micro glass

transmission

#1

i have to get the transmition for 100 micron glass of refractive index 1.5
it gives the transmition in all wavelength equal zero and it is impossible in reality
can any one help


#2

Hi @ahisham,

One important consideration is the wavelength range you are interested in. In the visible range for example, a thickness of 100 micron corresponds to more than a hundred wavelengths, so an FDTD simulation of the transmission can be challenging.

If you want to share your simulation file I can provide you more specific feedback.


#3

no problem
https://drive.google.com/file/d/0BwgTxzri39vAQzhFWXlhNUJLeEU/view?usp=sharing
again my problem is the transmission in all wavelength (visible and infrared ) equal zero and this is impossible


#4

Hi @ahisham,

In your simulation the glass layer has a thickness of 500 micron and the wavelength range is 0.3 to 1.1 micron. This means that for glass with index 1.5, there are 500um/(0.3um/1.5) = 2500 wavelengths (for light with wavelength 0.3um) which is a huge number. I think that the zero value for transmission was probably due to the simulation time not being long enough for the pulse to travel through the entire structure. However, even if we increase the simulation time, the main concern is the phase error due to the numerical dispersion. The phase error increases with the number of wavelengths along the propagation distance. You can minimize it by refining the mesh but this will increase the memory require for the simulation and the execution time significantly. In FDTD typically we can go up to 200 wavelengths without having to worry too much about the phase error, but beyond that point care is need to get the right results in FDTD and it can become unpractical to use this method if the number of wavelengths is too large.

In order to illustrate this points I modified slightly you simulation (see rrrr_modFG.fsp (343.7 KB)
). I reduced the glass thickness to 100um to reduce the simulation time and used only the wavelength range 0.4 to 0.41um. This is still a large thickness for the wavelengths of interest. To see the effect of the phase error we can compare the FDTD results with an analytical calculation using the script command stackrt. I did this below for mesh accuracy level 4:

Note that the FDTD simulation shows some Fabry-Perot oscillations but they do not match the analytical result. The reason is the accumulated phase error in the propagation using FDTD. The plots can be obtained using this script: analytical_transmission.lsf (644 Bytes)