Charges enter the ice age

Scattering experiments establish the partly disordered material CsNiCrF6 as the first verified example of a charge ice — and show that it supports Coulomb phases with correlations in three different degrees of freedom.

Frozen water, as common a material as it may be, is a bit of a headache for crystallographers. In ice, the arrangement of the water molecules is not periodic, but it is by no means random either. Ice presents an example of correlated disorder, which blurs crystallographic signatures, but also gives rise to some of the characteristic properties of water ice. In the past decade, a number of magnetic materials — known as spin ices — have been described and characterised in which the arrangement of the spin directions replicates the position of hydrogen atoms in water ice, leading to similarly peculiar behaviours. Writing this week in Nature Physics [1], Tom Fennell from the Laboratory for Neutron Scattering and Imaging at PSI, together with colleagues at the Swiss Light Source, University College London, University of Edinburgh, Institut Laue–Langevin (France) and Oak Ridge National Laboratory (US), reports a study in which they found in crystals of the material CsNiCrF6 ice-like correlations not only for both atoms and spins, but also for electric charges. These findings establish CsNiCrF6 as the first verified example of a charge ice — confirming the prediction of ice-type correlations in ionic systems made more than six decades ago [2].
 

Neutron scattering pattern of the material CsNiCrF6 showing magnetic correlations. Left: experiment; right: theory.
(Image reproduced from [1].)

The material CsNiCrF6 had been identified before as a candidate charge ice, but clear experimental confirmation was missing so far to establish that it indeed realises this state of matter. CsNiCrF6 has a pyrochlore crystal structure — prototypical for spin ices — and contains two ions of different charge, Ni2+ and Cr3+. The presence of these different charges opens up a rich, yet extraordinarily complex playing field for intriguing physics. To tackle this complexity, Fennell and his co-workers performed neutron and x-ray scattering experiments to separate the correlations due to structural, spin and charge degrees of freedom. Working at the Spallation Neutron Source SINQ and the Swiss Light Source SLS at PSI, as well as at the Institut Laue–Langevin in Grenoble, they obtained data that contained tell-tale signatures for a charge ice, arising from the specific arrangement of the ions in the crystal.

Beyond providing the first realization of a charge ice, CsNiCrF6 offers a range of further remarkable properties, as Fennell et al. also report. Associated with the correlations in the charge degree of freedom are structural correlations, in the form of displacement of fluoride ions from their average positions, not dissimilar to what is found in water ice. These structural correlations result in a ‘displacement ice’ phase, inherited from the charge ice. These phases are unusual in that they have no broken global symmetry, but a local gauge symmetry and a so-called closed-loop topology; these are the characteristics that define states of matter known as Coulomb phases. On top of the charge and structural Coulomb phases, the spin arrangement in CsNiCrF6 also gives rise to a magnetic Coulomb phase with antiferromagnetic character.

The finding of this complex interplay of properties hints at great scientific potential — there are many such compounds where the balance and strength of the effects can be modified, and, if the charges involved can be made mobile (not static as in the case of CsNiCrF6), perhaps even technological potential. But thanks to the study of Fennell et al. and other works in the field, it is clear already that correlated disorder in materials is far from being a nuisance, but instead the source of intriguing behaviour.