doi:10.1016/j.nima.2004.07.119

Nuclear Instruments and Methods in Physics Research A 535 (2004) 139–142

Design and status of IceCube

Martin Kestel

104 Davey Lab, PMB 29, State College, Penn State University, University Park, PA 16802, USA

For the IceCube collaboration

Available online 12 August 2004

Abstract

IceCube is a kilometer-scale high-energy neutrino detector that builds on the wealth of experience accumulated with

its smaller predecessor, AMANDA. An international collaboration has begun construction of key components of the

IceCube detector and deployment operations at the South Pole will begin in late 2004.

The underlying design of the IceCube detector and of the DAQ system are presented here, emphasizing

the digital optical modules (DOMs) as the smallest discrete IceCube building block. The event reconstruction

critically relies on a relative timing accuracy from DOM to DOM of a few nanoseconds over inter-DOM separations of

up to 1 km.

1. Physics goals

Through the detection of very high-energy

neutrinos (threshold a few 100 GeV), IceCube

[2,3] will open a new window on the universe. By

viewing astronomical sources with neutrinos as

astronomical messengers, it will address funda-

mental questions in high-energy astrophysics,

particle physics and cosmology. Through the

detection of surface electrons and muons, the

associated IceTop surface array will allow us to

study the chemical composition of high-energy

cosmic rays

and will also help

calibrate IceCube and provide a background

veto. IceCube and underwater neutrino telescopes

[4]

share scientific interests, such as searches

for steady or variable neutrino emission from

point-like source candidates like active galactic

nuclei (AGN), supernova remnants (SNR),

microquasars and gamma ray bursts (GRB). By

virtue of the low ambient noise level in the ice,

the ability to detect low-energy supernova neu-

trinos as an increase in the overall trigger rate is

unique to IceCube among all UHE neutrino

detectors. On the more speculative side, searches

for neutrinos from annihiliations of weakly inter-

acting massive particles (WIMPs), for magnetic

monopoles and other exotic particles like strange

quark matter or SUSY Q-balls can be listed (see

e.g. Refs. [5,6]).

ARTICLE IN PRESS

www.elsevier.com/locate/nima

0168-9002/$ - see front matter

doi:10.1016/j.nima.2004.07.119

Email address:

mka@phys.psu.edu (M. Kestel).

See

[1]

for a full author list.

2. Detector design and status

IceCube will consist of 4800 digital optical

modules (DOMs), organized in 80 strings, each

with 60 DOMs attached, buried in the ice at depths

of 1450–2450 m. DOMs will have a vertical

spacing of 17 m and the strings will be regularly

spaced horizontally by 125 m. At each string

location two IceTop tanks, each containing two

DOMs frozen in ice, will be deployed. The buried

DOMs will have an effective surface area of

around 1 km

;

promising optimal sensitivity for

neutrinos in the energy range of 1–10 000 TeV

while being able to trigger on all higher- and on

some of the lower-energy neutrinos, including

MeV bursts [8]. The positions of IceCube strings,

and the IceTop tanks deployed above them, are

shown in

Fig. 1. Simulations have shown that

IceCube’s sensitivity to possible signals is roughly

constant for a wide range of feasible configura-

tions.

DOMs form the fundamental building blocks of

the IceCube detector. Each DOM contains a 10

Hamamatsu R-7081 photo multiplier (PMT). The

high voltage for the PMT is converted in the DOM

from its 48 V DC power supply to achieve the

design gain of around 5

Within a DOM, the

PMT signal is split into two copies, with one used

for triggering and the other delayed and then

digitized if the threshold condition is met. Digiti-

zation occurs in two types of DOM-resident

digitizers, to extend the digitization time while

keeping the resolution high at early times. There is

a set of two four-channel ASIC analog transient

waveform digitizers (ATWDs) and a commercial

40 MHz FADC with up to 256 samples and 16-bit

resolution available. The two ATWDs operate at

300 MHz and at none, 16, 32, 64 or 128 sample

depths with 8- or 16-bit resolution. The two

ATWDs are fed signals in a ping-pong manner,

reducing DOM dead time to less than 1%. The

first three channels of each of the ATWDs are fed

signals that have been amplified with factors of 16,

4 and 2/3. This combination of ATWDs and

FADC ensures the design dynamic range of up to

200 photo electrons (p.e.) within the first 15 ns and

up to 2000 p.e. within the first 5

The fourth

ATWDchannel can be connected to various

inputs like the DOM-clock ticks or LED driving

currents, creating a versatile diagnosis and cali-

bration tool. The DOMs further contain a 405 nm

LEDflasher board, producing programmable light

flashes of various intensities detectable by other

modules in the array. This capability is useful for

studying ice properties and calibrating the relative

positions of DOMs. A preliminary version of a

DOM, deployed in an IceTop tank in January

2004, is shown in Fig. 2.

The DOM mainboard has a free-running timer

which needs to be synchronized with nanosecond

accuracy to GPS time, requiring re-calibration

roughly every minute. As shown in

Fig. 3,a

surface circuit sends a bipolar signal at a GPS-

latched time

;

received at a time

After a

certain, fixed time interval

;

an identical circuit in

the DOM sends an identical bipolar pulse to the

ARTICLE IN PRESS

Fig. 1. Aerial view of South Pole and positions of the IceTop tanks resp. the IceCube strings (black), Spase-2 stations (grey, dense,

regular foreground pattern

[7]) and the AMANDA strings (larger grey pattern). Courtesy V. Papitashvili.

M. Kestel / Nuclear Instruments and Methods in Physics Research A 535 (2004) 139–142

140

surface, detected at a time

The cable transmis-

sion time is then:

Down

¼ð

This calibration reduces signal time spread to the

inevitable contribution from light scattering in the

Antarctic ice.

Currently a fully digital, TCP/IP-based ap-

proach for the DAQ system is under development,

following closely the modular structure of the

experimental setup: each

String Processor

stores

DOM-data and passes trigger primitives on to the

InIce Trigger

, which, after examining trigger

primitives from all

String Processors

, sends its

trigger decisions to the

Global Trigger

. The

Global

Trigger

combines

InIce Trigger

IceTop Trigger

and other (external) information to form its

decision. If positive, the

Event Builder

is instructed

to retrieve DOM data from the

String Processors

and assembles them to IceCube events that get

passed to the

Online Filter Cluster

for further

processing. All of these DAQ system elements are

implemented in commercial computers.

The drilling process has been improved in

several aspects compared with the AMANDA

procedure: setup time for a season will be only 3–5

weeks; 60 cm diameter holes will be drilled with

water of 90

C from a number of heaters with a

total power of 5 MW (vs. 2 MW for AMANDA), a

larger hose diameter will reduce drill time to 40 h,

and the fuel consumption will be lowered by about

30%. With an estimated string drop time of

around 20 h, it should be possible to deploy 16

or more strings per austral summer season, leading

to a construction time of 5–6 years for the entire

detector.

3. Summary

With the assembly and testing of the first batch

of DOMs under way, the IceCube collaboration is

on track for deployment of the first set of strings at

the end of 2004. The digital approach to readout

and triggering, together with the sophisticated

time calibration, will help to overcome the

challenges posed by the sheer size of the detector

and the time spreads induced by the Antarctic ice

as a detector medium, enabling IceCube to

produce useful data for scientific purposes just

after the first few deployments.

Acknowledgements

This research was supported by the following

agencies: National Science Foundation—Office of

Polar Programs, National Science Foundation—

Physics Division, University of Wisconsin Alumni

Research Foundation, USA; Swedish Research

Council, Swedish Polar Research Secretariat,

Knut and Alice Wallenberg Foundation, Sweden;

German Ministry for Education and Research,

Deutsche Forschungsgemeinschaft (DFG), Ger-

many; Fund for Scientific Research (FNRS-

FWO), Flanders Institute to encourage scientific

and technological research in industry (IWT),

ARTICLE IN PRESS

Fig. 2. The first DOM frozen into a prototype IceTop tank at

South Pole (January 2004). Photo by John Kelley/NSF.

DOM Time

DOM

sends

DOM

receives

GPS Time

Surface receives

Surface

sends

down

T fixed interval

Fig. 3. DOM time calibration, see text.

M. Kestel / Nuclear Instruments and Methods in Physics Research A 535 (2004) 139–142

141

Belgian Federal Office for Scientific, Technical and

Cultural affairs (OSTC), Belgium; Inamori Science

Foundation, Japan; FPVI, Venezuela; The Nether-

lands Organization for Scientific Research

(NWO).

References

[1] J.A. Ahrens, et al., Astroparticle Phys. 20 (2004) 507.

[2] IceCube Homepage, http://icecube.wisc.edu.

[3] S. Yoshida, for the IceCube Collaboration, Proceedings of

the 28th ICRC, 2003, p. 1369.

[4] Antares,

http://antares.in2p3.fr/; Baikal,

http://www-

zeuthen.desy.de/baikal/baikalhome.html; Nemo,

http://ne-

moweb.lns.infn.it/ Nestor, http://www.nestor.org.gr/.

[5] C. Spiering, for the Amanda collaboration, Proceedings of

the 27th ICRC, 2001, p. 1242.

[6] F. Halzen, astro-ph/0311004.

[7]

http://ast.leeds.ac.uk/haverah/spase2.shtml.

[8] M. Leuthold, Proceedings of the Workshop on Large

Neutrino Telescopes, Zeuthen, 1998,

http://www.ifh.de/

nuastro/publications/conferences/proc.shtml.

ARTICLE IN PRESS

M. Kestel / Nuclear Instruments and Methods in Physics Research A 535 (2004) 139–142

142