In their two-dimensional (2D), monolayer form, transition metal dichalcogenides (TMDs) undergo an indirect-to-direct band-gap transition, which greatly enhances their radiative efficiency and is of high interest for optoelectronics. Furthermore, 2D-TMDs present clear evidence of polymorphism: in their most stable phase, the 2H structure, they are indeed direct-gap semiconductors, but their (semi)metallic 1T' phase can be easily stabilized by chemical treatments, mechanical deformations, or laser/electron-beam exposure. The possibility to locally induce an insulator-to-metal transition in 2D-TMDs has obvious technological implications, with tantalizing prospects for the realization of ultrathin devices. In this project, we propose to selectively trigger the 2H-1T' transition by exposing the sample surface to a controlled flux of hydrogen ions. Indeed, H chemisorption plays a key role in the chemical treatments that stabilize the 1T' phase. Moreover, H exposure leads to the formation of atomically thin micro- and nanobubbles on the surface of bulk TMDs. The formation of these bubbles, whose size and position can be controlled via electron-beam lithography, results in (1) a local exfoliation of the uppermost TMD layer, i.e., in a huge boost of the radiative efficiency and (2) in a significant lattice expansion, allowing for the introduction of large, site-controlled mechanical stresses in the sample. First of all, such stresses will be exploited to further manipulate the (strain-dependent) 2H-1T' transition. Also, the effects of the giant, strain-induced pseudo-magnetic fields associated with nanobubble formation should allow for the realization of valleytronic elements such as valley filters and beam splitters. This is particularly important in the context of 2H-1T' hybrid devices, where topological metallic states are expected to form at the edges of 1T' domains and would thus represent preferential channels for topologically protected valley transport.
The present project aims at controlling the phase and electronic properties of 2D TMDs on a nanometer scale, providing the perfect playground for addressing, e.g., the electronic states that we expect to form at the domain walls between different phases (see O.3 in the previous section). Notably, it has been shown that different layer stacking in bilayer graphene can lead to topological valley transport at the domain walls [Ju15]. TMDs are also candidates for hosting topologically protected electronic states, in particular at the edge of 1T' domains. Given that nanobubbles have been proposed for the realization of valley filters and beam splitters, the nanofabrication tools developed during the present project might lay the foundations for the realization of a new generation of valleytronic devices.
At a more fundamental level, up to now several aspects of the behavior of TMDs are unexplored and poorly understood: among others, (i) an experimental proof of the role of H adsorbates on stabilizing the sample phase [Cal13] is still lacking, and (ii) the microscopic mechanism leading to the formation of nanobubbles is largely unknown. On the other hand, the experimental verification of the effects of the large elastic strains induced by the formation of nanobubbles on (iii) the emergence of giant pseudomagnetic fields [Caz14, Set16, Och17] and (iv) the stabilization of the 1T' polytype with respect to the 2H one [Due14] would be crucial for the development of the proposed applications of such effects. Within the present project, the properties of the nano-engineered samples will be studied by means of micro Raman and micro-PL -both available at ONSM- and micro- and nano-IR spectroscopy (performed at IRS). The 2H-1T' transition should induce dramatic changes in the micro-PL signal, whereas the 1T' phase is characterized by fairly distinct Raman features [Cho15]. However, the lateral resolution of micro-Raman (~500 nm) does not allow for the study of the predicted metallic edge states (which should extend for ~10 nm from the edge itself).
With the nanoIR setup (see Fig. 4 in the section above), on the other hand, it is possible to overcome the limits imposed by diffraction and measure the absorbance of the sample below the AFM tip with a ~10-nm lateral resolution. The IR absorbance spectra are obtained with high sensitivity by detecting the mechanical forces exerted on the atomic force microscope tip upon light excitation. We will first aim at the identification of 1T' (semi-metallic) and 2H (insulating) regions by means of their spectroscopic signatures in the mid-IR. The possibility of studying the coexistence of deeply sub-wavelength metallic and insulating regions has been enabled only in recent years with the development of tip-enhanced near-field approaches [Ju15]. In order to study the presence of metallic edge-state in the semimetallic 1T' phase domains we will implement the nano-infrared setup by coupling it to a quantum cascade laser in the far-infrared, emitting at 4.7 THz. This non-commercial cryogenic laser is provided within an active collaboration with the ETH Zurich.
Further, the effects of bubble size on the optical properties of TMDs are still unexplored. This is particularly true for nanobubbles in the
Finally, on the theoretical side, we will use first-principles DFT calculations to interpret and guide the experimental approaches described above. In particular we will model the impact of non-uniform strain, of H chemisorption and of the presence 2H-1T' interfaces on the atomistic, electronic and vibrational properties of TMDs. Moreover, we will simulate directly the spectroscopic signatures of such modifications in the Raman, PL, and IR spectra. For IR spectroscopy we will consider both the vibrational absorption bands and the electronic ones expected at the 2H-1T'' interfaces.