The nature of the interface in lateral heterostructures of 2D monolayer semiconductors including its composition, size, and heterogeneity critically impacts the functionalities it engenders on the 2D system for next-generation optoelectronics. Here, we use tipenhanced Raman scattering (TERS) to characterize the interface in a single-layer MoS 2 /WS 2 lateral heterostructure with a spatial resolution of 50 nm. Resonant and nonresonant TERS spectroscopies reveal that the interface is alloyed with a size that varies over an order of magnitudefrom 50 to 600 nmwithin a single crystallite. Nanoscale imaging of the continuous interfacial evolution of the resonant and nonresonant Raman spectra enables the deconvolution of defect activation, resonant enhancement, and material composition for several vibrational modes in single-layer MoS 2 , Mo x W 1−x S 2 , and WS 2 . The results demonstrate the capabilities of nanoscale TERS spectroscopy to elucidate macroscopic structure−property relationships in 2D materials and to characterize lateral interfaces of 2D systems on length scales that are imperative for devices.

Recent studies on atomically thin lateral heterostructures have demonstrated the formation of complex interfaces that can be exploited for tailoring the properties of 2D semiconductor systems for optoelectronic applications. In order to understand the compositional disorder and the resulting optical and electronic properties at these interfaces, tip-enhanced Raman scattering (TERS) imaging and spectroscopy have been used to characterize 2D lateral heterostructures at different sites across the interface, showing a continuous evolution of the Raman-active modes when transitioning from one pristine material to the other. Here, we use density functional theory (DFT) for calculating the evolution of vibrational modes, nonresonant Raman spectra, optical absorption, and electronic structure of 1L-MoS2, WS2, and Mo x W1–x S2 alloys. The calculations reproduce the evolution of the Raman modes observed in the TERS measurements, explicitly confirm the extended alloyed nature of the heterostructure interface, and provide a direct mapping of the TERS spectra to local nanoscale alloy compositions. We further elucidate how S vacancies activate a defect mode in these systems and how the mode evolves with composition, providing a second direct comparison to the nonresonant TERS measurements. Leveraging the explicit determination of the composition, we calculate how the realistic interfacial composition affects the band alignment between the two 2D materials. Our study serves as a roadmap for how the same computational approach can predict the compositional-dependent properties of additional lateral heterostructures, providing a valuable resource for quantitatively interpreting state-of-the-art nanoscale characterization measurements such as TERS imaging and spectroscopy.

Computer vision algorithms can quickly analyze numerous images and identify useful information with high accuracy. Recently, computer vision has been used to identify 2D materials in microscope images. 2D materials have important fundamental properties allowing for their use in many potential applications, including many in quantum information science and engineering. One such material is hexagonal boron nitride (hBN), an isomorph of graphene with a very indistinguishable layered structure. In order to use these materials for research and product development, the most effective method is mechanical exfoliation where single-layer 2D crystallites must be prepared through an exfoliation procedure and then identified using reflected light optical microscopy. Performing these searches manually is a time-consuming and tedious task. Deploying deep learning-based computer vision algorithms for 2D material search can automate the flake detection task with minimal need for human intervention. In this work, we have implemented a new deep learning pipeline to classify crystallites of hBN based on coarse thickness classifications in reflected-light optical micrographs. We have used DetectoRS as the object detector and trained it on 177 images containing hexagonal boron nitride (hBN) flakes of varying thickness. The trained model achieved a high detection accuracy for the rare category of thin flakes ($$<50$$ < 50 atomic layers thick). Further analysis shows that our proposed pipeline could be generalized to various microscope settings and is robust against changes in color or substrate background.

Computer vision algorithms can quickly analyze numerous images and identify useful information with high accuracy. Recently, computer vision has been used to identify 2D materials in microscope images. 2D materials have important fundamental properties allowing for their use in many potential applications, including many in quantum information science and engineering. In order to use these materials for research and product development, single-layer 2D crystallites must be prepared through an exfoliation procedure and then identified using reflected light optical microscopy. Performing these searches manually is a time-consuming and tedious task. Deploying deep learning-based computer vision algorithms for 2D material search can automate the flake detection task with minimal need for human intervention. In this work, we have implemented a new deep learning pipeline to classify crystallites of 2D materials based on coarse thickness classifications in reflected-light optical micrographs. We have used DetectorRS as the object detector and trained it on 177 images containing hexagonal boron nitride (hBN) flakes of varying thickness. The trained model achieved a high detection accuracy for the rare category of thin flakes (<50 atomic layers thick). Further analysis shows that our proposed pipeline could be generalized to various microscope settings and is robust against changes in color or substrate background.

Computer vision algorithms can quickly analyze numerous images and identify useful information with high accuracy. Recently, computer vision has been used to identify 2D materials in microscope images. 2D materials have important fundamental properties allowing for their use in many potential applications, including many in quantum information science and engineering. In order to use these materials for research and product development, single-layer 2D crystallites must be prepared through an exfoliation procedure and then identified using reflected light optical microscopy. Performing these searches manually is a time-consuming and tedious task. Deploying deep learning-based computer vision algorithms for 2D material search can automate the flake detection task with minimal need for human intervention. In this work, we have implemented a new deep learning pipeline to classify crystallites of 2D materials based on coarse thickness classifications in reflected-light optical micrographs. We have used DetectorRS as the object detector and trained it on 177 images containing hexagonal boron nitride (hBN) flakes of varying thickness. The trained model achieved a high detection accuracy for the rare category of thin flakes (< 50 atomic layers thick). Further analysis shows that our proposed pipeline could be generalized to various microscope settings and is robust against changes in color or substrate background.