Menu

Discussion Forum

New Discussion
Filtern

SAW Delay Line Simulation with Large Number of IDT Pairs

Posted 25.08.2025, 16:09 MESZ MEMS & Piezoelectric Devices, Geometry, Physics Interfaces Version 6.3 2 Replies

TM
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hello,

I am trying to simulate a SAW delay line device in COMSOL with 100 IDT pairs on each side (input and output). Since modeling the full geometry with all these pairs would be very large and computationally expensive, I tried to simplify the model.

Currently, I built the geometry with only one IDT pair and applied periodic boundary conditions on both lateral sides of the device, assuming that this would mimic the effect of multiple pairs.

However, I am not sure if this approach is physically correct for a delay line with so many electrodes. Also, the results I am getting contain a lot of noise, which makes me wonder if my simplification is not valid.

My questions are:

Is it physically correct to model only one IDT pair with periodic conditions to represent many pairs in a SAW delay line?

Are there better approaches in COMSOL to simulate a large number of IDTs (e.g., 100 pairs per side) without building the full geometry explicitly?

Could the noise I observe be related to this simplification?

Any guidance or references would be very helpful.

Thank you!

Reply
2 Replies Last Post 10.09.2025, 21:13 MESZ
Edgar J. Kaiser Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 2 weeks ago 26.08.2025, 16:54 MESZ

A single IDT pair with periodic BC behaves like an assembly with an infinitely large number of pairs. I don't see how this would represent a delay line. To simplify things consider to do it in 2D instead of 3D. You can also consider to start with a smaller number of pairs and stepwise increase the number to see if it converges some way for larger numbers.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
A single IDT pair with periodic BC behaves like an assembly with an infinitely large number of pairs. I don't see how this would represent a delay line. To simplify things consider to do it in 2D instead of 3D. You can also consider to start with a smaller number of pairs and stepwise increase the number to see if it converges some way for larger numbers. Cheers Edgar
TM
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 4 hours ago 10.09.2025, 21:13 MESZ

A single IDT pair with periodic BC behaves like an assembly with an infinitely large number of pairs. I don't see how this would represent a delay line. To simplify things consider to do it in 2D instead of 3D. You can also consider to start with a smaller number of pairs and stepwise increase the number to see if it converges some way for larger numbers.

Cheers Edgar

Thank you for the earlier tip about simplifying to 2D and adding the IDT pairs gradually—that really helped; my S11 looks much better now. However, S12 is still a nightmare. I suspect I’m doing something wrong with the boundary conditions and with how I model the metal stack. I’ve attached a sketch of the device (GaN on sapphire, Cu/Ti IDTs). In my stack, the Cu IDTs sit on a very thin Ti adhesion/barrier layer on GaN. The Ti is only for adhesion.

Questions:

What is the best way to treat the Cu–Ti and Ti–GaN interfaces in COMSOL? Should I keep Cu/Ti as solid domains for mass loading, or model the electrodes as PEC boundaries and account for mass with an equivalent metal thickness?

Is there any benefit to keeping the thin Ti layer explicitly in the model for S-parameters, or is it safe to lump Cu/Ti into a single effective electrode?

Any tips to improve S12 (e.g., absorbing boundaries/damping to suppress edge reflections, meshing, etc.)?

Thanks again for the guidance.

>A single IDT pair with periodic BC behaves like an assembly with an infinitely large number of pairs. I don't see how this would represent a delay line. >To simplify things consider to do it in 2D instead of 3D. >You can also consider to start with a smaller number of pairs and stepwise increase the number to see if it converges some way for larger numbers. > >Cheers >Edgar Thank you for the earlier tip about simplifying to 2D and adding the IDT pairs gradually—that really helped; my S11 looks much better now. However, S12 is still a nightmare. I suspect I’m doing something wrong with the boundary conditions and with how I model the metal stack. I’ve attached a sketch of the device (GaN on sapphire, Cu/Ti IDTs). In my stack, the Cu IDTs sit on a very thin Ti adhesion/barrier layer on GaN. The Ti is only for adhesion. Questions: What is the best way to treat the Cu–Ti and Ti–GaN interfaces in COMSOL? Should I keep Cu/Ti as solid domains for mass loading, or model the electrodes as PEC boundaries and account for mass with an equivalent metal thickness? Is there any benefit to keeping the thin Ti layer explicitly in the model for S-parameters, or is it safe to lump Cu/Ti into a single effective electrode? Any tips to improve S12 (e.g., absorbing boundaries/damping to suppress edge reflections, meshing, etc.)? Thanks again for the guidance.

Attachments:

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.

Suggested Content