All-sky Sea ch o High-Ene gy Neu inos om G a i a ional Wa e
E en GW170104 wi h he An a es Neu ino Telescope
A. Albe 1, M. And ´e2, M. Anghinol i3, G. An on4, M. A did5, J.-J. Aube 6, T. A gi as7,
B. Ba e 7, J. Ba ios-Ma ´ı8, S. Basa9, B. Belho ma10, V. Be in6, S. Biagi11,
R. Bo mu h12,13, S. Bou e 7, M.C. Bouwhuis12, H. B ˆanza¸s14, R. B uijn12,15, J. B unne 6,
J. Bus o6, A. Capone16,17, L. Ca ame e14, J. Ca 6, S. Celli16,17,18, R. Che kaoui El
Mou sli19, T. Chia usi20, M. Ci cella21, J.A.B. Coelho7, A. Colei o7,8, R. Coniglione11,
H. Cos an ini6, P. Coyle6, A. C euso 7, A. F. D´ıaz22, A. Deschamps23, G. De Bonis17,
C. Dis e ano11, I. Di Palma16,17, A. Domi3,24, C. Donzaud7,25, D. Do nic6, D. D ouhin1,
T. Ebe l4, I. El Bojaddaini26, N. El Khaya i19, D. Els¨asse 27, A. Enzenh¨o e 6, A. E ahi i19,
F. Fassi19, I. Felis5, L.A. Fusco20,28, P. Gay29,7, V. Gio dano30, H. Glo in31,32, T. G ´egoi e7,
R. G acia Ruiz7, K. G a 4, S. Hallmann4, H. an Ha en33, A.J. Heijboe 12, Y. Hello23, J.J.
He n´andez-Rey8, J. H¨oßl4, J. Ho es ¨ad 4, C. Hugon3,24, G. Illumina i8, C.W. James4, M.
de Jong12,13, M. Jongen12, M. Kadle 27, O. Kalekin4, U. Ka z4, D. Kießling4,
A. Kouchne 7,32, M. K e e 27, I. K eykenbohm34, V. Kuliko skiy6,35, C. Lachaud7,
R. Lahmann4, D. Le `e e36, E. Leono a30,37, M. Lo ze8, S. Louca os38,7, M. Ma celin9,
A. Ma gio a20,28, A. Ma inelli39,40, J.A. Ma ´ınez-Mo a5, R. Mele41,42, K. Melis12,15,
T. Michael12, P. Migliozzi41, A. Moussa26, S. Na as43, E. Nez i9, M. O ganoko 44,
G.E. P˘a ˘ala¸s14, C. Pelleg ino20,28, C. Pe ina16,17, P. Pia elli11, V. Popa14, T. P adie 44,
L. Quinn6, C. Racca1, G. Riccobene11, A. S´anchez-Losa21, M. Salda˜na5, I. Sal ado i6, D. F.
E. Sam leben12,13, M. Sanguine i3,24, P. Sapienza11, F. Sch¨ussle 38, C. Siege 4,
M. Spu io20,28, Th. S ola czyk38, M. Taiu i3,24, Y. Tayala i19, A. T o a o11, D. Tu pin6,
C. T¨onnis8, B. Vallage38,7, V. Van Elewyck7,32, F. Ve sa i20,28, D. Vi olo41,42,
A. Vizzoca16,17, J. Wilms34, J.D. Zo noza8, and J. Z´u˜niga8
1GRPHE - Uni e si ´e de Hau e Alsace - Ins i u uni e si ai e de echnologie de Colma , 34 ue du G illenb ei BP 50568 -
68008 Colma , F ance
2Technical Uni e si y o Ca alonia, Labo a o y o Applied Bioacous ics, Rambla Exposici´o, 08800 Vilano a i la Gel ´u,
Ba celona, Spain
3INFN - Sezione di Geno a, Via Dodecaneso 33, 16146 Geno a, I aly
4F ied ich-Alexande -Uni e si ¨a E langen-N¨u nbe g, E langen Cen e o As opa icle Physics, E win-Rommel-S . 1, 91058
E langen, Ge many
5Ins i u d’In es igaci´o pe a la Ges i´o In eg ada de les Zones Cos ane es (IGIC) - Uni e si a Poli `ecnica de Val`encia. C/
Pa anim 1, 46730 Gandia, Spain
6Aix Ma seille Uni , CNRS/IN2P3, CPPM, Ma seille, F ance
7APC, Uni Pa is Dide o , CNRS/IN2P3, CEA/I u, Obs de Pa is, So bonne Pa is Ci ´e, F ance
8IFIC - Ins i u o de F´ısica Co puscula (CSIC - Uni e si a de Val`encia) c/ Ca ed ´a ico Jos´e Bel ´an, 2 E-46980 Pa e na,
1
a Xi :1710.03020 1 [as o-ph.HE] 9 Oc 2017
Valencia, Spain
9LAM - Labo a oi e d’As ophysique de Ma seille, Pˆole de l’´
E oile Si e de Chˆa eau-Gombe , ue F ´ed´e ic Jolio -Cu ie 38,
13388 Ma seille Cedex 13, F ance
10Na ional Cen e o Ene gy Sciences and Nuclea Techniques, B.P.1382, R. P.10001 Raba , Mo occo
11INFN - Labo a o i Nazionali del Sud (LNS), Via S. So ia 62, 95123 Ca ania, I aly
12Nikhe , Science Pa k, Ams e dam, The Ne he lands
13Huygens-Kame lingh Onnes Labo a o ium, Uni e si ei Leiden, The Ne he lands
14Ins i u e o Space Science, RO-077125 Bucha es , M˘agu ele, Romania
15Uni e si ei an Ams e dam, Ins i uu oo Hoge-Ene gie Fysica, Science Pa k 105, 1098 XG Ams e dam, The Ne he lands
16INFN - Sezione di Roma, P.le Aldo Mo o 2, 00185 Roma, I aly
17Dipa imen o di Fisica dell’Uni e si `a La Sapienza, P.le Aldo Mo o 2, 00185 Roma, I aly
18G an Sasso Science Ins i u e, Viale F ancesco C ispi 7, 00167 L’Aquila, I aly
19Uni e si y Mohammed V in Raba , Facul y o Sciences, 4 a . Ibn Ba ou a, B.P. 1014, R.P. 10000 Raba , Mo occo
20INFN - Sezione di Bologna, Viale Be i-Picha 6/2, 40127 Bologna, I aly
21INFN - Sezione di Ba i, Via E. O abona 4, 70126 Ba i, I aly
22Depa men o Compu e A chi ec u e and Technology/CITIC, Uni e si y o G anada, 18071 G anada, Spain
23G´eoazu , UCA, CNRS, IRD, Obse a oi e de la Cˆo e d’Azu , Sophia An ipolis, F ance
24Dipa imen o di Fisica dell’Uni e si `a, Via Dodecaneso 33, 16146 Geno a, I aly
25Uni e si ´e Pa is-Sud, 91405 O say Cedex, F ance
26Uni e si y Mohammed I, Labo a o y o Physics o Ma e and Radia ions, B.P.717, Oujda 6000, Mo occo
27Ins i u ¨u Theo e ische Physik und As ophysik, Uni e si ¨a W¨u zbu g, Emil-Fische S . 31, 97074 W¨u zbu g, Ge many
28Dipa imen o di Fisica e As onomia dell’Uni e si `a, Viale Be i Picha 6/2, 40127 Bologna, I aly
29Labo a oi e de Physique Co pusculai e, Cle mon Uni e si ´e, Uni e si ´e Blaise Pascal, CNRS/IN2P3, BP 10448, F-63000
Cle mon -Fe and, F ance
30INFN - Sezione di Ca ania, Viale And ea Do ia 6, 95125 Ca ania, I aly
3131, Aix Ma seille Uni e si ´e CNRS ENSAM LSIS UMR 7296 13397 Ma seille, F ance; Uni e si ´e de Toulon CNRS LSIS
UMR 7296, 83957 La Ga de, F ance
32Ins i u Uni e si ai e de F ance, 75005 Pa is, F ance
33Royal Ne he lands Ins i u e o Sea Resea ch (NIOZ) and U ech Uni e si y, Landsdiep 4, 1797 SZ ’ Ho n je (Texel), he
Ne he lands
34D . Remeis-S e nwa e and ECAP, Uni e si ¨a E langen-N¨u nbe g, S e nwa s . 7, 96049 Bambe g, Ge many
35Moscow S a e Uni e si y, Skobel syn Ins i u e o Nuclea Physics, Leninskie go y, 119991 Moscow, Russia
36Medi e anean Ins i u e o Oceanog aphy (MIO), Aix-Ma seille Uni e si y, 13288, Ma seille, Cedex 9, F ance; Uni e si ´e du
Sud Toulon-Va , CNRS-INSU/IRD UM 110, 83957, La Ga de Cedex, F ance
37Dipa imen o di Fisica ed As onomia dell’Uni e si `a, Viale And ea Do ia 6, 95125 Ca ania, I aly
38Di ec ion des Sciences de la Ma i`e e - Ins i u de eche che su les lois ondamen ales de l’Uni e s - Se ice de Physique des
Pa icules, CEA Saclay, 91191 Gi -su -Y e e Cedex, F ance
39INFN - Sezione di Pisa, La go B. Pon eco o 3, 56127 Pisa, I aly
40Dipa imen o di Fisica dell’Uni e si `a, La go B. Pon eco o 3, 56127 Pisa, I aly
41INFN - Sezione di Napoli, Via Cin ia 80126 Napoli, I aly
42Dipa imen o di Fisica dell’Uni e si `a Fede ico II di Napoli, Via Cin ia 80126, Napoli, I aly
43Dp o. de F´ısica Te´o ica y del Cosmos & C.A.F.P.E., Uni e si y o G anada, 18071 G anada, Spain
44Uni e si ´e de S asbou g, CNRS, IPHC UMR 7178, F-67000 S asbou g, F ance
Oc obe 10, 2017
2
Abs ac
Ad anced LIGO de ec ed a signi ican g a i a ional wa e signal (GW170104) o igina ing
om he coalescence o wo black holes du ing he second obse a ion un on Janua y 4 h,
2017. An all-sky high-ene gy neu ino ollow-up sea ch has been made using da a om he
An a es neu ino elescope, including bo h upgoing and downgoing e en s in wo sepa a e
analyses. No neu ino candida es we e ound wi hin ±500 s a ound he GW e en ime no any
ime clus e ing o e en s o e an ex ended ime window o ±3 mon hs. The non-de ec ion is
used o cons ain iso opic-equi alen high-ene gy neu ino emission om GW170104 o less han
∼4×1054 e g o a E−2spec um.
1 In oduc ion
The i s wo con i med obse a ions o g a i a-
ional wa es (GWs) p oduced by he me ge o
bina y black holes (BBHs) we e ecen ly made
by he Ad anced LIGO in e e ome e s du ing
hei obse a ion un O1 [1, 2]. The second ob-
se a ion un o Ad anced LIGO (O2) began
in No embe 2016 and s opped on Augus 25 h,
2017. A BBH signal, GW170104, was eco ded
du ing O2 on Janua y 4 h, 2017 a 10:11:58.6
UTC [3]. The alse ala m a e co esponding o
he signal p oduced by his e en is less han
one e en o e 70 000 yea s. The signal was p o-
duced by he coalescence o wo black holes o
in e ed masses o 31.2+8.4
−6.0Mand 19.4+5.3
−5.9M
a a luminosi y dis ance o 880+450
−390 Mpc. The
GW sou ce loca ion was cons ained o wi hin
1608 deg2o he sky a 90% c edible le el ( e-
gion he ea e deno ed as GW e o box) by he
LALIn e ence econs uc ion algo i hm [3].
Black holes wi h acc e ion disks can igge
ela i is ic ou lows whe e high-ene gy (TeV–
PeV) neu inos (HENs) can be p oduced, i
had onic pa icles a e accele a ed wi hin he je s
[4, 5, 6]. Such an accele a ion p ocess can ake
place i magne ic ields and a long-li ed deb is
disk emain om he s ella e olu ion o he
black-hole p ogeni o s o i he bina y sys em
esides in a dense gaseous en i onmen (see e.g.
[7, 8, 9, 10]). Since he p esence o an acc e ion
disk was no excluded in he case o GW170104,
he sea ch o muon HENs emi ed be o e o a -
e he me ge could b ing aluable in o ma ion
abou he o ma ion o ela i is ic ou lows.
The An a es Collabo a ion has joined he
ollow-up p og am o LIGO/Vi go de ec ions
and has ecei ed GW ale s du ing he whole
O2 un pe iod. The angula esolu ion o he
An a es neu ino elescope (∼0.4◦a ∼10 TeV
o muon neu inos) compa ed o he size o he
GW e o box o e s he possibili y o d as ically
educe he size o he egion o in e es in case
o a coinciden muon neu ino de ec ion.
The An a es ield-o - iew (FoV), when e-
s ic ed o upgoing e en s, enclosed 51% o he
GW170104 e o box p o ided by LIGO/Vi go
a he ale ime. Coincidences in ime and di ec-
ion be ween he GW signal and econs uc ed
muon HEN candida es we e sea ched o in a
da as eam o abou 1.2 e en s pe day, selec ed
om a o al o O(100) upgoing neu ino ack
candida es igge ed by ANTARES pe day [11].
No neu ino coun e pa was ound and he e-
sul s o his eal- ime analysis we e ansmi ed
ia he Gamma- ay Coo dina es Ne wo k (GCN)
3
ci cula #20370 [12] o he LIGO/Vi go ollow-
up communi y in less han 24 hou s a e he
elease o he ale . The esul s p o ided by
he An a es Collabo a ion we e he only eal-
ime neu ino ollow-up ela ed o his e en .
The absence o neu ino candida es bo h em-
po ally and spa ially coinciden wi h GW170104
allowed o de i ing a p elimina y uppe limi on
he spec al luence emi ed in neu inos by he
sou ce a 90% con idence le el (CL). This up-
pe limi is exp essed as a unc ion o he loca-
ion o he sou ce in equa o ial coo dina es and
assuming a s anda d neu ino spec al model
dN/dE ∝E−2. This esul was ansmi ed
o he LIGO/Vi go ollow-up communi y in he
GCN ci cula #20517 [12].
The esul s o an upda ed high-ene gy neu-
ino ollow-up o GW170104 using he An a es
neu ino elescope a e p esen ed in his pape .
The sea ch o a ansien neu ino coun e pa
has been ex ended o he ull sky wi h di e en
ene gy h esholds o e en s o igina ing om be-
low and abo e he An a es ho izon, and o a
la ge emission imescale. The sea ch desc ibed
he ea e was pe o med wi h he mos ecen
o line- econs uc ed da ase , inco po a ing ded-
ica ed calib a ions o posi ioning [13], iming [14]
and e iciency [15]. The analysis has been op-
imized o inc ease he sensi i i y o he de-
ec o a he ime o he ale . Two neu ino
spec al models we e assumed: a gene ic model
dN/dE =φ0E−2 ypically expec ed o Fe mi
accele a ion and a model wi h a high-ene gy
cu o dN/dE =φ0E−2exp h−p(E/100TeV)i.
The second model is expec ed o sou ces wi h
exponen ial cu o in he p ima y p o on spec-
um [16]. Finally, sys ema ic e o s a ec ing
he co esponding uppe limi s on neu ino emis-
sion a e accoun ed o .
The capabili ies o he An a es de ec o and
he sea ch p ocedu es a e summa ized in Sec ion
2. The cons ain s on he neu ino luence and
o al ene gy emi ed in neu inos de i ed om
he non-de ec ion o a neu ino coun e pa o
GW170104 a e p esen ed in Sec ion 3. The con-
clusions a e epo ed in Sec ion 4.
2 High ene gy neu ino sea ch
An a es [17] is an unde wa e neu ino ele-
scope loca ed in he Medi e anean Sea, o sho e
Toulon (F ance). I is composed o an a ay
o pho omul iplie ubes (PMTs), ancho ed a
a dep h o 2475 m unde he sea le el. Neu-
inos wi h ene gies abo e ∼102GeV a e de-
ec ed h ough he Che enko ligh induced by
ela i is ic pa icles c ea ed om he in e ac ion
o neu inos wi h ma e . In addi ion o as-
ophysical neu ino signals, bo h a mosphe ic
muons and neu inos can lead o de ec able ligh
in he de ec o and a e conside ed as backg ound
e en s. Howe e , only neu inos can a e se he
Ea h. Looking a upgoing pa icles in he de ec-
o e e ence ame allows o emo ing a la ge
pa o he downgoing a mosphe ic muon back-
g ound. Remaining mis- econs uc ed downgo-
ing muons a e u he ejec ed by applying cu s
on he econs uc ion quali y pa ame e s. In ad-
di ion, he in ense backg ound o downwa d go-
ing a mosphe ic muons is d as ically educed by
he equi emen o a join ime and space coin-
cidence wi h he GW ime and spa ial e o box.
This allows o sea ching o a neu ino coun e -
pa o GW170104 in bo h upgoing and downgo-
ing da ase s, which consis o e en s o igina ing
espec i ely om below and abo e he An a es
ho izon.
Conside ing he e ined loca ion p obabili y
4
p o ided by he LIGO/Vi go LALIn e ence so -
wa e [18], he e is a 52% chance ha he GW
emi e was below he An a es ho izon whe e
any neu ino e en s om his pa o he sky
would be seen as upgoing in he de ec o ame.
This co esponds o a 45% p obabili y o he
sou ce o be loca ed inside he GW e o box and
below he An a es ho izon (see Fig. 1). To ex-
end he o e lap be ween he An a es FoV and
he GW e o box, downgoing e en s ha e been
added o he sea ch in an independen analysis.
All-sky An a es da a ha e been sea ched o
ack e en s p oduced by νµand ¯νµcha ged cu -
en in e ac ions coinciden wi h GW170104 us-
ing a ime window o ±500 s a ound he GW
ansien (Sec ions 2.1 and 2.2). This ime win-
dow was adop ed as he s anda d sea ch win-
dow o p e ious join GW-HEN sea ches [19],
o ins ance in he case o GW150914 and
GW151226 [20, 21]. A sea ch o a neu ino
coun e pa wi hin an ex ended ime window o
±3 mon hs has also been done using he online
da as eam o An a es (Sec ion 2.3).
2.1 Sea ch below he An a es ho izon
A binned sea ch o coinciden upgoing neu-
inos was pe o med ollowing a blind p oce-
du e. The ack econs uc ion algo i hm com-
pu es bo h he neu ino di ec ion, oge he wi h
an es ima ed e o β, and a quali y pa ame e
Λ [22]. This sample is domina ed by backg ound
e en s om mis- econs uc ed downgoing a mo-
sphe ic muons, which deposi ene gy in he de-
ec o h ough s ochas ic p ocesses. The da ase
was educed by adjus ing Λ such ha any e en
passing he sea ch c i e ia and loca ed wi hin
he GW e o box, below he An a es ho i-
zon, would lead o a de ec ion wi h a signi icance
le el o 3σ. This op imiza ion was ca ied ou on
Figu e 1: Visibili y map o GW170104 in equa-
o ial coo dina es. The sky egions below and
abo e he An a es ho izon a he ale ime
a e shown in blue and whi e espec i ely. E en s
ha o igina e om he blue (whi e) egion will
be seen as upgoing (downgoing) in he de ec o
ame. The ed and black con ou s show he e-
cons uc ed p obabili y densi y con ou s o he
GW e en a 50% and 90% c edible le el espec-
i ely.
da a ou side he 1000 s ime window used in his
sea ch. A Mon e Ca lo simula ion o he de ec-
o esponse [23, 24] a he ale ime allows o
es ima ing he ela i e con ibu ion o he a mo-
sphe ic neu inos and he mis- econs uc ed a -
mosphe ic muons o he backg ound a e below
he An a es ho izon and wi hin ±500 s. A o al
o 2.2×10−2a mosphe ic neu ino candida es a e
expec ed while he numbe o mis- econs uc ed
downgoing muons amoun s o 3.7×10−2e en s
o e 2πs .
A e unblinding o he da ase , no e en em-
po ally coinciden wi h GW170104 was ound.
5
2.2 Sea ch abo e he An a es ho izon
A sea ch o coinciden neu ino candida es de-
ec ed abo e he An a es ho izon was ca ied
ou by selec ing downgoing e en s wi h βsmalle
han 1◦. The cu s a e op imized on a com-
bina ion o Λ and he numbe o hi s used in
he econs uc ion, whe e a hi co esponds o
a PMT signal abo e a gi en h eshold. The
numbe o hi s can be conside ed as a p oxy
o he muon/neu ino ene gy. Indeed, downgo-
ing a mosphe ic muons a e less likely o p oduce
a numbe o hi s as la ge as ha p oduced by
e y high-ene gy cosmic neu inos. Anew, he
selec ion c i e ia we e op imized such ha one
e en occu ing wi hin he signal ime window
o 1000 s and loca ed inside he GW e o box
loca ed abo e he An a es ho izon would lead
o a de ec ion wi h a signi icance le el o 3σ. The
inal se o cu s is chosen as he one maximizing
he ac ion o su i ing signal e en s. The i-
nal sample is mos ly composed o a mosphe ic
muons wi h a o al o 8.2×10−2backg ound
e en s expec ed abo e he An a es ho izon
wi hin ±500 s. The median neu ino ene gy ha
would be de ec ed by An a es o a E−2signal
spec um is abou a ac o o 10 highe o his
analysis compa ed o he sea ch below he ho i-
zon desc ibed in Sec ion 2.1.
A e unblinding o he da ase , no e en em-
po ally coinciden wi h GW170104 was ound.
2.3 Ex ended ime window sea ch
The ime window o ±500 s was chosen by as-
suming ha i compac bina y me ge s a e e-
la ed o gamma- ay bu s s hen he neu ino sig-
nal should occu close in ime o he GW emis-
sion. This ime window is la ge enough o ca ch
po en ial p ecu so neu ino emission and ime
o se s wi h espec o he GW signal [19]. Fo
comple eness, o p obe non-s anda d p opaga-
ion scena ios simila o hose desc ibed in [25],
a sea ch o shi ed and/o longe -las ing emis-
sion o e ±3 mon hs a ound he GW ale was
pe o med by looking o ime clus e ing o up-
going neu ino e en s.
The e en s selec ed om he online da as-
eam used in he An a es eal- ime ale p o-
g am [11] a e in es iga ed o ime clus e ing.
The spa ial clus e ing o he e en s and hei co-
incidence wi h he GW e o box we e in es i-
ga ed a pos e io i.
An unbinned likelihood sea ch was pe o med
ollowing he me hodology applied in p e ious
analyses [26, 27]. Fo each combina ion o wo
e en s aand b, a signal p obabili y o he i h
e en is de ined as:
Sa,b
i=H( b− i)H( i− a)
b− a
,(1)
wi h H he Hea iside unc ion1and aand b,
he de ec ion ime o e en s aand b(wi h a<
b). The backg ound ime p obabili y o he i h
e en , Bi, is de i ed di ec ly om he p obabili y
densi y unc ion (PDF) o he downgoing econ-
s uc ed e en s. In his way, he backg ound dis-
ibu ion e lec s he e olu ion o he e en a e
due o he a iabili y o he da a aking condi-
ions.
Gi en a da ase o Ne en s, he likelihood
unc ion La,b(ns) o a gi en pai o e en s oc-
cu ing a imes aand bis de ined as:
La,b(ns) =
N
Y
i=1 hns
NSa,b
i+1−ns
NBii,(2)
whe e nsis he unknown numbe o signal e en s.
Fo each pai o e en s occu ing a imes aand
1H(0) is de ined as 1.
6
b, he likelihood is maximized wi h espec o ns
o p o ide he bes - i numbe o e en s ˆns. The
es s a is ic (TSa,b) is compu ed om he likeli-
hood a io o he backg ound-only (null) hypo h-
esis o e he signal-plus-backg ound hypo hesis
as:
TSa,b =−2·log T
b− a
L(ns= 0)
La,b(ˆns),(3)
whe e he e m T
b− ain he squa e b acke s
is a ial ac o . This quan i y is needed o
co ec o he ac ha he e a e many mo e
independen small ime windows han la ge
ones, which end o a ou e y sho la es.
The pa ame e Tco esponds o he da ase
li e ime. Gi en a sample o Ne en s, N(N−1)
2
alues o TS we e compu ed (one o each pai o
e en s aand b). The clus e ha maximizes he
TS is inally conside ed as he mos signi ican
one.
To compu e he p- alue o he mos signi i-
can clus e , O(10 000) pseudo-expe imen s we e
gene a ed, each o hem consis ing o Ne en s
d awn andomly om he ime PDF o he back-
g ound. The ac ion o ials o which he TS
alue is la ge han he one ob ained om he
da a is e e ed o as he p- alue.
The mos signi ican ime clus e has been
ound o con ain ˆns=8.3 i ed signal e en s, oc-
cu ing be ween min(MJD)=57682.73398 and
max(MJD)=57685.62900 ( max − min =2.89
days). I leads o a pos - ial p- alue o 70%
and is hus consis en wi h he backg ound-only
hypo hesis. In addi ion, he An a es e en s
con ained in his ime window a e no spa ially
compa ible and do no o e lap wi h he GW e -
o box.
3 As ophysical cons ain s
The non-de ec ion o join GW and neu ino sig-
nals is used o cons ain neu ino emission om
he GW sou ce. Uppe limi s on bo h he lu-
ence and he o al ene gy emi ed in neu inos
a e p esen ed in he o m o skymaps since he
sensi i i y o An a es depends on he sou ce
di ec ion.
3.1 Cons ain s on he neu ino spec-
al luence
Uppe limi s a 90% CL on he neu ino luence
om a poin sou ce wi hin ∆ =±500 s we e
calcula ed using he null esul and he de ec o
accep ance, es ima ed ia a Mon e Ca lo simu-
la ion o he de ec o esponse a he ime o he
GW signal. This simula ion is p oduced on a
un-by- un basis [23, 24] o accoun o he a i-
a ion o he da a aking condi ions unde he sea.
The wo spec al models desc ibed in Sec ion 1
we e conside ed.
The numbe o neu ino e en s expec ed o be
obse ed by An a es om a poin sou ce a
declina ion δand wi h a neu ino lux dN/dE
(in GeV−1cm−2s−1) in a ime window ∆ is
gi en by:
Ne en s = ∆ ZdN
dE(E)Ae (E, δ)dE, (4)
whe e Ae (E, δ) is he e ec i e a ea o An a es
a he ale ime which ake in o accoun he ab-
so p ion o neu inos by he Ea h and depends
on he neu ino ene gy E, he sou ce declina ion
δand he applied cu s.
The uppe limi on he luence, φ90%
0, is de-
ined as he luence alue ha on a e age would
p oduce 2.3 de ec ed neu ino e en s. Assuming
7
adN/dE =φ0E−2neu ino spec al model, i is
de i ed as:
φ90%
0=2.3
∆ RE−2Ae (Eν, δ)dE,(5)
whe e he denomina o e e s o he ins an a-
neous accep ance o he An a es de ec o a
he ime o he ale compu ed be ween 1 GeV
and 100 PeV. Equa ion 5also applies o he sec-
ond spec al model conside ed in his s udy. The
same me hodology is used o de i e φ90%
0in he
case o he second conside ed spec al model and
on he sea ch abo e he ho izon.
Fig. 2shows he neu ino spec al luence up-
pe limi (φ90%
0) o GW170104 as a unc ion
o he sou ce di ec ion o bo h spec al mod-
els. Compu ed om he Mon e Ca lo simula-
ion, he ene gy ange co esponding o he 5%–
95% quan iles o he neu ino lux below he
An a es ho izon is equal o [3.2 TeV; 3.6 PeV]
o he dN/dE =φ0E−2spec al model and
[1.4 TeV; 270 TeV] o he model wi h exponen-
ial cu o a 100 TeV. Abo e he An a es ho i-
zon, he 5%–95% quan iles o he neu ino lux
a e espec i ely equal o [120 TeV; 22 PeV] and
[53 TeV; 950 TeV].
The sys ema ic unce ain ies on he luence
uppe limi s ha e been es ima ed by summing
quad a ically i) he sys ema ic e o on he ac-
cep ance o he de ec o and ii) he unce ain y
ela ed o he abili y o he An a es un-by- un
Mon e Ca lo app oach o accu a ely ep oduce
he a iable da a aking condi ions on sho ime
scales. This la e e ec can become dominan
when looking o ansien neu ino sou ces.
The sys ema ic e o on he accep ance has
been compu ed by educing he e iciency o each
op ical module by 15% in he de ec o simula-
ions. This leads o a 15% unce ain y on he
accep ance as de ailed in [22].
30
◦
60
◦
90
◦
120
◦
150
◦
180
◦
210
◦
240
◦
270
◦
300
◦
330
◦
-75°
-60°
-45°
-30°
-15°
0°
15°
30°
45°
60°
75°
100101102
E
2
dN/dE
[
GeV cm
−
2]
30
◦
60
◦
90
◦
120
◦
150
◦
180
◦
210
◦
240
◦
270
◦
300
◦
330
◦
-75°
-60°
-45°
-30°
-15°
0°
15°
30°
45°
60°
75°
100101102103
E
2
dN/dE
[
GeV cm
−
2]
Figu e 2: All-sky uppe limi on he neu-
ino spec al luence (νµ+ ¯νµ) om GW170104
as a unc ion o sou ce di ec ion assum-
ing dN/dE∝E−2( op) and dN/dE∝
E−2exp h−p(E/100TeV)i(bo om) neu ino
spec a. The ed and black lines show he GW
skymap con ou s a 50% and 90% c edible le els,
espec i ely. Skymaps a e de ined in equa o ial
coo dina es.
The second sou ce o unce ain y has been
cons ained by quan i ying he abili y o he un-
8
by- un Mon e Ca lo o accu a ely ep oduce he
e olu ion o he e en a e obse ed in he da a
om one un o ano he . Due o he low numbe
o e en s passing he op imized quali y cu s, he
un-by- un ag eemen be ween da a and Mon e
Ca lo can only be assessed by loosening he cu s,
on a da a sample domina ed by a mosphe ic
muon e en s. The a ia ions o he muon a e
a e well ep oduced wi h a median ela i e a ia-
ion be ween da a and Mon e Ca lo smalle han
20%. In addi ion, he sho - imescale luc ua-
ions o he e en a e a e smalle o signal neu-
inos wi h a E−2spec al model han o a mo-
sphe ic muons. This was expec ed since he de-
ec o geome y is op imized o upgoing e en s.
Thus, he sys ema ic e o o 20% is conside ed
a conse a i e alue.
The esul ing o al sys ema ic unce ain y on
he luence uppe limi s, which applies o bo h
upgoing and downgoing e en sea ches, is 25%.
3.2 Cons ain s on he o al ene gy
emi ed in neu inos
The GW signal con ains also in o ma ion on
he sou ce dis ance which can be econs uc ed
a ound he GW e o box [28]. This in o ma ion
can be used o de i e an uppe limi on he o al
ene gy adia ed in neu inos as a unc ion o he
di ec ion as pe o med in [21].
The mos likely alue D(~x) o he dis ance
o each di ec ion ~x is used o calcula e he up-
pe limi on he o al iso opic-equi alen ene gy
emi ed in neu inos by he sou ce as:
EUL
ν,iso (~x)=4π[D(~x)]2ZdN
dE(E, ~x)EdE. (6)
Uppe limi s on he o al ene gy a e
compu ed o bo h dN/dE∝E−2and
dN/dE∝E−2exp h−p(E/100TeV)ineu ino
spec al models. The spec um is in eg a ed o e
he ange [100 GeV; 100 PeV]. The uppe limi s
as a unc ion o sou ce di ec ion a e shown in
Fig. 3 o he egion co esponding o upgoing
e en s o An a es. I can be seen in Fig. 3
ha he de i ed cons ain s depend on he posi-
ion on he sky as bo h he luence uppe limi s
and he dis ance cons ain s do. The 5%–95%
ange o alues is [1 ×1054; 4 ×1054] e g and
[6 ×1053; 4 ×1054] e g, o he E−2and he 100
TeV cu o models espec i ely. The s onges
cons ain is ob ained a declina ion δ∼ −17◦
wi h E < 5×1053 e g o a E−2spec um and
E < 3×1053 e g o he spec al ene gy dis i-
bu ion wi h cu o a 100 TeV. These alues a e
abou 10% o he o al ene gy o ∼3.6×1054 e g
emi ed om GW170104 in g a i a ional wa es.
The esul s ob ained o downgoing e en s a e
no p o ided since he neu ino luence uppe
limi s a e much weake . The unce ain y on he
o al ene gy emi ed in neu inos is es ima ed by
accoun ing o bo h he sys ema ic e o on φ90%
0
(compu ed abo e) and he 1σs anda d de ia ion
on he dis ance p o ided by he LIGO/Vi go
GW e en econs uc ion, leading o an a e age
alue o ∼40%.
4 Conclusion
No neu ino emission associa ed wi h he hi d
con i med bina y black hole me ge , GW170104
was de ec ed in he An a es da a. This
non-de ec ion was used o de i e an uppe limi
o he o al neu ino emission om GW170104
o ∼4×1054 e g, o a gene ic E−2neu ino
spec um and o a high-ene gy cu o spec um
E−2exp h−p(E/100TeV)ias expec ed o
sou ces wi h an exponen ial cu o in he p o on
9
