Guiding Waves

Photonic devices are the future of data transfer. Compared to conventional electronics, photonics offer the advantages of being lower power, and having a longer range and larger bandwidth. Fiber optics transmit data over vast distances with minimal loss. However, when it’s time to connect the cable to a waveguide, there are significant losses. As much as 90% of the energy put into the cable is lost at the connection point. I’m working on simulating this coupling loss at the fiber-waveguide interface.  Using specialized software, I’m designing a tapered waveguide which will focus more of the energy to and from the optical fiber. It requires a huge amount of energy to process all the data being generated by the internet. We can therefore cut down on the internet electricity usage by designing more efficient waveguides.

Waveguides are used to guide an electromagnetic wave over short distances where it will be modulated and converted from a light signal to a purely electrical one. My task is to use a CAD software called RSoft BeamPROP to design a waveguide structure that would minimize losses. Photonic design is basically applied physics; in order to efficiently construct and test the design, knowledge of classical electromagnetic theory is essential. To understand how the allowed energy states of the structure form, Maxwell’s equations must be matched along the boundary of the material. The solutions to Maxwell’s equations in waveguides are analogous to the quantum mechanical solutions to the finite-square-well.

Many undergraduate researchers that I know work on simulations. Simulations are very important because they allow scientists to better visualize the structure they are looking at and work out any design flaws before the device is fabricated. Manufacturing devices can be very expensive; a single wafer of certain semiconductors materials can cost up to $10000. BeamPROP uses solutions to Maxwell’s equations to model an electromagnetic wave propagating through different materials. By simply defining the structure geometry and specifying its index of refraction, the software can predict which solutions are permitted.

If anyone is interested in doing simulation work related to waveguides, I have a few suggestions. If you have time to read and comprehend how the physics simulation is programmed, you will vastly increase your efficiency. The greatest difficulty I faced was trying to understand where the software went wrong when it gave me a solution that didn’t make sense. By understanding the machinery behind the program, you can much better tailor your simulation to work with the software rather than against it. It will also give you a greater appreciation of how the physical principles of electromagnetism are coded into the program. Simulation software isn’t perfect. My PI has noticed that many people will place overconfidence in the software. This is why one must never turn off their intuition when performing simulations. Just because the software outputs a certain value, it doesn’t mean it is physical. A good way to reason out which solutions are legitimate is to run the same simulation using two different pieces of software. Until both are in agreement, it’s safe to assume that one or both programs aren’t giving you physical solutions.