In-depth characterization of the aggregation state of cellulose nanocrystals through analysis of transmission electron microscopy images

Dispersion of cellulose nanocrystals (CNCs) is of utmost importance to guarantee their reliable application. Nevertheless, there is still no consensual method to characterize CNC aggregation. The hypothesis of this paper is that dispersion could be quantified through the classification of aggregates detected in transmission electron microscopy images. k-Means was used to classify image particulate elements of five CNC samples into groups according to their geometric features. Particles were classified into five groups according to their maximum Feret diameter, elongation, circularity and area. Two groups encompassed the most application-critical aggregates: one integrated aggregates of high complexity and low compactness while the other included elongated aggregates. In addition, the characterization of CNC dispersion after different levels of sonication was achieved by assessing the change in the number of elements belonging to each cluster after sonication. This approach could be used as a standard for the characterization of the aggregation state of CNCs.

Automated summary

This study proposed a method to quantify the dispersion of cellulose nanocrystals (CNCs) by classifying aggregates detected in transmission electron microscopy images using k-Means. The classification was based on geometric features, and the change in cluster elements after sonication was used to characterize CNC dispersion. This approach could serve as a standard for assessing the aggregation state of CNCs.