Designing an Optical Waveguide

In OPT262, Electromagnetic Theory, we did a problem based learning assignment centered around designing a slab waveguide. I, along with my partners Icel Sukovaty '23, Thomas Karpishin '23, and Joshua Teague '23, designed and optimized a single-mode optical slab waveguide that would communicate in a binary fashion (single pulse design, 1's and 0's).

The things that I think went very well for this project is our material dispersion optimization, which was my main contribution to the project, and our pulse creation. We spent many hours going over code in order to adequately fit curves to expected values, which we finally arrived at after employing extensive dimensional analysis. For example, it was up to us to decipher that we needed to use temporal frequency units instead of spatial frequency units in order to make our math work out. The related MATLAB files will be attached if you are interested in running/seeing our code.

For optimizing material dispersion, the first step was to calculate the effective index of the core/cladding pair that we chose. By using effective index, you are able to optimize for both materials at once instead of trying to balance two curves together. After that was to find at which wavelength the dispersion coefficient was zero. By then inputting this wavelength a temporal dispersion formula, we were then able to plot and find at which pulse width our temporal dispersion would be minimized. Using this data, we were then able to move onto defining the dimensional characteristics of our waveguide and designing the pulse.

One thing that went wrong was our attenuation calculations. For some reason we ended up at a very large amount of attenuation per unit distance. We chose to take a geometrical approach factoring in the Goos-Hanchen effect and calculating the amount of time that the light spent in the cladding. Essentially we defined a period of light travel, which is essentially the distance over which light goes from the middle of the fiber, reflects twice, and returns to the middle of the fiber again travelling at the same angle. This period was defined by distance along the x-axis. Since we were using fused silica in our core region (and it has negligible attenuation), we calculated the optical path length in the cladding per period and calculated how much attenuation would be seen over that distance. By then calculating how many periods were present over our travel distance, our belief was that multiplying our single period value by this scalar quantity would give us our final attenuation. There is clearly a flaw to this approach, but I have yet to meet with our professor to discuss to proper approach.

If you are interested in seeing any of the supporting documents, feel free to contact me.