Defect Etching in Carbon Nanotube Walls for Porous Carbon Nanoreactors: Implications for CO2 Sorption and the Hydrosilylation of Phenylacetylene

A method of pore fabrication in the walls of carbon nanotubes has been developed, leading to porous nanotubes that have been filled with catalysts and utilized in liquid- and gas-phase reactions. Chromium oxide nanoparticles have been utilized as highly effective etchants of carbon nanotube sidewalls. Tuning the thermal profile and loading of this nanoscale oxidant, both of which influence the localized oxidation of the carbon, have allowed the controlled formation of defects and holes with openings of 40–60 nm, penetrating through several layers of the graphitic carbon nanotube sidewall, resulting in templated nanopore propagation. The porous carbon nanotubes have been demonstrated as catalytic nanoreactors, effectively stabilizing catalytic nanoparticles against agglomeration and modulating the reaction environment around active centers. CO2 sorption on ruthenium nanoparticles (RuNPs) inside nanoreactors led to distinctive surface-bound intermediates (such as carbonate species), compared to RuNPs on amorphous carbon. Introducing pores in nanoreactors modulates the strength of absorption of these intermediates, as they bond more strongly on RuNPs in porous nanoreactors as compared to the nanoreactors without pores. In the liquid-phase hydrosilylation of phenylacetylene, the confinement of Rh4(CO)12 catalyst centers within the porous nanoreactors changes the distribution of the products relative to those observed in the absence of the additional pores. These changes have been attributed to the enhanced local concentration of phenylacetylene and the environment in which the catalytic centers reside within the porous carbon host.


Page S2
(Cr 2 O 3 @C)/GNF is likely due to small chromium nanoparticles that are similar in structure to the Cr 2 O 3 but may possess a higher number of oxygen vacancies.

Figure S2
-XPS analysis of chromium and oxygen of (Cr 2 O 3 @C)/GNF before (top) and after the air treatment resulting in the etching of the carbon surface (bottom). These evidence that chromium is in the 3+ oxidation state throughout this process; the oxidised chromium being simultaneously reduced by the carbon and resulting in localised combustion of the GNF. Scheme 1 -The decomposition of Cr(acac) 3 under an Ar atmosphere results in the formation of carbon. Although gaseous products are formed the as-produced carbon cannot be decomposed in the absence of further oxygen and therefore results in the formation of a Page S3 stabilising, carbon shell around the small chromium oxide nanoparticles (Cr 2 O 3 @C). Scheme produced based on results from J. Phys. Chem.

Figure S3
-Raman spectra of Cr(acac) 3 (blue) and the Cr(acac) 3 treated under the same conditions as the composite formation (red) in the absence of carbon nanotubes. This shows the formation of a carbon shell, most likely covering the small chromium oxide nanoparticles formed during decomposition of Cr(acac) 3 . The carbon observed here is likely to be the same as that observed in (Cr 2 O 3 @C)/GNF, accounting for the weight loss noted in the TGA of the composite material (~330 ᴼC), which acts as a protective layer, stabilising the small nanoparticles against undesirable further growth. Figure S4 -Thermogram (black) and derivative weight with respect to temperature (blue) of (Cr 2 O 3 @C)/GNF with corresponding mass spectroscopy gas analysis (red) for CO 2 (44 m/z). This highlights that the weight loss at 336 ºC is due to the oxidation of carbon, most likely associated with less thermally stable amorphous carbon passivating the metal oxide nanoparticles, rather than the carbon of GNF itself which oxidises at a much higher temperature. Figure S5 -TGA of GNF (red), the composite material of (Cr 2 O 3 @C)/GNF (green) and the final porous carbon nanoreactor after acid treatment (blue). This shows that chromium appears to have a catalytic effect in lowering the GNF combustion temperature and that the pGNF are more thermally stable following its removal. Due to the formation of the holes in the structure, the onset temperature of combustion of the GNF has also decreased and forms a much broader derivative consistent with the level and range of defect formation within the sample.       respectively. However, we cannot really comment on the nature of the defects using these techniques due to the significant number of layers of carbon within the GNF system which makes the concentration of the defects quite small. Spectra were acquired using a 532 nm laser (at 0.3 mW power) and a 600 lines mm −1 rotatable diffraction grating, conferring a spectral resolution of better than 1.8 cm −1 . Table S1 -Experimental condition for pore formation in carbon nanoreactors a) Ramp rates were all performed under air. Thermal treatments for samples A-F are all performed using small batches in a thermogravimetric analyser which allowed for confident control of the parameters. The thermal treatments for G-i represent a scaled-up procedure which was performed in a tube furnace where gas flows and ramp rates were less controllable. b) Once isotherms were completed, samples were cooled in an Ar atmosphere. Page S9  Figure S16 -TPD profiles of CO 2 with 15 mg of pure GNF at β=8 K min -1 , β=6 K min -1 and β=4 K min -1 . As pure GNF do not display any CO2 desorption processes, any desorption process observed in the following must be associated with Ru NP in the bespoke materials.

S2 Porous carbon nanoreactors in gas-phase reactions
flow rate of 5 mL min -1 . An optimal conversion factor is found using a linear regression. The figures below are desorption profiles of Ru@AC, Ru@GNF and Ru@pGNF depicted.    The tables below summarise the activation energies found for the individual desorption processes with their corresponding orders and R 2 values. In bold are values highlighted which are used for analysis.

Ru@AC Ru@GNF Ru@pGNF
Here the highest R 2 values are found   procedure. Small increase in nanoparticles size from 1.95 nm to 2.35 nm for the RuNP@GNF and 1.93 nm to 2.46 nm for the RuNP@pGNF associated to the high temperature conditions used during the desorption process. It is clear that both nanoreactors seem to be able to stabilise the small nanoparticles and prevent unprecedented growth of the nanoparticles.