Code and models to reproduce the calculations in "Comment on 'Angle insensitive filters based on Fabry–Pérot resonance structures' [J. Appl. Phys. 136, 193102 (2024)]"

In a Comment submitted to Journal of Applied Physics and posted to Arxiv, we have pointed out that we are unable to reproduce the calculations shown in Fig. 7 of the article "Angle insensitive filters based on Fabry–Pérot resonance structures" by Cao et al. (https://pubs.aip.org/aip/jap/article/136/19/193102/3320668/Angle-insensitive-filters-based-on-Fabry-Perot). In particular, we have attempted to replicate their calculations by using three different electromagnetic simulation packages: (1) The same FDTD-based software used by Cao et al. (Lumerical), (2) a finite-element-method solver (Comsol), and (3) a Matlab library which implements RCWA (Reticolo). All these three methods give identical results (within the accuracy of numerical simulations), but in strong disagreement with Fig. 7 of the paper by Cao et al.

Moreover, in our Comment we have pointed out that the results by Cao et al. can be well reproduced by using incorrect boundary conditions in the simulations, specifically, by using periodic boundary conditions (instead of Bloch/Floquet ones) when injecting a tilted excitation in the simulation domain. Using periodic boundary conditions leads to a scenario where the input wave is always injected at normal incidence (even if the port "tries" to inject it at an angle), which can well explain the perfectly angle-independent results obtained in Fig. 7 of their paper.

Description of the files in this repository:

We uploaded the scripts/files that we have used to attempt to replicate the author's calculations [Figs. 2(a-c) of our Comment], with RCWA (Reticolo), Comsol and Lumerical, by using the geometrical parameters provided by the authors, and the correct boundary conditions (Bloch/Floquet):

Code for Reticolo simulations ( Reticolo_Code.m );

Comsol file ( Comsol_Correct_BC.mph );

Lumerical File (Lumerical_CorrectBC.fsp) ;

We uploaded the Lumerical file that we used to perform calculations with incorrect boundary conditions (See our Comment for additional detail)

Lumerical File with wrong BC (Lumerical_WrongBC.fsp)

The zip file Data Provided by Authors + Our Analysis.zip contains:

The raw data provided by Cao et al. for Fig. 7 of their paper (as three separated excel files);

A Matlab script written by us which performs some analysis on the raw data;

Two figures showing the main results of our analysis;

A readme file describing how and when we received the raw data file;

The zip file TransmissionMaps_for_different_values_of_dm_dg_dw.zip contains additional plots generated by us (which would be too long to include in the Supplementary Material of the submitted comment) showing the transmission maps (i.e. transmission versus wavelength and angle) of the device for different values of the three geometrical parameters $d_m, d_w, d_g$. In particular, we calculated the transmission map of the device for any possibly value of the triple $(d_m, d_w, d_g)$ in the range  $[0.2\;\mu m,0.4 \mu m] \times [0.2\;\mu m,0.4 \mu m] \times [0.2\;\mu m,0.4 \mu m]$, with each parameter varying in steps of 10 nm. The results are collected into pdf files inside the zip file . Each pdf file shows the transmission maps of the device for a fixed value of $d_m$ (mentioned in the filename) and for different values of $d_g$ and $d_w$, as indicated in above each column and to the left of each row. Given the large number of panels in these plots, and to save space, the ticks, labels and colorbars have been removed from all panels. The axis labels, axis ranges, and colorbars of each panel in these plots are the same as in Fig. 2 of our Comment.