Invited
Antiferroelectricity in perovskites: is this a physical property ?

When looking for the definition of a physical property, one usually encounters a series of examples rather a definition. A property, philosophy tells us, is an entity that can be predicated of (attributed to) "things". These things instantiate (exemplify) their properties. Properties are introduced to account for or explain phenomena, explain qualitative resemblance and recurrence, and are used to recognize a new occurrence.

In the case of antiferroelectricity, the question of its nature as a physical property is whether it is possible to define what phenomenon is unique to antiferroelectrics, to sort out antiferroelectrics from other polar materials and to determine whether a novel material is antiferroelectric or not.

From the general property point-of-view, the set of things that instantiate a property corresponds to the extension of this property. The definition of a property may emerge from what the members of its set have in common, and that is not defined by another property. The balance between comprehensiveness and simplicity determines the extent and broadness of the definition.

To do so, the identifying conditions for properties is that they are identical only if they confer the same "causal powers" on their instances. For antiferroelectricity, the usually assumed causal powers revolve around the symmetry of the antiferroelectric phase (antiparallel arrangement of dipoles having a net zero sum), its reaction to an applied electric field (with a maximum of the permittivity at the phase transition under low field and double hysteresis loops under high fields), or the occurrence of a phase transition toward a centrosymmetric phase that can be transformed into a ferroelectric one under the application of an electric field.

Additional characteristics include a negative electro-caloric effect when the material undergoes an AFE-to-FE phase transition, a smaller volume of the antiferroelectric phase compared to its parent phase, a stabilization under pressure... We shall examine which one(s) are key to defining AFE.

However, there are several difficulties in defining unambiguously antiferroelectricity in perovskites: the nature of the dielectric dipoles in crystals complicates the definition of their antiparallel arrangements compared to liquid crystals or ice, for example. We shall show that a purely structural definition of AFE cannot help distinguish between AFE and anti-polar materials. The behavior under an electric field alone is also not enough: neither the increase in the permittivity at the phase transition nor a double polarisation or strain hysteresis loop is specific to AFE.

We shall, therefore, first review the definitions and phenomena associated with antiferroelectric perovskites, and the question they raise, before examining to which extent the concept of antiferroelectricity in perovskites is of any use or relevance.