Due to the substantial complexities of atomic force microscope (AFM)-based dynamic manipulation models, these models require the identification of many parameters and inputs with varying levels of sensitivity. Various sensitivity analysis (SA) methods can be used to achieve such a goal. The Sobol method is one of the famous SA approaches based on variance, which is widely used today in various study and scientific fields. The dynamic models for the manipulation of micro/nanoparticles, which have been investigated in previous works by means of the SA methods, have involved spherical nanoparticles with smooth surfaces. Since different micro/nanoparticles have a variety of geometries and since surface roughness plays a significant role in the contact of these particles with different surfaces, in this paper, the sensitivity of the AFM-based dynamic manipulation model of rough micro/nanorods has been analyzed. By considering more diverse geometrical conditions for the target micro/nanoparticle, including a cylindrical geometry and rough surfaces, many geometrical parameters enter the model. Therefore, in this study, simulations have been performed by the Sobol SA approach in order to determine the sensitivity of the input manipulation parameters to the critical output values of manipulation, for the two groups of AFM and environmental parameters. Based on the results obtained from these simulations, cantilever thickness constitutes the most sensitive parameter in the manipulation of rough cylindrical micro/ nanoparticles. By analyzing the sensitivities of environmental parameters, it was found that the parameter of surface roughness in the range of 0–0.04 is highly sensitive for rough cylindrical micro/nanoparticles.
