Steven G. Johnson <[EMAIL PROTECTED]> writes: > > On Aug 6, 2008, at 7:02 PM, Chen wrote: > > 1.Normally you would want time-average power for this application. > > You don't need complex fields for this. You could just average the > power over one period, after all. Or you could use the Fourier > transform (if you are careful about normalization). > > > 2. Yes, you have to do multiple simulations, one for each frequency. One > way is to use a CW source at each frequency and wait for the system to > reach steady-state. Another is to use a very narrow-band Gaussian > source and wait until the middle of the pulse. > > We often prefer to use semi-analytical methods for this, such as > coupled-mode theory, based on perturbative methods. If you are > careful, these can be practically exact -- perturbative methods can be > very accurate here because the nonlinear change in the refractive > index is very small in most practical materials. This is true even if > the overall phenomenon that results from the nonlinearity is quite > dramatic (e.g. optical bistability). > > Steven >
Dear Steven, Thank you for your comments about nonlinear transmission curves. In comment 1, you suggested I should use Fourier transform if I am careful about normalization. Can you give more details about how to execute to obtain the average power? In comment 2, can I use a delta funtion source instead of a very narrow-band Gaussian source? We all know that the delta funtion source is a constant in frequency domain after Fourier transform. Thanks a lot. best, Chen _______________________________________________ meep-discuss mailing list [email protected] http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
