Pictures I took with a high speed camera of different regions of droplet formation (2022)
Summer 2022
I was honored to join the MIT Hatsopoulos Microfluids Lab as a summer researcher in Professor Gareth McKinley's Non-Newtonian Fluid Dynamics Research Group. I had the opportunity to work on a project about the formation of droplets and their rate of production. Our goal was to better understand how to produce emulsions with a below 5% coefficient of variation in droplet size. Methods for creating emulsions with low variability can be valuable in the pharmaceutical industry as a method for drug delivery and in the food and medicine industries for mixing immiscible fluids. For this research I helped design the experiment, independently constructed the components of the experimental setup, ran the experiment, processed the data, created informational graphs, and drew conclusions from my observations.
Original setup
Revised setup
I spent several weeks in Fusion360, designing devices of approximately 70 mm x 70 mm x 30 mm to hold glass microcapillaries. I then fabricated the capillary holders using a Form3L 3D printer. I designed the capillary holder to ensure experimental consistency for varying capillary sizes. There were six to ten different capillaries so identical holders were needed for each. The new design connects to the tank so that the Luer lock can be connected to the holder and the capillary will be at a fixed location within the tank for all the different capillary sizes.
I also worked in Python to write a script that analyzed the experimental data videos, utilizing object tracking to gain position and velocity data from dozens of droplets in each video. This enabled me to analyze hundreds of data points, each one representing data from twenty or more droplets, over the span of only a few weeks.
To use the script, I inserted a path to all the videos that I wanted to analyze. Then it took a screen shot of the first frame of video and found all the circles in that frame. Then the script used the location of each circle as an input for an object to track for the rest of the time. It followed each droplet as it fell and calculated its velocity, given the pixel to mm ratio. Finally, it saved the screenshots of the first frame in each video into a folder and output a list of all the droplet diameters and velocities.