Imaging of Soft Matter Particulate Systems in Aqueous Solutions by Scanning Electron Microscopy

When using a scanning electron microscope (SEM) to image soft-matter systems, the imaging dose is limited due to the potential radiation damage to the sample. This limits the signal-to-noise ratio making it challenging to distinguish regions with similar contrast from each other. In our case study, while taking a SEM video of polymeric microgels diffusing in a suspension of ionic liquid, the microgels did not contrast significantly with the surrounding ionic liquid. The noise also made it challenging to visually identify single microgels. Additionally, manual tracking of particles can be a laborious process. A three-stage program was developed to optimize this process. First, in order to improve the visibility of the particles, the program enhances the contrast by squaring intensity values, reduces noise with a series of filters, and converts each video frame to black and white image data using thresholding. A new video is reconstructed from the edited frames. Secondly, once particle visibility is improved, a custom machine learning model is used to automatically identify the xy position of every particle. This model is trained using thousands of positive and negative images. Lastly, the program automates the analysis of the data.