The properties of hafnium dioxide (commonly known as hafina) may seem unremarkable on the surface. However, when this material is made into ultra-thin layers, it exhibits fascinating properties: By switching dipoles under the influence of an electric field, such ultra-thin layers can be used as non-volatile computer memory. Furthermore, because the strength of these dipoles is affected by the electric fields they have experienced in the past, they are ideally suited for memristors, which can be used to build "brain-like" computer architectures.

This image shows an artistic impression of the Haffner atomic structure. Image source: Reprinted with permission from Springer-Nature

Beatriz Noheda, professor of functional nanomaterials at the University of Groningen, has studied this material and recently wrote a perspective article on its properties for the magazine Nature-Materials. "Even though we don't understand all the physics, it's already being used in devices," she says.

This is Professor of Functional Nanomaterials at the University of Groningen in the Netherlands and Cognitive Groningen Beatriz Noheda, Scientific Director of the Center for Systems and Materials. She is the lead author of a perspective paper on hafnium oxide ferroelectrics published in Nature Materials. Source: University of Groningen.

To create more efficient computers, fast non-volatile random access memory (RAM) appears to be a good candidate. These materials are composed of cells with dipoles that collectively switch under the action of an electric field. However, if the number of units is too small, their properties break down; spontaneous depolarization occurs below about 90 nanometers. The exception is the

oxygen vacancy, which was discovered more or less by accident, says Beatriz Noheda. Hafner is very stable in high temperatures and harsh environments and has been traditionally used in the metallurgical and chemical engineering industries. However, when amorphous Hafner proved to be a very efficient gate insulator in transistors, it caught the attention of microchip manufacturers. Replacing traditional silicon oxide with Hafner can make transistors smaller. "

Noheda's interest in this material stems from her work at the Center for Cognitive Systems and Materials in Groningen (CogniGron), where she is scientific director. CogniGron's goal is to create neuromorphic computing architectures. Hafnia is one of the materials studied at the centre." In a paper published in Science in 2021, we describe how switching occurs not just through dipoles. We found that the movement of oxygen vacancies also plays a role," said Noheda. Based on her experience, she was invited to discuss lessons learned from Hafnia in a perspective article in Nature Materials.

Hafner behaves like a ferroelectric, but it only maintains its properties at the nanometer scale. "Ferroelectrics seemed to have dropped out of the race for ultra-small non-volatile RAM, but with hafnia they are now in the lead." Nonetheless, Hafner does not appear to behave exactly like a ferroelectric, and as mentioned earlier, the movement of oxygen vacancies appears to be critical to its properties.

Noheda also pointed out another concept to consider: the surface energy of the nanoparticles. "Phase diagrams show that the relatively large surface area of these particles creates what amounts to extremely high pressures in hafnium dioxide, which appears to play a role in the material's properties. This type of knowledge is important in finding other materials that behave similarly to hafnium. Because the global supply is too small, hafnium is not a microchip producer. The most sustainable option in manufacturing. By looking for materials with similar properties, we may find better candidates. "

Finding sustainable alternatives to hafnium could accelerate the use of ferroelectric materials in RAM memory. Since the strength of a dipole depends on the history of the electric field that created it, it would be an ideal material for producing memristors. Such simulated devices behave similarly to the neurons in our brains and are candidates for neuromorphic computer architectures. "We are working hard to develop this neuromorphic chip. But first, we must fully understand the physical properties of hafnium dioxide and similar materials."