Error Recursion limit reached: segment bisection

Hi Lumerical Team

I am simulating a ring resonator modulator. In Device environment, I faced an error that I could not solve. The error is:

I have attached my simulation. It would be appreciate if you let me know how I could fix it.

n.mat (6.9 KB)
p.mat (3.2 KB)
Ring Resonator Modulator.ldev (6.6 MB)

Hi @omid.jafari.1, The geometry that you have here is a 3D geometry so you cannot simulate it with a 2D solver. You will have to use a 3D solver. However, a 3D simulation of such a large structure can be very time and memory consuming. I would recommend performing a 2D simulation on a slice of the structure (see attached file: Ring Resonator Modulator_2D.ldev (6.7 MB)). I have set the simulation region so that only the cross-section between the anode and the cathode is simulated.

This is how we simulated the performance of this pin MZI structure (PIN Mach-Zehnder) which is commonly used in ring modulators. Once you calculate the effective index of the 2D device (cross-section) you can use that to find out the phase shift in the entire ring.

​Hi @aalam, Thank you very much for your helping . After analyzing what you said, I was able to run this simulation thoroughly. Of course, there is a problem by simulating 2D, in fact I cannot see the device behavior in a part of the ring where is the closest area to the straight waveguide because there is no metal (NCathod) in the one side of this area. I think, I have to use a 3D geometry for considering it, but I faced a problem. I would be grateful if you tell me How I could fix 3D geometry for this structure.


From: Ahsan Support Alam <lumerical@discoursemail.com>

Sent: Monday, March 6, 2017 3:20 PM

To: Omid Jafari

Subject: [Lumerical Knowledge Exchange] [Solvers & Features/Charge Transport] Error Recursion limit reached: segment bisection aalam
Lumerical Team

March 6

Hi @omid.jafari.1, The geometry that you have here is a 3D geometry so you cannot simulate it with a 2D solver. You will have to use a 3D solver. However, a 3D simulation of such a large structure can be very time and memory consuming. I would recommend performing a 2D simulation on a slice of the structure (see attached file: Ring Resonator Modulator_2D.ldev (6.7 MB)). I have set the simulation region so that only the cross-section between the anode and the cathode is simulated.

This is how we simulated the performance of this pin MZI structure (PIN Mach-Zehnder) which is commonly used in ring modulators. Once you calculate the effective index of the 2D device (cross-section) you can use that to find out the phase shift in the entire ring.


Visit Topic or reply to this email to respond.

To unsubscribe from these emails, click here.

Hi @omid.jafari.1, my apologies for the late reply. If you want to run the 3D simulation then here is what you have to fix:

  1. The solver edges in x and y direction is touching the ring. This is making it very difficult for the solver to build the entire 3D geometry. I have made the simulation region slightly larger than the ring (on the XY plane) and that solved the geometry building issue in 3D.

  2. Another thing that I noticed is that the “scale factor” in your import doping objects are set to zero. This means that those dopings are not going to be applied. I have set these values to 1.

  3. Finally, I noticed that the mesh was not getting refined properly so I increased the “max refine steps” and increased the “min edge length” a little as well. Please take a look at this post for a discussion about how to get the mesh refined to the full extent: How to resolve warning message: “The mesh is valid but may not be adequately refined.

The modified file is here: Ring Resonator Modulator_MOD_3D.ldev (6.7 MB). The solver is able to mesh it so I believe it should be able to run it as well (provided that you have the required computational resources).