TC-SAW Design Study


Temperature compensation technology in Surface Acoustic Wave (TC-SAW) devices allows for improvement in their performance. OnScale enables designers to explore many TC-SAW device design spaces to offer a greater insight into the optimum design.

In this example, we simulate a 3D unit cell model of a TC-SAW device to analyse the effect of electrode and Si02 thickness on the electrical performance.

Model Setup

A schematic of the model and the two input variables can be seen below:


Basic Model Schematic

Parametric Variables

The key model parameters were as follows:



Default Value


Thickness of substrate

10 µm


Thickness of the electrodes

100 nm


Thickness of Si02 layer

250 nm


SAW finger pitch

1 µm


Number of finger pairs



Simulation length in cycles



Centre frequency

1.9 GHz

The design variable electrode thickness was swept from 100nm to 200nm in 14 steps and the Si02 thickness was swept from 250nm to 400nm in 14 steps. This resulted in a total of 196 simulations which were run in parallel.

Simulation Sweep Results

The full simulation sweep was completed in 2 minutes when using 8 cores per simulation and had a total cost of 52 core hours.

Using the outputs from the simulations, it is possible to calculate the device impedance, Q factor, resonance frequency and coupling coefficient. These outputs can all be plotted against the input variables to provide an insight into the device performance as shown in the image below:


0 out of 0 found this helpful


  • Hi, thanks for this example.
    I have two questions about the code: 1. Why is it that for the TC Unit Cell the electric solver window only spans $k1-> $k2 when SiO2 (defined in the .prjmat as a dielectric) is present up to $k5? Does this not partially ignore the effect of the dielectric SiO2 layer? 2. What parameters are used to calculate "coupling coefficient" in the sweep results?
    Thank you.

    Comment actions Permalink
  • Hi Lucas,

    The dielectric properties of Si02 will still be considered, the electrical window is just used to specify where the voltage conditions are applied, in this case between LiTaO3 and the electrodes. To calculate coupling coefficient, we use a Review script - Here is our documentation on Review if you haven't used it before -

    Review is basically our post processing language which allows us to extract these types of outputs from the output files which are generated in simulations (.flxhst files).

    Best Regards,

    Comment actions Permalink
  • Hi Chloe,

    Thanks for the speedy response and the Review script, this has been very helpful! I have a couple more outstanding questions:

    I'm wondering what additional processing (if any) was used to produce the plot of coupling coefficient? My understanding is that in line 80 the given Review script is calculating a variable called 'kt' which is actually the square root of what is traditionally called the coupling coefficient (K^2) (equation 8

    Furthermore, I'm wondering how the units of the plot for "kt" are in (%)? When I run this simulation with SiO2 thickness at 400nm and sweep electrode thickness in 14 steps from 100nm -> 200nm (as outlined in the study), the attached Review script yields kt ~ 7 which corresponds to kt ~ 700%, much greater than what is shown.

    Please let me know what I'm missing / where these assumptions have gone wrong.
    Best wishes

    Comment actions Permalink
  • Hi Lucas,

    Sorry for the delayed reply. I have looked at a few sources and it seems coupling coefficient is often referred to as k^2 which would mean our example is not correct. We are going to do some further research, I will update you when we are confident the coupling coefficient is correct in this case study.

    Best Regards,

    Comment actions Permalink

Please sign in to leave a comment.