UC Berkeley

MoS2 is one of a class of layered, semiconducting materials known as the transition metal dichalcogenides (TMDCs). When a monolayer is isolated from the multilayer crystal, the electronic properties of the TMDC can undergo a dramatic shift. In MoS2, the energy band gap changes from an indirect gap of 1.2 eV in the bulk to a direct gap of 1.8 eV in the monolayer. This intriguing feature has driven, in part, the recent surge in investigations into MoS2 as a candidate for atomically thin electronic and photonic devices. The last fifteen years have seen an accelerating effort to exfoliate large, high-quality monolayers of MoS2 and the other TMDCs, both from their natural crystalline forms and from lab-grown multilayers. The broader context of this dissertation is this effort to grow, exfoliate, characterize, and construct simple devices from MoS2 and WS2, including both natural and engineered bi-layers and heterostructures. Motivating the close study of Au on MoS2, it is noted that significant advances in selective monolayer TMDC exfoliation were achieved utilizing a Au `handle' layer and a compliant temporary `superstrate.'

The gold-assisted high-yield exfoliation of lithographically defined features motivates the study of the mechanical and structural relationship between Au layers and few-layer MoS2. Using a 100-nm layer of deposited Au as a handle layer, large areas of monolayer MoS2 are successfully exfoliated from a multilayer stack and characterized using photoluminescence spectroscopy and atomic force microscopy. STEM characterization identifies the preferred (111) rotational alignment of epitaxially grown Au on the (0001) basal plane of MoS2. While Au nuclei that form on the surface of MoS2 could be initially strained to form a coherent interface when they remain below a critical size, in Au crystallites with an approximate thickness of 1-10 nm and a median diameter of 15 nm, the underlying MoS2 substrate is not found to be strained when measured using nanobeam electron diffraction.

The deformation phenomena of the epitaxial Au-MoS2 system, and approaches to elucidating them, are explored in this dissertation using transmission electron microscopy techniques, particularly, 4D Scanning Transmission Electron Microscopy (4D-STEM). 4D-STEM involves recording a grid of nanobeam electron diffraction (NBED) patterns over a two-dimensional sample area, and the result is a 2D array of 2D diffraction patterns, comprising a 4D dataset. MoS2 diffraction patterns are compared to their corresponding patterns in a generated database of diffraction patterns simulated using the PRISM multislice algorithm. The thickness of deposited Au is approximated using high-angle annular dark field (HAADF) images of the Au-MoS2 system. Using the Z-contrast mechanism of the HAADF images, the thickness of a nanoporous Au film is estimated to be between 1-10 nm, slightly lower than the intended mean thickness of 15 nm of Au calibrated during e-beam deposition using quartz crystal microbalancing (QCM). Accordingly, variations in Au thickness are discussed in the context of the e-beam evaporation of Au onto MoS2. Finally, a workflow is outlined for examining the Moiré fringes and patterns in experimental HAADF images in order to identify rotational and lattice alignments.

These analyses reveal the dearth of measurable strains in the epitaxial Au-MoS2 system, the presence of rippling in ultra-thin MoS2 with an approximate radius of curvature of 50 nm, the self-avoiding growth morphology of thin-film Au, and the rotation of up to 3° between these materials. Data from nanobeam electron diffraction and dark-field STEM imaging show that gold typically grows on MoS2(0001) with its [111] vector normal to the interface. However, symmetric axial rotations about this vector direction with typical magnitudes of 1-2° can occasionally arise in Au domains, which range from 5-15 nm in diameter. The use of evaporated Au as a compliant handle layer in TMDC exfoliation processes is confirmed to be empirically useful for increasing the size and selectivity of single layers. Exfoliated mono- and bi-layers produced using this method are evaluated using photoluminescence spectroscopy, and sulfur vacancies are repaired using a TFSI superacid treatment which improves the photoluminescence quantum yield by a factor of 25 and 100 for MoS2 and WS2, respectively. These results have implications for future architectures and processes for nanoelectronics based on monolayer TMDCs.