Skip to main content

MMAs

Parametric Study of a Microwave Absorber Based on Metamaterials
#

this post is a work in progress
#

In the fall semester of academic year 2024-2025 I decided took upon a project in order to design a microwave absorber based on metamaterials. In this post I document the progress as per the development of the cell and the modeling in order to lay everything publicly available and understand better as I try explaining the process to this figurative rubber 🦆 that is my editor. Finally I have opened a GitHub repository for the report that will eventually be submitted to my university and based upon I’ll be eventually credited; it can be found here.

So to start by designing a basic layout in CST I’ll implement a three-layer structure:

  • A dielectric Substrate w/ a metal Resonance Layer*
  • An Air Layer
  • A Metal Copper Back-Plate

*The metal resonance layer is technically an extra layer on top of the substrate and as a matter of fact it’s the only layer above Z=0 for reasons that’ll become obvious later on.

At first placing the substrate without the resonance layer, then I’ll place the two other layers, turn on the orthographic side view to remove shadows and voila:

verticaLayout
I think it really gives a sense of scale as the air layer truly dwarfs the others.

Now its time to add the ring that is of the same material and thickness as the backplate and lies on top of the dielectric substrate.

ring

For the arrows I make the assumption that both the arrow body and point are \(\alpha = 0.5mm\) of width.

nonparallel
In order to accurately place all the curve points that define the arrow some basic calculations shall be made. The two points of the arrow base lay exactly on the arc of the ring (w/ r=2.7mm) and are equidistant from curve y=x so the in order to find their cartesian coordinates the following system shall be solved.

syms x1 x2

eq1 = 2*(x1 - x2)^2 == .5^2;
eq2 = sqrt(x2^2 + x1^2) == 2.7;

sol = solve([eq1, eq2], [x1 x2]);
disp([sol.x1 sol.x2]);

$$ \displaystyle \begin{array}{l} \left(\begin{array}{cc} \sigma_3 -\frac{2916,\sigma_1 }{1433} & -\sigma_1 \newline \sigma_4 -\frac{2916,\sigma_2 }{1433} & -\sigma_2 \newline \frac{2916,\sigma_1 }{1433}-\sigma_3 & \sigma_1 \newline \frac{2916,\sigma_2 }{1433}-\sigma_4 & \sigma_2 \end{array}\right)\newline\newline \textrm{where}\newline \sigma_1 =\sqrt{\frac{729}{200}-\frac{7,\sqrt{59}}{80}}\newline \sigma_2 =\sqrt{\frac{7,\sqrt{59}}{80}+\frac{729}{200}}\newline \sigma_3 =\frac{400,{{\left(\frac{729}{200}-\frac{7,\sqrt{59}}{80}\right)}}^{3/2} }{1433}\newline \sigma_4 =\frac{400,{{\left(\frac{7,\sqrt{59}}{80}+\frac{729}{200}\right)}}^{3/2} }{1433}\end{array} $$

Which results in two points per quadrant, picking out the two points of the 1st quadrant and inserting them to CST the arrow body is parallel again

corrected

Then the arrow is mirrored against the X, the Y and the XY planes in order to reach all four sides of the cell, then the face is covered with copper and a height of d=0.035mm is also attributed, which is why it was important to move all other layers below Z=0.

mirrored
rna

Now I’ll try and perform a simulation using the frequency solver in CST from 2.7 to 12.7 GHz, adding a port with space (\(Z_{max}\)) in front of the cell and setting the orientation to negative so that is faces the absorber and the coordinates as full-plane the boundaries will be periodic along the XY plate and I will add an absorbing condition (\(Z_{min}\)).

The mesh of after of the structure after the simulation is as:

mesh

For the Electrical Field Simulation > and for the frequency [2.7, 7.7, 12.7] GHz and for \(Z_{max}(1)\) is as:

UnitCell_E_Zmax1_2.7GHz
UnitCell_E_Zmax1_7.7GHz
UnitCell_E_Zmax1_12.7GHz

There are no articles to list here yet.