Visualizing and characterizing scientific data from observations at the nanoscale is an essential and challenging task.
Using computer vision technologies, and creative programming open-source environments, we developed a series of software visualizations which achieve significant advances in the characterization of observations obtained by the Scanning Electron Microscope (SEM), an instrument that produces images of a sample by scanning the surface with a focused beam of electrons. Observed material are nanopillar structures. Nanopillars is an emerging technology, within the field of nanostructures, placed together in lattice-like arrays. Each nanopillar has a pillar shape at the bottom and a tapered pointy end on top. This shape, in combination with nanopillars’ ability to be grouped together, exhibits many useful properties, having many applications including efficient solar panels, high-resolution analysis, antibacterial surfaces and water reduction.
Dense and large-area arrays of nanometric scale structures are regularly fabricated at INL Cleanroom. The characterization of sub-micron structures (below 500 nm) at a large scale and its post data analysis is a very challenging task. The observed subject consists of approximately ten US billion (10.000.000.000) pillars of 1 um height and 500 nm in diameter, with 1 um period equally distributed among a diameter of 110 mm. With the obtained raw data, geometric parameters were extracted from the images with the development of our custom software using computer vision algorithms. A metrological analysis has been performed to determine the patterning accuracy in the XYZ dimensions. A method was developed and established to automatically determine the diameter of each pillar, its circularity factor, and neighborhood distances, in each observation. For this, a large set of SEM images at high and low magnification were obtained. For low magnification top view images (250 X), more than 2×10^8 pillars (corresponding to ~ 2%) were analyzed.
Nanotechnology is mostly based on physical processes, and it is interesting to note that this more analogical effect causes interesting visual outputs, similar to generative images, usually produced by computational noise or even random effects. Those geometries, nanopillar parameters, neighborhoods, shape characteristics, distances, scales, and codes, will be presented in the form of an interactive installation.
Here, we observe and characterize at the sub-micron level, with incredible accuracy, facing an immense quantity of data. The developed software is used to generate those visuals, and each number, shape, distance or formula is authentic, invisible to the naked eye, and only processed by top-notch and accurate code.