Learning Forum

[WANG JINGXIN]

I want to simulate the dark current of avlanche photodiode using lumerical CHARGE.

However,  when I open the impact ionization model and start to run the program, the voltage scan start backtracking very early, which is stil far from the breakdown voltage in the reference.(for example, -25 V voltage is correct but the voltage backtrack at -1V)  And when the job error (troubleshooting convergence error) is reported, no breakdown can be observed in the simulated dark current. 

If I close the impact ionization model, no error reported but of course no breakdown observed. I have downloaded the example "avalanche photodetector" which is given by lumerical and successfully observe a breakdown voltage but I cannot observe breakdown in the model I made by myself.

I want to know what cause the problem and how to get a result in which breakdown can be observed. I have set the voltage step 100mV(not big) and opened gradient mixing and increased the global iteration limit but it still does not help. Is it because the doping type is set to be constant doping?( in the example the doping type is diffusions doping)

[Khashayar Ghaffari]

Hi,

Thank you for the detailed description of the issue. As you pointed out the core issue is the simulation stopping before the expected breakdown. it's hard to troubleshoot without inspecting the file (which is unfortunately not allowed on the forum). However, one thing I notice based on your screenshot is that you are running a 2D simulation. Please note that this will not accurately capture the circular geometry of your photodiode, so a 3D simulation is needed. Hopefully, this is causing the unexpected IV behavior. Of course, running 3D simulation is more time consuming and might cause more convergence errors compared to 2D.

If switching to 3D doesn't resolve your issue, I recommend inspecting the carrier density and electric field results. This can be insightful at finding more information about the behavior of the device and in some cases helps you find the necessary updates to your model.

The general guidelines for resolving troubleshooting issues can also be helpful. I suspect you might have already checked this out but just in case: https://optics.ansys.com/hc/en-us/articles/8355464963987-Troubleshooting-convergence-errors-in-CHARGE

Best regards