We can only imagine how intrigued and tantalized ancient observers became when they discovered the properties of magnetite (Fe3O4) more than 2,000 years ago. This material has since become a common source of iron ore, but it retains its capacity for surprise, as a group of researchers in Austria found after conducting a detailed scan of its surface structure.

“There was a well understood model of this structure out there,” says Eamon McDermott, a Canadian materials chemistry PhD student who took part in this research at the Vienna University of Technology. “The problem was that it conflicted with some of the experimental evidence that our collaborators were collecting,” says McDermott, who studies condensed matter systems using density functional theory calculations. Subsequent analysis with a combination of electron diffraction, scanning tunnelling microscopy and robust theoretical calculations revealed that magnetite’s surface does not resemble the tidily packed lattice depicted in most textbooks. Instead, the atomic arrangement consisted of an unexpected landscape with irregular displacements and interstitial openings. 

McDermott co-authored a Science paper on the investigation, which concluded that magnetite retains its stable structure in spite of these irregularities thanks to a mechanism that redistributes cations in response to the surrounding chemical environment. This observation sheds new light on the single-atom action that drives processes such as heterogeneous catalysis, hydrogen production and drug delivery, where iron oxides already have an important role. “You can even get single gold, palladium or platinum adsorption on magnetite, when normally these metals clump or cluster on surfaces,” McDermott says.

An accompanying commentary cites this work as an acknowledgement that there could be many more metal oxides with similarly complex surfaces and interfaces. As for why it took so long to tease out the true nature of magnetite, McDermott notes that experimental methods had to get ahead of what had been a satisfying chemical description. “Over the past 10 or 15 years the methods to prepare these materials have gotten so much better, which means that the images of these materials have gotten better,” he says. “We have new experimental phenomena that we have to explain.”