Antiferromagnetic ordering and possible lattice response to dynamic uranium valence in U3O8

Determining the correct electronic structure of ${\mathrm{U}}₃{\mathrm{O}}₈$ remains a formidable experimental and theoretical challenge. In the low-temperature phase, two crystallographic U sites are separated into a distinct 2U(V)$+1\mathrm{U}$(VI) oxidation configuration. At low temperatures, the U(V) sites form a distorted honeycomb lattice, but the U(VI) sit on a triangular sublattice, suggesting potential for magnetic frustration effects. The spin configuration of the unpaired $f$ electrons on the U(V) sites is likely antiferromagnetic (AFM) from susceptibility measurements, but this has not been confirmed. Here, we present a neutron scattering investigation of the structure and dynamics of ${\mathrm{U}}₃{\mathrm{O}}₈$ from 1.7 to 600 K. We confirm static AFM ordering onset at between 22 and 25 K, which is present down to at least 1.7 K with AFM peaks corresponding to [0.5 1 1] and [0.5 2 2] in the orthorhombic phase. These measurements rule out static AFM order along the $a$ axis of the $Amm2$ phase, a configuration previously suggested by theory. Above 100 K a quasielastic scattering channel opens that we speculate arises from a lattice relaxation response to thermally activated electron hopping. This term does not conform to a magnetic form factor, so it is not related to spin relaxations. If correct, this mechanism stabilizes a continuous valence transition from $2\mathrm{U}(\mathrm{V})+1\mathrm{U}(\mathrm{VI})$ in the low-temperature ($T<600$ K) orthorhombic phase to the hexagonal phase that contains only one degenerate U site, wherein the U valence can be dynamically stabilized between $\mathrm{U}(\mathrm{V})\ensuremath{\leftrightarrow}\mathrm{U}(\mathrm{VI})$ by phonon-assisted electron hopping.