Order-of-Magnitude Increase in Carbon Nanotube Yield Based on Modeling Transient Diffusion and Outgassing of Water from Reactor Walls

Abstract While reactor wall preconditioning was previously shown to influence the yield in chemical vapor deposition (CVD), especially for coatings of carbon nanotubes (CNTs), it was limited to studying accumulating deposits over a number of runs. However, the effects of temperature and duration as the reactor walls are exposed to hot humidity for an extended period of time between growth runs was not previously studied systematically. Here, we combine experimental measurements with a mathematical model to elucidate how thermochemical history of reactor walls impacts growth yield of vertically aligned CNT films. Importantly, we demonstrate one-order-of-magnitude higher CNT yield, by increasing the interim, i.e., the time between runs. We explain the results based on previously unexplored process sensitivity to trace amounts of oxygen-containing species in the reactor. In particular, we model the effect of small amounts of water vapor desorbing from reactor walls during growth. Our results reveal the outgassing dynamics, and show the underlying mechanism of generating growth promoting molecules. By installing a humidity sensor in our custom-designed multizone rapid thermal CVD reactor, we are able to uniquely correlate the amount of moisture within the reactor to real-time measurements of growth kinetics, as well as ex situ characterization of CNT alignment and atomic defects. Our findings enable a scientifically grounded approach toward both boosting growth yield and improving its consistency by reducing run-to-run variations. Accordingly, engineered dynamics recipes can be envisioned to leverage this effect for improving manufacturing process scalability and robustness.