On Wed, 1 Aug 2007, F.S.F. Brossard wrote:
> I am trying to define a structure with a defect and eventually look at
> the Q. However, I already obtain inconsistent results in the defined
> perfect crystal depending on whether I set the symmetry or not (please
> see attachments). One would expect half a hole at all sides of the cell,
> not just two of them (in the case without defined symmetry). Hence we
> should have 8 periods in the (1 0) direction. I assume setting the
> symmetry takes the mirror symmetry of the two sides without the
> additional half-hole, hence we end-up with 6 periods instead of the
> expected 8. Thanks for any advice or solution to this problem.
The problem is that your structure, without symmetry, is not symmetric.
Hence when you impose symmetry you get a different results.
The reason it is not symmetric is that
geometric-objects-lattice-duplicates was really designed for MPB, where we
always have periodic structures. Therefore, it only puts objects on one
side of the boundary, assuming they will wrap around. However, in your
Meep simulation you are *not* using periodic boundaries, so the objects
don't wrap around. You can set periodic boundaries by doing
(set! k-point (vector3 0 0 0))
in which case you will get the structure you expect to see. (Note that it
is perfectly possible to use both periodic boundaries and PML at the same
time.)
If you don't have/want periodic boundaries, then you probably shouldn't be
using geometric-objects-lattice-duplicates. Instead, use
geometric-objects-duplicates, which gives you more control over how many
duplicates you have and where they are placed.
Regards,
Steven G. Johnson
_______________________________________________
meep-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
