Lumerical CHARGE: Import Doping Simulation Discontinuity

junqianljunqianl Member Posts: 3

I am trying to import a doping profile and solve the charge profile based on the phase shifter example under various biases (https://support.lumerical.com/hc/en-us/articles/360042327854-PIN-Mach-Zehnder-modulator). However I noticed that the charge distribution is entirely different after the first voltage step for both positive and reverse bias despite it looking normal at the initial 0 V state (see images attached). I have also added my simulation file for reference.


Comments

  • gsungsun Posts: 729Ansys Employee

    I downloaded the example file, modified the voltage step from 0.035 to 0.05, and checked all the voltage values, the results are quite regular as this one:

    Not sure what computer operation system is, and how did you get that result. Then I just simulate two points, 0 and 0.05v, and I got




    Please check if your computer OS is supported: https://www.lumerical.com/licensing/system-requirements/

    or you maybe need to restart your computer; reinstall the latest version. If the problem still exists, please send us email so we can have a close look, with your computer OS, software version information, and the file you are using. Please mote the Mac system is no longer supported.

  • junqianljunqianl Posts: 8Member

    I am running on Windows 10 so that shouldn't be an issue. However I noticed the PN concentrations in your profiles are flipped from mine. Are you using the original file from the website link or the .7z file of the modified example that I provided? The original example file works fine for me as well. The issue only seems to come up when I attempt to import the doping profile from a .mat file though I can't see any issues with the format I am using.

  • gsungsun Posts: 729Ansys Employee

    Thank you for letting me know that the online example works fine for you. I tested the original online example file. As you may know, we are not allowed to open customer's files here at this Forum.

    As long as the online example works fine, the issue you have might be duo to your np doping: please check and make sure the doping is exactly the same online, and as what you expect. A simple test would be to use the online file, but set the doping from importing instead of the standard doping.

  • junqianljunqianl Posts: 8Member

    I tried importing my doping profiles into the original example file but the discontinuity still exists when sweeping the voltages. Is there another example showing the import of a PN MZM doping profile that I could try?

  • gsungsun Posts: 729Ansys Employee

    This example has also imported doping 3D pillar silicon solar cell – Lumerical Support

    I would suggest that you simulate a simple device such as a PN junction, with import doping and see if it works. In addition, please check your doping profile and make sure it is correct.

  • junqianljunqianl Posts: 8Member
    edited May 19

    Yes, the device I am currently importing is a simple PIN structure with 2 heavy doping layers under the metal contacts and a background P-doping so I am having trouble trying to find the root issue. The net doping profile (top image) is quite smooth so I don't see why the space charge distribution (2nd image) is so uneven. I have tried importing the structure into Sentaurus device simulations as well and don't see these issues.


  • gsungsun Posts: 729Ansys Employee

    Since you have heavy doping, you may need to lower the applied voltage. You may also try to reduce the doping density, begin with regular doping and watch how the result change with doping/voltage.

  • junqianljunqianl Posts: 8Member

    Alright, turns out the issue which was the scaling constant in the import doping object gave a mismatch in units since my profiles were in cm^-3 rather than m^-3. The quantitative results in phase change still don't align perfectly but at least the carrier distribution profiles are consistent now.

  • gsungsun Posts: 729Ansys Employee

    Thank you for the update! This explains why my result is much higher than yours and why I gave the previous suggestions. Yes, in general, Lumerical script will use mks unit which is meter.

Sign In or Register to comment.