Josh Kvavle and Dr. Aaron Hawkins, Electrical and Computer Engineering
The computer has changed every aspect of the modern world. It almost seems that this burst of technological advance in computers will continue without limit. One can go and buy a computer with more “gigahertz” today for less that another bought one for the week before. Gordon Moore, of Intel Corp., predicted this trend of scaling the transistor down every 18 months . This scaling is the reason we have seen prices continue to drop, number of transistors per chip to increase, and the personal computer become an everyday thing in most people’s lives. Though many have predicted an end to “Moore’s Law” of scaling the transistor, it has continued since the seventies.
This scaling cannot continue forever and many have their eyes wide open for other technologies that can keep up with our need for computing power, one of these technologies in integrated optics.
Integrated optics is the combination of both electrical and optical devices to perform tasks, and one of the most important types of optical device for use in integrated optics is the “coupler.” A coupler is a way to make light that is traveling in one material at a particular angle, begin to travel in another material. This is really important for integrated optics as one must be able to direct and confine light so as to make it go in the proper places.
One method of coupling light is to use a diffraction grating, an array of periodic grooves or changes in refractive index. Because of the periodic nature of these grooves or refractive index changes, the light that interfaces with this surface will undergo phase shifts, causing what are called diffracted orders. This basically means that the light will hit the surface and split into several beams propagating in different directions. Below is a picture of several types of diffractions gratings.
The aim of my research was to fabricate in-coupling and out-coupling gratings in Silicon Dioxide on a Silicon Substrate with dielectric waveguides. I was to use photolithography and reactive ion etching techniques for fabrication, and was to vary grating parameters to find optimal power efficiency.
The fabrication of the gratings went through many changes up until today. The first method I used to fabricate gratings was using photolithography and a photomask to generate .5 micron features. The size of these features was not small enough to really get light coupling. I looked at using an electron beam microscope to generate the small feature sizes that I needed, but found it too inconsistent, and the producible grating area was very small, and it was time consuming to generate gratings this way. The next phase was to modify the holography setup that had been used in previous years to be a Lloyd mirror holography setup.
Holography showed the best results and I began producing .25 um features with a .5 um period. To find the best parameters to maximize the efficiency of the grating, I varied time of exposure, and development time. Grating efficiency is how much of the incident light is transferred into diffracted beams. I used glass slides so I could measure the transmitted power of a 633nm LASER and using the ratio of incident vs. transmitted, optimized the parameters.
With consistent results in producing diffraction gratings in a photosensitive chemical called photoresist, I then tried to transfer that pattern into another more durable material called silicon dioxide. I had problems with this because the etching machine that I used tended to wash out the diffraction gratings before it would begin to etch the silicon dioxide. To solve this problem I used a solvent to make the thickness of the photoresist coating that I was using much small. With some very short high intensity etch steps, the gratings transferred into silicon dioxide.
I then had diffraction gratings that could bend the light into a waveguide. The next step was to characterize the waveguide material and thickness. I used a dielectric called silicon nitride for this at first. Then I tested the grating couplers to see if they worked. They didn’t work. Since then I’ve tried all kind of waveguide materials. Since the coupling is so sensitive to the index of refraction and thickness of the waveguide I’m still struggling at trying to get the light to couple in and then couple out again.
I’ve used some MATLAB modeling code to simulate the diffraction of the grating that I’ve made, and can’t see to figure out why it isn’t working like it should. Recently I contacted a man who has frequently produced this type of grating coupler, and as him many questions about the process. That was helpful in some ways. I now have a new plan. Rather than using a dielectric that we make here at BYU, I’m going to use a polymer for the waveguide and try to see if it will couple in and out.
The research has been successful in a few ways, but hasn’t materialized into working diffraction grating couplers yet. I hope that in the next months I can make some breakthroughs and achieve my goals.