Mahdi Siasifar An eigenanlysis will only give you an approximate solution, since you apparently are using a nonuniform array. That being said, you should try to make the background the same as your driven simulation (same dielectric, surrounding material, etc.).

You can make the boundaries on the transverse X-axis periodic, or perfect-electric conductors, since I assume you want the SPP polarized along this axis. The boundaries can probably be close to (or touching) the edge of the metal. Similarly on the transverse Z-axis, you can use periodic boundary conditions of perfect-magnetic conductors -- but ensure they are appropriately far from the metallic surface so that they approximate open space. Of course, you want periodic boundaries on the longitudinal (Y) axis, with a specified phase angle.

Stuart Barth Regarding eigenvalue analysis, I was forced to use it as I need dispersion curves of my structure to calculate beta for certain frequencies. Do you think it will be a good method?

Mahdi Siasifar An eigenanalysis will give you an approximate solution of the observed Beta. Since it looks like your inner cells are nearly uniform, the results should be fairly close. You could repeat the analysis for various unit cell parameters if you need additional accuracy.

Stuart Barth Thanks again for answer, I tried your response, the ports on y axis can not touch periodic axis and I get an error. My ultimate goal is to use Eigenvalue analysis for drawing dispersion diagram of the SSPP, do have any idea how I can solve it?

Mahdi Siasifar In most simulators, there shouldn't be any ports in an eigenmode simulation, only periodic boundary conditions. There are many helpful documents online for performing these analyses in various software, see for example here: http://www.emtalk.com/tut_2.htm

For SSPP design ,draw one more box with same dimensions so that the design should be exactly within that box, after placing the design assign primary and secondary boundaries and follow these steps http://www.emtalk.com/tut_2.htm