KAHIBARO
12.7 Choosing a Physics List
Table of Contents
Accuracy
Selecting a physics list is always a compromise between realism and cost. For a beginner, the safest habit is to start from a reference list recommended in the Geant4 manual, then adjust only if you have a clear reason.
A physics list controls which particles exist in your simulation, which physical processes they undergo, and how those processes are modeled. Different reference physics lists implement different models, energy ranges, and tuning choices. As a result, they can produce slightly different particle ranges, energy deposition, and secondary spectra, even for the same geometry and source.
For accuracy, two points are especially important. First, the energy range and type of particles you care about must be covered by suitable models. For example, hadronic lists that are optimized for LHC energies are not automatically ideal for a low‑energy neutron shielding study, and vice versa. Second, the physics list should be validated for a domain that is as close as possible to your application. The Geant4 collaboration and user community provide validation plots comparing simulation to experimental data, often for standard reference lists such as FTFP_BERT, QGSP_BERT, and specialized lists like QGSP_BIC_HP or Shielding.
When you choose a list, check what it is known for. For example, FTFP_BERT is widely used for general high‑energy applications and calorimetry. Lists with the suffix _HP include high precision neutron models and cross section data, which are more accurate below about 20 MeV but are heavier to run. Electromagnetic “Option3” or “Option4” variants offer more detailed modeling of multiple scattering and low‑energy processes, and are often recommended for medical or dosimetry work where detailed electron transport matters.
For many beginner projects, the physics list will be one of the main sources of systematic uncertainty. You are not expected to know the best list from the start, but you should document which one you use and keep in mind that switching to a different, more appropriate list can change your results by a few percent, sometimes more.
For any serious study, treat the physics list choice as a source of systematic uncertainty and compare at least two reasonable lists during validation.
Performance
More detailed physics usually requires more CPU time and memory. The way you choose a physics list can have a large impact on how long your simulation takes to run, especially in multithreaded or large‑scale productions.
Physics lists that include very low energy thresholds, detailed models over a wide energy range, or high‑precision neutron data need to perform more calculations per step. This is helpful where it is physically justified, but it slows down transport for every track in your simulation. For example, a list with low energy electromagnetic options and high precision neutrons can be significantly slower than a simpler list that omits them. Similarly, adding specialized optical physics or detailed hadronic models will also cost CPU time.
In practice you balance performance and realism by asking where the extra detail truly matters. If your particles never go below a few MeV, very low energy extensions are not needed. If your problem does not rely on thermal neutrons, you do not need a _HP list. If you only care about the total absorbed dose averaged over a large volume, you do not need the most detailed microdosimetric setup.
You can also combine physics choices with production cuts to control performance. Production cuts decide when to create secondary particles explicitly and when to treat them as continuous energy loss. Coarser cuts reduce the number of tracks, which interacts with your physics list to change runtime. For a given physics list, you can often gain much more speed through appropriate cuts and geometry simplifications than by switching to a “lighter” physics list that may be less accurate.
Do not use a more complex physics list than your problem justifies. Prefer the simplest list that is still physically adequate for your particles and energy range.
Application requirements
The most practical way to choose a physics list is to start from your application domain and work backward to the physics you need, then map that to a known reference list.
Different applications emphasize different interactions:
In high‑energy particle physics and calorimetry, you usually care about hadronic showers, electromagnetic showers, and secondary production over a wide energy range. Reference lists like FTFP_BERT or QGSP_BERT are commonly used. They balance accuracy and performance for proton and pion interactions from sub‑GeV to many TeV, and include standard electromagnetic physics for electrons and photons.
In medical physics, such as radiotherapy or dosimetry, detailed electromagnetic physics is often the priority, especially for electrons and photons down to low energies. Here, variants with extended electromagnetic options are often used. For hadron therapy or proton transport, you may need a list that has suitable proton and ion models in the energy range of interest. High precision neutrons can become important when you are interested in out‑of‑field doses or shielding.
In nuclear physics or shielding studies, the focus is often on low and intermediate energy neutrons and gamma production. High precision neutron lists and specialized shielding lists are better suited to these tasks than generic high‑energy collider lists, because they include evaluated cross sections and detailed treatments of thermal and resonance behavior.
In space and radiation effects applications, broad energy coverage for protons, heavy ions, and secondaries is important. You may choose a physics list designed for space or radiation analysis, or a general list that is documented as suitable for cosmic rays and space environments.
For detector development with scintillators, semiconductors, or other sensor materials, you typically need accurate electromagnetic stopping powers, multiple scattering, and sometimes optical photon processes. A general list with a detailed electromagnetic component is appropriate, and you can add optical physics if your study includes light transport. If you are studying timing and pulse height, the details of electromagnetic interactions in your detector material become especially important.
A practical approach for an absolute beginner is as follows. First, identify your main particle types and approximate energy range. Second, find in the Geant4 documentation which reference lists are recommended for that domain. Third, choose one recommended list and use it without modification at first. Only after you have a working simulation and begin validation should you consider switching lists or fine‑tuning physics constructors.
Always match your physics list to your particles, energy range, and domain. Do not invent a custom list until you clearly understand why the standard reference lists are insufficient.
Views: 7