Like a pesky cowlick that can’t be tamed no matter how much you threaten it with a comb, tiny whorls of magnetic moments (spins), known as skyrmions and found in magnetic materials, can be extremely persistent, thanks to their specific topology1. And, just like hairdos, skyrmions and their various relatives come in many shapes and sizes, and with a mishmash of unusual names, such as hedgehogs, anti-hedgehogs and skyrmion bubbles. Skyrmions have been thoroughly studied since their experimental observation a decade ago2,3 and promise denser and faster magnetic data-storage devices, but their electrical analogues have been elusive. Now, in a combined experimental and theoretical study in Nature, Das et al.4 demonstrate that ordered arrays of polar-skyrmion bubbles — electrical cousins of magnetic-skyrmion bubbles — can be stabilized in artificially layered oxide materials.
Despite their fundamentally different physical origins, materials called ferroelectrics and their eponymous magnetic counterparts, ferromagnets, have many similarities. The defining property of a ferroelectric (or ferromagnet) is a spontaneous electrical polarization (or magnetization) that can be reversed by the application of an electric (or magnetic) field. This attribute makes both of these materials extremely useful for data storage, as well as for a multitude of other applications. However, unlike spins in magnets, which can often rotate with relative ease to give complex, swirling patterns, electric dipoles that arise from the relative displacements of positive and negative ions in a crystal cause a deformation of the crystal lattice, and must pay a hefty price in elastic energy to bend outside the ordered ranks.
Nevertheless, this difference hasn’t stopped researchers from looking for patterns of rotating polarization in ferroelectrics, and one way to bend the rules is to go small5–8. When a ferroelectric is confined to the nanometre scale, it can be subject to large internal electric fields and stresses. These can strongly perturb the local polarization orientation and produce highly non-uniform distributions of electric dipoles, especially near surfaces or interfaces and domain walls, which are boundaries separating regions (domains) that have uniform polarization orientation9–11 (Fig. 1a). Such nanometre-scale confinement is the first ingredient for Das and colleagues’ polar-skyrmion bubbles.
The second ingredient comes courtesy of a decade of breakthroughs in our understanding of the structure of ferroelectric domain walls12,13. It turns out that these walls can harbour polarization components perpendicular to those in adjacent domains, reminiscent of boundaries called Néel and Bloch walls in ferromagnets. Therefore, if a ferroelectric domain wall is looped, it can form a ring of polarization14 (Fig. 1b). The direction of rotation of this polarization imparts a handedness (chirality) to the overall pattern of dipoles, making it distinct from its mirror image. Together, the rotating polarization across the domain wall and the ring of polarization within it produces the pattern of electric dipoles observed by Das et al. (Fig. 1c), and which has the same topology as a magnetic-skyrmion bubble.
To obtain such polar-skyrmion bubbles, Das and colleagues used artificially layered crystals called superlattices, which form ordered nanometre-scale domains. These crystals consist of alternating layers of ferroelectric and non-ferroelectric oxides, each just a few nanometres thick. To image the resulting polarization pattern, the authors used state-of-the-art electron microscopy that could resolve individual atomic displacements and produce stunning pictures of the local arrangements of electric dipoles.
A top-down view of the superlattices reveals a relatively ordered array of nanometre-scale bubbles, with in-plane polarization components converging towards the bubbles’ north poles. Cross-sectional images beautifully resolve a gradual rotation from ‘up’ polarization to ‘down’ polarization across each bubble. To complete the picture, Das et al. used a technique known as four-dimensional scanning transmission electron microscopy to probe the swirling in-plane dipoles around the equator of a bubble that determine the bubble’s chirality. The observed polarization structure is in remarkable agreement with atomic-scale simulations that reveal further details of the 3D dipole pattern.
Analogy with magnetic skyrmions is not without its subtleties. A key feature of magnetic skyrmions is their chirality, which determines their properties, including their stability and direction of motion under applied forces. This chirality stems from specific interactions that dictate whether neighbouring spins rotate in a clockwise or anticlockwise manner. Such interactions, however, have no electric counterpart.
Surprisingly, X-ray diffraction measurements by Das et al. reveal that the ordered polar-skyrmion bubbles exhibit macroscopic chirality. This finding poses questions about the origin of this unexpected handedness, whether it could be reversed using an applied electric field15 and whether it affects the stability of the bubbles. Crucially, unlike spins, electric dipoles can grow and shrink in magnitude, or even disappear entirely. This property might have other implications for the stability of polar-skyrmion bubbles; these require further investigation7,14.
The observed chiral patterns should be present in the many similar structures that have nanometre-scale domains, raising questions about the possible role of these patterns in the overall behaviour of previously investigated systems and whether their properties could be harnessed to increase functionality. Does the chiral nature of the patterns lead to any useful optical properties? Do polar-skyrmion bubbles have higher mobilities than those of stripe-like ferroelectric domains, and could they enhance the sought-after ‘negative capacitance’ behaviour observed in similar superlattices16 that might help to reduce the power consumption of transistors? Perhaps they even have other unexpected properties, such as conductivity or magnetism, that are analogous to, or entirely different from, those discovered in ferroelectric domain walls17.
The polar-skyrmion bubbles discovered by Das et al. necessarily form an ordered lattice to minimize the system’s electrostatic energy. However, any future devices akin to the technology known as magnetic racetrack memory18 will require the stabilization of individual polar skyrmions, as well as precise control over their injection and motion in applied fields16. Work in this direction will undoubtedly be a priority in this field. Irrespective of whether or not such polar-skyrmion bubbles ultimately translate into new technologies, Das and colleagues’ work will kindle further excitement in the emergent study of topological ferroelectrics.
Nature 568, 322-323 (2019)