Understanding Contact at the Nanoscale Using Molecular Dynamics Simulations

Nanoscale contacts between materials are found in advanced technologies from nanomanufacturing to nanodevices to scanning probe microscopy. These applications require quantitative understanding of the nanoscale contact properties, such as contact area and adhesion. Continuum mechanics models are extensively used to describe and predict the behavior of such contacts, but are based on assumptions that may not hold true at the nanoscale. Additionally, the buried nature of the contact interface makes direct measurement of nanocontact a challenge. Alternatively, molecular dynamics (MD) simulations provide atomic details of the contact interface and can be used to investigate contact at the nanoscale. Here, physically realistic models of nanoscale contact were created with identically matched materials, geometry, crystallographic orientation, degree of amorphization, adhesion and loading conditions. The contact area was then computed from the positions and interactions of atoms. The results demonstrated that significant variation in the value of contact area can be obtained, depending on the technique used to determine it. Next, contact area from MD simulations was compared with predictions from electron transport theories. Electron transport theories were shown to underpredict the contact area by more than an order of magnitude. This low conductance of the nanocontact could not be explained by electron scattering off of defects, and was instead attributed to approximately a monolayer of insulating oxide on the contact surfaces. The effect of native oxide on contact area measurements was further investigated using a newly developed method to approximate conductance based on the distance between atoms in channels across the contact. Then, we compared the contact area calculated with ballistic transport and tunneling theories to that obtained using the known positions of atoms in the contact. The difference was small for very thin (<0.1 nm) or thick (>1.0 nm) oxides, where ballistic transport and tunneling theories work well; however, the difference was significant for oxides between these limits, which is expected to be the case in many practical applications. The prediction of contact area from continuum mechanics models also relies on accurate knowledge of the work of adhesion between the surfaces. An important assumption in these models is that the work of adhesion is constant for a given pair of materials. Our results challenge this assumption, instead showing a significant increase in work of adhesion with increasing pressure, and analyses suggest the presence of stress-driven chemical reactions in the contact. Overall, this research defined and computed contact properties using MD simulations and the results demonstrated the limits of empirical models, including continuum theories and electron transport theories, to describe contact at the nanoscale