(217av) Topographic Effects On Chemo-Epitaxy in Directed Self-Assembly of Block Copolymer Films

Directed self-assembly (DSA) of block copolymers is a very promising technique for producing sub-30 nm pitch regular pattern. These patterns could be used as a lithographic technique in integrated circuit manufacturing. Often the DSA of block copolymers is guided by chemo-epitaxy induced by a patterned underlayer. Chemo-epitaxy is caused by different sections of the underlayer having different chemical interactions with the film, e.g. one section of the underlayer has a stronger chemical attraction to one block of the copolymer while another section of the underlayer has equally attractive interactions with both blocks of the copolymer, causing it to be seen as neutral to the block copolymer. However, there is concern that in creating the necessary chemo-epitaxy some topography is created in the underlayer, which in turn adds a grapho-epitaxial effect to the DSA. This topography may be caused by a variety of phenomena, for example some mass may be lost from the underlayer during the lithography process. Grapho-epitaxy has been used to guide DSA by itself, so having unintended topography in combination with the intended chemo-epitaxy could potentially change the level of defectivity found in the self-assembled block copolymers, for better or worse. In this work, the effect of chemo-epitaxy with topography on the DSA of block copolymers is explored using many different underlayers with varying chemical interactions with the block copolymer film. Examples of underlayers include underlayers with regular patterns, as well as underlayers with defects such as an underlayer with a particle on top or a hole in the underlayer. All explorations are performed using mesoscale molecular dynamics simulations, in which multiple monomers of a polymer are combined and treated as a single bead and then used in a molecular dynamics simulation. These simulations are run using parallel GPU computing in order to model very large systems for long simulation times.