Extraction Efficiency and Angular Distribution Calculations

I’m having issues with extaction efficiency and getting angular distribution of radiation. I’ve been looking at two papers with triangular lattices:

hybrid phc paper: http://link.springer.com/article/10.1007/s11082-015-0140-3
yellow enhance paper: http://ieeexplore.ieee.org/document/7414383/

A few notes on the simulations:
-The yellow enhancement is with pillar structures for lambda = 550nm, x,y, and z oriented dipoles averaged in 16 places according to Lumerical’s help page on hexagonal symmetric lattices.
-Hybrid PhC is with structured air holes for lambda = 465nm, only in the x direction (unless I’m reading E//x incorrectly) for two locations: under a hole and away from a hole (as quoted in the paper).

For both of them, I’ve set a monitor to measure the transmission close to the surface (0.5 um) of the photonic crystal. When I run simulations for the patterned cases, I get very low extraction efficiency enhancements, something like 12% for the yellow enhancement and 4% enhancement for the hypbrid PhC. The yellow enhance paper doesn’t specify a number for the enhancement but the hybrid PhC case quotes an enhancement of 20% for the given parameters of the model. I chose these because both of them are using all PML boundaries but I’m not seeing significant improvements in extraction efficiency enhancement.
I thought I was doing things correctly from the unpatterned of the yellow enhancement paper as the given index of refraction gives me 9% (close to the calculated 8%).

For the pillar structured one, Ive tried increasing FDTD sizes, varying monitor heights, varying dipole depths but to no avail. The hybrid phc is less flexible because the paper was more explicit about its setup.

Additionally, I need to figure out the angular distribution of radiation to see how light emits from the near field to the far field but not entirely sure what represents that or how it is calculated.

Here are the files I have been working with

hybridphc.fsp (5.0 MB)

yellowenhance.fsp (7.8 MB)

Any help would be greatly appreciated.

Dear @ejfern

Thank you very much for reaching out and sorry for my late reply.

I had sometime to go over your simulation file and the paper. It looks like your geometry is slightly different that the one in the paper and I am not sure how you implement the deep PhCs and shallow PhCs mentioned in the paper (please correct me if I am wrong).

Also, it will be good to use smaller mesh size and trans-box to make sure that your dipole source power normalization is done correctly. With your current simulation file, dipole source size is almost the size of the PhC hole. You can read more about dipole box approach here and here for how to use it properly.

Also, in the paper they mention that they optimized the dipole source position (?), so I am not quite sure where the final position was. I think if you start with our KB example, and try to see how changing the mesh, source location, and other parameters effect the results, you can probably get a feeling for your simulaitons.