Novel carbon materials can store and sieve hydrogen

The current prototype H₂ tanks are high-pressure (700bar) bottles and liquid hydrogen cryodevices. Both solutions are awkward for routine mobile applications, prompting an ongoing search for alternative storage materials.^1,2 The most likely possibilities are materials in which the H₂ is either chemisorbed or physisorbed. In chemisorption, H₂ is chemically bound in a host material, such as a metal hydride. In physisorption the gas is bound to the host through weak van der Waals forces. Chemisorption is often hampered by slow loading and/or unloading due to activation barriers for the chemical reactions involved. Devices based on H₂ physisorption have thus far demonstrated insufficient mechanical stability and persistence as well low volumetric storage capacities, although the latter is debated.^2,3

We aim to better understand the guest-host interaction of carbon nanostructures with H₂ and other gases, as such nanostructures are among the most promising physisorption-based, lightweight H₂ storage materials. Our work predicts maximum theoretically possible storage capacities as well as suggests new materials that may be useful as H₂ storage media.

We employ multi-scale simulation techniques to investigate the H₂-host system and its properties. Using an approximate density-functional scheme with van-der-Waals corrections,^4,5 we determine the mechanical and structural properties of carbon nanostructures. Then, we model H₂ interactions with the host using an additive Rutherford potential, with parameters fitted to high-level ab initio calculations (second order M⊘ller-Plesset perturbation theory, coupled-cluster with single and double perturbative triple excitations) on model systems. We solve the stationary Schrödinger equations for the motion of the H₂ center of mass inside the nanostructure on a numerical grid, with periodic boundary conditions applied. We neglect the internal degrees of freedom of the hydrogen molecule and interactions between the molecules. The result is the the energy spectrum ε_i of bound H₂, together with corresponding eigenstates Ψ_i, determined in the ideal gas approximation. The equilibrium constant K_eq for hydrogen adsorption is then calculated as a function of temperature T using standard thermodynamic expressions. For details, we refer to the literature.^6,7 Finally, to estimate storage capacities we partially correct for the neglected H₂–H₂ interactions using the experimental equation of state⁸ when computing the density of H₂ inside the nanostructure at the internal pressure P_int=K_eq,P_ext.

The strength of the H₂–host interaction is proportional to the polarizability of the host material. In a graphene layer, our prototype structure, the H₂–graphene interaction energy is 7kJ mol⁻¹.⁹ Our simulation shows that graphene increases H₂ density by only 60% (K_eq=1.6) at ambient temperature, however: hardly useful for practical applications. In contrast, a graphene bilayer with an optimized interlayer distance exhibits remarkable hydrogen storage potential, having K_eq≥50 at room temperature (see Figure 1). We found, in agreement with recent experiments,¹⁰ that optimal slit pores have a width of ≈0.6–0.7nm.⁶