This summer, I interned at an Austin-based scientific computing company called Enthought. While Covid-19 forced many people to cancel their plans, I am extraordinarily lucky in that my father works for Enthought, and I was able to participate in the internship from home.
The goal of my internship this summer was to provide a better understanding of physics and engineering through both a classroom and hands-on setting. The final aim was to create a maskless photolithography machine, but before we could begin, I needed a much better handle on physics and coding. We also needed a proper workspace (for some reason my mom refused to let us use the welding torch and hazardous chemicals in our living room…) so we set out to transform our garage into a laboratory. For the first month of the internship, I spent half of my workday in “class” with my dad, learning to code and studying electromagnetics and optics. For the rest of the day, I got my first taste of mechanical engineering as we designed and built tables for our new laboratory. After several weeks of cutting, drilling, and welding and hours of pulling metal shavings from my hair, we finished four tables, one of which is pictured.
Once I was adequately educated and our lab was completely built, I could begin researching and designing the maskless photolithography machine. Photolithography is a process used in microfabrication to etch a pattern onto a substrate, generally a silicon wafer. The wafer is coated with a substance called photoresist, which degrades when exposed to UV light. A patterned “mask” is placed over the substrate to block UV light, leaving only the unmasked areas to be exposed and degraded. Then, the wafer is coated in a solvent that dissolves the degraded areas but leaves the rest of the photoresist intact. This process enables manufacturers to etch extremely small patterns onto silicon wafers. However, masks are extremely expensive and cannot be altered, so they aren’t an ideal solution. In my internship, I investigated a newer idea called “maskless photolithography,” which uses a projector to shine a pattern directly onto the substrate, eliminating the need for a mask.
To give an idea of what the project was like, I have documented the main three challenges— though there were certainly more— that I faced when building the machine.
The first hurdle was to coat the substrate with photoresist (while prototyping, I used a microscope slide as my substrate instead of a silicon wafer because it is cheaper). In order to perform photolithography, the substrate must have an extremely even coating of photoresist. To achieve this, I found I must first rigorously clean the microscope slide using various chemicals and baking it on a hot plate. Then, I would place the slide on a spin-coater and apply the photoresist liquid to its center. As the slide spun, centrifugal force would push the photoresist outward so that it formed an even layer on the slide’s surface. In the picture provided, I am completing the spin-coating process. Note the light in the room is orange— I placed a filter on the lightbulb that blocks all light under 450nm so that the photoresist is not degraded as I am spin-coating. (IMAGE #2) The resulting photoresist coat isn’t completely perfect, especially near the edges of the slide, but it is plenty accurate for our project. In fact, we were able to use data from a spectrometer to discover the photoresist coating was a constant 7 microns thick in the center of the slide, which was exactly our goal!
The next step was to design an adjustable mount for the projector. Since photolithography is used for making extremely tiny objects such as computer chips, the projected image needs to be extremely small with an accuracy of several microns— which is a fraction of the width of a human hair. In order to condense the projector’s image to this size, I mounted the projector above a microscope so that it shines through the microscope’s lenses and produces a tiny image through the objective lens. After much trial and error, we created a mount that can adjust the projector’s position in the x,y, and z directions, as well as two rotating axes which allow us to adjust the projector’s angle. For context, I have provided an image of my initial design for the mount accompanied by my notes for improvement. I have also included an image of the final product we created.
Lastly, I adjusted the lens path of the projector and microscope in order to focus and center the image through the microscope’s objective lens. The projector’s initial image was far too large to fit into the small microscope lens, so I 3D printed a device to hold a series of condenser lenses in order to reduce the image. However, each lens attenuates the UV light, which activates the photoresist, so I adjusted my final design to use only a single condenser lense which I positioned in an adjustable 3D printed mount. After making a few additional adjustments to the projector— such as removing the color wheel and UV filter, I was ready to test my machine.
As with every science experiment, the first attempt was largely a failure. Though the projector and lens mounts were completely adjustable, I still couldn’t seem to position the projector so that the image was both centered and focused beneath the objective lens. After designing several new iterations of a 3d-printed mount for the condenser lens— ultimately deciding to mount it to the microscope rather than the projector— I have nearly attained success. I am hopeful that several more iterations in the last few weeks of my internship will prove successful and the project will be complete.
I have thoroughly enjoyed my internship this summer, and feel so lucky to have been able to participate in it!