skip to main content

Title: Physically informed machine-learning algorithms for the identification of two-dimensional atomic crystals

After graphene was first exfoliated in 2004, research worldwide has focused on discovering and exploiting its distinctive electronic, mechanical, and structural properties. Application of the efficacious methodology used to fabricate graphene, mechanical exfoliation followed by optical microscopy inspection, to other analogous bulk materials has resulted in many more two-dimensional (2D) atomic crystals. Despite their fascinating physical properties, manual identification of 2D atomic crystals has the clear drawback of low-throughput and hence is impractical for any scale-up applications of 2D samples. To combat this, recent integration of high-performance machine-learning techniques, usually deep learning algorithms because of their impressive object recognition abilities, with optical microscopy have been used to accelerate and automate this traditional flake identification process. However, deep learning methods require immense datasets and rely on uninterpretable and complicated algorithms for predictions. Conversely, tree-based machine-learning algorithms represent highly transparent and accessible models. We investigate these tree-based algorithms, with features that mimic color contrast, for automating the manual inspection process of exfoliated 2D materials (e.g., MoSe2). We examine their performance in comparison to ResNet, a famous Convolutional Neural Network (CNN), in terms of accuracy and the physical nature of their decision-making process. We find that the decision trees, gradient boosted decision more » trees, and random forests utilize physical aspects of the images to successfully identify 2D atomic crystals without suffering from extreme overfitting and high training dataset demands. We also employ a post-hoc study that identifies the sub-regions CNNs rely on for classification and find that they regularly utilize physically insignificant image attributes when correctly identifying thin materials.

« less
; ; ; ;
Award ID(s):
1846747 1749774
Publication Date:
Journal Name:
Scientific Reports
Nature Publishing Group
Sponsoring Org:
National Science Foundation
More Like this
  1. Two-dimensional (2D) materials assembled into van der Waals (vdW) heterostructures contain unlimited combinations of mechanical, optical, and electrical properties that can be harnessed for potential device applications. Critically, these structures require control over interfacial adhesion for enabling their construction and have enough integrity to survive industrial fabrication processes upon their integration. Here, we promptly determine the adhesion quality of various exfoliated 2D materials on conventional SiO 2 /Si substrates using ultrasonic delamination threshold testing. This test allows us to quickly infer relative substrate adhesion based on the percent area of 2D flakes that survive a fixed time in an ultrasonic bath, allowing for control over process parameters that yield high or poor adhesion. We leverage this control of adhesion to optimize the vdW heterostructure assembly process, where we show that samples with high or low substrate adhesion relative to each other can be used selectively to construct high-throughput vdW stacks. Instead of tuning the adhesion of polymer stamps to 2D materials with constant 2D-substrate adhesion, we tune the 2D-substrate adhesion with constant stamp adhesion to 2D materials. The polymer stamps may be reused without any polymer melting steps, thus avoiding high temperatures (<120 °C) and allowing for high-throughput production. We showmore »that this procedure can be used to create high-quality 2D twisted bilayer graphene on SiO 2 /Si, characterized with atomic force microscopy and Raman spectroscopic mapping, as well as low-angle twisted bilayer WSe 2 on h-BN/SiO 2 /Si, where we show direct real-space visualization of moiré reconstruction with tilt-angle dependent scanning electron microscopy.« less
  2. Abstract

    Advances in nanoscience have enabled the synthesis of nanomaterials, such as graphene, from low‐value or waste materials through flash Joule heating. Though this capability is promising, the complex and entangled variables that govern nanocrystal formation in the Joule heating process remain poorly understood. In this work, machine learning (ML) models are constructed to explore the factors that drive the transformation of amorphous carbon into graphene nanocrystals during flash Joule heating. An XGBoost regression model of crystallinity achieves anr2score of 0.8051 ± 0.054. Feature importance assays and decision trees extracted from these models reveal key considerations in the selection of starting materials and the role of stochastic current fluctuations in flash Joule heating synthesis. Furthermore, partial dependence analyses demonstrate the importance of charge and current density as predictors of crystallinity, implying a progression from reaction‐limited to diffusion‐limited kinetics as flash Joule heating parameters change. Finally, a practical application of the ML models is shown by using Bayesian meta‐learning algorithms to automatically improve bulk crystallinity over many Joule heating reactions. These results illustrate the power of ML as a tool to analyze complex nanomanufacturing processes and enable the synthesis of 2D crystals with desirable properties by flash Joule heating.

  3. Abstract

    Van der Waals (vdW) materials are an indispensable part of functional device technology due to their versatile physical properties and ease of exfoliating to the low‐dimensional limit. Among all the compounds investigated so far, the search for magnetic vdW materials has intensified in recent years, fueled by the realization of magnetism in 2D. However, metallic magnetic vdW systems are still uncommon. In addition, they rarely host high‐mobility charge carriers, which is an essential requirement for high‐speed electronic applications. Another shortcoming of 2D magnets is that they are highly air sensitive. Using chemical reasoning, TaCo2Te2is introduced as an air‐stable, high‐mobility, magnetic vdW material. It has a layered structure, which consists of Peierls distorted Co chains and a large vdW gap between the layers. It is found that the bulk crystals can be easily exfoliated and the obtained thin flakes are robust to ambient conditions after 4 months of monitoring using an optical microscope. Signatures of canted antiferromagntic behavior are also observed at low‐temperature. TaCo2Te2shows a metallic character and a large, nonsaturating, anisotropic magnetoresistance. Furthermore, the Hall data and quantum oscillation measurements reveal the presence of both electron‐ and hole‐type carriers and their high mobility.

  4. Abstract

    Individual atomic defects in 2D materials impact their macroscopic functionality. Correlating the interplay is challenging, however, intelligent hyperspectral scanning tunneling spectroscopy (STS) mapping provides a feasible solution to this technically difficult and time consuming problem. Here, dense spectroscopic volume is collected autonomously via Gaussian process regression, where convolutional neural networks are used in tandem for spectral identification. Acquired data enable defect segmentation, and a workflow is provided for machine-driven decision making during experimentation with capability for user customization. We provide a means towards autonomous experimentation for the benefit of both enhanced reproducibility and user-accessibility. Hyperspectral investigations on WS2sulfur vacancy sites are explored, which is combined with local density of states confirmation on the Au{111} herringbone reconstruction. Chalcogen vacancies, pristine WS2, Au face-centered cubic, and Au hexagonal close-packed regions are examined and detected by machine learning methods to demonstrate the potential of artificial intelligence for hyperspectral STS mapping.

  5. BACKGROUND Optical sensing devices measure the rich physical properties of an incident light beam, such as its power, polarization state, spectrum, and intensity distribution. Most conventional sensors, such as power meters, polarimeters, spectrometers, and cameras, are monofunctional and bulky. For example, classical Fourier-transform infrared spectrometers and polarimeters, which characterize the optical spectrum in the infrared and the polarization state of light, respectively, can occupy a considerable portion of an optical table. Over the past decade, the development of integrated sensing solutions by using miniaturized devices together with advanced machine-learning algorithms has accelerated rapidly, and optical sensing research has evolved into a highly interdisciplinary field that encompasses devices and materials engineering, condensed matter physics, and machine learning. To this end, future optical sensing technologies will benefit from innovations in device architecture, discoveries of new quantum materials, demonstrations of previously uncharacterized optical and optoelectronic phenomena, and rapid advances in the development of tailored machine-learning algorithms. ADVANCES Recently, a number of sensing and imaging demonstrations have emerged that differ substantially from conventional sensing schemes in the way that optical information is detected. A typical example is computational spectroscopy. In this new paradigm, a compact spectrometer first collectively captures the comprehensive spectral information ofmore »an incident light beam using multiple elements or a single element under different operational states and generates a high-dimensional photoresponse vector. An advanced algorithm then interprets the vector to achieve reconstruction of the spectrum. This scheme shifts the physical complexity of conventional grating- or interference-based spectrometers to computation. Moreover, many of the recent developments go well beyond optical spectroscopy, and we discuss them within a common framework, dubbed “geometric deep optical sensing.” The term “geometric” is intended to emphasize that in this sensing scheme, the physical properties of an unknown light beam and the corresponding photoresponses can be regarded as points in two respective high-dimensional vector spaces and that the sensing process can be considered to be a mapping from one vector space to the other. The mapping can be linear, nonlinear, or highly entangled; for the latter two cases, deep artificial neural networks represent a natural choice for the encoding and/or decoding processes, from which the term “deep” is derived. In addition to this classical geometric view, the quantum geometry of Bloch electrons in Hilbert space, such as Berry curvature and quantum metrics, is essential for the determination of the polarization-dependent photoresponses in some optical sensors. In this Review, we first present a general perspective of this sensing scheme from the viewpoint of information theory, in which the photoresponse measurement and the extraction of light properties are deemed as information-encoding and -decoding processes, respectively. We then discuss demonstrations in which a reconfigurable sensor (or an array thereof), enabled by device reconfigurability and the implementation of neural networks, can detect the power, polarization state, wavelength, and spatial features of an incident light beam. OUTLOOK As increasingly more computing resources become available, optical sensing is becoming more computational, with device reconfigurability playing a key role. On the one hand, advanced algorithms, including deep neural networks, will enable effective decoding of high-dimensional photoresponse vectors, which reduces the physical complexity of sensors. Therefore, it will be important to integrate memory cells near or within sensors to enable efficient processing and interpretation of a large amount of photoresponse data. On the other hand, analog computation based on neural networks can be performed with an array of reconfigurable devices, which enables direct multiplexing of sensing and computing functions. We anticipate that these two directions will become the engineering frontier of future deep sensing research. On the scientific frontier, exploring quantum geometric and topological properties of new quantum materials in both linear and nonlinear light-matter interactions will enrich the information-encoding pathways for deep optical sensing. In addition, deep sensing schemes will continue to benefit from the latest developments in machine learning. Future highly compact, multifunctional, reconfigurable, and intelligent sensors and imagers will find applications in medical imaging, environmental monitoring, infrared astronomy, and many other areas of our daily lives, especially in the mobile domain and the internet of things. Schematic of deep optical sensing. The n -dimensional unknown information ( w ) is encoded into an m -dimensional photoresponse vector ( x ) by a reconfigurable sensor (or an array thereof), from which w′ is reconstructed by a trained neural network ( n ′ = n and w′   ≈   w ). Alternatively, x may be directly deciphered to capture certain properties of w . Here, w , x , and w′ can be regarded as points in their respective high-dimensional vector spaces ℛ n , ℛ m , and ℛ n ′ .« less