Dear Steven,

Thanks a lot for your detailed explanation on the dielectric constant stuff
in MPB. It really solved my puzzle. I have another quick question here if
you don't mind. Since MPB uses the same grid for E and H field, how does it
actually calculate the derivatives? Suppose it wants to know the curl of H
field at point A, does it just calculate the differences between its
neighboring points and then divide by 2dr? In addition, is the output tensor
epsilon_inverse the same one used by MPB internally to solve H field, while
epsilon the one used for E field?

Best regards,
Minghui Han


On 10/2/06 4:27 PM, "Steven G. Johnson" <[EMAIL PROTECTED]> wrote:

> On Sun, 1 Oct 2006, Minghui Han wrote:
>> Thanks Steve. I actually already tried to output both E and D, but then
>> the epsilon inferred from D/E varies for different Bloch modes. Since
>> what I want is the actual epsilon for Ex and Ey used in MPB, and I also
>> want it to be the same for all Bloch modes for a given structure, do you
>> think it is possible to implement the smoothing method externally by
>> myself, just following what is described in your MPB paper and Meade's
>> 93 paper?
> 
> First off, I should apologize - I confused this with the Meep mailing
> list, so I mis-answered your question.  MPB does indeed output the full
> dielectric tensor: do an 'h5ls' on epsilon.h5 and you will see all of the
> tensor components.  Also, there is no Yee grid or interpolation in MPB, so
> this is not an issue.
> 
> Second of all, you are inferring epsilon incorrectly if you think it is
> different for different Bloch modes.  MPB (and Meep) uses the same epsilon
> tensor regardless of the field pattern.  You're perhaps confused by the
> fact that epsilon is a *tensor*, so you can't infer it merely by dividing
> D/E.  (You need 3 D's and 3 E's at a given point to infer the tensor.)
> 
> It is certainly possible to implement the smoothing method by yourself,
> but it shouldn't be necessary.
> 
> Steven



_______________________________________________
mpb-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/mpb-discuss

Reply via email to