Could you share the cross-sec result along with simulation details (material properties, simulation lambda range, and geometry). It will be helpful to provide some reasons behind this!.

Hello Vasanthan Devaraj ,

attach to this message an example of file with negative absorption cross section.

I try various dielectric function home made or provide by the toolbox.

But it seems that at a certain size negative abs Cs appears.

I have not found any documentation on it

Thanks for reading and thinking about it

Regards

Julien

Hello Julien Cardin,

Thanks for sharing the file. I'm not using MNBPEM toolbox. Can you share your result as image file (Cs) and well as geometry schematic. I can cross check with FDTD and compare. It can provide some understanding.

Dear Vasanthan Devaraj,

Thank you for your proposal, I am pretty sure that you will find a difference by using FDTD method, I attach the files of CS obtained. However, I think it's more in the technical use of the MNPBEM toolbox that I am doing something out of the range/limit of the method. The problem arrize when I start to use larger object and larger refractive index of the surrounding materials. My example was a silver cylinder of 30 nm diameter of height 20nm on silicon (eps=4^2) and surrounded bu Si3N4 (eps=2^2).

thanks again.

Regards.

Julien

Some news it seems that problem arise for large particles in quasi-static (bemstat) approximation and not with BEM solver for full Maxwell equations (bemret)

thanks

regards

Dear Julien Cardin,

From your CS results, it was not negative (I assume, for the mentioned smaller Ag cylinder size - 30 nm diameter/height 20nm , results are fine). Currently checking with larger Ag cylinder size as function of surrounding medium index. When problems arise? (for example For Ag cylinder diameter > XX nm /height > YY nm)

Thank you.

Vasanth

Dear Julien Cardin,

I have one doubt in your structure.

Is below structure model is correct?

I assume SiN as background medium instead of air!

Thanks,

Vasanth

Dear Vasanthan Devaraj,

Yes it's that I varied height from 1 diameter to 1/2 diameter

I used real part of refractive index for Si and Si3N4

(i have files with dispersed values)

You may use constant to be easier nSi3N4=2 and nSi=4

I used a home made oscillator model for silver but based on Rakic data

thanks

Julien

Dear Julien Cardin,

I had checked the CS results for Silver cylinder (d = 30 nm) and height variation for 10 - 30 nm using FDTD. In case of Silver I cross checked both with Rakic and J&C index data. For SiN - Luke data base and Si Palik database. In my simulations, I could not find any issues like negative abs CS. I'm not sure why negative CS can obatined. But I can give few suggesstion, which you can consider:

1) The planewave source to structure distance.

2) Power monitor to Structure distance which records the CS data.

3) Mesh size

I'm not sure about the toolbox methodology you use. But in general electromagnetic simulations, above 3 play crucial part as they need to be optimized to get the accurate result.

I hope this might help.

Thanks

Vasanth