Accurate and inexpensive computational tools are indispensible for screening and predicting new sorbent materials for hydrogen storage. While Density Functional Theory (DFT) is moderately inexpensive, its accuracy strongly depends on the form of the specific Density Functional Approximation used. In this work, we constructed a database consisting of binding moieties -- including bare ions, organic linkers, and open metal sites -- interacting with hydrogen. The binding moieties in our database comprise of a variety of binding mechanisms, interaction strengths, and broad chemical diversity. The moieties model the immediate vicinity of the bound hydrogen in sorbent materials, thus accurately representing the hydrogen storage problem. The density functionals were assessed on two criteria: (i) errors in binding energies at and away from equilibrium and (ii) errors in vibrational frequencies. In general, the errors associated with density functionals decrease with the cost of computation, but we have found notable exceptions to this rule at both ends of the spectrum. In absence of a permanent dipole, hydrogen is sometimes bound only with dispersion. Inclusion of empirical or non-empirical dispersion corrections, in general, increases the accuracy of density functionals. We found that the error caused by using a smaller basis set is much larger than the error caused by compromising on DFT quadrature grid. While the density functional wB97M-V gives the best performance, the significantly less expensive functional B97-D3(BJ) provides the best compromise between accuracy and cost. Modern density functionals perform reasonably well for the hydrogen binding problem and are able to correctly capture various interactions between hydrogen and its binding moiety. The functionals suggested in this study can be used to screen and/or design binding sites that can be incorporated into sorbents. These functionals can also be reliably used to parametrize force fields, which can enable the study of dynamical aspects of hydrogen storage and help compute usable capacities.

Assessment of the performance of density functionals for hydrogen storage in sorbents

2019 Spring National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.