Suppression of Catalyst Diffusion Into Alumina Support in Dynamic Chemical Vapor Deposition of Carbon Nanotube

Abstract

The utilization of rapid thermal processing capabilities in chemical vapor deposition of carbon nanotubes (CNTs) enables creating dynamic growth recipes engineered to independently control each of the following stages: catalyst nanoparticle preparation by thin film dewetting, and nucleation of CNTs by catalytic growth. While higher temperature pretreatment during the catalyst preparation step enables extending the catalytic lifetime to greatly increase the growth yield by more than three folds, revealing the mechanism of such improvement is crucial. Here, we combine multilayer modeling of ellipsometry results with cross-section scanning transmission electron microscopy (STEM), X-ray diffraction, and scanning electron microscopy (SEM), in order to quantify the evolution of the iron nanocatalyst on alumina thin film stacks as a function of both temperature and time of heating in a reducing environment. Cross section STEM shows the diffusion of iron into the underlying alumina support layer, and our ellipsomettry results confirm that this diffusion is suppressed at high temperature of 900 degrees C, compared to 700 degrees C. Importantly, our measurements support the hypothesis that both the loss of iron catalyst from the surface and the rate of subsurface diffusion are higher at the low annealing temperatures, because of a lower density alumina phase. We utilize SEM to estimate the percentage of iron on the surface after dewetting, which is used in the top-most layer in the stack for ellipsometry modeling. Moreover, we show that an interface layer of both iron and alumina is necessary to include in order to accurately model the multilayer system. By comparing two different sets of samples with 1 nm iron on 10 nm alumina as well as with 10 nm iron on 100 nm alumina, we confirm that these findings can be extended to different ranges of thickness.