KAHIBARO
11.2. Position Distributions
Table of Contents
Point
Position distributions in the General Particle Source (GPS) control where primary particles are created in your simulation. The simplest case is a point source. Here, all particles start from a single position in space, for example the center of your detector or a source placed at a fixed distance.
With GPS, a point source is selected by choosing the point type for the position distribution. In a macro, you typically first choose GPS as your source and particle, then configure the position type and coordinates. Conceptually, you have a single 3D point $\vec{r}_0 = (x_0, y_0, z_0)$ expressed in Geant4 units such as mm or cm. Every event will then have the same initial position for its primary particle, unless you later change the configuration during the run.
A point source is useful when you want to study the ideal response of a detector to radiation coming from a very small region, for example a sealed calibration source or a beam that is already very narrow. It is also a good starting point when you are first testing your geometry and physics list, because the geometry of the source itself does not introduce any extra complexity.
When you set a point source, you usually combine it with a separate angular distribution to define the direction of the emitted particles. For example, you can have a point source that emits all particles along the positive $z$ axis, or a point source that emits isotropically. The position distribution and the angular distribution are independent parts of the GPS configuration, and the point type only controls where the source is located, not how particles are directed.
For a point source, all primaries start at exactly the same position, so any spatial spread in the detector response must come from the detector geometry and physics, not from the source size.
Plane
A plane position distribution in GPS allows you to create particles over a flat surface instead of a single point. This is typically used to model beams or sources that cover an area, such as a wide beam entering a phantom or a uniform radioactive contamination on a surface.
In GPS, a plane is defined by choosing a planar position type and then specifying the shape and size of the plane, as well as its orientation in space. For beginners, it is common to start with simple shapes like a circle or a rectangle. Conceptually, you define a reference point on the plane (for example its center), and then define two perpendicular directions that lie on the plane. GPS then samples random coordinates within the chosen shape to generate a uniform distribution over the surface, unless you choose a different distribution.
A planar distribution is useful when the physical source is extended in one or two dimensions, but still thin in the third. For example, to simulate a parallel beam entering a water phantom, you can place a rectangular plane just in front of the phantom and restrict the emission direction to be perpendicular to the plane. In that case, the initial positions are spread over the plane, but the direction is fixed and normal to the surface, approximating a uniform illumination.
You can also rotate the plane to match your detector or beam geometry. The orientation is controlled by direction vectors in GPS, not by the geometry description, so the plane is purely a property of the source configuration. If you move your detector but forget to update the source plane, particles can start outside the intended region, so you must keep the GPS plane consistent with your geometry.
When using a plane distribution, the source has a finite size. The effective fluence (particles per unit area) depends on both the number of primary particles and the plane area, so changing the plane size changes the fluence even if the number of events stays the same.
Volume
A volume position distribution in GPS creates particles inside a 3D region. Instead of being confined to a surface, primaries are generated throughout the interior of a shape such as a box, cylinder, or sphere. This is essential when you want to simulate radioactive material distributed inside an object, for example activity inside a tumor, a bulk scintillator crystal, or a contaminated volume of air.
With GPS, you select a volume type for the position distribution, then choose the geometric shape and its dimensions in Geant4 units. For a box volume you define half lengths in $x$, $y$, and $z$. For a cylinder you specify radius and half length. For a sphere you define inner and outer radii and angular ranges if needed. GPS then samples positions uniformly inside this volume by default, so the probability density is constant throughout the region.
Uniform sampling in a volume is very useful for simulating homogeneous activity distributions. If you need nonuniform sources, for example activity that decreases with radius, you can still start from a volume distribution but add extra constraints or weights using GPS configuration commands. For absolute beginners, however, a simple uniform volume distribution already covers many realistic applications.
One important point is that the GPS volume definition is independent of the Geant4 detector geometry, even if you choose similar shapes. GPS does not automatically know about your logical volumes. You must manually set the same position, dimensions, and orientation if you want the source to coincide with a physical volume in your geometry. A mismatch can easily lead to primaries being generated outside the intended region or partially outside the detector.
When using a volume distribution to model activity inside a detector or phantom, always ensure that the GPS volume is fully contained inside the corresponding Geant4 volume, with consistent size, position, and orientation.
A volume source is especially important in medical physics and detector studies, where radiation is often emitted from within the material itself. In such cases, a point or plane source would give unrealistic energy deposition patterns, while a properly configured volume source can approximate the true three-dimensional activity distribution.
Views: 9