I am currently working with lumerical device.After I run the charge simulator and export the charge data using the charge monitor , I run FEEM solver while making an nk import for the charge distribution and check out the mode an its effective index.The problem is that the effective index of the mode does not change even if I change the doping concentration or If I import an optical generation source which is completely not logical.Note that I do this without applying a voltage because I would like to find the effect of change of the effective index when changing the doping concentration only.
Can you please provide simulation files for a review?
Also, can you try to run the simulations for different charge distribution and then visualize the effective index results?
lumerical.ldev (9.7 MB)
Kindly find attached the simulation file.If I remove the doping or the solar generation I get the same effective index for the mode.
It looks like the applied voltage in your nk-import is set to zero. This maybe one reason why you are not seeing change in the effective index. The other thing is that change in effective index is in the order 1e-4–1e-5 and you need to lock the mesh to make sure that your results are not affected by mesh mismatch between different applied voltages. Please visit the link below for MZI application example:
Thank for your reply. But I am actually setting the voltage =0 on purpose to find the change in effective index due to doping and solar generation not due to applied voltage .So how can I do that ?
The imported nk object should have the perturbed index data while in your case it seems to be just the charge data. Please use the script file provided in the MZI application example to create the perturbed index data (see the Index Perturbation of this direct: https://kb.lumerical.com/en/index.html?pic_modulators_getting_started_mach_zehnder_discussion.html). Other than this make sure that mesh is refined and you are using the lock mesh option.
Hope this solves the problem.
Hi @bkhanaliloo ,
Thank you for your reply I now understand where the problem is .I just have a problem that I am working with wavelength 5 Microns where as the silicon model used by the script in The mach_zender example only has data for wavelengths 1.3 and 1.55 Microns. I tried to write an equivalent script for the drude expansion model and I would be grateful if you have a look and let me know if it is correct that way or not.Also I would like to know does the NP Density Grid Attribute in mode and FDTD solvers work for Silicon model only or can I use it to calculate the perturbed index at 5 Microns.
Kindly find attached the script.
My apologies for a late reply. I had very busy week and missed to get back to you.
That is correct, since your operating wavelength is different you need to use Drude model. Now you have two options to either use the generalized Drude or Druce expansion models:
We have examples in both cases. You can double check your equations and use the calculations done in the script to calculate the results for 1.3 and 1.55um operating wavelengths to calculate the index perturbation data.
Please let me know if you have any further questions.
Thank you very much for your reply it really helped a lot.I just have one final request to ask .I searched for an example that uses the drude expansion model but found nothing except for the example of charge to index conversion that explains the three model.So if you know a link for an example that uses the drude expansion model I would be grateful.
Thanks a lot.
I was wondering if you checked the WgImport.fsp simulation file. The WgImport_Drude uses the Druce_exapnsion model. If you have any problem regarding the file or you have further inquires please let me know.
Yes I checked it out and took parts of the script in my script.I just would have preferred if that was an example like that MZI modulator.
I just have one final question . Is the drude model for the index perturbation material the normal or the expansion one ? And is the effective mass of the electron the conductivity effective mass or the effective mass for density of states ?
I also have one question concerning the default values of drude model for the index perturbation material .I noticed that for the values for Silicon are switched for the effective mass of the electron and the hole but it was written that these are the default values for Silicon which made me wonder about the type of this effective mass . It was written that effective mass for electrons are 0.8 and for holes 1.1 where they should be switch am I right ?
There are three imported material based on generalized Drude, silicon, and Durde expansion models. Their differences can be found from the script files for the structure groups. The parameters used in the scripts are summarized below:
Can you please elaborate on your second inquiry? Where do you get the data for the electron and holes effective mass and where are they different?
Concerning my first part of question ,we have two perturbed index material options that we could choose from in material models in FDTD or mode .One is Silicon Soref and the other is drude model.So my question basically was is this drude model the generalized one or the expansion one ?
My second question is the mn* and mp* which are the parameters to fill in for the model represent m :effective mass or m*c conductivity effective mass.Also I don’t think that the effective mass for holes and electrons are wavelength dependent since they are property of material am I right ?
Finally I was referring to default values when you add this model i think that the hole and electron effective masses for Silicon are reversed.
Kindly find attached a picture to clarify more what I mean.
Appreciate your help.
Sorry for my late reply. I was quite busy with few other tasks.
The “Drude model” for the “np Density” of Index perturbed material follows the generalized Drude model. We do not use any expansion here.
You are right, they are not wavelength dependent. I think we refer to coefficients with parameters defined by red box in the image below:
Here the n (refractive index) and \( \lambda \) are wavelength dependent.
Hope this answers your inquiry.