Search for: All records

High-throughput scanning electron microscopy (SEM) coupled with classification using neural networks is an ideal method to determine the morphological handedness of large populations of chiral nanoparticles. Automated labeling removes the time-consuming manual labeling of training data, but introduces label error, and subsequently classification error in the trained neural network. Here, we evaluate methods to minimize classification error when training from automated labels of SEM datasets of chiral Tellurium nanoparticles. Using the mirror relationship between images of opposite handed particles, we artificially create populations of varying label error. We analyze the impact of label error rate and training method on the classification error of neural networks on an ideal dataset and on a practical dataset. Of the three training methods considered, we find that a pretraining approach yields the most accurate results across label error rates on ideal datasets, where size and other morphological variables are held constant, but that a co-teaching approach performs the best in practical application.

Despite persistent and extensive observations of crystals with chiral shapes, the mechanisms underlying their formation are not well understood. Although past studies suggest that chiral shapes can form because of crystallization in the presence of chiral additives, or because of an intrinsic tendency that stems from the crystal structure, there are many cases in which these explanations are not suitable or have not been tested. Here, an investigation of model tellurium nanocrystals provides insights into the chain of chirality transfer between crystal structure and shape. We show that this transfer is mediated by screw dislocations, and shape chirality is not an outcome of the chiral crystal structure or ligands.