The locations of atoms in a metallic alloy nanoparticle have been determined using a combination of electron microscopy and image simulation, revealing links between the particle's structure and magnetic properties. See Letter p.75
The physical and chemical properties of binary alloys — metals that consist of two elements — can be tuned by varying the composition and arrangement of their atoms. Alloys made of iron (Fe) and platinum (Pt) are of particular interest because they can yield one of the magnetically 'hardest' materials (magnetization is retained for a long time) as well as one of the magnetically softest. On page 75, Yang et al.1 report a quantitative experimental study of a binary alloy at the single-atom level. Using an 8.4-nanometre FePt nanoparticle, the authors demonstrate a state-of-the-art approach for determining the 3D atomic arrangement of such particles. Their technique could also lead to a better understanding of how nanoparticles form from initially small clusters of atoms.
At the nanoscale, every atom counts. For example, changing the relative positions of a few Fe and Pt atoms in a FePt nanoparticle dramatically alters the particle's properties, such as its response to a magnetic field. Why is this? For a particle that has a diameter smaller than 9 nm, 8% or more of its atoms reside on the surface2. Pt is a very good catalyst, so moving some of these atoms to the surface increases the particle's catalytic activity3. At the same time, this would strengthen such a particle's magnetism because Fe atoms would have been pushed to the centre. Bifunctional (hybrid) nanoparticles can therefore be obtained in which catalytic and magnetic properties can be tuned by repositioning the respective elements.
Yang and colleagues used an imaging technique called atomic electron tomography to determine the shape of an 8.4-nm FePt particle and the 3D positions of its 23,196 atoms. First, the authors passed an electron beam through the nanoparticle at a fixed angle and used the scattering of the beam to produce a 2D image, or projection, of a slice of the particle (Fig. 1a). They then rotated the particle through angles of −65.6° to +64.0° and recorded a series of 68 such projections. Finally, Yang et al. used impressive numerical simulations to combine the 2D images, producing a 3D reconstruction of the particle, in which the individual Fe and Pt atoms are clearly distinguishable.
The authors achieved a remarkable spatial resolution of 22 picometres (1 pm is 10−12 m), corresponding to roughly one-tenth of the diameter of the atoms — this allowed the position of each atom to be determined. This resolution, however, is insufficient for analysing small changes in atomic positions owing to strain or surface-relaxation effects4. Nevertheless, the authors' experimental data provide us with a much better understanding of how small clusters of atoms combine to form nanoparticles.
Yang et al. also looked at the magnetic properties of their FePt nanoparticle. The fraction of atoms that exist on the surface of such particles is large. These surface atoms have a smaller number of nearest neighbours than atoms in the bulk, causing the interatomic distances to oscillate between the particle's outer three or four layers5. Because a nanoparticle's magnetic properties are highly sensitive to interatomic distances in the bulk, numerical simulations of such particles should, in principle, take this effect into account. The authors therefore used a well-established theoretical approach called density functional theory6 to determine local magnetic properties from the measured atomic coordinates.
The magnetic moment and magnetic anisotropy energy density — the dependence of the internal energy of a material on the direction of its magnetic moment — rely on composition and local symmetry. Structural distortions of a few per cent can cause the magnetic anisotropy to vary by as much as a factor of 1,000 (ref. 7). Consequently, undistorted FePt is magnetically soft whereas distorted (ordered) FePt is one of the magnetically hardest materials, comparable to samarium cobalt or neodymium iron boron8. Yang et al. found that their FePt nanoparticle is not homogeneous, but instead has locally varying order and composition, yielding different local magnetic anisotropies (Fig. 1b). This experimental confirmation of previous speculation9 explains why the magnetic anisotropy generally observed in FePt nanoparticles is smaller than expected, and provides an opportunity to design magnetic nanoparticles that have tailored properties.
Although Yang and collaborators' work is an essential step forward in the determination of 3D atomic arrangements, experimental limitations remain. The spatial resolution of 22 pm achieved by the authors would need to be improved by a factor of about 100 to analyse strain and surface-relaxation effects, which typically alter interatomic distances by 0.1–1 pm (refs 4, 7). Another limitation is that Yang and colleagues' nanoparticle was embedded in a carbon matrix. This might have modified the particle's surface energy (and, consequently, its shape and magnetism) compared with a 'naked' particle6,10. The matrix might also have allowed carbon, which couldn't be resolved in the authors' experiment, to diffuse into the bulk of the particle.
Researchers studying magnetic nanoparticles are waiting for a '7D' transmission electron microscope that would allow the 3D position and 3D magnetic moment of a particle to be determined with sub-nanometre spatial resolution and picosecond time resolution. Work is in progress across the world to separately address these challenges.Footnote 1
Yang, Y. et al. Nature 542, 75–79 (2017).
Vollath, D. Nanomaterials: An Introduction to Synthesis, Properties and Applications (Wiley, 2008).
Pohl, D., Wiesenhütter, U., Mohn, E., Schultz, L. & Rellinghaus, B. Nano Lett. 14, 1776–1784 (2014).
Wang, R. M. et al. Phys. Rev. Lett. 100, 017205 (2008).
Cho, J-H., Ismail, Zhang, Z. & Plummer, E. W. Phys. Rev. B 59, 1677–1680 (1999).
Gruner, M. E., Rollmann, G., Entel, P. & Farle, M. Phys. Rev. Lett. 100, 087203 (2008).
Hjortstam, O., Baberschke, K., Wills, J. M., Johansson, B. & Eriksson, O. Phys. Rev. B 55, 15026 (1997).
Gutfleisch, O. J. Phys. D: Appl. Phys. 33, R157–R172 (2000).
Antoniak, C. et al. J. Phys. Condens. Matter 21, 336002 (2009).
Antoniak, C. et al. Nature Commun. 2, 528 (2011).
Related links in Nature Research
About this article
Artificial Cells, Nanomedicine, and Biotechnology (2018)
Composition-Tunable Optical Properties of Zn x Cd(1 − x)S Quantum Dot–Carboxymethylcellulose Conjugates: Towards One-Pot Green Synthesis of Multifunctional Nanoplatforms for Biomedical and Environmental Applications
Nanoscale Research Letters (2017)