(417i) Evaluation of Three-Dimensional Line-Edge Roughness of Pre-and Post-Dry Etched Line and Space Patterns of Block-Copolymer Lithography

Over the last decade, Directed-Self-Assembly (DSA) of block copolymers has gained relevance as a promising ‘bottom-up’ technique to produce nano-lithographic patterns. While optical lithography has conventionally been the technology favored by the industry, it has inherent diffraction-based limitations to produce smaller nano-features. Moreover, it requires to use enhancement techniques which can prove to economically unfeasible. DSA produces features of the order of 5 - 50nm, quite similar to Extreme Ultra-Violet Lithography, an alternative non-conventional patterning technique, but with significantly lower total cost of ownership. To make DSA a reliable and robust technique, much of the research is focused on reducing defectivity and mitigating Line Edge Roughness (LER) and Line Width Roughness (LWR) of the pre-and post-etched polymer blocks. LER can be defined as the 3σ deviation in the edge roughness of a feature edge in magnitude and frequency from the ideal desired shape. High roughness values often inhibit the smooth functioning of the transistors by constraining the electron flow through the spaced channels. It is thus crucial to develop novel modeling and simulations techniques for DSA to harness the full potential of this technique and thus meet the International Technology Roadmap for Semiconductors standards for LER.

In this talk, we will focus on the comparative study of the experimental LER / LWR calculations with our coarse-grained molecular dynamics (CGMD) simulation results. In our recent work, we have outlined our CGMD framework and subsequent etching methodology used to form thin-film line and space patterns created by symmetric polystyrene-block-polymethyl methacrylate (PS-b-PMMA) with periods of about 28 nm, on a patterned Liu-Nealey (LiNe) flow substrate. The defect free lamellae patterns in this study were etched using a wet etching schematic which have been experimentally observed to favor patterns collapse especially at lower BCP film thicknesses. In an effort to closely match the experimental conditions, we have developed a novel dry-etching simulation methodology to study the selective removal of PMMA with O₂/Ar etch chemistry. Subsequently, we perform a three-dimensional edge detection on the resist domains (PS) to evaluate the LER and LWR and compare it with the pre-etched BCP interfacial fluctuations. In addition to the default etching, we also evaluate roughness contribution of only the etching process by kinetically freezing the BCP beads, thereby nullifying the post-etch polymer relaxation. The roughness evaluation is done both in the space as well as Fourier domains for a more accurate understanding of the low and high frequency contributions of the deviations. Strategies in the form of short-chained polymeric additives, that are resistive to O_2­/Ar etch chemistry and thermodynamically prefer to self-assemble at the BCP interface are assessed. The efficacy of these additive to protect the BCP interface from the etch-ions is studied by examining their edge roughness. Lastly, we compare our simulation results to experimental SEM images of equivalent line and space patterns using an in-house image processing code.