Transition Metal Dichalcogenide Defect Functionalization with Magnetic Impurities—Defect Introduction and Identification

Two-dimensional transition metal dichalcogenides (2D TMDs) are a remarkable class of materials with many applications, including electronic and optoelectronic devices, quantum information systems, and new catalytic materials. In particular, the introduction of zero-dimensional defects exhibit strong confinement and when combining 3<i>d</i> transition metal nanostructures with monolayer TMDs, magnetic interactions may allow spin-valley modulation. 2D TMDs have the benefit of large synthetic variability and the possibility of bottom-up or top-down impurity incorporation, which facilitates precise defect engineering. However, for both understanding how defects alter the material properties, and how engineered defects introduce specific functionalities, it is critical to study the correlation between atomic-scale morphology, altered electronic structure, and macroscopic functionality that is illusive for many of these systems. High resolution scanning probe microscopy (SPM) techniques have the unique ability to examine the structural, electronic, and optoelectronic properties of these materials on the atomic length scale. Previously, we have extensively used low temperature, high resolution SPM to study both intrinsic and induced point defects in TMDs, predominantly in monolayer WS₂, and here, we utilize scanning tunneling microscopy and spectroscopy (STM/S) to understand the electronic and structural properties and combine our experimental results with <i>ab initio</i> calculations to confirm defect identity. We are able to image and identify the adsorbed system at 4 K and show magnetic contrast between defective states due to the nature of our STM tip after preparation. Previously, the introduction of cobalt clusters on TMD by thermal deposition, and subsequent atomic manipulation of adsorbed atoms has been limited to 77 K. However, we are able to show both theoretically and experimentally that single-atom adsorbed states at liquid helium temperatures produce different magnetic anisotropy energies that contribute to the visualized contrast with STM/S. We also employ autonomous exploration measurement techniques, where dense spectroscopic volume is collected via Gaussian process regression and convolutional neural networks are used in tandem for spectroscopic classification and subsequent image segmentation.