Search for neutrino-induced particle showers with IceCube-40

TLDR: In this article, the search for neutrino-induced particle showers, so-called cascades, in the IceCube-40 detector was performed, and a total of 67 events were found consistent with the expectation of 41 atmospheric muons and 30 atmospheric neutrinos events.

Abstract: We report on the search for neutrino-induced particle showers, so-called cascades, in the IceCube-40 detector. The data for this search were collected between April 2008 and May 2009 when the first 40 IceCube strings were deployed and operational. Three complementary searches were performed, each optimized for different energy regimes. The analysis with the lowest energy threshold (2 TeV) targeted atmospheric neutrinos. A total of 67 events were found, consistent with the expectation of 41 atmospheric muons and 30 atmospheric neutrino events. The two other analyses targeted a harder, astrophysical neutrino flux. The analysis with an intermediate threshold of 25 TeV leads to the observation of 14 cascadelike events, again consistent with the prediction of 3.0 atmospheric neutrino and 7.7 atmospheric muon events. We hence set an upper limit of E-2 Phi(lim) <= 7.46 x 10(-8) GeV sr(-1) s(-1) cm(-2) (90% C.L.) on the diffuse flux from astrophysical neutrinos of all neutrino flavors, applicable to the energy range 25 TeV to 5 PeV, assuming an E-nu(-2) spectrum and a neutrino flavor ratio of 1: 1: 1 at the Earth. The third analysis utilized a larger and optimized sample of atmospheric muon background simulation, leading to a higher energy threshold of 100 TeV. Three events were found over a background prediction of 0.04 atmospheric muon events and 0.21 events from the flux of conventional and prompt atmospheric neutrinos. Including systematic errors this corresponds to a 2.7 sigma excess with respect to the background-only hypothesis. Our observation of neutrino event candidates above 100 TeV complements IceCube's recently observed evidence for high-energy astrophysical neutrinos.

Figures

FIG. 4 (color online). Proton air showers are simulated according to the Glasstetter spectrum illustrated in Fig. 3. Bremsstrahlung cascades from muons originating in these showers are studied in order to assess the probability to obtain a bremsstrahlung cascade bright enough to pass event selection cuts from air showers below a given energy threshold. The plot shows the fraction of proton air showers with primary energy Eprimary below 10, 100, 1000 TeV that exhibit catastrophic energy losses of Ecascade above a given energy.

FIG. 3 (color online). The cosmic-ray spectrum is modeled with two broken power laws for proton and iron (red and magenta dashed lines, respectively). Their parameters are taken from Ref. [38]. The Hörandel model and the data from which it has been derived are shown for comparison (both taken from [39]). The Hörandel model comprises components for each element from hydrogen up to iron illustrated by solid lines.

FIG. 7 (color online). At level 6, 11 input variables are combined to train a boosted decision tree. Individually the variables provide only limited separation power, especially for distinguishing atmospheric neutrinos and muons. The four most important variables according to the classifier are shown here at level 6. Definitions of the variables can be found in Sec. VA. The shaded regions in the plot denote the statistical uncertainties.

TABLE III. Overview on systematics uncertainties on the event count for analyses Ia and Ib. The systematic uncertainties associated with sample II are similar to those of sample Ib except for the atmospheric muon prediction. Unlike sample Ib both samples Ia and II estimate the muon background from simulation for which similar systematic uncertainties apply. See Sec. VI for further details.

FIG. 5 (color online). The IceCube-40 online filter, the data are shown by the black filled circles, simulated atmospheric muon background by the red line, simulated atmospheric electron neutrinos by the blue line, and simulated electron neutrino signal with spectrum E2dN=dE ¼ 3.6 × 10−8 GeV sr−1 s−1 cm−2 by the green line.

FIG. 6 (color online). The IceCube-40 level 3 filter variables. (a) The level 3 zenith angle and energy reconstructions. The two panels show the data and E−2 spectrum electron neutrino signal. The level 3 cuts are represented by the black lines at zenith Θtrack ¼ 80° and energy EACER ¼ 10 TeV. Events in the hatched upper left quadrant of each of the panels were removed. (b) The reduced log-likelihood rlogLCSCD. The data are shown by the black filled circles, simulated atmospheric muon background by the red line, simulated atmospheric electron neutrinos by the blue line, and simulated electron neutrino signal with spectrum E2dN=dE ¼ 3.6 × 10−8 GeV sr−1 s−1 cm−2 by the green line. Events with rlogLCSCD < 10 were removed.

Full text

Search for neutrino-induced particle showers with IceCube-40
M. G. Aartsen,
2
R. Abbasi,
29
M. Ackermann,
45
J. Adams,
15
J. A. Aguilar,
23
M. Ahlers,
29
D. Altmann,
22
C. Arguelles,
29
T. C. Arlen,
42
J. Auffenberg,
29
X. Bai,
33,*
M. Baker,
29
S. W. Barwick,
25
V. Baum,
30
R. Bay,
7
J. J. Beatty,
17,18
J. Becker Tjus,
10
K.-H. Becker,
44
S. BenZvi,
29
P. Berghaus,
45
D. Berley,
16
E. Bernardini,
45
A. Bernhard,
32
D. Z. Besson,
27
G. Binder,
8,7
D. Bindig,
44
M. Bissok,
1
E. Blaufuss,
16
J. Blumenthal,
1
D. J. Boersma,
43
C. Bohm,
36
D. Bose,
38
S. Böser,
11
O. Botner,
43
L. Brayeur,
13
H.-P. Bretz,
45
A. M. Brown,
15
R. Bruijn,
26
J. Casey,
5
M. Casier,
13
D. Chirkin,
29
A. Christov,
23
B. Christy,
16
K. Clark,
39
L. Classen,
22
F. Clevermann,
20
S. Coenders,
1
S. Cohen,
26
D. F. Cowen,
42,41
A. H. Cruz Silva,
45
M. Danninger,
36
J. Daughhetee,
5
J. C. Davis,
17
M. Day,
29
J. P. A. M. de André,
42
C. De Clercq,
13
S. De Ridder,
24
P. Desiati,
29
K. D. de Vries,
13
M. de With,
9
T. DeYoung,
42
J. C. Díaz-Vélez,
29
M. Dunkman,
42
R. Eagan,
42
B. Eberhardt,
30
B. Eichmann,
10
J. Eisch,
29
S. Euler,
1
P. A. Evenson,
33
O. Fadiran,
29
A. R. Fazely,
6
A. Fedynitch,
10
J. Feintzeig,
29
T. Feusels,
24
K. Filimonov,
7
C. Finley,
36
T. Fischer-Wasels,
44
S. Flis,
36
A. Franckowiak,
11
K. Frantzen,
20
T. Fuchs,
20
T. K. Gaisser,
33
J. Gallagher,
28
L. Gerhardt,
8,7
L. Gladstone,
29
T. Glüsenkamp,
45
A. Goldschmidt,
8
G. Golup,
13
J. G. Gonzalez,
33
J. A. Goodman,
16
D. Góra,
22
D. T. Grandmont,
21
D. Grant,
21
P. Gretskov,
1
J. C. Groh,
42
A. Groß,
32
C. Ha,
8,7
A. Haj Ismail,
24
P. Hallen,
1
A. Hallgren,
43
F. Halzen,
29
K. Hanson,
12
D. Hebecker,
11
D. Heereman,
12
D. Heinen,
1
K. Helbing,
44
R. Hellauer,
16
S. Hickford,
15,†
G. C. Hill,
2
K. D. Hoffman,
16
R. Hoffmann,
44
A. Homeier,
11
K. Hoshina,
29
F. Huang,
42
W. Huelsnitz,
16
P. O. Hulth,
36
K. Hultqvist,
36
S. Hussain,
33
A. Ishihara,
14
E. Jacobi,
45
J. Jacobsen,
29
K. Jagielski,
1
G. S. Japaridze,
4
K. Jero,
29
O. Jlelati,
24
B. Kaminsky,
45
A. Kappes,
22
T. Karg,
45
A. Karle,
29
M. Kauer,
29
J. L. Kelley,
29
J. Kiryluk,
37
J. Kläs,
44
S. R. Klein,
8,7
J.-H. Köhne,
20
G. Kohnen,
31
H. Kolanoski,
9
L. Köpke,
30
C. Kopper,
29
S. Kopper,
44
D. J. Koskinen,
19
M. Kowalski,
11
M. Krasberg,
29
A. Kriesten,
1
K. Krings,
1
G. Kroll,
30
J. Kunnen,
13
N. Kurahashi,
29
T. Kuwabara,
33
M. Labare,
24
H. Landsman,
29
M. J. Larson,
40
M. Lesiak-Bzdak,
37
M. Leuermann,
1
J. Leute,
32
J. Lünemann,
30
O. Macías,
15
J. Madsen,
35
G. Maggi,
13
R. Maruyama,
29
K. Mase,
14
H. S. Matis,
8
F. McNally,
29
K. Meagher,
16
M. Merck,
29
T. Meures,
12
S. Miarecki,
8,7
E. Middell,
45,‡
N. Milke,
20
J. Miller,
13
L. Mohrmann,
45
T. Montaruli,
23
R. Morse,
29
R. Nahnhauer,
45
U. Naumann,
44
H. Niederhausen,
37
S. C. Nowicki,
21
D. R. Nygren,
8
A. Obertacke,
44
S. Odrowski,
21
A. Olivas,
16
A. Omairat,
44
A. O’Murchadha,
12
T. Palczewski,
40
L. Paul,
1
J. A. Pepper,
40
C. Pérez de los Heros,
43
C. Pfendner,
17
D. Pieloth,
20
E. Pinat,
12
J. Posselt,
44
P. B. Price,
7
G. T. Przybylski,
8
M. Quinnan,
42
L. Rädel,
1
M. Rameez,
23
K. Rawlins,
3
P. Redl,
16
R. Reimann,
1
E. Resconi,
32
W. Rhode,
20
M. Ribordy,
26
M. Richman,
16
B. Riedel,
29
S. Robertson,
2
J. P. Rodrigues,
29
C. Rott,
38
T. Ruhe,
20
B. Ruzybayev,
33
D. Ryckbosch,
24
S. M. Saba,
10
H.-G. Sander,
30
M. Santander,
29
S. Sarkar,
19,34
K. Schatto,
30
F. Scheriau,
20
T. Schmidt,
16
M. Schmitz,
20
S. Schoenen,
1
S. Schöneberg,
10
A. Schönwald,
45
A. Schukraft,
1
L. Schulte,
11
O. Schulz,
32
D. Seckel,
33
Y. Sestayo,
32
S. Seunarine,
35
R. Shanidze,
45
C. Sheremata,
21
M. W. E. Smith,
42
D. Soldin,
44
G. M. Spiczak,
35
C. Spiering,
45
M. Stamatikos,
17,§
T. Stanev,
33
N. A. Stanisha,
42
A. Stasik,
11
T. Stezelberger,
8
R. G. Stokstad,
8
A. Stößl,
45
E. A. Strahler,
13
R. Ström,
43
N. L. Strotjohann,
11
G. W. Sullivan,
16
H. Taavola,
43
I. Taboada,
5
A. Tamburro,
33
A. Tepe,
44
S. Ter-Antonyan,
6
G. Tešić,
42
S. Tilav,
33
P. A. Toale,
40
M. N. Tobin,
29
S. Toscano,
29
M. Tselengidou,
22
E. Unger,
10
M. Usner,
11
S. Vallecorsa,
23
N. van Eijndhoven,
13
A. Van Overloop,
24
J. van Santen,
29
M. Vehring,
1
M. Voge,
11
M. Vraeghe,
24
C. Walck,
36
T. Waldenmaier,
9
M. Wallraff,
1
Ch. Weaver,
29
M. Wellons,
29
C. Wendt,
29
S. Westerhoff,
29
B. Whelan,
2
N. Whitehorn,
29
K. Wiebe,
30
C. H. Wiebusch,
1
D. R. Williams,
40
H. Wissing,
16
M. Wolf,
36
T. R. Wood,
21
K. Woschnagg,
7
D. L. Xu,
40
X. W. Xu,
6
J. P. Yanez,
45
G. Yodh,
25
S. Yoshida,
14
P. Zarzhitsky,
40
J. Ziemann,
20
S. Zierke,
1
and M. Zoll
36
(IceCube Collaboration)
1
III. Physikalisches Institut, RWTH Aachen University, D-52056 Aachen, Germany
2
School of Chemistry and Physics, University of Adelaide, Adelaide SA, 5005 Australia
3
Department of Physics and Astronomy, University of Alaska Anchorage, 3211 Providence Drive,
Anchorage, Alaska 99508, USA
4
CTSPS, Clark-Atlanta University, Atlanta, Georgia 30314, USA
5
School of Physics and Center for Relativistic Astrophysics, Georgia Institute of Technology, Atlanta,
Georgia 30332, USA
6
Department of Physics, Southern University, Baton Rouge, Louisiana 70813, USA
7
Department of Physics, University of California, Berkeley, California 94720, USA
8
Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA
9
Institut für Physik, Humboldt-Universität zu Berlin, D-12489 Berlin, Germany
10
Fakultät für Physik und Astronomie, Ruhr-Universität Bochum, D-44780 Bochum, Germany
11
Physikalisches Institut, Universität Bonn, Nussallee 12, D-53115 Bonn, Germany
12
Université Libre de Bruxelles, Science Faculty CP230, B-1050 Brussels, Belgium
13
Vrije Universiteit Brussel, Dienst ELEM, B-1050 Brussels, Belgium
14
Department of Physics, Chiba University, Chiba 263-8522, Japan
PHYSICAL REVIEW D 89, 102001 (2014)
1550-7998=2014=89(10)=102001(20) 102001-1 © 2014 American Physical Society

15
Department of Physics and Astronomy, University of Canterbury, Private Bag 4800,
Christchurch, New Zealand
16
Department of Physics, University of Maryland, College Park, Maryland 20742, USA
17
Department of Physics and Center for Cosmology and Astro-Particle Physics, Ohio State University,
Columbus, Ohio 43210, USA
18
Department of Astronomy, Ohio State University, Columbus, Ohio 43210, USA
19
Niels Bohr Institute, University of Copenhagen, DK-2100 Copenhagen, Denmark
20
Department of Physics, TU Dortmund University, D-442 21 Dortmund, Germany
21
Department of Physics, University of Alberta, Edmonton, Alberta, Canada T6G 2E1
22
Erlangen Centre for Astroparticle Physics, Friedrich-Alexander-Universität Erlangen-Nürnberg,
D-91058 Erlangen, Germany
23
Département de physique nucléaire et corpusculaire, Université de Genève,
CH-1211 Genève, Switzerland
24
Department of Physics and Astronomy, University of Gent, B-9000 Gent, Belgium
25
Department of Physics and Astronomy, University of California, Irvine, California 92697, USA
26
Laboratory for High Energy Physics, École Polytechnique Fédérale, CH-1015 Lausanne, Switzerland
27
Department of Physics and Astronomy, University of Kansas, Lawrence, Kansas 66045, USA
28
Department of Astronomy, University of Wisconsin, Madison, Wisconsin 53706, USA
29
Department of Physics and Wisconsin IceCube Particle Astrophysics Center, University of Wisconsin,
Madison, Wisconsin 53706, USA
30
Institute of Physics, University of Mainz, Staudinger Weg 7, D-55099 Mainz, Germany
31
Université de Mons, 7000 Mons, Belgium
32
T.U. Munich, D-85748 Garching, Germany
33
Bartol Research Institute and Department of Physics and Astronomy, University of Delaware, Newark,
Delaware 19716, USA
34
Department of Physics, University of Oxford, 1 Keble Road, Oxford OX1 3NP, United Kingdom
35
Department of Physics, University of Wisconsin, River Falls, Wisconsin 54022, USA
36
Oskar Klein Centre and Department of Physics, Stockholm University, SE-10691 Stockholm, Sweden
37
Department of Physics and Astronomy, Stony Brook University, Stony Brook,
New York 11794-3800, USA
38
Department of Physics, Sungkyunkwan University, Suwon 440-746, Korea
39
Department of Physics, University of Toronto, Toronto, Ontario, Canada, M5S 1A7
40
Department of Physics and Astronomy, University of Alabama, Tuscaloosa, Alabama 35487, USA
41
Department of Astronomy and Astrophysics, Pennsylvania State University, University Park,
Pennsylvania 16802, USA
42
Department of Physics, Pennsylvania State University, University Park,
Pennsylvania 16802, USA
43
Department of Physics and Astronomy, Uppsala University, Box 516, S-75120 Uppsala, Sweden
44
Department of Physics, University of Wuppertal, D-42119 Wuppertal, Germany
45
DESY, D-15735 Zeuthen, Germany
(Received 30 November 2013; published 1 May 2014)
We report on the search for neutrino-induced particle showers, so-called cascades, in the IceCube-40
detector. The data for this search were collected between April 2008 and May 2009 when the first 40
IceCube strings were deployed and operational. Three complementary searches were performed, each
optimized for different energy regimes. The analysis with the lowest energy threshold (2 TeV) targeted
atmospheric neutrinos. A total of 67 events were found, con sistent with the expectation of 41 atmospheric
muons and 30 atmospheric neutrino events. The two other analyses targeted a harder, astrophysical
neutrino flux. The analysis with an intermediate threshold of 25 TeV leads to the observation of 14
cascadelike events, again consistent with the prediction of 3.0 atmospheric neutrino and 7.7 atmospheric
muon events. We hence set an upper limit of E
2
Φ
lim
≤ 7. 46 × 10
−8
GeV sr
−1
s
−1
cm
−2
(90% C.L.) on the
diffuse flux from astrophysical neutrinos of all neutrino flavors, applicable to the energy range 25 TeV to
5 PeV, assuming an E
−2
ν
spectrum and a neutrino flavor ratio of 1∶1∶1 at the Earth. The third analysis
*
Also at Physics Departme nt, South Dakota School of Mines and Technology, Rapid City, SD 57701, USA.
†
Corresponding author.
stephanie.v.hickford@gmail.com
‡
Corresponding author.
eike.middell@desy.de
§
Also at NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA.
M. G. AARTSEN et al. PHYSICAL REVIEW D 89, 102001 (2014)
102001-2

utilized a larger and optimized sample of atmospheric muon background simulation, leading to a higher
energy threshold of 100 TeV. Three events were found over a background prediction of 0.04 atmospheric
muon events and 0.21 events from the flux of conventional and prompt atmospheric neutrinos. Including
systematic errors this corresponds to a 2.7σ excess with respect to the background-only hypothesis.
Our observation of neutrino event candidates above 100 TeV complements IceCube’s recently observed
evidence for high-energy astrophysical neutrinos.
DOI: 10.1103/PhysRevD.89.102001 PACS numbers: 95.85.Ry, 14.60.Lm, 29.40.Ka
I. INTRODUCTION
One century after the discovery of cosmic rays the search
for their sources is still ongoing. Astrophysical objects
which are either confirmed or expected to be able to
accelerate hadrons to the observed energies include super-
nova remnants [1], active galactic nuclei, gamma-ray bursts,
and shocks in star formation regions of galaxies. The cosmic-
ray nuclei interact with ambient matter and radiation fields
close to their source [2]. Charged pions produced in these
interactions decay into neutrinos. Therefore, the detection of
high-energy neutrinos from such objects provides a unique
possibility to identify individual astrophysical objects as
cosmic-ray sources. However, low fluxes and small inter-
action probabilities make the detection of high-energy
neutrinos challenging. To date no astrophysical object has
been conclusively identified as a source of TeV neutrinos.
Previous searches have established limits enabling astro-
physical models to be constrained [3,4].
While individual neutrino sources might be too weak
to be detectable with current instruments, they would still
contribute to a collective astrophysical neutrino flux. Fermi
shock acceleration is thought to be the main acceleration
mechanism for cosmic-ray nuclei and therefore a power-
law spectrum with an index of about −2 is expected for
the nuclei in the interaction regions where the neutrinos are
produced. Based on the energy density of ultrahigh-energy
cosmic rays and assuming the cosmic-ray sources are
transparent, the all-flavor diffuse neutrino flux can be
constrained theoretically to be lower than the Waxman-
Bahcall bound of E
2
ν
Φ⪅3 × 10
−8
GeV sr
−1
s
−1
cm
−2
[5,6].
As neutrinos are assumed to originate mainly from pion
decays, at the source a flavor ratio of ν
e
∶ν
μ
∶ν
τ
¼ 1∶2∶0 is
expected. This ratio would transform to 1∶1∶1 on Earth due
to neutrino oscillations [7,8]. However, observing unequal
or energy-dependent flavor contributions would be inter-
esting, since for example the flavor ratio is sensitive to the
assumed production mechanism at the source [9].
Recently, evidence for this diffuse astrophysical neutrino
flux was found. Its all-flavor intensity is estimated to be
E
2
ν
Φ ¼ð3.6  1.2Þ × 10
−8
GeV sr
−1
s
−1
cm
−2
with indica-
tions for a cutoff at ∼2 PeV. It is consistent with an isotropic
flux and a flavor ratio of 1∶1∶1 ([10,11] and Fig. 1).
In order to measure the diffuse astrophysical neutrino
flux at TeV energies, it has to be separated from two main
sources of background, which both originate from the
Earth’s atmosphere. These are atmospheric muons and
neutrinos produced in cosmic-ray air showers. The atmos-
pheric neutrino flux has two components. The so-called
conventional atmospheric neutrinos are produced in decays
of pions and kaons. Their intensity is well measured up to
6 TeV for ν
e
and up to 400 TeV for ν
μ
[12,13]. At higher
energies the poor knowledge of the composition of the
cosmic-ray flux, creating the neutrinos, causes significant
uncertainties on the intensity. The spectrum of the conven-
tional atmospheric neutrinos is steeper than the cosmic-ray
spectrum due to pion and kaon energy losses in the
atmosphere. The second component originates from the
decay of charmed mesons, which have lifetimes several
orders of magnitude smaller than charged pions and kaons.
Accordingly, neutrinos from these decays are called prompt
atmospheric neutrinos. Because of the short lifetime of the
parent mesons the energy spectrum of the prompt atmos-
pheric neutrinos is expected to follow the spectrum of the
cosmic rays that create them. However, their intensity
has never been measured and uncertainties in the relevant
FIG. 1 (color online). Neutrino energy spectrum above
100 GeV. The theoretical predictions and measurements for
the atmospheric neutrino flux are shown, as well as the current
estimate for the diffuse astrophysical neutrino flux.
SEARCH FOR NEUTRINO-INDUCED PARTICLE SHOWERS … PHYSICAL REVIEW D 89, 102001 (2014)
102001-3

production cross sections lead to large uncertainties in the
predicted flux. The presence of a prompt neutrino compo-
nent, like an astrophysical neutrino component, introduces
a break into the neutrino energy spectrum. Given the large
uncertainties in the prompt neutrino predictions, identifi-
cation and separation of the astrophysical and prompt
components needs to be made through their respective
spectral signatures (see Fig. 1).
The IceCube Neutrino Observatory is located at the South
Pole and is the first kilometer-scale Cherenkov neutrino
telescope. An optical sensor array observes the Cherenkov
radiation from secondary charged particles produced in
neutrino interactions deep in the ice. These are dominantly
neutrino-nucleon interactions except for the Glashow reso-
nance [14] for electron antineutrinos at 6.3 PeV. Based on the
signature of the neutrino interaction, which depends on the
flavor of the incident neutrino and the type of the interaction,
two main detection channels exist. Searches in the muon
channel look for charged-current muon neutrino interactions.
These have a muon in the final state whose direction is
reconstructible with a resolution of about 1° [15]. The large
muon range also allows one to detect neutrino interactions
outside the instrumented volume. The cascade channel
comprises all other interaction scenarios which have particle
showers in the final state. Above PeV energies charged-
current ν
τ
interactions exhibit more complex event signa-
tures, for which tailored analyses are developed. But at TeV
energies ν
τ
can also be detected through the cascade channel.
Consequently, an astrophysical flux with equal neutrino
flavor contributions would yield more cascade than track
events starting inside of IceCube. If the neutrino interaction
happens inside the detector, the Cherenkov light yield of
particle showers scales nearly linearly with the deposited
energy, leading to an energy resolution that is better than in
the muon channel. On the other hand the angular resolution
is rather poor (> 10° for the completed IceCube detector).
Overall, the cascade channel is best suited for searches
for diffuse astrophysical neutrinos in which the neutrino
energy measurement is more important than pointing
capabilities [16].
This paper presents searches for neutrino-induced
cascades in one year of data taken during the construction
phase of IceCube, when about half the detector was opera-
tional (IceCube-40). The main objective of the searches was
to identify an astrophysical flux of neutrinos. In addition, a
sensitivity to atmospheric neutrinos in the few TeV energy
range was maintained to allow a validation of the anticipated
backgrounds in the data set.
A 2.7σ excess of events above 100 TeV was found,
compatible with the all-flavor astrophysical diffuse neu-
trino flux estimate obtained in IceCube’s high-energy
starting events analysis [11]. In comparison to that analysis,
the IceCube-40 cascade analysis provides an event sample
with unprecedented low background contamination
between 100 and 200 TeV. This is possible because both
searches employed rather different event selection strate-
gies. Methods outlined in this paper also prove powerful in
cascade searches with later IceCube configurations [17,18].
The paper is organized as follows: The IceCube detector
and IceCube-40 data set are described in Sec. II.The
simulation used is presented in Sec. III followed by a
description of the cascade reconstruction in Sec. IV.The
details of the event selection and expected sensitivity are
presented in Sec. V. A survey of the systematic uncertainties
follows in Sec. VI before the results and implications are
discussed in Sec. VII. A conclusion is given in Sec. VIII.
II. THE ICECUBE DETECTOR
The IceCube Neutrino Observatory [19] consists of an
in-ice array of optical sensors and a complementary surface
air shower detector called IceTop. The analyses presented
here utilized only the in-ice component so the following
detector description will be limited to that.
The optical sensors, called digital optical modules
(DOMs) [20], are sensitive to Cherenkov photons between
350 and 650 nm. The DOMs are deployed between depths
of 1450 and 2450 m and are attached to strings that are
formed by the readout cables. Each string has 60 DOMs
attached. The vertical string spacing of the DOMs is
approximately 17 m and the horizontal spacing between
the strings is approximately 125 m. The data for this analysis
were collected between April 2008 and May 2009 with a
total of 367.1 days live time. In this period 40 strings were
deployed and operational. The detector layout is shown in
Fig. 2. Before IceCube’s completion in 2010, the IceCube-
59 and IceCube-79 configurations took data with 59 and 79
deployed strings, respectively.
Each DOM consists of a 25 cm diameter photomultiplier
tube (PMT) [21], made by Hamamatsu Photonics, and a
data acquisition system housed within a pressure sphere
made of 13 mm thick borosilicate glass. The PMT’s dynamic
range is 200 photoelectrons per 15 ns and it is designed to
accurately record the amplitudes and widths of the pulses
with a timing resolution of 5 ns. Their peak quantum
efficiency is approximately 25% and they are operated at a
gain of 10
7
to resolve single photoelectrons.
The time-resolved PMT signal (waveform) is digitized in
the DOM. For this purpose two digitization devices are
available on the DOM mainboard: two analog transient
waveform digitizers (ATWDs) and a fast analog-to-digital
converter (fADC). The ATWDs have three channels oper-
ated in parallel at different gains to provide a large dynamic
range (a fourth channel is used only for calibration purposes).
Because of scattering in the ice the arrival times of photons
emitted at the same point and time can vary by microseconds.
The ATWDs provide a sampling rate of 300 Megasamples=s
over a time window of 425 ns allowing them to record the
earliest photons (i.e. those least affected by scattering in the
ice) with high precision. The second digitizer, the fADC, has
a coarser sampling of 40 Megasamples=s recording data over
M. G. AARTSEN et al. PHYSICAL REVIEW D 89, 102001 (2014)
102001-4

a longer time period for photons with larger delays of
up to 6.4 μs. In order to reduce data readout volume due
to noise, in IceCube-40 a local coincidence criterion is
required. Only if a neighboring DOM on the same string
also detects light within the local coincidence timewindow of
1000 ns, the PMT response is digitized, time stamped, and
transmitted to the surface for analysis. The surface data
acquisition system combines the individual PMT responses
and forms events when one of the several possible triggering
criteria is fulfilled.
The trigger requirement for the IceCube-40 cascade
search was the simple multiplicity trigger, which requires
that eight DOMs were hit within a 5000 ns time window.
The data rate for IceCube-40 from this trigger was
approximately 1000 Hz.
III. SIMULATION
Interactions of all flavors of neutrinos were simulated
to model atmospheric and astrophysical neutrinos. The
N
U
GEN software package maintained by the IceCube
Collaboration was used. It is based on the ANIS [22]
neutrino generator, which produces neutrinos isotropically
over the Earth’s surface and propagates them to interact
in or near the detector volume. Neutrino attenuation and
ν
τ
regeneration are accounted for using the preliminary
reference Earth model [23]. CTEQ5 structure functions
[24] were used to model the deep-inelastic neutrino-
nucleon scattering cross section.
Throughout this paper the diffuse astrophysical neutrino
flux is simulated isotropically, with a flavor ratio of 1∶1∶1
and, if not stated otherwise, with an unbroken power-law
spectrum with index of − 2 and an all-flavor intensity
of 3.6 × 10
−8
GeV sr
−1
s
−1
cm
−2
.
Rate predictions for the atmospheric neutrinos are based
on the HKKMS07 model [25] for conventional atmos-
pheric neutrinos and the ERS model [26] for prompt
atmospheric neutrinos. Extrapolations of the original cal-
culations to higher energies provide rate predictions at the
energy range relevant to this work. The steepening of the
cosmic-ray spectrum around a few PeV (the so-called
“knee”) causes a similar feature in the atmospheric neutrino
spectrum which is not accounted for in the HKKMS07
model. A modification to the HKKMS07 model [27,28]
was applied to account for the knee. For one of the presented
analyses the Bartol model [29] was used to estimate the
conventional atmospheric neutrino flux. Compared to the
modified HKKMS07 model it predicts a higher ν
e
contri-
bution (see Fig. 1).
The propagation of muons and taus through the detector
and their energy losses were simulated using the MMC
program [30] and the cascade simulation inside the detector
was handled by the CMC program [31]. Neutrino-induced
cascades below a threshold of 1 TeV were simulated as
pointlike light sources, emitting an angular Cherenkov light
profile typical of an electromagnetic shower [32]. Cascades
of higher energies are split into segments along the direction
of the shower development. Each cascade segment is then
approximated by a pointlike subshower with a light yield
proportional to the light yield in the corresponding segment
of the electromagnetic cascade. The elongation of electro-
magnetic cascades due to the suppression of bremsstrahlung
and pair production cross sections above PeV energies
(Landau-Pomeranchuck-Midgal effect [33]) is accounted
for. Hadronic cascades are simulated as electromagnetic
cascades with a smaller light yield per deposited energy to
account for the neutral shower components which do not
generate Cherenkov light [34].
The contribution from atmospheric muon events is
estimated from simulations done with a modified version
[35,36] of the CORSIKA air shower simulation software
[37]. A large number of background events must be
generated due to the high background suppression that is
necessary to reach an event sample dominated by neutrinos.
Providing a large background sample is computationally
challenging, mostly because of the sheer number of air
showers needed but also due to the simulation of light
propagation in the optically inhomogeneous ice. The figure
of merit used to quantify the statistics of a simulated data
sample is the effective live time T
eff
, i.e. the time that one
would have to run the real experiment to obtain the same
statistical error as in the simulated data set.
The chemical composition of the cosmic rays is impor-
tant for an estimate of the muon background. Previous
FIG. 2 (color online). The IceCube-40 detector configuration as
viewed from above. The circles and squares are the positions of
the strings, which are deployed vertically into the ice. The point
ðx; yÞ¼ð0; 0Þ is the cent er of the complete 86-string detector.
Particle showers with reconstructed vertices inside the instru-
mented volume are called contained events. The analyses
presented in this work reject noncontained events in order to
suppress atmospheric muons that enter the detector from the
outside. The blue dashed and solid lines show the two differently
tight contai nment requirements that are used. The strings denoted
by red squares form the outer layer of the detector. They are used
to veto incident atmospheric muons.
SEARCH FOR NEUTRINO-INDUCED PARTICLE SHOWERS … PHYSICAL REVIEW D 89, 102001 (2014)
102001-5