Thin film growth of isotropically and anisotropically interacting particles with lattice KMC simulations

Abstract

We performed the simulation studies presented in this thesis with the goal of better understanding the behavior of a thin film growing on a substrate. We used the Kinetic Monte Carlo (KMC) method to simulate cubic particles on a cubic lattice using two different approaches: One in which neighboring particles interact isotropically and one where nearest-neighbor interactions were anisotropic. The KMC method is widely used to study the temporal evolution of non-equilibrium systems. Here we use it first to study the growth of a film when cubic particles are deposited onto a substrate comprised of a different material. Specifically, we compare two models: The well-known solid-on-solid (SOS) model, in which no cavities are allowed to form inside the film, and a newly developed colloidal growth model (CGM). In the CGM, particles are allowed to desorb from the film into the gas phase, and possibly re-adsorb at a different site, leading to the formation of cavities and overhangs. We find that in the intermediate regime (O(10) monolayers), the systems will show one of several growth modes, depending sensitively on the ratios of inter-particle interaction to substrate interaction and diffusion speed to deposition rate. Specifically, the film can initially either wet the substrate or desorb from it. The point of the transition between these two modes is shifted with respect to the equilibrium value, hence we call this a dynamic layering transition. At longer times, the roughness of an initially dewetting film will decrease. The main difference between the models is whether the film will become completely smooth or retain a constant roughness. For 𝑡→∞, each film will eventually roughen, independent of the initial growth mode. Finally, we tie these findings together into a global phase diagram denoting the possible growth modes and the conditions for their occurrences. Comparison to experimental data shows a good qualitative agreement in the general growth modes. The main goal of such simulations is the comparison to experimental results. Since the previous model is extremely simplified (isotropic particles and interactions), it is not suitable to explain many phenomena occurring during the growth of films of anisotropic organic molecules, e.g. ordering of particles. Thus we extend the model by implementing anisotropic interactions between particles. Each particle now had an internal orientation, and the interaction strength now depended on the orientations and relative positions of neighboring particles. These interactions can be implemented to either model disc-shaped molecules (e.g. benzene) or rod-shaped molecules (e.g. pentacene). Here we find two new ordering transitions upon increasing 𝜂, the strength of the anisotropy, with the film first going from an unordered phase to one in which two orien tations dominate, and finally to one in which the third species starts dominating. The point of the first transition is close to the equilibrium value, while the second transition is a non-equilibrium one. At low values of 𝜂, before the second transition, the ordering behavior of the film is independent of the film growth mode, which depends on the en ergetic and kinetic parameters. For strong values of 𝜂, the change in ordering coincides with a very strong roughening if the film, which is due to the occurrence of long needles perpendicular to the substrate. We again tie these results together in a dynamic phase diagram, and comparison to experimental results now shows good agreement concerning the ordering behavior of many real-world molecules. Finally, we explore the behavior of mixed films consisting of particles of two species: An anisotropically and an isotropically interacting one. These simulations were performed specifically to model the behavior of a 1:1 mixed film of CuPC and C60.