I'm trying to simulate breakdown voltage characteristics of 2D Silicon PIN diodes in cylindrical-shaped diode. I've used cylindrical keyword (Yaxis=0) in the Math section in Sentaurus Device in order to rotate the half-device built in SDE around the vertical axis but the BV simulation doesn't converge.
I tried to change mesh-sized in SDE and decreased min-step in Sentaurus Device but it didn't converge. However, my BV simulation worked in planar coordinate.
There exist complete Examples about cylindrical 3D devices in Silvaco web site (examples) : https://silvaco.com/examples/tcad/index.html
Just download and check how the cylindrical structure are built.
Dear Pathom Jai,
regarding your problem I would like to provide two comment/tips:
a) cylindrical structure might lead to numerical problems. It might be helpful to avoid the very near surrounding of symmetry lines, e.g. start with slight offset of 0.01 to 0.1um which should effect the validiy of simulation results (dependent of size of pn-diode).
b) in general BV might have convergence problems; in most cases it helps to switch from Quasistationary mode to Transient mode and use the BE method (Backward-Euler; switch inside Math section) which is most robust and deactivate Extrapolation mode.
This helps especially if you have incorporated floating zones or super-junction concept inside your pin-diode.
BR,
Thomas
Thank you Thomas Rotter for your suggestion.
As you mentioned, I switched Quasistationary mode to Transient mode and used keyword Transient=BE in the Math section, and the graph converge! but my BV graph can’t reach to the breakdown voltage, so I am trying to adjust min-maxstep but it still doesn’t work. ( I have attached my BV graph picture to this massage)
Dear Pathom Jai ,
welcome!
In cylindrical junctions, if you have a curvature from cylindrical junction to a planar junction there will be a concentration of the electric field lines causing early breakdown of the junction.
Therefore you have to increase the applied voltage slowly and map the the potential distribution and the the electric field. You have to locate the point of the maximum electric filed and the magnitude of the electric field in these points.
If this is the case you will have only breakdown in a very limited area of the device.
The other point which you have to consider is meshing problem. You have to make the mesh very dense near the curvatures of the junction.
I could not get put the shape of your junction.
Best wishes
- Badges
- Science topic