Particle identification (PID)
Determining which type of particle produced a signal.
Methods include Cherenkov ring shape (sharp µ-like vs fuzzy e-like), dE/dx in TPCs, range-momentum relations, pulse shape in scintillators and germanium, and multivariate or deep-learning classifiers. PID enables flavour identification (νe vs νµ CC) and background rejection.
In plain words
A neutrino detector records flashes of light, pulses of charge or tiny amounts of heat. On their own these signals do not say which particle caused them. Particle identification is the detective work that turns a pattern of hits into a label: muon, electron, photon, proton, neutron, alpha particle or nuclear recoil. Without it, a detector could count events but could not tell which neutrino flavour arrived or whether the event was a neutrino at all.
The main techniques
- Cherenkov patterns: muons give sharp rings, electrons fuzzy ones; this is ring imaging in water detectors.
- Ionisation density (dE/dx): in a liquid-argon TPC a stopping proton deposits far more energy per centimetre than a muon, and a single electron leaves half the charge of a converted photon pair, the basis of electron–photon separation.
- Range versus momentum: in tracking calorimeters muons travel much farther than pions of the same energy, which shower hadronically.
- Time profile of the signal: pulse-shape discrimination separates alphas and nuclear recoils from electrons in scintillators, and single-site from multi-site events in germanium.
- Delayed tags: a Michel electron a few microseconds later marks a stopped muon, and a captured neutron marks inverse beta decay (neutron tagging).
From cuts to classifiers
Early experiments applied simple cuts on one or two variables. Modern analyses combine dozens of variables in likelihood ratios, boosted decision trees or neural networks. NOvA introduced a convolutional neural network that classifies whole event images, and similar deep-learning methods are now standard in liquid-argon experiments such as MicroBooNE and in IceCube.
Why it matters for neutrino physics
Flavour is only visible through the charged lepton. Measuring appearance of νe in a νµ beam requires picking out a few electron events among many muons and neutral-current π⁰ events, which makes PID one of the leading inputs to CP-violation searches. In double beta decay and dark-matter-style searches, PID is mainly a tool for background rejection.
Frequently asked questions
What is the hardest particle-identification problem in accelerator experiments?
Separating electrons from photons. A neutral-current interaction can produce a π⁰ whose decay photons shower just like an electron. Water detectors look for a second ring, while liquid-argon TPCs measure the charge in the first few centimetres and the gap between the vertex and the shower start.
Is particle identification the same as flavour identification?
Not quite. Particle identification labels the particles in an event. Flavour identification is the next step: inferring from the charged lepton which neutrino flavour interacted. Neutral-current events carry no flavour information at all.
Go deeper
- Cherenkov ring imaging (e-like vs µ-like)Classifying events by the pattern of their Cherenkov ring on the detector wall.
- Pulse-shape discriminationSeparating particle types by the time profile of their scintillation or charge signal.
Related terms
- Electron–Photon separation (dE/dx)
- Pulse-shape discrimination
- Cherenkov ring imaging (e-like vs µ-like)
People
On the timeline
Mentioned in 20 pages
Comparisons 2
Milestones 2
Last updated: · Editorial standard
Cite this page
Text
Neutrino Atlas (2026). “Particle identification (PID)”. Retrieved 22 November 2026, from https://neutrino-atlas.com/terms/particle-identification/BibTeX
@misc{neutrinoatlas2026msparticleidentification,
title = {Particle identification (PID)},
author = {{Neutrino Atlas}},
year = {2026},
url = {https://neutrino-atlas.com/terms/particle-identification/},
note = {Accessed 2026-11-22},
language = {english}
}