I stayed on campus for my sophomore summer, where I had the honor to gain hands on experience in semiconductor fabrication. I have always been interested in the precise work being done in amber-lit cleanrooms, and I jumped on an offer to experiment with photolithography techniques.
2025
Manufacturing Co-Lead
Cleanroom Protocol
Communication through Failures
Hand-off Teamwork
As part of an exploration into the Material Science department, a team of three others and I were given access to the resources available to us in micro-engineering. A lab instructor mentored us on how to safely and effectively use various equipment, but left us to our own devices on what to investigate by the end of the summer. The team decided to examine how dosage and deposition parameters influenced the sheet resistance of micro-channels, as measured by a four-point probe. Through the course of the project, I learned cleanroom and SEM protocol that bridged the gap between theory that I have learned of how tiny circuits are made and seen on silicon with the physical process of how it is actually done!
The team organized ourselves into roles so we could hand-off work like in industry: Joe would be the mask (the "print") designer, Jaden and I both would head the manufacturing process, and Aksel would take on the imaging process. After learning how the photolithography machine worked and how to use it, we experimented with a checkerboard design to see how different beam doses and wavelengths impacted the "print" of closely placed right angle edges. Using error analysis between tests, we determined 405nm wavelength light yielded the optimal exposure for the size of micro-channel we were hoping to print (~10 μm), with a dose of 130 mJ/cm² resulting in the most detailed line edges at this wavelength. After this, Joe designed our varying resistor designs in KLayout for us to etch into our silicon wafer.
The manufacturing process, like the nature of physics, was cruel and apathetic. The development process was very meticulous; one small piece of dust or skin cells ruined the photoresist and development stage (see image 2 of left). The photolithography process is funnily enough much like analog photography (it is a hobby of mine to work in a camera darkroom): the silicon wafer is covered with a "film" that light (photolithography) etches into, until put into a developer bath where the etched space (chemical change) washes away.
Once we got a good print, we needed to deposit metal over the wafer. We attempted many methods (gold sputtering (see image 1 of below), copper e-beam deposition, and aluminum evaporation (see image 2 of below)) that needed a lot of troubleshooting. Each failure required a new wafer to be processed, proving to be a tedious, but insightful task. The deposition time and material lead to different failures, such as improper adhesion and coverage (see images 1, 3, and 4 of left). We imaged each method to determine aluminum evaporation to be the best for our use case (perpendicular walls). We then put the wafer into an acetone bath to "lift off" the aluminum not in the channel (remaining photoresist peels off).
One of successful resistors was imaged seen below. Using the four-point probe method of calculating sheet resistance of our 10 μm wide, 50 nm thick aluminum channel, we found our resistor to have a resistivity around 6 times that of bulk aluminum. This was on the higher level of thin-film increase in comparison to established literature, which we theorized to be a result of surface oxidation. However, the increase in resistivity demonstrated the effect of surface and grain scattering of electrons on the microstructure level. This property is extremely important to semiconductor manufacturing, especially in modern nm-process scales.
This project gave me an insight into semiconductor manufacturing by learning and troubleshooting machines relevant to the field. Our ability to experiment greatly furthered my understanding of the material science behind our results at a microscopic level that is sometimes hard to grasp due to its conceptual nature. I enjoyed working in a group where team members took on separate roles, but understood the flow of process, allowing us to be on the same page across disciplines. We could hand-off our wafers and pick up individually right where we needed to. Working in the cleanroom was fun and taught me about the situational awareness and responsibility needed on industrial processes. Thank you to our lab instructor Phil Chapman for help and access to the lab!
I stayed on campus for my sophomore summer, where I had the honor to gain hands on experience in semiconductor fabrication. I have always been interested in the precise work being done in amber-lit cleanrooms, and I jumped on an offer to experiment with photolithography techniques.
2025
Manufacturing Co-Lead
Cleanroom Protocol
Communication through Failures
Hand-off Teamwork
As part of an exploration into the Material Science department, a team of three others and I were given access to the resources available to us in micro-engineering. A lab instructor mentored us on how to safely and effectively use various equipment, but left us to our own devices on what to investigate by the end of the summer. The team decided to examine how dosage and deposition parameters influenced the sheet resistance of micro-channels, as measured by a four-point probe. Through the course of the project, I learned cleanroom and SEM protocol that bridged the gap between theory that I have learned of how tiny circuits are made and seen on silicon with the physical process of how it is actually done!
The team organized ourselves into roles so we could hand-off work like in industry: Joe would be the mask (the "print") designer, Jaden and I both would head the manufacturing process, and Aksel would take on the imaging process. After learning how the photolithography machine worked and how to use it, we experimented with a checkerboard design to see how different beam doses and wavelengths impacted the "print" of closely placed right angle edges. Using error analysis between tests, we determined 405nm wavelength light yielded the optimal exposure for the size of micro-channel we were hoping to print (~10 μm), with a dose of 130 mJ/cm² resulting in the most detailed line edges at this wavelength. After this, Joe designed our varying resistor designs in KLayout for us to etch into our silicon wafer.
The manufacturing process, like the nature of physics, was cruel and apathetic. The development process was very meticulous; one small piece of dust or skin cells ruined the photoresist and development stage (see image 2 of left). The photolithography process is funnily enough much like analog photography (it is a hobby of mine to work in a camera darkroom): the silicon wafer is covered with a "film" that light (photolithography) etches into, until put into a developer bath where the etched space (chemical change) washes away.
Once we got a good print, we needed to deposit metal over the wafer. We attempted many methods (gold sputtering (see image 1 of below), copper e-beam deposition, and aluminum evaporation (see image 2 of below)) that needed a lot of troubleshooting. Each failure required a new wafer to be processed, proving to be a tedious, but insightful task. The deposition time and material lead to different failures, such as improper adhesion and coverage (see images 1, 3, and 4 of left). We imaged each method to determine aluminum evaporation to be the best for our use case (perpendicular walls). We then put the wafer into an acetone bath to "lift off" the aluminum not in the channel (remaining photoresist peels off).
One of successful resistors was imaged seen below. Using the four-point probe method of calculating sheet resistance of our 10 μm wide, 50 nm thick aluminum channel, we found our resistor to have a resistivity around 6 times that of bulk aluminum. This was on the higher level of thin-film increase in comparison to established literature, which we theorized to be a result of surface oxidation. However, the increase in resistivity demonstrated the effect of surface and grain scattering of electrons on the microstructure level. This property is extremely important to semiconductor manufacturing, especially in modern nm-process scales.
This project gave me an insight into semiconductor manufacturing by learning and troubleshooting machines relevant to the field. Our ability to experiment greatly furthered my understanding of the material science behind our results at a microscopic level that is sometimes hard to grasp due to its conceptual nature. I enjoyed working in a group where team members took on separate roles, but understood the flow of process, allowing us to be on the same page across disciplines. We could hand-off our wafers and pick up individually right where we needed to. Working in the cleanroom was fun and taught me about the situational awareness and responsibility needed on industrial processes. Thank you to our lab instructor Phil Chapman for help and access to the lab!
I stayed on campus for my sophomore summer, where I had the honor to gain hands on experience in semiconductor fabrication. I have always been interested in the precise work being done in amber-lit cleanrooms, and I jumped on an offer to experiment with photolithography techniques.
2025
Manufacturing Co-Lead
Cleanroom Protocol
Communication through Failures
Hand-off Teamwork
As part of an exploration into the Material Science department, a team of three others and I were given access to the resources available to us in micro-engineering. A lab instructor mentored us on how to safely and effectively use various equipment, but left us to our own devices on what to investigate by the end of the summer. The team decided to examine how dosage and deposition parameters influenced the sheet resistance of micro-channels, as measured by a four-point probe. Through the course of the project, I learned cleanroom and SEM protocol that bridged the gap between theory that I have learned of how tiny circuits are made and seen on silicon with the physical process of how it is actually done!
The team organized ourselves into roles so we could hand-off work like in industry: Joe would be the mask (the "print") designer, Jaden and I both would head the manufacturing process, and Aksel would take on the imaging process. After learning how the photolithography machine worked and how to use it, we experimented with a checkerboard design to see how different beam doses and wavelengths impacted the "print" of closely placed right angle edges. Using error analysis between tests, we determined 405nm wavelength light yielded the optimal exposure for the size of micro-channel we were hoping to print (~10 μm), with a dose of 130 mJ/cm² resulting in the most detailed line edges at this wavelength. After this, Joe designed our varying resistor designs in KLayout for us to etch into our silicon wafer.
The manufacturing process, like the nature of physics, was cruel and apathetic. The development process was very meticulous; one small piece of dust or skin cells ruined the photoresist and development stage (see image 2 of left). The photolithography process is funnily enough much like analog photography (it is a hobby of mine to work in a camera darkroom): the silicon wafer is covered with a "film" that light (photolithography) etches into, until put into a developer bath where the etched space (chemical change) washes away.
Once we got a good print, we needed to deposit metal over the wafer. We attempted many methods (gold sputtering (see image 1 of below), copper e-beam deposition, and aluminum evaporation (see image 2 of below)) that needed a lot of troubleshooting. Each failure required a new wafer to be processed, proving to be a tedious, but insightful task. The deposition time and material lead to different failures, such as improper adhesion and coverage (see images 1, 3, and 4 of left). We imaged each method to determine aluminum evaporation to be the best for our use case (perpendicular walls). We then put the wafer into an acetone bath to "lift off" the aluminum not in the channel (remaining photoresist peels off).
One of successful resistors was imaged seen below. Using the four-point probe method of calculating sheet resistance of our 10 μm wide, 50 nm thick aluminum channel, we found our resistor to have a resistivity around 6 times that of bulk aluminum. This was on the higher level of thin-film increase in comparison to established literature, which we theorized to be a result of surface oxidation. However, the increase in resistivity demonstrated the effect of surface and grain scattering of electrons on the microstructure level. This property is extremely important to semiconductor manufacturing, especially in modern nm-process scales.
This project gave me an insight into semiconductor manufacturing by learning and troubleshooting machines relevant to the field. Our ability to experiment greatly furthered my understanding of the material science behind our results at a microscopic level that is sometimes hard to grasp due to its conceptual nature. I enjoyed working in a group where team members took on separate roles, but understood the flow of process, allowing us to be on the same page across disciplines. We could hand-off our wafers and pick up individually right where we needed to. Working in the cleanroom was fun and taught me about the situational awareness and responsibility needed on industrial processes. Thank you to our lab instructor Phil Chapman for help and access to the lab!