Non-linear Energy Calibration for Particle Detectors

This direction is split into other directions: scintillator non-linearity. Its last version stays readable and citable here.

Version 1 of 121 papers35 conceptsfrozen 2026-10-04

In an ideal world, a particle detector would be a perfect ruler: if a particle with twice the energy hits the sensor, the sensor should produce exactly twice the signal. In reality, detectors are messy. The relationship between the energy deposited and the digital signal recorded is often warped by the physics of the sensing material and the quirks of the electronics.

To reach the sub-percent precision required to solve mysteries like the neutrino mass hierarchy or the existence of sterile neutrinos, researchers must move beyond simple linear scaling. They must build sophisticated models that account for the "non-proportionality" of the detector’s response.

The Physics of Non-proportionality

The first hurdle in energy reconstruction is that the sensing medium itself—often a liquid or crystal scintillator—does not produce light linearly.

At low energies, a phenomenon known as quenching dominates. When a particle deposits energy in a small volume, the density of excited states becomes so high that they interfere with one another, "quenching" the light output. As Songbai Zhang and Y. Qin note in their work on neutron spectra, this makes it particularly difficult to distinguish true signal from instrumental noise at the lower bounds of a detector's range.

Furthermore, the atomic structure of the detector material introduces specific "dips" in light production. Jiaxing Wen and Chao Zheng have shown that scintillation crystals exhibit non-linearities near the K-shell binding energies of their constituent elements, such as Iodine or Lanthanum. When an incoming photon has just enough energy to eject a K-shell electron, the resulting cascade of secondary particles changes the light yield, creating a localized non-linearity that must be modeled to prevent systematic errors in spectral analysis.

Decoupling Physics from Electronics

Even if the light production were perfectly linear, the electronics used to read out the signal introduce their own distortions.

The Daya Bay collaboration has pioneered methods to decouple these effects. They distinguish between:

  1. Scintillator non-linearity: Caused by quenching and Cherenkov radiation.
  2. Electronics non-linearity: Caused by the Analog-to-Digital Converters (ADCs) and signal processing chains.

In experiments like the MAJORANA DEMONSTRATOR, researchers like N. Abgrall and I. J. Arnquist have found that ADC artifacts can shift or broaden the narrow energy peaks they are looking for. By implementing comprehensive non-linearity correction protocols, they can reduce systematic uncertainties by an order of magnitude.

At the extreme end of the scale, high-energy particles can "saturate" the electronics, clipping the tops of signal waveforms. J. J. Zhu and Neil Roberts have demonstrated that artificial neural networks can be trained to "recover" these saturated waveforms by learning the underlying pulse shapes that the hardware failed to capture.

Self-Calibration and Spectral Anchors

How do we calibrate a detector when the "true" energy is what we are trying to measure? One powerful intuition is to use the physics of the signal itself as a reference.

Jun Cao and Y. F. Wang proposed a self-calibration mechanism for reactor neutrino experiments. Because an oscillation spectrum contains multiple peaks, and non-linearity affects those peaks differently, the detector's response function can be treated as a fit parameter. The "wiggles" in the data act as anchors, allowing the detector to calibrate itself against the known laws of physics.

From Linear Sums to Machine Learning

Traditional energy reconstruction often relies on a simple sum of "hits" or light pulses. However, this fails when energy "leaks" out of the detector or when different types of particles (like electrons vs. hadrons) create different shower shapes.

In the NOvA experiment, researchers have moved toward Convolutional Neural Networks (CNNs) to classify particle tracks and apply specific non-linear corrections for different particle types. Similarly, Suzuna Morimasa and Taikan Suehara are developing machine learning models for the International Linear Collider that map raw, low-level detector data directly to energy. These models naturally compensate for non-linearities like energy leakage and geometric biases that simple linear summation cannot handle.

By treating energy reconstruction as a complex, non-linear regression task rather than a simple counting problem, these researchers are pushing the precision of particle physics into a new era.

Papers in this version

History & citationVersion 1 · 2026-10-04

Split into other directions · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_non_linear_energy_calibration_for_partic_2fbfe4c4,
  title = {Non-linear Energy Calibration for Particle Detectors},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/non-linear-energy-calibration-for-particle-detectors/rdr_2fbfe4c470584de0942e324c8ff3f9c0},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_non_linear_energy_calibration_for_partic_2fbfe4c4_v1,
  title = {Non-linear Energy Calibration for Particle Detectors},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/non-linear-energy-calibration-for-particle-detectors/rdr_2fbfe4c470584de0942e324c8ff3f9c0/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}