Mechanical cleaning of graphene using in situ electron microscopy

Maintaining cleanliness is crucial across various scientific fields, from semiconductor manufacturing to forensic science. The challenge is amplified with two-dimensional (2D) materials like graphene, where even a single adatom can drastically affect properties and experiments. Existing graphene cleaning methods, including heating, plasma treatment, laser cleaning, chemical activation, and current-driven cleaning, haven't achieved atomic-scale, site-specific cleanliness. While some approaches, like high-temperature aging in ultra-high vacuum, yield improved results, they lack control over the size and location of clean areas. This research addresses this gap by introducing a novel site-specific mechanical cleaning method combined with in situ electron microscopy, aiming for atomic-scale cleanliness and controlled area cleaning.

The paper reviews several existing methods for cleaning graphene, including heating, plasma treatment, laser cleaning, chemical activation, and current-driven cleaning. It notes that while these methods have shown some success, none have achieved site-specific, atomic-scale cleanliness. Attempts at mechanical cleaning using atomic force microscopy (AFM) also fell short of this goal. The authors highlight the limitations of existing techniques in controlling the size and location of cleaned areas and their potential to damage sensitive samples. The literature review sets the stage for the introduction of the authors' innovative approach.

The researchers developed a site-specific mechanical cleaning approach integrated with both scanning electron microscopy (SEM) and transmission electron microscopy (TEM). The technique mimics the action of a broom, using a fine-metal tip (tungsten in SEM, gold in TEM) controlled by a piezo-driven nanopositioning system to sweep across the graphene surface, removing loosely bound contaminants. In SEM, two independent manipulators allow simultaneous cleaning of both sides of the suspended membrane. The TEM setup utilizes a single manipulator with a double tip for the same purpose. The process relies on the difference in bond strength between strong intralayer graphene bonds and weaker van der Waals bonds between contaminants and graphene. Force measurements during cleaning using a spring table system determined the lateral (~76 nN) and normal (~50 nN minimum) forces required. Raman spectroscopy confirmed no additional defects after cleaning. High-resolution TEM imaging and electron energy loss spectroscopy (EELS) were used to verify atomic-scale cleanliness and identify contaminants (primarily PMMA residuals and airborne hydrocarbons). Experiments exploring recontamination were conducted both ex situ (exposure to ambient conditions) and in situ (under electron beam irradiation). To study in situ recontamination mechanisms, time-lapse TEM imaging was performed at varying scan speeds, allowing the researchers to investigate the influence of surface diffusion and electron-beam-induced cross-linking on contaminant buildup. Finally, the method was used for the in situ synthesis of a nanocrystalline graphene layer, using copper(II)-tetraphenylporphyrin as a precursor molecule. The deposition, cleaning, and electron-beam-induced growth of the new graphene layer were all observed in situ.

The in situ mechanical cleaning method successfully removed contamination from both sides of suspended graphene membranes, achieving atomic-scale cleanliness in areas of several µm² within minutes. The process doesn't involve thermal energy or reactive species. Force measurements showed that cleaning requires lateral forces around 76 nN and normal forces above 50 nN, but well below the graphene membrane rupture threshold. Raman spectroscopy confirmed that cleaning didn't introduce additional defects. EELS analysis revealed the removal of PMMA residuals and a shift in the plasmon loss peak indicative of pristine graphene. Studies of recontamination showed that surface diffusion is the main factor for in situ contamination during electron microscopy. The in situ observation of recontamination using time-lapse imaging revealed that contamination seeds nucleate and grow, mainly at defects. The rate of contamination is strongly dependent on the electron beam scan speed, with fast scans limiting contamination buildup compared to slow scans. The analysis revealed a diffusion coefficient of 4.88 × 10⁻¹² m² s⁻¹, suggesting larger, less mobile molecules are responsible for recontamination. Finally, the researchers demonstrated the electron-beam-assisted synthesis of a nanocrystalline graphene layer by supplying copper(II)-tetraphenylporphyrin to a cleaned area. This second layer exhibited small crystalline patches with different orientations, separated by grain boundaries with typical dislocation structures.

The findings address the long-standing challenge of achieving atomic-scale cleanliness in graphene samples. The developed mechanical cleaning method offers site-specific control and avoids the limitations of existing techniques. The identification of surface diffusion as the main recontamination mechanism in electron microscopy provides important insights into sample handling and experimental design. The ability to perform in situ synthesis of nanocrystalline graphene demonstrates the versatility of the method. The results have significant implications for the study of 2D materials and high-vacuum environments, opening possibilities for precisely controlled experiments and synthesis at the atomic level.

This research successfully demonstrates a novel in situ mechanical cleaning method for achieving atomic-scale cleanliness in 2D materials, specifically graphene. The technique provides site-specific control, avoids the use of high temperatures or reactive species, and is compatible with electron microscopy. The characterization of recontamination mechanisms highlights the importance of surface diffusion in electron microscopy and enables better control over sample preparation. The demonstration of in situ nanocrystalline graphene synthesis opens new avenues for controlled material growth and modification at the atomic scale. Future research could explore the applications of this technique to other 2D materials and investigate advanced nanofabrication techniques based on this controlled surface manipulation.

While the method shows high success rates on freshly prepared membranes, the presence of defects or pre-existing strong contamination could reduce the effectiveness of cleaning and potentially damage the sample. The study focuses primarily on graphene, although it is shown to work on MoS2 as well; further investigations on the efficacy with other 2D materials are needed. The exact chemical composition of recontaminating molecules in situ was not fully determined, although PMMA fragments were a strong candidate. More research into the precise mechanisms of electron-beam-induced nanocrystalline graphene growth is also warranted.