Discussion Forum

Dear all-

I am trying to model the electromagentic radiation generated by a surface acoustic wave (SAW) in a piezoelectric material. These radiation losses have previously been modelled in COMSOL, and are published for example, here: ieeexplore.ieee.org/document/7748525. I'm just trying to get a similar simulation to work, and I'm having some troubles.

The radiation problem in this case cannot be solved for using the in-built piezoelectricity module in COMSOL, as this couples to the electrostatics module and therefore both assumes no displacement field (therefore no impressed magnetic field) AND that the electric field produced is irrotational. Because of this, the radiated energy calculated by the piezoelectricity module is clearly inaccurate. I have tried multiple work around's on the advice of others including COMSOL support (including using external current densities etc) and none have solved this problem, so I am convinced that the standard piezoelectricity module isn't capable of handling it.

I therefore believe that i need to set up my own set of multiphysics couplings, to couple the solid mechanics module (producing the SAW) to the RF module (to model the EM mode). The required couplings are described by equations 9 and 10 in the paper i link to above.

Taking Equation 9 as an example. This equation takes the form of the standard frequency domain electromagnetic wave equation, but with a modified permittivity tensor, and a driving term on the right hand side -which would normally be 0, as in the 'Wave equation, Electric' interface available in the frequency domain branch of the RF module-. I believe i can handle the modification to permittivity tensor quite simply by providing a user defined relative permittivity, and so I'm happy with that part. What I can't figure out how to do however, is to provide the driving term on the right hand side of this equation, I.E., to replace the 0 on the right hand side of the electric wave equation with a term of my choosing. This term would necessarily be a vector describing an impressed electric field, and i have found the expression for the components of this using COMSOL's internal definitions. However, I don't know how to feed this information into the calculation, and this is where i need some help.

I have tried to add this in as a 'weak contribution' in the RF module, but i found this wouldn't work as it won't accept vectors. I have tried various support pages and documentation sources online for an answer to this, but i cannot find any examples of adding a new multiphysics coupling in this manner. I'd just like to know if this is possible, and if so, what interface/feature would i need to make use of to achieve it?

Thanks very much for your timeand any help offered!

Michael