Additional results and updates to adjoint-optimization tutorial involving waveguide mode converter

[oskooi]

Additional results including two new figures for the tutorial Broadband Waveguide Mode Converter with Minimum Feature Size originally added in #2365.

[stevengj]

Would be good to check if the constraint is active (i.e. near zero) for the case with a lengthscale of 225 where the Wenchao's a posteriori measurement is returning 275. If the constraint is active, then this may be limited by the accuracy of the constraint functions?

Also, make sure to run Wenchao's ruler code for both the linewidth and line spacing?

[oskooi]

Would be good to check if the constraint is active (i.e. near zero) for the case with a lengthscale of 225

Looks like the constraint function is not active for the line width and spacing in the final design because the constraint function values are both ~0.001 (i.e., positive). Because the constraint is inactive, this means that the optimizer therefore likely got trapped in a local optima with minimum feature size of 275 nm.

Separately, I have reverted the horizontal axis of the plot of the worst-case reflectance and transmittance back to the imposed constraint from the measured feature size.