Computational Modeling of Physical Surface Reactions of Precursors in Atomic Layer Deposition by Monte Carlo Simulations on a Home Desktop Computer

Abstract

The continuously increasing demand for miniaturized devices in the semiconductor industry has increased the need for ultrathin films. Atomic layer deposition (ALD) is the most favorable technique for this purpose and has attracted significant interest. Prior to experimentation, understanding the reaction mechanism of the precursor with the substrate surface is crucial. However, research on the growth mechanism of ALD is limited as compared with research on its process development. Currently, ALD reaction mechanisms are typically studied using computational methods, such as molecular dynamics. However, they are complex and focus predominantly on chemical reactions between molecules and surfaces. In this study, the ALD reaction mechanisms are investigated using Monte Carlo (MC) simulations based on a simple steric hindrance model. The physical surface reactions of the precursors are modeled using MC simulations. Therefore, the steric hindrance effect of precursor adsorption on surfaces is easily predicted using a home desktop computer without requiring significant computing resources. The proposed MC simulation models yield highly consistent results with experimental data and theoretical results obtained from density functional theory calculations. We believe that this simulation method can be a useful tool with a laptop-scale computer for researchers and students working on understanding surface reactions.

Atomic Layer Deposition of Zinc Oxide on Mesoporous Zirconia Using Zinc(II) Acetylacetonate and Air

“…The amount of metal obtained by ALD can be defined by steric hindrance of bulky ligands. ,,,, Puurunen demonstrated via a ball model that the GPC (nm –2 ) decreases with increasing ligand size, and the GPC is typically less than 50% of a monolayer because of steric hindrance of ligands. Gu et al showed the effect of bulky reactants on areal coverage (%) by Monte Carlo simulations using various reactants such as trimethyl­aluminum (TMA) and dimethyl­aluminum isopropoxide (DMAI). The areal coverage of DMAI was lower than that of TMA because DMAI has a larger size.…”

Organosulfide Inhibitor Instigated Passivation of Multiple Substrates for Area-Selective Atomic Layer Deposition of HfO2

“…During DES exposure, a dense layer of DES molecules was exposed to the Cu and SiO 2 substrates modeled in a 3D plane. The simulation algorithm as previously proposed by our research group comprised of two steps: (a) SMI settlement on adsorption points on the model surface and (b) TDMAH precursor rearrangement and settlement on vacant adsorption points . The areal and point densities in the MC were calculated using the following equations: normalareal .25em normalcoverage = N adsorbed × s molecule s substrate × 100 % normalpoint .25em normalcoverage = N adsorbed N points × 100 % where s molecule is the molecular size of the top view, s substrate is the substrate size, N adsorbed is the total number of molecules adsorbed, and N points is the number of adsorption points.…”

Area-Selective Deposition of Ruthenium Using Homometallic Precursor Inhibitor

“…The MC simulations were performed using a modified algorithm ,, that ran on the “R” simulation program (version 3.6.1; GUI 1.70; EL Capitan build 7684). We demonstrated the adsorption behavior of Ru­(EtCp) 2 on the Ru substrate surface through areal coverage in an ideal case.…”