Simulating time-dependent radioactive decay instead of event-based

Particles, Track, Event, Run and Biasing
1

Hello,

I’m currently simulating F-18 decays in Geant4. When I run the simulation with /run/beamOn N, I understand that Geant4 treats this as N independent events - each event starting at t=0.

I’d like to simulate a time-dependent scenario, where decays occur continuously over a certain period (several seconds or minutes), closer to a real radioactive source.

My questions are:

  1. Is it possible to make Geant4 evolve “in real time,” or is it fundamentally event-based?

  2. If not, what’s the correct approach to model a continuous source - should I assign SetParticleTime() manually for each primary, or use a different run structure?

  3. How do users typically handle this when simulating activities or decay rates over time (especially for PET/PEPT-type setups)?

Thank you!

Hajji Oussama

2

Geant4 gives you the timestamps relative to events. You can add a global timestamp reference when processing the data by sampling N events from a Poisson distribution with mean given by your activity and then just running a cumulative sum that you add to each event.

3

Thanks a lot, that makes perfect sense. So Geant4 stays event-based, but adding timestamps afterward (sampled from a Poisson/exponential law using the source activity) makes the dataset statistically time-realistic.

I was just hoping there might be a way to make it more time-continuous within Geant4 itself, since the events are still independent.

4

The philiosophy here is that you can use the same simulation results with different “activities” very easily. It is just that, for a fixed amount of time, you will need to scale the simulation accordingly to get the total number of events you’d desire.

5

Thanks, that really clarifies the philosophy. I understand the benefit of keeping events independent for scaling activities, but my goal is to simulate a moving radioactive source, and I’m not sure how to model that if events are time-independent.

If the source follows a trajectory, the emission position should depend on time, which seems at odds with Geant4’s event-by-event reset. Would the right approach be to define the source position externally as a function of time and assign it (along with SetParticleTime())?
Or is there a built-in way to let the geometry or source evolve dynamically during a run?

6

See example extended/radioactivedecay/Activation :

README, PrimaryGeneratorAction.cc lines ~98-101, Co60.mac and Bi209.mac

It is interesting to see the histograms with and without the command /testhadr/gun/beamTime

7

Thanks a lot, I will check that. Appreciate your time and insights!