The deposition techniques of silicon nanoparticles (Si-NPs) have been widely researched because of the plethora of promising applications. An accurate calculation of the nanoparticle density is of paramount significance for the understanding of device performance. In this work, we statistically evaluate the deposition of 3 nm Si-NPs via spin coating and drop casting. We image the Si-NP samples using optical microscopy combined with ultra-violet illumination, under which they exhibit red photoluminescence. We then perform image segmentation through a computer vision technique, particularly Otsus method, which works by detecting an automatic threshold value to segment the micrograph into its foreground (Si-NPs) and background (substrate). Finally, we utilize a counting algorithm to determine the particle density, surface coverage and cluster size distribution (CSD) on each sample, and hence evaluate the outcome of the two deposition techniques. For spin coating, the effect of drop volume (from 505 L to 2505 L) at constant speed, and the effect of spin speed (100, 300, 400 and 500 RPM) are both investigated while for drop casting, drop volumes of 100, 300 and 500 L were examined. It was found that the nanoparticle cluster density increases with increasing volume until 300 L while increasing the spin speed decreases the cluster of NPs deposited but results in higher uniformity of particle distribution. For spin coating, the highest cluster density recorded was 410<sup>6</sup> particles/cm<sup>2</sup> for the sample with 1005 L at a speed of 400 RPM while for drop casting it was 8.7510<sup>6</sup> particles/cm<sup>2</sup> at a volume of 300 L. Moreover, the relation between particle density, size distribution and the deposition techniques is discussed in terms of the strong centrifugal force the particles experience during spin coating compared to drop casting, as well as the interdependency of bonding of the Si-NPs with the solvent molecules on the deposition efficiency.
Read full abstract