Learning Forum

[HAOYU LIU]

When I use 2D transient mode to calculate the bandwidth, it is around 12GHz. When I use the ssac mode to do the 3D simulation with the whole structure, it's 50GHz, I use the script from the example in the vertical photodiode. It's strange for ssac simulation.

[Guilin Sun]

I thought this post was replied however it seems the reply is gone to the wind.

I believe your result is reasonable: in 2D simulation by default we assume the 3rd dimension "norm length" is 1cm thick, if you did not change it. However in 3D I think you may not use 1cm thickness. The RC will be very different in the two cases. 

 

[HAOYU LIU]

 

Actually, my normal length in 2D is the same with the size in 3D ssac. However, when I decrease the electrical field in the ssac mode, the bandwidth should decrease because the tau (transit time) increase with the decrease of speed of the carrier. But the simulated bandwidth actually increase a lot, it is very strange. 

 

[Guilin Sun]

I believe that the badwidth is not a linear function of voltage or electric fields. There are other factors that impacts the badwidth, in addition to the transient behavior and charges. Khash has replied in this post: How to test the saturation performance of the photodiode?  

The best way is to find a benchmark example for verification. 3D and 2D is different even though the device size is the "same". The distributed RC may not be the same.