AVS 70 Session 2D-ThM: 2D Materials: Defects, Dopants, Edges, Functionalization, and Intercalation

Transition metal dichalcogenides (TMDCs) are materials with unique electronic properties due to their two-dimensional nature. Recently, there is a large and growing interest in synthesizing ferromagnetic TMDCs for applications in electronic devices and spintronics. Apart from intrinsically magnetic examples, modification via either intrinsic defects or external dopants may induce ferromagnetism in non-magnetic TMDCs [1-2]. Our study focuses on intrinsic defects in TiS₂, which is a system known for potential applications in energy storage. We use scanning tunneling microscopy (STM) and superconducting quantum interference device (SQUID) magnetometry to characterize the crystal structure and magnetic properties of TiS₂ crystals. Atomically resolved STM images suggest the formation of sulfur vacancies and possibly interstitial defects creating brighter triangular shape regions. Preliminary analysis of magnetic data indicates low-spin paramagnetic response, with a saturated magnetization of ~0.16 emu/g and 80% saturation by ~2.5 T. To rule out ferromagnetism, hysteresis loops were analyzed and showed the coercive field to be zero within experimental error. Concurrently, DFT calculations on formation energy and electronic density were also being performed for proper identification of defect formations. Additionally, simulated STM images were generated by calculations that map the electronic density of the surface for the energetically favorable defects. An initial comparison to experimental STM images corroborates with the initial hypothesis of sulfur vacancies and titanium interstitial defects. Further studies include doping of TiS₂ with transition metals and further investigation of the electronic and magnetic properties of these doped TMDC systems.

[1] Magnetic doping in transition metal dichalcogenides, PM Coelho, J. Phys.: Condens. Matter 36 203001 (2024).

[2] Room-Temperature Ferromagnetism in MoTe₂ by Post-Growth Incorporation of Vanadium Impurities, PM Coelho, HP Komsa, K Lasek, V Kalappattil, J Karthikeyan, MH Phan, AV Krasheninnikov, and M Batzill, Adv. Electron. Mater., 1900044 (2019).

Titanium disulfide (TiS₂) belongs to a family of materials known as transition metal dichalcogenides (TMDs). These materials exhibit relatively weak van der Waals forces between layers, allowing them to be reduced down to thin films and even one monolayer in thickness. While graphene is the most well researched 2D material, it does not possess a bandgap and thus has limited applications with regard to electronic devices. TMDs often have tunable bandgaps, filling a clear need within 2D materials. They also exhibit sandwich-like structures, with one layer of transition metal atoms sandwiched by two layers of chalcogen atoms.

The most researched applications of 2D TiS₂ are those regarding ion batteries, photovoltaic devices, and spintronics. This study investigates the effects of intrinsic point defects within a 2D (TiS₂) lattice on the electronic and magnetic properties. By comparing post-annealed scanning tunneling microscopy (STM) images with SGGA+U-based density functional theory (DFT) calculations, we examine which defects are more likely to occur by connecting specific geometries that arise from typical intrinsic defects. We also compare a base, benchmarked 2D TiS₂ supercell with supercells containing various point defects, allowing us to interpret the total energy differences and identify which are most energetically favorable.

Like in any other semiconductor, point defects in transition-metal dichalcogenides (TMDs) are expected to strongly impact their electronic and optical properties. However, identifying defects in these layered two-dimensional materials has been quite challenging with controversial conclusions despite the extensive literature in the past decade. Using first-principles calculations, we revisit the role of chalcogen vacancies and hydrogen impurity in bulk TMDs, reporting formation energies and thermodynamic and optical transition levels. We show that the S vacancy can explain recently observed cathodoluminescence spectra of MoS₂ flakes and predict similar optical levels in the other bulk TMDs. In the case of the H impurity, we find it more stable sitting on an interstitial site in the Mo plane, acting as a shallow donor, and possibly explaining the often observed n-type conductivity in some bulk TMDs. We also predict the frequencies of the local vibration modes for the H impurity, aiding its identification through Raman or infrared spectroscopy.

Our results show that the chalcogen vacancies are deep acceptors and cannot lead to n-type or p-type conductivity in monolayer TMDs. Both the (0/-1) and (-1/-2) transition levels occur in the gap, leading to paramagnetic charge states S=1/2 and S=1, respectively, in a collinear-spin representation. We discuss trends in terms of the band alignments between the TMDs, which can serve as a guide to future experimental studies of vacancy behavior.

As silicon-based electronics approach their physical limitations, two-dimensional (2D) transition-metal dichalcogenides (TMDs) have emerged as potential alternatives due to their unique mechanical and electrical properties. Among TMDs, WS₂ exhibits layer-dependent bandgaps and impressive electron and hole mobilities. However, to effectively integrate 2D TMDs into advanced applications, reliable doping strategies must be developed.

Conventional ion implantation doping methods used in silicon and GaAs-based semiconductors are not suitable for 2D TMDs due to lattice damage and high post-annealing temperatures. Alternative techniques, such as substitutional doping and charge transfer doping, have been explored but face limitations related to high-temperature processes and dopant carrier concentration control.

This study investigates the n-type doping effect of high-energy proton beam irradiation on intrinsic n-WS₂ using various fluences. Previous studies have focused on post-fabricated devices, making it difficult to isolate the specific interactions between the substrate and TMD materials. By examining the effects of proton irradiation on bulk WS₂ crystals, this study aims to understand the intrinsic impacts with different proton fluences on 2D WS₂.

Electrical properties, defect state formation, and work function variations were assessed using transmission line measurement, Raman spectroscopy, low-temperature photoluminescence spectroscopy, and Kelvin probe force microscopy. Density Functional Theory (DFT) calculations suggest that sulfur vacancies generated by proton irradiation contribute to the formation of defect states near the conduction band, facilitating an n-type doping effect.

To validate the effects of proton beam irradiation-induced defects, multilayer WS₂ photodetectors exposed to varying proton fluences were fabricated. Increasing proton fluence led to a significant improvement in internal photo-gain under 530 nm wavelength illumination, attributable to the creation of mid-gap states. This proton-mediated doping technique offers insights into the direct effects of high-energy protons on WS₂, potentially informing the development of advanced ultra-large-scale integration devices with controllable proton irradiations.

This work was supported by the Korea Research Institute for defense Technology planning and advancement (KRIT) grant funded by Defense Acquisition Program Administration (DAPA) (KRIT-CT-21-034).

Transition metal dichalcogenides (TMD) MX₂ offer unique and versatile functionality for a wide range of electronic, photonic, and quantum devices. It remains a significant challenge to achieve electronic or magnetic doping. In conventional semiconductors such as Si, or III-V materials ion implantation is well-established as a versatile method of doping but TMDs are much more susceptible to damage. We introduce a new concept termed “backdoor doping” which mimics low energy ion irradiation with an energy below 50 eV with the goal of doping of the metal sub-lattice with a wide range of elements while at the same time minimizing damage.

We combine experimental studies of defects in TMDs with computational models describing the ion-matter interactions, and defect signatures using DFT. We will compare defects created by (i) annealing, (ii) ion irradiation, and (iii) plasma exposure in a semiconducting and a metallic TMD (e.g. MoS₂, TaS₂). We observed defects created by annealing in WSe₂ with scanning tunneling microscopy and spectroscopy (STM, STS) and density functional theory (DFT). Initially chalcogen vacancies dominate but larger, extended defects with multiple damaged bonds develop rapidly. The defects introduce numerous states in the gap and damage the structural and bonding integrity of the TMD layer. A small background pressure of the chalcogen might help to repair broken M-X bonds and offer a path to defect repair. We will systematically compare the different modes of defect formation using experiment and DFT modeling. If time permits, STEM (scanning transmission electron microscopy) will be included for detailed structural and phase information.

The low energy ion irradiation is achieved by using a “sandwich structure” made of a TMD layer which is capped with a thin metal layer made of the desired dopant. For proof of concept we are using Au and Fe metal layers. The sandwich structure is irradiated from the metal layer side and the projectile (a noble gas ion with an energy <5 keV) will initiate a collision cascade. The collision cascade is adjusted to eject metal atoms from the thin layer into the TMD while limiting transmission of the projectile. The defect formation in the metal is modeled with SRIM/TRIM (stopping and range of ions in matter) and SDTrimSP which are Monte Carlo simulations of collision events in matter. These simulations predict the energy of ions leaving the metal layer and entering the TMD. The TMD layer is then studied with STM to assess the defect inventory, XPS to understand compositional variations, and other methods to fully characterize the TMD.

Heterogeneities can be selectively engineered in 2D materials from the interaction of charged projectiles at varying velocities, angle of incidences, and stage temperatures. These heterogeneities exist as intrinsic structural defects, impurities in lattice sites, topological disorder, and strain-driven interfaces. So, how can we study these defects in detail to correlate their impact on materials properties? In this talk, aberration-corrected scanning transmission electron microscopy (STEM) is combined with monochromatic electron energy loss spectroscopy (EELS) to combine unprecedented spatial resolution with world-class energy resolution from an electron probe to decouple the nature of bosons arising from subtle heterogeneities. On-axis EELS, where our bright field disc is perfectly aligned with the EELS entrance aperture, leaves our signal being dominated by dipole scattering, so we get a delocalized signal. To get a localized signal from adatoms and vacancies in our material, we must go off-axis to suppress the long-range signals and highlight hyperbolic phonon polaritons. Hereby, off-axis EELS is employed, where a bright-field disc is electrically shifted outside of the EELS entrance aperture by the projector lenses such that a portion of the dark field disc now enters the spectrometer window to give information on the impact scattering in the off-axis regime. We use the selection rules inside an electron microscope to decouple the optical transitions arising from single anomalies in 2D materials. Ultimately, this work not only pushes boundaries in electron microscopy, but provides avenues to the entire scientific landscape on decoupling the defect-property relationship in solids for the better design of next-generation nanoelectronics.

In this talk we’ll present the results of an extensive collaborative project aimed at the cross-correlated nanoscale AFM (atomic force microscopy) and TERS (tip enhanced Raman scattering) characterization of various defects appearing in the course of synthesis of Janus transition metal dichalcogenides (TMDs). Optoelectronic and catalytic properties of Janus TMDs differ from normal TMDs, thus Janus materials are of great interest for the research community.

In the course of our study, we characterized MoSSe and MoSeS Janus materials derived from MoS₂ and MoSe₂ correspondingly via the cold plasma-assisted replacement of the top chalcogen layer. Preliminary Raman characterization of the as-synthesized crystal performed in a single point with the single excitation wavelength (532nm) showed weak Raman peaks of the precursors, what was misinterpreted as incomplete conversion to Janus material.

Cross-correlated AFM and TERS imaging revealed that the precursor monolayers, both MoS₂ and MoSe₂ featured noticeable number of nanoscale bi-, tri- and higher number of layer islands. These islands have been identified in the Janus crystals transferred to gold or silver via the Kelvin probe imaging and their composition was confirmed by TERS imaging even for the islands of just 20-30 nm in diameter. It was these multi-layer islands which were responsible for the weak Raman bands of the precursor.

The morphology of the Janus crystals derived from MoS₂ and MoSe₂ was also fundamentally different. In the course of conversion of MoSe₂ to MoSeS pre-existing tensile strain in MoSe2 was complemented by additional tensile strain resulting from the replacement of the selenium atoms with sulfur which led to physical breakage of the crystals. TERS imaging demonstrated that the gaps between the domains in MoSeS monolayers seen in topography and the surface potential images were physical cracks.

Conversely, compressive strain appearing in MoSSe Janus crystals converted from MoS₂ results in the formation of wrinkles that after the transfer to gold or silver looked like cracks in MoSeS, but in reality there was no physical breakage in these crystals. Interestingly, by varying the substrate on which the precursor crystals are grown, nice wrinkle- and crack-free Janus monolayers can be produced.

Finally, we’ll briefly discuss the TERS spectral peculiarities with 785 nm and 473 nm excitation.

With increasing interest in transition metal tellurides (TMTe_x) for optoelectronic applications, physical vapor deposition (PVD) thin film growth techniques offer unique opportunities to decouple factors influencing material characteristics during growth. TMTe_x thin films are commonly engineered via modulation of their domain morphology and dimensionality through deposition process controls for regulating the crystallinity. Film crystallinity regulation can be achieved using PVD techniques, like molecular beam epitaxy (MBE), through its unique co-deposition process controls. While co-deposition, or simultaneous deposition, of both the transition metal and tellurium (Te) is a common approach, this process obscures the influence of individual precursors due to the complex thermodynamic and kinetic growth conditions. The simultaneous release of precursors directly influences the vapor flux which is correlated to the film roughness and thickness. Thin film growth is further dictated by the interaction between adatoms and the substrate surface. Particularly, lattice mismatch at the film-substrate interface is influenced by the crystalline/amorphous surface nature of Asaro–Tiller–Grinfeld instability, indicating a significant impact on the grown crystalline domain size and morphology. We demonstrated sequential deposition of precursors on substrates with varying crystal structure and chemical composition to decouple the influence of different factors, yielding TMTe_x with distinguishable morphology.

Specifically, we demonstrated control over the film morphology which is dependent on the initial precursor (TM/Te) deposited on either crystalline or amorphous substrates. Initial deposition of either of the precursors on muscovite mica (crystalline) resulted in the formation of fully oriented nanoribbon films with significantly different domain morphology. In contrast, either TM-initiated or Te-initiated deposition on oxidized/nitrided silicon (amorphous) surfaces, led to either disordered films or randomly oriented nanoribbons, respectively. This demonstrates that MBE-grown domain morphology of TMTe_x thin films is highly dependent on precursor sequence and substrate selection, permitting transition between uniform and nanoribbon films for future applications in field-effect transistor or gas sensor device development.

Tin selenide (SnSe) is predicted to exhibit a d₁₁ piezoelectric response of 250 pm/V when scaled down to the monolayer limit [1], but as the film thickness is decreased surface interactions can have a greater impact on performance. For example, in 2D transition metal dichalcogenides (TMDs), oxides form not only on the surface but also propagate between layers, degrading the thin film [2]. Oxide layers have been shown to form at the surface of bulk SnSe above 200 °C [3], but the impact of oxygen on the surface and layer stability at room temperature has yet to be fully explored. Therefore, it is critical to understand the stability of the surface and interlayer structure in SnSe under standard operating condition.

This talk describes the chemical and structural stability of SnSe thin films to atmospheric exposure. The structure and composition of thin films of SnSe grown by molecular beam epitaxy is compared immediately upon growth and after a two-year exposure to ambient atmosphere. X-ray diffraction (XRD) taken before and after atmospheric exposure shows no measurable change in the crystallographic phase, orientation, or layer spacing. Similarly, the characteristic Raman vibrational modes of SnSe are unchanged with exposure, and there are no indications of modes corresponding to SnSe₂, SnO or SnO₂. These measurements show that the bulk of the SnSe thin film is not degrading over time as incorporating oxygen between layers would have changed the peak spacing in XRD and Raman spectroscopy. The chemical stability of the bulk phase is further supported by x-ray photoelectron spectroscopy (XPS) measurements that show that the film maintains a 1:1 Sn:Se stoichiometry. XPS cross-sections show indications of SnO₂ but only at the surface of the SnSe film. The oxide layer was limited to the surface within approximately 3.5 nm, as confirmed via x-ray reflectivity (XRR) measurements of the layer thicknesses. Resistivity measurements show an electrical response dominated by SnSe, not SnO₂. These XPS, Raman, XRR, and resistivity results suggest that exposure to atmosphere creates a passivated layer of SnO₂ on the surface of SnSe but does not impact the bulk. Overall, SnSe demonstrates long-term chemical stability under atmospheric conditions, rendering it a suitable option for device applications that employ the protective metal oxide layer found in SnSe and possibly other layered chalcogenide structures.

1. R. Fei, W. Li, J. Li, L. Yang, Appl. Phys. Lett. 107, 173104 (2015).
2. Y. Guo, S. Zhou, J. Zhao, ChemNanoMat 6, 838–849 (2020).
3. Sci. 21, 3333–3338 (1986).

Due the ever decreasing device size driven by the electronics industry, the future of channel materials for small transistors is heading toward the use of two dimensional transitional metal dichalcogenide (2d TMD) materials. 2d TMDs are beneficial in the short channel regime because of their potential high on-to-off current ratios and because of their potentially high channel mobilities, or conductance, between the source and the drain.Beyond TMD electrical and structural quality, two large challenges exist for the implementation of TMDs into transistors. First is the electrical contact of source and drain metals to the TMD and second is the deposition of a nanometer thin, high quality, high k, defect free dielectric material on top of the TMD to form the critical transistor gate dielectric. To tackle the second challenge, the surface of WS₂ TMD was functionalized with a low temperature remote oxygen plasma which enabled nucleation sites to form on the surface of multi-layer WS₂ without catastrophic destruction of the long-range order of the WS₂. From these nucleation sites, ALD HfO₂ and Al₂O₃ using TMEAH (tetrakis(methylethylamido)hafnium)/H₂O and TMA (trimethylaluminum)/H₂O at 200C was deposited as the gate dielectric and a 5 nm thin layer with 100% coverage could be achieved at increased oxygen plasma exposures. This WS₂ functionalization / nucleation process was explored by measuring the effects of oxygen exposure and temperature and their resulting effect on the ALD deposited gate dielectric film. The ALD gate dielectric coverage and its effects on the underlying 2D materialwas quantified with AFM, XPS, photoluminescence and Raman spectroscopy. This research has led to a process where multi-layer WS₂ can be functionalized and a high-quality gate dielectric can be successfully deposited on the TMD channel materials.