CHARGE'S Geometry


#1

Dear sir,

Based on what we can choose the geometry in charge. I have structure, when I am changing the dimension of geometry I have in each one different results, in contrary the FDTD. How could I know that the dimension of the geometry for CHARGE that I chose is correct?

Thanks in Advance


#2

Choosing the simulation volume in CHARGE just like in FDTD depends on the device you are simulating and the parameter you are measuring. For example, if you think about the simple pn-junction diode below; if you make the simulation region wider, you will get more current (since you have a bigger diode now). However, if you want to calculate the current density J = I/A, you will get the same value since the current density cancels the effect if width by dividing the current with area.

If you can share your device geometry and the parameter you are measuring, I should be able to share some more specific comments.


#3

Dear Sir,

I appreciated what you sent. I share with you my structure to choose the suitable CHARE’s geometry. When I change the x and y dimensions, I get the different results in each one. Please, Define the suitable CHARE’s geometry for my structure and explain why.
Note, I validated the previous structure from research article that was stimulated with another program and the eta for it=12.3%

Thanks in advance


#4

Thanks for sharing the file. In your structure, you have optical absorption (and generation of electron-hole pairs) in the entire device and it is position dependent (non-uniform). As a result you will need to simulate the entire device or otherwise you are not counting some of the photogenerated electron-holes. When you use a smaller simulation volume you are missing out on some of the photogenerated electron-holes and so you should expect to see an effect for the simulation volume on the result. Let me know if this answers your question.


#5

Dear Sir,

Could you help me to fix my structure?

Thank you


#6

Dear sir,

I’m trouble sorry. The previous structure was sent by mistake. Could you discard it?
I’m sharing with you the correct structure and I have problem when run it as 3D, Could you help me to find my mistake to get the proper results?

Thanks in Advance


#7

Thanks for sharing the file. I have taken a look and there are a few points to mention,

  1. Geometry building: When it comes to 3D simulation in DEVICE, building a complex geometry can be hard. In your file, the solver was having trouble creating the structure. I have fixed it by following the suggestions on this post: Geometry building issue in 3D simulation for DEVICE.

  2. The scale factor for both the import doping and generation object was set to zero. I’ve made these values 1.

  1. The import doping value seemed to be too low (the unit of the data should be in /m^3). I have updated the value so that it is 5e24 /m^3 (~5e18 /cm^3). I have also made the minimum zero since I do not believe you want to dope the nanowire so heavily as it will incur recombination loss.

  1. Finally, I have used a ring geometry for the emitter instead of the rectangle since you only want to emitter to couple to the heavily doped n-region.

The file now runs fine and the I-V looks reasonable. You may need to fine tune your doping to improve the collection efficiency though. For example place a rear diffusion doping layer at the base (as well as at the top surface near the emitter) to reject minority carriers (see the discussion here: https://kb.lumerical.com/prior_en/en/index.html?solar_cells_device_getting_started.html)

I have shared the modified file with you via messages as it is too large.