(169bg) Can Classical Nucleation Theory Describe Heterogeneous Crystal Nucleation on Non-Uniform Surfaces?

Crystal nucleation stands as a pivotal phenomenon in numerous natural and industrial processes. Its profound implications extend to high-end industries, including semiconductors, solar cells, and aviation, where the understanding alongside the manipulation of heterogeneous crystal nucleation is of utmost importance as it is at the heart of the production of single crystalline materials. The time and length scales at which nucleation proceeds are too small to capture using most experimental techniques, presenting a challenge to experimental investigations of nucleation. As such, computer simulations are utilized to unveil the molecular details of the nucleation process. Historically, classical nucleation theory (CNT) has been used extensively to describe the physics of both homogeneous nucleation and heterogeneous nucleation. Despite its success, its validity or lack thereof under different circumstances is still debated.

In this work, we use molecular simulations and advanced path sampling techniques to probe heterogeneous nucleation of the Lennard-Jones (LJ) crystal on different uniform and patterned surfaces. To rigorously assess the applicability of CNT, we first compute heterogeneous nucleation rates using the jumpy forward flux sampling algorithm and examine their dependence on temperature. We also assess the stability of the FCC and HCP crystals next to each surface by conducting conventional MD simulations of FCC and HCP crystals at its vicinity and characterizing the extent of pre-melting. Finally, we also analyze crystalline nuclei along the nucleation pathway and find them to change shape upon enlargement, directly contradicting one of the core assumptions of CNT that crystalline nuclei are spherical caps with fixed contact angles. Overall, our findings underscore the nuances and intricacies of applying CNT to probe the kinetics of heterogeneous nucleation.