Surface kinetics and feature scale particle model of Si x N y atomic layer deposition using Si2Cl6 precursor

Abstract

One of the important steps in the fabrication of semiconductor memory devices is the deposition of ultrathin silicon nitride films with uniform film thickness and electrical properties. Such high-quality films have made atomic level control during deposition a necessity and can be achieved via atomic layer deposition (ALD) with excellent step coverage. While ALD has been studied experimentally by many authors, there exist significant gaps between their observations and the practical application of the ALD process in large-scale manufacturing. In this work, a computational model of thin film deposition for a silicon based ALD application was developed. The model includes a surface chemistry mechanism for the deposition of hexachlorodisilane (Si2Cl6) on a growing SixNy film. This mechanism quantifies the sticking probability of the Si2Cl6 precursor on the growth surface as well as an effective active reaction site density on these surfaces. This surface reaction chemistry was used in the context of a feature scale particle transport model to simulate ALD in 100 nm–1 μm critical dimension, ∼10–100 high aspect ratio holes. The model demonstrates the effects of hole size and aspect ratio dependence on the overall kinetics of the deposition process. An increase in the completion time for ALD processes with increasing hole aspect ratio and the increasing statistical nature of the deposition process with smaller critical dimension of the hole were predicted.