Which method is better for analyzing metallic photonic crystals: FDTD or TMM?

Hi, the FDTD and the TLM methods, give approximately the same results. Try to use the CFDTD software: http://www.ecl.ee.psu.edu/demos_fss_cfdtd/cfdtd.zip

This is a PPT presentation: http://www.ecl.ee.psu.edu/ecl/newsoftware/cfdtd_ppt.zip

Good Luck!

P.S. If you are interested on the CFDTD, send me a private message and I will reply some material presented at lessons here at the Faculty of Telecommunication Engineering (University of Pisa).

Look at this IEEE Document [Application of a 2D-CFDTD algorithm
to the analysis of photonic crystal fibers (PCFs)
Yanjie Zhu; Yinchao Chen; Huray, P.; Xiaopeng Dong;
SoutheastCon, 2002. Proceedings IEEE - 5-7 April 2002 Page(s):215 - 219]

Do anybody have TMM source code for photonic crytsal?

what is TMM? maybe TLM?

TMM---transfer matrix method

Both of these methods will work fine for metals, but there are some important differences that may help you choose. These arise because FDTD is time-domain and TMM is frequency-domain. You should also note there are at least four different forms of TMM that are all quite different.

1. If the structure you are modeling is highly resonant or there are abrupt features in the spectral response, you will want to use a frequency-domain model because it will take a huge number of iterations in time to resolve the behavior.

2. If the structure you are modeling has many lattice periods in the z-direction, but periodic boundary conditions in x and y directions, you will want to use TMM because you can number crunch through one unit cell and then just keep stacking that transfer matrix onto itself to quickly model any number of lattice periods. FDTD much simulate the entire structure at once.

3. If you want to model the response of your structure over a huge range of frequencies, FDTD may be the better method because you will excite your device with an impulse and then record the impulse response wherever you are interested. So, in just one simulation you can FFT the impulse response to calculate the spectral response over a huge span of frequencies.

4. If the device you want model is very wide, you may have to use FDTD. Unless you have access to a cluster or supercomputer, the TMM method is essentially limited to modeling one lattice period in the x and y directions (any number in z). For 2D simulations, you can perhaps handle several lattice periods in x using TMM.

Hope this helps!
-Tip

Added after 3 minutes:

If you are still comparing methods, you may want to look into a somewhat new method I recently developed that I called the "Slice Absorption Method." For the types of metallic structures that I simulate, it has virtually replaced FDTD and TMM. For a short time, you can freely download a paper on the method from the Optical Society of America. Just copy and past "Slice Absorption Method" in quotes into Google and a link to the pdf will come up. Unfortunately, this version is still in the crazy format required for paper submissions. It is double spaced and figures are at the end.

Hope this helps!
-Tip