Until recently, many intriguing phenomena that appear in transition metal dichalcogenides (TMDs), a family of two-dimensional (2D) semiconductors, were associated with chalcogen vacancies, the commonly assumed “dominant” type of defect in these materials. Defects in TMDs, as in most semiconductors, have been used extensively to tune their electronic, optical, and catalytic properties. In a study published in a recent issue of Physical Review Letters, an international team of researchers from the US, Israel, Germany, and Spain selectively engineered sulfur vacancies in their samples, thus offering a new form of defect engineering for 2D TMDs. The researchers discovered that chalcogen vacancies exhibit extraordinarily large spin-orbit coupling, which could allow tuning of the opto-spintronic properties of these materials.  By using a combination of scanning probe microscopy techniques and ab initio theory, they were able to link unequivocally the atomic and electronic structure of a single sulfur vacancy in the 2D material. 

Transition metal dichalcogenides, like WS₂ used in this study, are semiconductors of the type MX₂, where M is a transition metal atom (in this case tungsten or W) and X is a chalcogen atom (elements of the oxygen group, here sulfur or S). Because of the reduced dimensionality and quantum confinement, TMDs are particularly sensitive to certain types of defects, like substituents and vacancies. Vacancies are suspected, for example, to induce catalytic functionality to the otherwise inert surface of a TMD and are also widely believed to be the reason behind the modification of the optoelectronic properties in these materials. This is because defects can add additional states in the bandgap, an energy forbidden region for electrons in the pristine material.

According to Bruno Shuler, group leader at the Lawrence Berkeley National Laboratory and first author of the article, not detecting any chalcogen vacancies in their original as-grown pristine WS₂ samples came as a big surprise. “This result made us initially very skeptical about our interpretation of the experimental data, since it was completely against the common perception in the field,” Schuler says. 

Chemical vapor deposition was used to grow the WS₂ material, which consists of a chalcogen layer, a transition metal layer, and a second chalcogen layer. Sulfur vacancies were then selectively introduced by in-vacuum annealing (high-temperature heating in vacuum) in the top and bottom chalcogen layers. 

To identify the defects, the team chose atomic force microscopy (AFM) with a single carbon monoxide atom at the end as a sensor. Schuler says that the reason they opted for AFM instead of a transmission electron microscope (TEM) that has been previously used was that the high electron beam of TEM can induce damage and create defects, which cannot be distinguished from the ones that pre-exist in a sample. AFM not only helped the researchers to get information about the atomic structure of single defects, but also to locate each defect on the lattice, thus helping their identification as sulfur vacancies. 

Finding out the type and position of defects that grow in these materials was only part of the problem. Assessing the functionality of the defects, which is mainly connected to their electronic structure, was an even more important puzzle to solve. For this, the researchers used low temperature scanning tunneling microscopy (STM). 

“Optical-based techniques, like optical spectroscopy, have been used before to uncover the electronic states of optically active defects that exist in a sample,” Schuler says. “The resolution of such techniques is, however, limited by the diffraction limit of the size of the light. As a result, ensembles of thousands of defects are simultaneously measured, without being possible to tell two defects apart or recognize which atomic site the signal is coming from,” he says. 

By resolving the electronic orbitals associated with the signal from the defects using scanning tunnelling spectroscopy, the team measured a very large energetic splitting induced by spin-orbit coupling. “If you take out one sulfur atom to create a vacancy, then you have three ‘dangling’ or broken bonds from tungsten atoms, which were formerly connected to the sulfur. The vacancy states are thus made of tungsten d-orbitals. These d-electrons have a high orbital momentum, which results in strong spin-orbit coupling,” Schuler says. 

“What a defect does is to introduce additional electronic states in the bandgap of a semiconductor and this is very important for the optical transitions or the catalytic functionality of a defect. Knowing the number and the location of the defect states within the bandgap is very important to figure out what their functionality is,” Schuler explains. 

Vacancies in WS₂ were often predicted to have just one additional electronic state in the bandgap of the pristine material. However, because of the spin-orbit splitting, two electronic states were measured at 252 (±4) meV apart. 

If the vacancies are absent, then which defects are responsible for the properties of the materials? “What we will do in the future is a targeted introduction of defects, to see how properties change,” Schuler, says. “Correlating the electronic properties with the optical signatures of chalcogen vacancies will be the next important step for us. Chalcogen vacancies are also ideal sites for atomically precise functionalization of 2D TMDs with arbitrary dopants. We plan to use the highly reactive vacancy sites as an ‘anchor point’ for atomic impurities, which is of particular interest for next-generation color centers,” he says. 

Nacho Pascual, an expert in STM and non-contact-AFM and group leader of the nanoimaging group at NanoGUNE, a research center in San Sebastián, Spain, who did not participate in the study, says, “These results open the door to selective functionalization of TMDs with specific species, and develop a new class of functional materials. The survival of such large spin-orbit coupling in the localized states and site-selective access with local probes will allow the [researchers] to create a single photon source with controlled chirality.” 

For Blanca Biel, professor in the University of Granada, Spain and expert in 2D materials, who also did not participate in the study, this work is a call to revisit long-time assumed ideas on defects in 2D-TMDs. “Works like this one allow us to gain access to much needed accurate experimental data that will either confirm or falsify the current theoretical predictions. It also facilitates the path to controlled chemical doping or selective functionalization with ‘elusive’ molecules, by, for instance, anchoring them to defects created in specific sites of the TMD lattice,” she says. 

Read the abstract in Physical Review Letters.