Create a radial diffusion doping profile in DEVICE



In an earlier post, I had shared a script on how to create a radial doping profile in DEVICE (CHARGE transport solver) using the import doping object. In a lot of cases, the user would prefer a radial doping profile that has a diffusion pattern, i.e. the dopant concentration is high at the outer surface and exponentially decreases to a low value at the core. This is a much realistic representation of doping on a wire in compared to the constant profile in that earlier post. The new erfc function in Lumerical’s S2016A release allows us to write a script where the radial doping profile can have a diffusion of dopants from the surface to the core. In the attached script file, I have created a radial dopign profile that exponentially decays from a peak concentration to a low value over a certain length (junction width). It is similar to the diffusion doping object available by default in DEVICE which is limited to rectangular shapes only.

I have also attached a DEVICE project file to help visualize the resulting profile. To see the doping profile, open the project file, load the attached .mat file into the (import) doping object, calculate the mesh and visualize the grid.

NOTE: Both the script and project file will run only in the latest version (DEVICE 5). You can download the latest version from here .

radial_diffusion_doping.lsf (2.4 KB)
readial_diffusion_doping.ldev (5.4 MB)
rad_diff_doping.mat (2.5 MB)

How to define doping in DEVICE?
Can doping boundary be curved?
Unknown problem in doping
Unknown problem in doping
complex doping
Error: Initialization failed to converge electrostatic potential update

In the code, the peak concentration is in the order of 25. However, as shown in the image, the max. doping concentration is the order of 19 (in the color bar). why is that?!


In the scripting language, everything is defined in SI unit. So when we define the concentration in script, the unit is /m^3. However, if you look at the “Results” tab in the CHARGE solver properties, you will see that carrier densities are reported in units of /cm^3 which is the more common practice. This is why the peak doping concentration in the image is a factor of 1e6 less than seen in the script.


The previous code gets the good results when the raduis of radial doping is greater than or equal 0.5e-6 but when it is below 0.5e-6, the followoing messege appears:

radial_diffusion_doping.lsf line 31: matrix arguments must be greater than or equal to 1

How can I get results for the raduis of radial doping=0.15e-6?


Hi. The script uses a junction width of 0.5 micron. This is the distance over which the doping goes from the peak density to the lower density. So the size of the radial profile needs to be larger than this. If you want to make the profile 0.15 micron in length then the junction width will need to be smaller than that.


A post was split to a new topic: Creating a ring doping profile for a nanowire