Control of growth morphology of deposited fcc metals through tuning substrate-metal interactions

TL;DR

This study uses kinetic Monte Carlo simulations to demonstrate how tuning substrate-metal interactions can control the growth morphology of fcc metal thin films, influencing properties like roughness and island formation.

Contribution

It introduces a simulation framework for screening and controlling thin film growth modes by varying substrate-metal interaction strengths and thermal annealing conditions.

Findings

Stronger substrate interactions promote layer-by-layer growth and reduce surface roughness.

Thermal vacuum annealing enhances flatness and reduces defect density in deposited metals.

Au, Pd, and Pt are highly sensitive to substrate interaction variations, affecting morphology significantly.

Abstract

Precise control over thin film morphology is critical for optimizing material properties across diverse technological applications, as the growth mode (whether 2D layer-by-layer or 3D island formation)determines key functional properties such as electrical conductivity in CMOS interconnect applications and catalytic activity, where island distribution and size dictate performance. To explore the role of the substrate on the morphology of deposited metals, we present extensive kinetic Monte Carlo simulations on six fcc metals growing in the (111) direction: Ag, Au, Cu, Ni, Pd and Pt. Our simulation framework enables screening and evaluation of their growth mode under homoepitaxial growth scenarios and proposes morphology control strategies by variation of substrate-metal interaction strengths, modeled by modifying the activation energies for upward and downward migration, combined with thermal vacuum annealing within typical back end of line (BEOL) integration thermal budget. Our simulation results demonstrate that modulation of the substrate interaction strength can be effectively employed to promote island formation or layer-by-layer growth modes overcoming limitations in achieving large flat surface areas. Au, Pd and Pt exhibit the highest sensitivity to substrate interaction strength variations, followed by Ag, showing that strongly interacting substrates decrease the root mean square (RMS) roughness, (uncovered) substrate exposure, island number and island aspect ratios, with moderate increases in flat surface areas and atomic coordination numbers. Additionally, interconnect relevant metrics are improved through thermal vacuum annealing particularly when sufficiently strong metal-substrate interactions are employed, reducing surface roughness, achieving larger flat surface areas, merging and smoothing islands, and decreasing defect density...