I-V characteristics become nonlinear with the introduction of grating


#1

Hi @aalam,

We are trying to get the electrical characteristics i.e. IV curves for the following design.

The file for the design with grating and without gratings are attached
Without Grating - https://drive.google.com/file/d/0B42M3soVszILNXhmQlVnRGMxSWs/view?usp=sharing
With Grating - https://drive.google.com/file/d/0B42M3soVszILaUx2dEJrM3NjeEE/view?usp=sharing

Since we are getting a straight line IV curve graph for the not-grating structure in without grating structure, could you check if the structure in with grating file is wrong, because of which we are not getting a straight line IV response.

Thanks


How to apply constant doping to a material layer?
#2

Hi @Arastu, I have taken a look at your file and there are a few issues that need to be fixed. However I haven’t been able to fix them all yet. I will spend some more time on this tomorrow and try to get back to you with some answers.


#3

Hi @Arastu, as I mentioned before, there are a couple of things to look at here.

  1. First of all there is a problem in the latest version of DEVICE where any geometry in a structure group gets a lower priority in the model tree. As a result the geometries in the structure group are getting overwritten by the other rectangles even when you have placed them in the correct order in the objects tree. The solution to this problem is to use the “mesh order” option in all the geometries and specifically tell the solver which geometry overwrites which one. In your file I have set the mesh order of all the geometries by using largest value for the elements at the top and smallest values for the ones at the bottom. If you mesh the structure and look at the “ID” dataset (which shows all the different region) you will see that the final structure is now what you want it to be (note that you air rectangle was set up with SiO2 and I have switched it to use Air).

  1. The second point is that in your “without grating” file you have used only two bias points. That is why the I-V looks linear. You actually have a reverse biased pn junction here so the I-V should be nonlinear. When I sweep the bias from 0.1 V to 0.4 V in 4 steps I do get a nonlinear I-V with the “without grating” structure as well.

  1. Finally, the last point to mention is that you have noticed (in the plot above) that the current is your structure is extremely small (not something that you can measure experimentally). The extremely small current makes it hard for the solver to converge on a self-consistent solution (specially when the variation in the system with applied bias is close the level of numerical accuracy of the solver). I noticed that the current reported on the two contacts are not same and this mismatch is an artifact of the difficulty the solver faces due to the small amount of current in the system. After playing with the file a little I was able to get the contact current to agree (indicating that the result is valid) by applying a negative voltage on the bottom contact (instead of a positive one at the top). This biasing condition should be identical to what you had and hopefully will allow you to get useful results. I had to use the same biasing scheme on both files to get them to work.

The simulation is very sensitive to the mesh as well and I had to fine tune the mesh as well. I also had to increase the accuracy of the solver in the advanced tab for the reason mentioned above.

Here are the modified files. Hope these will help you solve your problems.

WithoutGrating_MOD.ldev (6.7 MB)
WithGrating_MOD.ldev (7.6 MB)


#4

Hey @aalam, your support has been really helpful.
On your running the files with your modifications, I got the following results.

IV-Curve from the “without grating” file

IV-Curve from the “Withgrating” file

Though the results from the structure in which the terminal is not grated the result is expected, while the result from the etched terminal file is looking vague. Do you think their is a problem in the file or the structure or whether the structure is not possible to be implemented in Device? If you think it’s correct could you suggest a reason this could be happening?