An unrestricted approach for the accurate calculation of the intermolecular potential of (O2)4: Implications for the solid epsilon phase.

Abstract

Oxygen in its elemental form shows a variety of magnetic properties in its condensed phases; in particular, the epsilon solid phase loses its magnetism. These phenomena reflect the nature of the intermolecular forces present in the solid and the changes that arise with variations in pressure and temperature. In this study, we use intermolecular potentials obtained with unrestricted ab initio methods to model the singlet state of the oxygen tetramer [(O2)4], which is the unit cell, consistent with the non-magnetic character of this phase. To do this, we perform an analysis of the coupled-uncoupled representations of the spin operator together with a pairwise approximation and the Heisenberg Hamiltonian. We start from unrestricted potentials for the dimer calculated at a high level as well as different density functional theory (DFT) functionals and then apply a finite model to predict the properties of the epsilon phase. The results obtained in this way reproduce well the experimental data in the entire pressure range below 60 GPa. Additionally, we show the importance of calculating the singlet state of the tetramer as opposed to previous DFT periodic calculations, where the unrestricted description leads to a mixture of spin states and a poor comparison with the experiment. This point is crucial in the recent discussion about the coexistence of two epsilon phases: one where the identity of each O2 with spin S = 1 is retained within the tetramer unit vs another at higher pressures where the tetramer behaves as a single unit with a closed-shell character.