Hi,
I am trying to get scattering cross-section from gold hemispherical nanoparticle (r = 20 nm) placed on the Silicon substrate. I am attaching my image of the geometry. Substrate is being modeled by a cylindrical domain inside air domain. The issue is I am not getting plasmonic resonance and just getting noise. I am aware of the "scattering from the substrate" model on COMSOL. However, I wanted to solve for scattering field instead of total field. And I think instead of representing entire lower half by the substrate, it could be represented by a slab too? What are your suggestions?
Regards,
Hamza
Dear Hamza,
may I ask how coarse is your mesh? If I need to bet on what is the problem I would say that very likely your mesh is very coarse and you cannot resolve with accuracy your simulation. Try to perform a h-refinement (mesh convergence study) to see if with finner meshes you get an asymptotic output.
Kind regards,
Julian
Dear Julian,
Thanks for your reply.
In the absence of scatterer I do get resonance (figure attached). So i believed meshing might not be the issue as there should be some resonant behavior.
I was wondering if using scattering field formulation is the source of problem?
But I will try what you are suggesting, can you tell me how to do mesh convergence study. My system has just 16 GB of RAM, I wonder if that will be a bottleneck?
I am setting global size of mesh to finer. The meshing of nanoparticle and substrate domain is set to extra fine.
Regards
Hamza
Hi J. Anaya . I edited the meshing to have atleast 6 elements per wavelength, but still there is no change in the results, any suggestions?
Dear Hamza,
It is very interesting that you are getting the strange behavior only when using the scattered field formulation. May I ask what are you using as background electric field? If you do not have that clear I would use the full field formulation --both are identical, but the scattered field separates the problem in the superposition both: Total field= Background field + scattered field. If you are not giving the right background field that can be the root cause of your problem.
Kind regards,
Julian
Sir J. Anaya . I am defining it as: E0*exp(-j*emw.k0*x) ..polarized in z direction.
I understand you must be busy, but I am gonna stretch my luck. I am inboxing you my file on your inbox, if you would be kind enough to see if I am making any bloopers.
Regards,
Hamza
J. Anaya . By changing the background field to: E0*exp(j*emw.k0*z) (x-pol.) I got the attached results.
Regards,
Hamza
Dear Hamza Humayun
I have looked the above discuss I think you have to excite your nano particle by defining the Ey and Hy component of the electric and magnetic field and you havent defined that. keep in mind port excitation is a crucial part in this simulation as i also struggled with it...Additionally, whatever dielectric contant you are using please crosscheck it once.....
Dear Hamza Humayun
Maybe the answer is a little late. Anyway, it may help other people who see the post. If you are working on the optical regime of electromagnetic waves, then a substrate of several millimeters or even larger is an infinite domain for the optical wavelengths. SO, if you represent it by a finite cylinder [ I guess that you set the dimensions in nm or several micrometers] then it does not correctly represent the substrate and instead, it represents a scattering object that alters at somewhat the optical field resulting in noise in the output. The model of the scattering on the substrate can be used to solve the scattered field as you wish.
Kind regards
M.H Elshorbagy
Dear Hamza Humayun ,
Please tell me in detail about what is your boundary condition you have used and did u get the final results?
kind regards
Zahra Razavi
M.H. Elshorbagy . Thanks for your recommendation. I must admit that I did not think the dimensions of block (substrate) will be impacting this much.
Best Regards,
Hamza
M.H. Elshorbagy I actually tried to form my model around a youtube tutorial as well as discussion on the link:
http://srdjancomsol.weebly.com/scattering-vs-total-field-click-for-page.html
I avoided the scatterer on the substrate model as I felt at that time that it was too complex.(I am coming from Lumerical background )
Regards
The substrate must be considered infinite. There is an exact example in COMSOL for this problem:
https://www.comsol.com/model/scatterer-on-substrate-14699
First one needs to compute the background Field due to the substrate (as an infinite substrate in x/y) using ports and symmetry boundary conditions (see attached link). Then one can introduce the scatterer in the resultant background Field to compute the scattered field due to the particle.
Kind regards,
Julian
Dear Hamza Humayun The models that you follow are old and were just an unsuccessful try. However, comsol added the solution to this problem in the library [ The scatterer on substrate problem]. If you need any help please feel free to ask.
Greeting,
M.Elshorbagy
- Badges
- Science topic
- Similar topics
- Chemistry
- Nanotechnology
- Nanostructured Materials
- Nanoparticles