Molecular Dynamics Modeling of Nanoparticle Impact

We are using the Large-scale Atomic/Molecular Massively Parallel Simulator package, LAMMPS, to model the impact of a liquid droplet on a Si surface. We plan to study a variety of problems such as sputtering and amorphization, the effects of projectile energy and diameter, angle of impact, projectile dose, etc. For now we are working on the mechanisms behind the observed amorphization of Si. Typically, we model the impact of a droplet with a diameter of 10 nm made of 1224 identical spheres distributed in a hexagonal close-packed arrangement. Each sphere has a radius of 0.422 nm, a mass of 391.31 amu, and represents a molecule of the ionic liquid 1-ethyl-3-methylimidazolium bis (trifluoro-methylsulfonyl) imide, C₈H₁₁F₆N₃O₄S₂, used in our nanodroplet impact research. The target is a rectangular slab with a cross section of 48.88 x 48.88 nm² and a thickness of 30.55 nm, filled with 4,390,200 Si atoms in the standard cubic-diamond arrangement. The interactions between liquid molecules and between liquid molecules and Si atoms are modeled with a Ziegler-Biersack-Littmark two-body potential, while the Si-Si interact ion is reproduced with the Stillinger-Weber potential.

 

The simulations shows that the amorphization results from the heating and subsequent melting of a thin layer of silicon surrounding the area of the impact, followed by an ultrafast quenching with cooling rates surpassing 10¹³ K/s. These conditions impede crystal growth in the supercooled liquid phase, which finally undergoes a glass transition to render a disordered solid phase. The figure and explanation below describe in detail the physics of the amorphization.

It takes 0.9 ps for the effects of the impact to arrive at the control volume, and then the stress and the temperature rapidly increase to 9.2 GPa and 1075 K. During this initial compression the material remains in the cubic diamond arrangement, as seen from its constant coordination number of 4. This phase is followed by a sustained increase of the temperature and coordination number, and a peak of the stress. The temperature increases well above the melting point and the coordination number reaches and fluctuates around 6, indicating the occurrence of a solid to liquid transition. At 8 ps the control volume is largely unloaded and the temperature reaches a maximum value of approximately 2080 K, before its decays at a rate well fitted by an exponential law. The point-source nature of the hot area surrounding the impact leads to cooling rates as high as 1.7x10¹² K/s at the normal melting point, and 9x10¹¹ K/s at 900 K, which several studies have identified as the probable glass transition temperature. Thus, the cooling rates during the quench of the liquid phase are some 2 orders of magnitude higher than the value of 7.0x10⁹ K/s needed to prevent the growth of the crystalline phase. The ultrafast quenching confirms the amorphous nature of the final state of the region surrounding the impact, regardless of the ability of the Stillinger-Weber potential for capturing the glass transition.