In oxide materials such as the transition metal perovskites (which have the formula ABO3, where A and B are metals), atomic bonds are usually ionic, explains team leader Manuel Bibes of CNRS/Thales. This means that electrons sit either on the metal atoms or on the oxygen ions. However, in some perovskites, covalent bonding is favoured. Here, electrons are shared between the metal atoms and the oxygen ions.

In reality, things lie in between these two extremes and fine tuning between ionic and covalent bonding plays an important role when it comes to the emergence of high-temperature superconductivity in copper oxides (cuprates), for example, he says. “In the on-going quest for novel high-Tc superconductors built from oxide multilayers, characterizing and tuning the level of covalence in the different oxide layers may therefore be just as important as controlling (through doping, for instance) the overall number of electrons leaking from one layer to another.”

Trying to keep covalence levels steady

In this context, the main finding of our new study, published in Nature Physics doi:10.1038/nphys3627, is that at the interface between two strongly correlated oxides, electrons can be transferred in an amount that is regulated by the local level of covalence, or electron sharing,” he tells “Added electrons tend to disturb covalent bonds, which costs energy, so the material tries to keep the covalence level steady.”

The researchers obtained their results by looking at how electrons distribute at the interface between a titanium perovskite (GdTiO3) and a rare-earth nickel perovskite (RNiO3, where R = La, Nd or Sm) using a synchrotron technique called X-ray absorption spectroscopy. “We tuned the energy of the X-rays we applied to our samples and deduced the number of electrons on the different ions from the shape of the spectra produced,” explains Bibes. “We also performed first-principles calculations to estimate the amount of charge transferred between the two materials.”

Discovering high-Tc superconductors

While our interfaces are not superconducting, they do develop a novel ferromagnetic-like state, which does not exist in bulk nickelates, and whose properties are tuned by the covalence level,” he adds. “We now need to do more experiments to better understand the mechanisms behind the appearance of this interfacial magnetism.”

The findings could help guide future research on strongly correlated oxide heterostructures and develop novel strategies to engineer two-dimensional states in these materials through both doping and by adjusting the level of covalence, he says. Ultimately, we hope the work will help us discover high-Tc superconductors – a holy grail in this field.