A high-throughput methodology comprising of an automated micro-chemical vapour deposition (CVD) reactor and statistical design of experiments were used for screening and the first pass optimisation of the single- (SWCNT) and double-walled carbon nanotube (DWCNT) parameter space using an alumina supported iron-molybdenum catalyst.The parameters of reaction temperature, metal loading, bi-metallic ratio, total gas flow, and gas flow ratio were initially investigated using an L18 experimental design to identify the most significant reaction parameters. Product characterisation metrics included carbon yield (determined from thermogravimetric analysis), IG/ID (ratio of G band to D band from Raman spectroscopy), and the presence of SWCNTs/DWCNTs from radial breathing modes (Raman spectroscopy) and transmission electron microscopy (TEM).A good model fit for Reaction Carbon Yield and IG/ID was obtained for the L18 experimental design (R2–Q2<0.2–0.3, model validity>0.25, reproducibility>0.5), and temperature and metal loading were found to be the most influential variables. Temperature and metal loading were consequently optimised using a D-Optimal experimental matrix to maximise both output metrics. The optimisation generated model predicted a metal loading of 4.9wt% and a synthesis temperature of 900°C to result in a carbon yield of 11.4wt% and an IG/ID of 8.0, in excellent agreement with experimental data of 11.8wt% and an IG/ID of 8.1–8.7.This study confirms that the efficient small-scale optimisation protocol is a valid alternative to resource intensive large-scale experiential screening and offers a route towards completely mapping the CNT growth space. However, results from the high-throughput CVD reactor are yet to be transferred with high accuracy to a larger scale fluidized bed apparatus (52mm ID), and further research is required to scale-up the process.
Read full abstract