovespulsesfromstrongnoisesourcesnear
and 48 High Band Antennas (HBA), which cover the oneofthestations:averystrongnoisepulseinoneofthe
110−240MHzband.Currently,thearrayconsistsof24 stationbeamswillcauseaspikeinallthearraybeams.
corestations,concentratedinanareaof∼12km2,9re- Each of the LOFAR antennas is connected to a ring
mote stations, distributed over the northern part of the bufferthatstoresthelastfivesecondsofrawdata.When
Netherlands with baselines up to 80 km from the core, a pulse is found at the central processor that passes the
and8internationalstations. anti-coincidence veto, a trigger will be sent back to the
FortheNuMoonexperimentwewillusetheHBAan- stations and the buffers are read out and permanently
tennas operating in the 110−190 MHz window. The stored. This allows us to use the full resolution, full
beamforming is done in two stages because of limited bandwidthrawtime-seriesdataofallantennasforoffline
bandwidth between stations and the central processor. analysis[26].
At each station a beam of a few degrees opening an-
gle is formed in the direction of the Moon. This sig-
4. SENSITIVITY 5. CONCLUSIONS
In the past years, much progress had been made in im- TheNumoonexperimentusestheMoonasaultra-high-
proving the sensitivity to cosmic neutrinos over a large energy cosmic-ray and neutrino detector. It is different
range of energy, with a large range of techniques. In fromotherlunarAskaryanexperimentsasitsearchesfor
Fig.5existingupperlimitsontheall-flavorneutrinoflux radioflashesatlowerfrequencies(110−190MHz).This
are plotted in solid lines. The most stringent limits are has the advantage that a much higher sensitivity can be
setbyIceCube[27],ANITA[28]andtheWSRT[16]at reached, although the threshold energy is slightly lower
increasinglylargerenergies.Alsoplottedarelimitspub- at high frequencies (compare for example the sensitivi-
lishedbyRice[29]andAuger[30]. tiesoftheSKAlowandmiddlebandinFig.6).Observa-
Theexpectedsensitivitythatisreachedwithoneweek tionswiththeWSRThaveledtothemoststringentupper
ofLOFARdatawillimprovethepreviousWSRTlimitby limitsabove1023eV.Preparationsarenowinprogressto
anorderofmagnitude,whilealsodecreasingthethresh- performmeasurementswithLOFAR.Theexpectedsen-
olddetectionenergy.Inthefuture,theSquareKilometer sitivityisoveranorderofmagnitudebetterthanprevious
Array (SKA) [31] will provide an even better sensitiv- measurements.Inthefuture,theSKAwillmakeanother
ity using the same techniques as LOFAR. At lower en- bigleapinsensitivityusingthesametechniques.
ergies, better sensitivities will be reached by future ex-
periments ARA [32] and ARIANNA [33] (not plotted),
both searching for radio flashes of neutrinos in ice, and ACKNOWLEDGMENTS
theJEM-EUSOsatellite[34].
ThefluxofGZKneutrinosissensitivetovariousun- This work was supported by the Netherlands Organiza-
knownparametersincludingsourceevolution,CRcom- tion for Scientific Research (NWO), VENI grant 639-
positionandsourceenergycutoff.Figure5containspre- 041-130, the Stichting voor Fundamenteel Onderzoek
dictionscomputedby[35]basedonvaryinginputparam- derMaterie(FOM),theSamenwerkingsverbandNoord-
eters. The shaded band represents a region which con- Nederland (SNN) and the Netherlands Research School
tainsalargepartofthepossiblemodels.Themodelthat for Astronomy (NOVA). LOFAR, the Low Frequency
produces the largest neutrino flux is indicated by dash- ArraydesignedandconstructedbyASTRON,hasfacili-
dottedline. tiesinseveralcountries,thatareownedbyvariousparties
UHE neutrinos can also be created in the decay of (eachwiththeirownfundingsources),andthatarecol-
supermassive particles (top-down scenarios). The TD- lectivelyoperatedbytheInternationalLOFARTelescope
curve in Fig. 5 delimits the parameter space in which (ILT)foundationunderajointscientificpolicy.
neutrinos may be found originating from moduli —
weakly coupled scalar particles predicted in supersym-
metrictheories—fromthekinksofcosmicstringloops REFERENCES
[36,37].
AcascadeinducedbyaCRwillbeinitiatedjustbelow 1. J.Abrahametal.[AUGERcollaboration],Science318,
the lunar surface in contrast to neutrino induced show- 938(2007).
ers. Therefore, the radio signal suffers less attenuation. 2. R.Abbasietal.[HiRescollaboration],Phys.Rev.Lett.100,
101101(2008).
However,thequestionriseswhetherornottheradiative
3. T.Abu-Zayyadetal.[TAcollaboration],ApJ 757,26
field can properly develop in a volume that is smaller
(2012).
than the wavelength of the radiation. A detailed deriva- 4. J.Abrahametal.[AUGERcollaboration],Phys.Rev.Lett.
tion [18, 19] shows that such a formation zone is not 104,091101(2010).
needed. This becomes evident when one realizes that a 5. R.Abbasietal.[HiRescollaboration],Phys.Rev.Lett.104,
system of growing and decaying current would also ra- 161101(2010).
6. K.Greisen,Phys.Rev.Lett.16,748(1966).
diate if it were to develop in a vacuum. Hence, the di-
7. G.ZatsepinandV.Kuzmin,JETPLett.4,78(1966).
electricmediumisnotessentialfortheformationofthe
8. J.Abrahametal.[AUGERcollaboration],Phys.Rev.Lett.
radiative field, although it certainly affects it. In Fig. 6 101,061101(2008).
theWSRTupperlimitonCRflux[18]isplotted,aswell 9. R.DagkesamanskiiandI.Zheleznykh,PismaZhurnal
asexpectedsensitivitiesforLOFARandSKA(lowband EksperimentalnoiiTeoreticheskoiFiziki50,233(1989).
andmiddleband).Thedatapointsrepresentthehighend 10. J.Alvarez-Muñiz andE. Zas, Phys. Lett.B 434, 396
(1998).
oftheCRspectrumasmeasuredbyAuger[8].
11. L.GerhardtandS.Klein,Phys.Rev.D82,074017(2010).
12. G.Askaryan,Sov.Phys.JETP14,441(1962).
13. P.Gorhametal.,Phys.Rev.Lett.93,41101(2004).
14. C.Jamesetal.,Phys.Rev.D81,042003(2010).
1
-sr Auger Rice
1 WSRT
2- s 10-6
-m
c
V IceCube Anita
e
G 10-7 TD

LOFAR
En JEM-EUSO
d
/
Nn ARA-37
d
10-8
2
En
SKA
GZK
10-9
8 9 10 11 12 13 14 15 16
log E [GeV]
n
10
FIGURE5. All-flavourneutrinoupperlimitsandfuturesensitivitiesofseveralexperiments.Seetextfordetails.
15. T.Jaeger,R.MutelandK.Gayley,Astropart.Phys.34,
293(2010).
16. O.Scholtenetal.,Phys.Rev.Lett.103,191301(2009).
17. S.Buitinketal.,A&A521,A47(2010). 10-7
18. S.terVeenetal.,Phys.Rev.D82,103014(2010).
19. C.Jamesetal.,Phys.Rev.E84,056602(2010).
20. E.Zas,F.HalzenandT.Stanev,Phys.Rev.D45,362 s]
/
(1992). r
21. O.Scholtenetal.,Astropart.Phys.26,219(2006). 2/s 10-8
m
22. R.Gandhi,Nucl.Phys.B91453(2000). 46.7hWSRT
c
23. J.Alvarez-Muñizetal.,Phys.Rev.D74,023007(2006). V/
24. M.vanHaarlemetal.[LOFARCollaboration]submitted e
G
toA&A (2013).
25. M.Meviusetal.,NIMPA662,26(2012). E[ 10-9
26. K.Singhetal.,NIMPA664,171(2012). d
27. R.Abassietal.[IceCubecollaboration],Phys.Rev.D83, N/ 7dLOFAR
d
092003(2011).
2
28. P.Gorhametal.,Phys.Rev.D82,022004(2010). E
29. I.Kravchenkoetal.,Phys.Rev.D73,082002(2006). 10-10 SKA-m
30. J.Abrahametal.[AUGERcollaboration],Phys.Rev.Lett.
100,211101(2008).
SKA-l
31. R.Schilizzietal.,SPIE7733(2010).
32. P.Allisonetal.[ARAcollaboration],Astropart.Phys.35, 10-11
457(2012).
10-1 1 10 102 103 104
33. L.Gerhardtetal.,NIMPA624,85(2010).
34. Y.Takahashietal.[JEM-EUSOcollaboration],NewJ. 20
E[10 eV]
Phys.11,065009(2009).
35. K.Kotera,D.AllardandA.Olinto,JCAP10,13(2010).
36. V.Berezinsky,E.SabancilarandA.Vilenkin,Phys.Rev.
D84,085006(2011).
37. C.LunardiniandE.Sabancilar,Phys.Rev.D86,085008
FIGURE 6. CR flux limit of WSRT, as well as expected
(2012).
sensitivities of LOFAR and SKA. The dotted line indicates
a constant particle flux (cm−2sr−1s−1) corresponding to the
particlefluxofthehighest-energydatapoint.Afluxabovethis
linewouldhavebeenfoundbyAuger.