33.4. Scintillation Light

Table of Contents

Basic Idea of Scintillation in Geant4

Scintillation is the process where deposited energy in a material is converted into optical photons. In Geant4, scintillation is a physical process attached to particles, but its behavior is controlled by the optical properties of the material.

When a charged particle traverses a scintillator, it loses energy through ionization and excitation. A fraction of this energy is released as photons in the optical energy range. Geant4 models this by generating optical photons with a spectrum, yield, and time structure that you define through material properties.

In Geant4, scintillation is entirely controlled by material properties and the scintillation process in the physics list. If either is missing or incomplete, you will not get the expected scintillation photons.

Enabling Scintillation in the Physics List

To obtain scintillation photons, you must include the scintillation process as part of your physics configuration. Many reference physics lists that include optical physics already provide scintillation. For custom physics, you typically use an optical physics constructor that registers the scintillation process for relevant particles.

Scintillation is usually applied to charged particles, such as electrons, positrons, and protons, since they directly excite the medium. When the process is active, each step that deposits energy in a scintillating material can produce a number of optical photons according to the material’s scintillation yield and your chosen configuration.

You can also configure whether secondary optical photons are tracked with full detail or limited through production cuts and other performance settings. This becomes important when you simulate large detectors that produce many photons.

Material Properties for Scintillation

Geant4 describes scintillation behavior using entries in the material properties table associated with a G4Material. These properties tell Geant4 how many photons are produced per unit energy, what their spectrum is, and how they are emitted in time.

The key concept is the G4MaterialPropertiesTable that you attach to a material. For scintillation you typically define several named properties. Some are constant values, and some are arrays that give values as a function of photon energy.

A common minimal setup requires:

The scintillation photon spectrum.
The total light yield per unit energy.
The time constants that describe the decay of the light pulse.

Without these, scintillation will either not occur or will not behave realistically.

For all optical properties, photon energy (not wavelength) is the independent variable. Values are given as arrays of energy points with corresponding property values. All arrays for a given material must use the same energy grid and length.

Scintillation Yield and Light Output

The number of photons produced in a step is determined from the energy deposited in the material and the total light yield. Conceptually, the average number of generated photons is

$$N_{\gamma} = Y \times E_{\text{dep}},$$

where $Y$ is the light yield, often expressed in photons per MeV, and $E_{\text{dep}}$ is the deposited energy in MeV.

In Geant4, the light output is controlled by several material properties.

SCINTILLATIONYIELD defines the basic yield, in photons per unit energy. You also often specify a relative fluctuation in the yield, through RESOLUTIONSCALE, to model non statistical effects in the light output. A value greater than 1 increases the spread of the number of photons around the mean Poisson expectation.

You can use different yields for different particle types, for example electrons compared to heavier ions. This is represented by additional properties such as ELECTRONSCINTILLATIONYIELD or similar names in more advanced configurations. These allow you to approximate quenching and different response to various particles.

In many detectors, the fraction of deposited energy that is converted into scintillation light is small. Therefore, the light yield value must be chosen based on the material and application. Fast plastic scintillators typically have yields in the range of thousands of photons per MeV, while inorganic crystals can have higher or lower yields depending on composition and temperature.

The scintillation yield must match the energy units you use for $E_{\text{dep}}$. If you define SCINTILLATIONYIELD in photons/MeV, then $E_{\text{dep}}$ must be expressed in MeV when you think about the expected number of photons.

Scintillation Emission Spectrum

Scintillators do not emit a single wavelength, but a distribution of photon energies. In Geant4, you define this as a function SCINTILLATIONCOMPONENT1 or similar, with values that are relative intensities versus photon energy.

You supply an array of photon energies and a corresponding array giving the probability density for emission at each energy. Geant4 uses this distribution to sample the photon energies for each produced scintillation photon. You can define one or more components, for example a fast component and a slow component, each with its own spectrum.

The spectrum can be approximated as a broad peak centered around the characteristic emission wavelength of the material. For example, a typical plastic scintillator might emit mostly around 400 nm, which corresponds to a photon energy of about 3.1 eV. In the material properties table you would convert the wavelength range of interest into energy values, then define the relative emission probability at each point.

It is important that the emission spectrum range overlaps with the optical properties of other detector elements, for example the refractive index and absorption length of the scintillator itself, and the photon detection efficiency of a photodetector if you model it.

Scintillation emission properties must be defined as functions of photon energy, not wavelength. If you design your spectrum using wavelength, always convert to energy using $E = \frac{hc}{\lambda}$ before providing it to Geant4.

Time Profile and Decay Components

Scintillation light is not emitted instantly. Instead, the light output follows a time profile, often described with exponential decays. Many scintillators exhibit a fast component, and sometimes one or more slower components. The time profile affects detector timing, pileup, and coincidence measurements.

Geant4 models the time distribution of generated scintillation photons using decay time constants that you specify in the material properties.

FASTTIMECONSTANT and SLOWTIMECONSTANT define the characteristic times for the fast and slow components, typically in nanoseconds. These constants describe exponential decays of the form

$$I(t) \propto \exp\left(-\frac{t}{\tau}\right),$$

where $\tau$ is the relevant time constant. You also specify a YIELDRATIO to set the fractional contribution of the fast component compared to the slow component. For example, a yield ratio of 0.8 means that 80 percent of the light is assigned to the fast component and 20 percent to the slow component.

If you only need a single decay time, you can simply use the fast component and set the yield ratio to 1. More complex materials might require additional detail or multiple components, but the basic idea remains the same. Geant4 then assigns emission times to each scintillation photon according to the combined time distribution.

The resulting time of each optical photon is added to the global time of the event and can be used later to simulate detector timing, time resolution, or coincidence windows.

A realistic scintillation time profile requires at least one nonzero time constant. If all time constants are missing or zero, Geant4 cannot model realistic scintillation timing and all photons may appear at the same time.

Distinguishing Fast and Slow Components

Some scintillators produce light in multiple components that differ in speed and sometimes in spectrum. Geant4 allows you to define separate emission spectra for the fast and slow components. This is useful when the fast component is at slightly shorter wavelengths than the slow component, or when your photodetecting system has a different sensitivity to each.

In the material properties table you can define SCINTILLATIONCOMPONENT1 and SCINTILLATIONCOMPONENT2, one associated with the fast component and one with the slow. Each component uses the same energy grid, but the relative intensities can differ. The yield ratio then decides how many photons belong to each component, while the time constants determine how they are delayed in time.

This separation makes it possible to study how detector timing performance depends on the fraction of light in the fast component. It also allows you to approximate pulse shape discrimination techniques, where different particles produce different fast to slow ratios.

If you do not need such detail, you can provide only one component and a single time constant, and leave the slow component undefined or unused.

Birks Law and Quenching (Conceptual Overview)

In real scintillators, the light yield is not always proportional to the deposited energy, especially for highly ionizing particles. This behavior is known as quenching. Birks law is a commonly used empirical description of this effect.

Although the complete implementation details belong elsewhere, the basic idea of Birks law is that the light yield per unit path length, $dL/dx$, is reduced when the energy loss per unit path length, $dE/dx$, is large. A typical expression is

$$\frac{dL}{dx} = \frac{S \, \frac{dE}{dx}}{1 + k_B \, \frac{dE}{dx}},$$

where $S$ is the scintillation efficiency and $k_B$ is Birks constant. As $dE/dx$ increases, the denominator grows and the light output per unit energy decreases.

Geant4 supports Birks law through additional material properties. By setting an appropriate Birks constant, you can make the scintillator produce fewer photons for heavily ionizing tracks compared to minimum ionizing particles. This is especially relevant for simulations where particle identification relies on pulse shape or light yield differences.

Quenching effects mean that light yield is not always linear with energy deposition. For highly ionizing particles, expect fewer scintillation photons per MeV than for minimum ionizing particles if Birks law is active.

Practical Considerations for Scintillation Simulations

Simulating scintillation light in detail can create many optical photons. Each photon must be transported through the geometry, interact with optical surfaces, and possibly be absorbed or detected. This can significantly increase computation time and memory usage.

To manage this, you can adjust several aspects of your simulation. One approach is to scale down the scintillation yield while keeping the relative distributions of timing and spectrum, then apply a corresponding scale factor when interpreting detector signals. Another approach is to use light guides, optical surfaces, and realistic absorption lengths to limit the number of photons that reach your sensors.

You should ensure that the refractive indices, absorption lengths, and surface properties of your materials are defined over the same energy range as the scintillation emission spectrum. Missing or inconsistent ranges can lead to unphysical results or messages about missing optical properties.

Finally, a complete scintillation simulation connects the scintillator to a photodetector model, where you typically convert arriving optical photons into detector hits according to some quantum efficiency or detection probability. That part is usually handled by an optical surface definition and a sensitive detector attached to the photodetector volume.

For reliable scintillation simulations, keep these points in mind:

Define consistent optical properties (refractive index, absorption length, emission spectrum) over the same photon energy range.
Include the scintillation process in your physics configuration.
Monitor the number of optical photons and adjust yields if the simulation becomes too slow.