Multiscale morphology and contact mechanics of physisorbed Al and Cu nanoparticles

Using large-scale molecular dynamics simulations, we investigate the scaling of the morphological and contact mechanics properties of Al and Cu nanoparticles (NPs) physisorbed on suspended graphene. The characteristic linear size of a NP ranges from 1 nm to 49 nm, covering a length scale of 1.5 decades. The NPs were obtained using a procedure mimicking thermal dewetting of thin films. Calculations show that NPs with a surface area-to-volume ratio above about 1.8 nm^-1, or with a linear size under 3-6 nm, behave differently from larger particles. For these smaller NPs, scaling of their total surface area and volume with the linear size can deviate from quadratic and cubic dependencies, respectively. Their mean interfacial separation and relative contact area change rapidly with size, exhibiting substantial variation. In contrast, for larger NPs, these quantities approach the asymptotic value. The height distributions of all particles exhibit a narrow spike and a decaying tail, both of which can be fit to Gaussians for larger NPs. In contrast, the interfacial gap distributions are close to a single Gaussian. The height power spectrum density (PSD) heatmaps of the smaller NPs are smeared and do not manifest a clear structure in contrast to the six-fold symmetry of the PSD of the larger ones. The maximum spatial frequency of the hexagonal 2D PSD roughly corresponds to the nearest-neighbor atomic distance of Al and Cu. For NPs with diameters larger than 20-25 nm, the isotropic height PSD exhibits power-law regions. We also calculate the relative difference between the apparent contact area and the approximated area of the bottom atomic layer. Small NPs have errors above 10%, which decrease with size. Our simulations illustrate how surface topography evolves with NP size and suggest that larger NPs can have random surface roughness. These results highlight the size-dependent morphology and contact mechanics of Al and Cu NPs, which differ qualitatively at smaller length scales.