Learning Forum

[Rui_Xu]

Hi,

I am trying to simulate the behaviour of Fabry Perot cavity in THz region made by two Au mirrors. I start with the simplest situation, which is just two Au thin film with vaccum inside (attached file). The dielectric functions of Au in THz region is described by Drude model, which is Plasma in the database. The distance of two thin films are 7.5um, which corresponds to resonance at 15um. However, the result of reflection spectra just looks weird (attached file). Can anyone help me solve the problem? Thanks!

Rui

FPC au THz.rar

[Guilin Sun]

Hi, Rui, Ansys employees are not allowed to download ALF posts from customers. Maybe you can give some screenshots.
In THz usually it needs much longer simulation time than the default 1000fs. You may try 10~100ps or even longer. Please check this Chinese posts: Ansys Insight: Õªéõ¢òþƒÑÚüôFDTDþÜäõ╗┐þ£ƒµùÂÚù┤ÞÂ│ÕñƒÚò┐
In addition, you will need to set periodic BCs perpendicualar to the light propagation and use sufficient thick PML along the light propagation. Please try to modify the settings. In addition, you can add a time monitor and check the signal if it decays completely.

[Rui_Xu]

Hi gsun Thanks for the quick respince!
Here are some screenshots which may be helpful. I used anti-symmetric BC perpendicular to light propagation orientation and standard PML along (fig 1). The reflction spectra is smooth but just wierd, even over 1 at some points (fig2). The time limit I set was 25000fs, but the simulation autoshut off just at 2500fs. The time monitor at the center of the cavity is shown in fig 3. Seems there is still a little field. Does it mean I need to decrease the auto shutoff to be smaller than 1e-5?

[Guilin Sun]

yes, try 1e-6 or1e-7. Make sure the simulation does not terminate at 100%.

[Rui_Xu]

Hi gsun I changed to 1e-7 and time decay looks better (Fig 1). The simulation does not terminate at 100%. However the reflection still looks wierd (Fig 2). Any more suggestions?

[Guilin Sun]

The time monitor still shows strong signal, so smaller autoshutoff is further needed. You can try 1e-9 and longer simulation time. At the same time, make sure the PML has about half wavelength thick, which you can roughly measure it using the ruler in the left-hand panel. Since you are using Au material as mirrors, the reflectivity is very high, so the signal can be decay very slow and there is no other mechanism such as higher diffraction orders to leak energy.

I suggest that you monitor the transmission. You can also move the reflection monitor in between the source and the mirror to get the reflection. Please refer this post Ansys Insight: Õªéµ×£ÕÅìÕ░äþøæÞºåÕÖ¿õ¢ìõ║ÄÕàëµ║Éõ©Äþ╗ôµ×äõ╣ïÚù┤Õªéõ¢òÕ¥ùÕê░µ¡úþí«þÜäÕÅìÕ░äþÄç

[Rui_Xu]

Thanks for the suggestion!
Now the time signal has been decayed to ~1e-5, and the thickness of PML is larger than half of the wavelength. The reflection spectra are all smaller than 1, though still looks interesting. It is supposed to be a dip at ~15.6um, somehow half of it just inversed. Is it due to the meshing problem? Since I just meshed the region where the Au mirror is.

[Guilin Sun]

What is the thickness of the mirror? does the time monitor show the complete signal ? if so, the signal is still strong. The DFT monitor cannot get accurate result because of the signal truncation. Please make sure the autoshutoff level is small enough and the simulation time is long enough and it terminates at less than 100% in the progress.
If there is still problem, please give me all the device parameters so I can have a test if I have the time.

[Rui_Xu]

Hi gsun The thickness I used is 50nm, and the distance between the two are vaccum with T=7.8um. The dielectric function I used for Au is plasma model, with epsilon=6.88, resonance frequency= 8.96eV, gamma=0.0723eV. The light source is just plane wave from 10-20um, since the resonance should be @15.6um. Anti-symmetric BC are used in both x boundaries. Could you please have a try if you have time? Thanks a lot! If additional parameters are needed please let me know.
Best Rui