(495g) Coarse Grained Molecular Dynamics Model of Block Copolymer Directed Self-Assembly

Directed self-assembly (DSA) of block copolymers (BCPs) is a promising technique for producing sub-30 nm pitch regular patterns, and the development of these DSA techniques could benefit greatly from computer simulation of such methods. Current simulation methods such as mean field approaches suffer from a number of limitations that affect their accuracy and their level of detail. Unlike other models, molecular dynamics (MD) combined with realistic potentials for polymer behavior can potentially provide more accurate simulations of the inherent polymer behavior, dynamics, and equilibrium states without a need to guess modes of molecular movement and without oversimplifying interatomic interactions. This is especially important for polymer systems that have very different densities or interaction strengths. We have developed a novel model for simulation of BCP behavior and DSA processes based on molecular dynamics of coarse-grained polymer chains simulated using graphics processing units (GPUs) to perform the calculations. The polymers consist of a bead-spring type system where a single bead corresponds to multiple monomers. Each bead undergoes 3 interactions: a harmonic spring-like bonded potential, a harmonic angle (consisting of 3 consecutive bonded beads), and a non-bonded potential modeled using a Lennard-Jones like potential. This coarse graining can result in simulations ~100-200 times faster than atomistic simulations, depending on what scale of coarse graining is used in the polymer model. In addition to this, this work uses graphics processing units (GPUs) to perform the calculations. This is very effective because of the natural parallelizability of MD simulations. Using HOOMD-Blue, simulations can run up to 200 times faster on a GPU enabled computer as compared to a standard desktop computer. By combining the two speed gains from coarse graining the simulation and using GPUs for the calculations, the simulation executes at speeds at least 40,000 times faster than conventional atomistic simulations and approaches the speed of other more commonly used simulation techniques for BCPs. The simulation has shown comparable results to well-known mean field studies. For ideal symmetric copolymers, our model reproduces mean-field order-disorder transition curves. The model also shows similar results to mean-field studies of domain scaling degree of polymerization N in both the strong and weak segregation regimes. Once the MD simulation model was validated, it was used to investigate a series of practical questions related to DSA using block copolymers. This includes: (1) the effect of blending BCPs with other components such as homopolymers, (2) use of tri-block and higher multiplicity block copolymers for DSA, (3) predicting phase diagrams for BCPs where each block has large differences in density or cohesive energy density. Our model will be fully described and highlights of the extensive DSA studies performed using it will be discussed.