Deep learning based atomic defect detection framework for two-dimensional materials

Abstract

Defects to popular two-dimensional (2D) transition metal dichalcogenides (TMDs) seriously lower the efficiency of field-effect transistor (FET) and depress the development of 2D materials. These atomic defects are mainly identified and researched by scanning tunneling microscope (STM) because it can provide precise measurement without harming the samples. The long analysis time of STM for locating defects in images has been solved by combining feature detection with convolutional neural networks (CNN). However, the low signal-noise ratio, insufficient data, and a large amount of TMDs members make the automatic defect detection system hard to be applied. In this study, we propose a deep learning-based atomic defect detection framework (DL-ADD) to efficiently detect atomic defects in molybdenum disulfide (MoS2) and generalize the model for defect detection in other TMD materials. We design DL-ADD with data augmentation, color preprocessing, noise filtering, and a detection model to improve detection quality. The DL-ADD provides precise detection in MoS2 (F2-scores is 0.86 on average) and good generality to WS2 (F2-scores is 0.89 on average).

Advancements in defect engineering of two-dimensional nanomaterial-based membranes for enhanced gas separation

“…It can be used for testing in terms of deep defects in two-dimensional heterogeneous structures. 71 Although the STM method has excellent analytical performance, it is not sufficient for some applications due to its limited resolution.…”

Soft Sensing Modeling of Penicillin Fermentation Process Based on Local Selection Ensemble Learning

“…However, due to the strong nonlinear relationship between input and output caused by the high-level semantic concept of high-dimensional input represented by multiple targets, the soft sensor model cannot perform well in data prediction 24 . The study 25 proposes a deep learning-based atomic defect detection framework (DL-ADD) to efficiently detect atomic defects in molybdenum disulfide (MoS2) and generalize the model for defect detection in other TMD materials. A large-scale dataset of 3D solar magnetic fields of active regions is built by using the nonlinear force-free magnetic field (NLFFF) extrapolation from vector magnetograms of Helioseismic and Magnetic Imager (HMI) on the Solar Dynamics Observatory (SDO) 26 .…”

Probe conditioning via convolution neural network for scanning probe microscopy automation

“…1,2) The inherent instability of surfaces caused by thermal motion, particularly at room temperature, can be exploited to manipulate dopant atoms on semiconductor surfaces. 3) This technique is expected to synergize with SPM automation techniques, such as the use of algorithms or machine learning (ML) for instrument operation, [4][5][6] thereby eliminating the need for manual intervention. Through the automated positioning of foreign atoms on a Si surface, this advancement holds the potential for achieving large-scale selective doping and bringing in a new era in the development of smaller and faster electronic devices.…”