Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

How to calculate the Extinction Cross Section of two Ag side by side nano sphere(R=10nm) with a gap of 4nm in Comsol Multiphysics?

Please login with a confirmed email address before reporting spam

Hello all,

I'm trying to calculate the extinction cross-section of Ag sphere Dimer (two NPs side by side of R=10nm), with a gap of 4nm, in Comsol Multiphysics. Now, I know how to calculate the extinction cross-section of a single spherical Nanoparticle in Comsol and, also have analytically sloved it. For a single nanoparticle the Power-flow outwards of the Nanoparticle surface is integrated. e.g. nrelPoav = nxemw.relPoavx + nyemw.relPoavy + nz*emw.relPoavz; sigma_sc = (intop_surf(nrelPoav)/S_in)/sigma_geom; sigma_abs = intop_vol(emw.Qh)/sigma_geom;

where, intop_surf integrates over the surface of a single Nanoparticle and sigma_geom is the geometric cross-section. and S_in = E0^2/(2*Z0_const)

Now, I am confused, when calculating the Extinction Cross section of two NanoParticles should I integrate over both the NP surface in a similar manner but not able to calculate the accurate Extinction Cross section, or Kindly let me know which approach should be use to calculate the same???

Thankyou so much in advance.


1 Reply Last Post Oct 22, 2019, 2:48 a.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 22, 2019, 2:48 a.m. EDT
Updated: 5 years ago Oct 22, 2019, 5:49 a.m. EDT

Hi,

as far as I know, when calculating cross-sections from Poynting vector, one should integrate over a sphere that is furthest away from the source (as long as this sphere is between PML and the scatterer). When calculating scattering cross-section from Efar, one should use a sphere that lies the closest to the scatterer (that would be a sphere of far-field transform) since such a sphere usually has the best-resolved mesh (and the built-in Efar variable has an intrinsic error - refer to the definition in the manual). These I have learned on COMSOL conference of late.

Btw, where did you find the equations for cross-sections? I don't get why there is this division by geometrical cross-section?

S_in = E0^2/(2Z0const) has unit of and thus (intop_surf(nrelPoav)/S_in)/sigma_geom gives you units of , which is not the proper unit for cross-section (rather the efficiency).

intopvol(emw.Qh)/sigma_geom gives units of , which again is not the proper unit for cross-section (rather the power flow).

For sure I can tell you that the proper way to define absorption cross-section is intopvol(emw.Qh)/S_in. This yields proper values and units when the integration is over the near-field.

When talking about scattering - this is another story. I have some doubts, too, because I obtain electronic-resonance-like curve instead of plasmonic one - when calculating from Poynting vector (or the values are too small by 15 orders of magnitude when with proper plasmonic curve - when calculating from Efar).

EDIT: Now, I have found in my notes that for the scattering cross-section you need to use emw.nPoav, as this is time-average power outflow (which is what scattered light is, see Stratton's "Electromagnetic Theory" quote below):

The second term obviously measures the outward flow of the secondary or scattered energy from the diffracting sphere, and the total scattered energy is

(20)

Therefore

sigma_sca = intop_surf(emw.nPoav)/S_in

where integration is over boundary between far-field and PML. Now I also got rid of the problems with such a solution.

EDIT2: Just to mention: In my case the Far-field to PML boundary has wavelength-dependent size (radius = lambda/2), so (whether I'm right or not) I had to add geometry part as: , where r_eff is effective radius of all scatterers volume.

Cheers,

Radek

Hi, as far as I know, when calculating cross-sections from Poynting vector, one should integrate over a sphere that is furthest away from the source (as long as this sphere is between PML and the scatterer). When calculating *scattering* cross-section from Efar, one should use a sphere that lies the closest to the scatterer (that would be a sphere of far-field transform) since such a sphere usually has the best-resolved mesh (and the built-in Efar variable has an intrinsic error - refer to the definition in the manual). These I have learned on COMSOL conference of late. Btw, where did you find the equations for cross-sections? I don't get why there is this division by geometrical cross-section? S_in = E0^2/(2Z0const) has unit of W/m^2 and thus (intop_surf(nrelPoav)/S_in)/sigma_geom gives you units of [1], which is not the proper unit for cross-section (rather the efficiency). intopvol(emw.Qh)/sigma_geom gives units of W/m^2, which again is not the proper unit for cross-section (rather the power flow). For sure I can tell you that the proper way to define absorption cross-section is intopvol(emw.Qh)/S_in. This yields proper values and units when the integration is over the near-field. When talking about scattering - this is another story. I have some doubts, too, because I obtain electronic-resonance-like curve instead of plasmonic one - when calculating from Poynting vector (or the values are too small by 15 orders of magnitude when with proper plasmonic curve - when calculating from Efar). EDIT: Now, I have found in my notes that for the scattering cross-section you need to use emw.nPoav, as this is time-average power outflow (which is what scattered light is, see Stratton's "Electromagnetic Theory" quote below): > The second term obviously measures the outward flow of the secondary or scattered energy from the diffracting sphere, and the total scattered energy is > > (20) W_s=\frac{1}{2} Re \int_{0}^{\pi}\int_{0}^{2\pi} \left(E_{r\theta}\tilde{H}_{r\phi}-E_{r\phi}\tilde{H}_{r\theta}\right) R^2 \sin\theta d\theta d\phi > Therefore sigma_sca = intop_surf(emw.nPoav)/S_in where integration is over boundary between far-field and PML. Now I also got rid of the problems with such a solution. EDIT2: Just to mention: In my case the Far-field to PML boundary has wavelength-dependent size (radius = lambda/2), so (whether I'm right or not) I had to add geometry part as: \oint\frac{emw.nPoav\left(\pi\, r_{eff}^2\right)}{S_{in}\left(\pi\,\left(\frac{\lambda}{2}\right)^2\right)}, where r_eff is effective radius of all scatterers volume. Cheers, Radek

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.