This handy property mostly arises from distortions within the ferroelectrics’ crystal structure. These, in turn, create tiny electric dipoles, whose polarities, plus and minus, can be reversed by a subtle, locally applied electric field. This way, it is possible to manipulate two distinct polarisation states within the material — similar to the 0 and 1 states of a “bit” in magnetic memory devices. “This means that one can ‘write’ on the material with an electric field and use it to store information,” explains Gattinoni.
However, this adaptable polarisation remains robust only to a certain critical thinness of the material, which is a limiting factor for the miniaturisation of electronic systems. Gattinoni’s work provides insight into why the polarisation is lost and how this could be overcome — specifically, how ferroelectric materials could be engineered to sustain stable polarisation even in very thin nano-scale films.
Pinpointing the trouble in thin films
To that end, Gattinoni and her colleagues performed density functional theory (DFT) calculations using the “Piz Daint” supercomputer at CSCS. The scientists probed very thin films, above and below a thinness of only 7 unit cells, and compared the results to experiments performed by their collaborators at the lab for ferroic materials at ETH on a film of lead titanate (PbTiO3) — a prototypical ferroelectric material — placed on a strontium ruthenate metal waver. The combined results offer insights into the cause of the suppressed polarisation.
Two known phenomena contribute to this polarisation loss. One is the so-called depolarising field, which arises from opposite charges on the opposite surfaces of the material — a natural consequence of its dipole moment. These charges create an electric field that is contrary to the direction of the ferroelectric polarisation. The thinner the film, the closer these opposite charges come to one other, and, consequently, the stronger this destructive depolarising field becomes. The second phenomenon arises from the instability of the surface charges, per se. “In practice, the system has to find a way to compensate the surface charges in order to maintain the ferroelectric properties,” explains Gattinoni. Her DFT calculations indicate a way to achieve this even in thin films.
Gattinoni found that, contrary to previous suppositions, the chemical bonding at the interface between the ferroelectric and the underlying metal — in this case a strontium ruthenate layer — does not have a major influence on maintaining polarisation. Instead, the electronic structure and therefore the electrostatic properties of the ferroelectric and the metal layer play the dominant role in the system. “This means that for controlling the polarisation stability, researchers need to engineer the electrostatic properties, rather than focus on the chemical bonding,” Gattinoni says.
Her calculations also gave indications as to where such modifications would have the highest effect on charge compensation. According to the results, charge transfers at the interface between ferroelectric and metal, which had previously been assumed to be vital, really only play a minor role. In contrast, modifications at the ferroelectrics’ surface proved to be significantly more effective. Such modifications could consist of the introduction of structural defects to create vacancies that can adsorb the charges or of charge-adsorbing molecules in a device’s atmosphere. “Both of these methods can be used to stabilise the polarisation, even in very thin ferroelectric films,” Gattinoni says.
The magical material