master.png
Microscope image of one nanopillar parameter sweep area with approximately 1,600 individual pillars. Produced by confocal 3D laser scanning microscopy.
In the process of fabricating nanostructures, parameters such as the laser power and exposure time can greatly affect structure dimensions. Thus, parameter sweeps are often used to determine parameter-structure relationships.
However, parameter sweeps and samples alike often have upwards of hundreds of thousands of individual structures. Measuring these by hand would take weeks. 
picturetest1.png
Microscope image of a nanopillar parameter sweep with approximately 1,000,000 individual pillars, testing four different parameters. Produced by confocal 3D laser scanning microscopy.
To make this process both more efficient and more accurate, I developed an automated data processing software that extracts structure dimensions from microscope scans produced by confocal laser scanning microscopy.
Screen Shot 2020-08-18 at 4.18.15 PM.png
Visual depiction of my data processing software. Taken from 2019 science fair poster.
The area of interest doesn't always match up to the microscope scan area, so the software automatically detects and crops the area of interest. After that, the surface heights are segmented into a base level and a structure level, the difference between which becomes the structure height (in this case, the software processes nanopillars). Thresholding the birds eye scan produces a binary representation of the structures and further dimensions (diameter, circularity, etc.) can be extracted from there.
This software not only enables drastically more efficient data processing, but also objectifies the process and leaves no room for the human error that manual measurements are prone to.
Screen Shot 2020-08-18 at 4.30.28 PM.png
Other attributes of nanostructured samples can also be studied, one of which is color. To analyze sample colors, I developed a data analysis software that extracts RGB spectra from sections within microscope images.
Microscope image of one nanopillar parameter sweep area with approximately 1,600 individual pillars. Produced by confocal 3D laser scanning microscopy.
Screen Shot 2020-08-18 at 4.29.38 PM.png
Visual depiction of my data analysis software. Taken from 2020 science fair poster.
Again, the area of interest doesn't always match up to the microscope scan area, so the software automatically detects and crops the area of interest. After that, the RGB values of pixels within each area are analyzed to produce population statistics (average, standard deviation, etc.). In parameter sweeps, these results are correlated with various parameters using a least squares model.