Influence of mold and substrate material combinations on nanoimprint lithography process: MD simulation approach

Abstract

A molecular dynamics (MD) study was performed to examine the effect of mold substrate material composition on the pattern transferring and defects of the resist polymer in a thermal Nano Imprint Lithography (NIL) process. As candidate materials, single crystalline nickel (Ni), silicon (Si) and silica (SiO2, alpha-quartz) for the rigid mold substrate, and amorphous poly-(methylmethacrylate) (PMMA) thin film for the resist were considered for common applications in NIL processes. Three different material compositions of Si mold Ni substrate, Ni mold Si substrate, and quartz mold Ni substrate were considered. In accordance with a real NIL process, a sequence of indentation relaxation release processes was quasi-statically simulated using isothermal ensemble simulation on tri-layer molecular structures consisting of a mold, resist, and substrate. To correlate the deformed shape and delamination of PMMA resist from the substrate in indentation and release processes, non-bond interaction energy between a rigid mold and resist was calculated for each combination of mold and substrate materials. The Si mold Ni substrate combination shows successful pattern transfer to the resist polymer even without an anti-sticking layer as a result of the desirable balance of surface free energy for mold and substrate materials. However, Ni mold Si substrate combination shows a critical delamination of the resist in the release process due to strong van der Waals adhesion between the resist and Ni mold. Similarly, the quartz mold Ni substrate combination shows the same delamination in pattern transfer, but the adhesion of the resist to the quartz mold is attributed to electrostatic interaction. In order to provide guidelines for material selection in imprint-like processes where surface adsorption and wetting characteristics are critical design parameters, a simple PMMA-rigid plate model is proposed, with which consistent surface interaction characteristics in the full model NIL process simulation can be obtained. (C) 2014 Elsevier B.V. All rights reserved.