Electrostatic polarizability of platonic solids

In this example we use the scuff-static electrostatic module of scuff-em to compute the DC polarizabilities of dielectric bodies with regular shapes and varying dielectric permittivity. This example was inspired by this paper:

The files for this example may be found in the share/scuff-em/examples/PlatonicSolids subdirectory of your scuff-em installation.

gmsh geometry files for spherical, tetrahedral, and octrahedral particles

The gmsh geometry files Sphere.geo, Tetrahedron.geo and Octahedron.geo respectively describe spherical, tetrahedral, and octahedral particles. I turn them into .msh files describing surface meshes of two different finenesses by running the following bash script, which invokes the RenameMesh script:


for SCALE in 1.0 0.7
  for SHAPE in Sphere Tetrahedron Octahedron
    gmsh -clscale ${SCALE} -2 ${SHAPE}.geo -o ${SHAPE}.msh
    RenameMesh ${SHAPE}.msh

Note: The .geo files contain a variable parameter L that sets the sphere radius or the polygon edge length. In each case, the default value of L is chosen to ensure that the volume of the result particle matches the volume of a sphere of radius 1, but you may override this setting (to set L to a value of, say, 2.34) by adding the command-line option -setnumber L 2.34 to the gmsh command above.

Here are pictures of the finer of the three meshes produced by this procedure:


Polarizability of a dielectric sphere

Here's a simple scuff-em geometry file named E10Sphere.scuffgeo that describes a dielectric sphere with relative permittivity :

    MESHFILE Sphere_501.msh

To compute the DC polarizability of this object, we say

% scuff-static --Geometry E10Sphere.scuffgeo --PolFile E10Sphere.pol

This produces a file named E10Sphere.pol, which reports all 9 components of the DC polarizability tensor for the sphere:

# data file columns: 
# 01: object label 
# 02: alpha_{xx} 
# 03: alpha_{yx} 
# 04: alpha_{zx} 
# 05: alpha_{xy} 
# 06: alpha_{yy} 
# 07: alpha_{zy} 
# 08: alpha_{xz} 
# 09: alpha_{yz} 
# 10: alpha_{zz} 
Sphere 9.221300e+00 6.692939e-03 4.299017e-02 1.225110e-02 9.231265e+00 1.831022e-03 -1.860956e-02 1.039749e-02 9.233396e+00 

Here are some observations on this result:

  • The diagonal components here are in decent agreement with the theoretical value of .

  • The slight discrepancies between the three diagonal components arise because the surface mesh slightly breaks rotational symmetry of the sphere.

  • The nonzero values of the off-diagonal components are numerical noise.

In all three cases, the result is improved by running calculations with finer surface meshes.

Polarizability of a dielectric cavity

Here's an alternative geometry file describing a spherical vacuum cavity in a universe filled with dielectric material:


    MESHFILE Sphere_501.msh

The polarizability data predicted by scuff-static for this geometry are

Sphere -5.168105e+00 3.103541e-03 1.420259e-02 3.842746e-03 -5.164908e+00 -1.409712e-03 -6.971980e-03 4.036885e-03 -5.165247e+00 

The diagonal components here agree with the theoretical value of .

bash script to run calculations at many different permittivities

Ultimately we want to compute the polarizability at many different values of the interior and exterior permittivity, and it would be unwieldy to have to repeat the above process by hand in each case. Instead, it's convenient to write a a bash script to automate the process. This script is called GetPolVsTau.sh. It has a hard-coded list of surface meshes and reads pairs of interior and exterior permittivity values in the form (EpsOut, EpsIn) from a file named EpsFile. For each surface mesh and each pair of permittivity values, the script does the following:

  • It constructs a .scuffgeo file for the given mesh and permittivity values by appropriately modifying a template file called Template.scuffgeo.

  • It runs scuff-static to get the polarizability for this object with this pair of permittivity values.

  • It reads the resulting polarizability data from the output file written by scuff-static and writes the data, together with the (EpsOut, EpsIn) values, to an overall output file for the shape in question.

This script, which takes about 10 minutes to run on my laptop, produces a series of files with names like Tetrahedron_1179.AlphaVsTau. Here Tau refers to the quantity defined in the paper by Sihvola et al. referenced above.

Plotting polarizability vs. for the three shapes reproduces Figure 4 of the paper by Sihvola et al. referenced above, except for something funky that appears to be happening for the tetrahedron around


Incidentally, here's the gnuplot script that I used to produce this plot: Plotter.gp.