Open-Source Computational Photonics with Auto Differentiable Topology Optimization
Round 1
Reviewer 1 Report
This is a very relevant work in the field of nanophotonics. It has been recently recognized that the optimization of nanophotonic structure could lead to unconventional structures for which specific modelization tools should be available. The present work proposes such an implementation together with an algorithm for the computation of gradients that should be very useful to the community. It is important that open source codes be available to limit the use of black-box commercial programs. The paper is well-written and proposes interesting examples.
I recommend publication.
Author Response
We first thank the reviewer for recommending publication. We are delighted that the code may be useful to others and share the idea of supporting open-source software.
Reviewer 2 Report
The manuscript presents a detailed description of the open source tool developed by the Authors for the solution of the Maxwell's equations with applications in the photonics field. The Authors present very clearly their implementation of Finite Element method, Fourier Modal Method and Plane Wave expansion method. The mathematical formulation is accurate and precise. References are rich and complete. The tool, available on git, can be an important resource for other research groups, companies and students all over the world. I thank the Authors for so actively supporting the idea of open source software.
While the manuscript presents some interesting examples, there is no explicit comparison between the reported results and results obtained using other already established tools. I think is important to validate the correctness of the proposed solution and I think these cross validations are indispensable for the publication of the manuscript.
I found the following typos:
line 63: the mesh generation IS
line 125 nanOrodes
line 134 showS
line 135 allowS
While it is not part of the submitted work, I suggest to organise the git sample code clearly showing which files can be used to reproduce the results presented in the manuscript. Maybe the filenames can be indicated in the figure captions?
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 3 Report
The article is interesting and well written. I would expect ore details on the test systems used in the paper.
Author Response
We first thank the reviewer for its comment.
" I would expect ore details on the test systems used in the paper."
We assume this reads "more" details.
We however do not fully understand what is being asked and what "test systems means.
- if this is about the parameters for the simulation and optimization, we believe that there is enough description in the main text, and additional insight can be gained from reading the code that is freely available
- if this is about the testing of the code itself, there are unit tests being implemented
Round 2
Reviewer 2 Report
I appreciate the effort of the Authors to reproduce literature results with their code. However, the updated version of the manuscript does not contain any reference to those tests. In the answer they mention the sites:
- https://gyptis.gitlab.io/examples (FEM)
- https://nannos.gitlab.io/examples (FMM)
- https://protis.gitlab.io/examples (PWEM)
but those sites are not mentioned in the manuscript and do not seem to contain any clear comparison with published results or data from commercial tools, just a few examples of they own code.
In my opinion, comparisons must be clearly accessible in the manuscript for any reader.
Author Response
We have added an Appendix with checks and comparisons to existing results in the literature for each numerical method described in the main text. Links to online examples and further comparisons have been clarified and mentioned at the beginning of the Appendix.
We hope that including those in the manuscript will help the reader gain confidence in the validity of the code
Round 3
Reviewer 2 Report
I thank the Authors for the introduction of the Appendix; the manuscript can be published in its current version.
