KAHIBARO
37.6. Optical Photon Detection
Table of Contents
Photon counts
Optical photon detection in GATE connects the transport of individual optical photons to what a real photodetector would “see” as a count or signal. At this stage you no longer care about every photon’s detailed path, but about how many of them are converted into detectable photoelectrons and how this is recorded.
In GATE this behavior is controlled by a combination of optical properties of materials and surfaces, and by specific actors or digitizers for optical photons. Photons that reach a photodetector volume can either be absorbed without signal or converted to a count, depending on the quantum efficiency or detection efficiency that you define.
The most direct way to model photon counts is to attach an actor to the sensitive photodetector volume. For example, a hits or optical hits actor can be used to record each time an optical photon is detected. Each hit can correspond to a “detection event” such as the creation of a photoelectron in a PMT or SiPM pixel. From this list of hits you can then build histograms or time series of counts.
Alternatively, you can include optical photons in a more general digitizer chain, where hits in the photodetector are transformed into signals that resemble realistic electronics response. In this case you might group optical hits in time, apply a threshold in number of detected photons, and then count only signals that cross this threshold as valid detector counts.
The key link between transported photons and recorded counts is the detection probability. This can be represented as a wavelength dependent efficiency that is assigned to the optical material or surface of the detector. When a photon reaches the detector, GATE uses this efficiency to decide stochastically if the photon is detected.
Important rule for detection probability:
If the quantum efficiency (QE) or detection efficiency is $p$ (for example $p = 0.25$), and $N_{\text{photons}}$ photons reach the detector, then the expected number of detected photons is
$$
\langle N_{\text{detected}} \rangle = p \, N_{\text{photons}}.
$$
The actual number detected fluctuates according to binomial statistics.
These binomial fluctuations are often called photon statistics or counting statistics. They are an important source of noise in scintillation detectors and Cherenkov detectors. When you analyze simulation output, you will see these fluctuations in energy spectra or light yield distributions, even if you start from a fixed number of generated photons.
If you are simulating pixelated photodetectors, you typically define separate sensitive volumes for each pixel. The optical actor or digitizer records counts per pixel, which allows you to build spatial images of light distribution on the detector surface. This is useful for studying light sharing between crystals, optical cross talk, and position reconstruction algorithms.
You should always distinguish between two different “counting” levels. First, individual detected photons are counted by hits or low level digitization. Second, these photons may be combined into a higher level detector event such as a PMT pulse or a SiPM microcell avalanche burst. At the event level you may only keep the total number of detected photons in a given time window, or a charge proportional to this number, instead of every individual photon.
In practice you will often study how photon counts depend on geometry and optical properties. For instance, you might vary reflector type, surface roughness, or coupling layers and record the number of detected photons per deposited MeV in a scintillator. By comparing $N_{\text{detected}}$ across configurations, you can optimize light collection efficiency. GATE’s optical actors make this straightforward, since you can attach them to different detector volumes and compare the output files.
Because optical simulations can produce very large numbers of photons, it is important to limit which photons are recorded. You can use filters to restrict hits to specific volumes or to photons that satisfy certain conditions. This keeps output sizes manageable while preserving the information you need for count analysis.
Timing
Timing information for optical photons is critical in applications that rely on fast scintillators or time-of-flight measurements, such as TOF-PET or Cherenkov-based detectors. In GATE, every optical photon has an associated time that records when it is created and when it reaches the detector. The optical photon detection stage uses this information to build time distributions and timing observables.
The arrival time of a detected photon is influenced by several effects. First, the emission time of scintillation light follows one or more exponential decay components, which you define in the scintillator’s optical properties. Second, the path taken by each photon inside the crystal can vary significantly because of multiple reflections and refractions, so photons do not all travel the same distance. Third, the refractive index determines the speed of light in the material. Together, these factors cause a spread in arrival times at the photodetector, even for a single energy deposit.
When you record optical hits on a photodetector surface, each hit typically includes a time attribute. By analyzing these times you can construct time histograms and study the time profile of detected light. The first detected photons are particularly important for fast timing applications, because they carry the most precise information about the interaction time.
In a more realistic detector model you will not use each individual photon time directly, but you will build an electronic time stamp from the collection of detected photons. For example, you may define the event time as the time when the first photon is detected, or as the time when the integrated signal crosses a threshold. This process is often implemented in a digitizer that takes optical hits as input and produces detector events with a single time per event.
Key timing concepts for optical detection:
- The photon time of arrival $t_{\text{arrival}}$ is the sum of emission time and travel time in the medium.
- The detector time stamp $t_{\text{detector}}$ is a function of many photon arrival times, for example the time of the first photon or the time when a threshold is crossed.
- The timing resolution $\sigma_t$ is usually modeled as a Gaussian blurring of the ideal detector time:
$$
t_{\text{measured}} = t_{\text{detector}} + \mathcal{N}(0, \sigma_t)
$$
where $\mathcal{N}(0, \sigma_t)$ is a Gaussian random variable with standard deviation $\sigma_t$.
GATE allows you to introduce timing blurring at the detector level to approximate the finite timing resolution of PMTs or SiPMs. This is important when you want your simulated timing distributions to match experimental performance. You typically specify a time resolution parameter, and the digitizer will smear the ideal time stamp accordingly.
In coincidence systems such as PET, you are often interested in the time difference between two detector events. For optical photon detection this means that you must simulate the timing of light in both detectors, apply timing blurring, and then compute the coincidence time difference. Accurate modeling of optical photon timing is therefore essential if you want to predict the coincidence timing resolution of a scanner.
In position sensitive detectors, timing can also be used for depth or position estimation. For example, differences in arrival time distribution between light detected at two ends of a long scintillator can be used to estimate the interaction depth. In GATE you can study this by recording time information separately for different photodetectors attached to the same crystal and analyzing the relative timing.
Finally, timing information is crucial when you simulate systems that rely on fast gating, time windows, or pile up effects. By having precise photon and event times in your output, you can test how different coincidence windows, trigger conditions, or dead times affect the effective detection efficiency and count rate. Optical photon detection in GATE thus provides not only how many photons are detected, but also when they are detected, which is fundamental for time-based performance studies.
Views: 12