AmericanChemicalSociety.com

This research project—a new direction for the PI into the areas of nanomaterials and hydrogen storage—seeks to elucidate the fundamental dynamical interactions between hydrogen and carbon-based nanomaterials, which have not been previously well understood.  Certainly, traditional chemical intuition, in terms of local covalent bonding models and the like, is found to be lacking in this context, vis-à-vis providing a quantitative description of the electronic structure. Moreover, classical trajectory simulations (CTS), commonly used to model such systems, fail to capture quantum dynamical effects associated with the nuclear motion, which are sure to be important, given that hydrogen is the lightest and “most quantum” atom.  Perhaps even more important though, a new paradigm is needed for describing and understanding the essential physics of such systems, which requires new ways of framing the problem.  This research project addresses all of these limitations of conventional treatments, and has already shed important insights and mechanistic understanding that could well be of substantial benefit to material scientists and engineers working on hydrogen storage and related applications. 

Specifically, we have performed accurate first-principles theoretical investigation of hydrogen atom adsorbates, bound exohedrally to a single-walled carbon nanotubes (SWNT).  Molecular hydrogen adsorbates are not considered, as direct H₂ adsorption is no longer regarded a viable mechanism for hydrogen storage via SWNTs.  Atomic hydrogen adsorbates, on the other hand, are bound much more strongly, and are essentially incapable of direct desorption at ambient temperatures, thus preventing against storage loss.  However, atomic loading requires catalytic dissociation of free hydrogen, via the “spillover” mechanism.  Catalysis per se is not the focus of this investigation. Instead, our interest is understanding what happens to the H atom adsorbates once they are loaded onto the tube, e.g. by what mechanism are they able to migrate downwards, thus making room for additional adsorbates to be catalytically loaded as the adsorption site.  Experimentally, the overall spillover process has several “mysterious” features, still not well understood.

In particular, this project seeks to understand or shed some light on the following:

1. The per-atom binding energy for free H₂ is around 2.4 eV, whereas  for a single H atom adsorbate on the (5,5) SWNT, it is only ~1 eV.  How is this large energy gap overcome?
2. Classical diffusion models of the migration process predict rates that are orders of magnitude lower than experimental rates, at low pressure (P) and temperature (T). Why?
3. Experimental spillover is essentially reversible, implying that all of the many competing pathways (such as adsorbate collision-induced recombination of H2) that are in principle available, are in practice not accessible. Why?

Already, the research project has provided insight into all three questions above, as described in summary below, and in detail in two publications [J. L. McAfee and B. Poirier, J. Chem. Phys. (in press) and  J. L. McAfee and B. Poirier, J. Chem. Phys. 130, 064701 (2009)],  one of which (the latter) has also been selected for joint publication in the Virtual Journal of Nanoscale Science and Technology [19 (2009)]. Comprehensive, first-principles electronic structure and quantum dynamics calculations were performed for the one-adsorbate, two-adsorbate, and full-coverage-limit (one H atom adsorbate per substrate C atom) systems. In all cases, the (5,5) SWNT was used, as a standard benchmark, and also a good compromise between stability and curvature.  These calculations have enabled us to obtain an accurate and detailed dynamical characterization of all of the relevant mechanisms involved. Main conclusionst are highlighted in the summary below:

1. Curvature and corrugation effects matter greatly, with substantial potential barriers existing between neighboring C-atom wells.
2. At low T and P, migration is dominated by quantum tunneling, rather than diffusion.
3. Adsorbates migrate down the SWNT cylinder, rather than around.
4. Adsorbates can not get sufficiently close to pass each other, transfer energy, recombine into molecular hydrogen, or multiply bind at the same substrate C atom.
5. The full coverage limit gives rise to a global “multiple binding enhancement” effect, such that the per-adsorbate binding energy increases from ~1 eV to ~2.5 eV.
6. The dominant full coverage migration pathway occurs via hole defects, which only allow a single adsorbate to migrate at a time.

The work above was primarily performed by Jason McAfee, a recently graduated Ph.D. student, who gave oral presentations at two regional meetings, one of which led directly to a postdoctoral researcher position at U North Texas.  He was assisted by (and helped to train) two other individuals, undergraduate student Karl Gillenwater, and high school chemistry teacher Amber Allen, who joined the PI research group as part of a special Summer Academy run by the computational chemistry groups at TTU.  The PI himself has given three non-TTU seminars on this work, one at the spring 2009 national ACS meeting in Salt Lake City, one an invited seminar at the Nanoscience Center at U South Carolina, and the third as a special invited participant (all expenses paid) at the Workshop on Quantum Atomic and Molecular Tunneling in Solids and Other Condensed Phases in Darmstadt, Germany. The latter is likely to lead to several important collaborations.