HAADF, BF detectors, and EELS spectrometer. (b), and (c) Typical ADF-STEM image and core-loss EELS of doped monolayer MoS2. Reproduced from Robertson et al. (2016) with permission. (d) Monochromated valence EELS of monolayer MoS2. Reproduced from Suenaga et al. (2015) with permission.
While imaging by annular detectors, the post-column spectrometer receives the low-angle scattered and transmitted electrons after an energy transfer to the target atom in the electron–sample interaction, to yield the electron energy loss spectroscopy (EELS). Through an inelastic scattering, the energy transfer from the fast-incident electron to the target atom will excite the core electron in the K and L shells, or the valence electrons forming the band structure, into unoccupied electronic states above the Fermi level of the sample. According to the Fermi’s golden rule, the scattering intensity, or cross section, mainly depends on the energy gap and the density of unoccupied states to accommodate the excited electron. This will give rise to the cross section of low loss being several orders of magnitude higher than that of core loss. Generally, low-loss signals with a large cross section are dependent on the electronic structure of the sample and behave as interband transition (Fig. 1(d)) and plasmon excitation related with the valence electrons of the specimen. In the high-loss regime of EEL spectrum, chemical species (Fig. 1(c)) and unoccupied electronic states of the target nanostructures can be quantitatively determined.
In recent years, instrumental performance has advanced to atom-by-atom spectroscopic analysis of the crystal specimen even with single-atom accuracy, higher sensitivity, and energy resolution. Owing to the special versatility in the light element analysis, the valence state of the target elements can be determined from a quantitative measurement of the chemical shift of the characteristic K, L and M edges in the core-loss EEL spectrum. Meanwhile, chemical environment and magnetic structure of the target atom can also be deduced from the energy loss near edge fine structure (ELNES) at the single-atom level. In defective crystals, ELNES can be employed to unveil the defect-induced nanophysics such as valence/spin states and coordination crystal fields of single transition metal dopants or other defects. Specifically, high-energy resolution EELS has been demonstrated as a versatile technique to measure bandgaps, plasmon excitations, or even the vibrational modes of various crystal materials.
2.2. Principles of STM and STS
STM/STS is a common atom probe used in electronic state imaging and analysis in surface science, based on the quantum tunneling effect between the sample surface and the STM tip. It is particularly suited for studying the defects in surfaces due to its high spatial resolution as well as the fact that electronic states introduced by defects will contribute to the tunneling current and thus be revealed by the difference in STM contrast between defects and the surrounding defect-free region. More importantly, by performing scanning tunneling spectroscopy measurements, one can derive the local density of states (LDOS) by taking the differential conductance spectra at the defects and its localization, thus providing direct evaluations of the electronic properties of the defects. Instead of repeating on the working principles of STM/S here, readers are referred to a few excellent monographs or book chapters on this celebrated technique [‘Methods of Experimental Physics, Vol. 27, Scanning tunneling microscopy’ Joseph A. Stroscio & William J. Kasiser (eds.), Academic Press 1993; ‘Introduction to Scanning Tunneling Microscopy (2nd ed.)’ C. Julian Chen, Oxford University Press 2008; ‘Scanning Probe Microscopy, Atomic Force Microscopy and Scanning Tunneling Microscopy’, Bert Voigtlaender, Springer 2015; ‘Scanning Probe Microscopy, Analytical Methods’ Roland Wiesendanger (ed.), Springer 1998; ‘Scanning Tunneling Microscopy and Its Application’ Chunli Bai, Springer 2000].
3. Atomic Defects in 2D Transition Metal Dichalcogenides
In the post-graphene era of 2D materials, the diverse layered transition metal dichalcogenides (TMDs) [4–7] are a large 2D family with unique structure, opto/electronic, and valleytronic properties, especially in valleytronics [8–11] and electronics [12–14] application. Among them, IV–VI group MX2(M = Mo, W; X = S, Se) has three types of crystal phases: 1T phase as a metal and 2H and 3R phases as semiconductors. The hexagonal MoS2/MoSe2 in the 2H phase has been extensively investigated in materials science and finds wide applications in industries [15] as lubricants and hydrodesulfurization catalysts.
3.1. Point defects in monolayer MoS2
Two-dimensional layered MoS2 is a typical semiconductor with a well-known cross-over from indirect to direct bandgap when the thickness decreases from bulk to monolayer, as a result of quantum confinement effect. Semiconducting MoS2 (E.g., 1.3–1.8 eV) can be promising building blocks of photodetectors, gas sensors, and opto/electronic devices. To synthesize MoS2 atomic layers in large scale, chemical vapor deposition (CVD) [16, 17] has been demonstrated as a feasible route to realize the scalable nanoelectronic applications based on large-size high-quality thin films. However, plenty of point defects and grain boundaries are still inevitably present in the atomic thin layers after the CVD growth.
3.1.1. Vacancies and antisite defects
Diverse intrinsic point defects in CVD-grown MoS2 were first systematically characterized by Zhou et al. [1] by using atomic resolution HAADF imaging. Due to the large intensity difference of Mo and S2 column in this Z-contrast mode, it is easy to directly assign the type of the point defects. Single-site sulfur vacancies were found to be the most common defects, including mono-vacancy (VS) and double-vacancy (VS2) with only one or two S atoms missing from the S sublattice. Other less common defects observed include extended Mo vacancies such as VMoS3 and VMoS6, and antisite defects with Mo atom replacing S2 column (MoS2) or S2 occupying the Mo site (S2Mo), but with a much lower frequency. Through HRTEM imaging, Komsa et al. [18] observed the structure of single vacancies VS and VS2 in monolayer MoS2 with atomic resolution, which could be readily created by electron beam irradiation at an acceleration voltage of 80 kV. Atomic S vacancies get generated and agglomerated in the monolayer under the electron beam irradiation. These vacancy sites could accommodate impurity atoms to form substitutional dopants, such as N, P, As, and Sb in V-A group behaving as acceptors and F, Cl, Br, I in VII-A group as donors, respectively. This electron beammediated substitutional doping could serve as a route to engineer the local electronic structure of TMDs.
Using atomically resolved ADF imaging, Jin et al. found plenty of antisite defects emerging in physical vapor-deposited (PVD) MoS2 monolayers. Figure 2 is an image gallery to demonstrate all types of antisite defects in monolayer MoS2 including Mo replacing S sublattice (MoS, MoS2, Mo2S2) and S substituting Mo sublattice (SMo, S2Mo). In the ADF imaging mode, these two different categories of antisites can be easily distinguished and even quantitatively analyzed. The experimental atomically resolved ADF-STEM images of antisites agree well with the simulated images based on density functional theory (DFT) relaxed structures.