Interaction of acetone with single wall carbon nanotubes at cryogenic temperatures: A combined temperature programmed desorption and theoretical study

The interaction of acetone with single wall carbon nanotubes (SWCNTs) at low temperatures was studied by a combination of temperature programmed desorption (TPD) and dispersion-augmented density-functional-based tight binding (DFTB-D) theoretical simulations. On the basis of the results of the TPD study and theoretical simulations, the desorption peaks of acetone can be assigned to the following adsorption sites: (i) sites with energy of ∼75 kJ mol ^-1 (T_des ∼300 K)-endohedral sites of small diameter nanotubes (∼7.7 Å); (ii) sites with energy 40-68 kJ mol^-1 (T_des ∼240 K) - acetone adsorption on accessible interstitial, groove sites, and endohedral sites of larger nanotubes (∼14 Å); (iii) sites with energy 25-42 kJ mol^-1 (T_dex ∼140 K) - acetone adsorption on external walls of SWCNTs and multilayer adsorption. Oxidatively purified SWCNTs have limited access to endohedral sites due to the presence of oxygen functionalities. Oxygen functionalities can be removed by annealing to elevated temperature (900 K) opening access to endohedral sites of nanotubes. Nonpurified, as-received SWCNTs are characterized by limited access for acetone to endohedral sites even after annealing to elevated temperatures (900 K). Annealing of both purified and as-produced SWCNTs to high temperatures (1400 K) leads to reduction of access for acetone molecules to endohedral sites of small nanotubes, probably due to defect self-healing and cap formation at the ends of SWCNTs. No chemical interaction between acetone and SWCNTs was detected for low temperature adsorption experiments. Theoretical simulations of acetone adsorption on finite pristine SWCNTs of different diameters suggest a clear relationship of the adsorption energy with tube sidewall curvature. Adsorption of acetone is due to dispersion forces, with its C-O bond either parallel to the surface or O pointing away from it. No significant charge transfer or polarization was found. Carbon black was used to model amorphous carbonaceous impurities present in as-produced SWCNTs. Desorption of acetone from carbon black revealed two peaks at ∼140 and ∼180-230 K, similar to two acetone desorption peaks from SWCNTs. The characteristic feature of acetone desorption from SWCNTs was peak at ∼300 K that was not observed for carbon black. Care should be taken when assigning TPD peaks for molecules desorbing from carbon nanotubes as amorphous carbon can interfere.