Deciphering chemical order/disorder and material properties at the single-atom level

The three-dimensional coordinates of more than 23,000 atoms in an iron-platinum nanoparticle are determined with 22 picometre precision to correlate chemical order/disorder and crystal defects with magnetic properties. Material properties at the single-atom level FePt nanoparticles have practical potential in fields as diverse as catalysis and magnetic storage media. But far from being pristine crystalline materials, these nanoparticles are structurally heterogeneous with grain boundaries and other crystal defects. In this paper, Jianwei Miao and colleagues reveal the complex atomic-scale structure of a single FePt nanoparticle containing more than 22,000 atoms. They do this by generating a high-resolution tomographic tilt series of 68 images of the nanoparticle and reconstructing it using a new algorithm, achieving resolution with 22 picometre precision. The resulting structure reveals the complexity of the nanoparticle, and the chemistry and crystal structure of the grains within the material. When analysing the order/disorder character, the authors find that the grains are more ordered towards the core of the nanoparticle and less ordered towards the surface. They use data from the boundary between two grains to calculate local magnetocrystalline anisotropy energies using density functional theory, revealing how these energies vary across the grain with order parameter and across a grain boundary. Perfect crystals are rare in nature. Real materials often contain crystal defects and chemical order/disorder such as grain boundaries, dislocations, interfaces, surface reconstructions and point defects 1 , 2 , 3 . Such disruption in periodicity strongly affects material properties and functionality 1 , 2 , 3 . Despite rapid development of quantitative material characterization methods 1 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , correlating three-dimensional (3D) atomic arrangements of chemical order/disorder and crystal defects with material properties remains a challenge. On a parallel front, quantum mechanics calculations such as density functional theory (DFT) have progressed from the modelling of ideal bulk systems to modelling ‘real’ materials with dopants, dislocations, grain boundaries and interfaces 19 , 20 ; but these calculations rely heavily on average atomic models extracted from crystallography. To improve the predictive power of first-principles calculations, there is a pressing need to use atomic coordinates of rea