Question about running LWE in 2D Cartesian and avoiding NaN errors with GDD

[889562]

Hi Nick,

First of all, thank you for your great work on LWE!
Recently, I’ve been running a simulation and ran into some difficulties. I obtained preliminary results, but under my parameters, I found that I could only successfully run it in 3D radial symmetry, while 2D Cartesian always fails to run to completion.

In addition, I tried adding GDD (for example, 132000 fs^2) to the input pulse. I increased the time window accordingly, but it’s very hard to avoid the error message: NaN detected in grid！

May I ask if this issue can be avoided in some way?
Any suggestions or hints would be greatly appreciated.

Thanks a lot for your time and help!

Best regards

Name: KTA Type: 2 Sellmeier equation: 0 1st axis coefficients: 5.5552 0.04703 0 -0.04030 602.9734 0 -249.6806 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2nd axis coefficients: 5.70174 0.04837 0 -0.04706 647.9035 0 -254.7727 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3rd axis coefficients: 6.98362 0.06644 0 -0.05279 920.3789 0 -259.8645 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Sellmeier reference: SNLO chi2 type: 1 d: 0.0 0.0 0.0 0.0 0.0 2.3 2.3 15.5 3.66 0.0 0.0 0.0 0.0 0.0 0.0 3.64 0.0 0.0 d reference: Measurement of the tensors of KTiOPO4,KTiOAsO4, RbTiOPO4, and RbTiOAsO4 crystals chi3 type: 2 chi3: 1.6031302473498233e-21 0 0 chi3 reference: Opt. Express 30, 30507–30524 (2022) Spectral file: None Nonlinear reference frequencies: 2.8176e+14 2.8176e+14 5.6352e+14 2.910606e+14 2.910606e+14 2.910606e+14 2.910606e+14

CrystalDatabase.txt

KTA11.txt

[NickKarpowicz]

Hi there, thanks!
The divergence you're seeing is related to the nonlinear absorption and plasma generation (this is always the trickiest thing to converge). If I set the NL absorption parameter to 0, both modes work with your settings.

The 3D radial symmetry can be a bit more robust against things like this because it effectively have smaller step sizes in some spatial directions, so I guess that's why it was surviving in this case.

The 1.5e-19 value you were using seems to be quite strong - its making pretty strong nonlinear effects at an intensity where otherwise not much Kerr effect was happening. Maybe it's too large? Usually I don't add it unless I see that the beam is collapsing due to Kerr lensing first; the higher order nonlinearities like plasma generation usually only become important once third order ones are significant.

Do you have any nonlinear absorption data that could help to calibrate that number? i.e. transmission vs. intensity like one gets in an open-aperture z-scan?

Best wishes,
Nick