Model calculations are presented which predict whether or not an arbitrary gas experiences significant absorption within carbon nanotubes and/or bundles of nanotubes. The potentials used in these calculations assume a conventional form, based on a sum of two-body interactions with individual carbon atoms; the latter employ energy and distance parameters which are derived from empirical combining rules. The results confirm intuitive expectation that small atoms and molecules are absorbed within both the interstitial channels and the tubes, while large atoms and molecules are absorbed almost exclusively within the tubes.Comment: 9 pages, 12 figures, submitted to PRB Newer version (8MAR2K). There was an error in the old one (23JAN2K). Please download thi

An overview is presented of the various phases predicted to occur when gases are absorbed within a bundle of carbon nanotubes. The behavior may be characterized by an effective dimensionality, which depends on the species and the temperature. Small molecules are strongly attracted to the interstitial channels between tubes. There, they undergo transitions between ordered and disordered quasi-one dimensional (1D) phases. Both small and large molecules display 1D and/or 2D phase behavior when adsorbed within the nanotubes, depending on the species and thermodynamic conditions. Finally, molecules adsorbed on the external surface of the bundle exhibit 1D behavior (striped phases), which crosses over to 2D behavior (monolayer film) and eventually 3D behavior (thick film) as the coverage is increased. The various phases exhibit a wide variety of thermal and other properties that we discuss here.

No abstract

Helium atoms are strongly attracted to the interstitial channels within a bundle of carbon nanotubes. The strong corrugation of the axial potential within a channel can produce a lattice gas system where the weak mutual attraction between atoms in neighboring channels of a bundle induces condensation into a remarkably anisotropic phase with very low binding energy. We estimate the binding energy and critical temperature for 4 He in this novel quasi-onedimensional condensed state. At low temperatures, the specific heat of the adsorbate phase (fewer than 2% of the total number of atoms) greatly exceeds that of the host material.Low temperature research on Helium was initially stimulated by the challenge of determining the condensation temperature of bulk He [1]. In recent decades, two-dimensional He films, in which the superfluid transition differs qualitatively from that of the bulk [2], have been particularly intriguing. While once of only academic interest [3][4][5][6], Helium in one-dimensional or quasi-one-dimensional systems has received increased attention recently since the realization that such systems can be created in the laboratory. He atoms are very strongly bound within the hexagonal lattice of narrow interstitial channels between tubes within the triangular lattice of a bundle of carbon nanotubes [7][8][9][10]. Within this very narrow channel, the transverse degrees of freedom are frozen out even at relatively high temperatures of ∼50 K. The binding energy per atom, 340 K, is the highest known for He, almost twice that calculated for He within individual nanotubes [11] and 2.4 times higher than that on the basal plane of graphite [12,13]. It exceeds the ground state binding energy of bulk liquid 4 He by nearly fifty times [14].Here we describe how the strong axial confinement of the He wavefunctions within a single channel can produce a direct experimental realization of a lattice gas model, wherein the weak coupling between atoms in neighboring channels induces a finitetemperature transition into a remarkably anisotropic and extremely weakly bound condensed state. First we present a localized model wherein the Helium resides in periodic array of relatively deep potential wells; this "bumpy channel" approximation is supported by single-particle Helium band structure calculations. For comparison, we also describe a delocalized model which assumes translational invariance within each channel (a "smooth channel" approximation). The large difference between the models in 1

No abstract