Gas Adsorption on Heterogeneous Single-Walled Carbon Nanotube Bundles

Optimization of carbon nanotube bundles containing a distribution of nanotube diameters always gives structures with packing defects that form relatively large interstitial channels. Experimental data for ${\mathrm{C}\mathrm{H}}_4$, Ar, and Xe adsorption are compared with simulations. Low coverage experimental isosteric heats are in excellent agreement with simulations of gases adsorbing into interstitial channels of defective nanotube bundles, whereas adsorption onto perfect bundles does not agree with experiments. Thus, an accurate description of adsorption on nanotube bundles must account for interstitial adsorption.

Figure 1

Sample heterogeneous (left panel) and homogeneous (right panel) bundles optimized by the basin-hopping technique [20, 21, 22]. The red spheres represent ${\mathrm{C}\mathrm{H}}_4$ adsorbed in equilibrium with a bulk phase at 159.88 K and 0.05 bar.

Figure 2

Isosteric heats of adsorption for ${\mathrm{C}\mathrm{H}}_4$ from experiments [5, 7] (circles) and simulations. The diamonds (squares) are for adsorption onto heterogeneous (homogeneous) bundles. The triangles are for a homogeneous bundle with the solid-fluid potential increased by 45%. The inset shows the low coverage region. The lines are drawn as a guide to the eye. The estimated error bars for simulations are about the size of the symbols.

Figure 3

Experimental and simulated $q_{\mathrm{s}\mathrm{t}}$ for Ar on SWNT bundles. Circles and right triangles are experimental data from Wilson et al. [9] and Talapatra et al. [12], respectively. The diamonds (squares) are simulation data for adsorption onto heterogeneous (homogeneous) bundles. The inset shows the low coverage region.

Figure 4

Xenon $q_{\mathrm{s}\mathrm{t}}$ from experiments [8] (circles) and simulations on homogeneous (squares) and heterogeneous (diamonds) bundles.