Transition Metal Dichalcogenides—Self-Flux Synthesis, Characterization and Defect Computations

Two-dimensional transition metal dichalcogenides (TMDs) have emerged as a highly attractive class of materials due to their novel optical and electronic phenomena, scalability to sub-nanometer sizes, and potential for electronic and optoelectronic applications. Chemical vapor deposition (CVD) and chemical vapor transport (CVT) have been used to grow monolayers and bulk crystals, respectively, quickly and at high yields. However, these techniques produce highly defective TMDs, negatively impacting the novel properties [1-4]. An alternative synthesis method is self-flux crystal growth, which has been shown to grow bulk TMDs with point defect densities several orders of magnitude lower than those from CVD and CVT [5]. The self-flux-growth uses high-purity elemental precursors limiting impurities and occurs inside of a vacuum sealed quartz ampule. The transition metal dissolves in the excess liquid chalcogen flux at an elevated temperature, then the mixture is slowly cooled to allow for growth of high-quality crystals.<br/>In this work, we use the self-flux method to synthesize five different TMDs: MoSe₂, WSe₂, WTe₂, 2H-MoTe₂, and 1T’-MoTe₂. Temperature profiles for growth of the four stable phases are largely similar; however, for the metastable 1T’-MoTe₂ phase, the ampule is removed from the furnace at 900 °C and rapidly quenched instead of slow cooling to room temperature. The material and phase are confirmed using a combination of Raman spectroscopy and x-ray diffraction. The point defect density is determined by counting point defects in several scanning tunneling microscopy (STM) images taken from across a cleaved bulk TMD’s surface. To identify the defect chemical identities in self-flux TMDs, density functional theory calculations were used to simulate STM images for various intrinsic defects, and after comparison to experimental images, the simulations suggest the presence of chalcogen vacancies in MoSe₂ and WSe₂, and both metal and chalcogen vacancies in 2H-MoTe₂. Additionally, ab initio modeling is used in conjunction with experimental thermodynamic data to calculate temperature dependent thermal equilibrium defect densities for intrinsic defects of the four stable self-flux TMDs. In WSe₂, the thermal equilibrium defect density of selenium vacancies at the self-flux growth conditions was calculated to be two to three orders of magnitude less than experimental values determined by STM images, suggesting thermal equilibrium has not been reached, and kinetic or residual impurity factors drive increased point defect densities in self-flux TMDs.<br/><br/>[1] D. Rhodes, <i>et al.</i>, <i>Nat. Mater. </i>18 (2019), pp. 541-549, https://doi.org/10.1038/s41563-019-0366-8<br/><br/>[2] G. H. Han, <i>et al.,</i> <i>Nat. Comm. </i>6, 6128 (2015), https://doi.org/10.1038/ncomms7128<br/><br/>[3] C. H. Naylor, <i>et al.,</i> <i>Nano. Lett. </i>16, 7 (2016), pp. 4297-4304, https://doi.org/10.1021/acs.nanolett.6b01342<br/><br/>[4] C. H. Naylor, <i>et al.,</i> <i>2D Mater. </i>4 (2017), 021008, https://doi.org/10.1088/2053-1583/aa5921<br/><br/>[5] D. Edelberg, <i>et al.,</i> <i>Nano Lett. </i>19 (2019), pp. 4371-4379, https://doi.org/10.1021/acs.nanolett.9b00985