Modeling Incomplete Conformality during Atomic Layer Deposition in High Aspect Ratio Structures

TL;DR

This paper introduces an advanced model for atomic layer deposition that accounts for incomplete conformality in high aspect ratio structures, integrating gas-phase diffusivity and reaction reversibility, and validated against experimental data.

Contribution

The model uniquely combines Knudsen diffusion, Langmuir kinetics, Bosanquet formula, and reversibility, and is efficiently integrated into topography simulators, advancing the understanding of ALD conformality.

Findings

The model accurately reproduces experimental results for Al₂O₃ ALD.

Temperature dependence of H₂O step yields an activation energy of 0.178 eV.

Inclusion of Bosanquet formula improves TMA step accuracy.

Abstract

Atomic layer deposition allows for precise control over film thickness and conformality. It is a critical enabler of high aspect ratio structures, such as 3D NAND memory, since its self-limiting behavior enables higher conformality than conventional processes. However, as the aspect ratio increases, deviations from complete conformality frequently occur, requiring comprehensive modeling to aid the development of novel technologies. To that end, we present a model for surface coverage during atomic layer deposition where incomplete conformality is present. This model combines existing approaches based on Knudsen diffusion and Langmuir kinetics. Our model expands the state-of-the art by (i) incorporating gas-phase diffusivity through the Bosanquet formula as well as reaction reversibility in the modeling framework first proposed by Yanguas-Gil and Elam, and (ii) being efficiently integrated within level-set topography simulators. The model is manually calibrated to published results of the prototypical atomic layer deposition of Al$_2$O$_3$ from TMA and H$_2$O in lateral high aspect ratio structures. We investigate the temperature dependence of the H$_2$O step, thus extracting an activation energy of $0.178\,\mathrm{eV}$ which is consistent with recent experiments. In the TMA step, we observe increased accuracy from the Bosanquet formula and we reproduce multiple independent experiments with the same parameter set, highlighting that the model parameters effectively capture the reactor conditions.