(576b) Examination of Line Edge Roughness of Directed Self-Assembled Block Copolymers: A Coarse-Grained Molecular Dynamics Study

Over the past decade, with the ever-increasing demand for miniaturization of magnetics storage devices, flash memory devices, semiconductor capacitors among others, directed self-assembly (DSA) of block copolymers (BCPs) has attracted the interest of both academia and industry to develop nanolithography to form defect free patterns with optimal control and tolerance on the lithographic features. DSA is not restricted by inherent diffraction based limitations of conventional lithography and has a much lower cost of ownership compared to the existing alternatives. However, the theoretical harmonic spatial arrangement and periodicity of the micro-phases are seldom perfectly achieved due to a variety of defects causing potential limitations in micro-electronics applications. In particular, the lamellae forming symmetric BCPs used to create line / space patterns show roughness features at the interface of the two blocks. This roughness, quantified as Line Edge Roughness (LER) is defined as the deviation of a feature edge in magnitude and frequency from the ideal desired shape. Morphologies with high LERs often inhibit the smooth functioning of transistors by affecting the electron flow through the spaced channels. It is thus extremely crucial to minimize the LER to harness the full potential of DSA techniques to develop the next-gen devices.

In this work, we have adopted a large-scale coarse-grained molecular dynamics (CGMD) framework to analyze the LER variation for the line / space patterns formed by symmetric polystyrene-block-polymethyl methacrylate (PS-b-PMMA). Three types of substrate models have been considered to direct the BCP. 1) Flat patterned substrate 2) Wedged patterned substrate and 3) Wedged substrate with a random PS-b-PMMA polymeric brush. For the first type of substrate, the enthalpic contribution of the chemoepitaxial substrate is the driving force to form the lamellae. The wedged patterned has an entropic contribution from the topography of the substrate in addition to the patterning. The third type of substrate employs random PS-b-PMMA co-polymer brushes to mimic the neutral layer of the chemoepitaxial substrate. The BCP morphology formed by each type of the substrate has been optimized by varying the polymer and surface properties as well as the annealing conditions. The morphology obtained from large scale CGMD simulations is further processed by etching the PMMA domains resulting in the line / space patterns. Line edge roughness analysis is then performed to quantify the high and low frequency noise of the features by plotting the power spectrum density curves. We demonstrate that the correlation factors from fitting the power spectrum gives values comparable to experiments by accounting for instrumentation / measurement noise. Lastly, the effect of additives for reduction in LER and defect annihilation is also presented.