On this page
Objective
Detect high-energy astrophysical and atmospheric neutrinos, and use the Earth as an analyzing medium for oscillation and exotic-physics studies.
Method
A cubic-kilometer array of 5,160 digital optical modules deployed along 86 vertical strings at depths of 1,450–2,450 m in the Antarctic ice. Each module records Cherenkov light from charged particles produced in neutrino interactions in or near the instrumented volume. Muon tracks give directional information; electromagnetic and hadronic cascades give better energy resolution.
Key results
- 2013: first observation of high-energy astrophysical neutrinos — a diffuse flux above the atmospheric background.
- 2017: real-time identification of the blazar TXS 0506+056 as a likely cosmic-neutrino source, with coincident gamma-ray observation.
- 2022: identification of the nearby active galaxy NGC 1068 as a neutrino point source at 4.2σ.
- Galactic-plane neutrino emission detected at 4.5σ (2023).
- Standard oscillation measurements through atmospheric neutrinos at IceCube-DeepCore.
- Stringent constraints on keV-scale sterile neutrinos, cosmic-ray composition, and dark-matter annihilation.
Significance
IceCube opened neutrino astronomy as a multi-messenger discipline. Cosmic neutrinos complement photons (which scatter and are absorbed) and cosmic rays (which lose directional information in galactic magnetic fields) as messengers from the high-energy universe. The observatory has also become a competitive facility for oscillation physics, sterile searches, and indirect dark-matter detection.
Why the South Pole
Deep Antarctic ice is the clearest natural medium at scale, with optical attenuation length exceeding 200 m in the blue at depths below 1400 m. The ice is stable over tens of thousands of years, dust-banded but locally transparent, and sits above bedrock that shields upward-going atmospheric muons. The Amundsen–Scott Station provides year-round access.
Alternative sites — deep ocean (ANTARES, KM3NeT in the Mediterranean; Baikal-GVD in Russia) — exploit similar physics but in salt water, trading optical stability for freedom from the bubble-content of ice. Cross-calibration between these platforms is an active research area.
The instrumented array
IceCube consists of 86 strings, each carrying 60 Digital Optical Modules (DOMs), spaced every 17 m between depths of 1,450 and 2,450 m. The DOMs — each a 25-cm photomultiplier in a pressure-sphere with self-contained digitization electronics — were deployed by drilling holes with a hot-water drill, lowering strings into the water-filled column, and letting the ice refreeze around them.
The deployment took seven seasons (2004–2010). Construction was completed on 18 December 2010.
DeepCore and IceCube-Upgrade
At the bottom center of the array, a denser sub-array known as DeepCore lowers the energy threshold from ~100 GeV to ~10 GeV, enabling atmospheric neutrino oscillation measurements competitive with accelerator long-baseline experiments. The IceCube-Upgrade, under construction, adds seven further strings to further densify DeepCore and improve ice-model calibration, targeting sub-GeV thresholds for precision oscillation physics and improved directional reconstruction.
Discoveries
Astrophysical diffuse flux (2013). Analysis of very-high-energy “starting events” (vertex contained in the detector) revealed an excess above the expected atmospheric flux beginning around 30 TeV. The 2013 publication of 28 events over 988 days — including two at PeV energies, dubbed “Ernie” and “Bert” — was the first detection of a high-energy astrophysical neutrino flux.
TXS 0506+056 (2017). A 290 TeV muon-neutrino alert on 22 September 2017 was followed within minutes by a multi-wavelength follow-up campaign. The gamma-ray satellite Fermi and the imaging Cherenkov telescope MAGIC identified the source as the blazar TXS 0506+056 in a flaring state. An archival IceCube search uncovered a 2014–2015 flare of neutrinos from the same source at 3.5σ. The combined 2017 real-time + archival evidence — 3–4σ each — made this the first plausible identification of a cosmic-neutrino source.
NGC 1068 (2022). A point-source search using 10 years of through-going muon tracks identified the nearby Seyfert 2 galaxy NGC 1068 as the strongest hot spot in the sky, at 4.2σ after trials correction. The energy spectrum favors production in the obscured nuclear core, ruling out the jet as the dominant source.
Galactic plane (2023). A neural-network-based search for cascade-topology events identified a diffuse galactic neutrino emission at 4.5σ — the first detection of the Milky Way in neutrinos.
Oscillation and exotic-physics reach
Atmospheric neutrinos at DeepCore provide a measurement of and that is systematics-limited rather than statistics-limited, complementary to Super-K and accelerator measurements. IceCube has set the most stringent limits on eV-scale sterile neutrinos across a substantial parameter range, and on TeV-PeV dark matter annihilation in the Sun and galactic center.
IceCube-Gen2
The planned IceCube-Gen2 will expand the instrumented volume to 8 km³, add a surface cosmic-ray veto, and deploy a radio-Cherenkov array for ultra-high-energy (EeV) neutrino detection. First deployment is expected later this decade; full operation by the late 2030s.