A statistics-guided approach to precise characterization of nanowire morphology

Precise control of nanomaterial morphology is critical to the development of advanced nanodevices with various functionalities. In this paper, we developed an efficient and effective statistics-guided approach to accurately characterizing the lengths, diameters, orientations, and densities of nanowires. Our approach has been successfully tested on a zinc oxide nanowire sample grown by hydrothermal methods. This approach has three key components. First, we introduced a novel geometric model to recover the true lengths and orientations of nanowires from their projective scanning electron microscope images, where a statistical resampling method is used to mitigate the practical difficulty of relocating the same sets of nanowires at multiple projecting angles. Second, we developed a sequential uniform sampling method for efficiently acquiring representative samples in characterizing diameters and growing density. Third, we proposed a statistical imputation method to incorporate the uncertainty in the determination of nanowire diameters arising from nonspherical cross-section spinning. This approach enables precise characterization of several fundamental aspects of nanowire morphology, which served as an excellent example to overcome nanoscale characterization challenges by using novel statistical means. It might open new opportunities in advancing nanotechnology and might also lead to the standardization of nanocharacterization in many aspects.