Chemical vapour deposition of tungsten disulphide aided by thermodynamic and density functional theory modelling of growth.

Includes bibliographic references.

PhD;2023;Centre for Nanoscience and Engineering.

Transition metal dichalcogenides (TMDs) are versatile materials that offer a wide variety of applications in electronics, sensing, energy, etc. Tungsten disulphide (WS2) is a layered semiconductor TMD with impressive properties such as a tuneable bandgap, high on/off ratio, sub-nanometer layer thickness and electrocatalytic and sensing capabilities. These make it a good candidate for channel material, photodetectors, window layers in solar cells, electrode materials for catalysis, and sensing medium for gas sensors. Bottom-up synthesis techniques such as chemical vapour deposition (CVD), physical vapour transport (PVT), and atomic layer deposition (ALD) are commonly used to synthesize WS2. Unlike exfoliation from bulk, these synthesis techniques offer scalability of process as they ensure better control over morphology and uniformity of grown material. For growth of WS2 and other TMDs, a PVT technique using oxide precursors is commonly employed. However, as the oxide precursor must be sealed inside the furnace before starting the process, control over growth is compromised. In contrast, a true CVD method, where all precursors are in the vapour form, and so can be kept outside the reactor in tanks or bubblers and released into the reactor as and when required, permits more real-time control. In this work CVD synthesis of WS2 was achieved using tungsten hexacarbonyl (W(CO)6), a solid with a low sublimation point, which can be vapourized from an external bubbler, and hydrogen sulphide gas (H2S). Predominantly two different morphologies of WS2 were discovered during the course of this work at two different temperature conditions – Laterally oriented WS2 flakes at higher temperatures (> 800 ⁰C) and vertically oriented WS2 flakes at lower temperatures. Each of these are interesting for a separate set of qualities and growths of these were hence explored. In the first part of this work, synthesis of flat WS2 flakes oriented lateral to the substrate surface was aimed. Flat WS2 films with large flake sizes are desired for electronic and optoelectronic device applications. Large flake sizes are necessary as they reduce grain boundary defect densities that will otherwise degrade electrical quality of the film. As we embarked on CVD synthesis of WS2 using this multi-component W-C-O-H-S chemistry, one of the primary challenges encountered was the possibility of oxides, carbides and other sulphides forming as impurities alongside WS2. Predictive thermodynamic modelling was used to circumvent this obstacle, and successfully predict a window of growth conditions where pure WS2 grew. The modelling results showed thermodynamic behaviour that is identical to that of the similar Mo-C-O-H-S system, which was used to grow very large islands (>100 µm) of Molybdenum disulphide (MoS2) in a previously existing work. Further to this, flat WS2 films were successfully grown. Different sets of conditions from the predicted windows were explored to optimize growth and maximize flake sizes. The maximum flake size that was obtained was ~500 nm, which stood in stark contrast with the ~137 um large MoS2 flakes grown with the Mo-C-O-H-S system. Large differences in island sizes pointed to underlying kinetic differences between growth of WS2 and MoS2 in spite of the thermodynamic similarities. Understanding these became imperative to furthering knowledge about growth using this system. The fundamental physico-chemical quantities that determine nucleation densities and growth rates in CVD synthesis are supersaturation and the barriers to diffusion and desorption of adatoms over the growth surface. While supersaturation can be controlled using external handles such as precursor flows-rates, pressure and temperature, the barriers to desorption and diffusion are inherent to a particular set of adatoms and growth surface. In order to study growth, these barriers need to be found. Density functional theory (DFT) modelling was done to calculate diffusion energy and desorption energy barriers to growth of MoS2 and WS2 on sapphire surface which is a commonly used growth surface. The results of DFT calculations were used to predict nucleation densities and growth rates of these compounds. It was predicted that a larger diffusion barrier for W adatoms caused slower growth rates and larger saturation nucleation densities in WS2, which matched with experimental observations. Further insights derived from this investigation allowed a better understanding of the difference between growth of MoS2 and WS2. In the final part of this work, growth of vertically standing WS2 was investigated. Vertically standing WS2 grows in large swathes of parameter space of the W-C-O-H-S chemistry. This morphology allows packing of more flakes within a given substrate area. Here, WS2 flakes bend up and grow at an angle to the substrate surface, with edges exposed outwardly. Consequently, vertically standing morphology enhances exposed basal plane surface area and flake edge densities. Large surface area is desired in applications such as sensing, and large density of catalytically active edges can boost its catalytic capabilities in hydrogen evolution, H2S decomposition, etc. In a bid to enhance vertically standing WS2 for its prominent applications flake, densities, exposed edge densities, and basal plane surface area were maximized by changing conditions of pressure, temperature, H2S, and hexacarbonyl flows. It has been reported that defects in the basal plane of WS2, such as sulphur vacancies, are catalytically active and also enhance sensing capabilities. Successfully engineering sulphur vacancies into the abundant basal planes of vertically standing WS2 flakes may improve their functionality. Hence, additionally, stoichiometry tuning of flakes was done. For this, hydrogen annealing was used where S from the solid can be removed as hydrogen sulphide, leaving vacancy defects. Annealing conditions were determined prior to runs by thermodynamic modelling and by calculating most stable sulphur vacancy concentrations at different sets of annealing conditions. For this, Brouwer diagrams depicting dependence of S vacancy concentrations on partial pressures of H2S and H2 were constructed. To the best of this author’s knowledge, this is the first time Brouwer diagrams have been made for WS2-H2S-H2 system. Based on results of calculations, annealing was done at different hydrogen pressures, furnace temperatures, and annealing times. Stoichiometries were measured using XPS and optimized by changing annealing conditions. At predicted conditions, good results were obtained at a few minutes of annealing duration, which is significantly smaller annealing times compared to existing reports. Thus, through this work, CVD growth of WS2, study of growth mechanisms and subsequent tuning of morphology and stoichiometry towards enhancement of its features were achieved.