Targeted electron beam creates thousands of atomic crystal defects

In 1959, the physicist Richard Feynman posed the question: “What would happen if we could arrange the atoms one by one the way we want them?” 1 This vision represents the ultimate limit of both materials science and nanotechnology — the study and control of matter down to the atomic scale. Highly focused electron beams can be used to manipulate single atoms in a crystal, but so far it has only been possible to make a handful of rearrangements at a time. Now, writing in Nature, Klein et al. 2 report a method for manipulating thousands of lattice sites in a crystal in minutes using an electron beam, creating controllable patterns of defects that promise to have intriguing quantum properties.

In 1990, researchers at the multinational technology company IBM were the first to report the repositioning of single atoms 3 on the surface of a material. Their work was based on scanning tunnelling microscopy (STM), in which an ultrasharp metal tip is brought close to atoms on top of an atomically flat surface. The tip can pick up atoms and position them precisely in desired patterns. However, although this technique has been useful for basic science, it can manipulate only surface atoms, and the resulting structures are stable only at cryogenic temperatures, limiting its possible applications.

In 2014, a different approach was introduced 4 : using highly focused electron beams instead of a physical tip. This technique was based on scanning transmission electron microscopy (STEM), which records the scattering of an electron beam at each position of a material to study its properties at the atomic level. STEM atomic engineering uses the same type of electron beam to manipulate atoms. This was originally explored in the 2D material graphene, and the researchers’ simulations indicated that if an electron beam with the right energy is precisely directed at the material, it can move an individual silicon-atom impurity to an adjacent point in the hexagonal lattice.

The first experimental demonstrations showed that this could be done at room temperature 5 , 6 , and the technique was subsequently extended to manipulate atoms in 3D crystals 7 . However, the challenge of scaling up electron-beam manipulation to more than a few atoms had, until now, stymied the field 8 .

In this study, Klein et al. altered the atomic structure of a semiconductor material at tens of thousands of user-selected lattice sites (Fig. 1a). The resulting patterns reached into the mesoscopic range, which lies between the atomic and bulk-material length scales. To perform this feat, they chose chromium sulfur bromide (CrSBr), an air-stable magnetic semiconductor. CrSBr can host thermally stable pairs of defects in which a missing atom, or vacancy, in the lattice occurs next to an atom called an interstitial that occupies a normally empty lattice site — in CrSBr, the interstitial sits between lattice planes. The authors successfully automated many of the steps in their STEM atomic-engineering process, enabling them to achieve greater precision and speed compared with previous approaches.

Klein and colleagues showed that directing an electron beam at CrSBr could move the chromium atoms in a column of the lattice to interstitial sites, creating a stack of vacancy–interstitial defects in the depth direction (Fig. 1b). These defect complexes were generated by quickly moving the electron beam back and forth in a line that crossed the target column of chromium atoms.

The experimental findings are supported by advanced computer simulations, which hint that the defects are correlated and form collective quantum states. The work shows both impressive control over atomic-defect placement and a scaling-up that vastly surpasses the number of atomic manipulations that have previously been reported.

Before atomic engineering could begin, the positions of the atomic columns needed to be accurately detected. The authors developed a method for detecting the lattice that minimized the dose of irradiated electrons, which can damage the material. They further minimized unwanted changes during atomic manipulation by detecting the repositioning of the chromium atoms using the real-time electron scattering signal and automating the repositioning of the electron probe at the next site in the programmed pattern.

Defects in certain crystals have properties that make them attractive for emerging quantum technologies, but the potential for precisely positioning these defects has been limited to surface engineering using STM 9 . Programmable 3D arrays of defects could be used to perform simulations of interacting quantum systems and could open up paths towards scalable quantum computing, dense magnetic memory and atomic-scale logic devices. Many existing quantum platforms are based on the trapping of ultracold atoms or ions, but the authors showed that the structures they created remain stable at room temperature and outside the microscope, making them compatible with common optical, electrical and magnetic measurement techniques.

Some challenges remain. STEM imaging generates a 2D projected image of a 3D structure, which means the authors could not determine the exact depth of each defect in the crystal, let alone control the defects’ placement in the direction perpendicular to the surface. This limitation might be overcome by rapid ongoing progress in electron ptychography 10 , an advanced computational imaging technique that can yield depth resolution, but full 3D control of atomic manipulation might require fresh ideas and breakthroughs.

Finally, although the authors identified CrSBr as an excellent model system that has potentially tunable properties, and the technique is, in principle, applicable to a broad range of materials, extensive modelling and experimental testing will be required to find other systems that can be manipulated in this way.

The ability to engineer the properties of a material by purposefully removing or introducing various elements has enabled a plethora of technological applications, from high-performance semiconductors to quantum devices including single-photon emitters. The precise, top-down control enabled by the automated electron-beam manipulation is a powerful tool with which to engineer the properties of materials with unprecedented control.