Several studies have reported carbon nanotube (CNT) membranes that consist of aligned nanotubes sealed in a polymer or inorganic matrix1-3. These membranes had single gas selectivities that were approximately Knudsen, and they had high permeation fluxes for liquid and gas feeds in nanotubes. Because the aligned CNTs grew with a low density (~ 1011 CNTs per cm2 of surface area), only a few percent (0.08 ~ 2.7) of the membrane consisted of CNTs; most of it was the sealing material. Thus, although the fluxes per cm2 of CNT area were orders of magnitude higher than other types of membranes, the fluxes per actual membrane area (CNTs plus polymer or inorganic sealant) were much lower.

We have prepared vertically-aligned CNT membranes with a CNT density of 2.9 x 1012 CNTs/cm2, which is approximately 20 times higher than these previous studies by eliminating the need for a polymer or inorganic filler. Aligned CNTs were grown on a silicon wafer with catalyst thin films 1nm Fe/10 nm Al2O3 from e-beam evaporation, the nanotubes were removed from the silicon surface by in-situ water etching, and the nanotubes were collapsed by to about 5% of their original area by solvent evaporation. The tops of the CNT membrane are expected to be open due to water etching4, 5, and the bottoms are also open because the silicon wafers can be reused for several times for CNT growth after detaching the CNT arrays. This preparation is much simpler than that used for composite membranes, and the membranes have much higher fluxes because of the much higher CNT density and additional interstitial transport pathway between nanotubes. These membranes exhibit light gas selectivities that are equal to or greater than Knudsen selectivities, but their permeances are not independent of pressure. Instead, for most gases the permeances decrease with increasing pressure. The permeance at 1 bar pressure drop for N2 through a membrane that was 750 µm thick was 1.2 x 10-4 mol/m2•s•Pa. This corresponds to a permeability of 27 cc (STP)/ m2•s •atm. Thus, these permeabilities are 222, 391 1 and 330003 times higher than those reported for composite membranes.

Because these membranes do not contain a filler, the spaces between the nanotubes must also be available for transport, but apparently the CNTs are close enough together that the flux through these spaces has similar behavior to the flux through nanotubes. The flux of liquid n-hexane through these membranes was approximately 1,500 kg/m2•h at 1 bar pressure drop, which is 3 to 4 orders of magnitude higher than the flux of n-hexane through MFI zeolite membranes, even though these membranes are thicker (750 µm). n-Hexane flux increased linearly with pressure drop. The nanotubes were approximately 3 nm in diameter, as determined by TEM and calculated from N2 desorption isotherms at 77 K. The average space between nanotubes is appropriately 3 nm with a distribution from 1.4 to 7 nm, as calculated by BJH method from N2 desorption isotherms at 77 K.

1. Hinds, B. J.; Chopra, N.; Rantell, T.; Andrews, R.; Gavalas, V.; Bachas, L. G. Science 2004, 303(5654), 62-65.

2. Holt, J. K.; Park, H. G.; Wang, Y. M.; Stadermann, M.; Artyukhin, A. B.; Grigoropoulos, C. P.; Noy, A.; Bakajin, O. Science 2006, 312(5776), 1034-1037.

3. Kim, S.; Jinschek, J. R.; Chen, H.; Sholl, D. S.; Marand, E. Nano Lett 2007, 7(9), 2806-2811.

4. Ci, L. J.; Manikoth, S. M.; Li, X. S.; Vajtai, R.; Ajayan, P. M. Adv Mater 2007, 19(20), 3300-+.

5. Zhu, L. B.; Xiu, Y. H.; Hess, D. W.; Wong, C. P. Nano Lett 2005, 5(12), 2641-2645.