On Tue, 14 Nov 2006, Ashifi wrote:
Well, I thought that there was a way to simulate what happens when
microwaves are incident on a nanoslit. It is routinely done with the
Finite Element Method - I assumed FDTD would do it too.
FDTD, at least in its standard form, has uniform spatial resolution.
Finite-element methods have nonuniform spatial resolution, and are
therefore more appropriate for problems where you want to analyze
phenomena occurring with a vast disparity in lengthscales.
So, you are specifying a computational cell of 0.66e8 x 1.98e8 pixels;
it's not surprising that there is an error!
From previous emails and information in the archives, I thought this was
required: a large enough computational cell (at least a wavelength in
size), a large enough computational volume for the source to propagate
before it hits the target (hence the long rectangular shape) and a high
enough resolution to make out the tiny features.
You need a high enough resolution, but you have to formulate a problem
that is tractable in a uniform grid if you want to use Meep. It may be
that your problem is not of this form (although there are ways to break up
problems like this into a set of coupled problems that can be analyzed
individually, each at its own lengthscale).
Different computational methods have their strengths and weaknesses, and
I'm certainly not going to claim that Meep is the best tool for all tasks.
Cordially,
Steven G. Johnson
_______________________________________________
meep-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
