Why COMSOL?
Initial advantages (subjective point of view)
- No one was using it at the moment when I start my PhD studies
- GUI looked fairly simple and user friendly
- 30kg of manuals
- Based on Jiaming Jin´s book
- Thought it is going to be easy to simulate far-field spectra of nanoparticles,
and to compare it to nanofabricated
Reality
- Did not have enough time to dedicate myself to this, since I am an experimentalist stuck in the clean room most of the day
- 30kg of pretty much useless manuals
- Expired license-no comsol support
- Version COMSOL 3.5 had bug in PML definitions, at least when solving for scattering field (I did not know about this for 2 years)
- Started browsing forums, and asked for help at www.physicsforums.com
- Results in first 2 years: A lot of frustrations, slow server and not enough memory, and idea of implementing my own PML
Breakthrough
- Release of COMSOL 3.5a with some hot-fix that solved problem of PMLs, and company admitted that older versions had this problems
- Getting 3.5a plus hot-fix for free, as compensation for my wasted time
- Getting into touch with some people that knew how to use COMSOL, or were dealing with same problems (hml, VM, Ron, Tatarin), and sharing of tips and tricks, discussions, etc
- Stratton-Chu doesn´t work for substrates (special thanks to my friend VM)
At this moment
Many things are now clear
- How to do meshing and how dense it should be
- How to get good far-field spectra
- How to simulate particles on the substrates (the most relevant case)
- Symmetry and periodic boundary conditions
- How to define excitation in predefined GUI edit fields (it is not clear from manuals or examples where only the most simple cases are covered)