Jiadong Dan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jiadong Dan has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Ne-Te Duane Loh
National University of Singapore
4 shared publications - Stephen John Pennycook
Department of Materials Science and Engineering, University of Tennessee, Knoxville, USA
3 shared publications - Xiaoxu Zhao
School of Materials Science and Engineering, Peking University, Beijing 100871, China
2 shared publications Cheng Zhang
Dept. of Physics, National University of Singapore , Singapore
1 shared publication- Cheng. Zhang
Department of Materials Science and Engineering City University of Hong Kong Hong Kong China
1 shared publication - Chenghang Zhou
Department of Materials Science and Engineering, National University of Singapore, Singapore 117575, Singapore
1 shared publication
Latest Publications
- Open Access
- 1 January 2025
- Published by ScienceOpen
- p. e295
Less Is More And More Is Different: Learning the Emergent Language of Disorder In Complex Materials
Understanding complex materials often requires dissecting them into recurring structural motifs, which collectively dictate their physical properties. However, identifying these motifs in disordered systems remains a significant challenge, especially when the spatial arrangements of motifs lack periodicity and when intricate motif-motif interactions emerge. Traditional first-principles approaches struggle with these systems, as evidenced by fields like twistronics, where simple twists in bilayers can transform material properties drastically, making it crucial to identify relevant motifs under real-world conditions. The concept of “less is more” in relation to coarse-graining in disordered materials emphasizes how, by focusing on simplified, interpretable rules, one can capture essential patterns within structural complexity. Coarse-graining of atomic configurations into motif classes filters out extraneous noise and uninformative variations, enabling us to focus on meaningful structural features. Disordered systems, like the domain boundaries in potassium sodium niobate (KNN) piezoelectric thin films, present intricate and seemingly random structural motifs that defy straightforward reductionist modeling. Instead of mapping every detail, we prioritize certain recurring motifs and their transitions, which collectively convey the system’s fundamental characteristics while maximizing entropy—a process that coarse-graining allows. The principle of “more is different” comes into play in disordered materials when the increasing complexity of structural motifs and their hierarchical relationships leads to the discovery of emergent phenomena—properties and behaviors that cannot be predicted simply by examining individual motifs or lower-level interactions. As we incorporate more motifs and layers of hierarchy, distinct structural patterns and functional properties emerge, often as a result of configurational entropy and complex interactions across multiple length scales.
- 1 January 2025
- Published by ScienceOpen
- p. e295
Open AccessLess Is More And More Is Different: Learning the Emergent Language of Disorder In Complex Materials
Understanding complex materials often requires dissecting them into recurring structural motifs, which collectively dictate their physical properties. However, identifying these motifs in disordered systems remains a significant challenge, especially when the spatial arrangements of motifs lack periodicity and when intricate motif-motif interactions emerge. Traditional first-principles approaches struggle with these systems, as evidenced by fields like twistronics, where simple twists in bilayers can transform material properties drastically, making it crucial to identify relevant motifs under real-world conditions. The concept of “less is more” in relation to coarse-graining in disordered materials emphasizes how, by focusing on simplified, interpretable rules, one can capture essential patterns within structural complexity. Coarse-graining of atomic configurations into motif classes filters out extraneous noise and uninformative variations, enabling us to focus on meaningful structural features. Disordered systems, like the domain boundaries in potassium sodium niobate (KNN) piezoelectric thin films, present intricate and seemingly random structural motifs that defy straightforward reductionist modeling. Instead of mapping every detail, we prioritize certain recurring motifs and their transitions, which collectively convey the system’s fundamental characteristics while maximizing entropy—a process that coarse-graining allows. The principle of “more is different” comes into play in disordered materials when the increasing complexity of structural motifs and their hierarchical relationships leads to the discovery of emergent phenomena—properties and behaviors that cannot be predicted simply by examining individual motifs or lower-level interactions. As we incorporate more motifs and layers of hierarchy, distinct structural patterns and functional properties emerge, often as a result of configurational entropy and complex interactions across multiple length scales.
- Open Access
- 1 January 2025
- Published by ScienceOpen
- p. e284
Integrating Chemistry and Experimental Knowledge into Deep Learning Models for Classification of Structural Defects
Accurate identification of structural defects in materials using atomically resolved scanning transmission electron microscopy (STEM) is crucial for understanding materials structure property correlations. However, deep neural networks (DNNs) trained on simulated data for this classification task often struggle due to the issue of out-of-distribution when applied to experimental data. Moreover, nearly all previously supervised models fail to include additional information such as chemical composition and experimental conditions. To address this issue, we introduce a novel approach that employs a single multilayer perceptron (MLP) classifier trained on multislice simulated data, along with extra information, which we coin as context information. The context information, including chemical composition and experimental imaging conditions, is transformed into vectors and combined with Zernike moment representations of the image data [ 1, 2]. This ensures a single context-aware model can be trained on simulated data from various materials, as illustrated in Figure 1. To evaluate the effectiveness of this context-aware model, we train our MLP on multislice simulated images of all available 1H MX2 monolayer transition metal dichalcogenides with various dopants, where M is the transition metal and X denotes the chalcogen element. We show that the MLP model achieves 96% accuracy on both training and testing datasets. We also evaluate the model performance on experimental high-angle annular dark field (HAADF) images of monolayer WSe₂ doped Cr. The classification results are consistent with those obtained by human experts, as shown in Figure 2. Additionally, we applied an attention model to the same datasets, which produced comparable outcomes. Fig. 1. Integrating context into a unified model. a. Context information (chemical composition, beam energy, detector range in STEM) is often neglected in prior deep learning models. b. Experimental and simulated datasets are used for training. c and e. Previous models trained without context require manual interpretation of outputs. d and f. Our approach encodes context as vectors combined with Zernike moments, enabling a unified model to produce interpretable defect labels across different materials and conditions. Fig. 2. Unified MLP identifying point defects in monolayer TMDCs. a. Accuracy curves for 70 epochs reach 96% on training and test datasets, covering 13 materials, 10 dopants, 3 energies, and 2 detectors. b. Classification of an experimental HAADF image of Cr-doped WSe₂, identifying single vacancies (green), double vacancies (cyan), and dopants (red). c. Mean patches for the five defect classes.
- 1 January 2025
- Published by ScienceOpen
- p. e284
Open AccessIntegrating Chemistry and Experimental Knowledge into Deep Learning Models for Classification of Structural Defects
Accurate identification of structural defects in materials using atomically resolved scanning transmission electron microscopy (STEM) is crucial for understanding materials structure property correlations. However, deep neural networks (DNNs) trained on simulated data for this classification task often struggle due to the issue of out-of-distribution when applied to experimental data. Moreover, nearly all previously supervised models fail to include additional information such as chemical composition and experimental conditions. To address this issue, we introduce a novel approach that employs a single multilayer perceptron (MLP) classifier trained on multislice simulated data, along with extra information, which we coin as context information. The context information, including chemical composition and experimental imaging conditions, is transformed into vectors and combined with Zernike moment representations of the image data [ 1, 2]. This ensures a single context-aware model can be trained on simulated data from various materials, as illustrated in Figure 1. To evaluate the effectiveness of this context-aware model, we train our MLP on multislice simulated images of all available 1H MX2 monolayer transition metal dichalcogenides with various dopants, where M is the transition metal and X denotes the chalcogen element. We show that the MLP model achieves 96% accuracy on both training and testing datasets. We also evaluate the model performance on experimental high-angle annular dark field (HAADF) images of monolayer WSe₂ doped Cr. The classification results are consistent with those obtained by human experts, as shown in Figure 2. Additionally, we applied an attention model to the same datasets, which produced comparable outcomes. Fig. 1. Integrating context into a unified model. a. Context information (chemical composition, beam energy, detector range in STEM) is often neglected in prior deep learning models. b. Experimental and simulated datasets are used for training. c and e. Previous models trained without context require manual interpretation of outputs. d and f. Our approach encodes context as vectors combined with Zernike moments, enabling a unified model to produce interpretable defect labels across different materials and conditions. Fig. 2. Unified MLP identifying point defects in monolayer TMDCs. a. Accuracy curves for 70 epochs reach 96% on training and test datasets, covering 13 materials, 10 dopants, 3 energies, and 2 detectors. b. Classification of an experimental HAADF image of Cr-doped WSe₂, identifying single vacancies (green), double vacancies (cyan), and dopants (red). c. Mean patches for the five defect classes.
- Open Access
- 24 July 2024
- Published by Oxford University Press (OUP) in Microscopy and Microanalysis
- Vol. 30 (Supplement)
Can Conventional Classifiers Outperform Neural Networks in Identifying Structural Defects from Atomic Resolution Micrographs?
The identification of structural defects in materials through atomic resolution scanning transmission electron microscopy (STEM) images is pivotal for understanding material properties [ 1] and facilitating advancements in nanotechnology. However, the efficacy of deep neural networks (DNNs), which are commonly employed for such tasks, is significantly hampered by out-of-distribution drift when applied to experimental data. Furthermore, these DNN models often overlook the incorporation of available prior knowledge, such as the chemical composition of samples, into their analytical frameworks. This study proposes a novel approach that challenges the supremacy of DNNs by leveraging an ensemble of simple, yet highly effective, classifiers trained on data simulated using the multislice algorithm for various two-dimensional materials. By categorizing these materials into groups based on differences in atomic number, our methodology not only utilizes prior knowledge regarding chemical composition but also enhances the specificity and accuracy of defect identification.
- 24 July 2024
- Published by Oxford University Press (OUP) in Microscopy and Microanalysis
- Vol. 30 (Supplement)
Open AccessCan Conventional Classifiers Outperform Neural Networks in Identifying Structural Defects from Atomic Resolution Micrographs?
The identification of structural defects in materials through atomic resolution scanning transmission electron microscopy (STEM) images is pivotal for understanding material properties [ 1] and facilitating advancements in nanotechnology. However, the efficacy of deep neural networks (DNNs), which are commonly employed for such tasks, is significantly hampered by out-of-distribution drift when applied to experimental data. Furthermore, these DNN models often overlook the incorporation of available prior knowledge, such as the chemical composition of samples, into their analytical frameworks. This study proposes a novel approach that challenges the supremacy of DNNs by leveraging an ensemble of simple, yet highly effective, classifiers trained on data simulated using the multislice algorithm for various two-dimensional materials. By categorizing these materials into groups based on differences in atomic number, our methodology not only utilizes prior knowledge regarding chemical composition but also enhances the specificity and accuracy of defect identification.
- Open Access
- 20 October 2023
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 9 (42) , eadj0904
A multiscale generative model to understand disorder in domain boundaries
A continuing challenge in atomic resolution microscopy is to identify significant structural motifs and their assembly rules in synthesized materials with limited observations. Here, we propose and validate a simple and effective hybrid generative model capable of predicting unseen domain boundaries in a potassium sodium niobate thin film from only a small number of observations, without expensive first-principles calculations or atomistic simulations of domain growth. Our results demonstrate that complicated domain boundary structures spanning 1 to 100 nanometers can arise from simple interpretable local rules played out probabilistically. We also found previously unobserved, significant, tileable boundary motifs that may affect the piezoelectric response of the material system, and evidence that our system creates domain boundaries with the highest configurational entropy. More broadly, our work shows that simple yet interpretable machine learning models could pave the way to describe and understand the nature and origin of disorder in complex materials, therefore improving functional materials design.
- 20 October 2023
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 9 (42) , eadj0904
Open AccessA multiscale generative model to understand disorder in domain boundaries
A continuing challenge in atomic resolution microscopy is to identify significant structural motifs and their assembly rules in synthesized materials with limited observations. Here, we propose and validate a simple and effective hybrid generative model capable of predicting unseen domain boundaries in a potassium sodium niobate thin film from only a small number of observations, without expensive first-principles calculations or atomistic simulations of domain growth. Our results demonstrate that complicated domain boundary structures spanning 1 to 100 nanometers can arise from simple interpretable local rules played out probabilistically. We also found previously unobserved, significant, tileable boundary motifs that may affect the piezoelectric response of the material system, and evidence that our system creates domain boundaries with the highest configurational entropy. More broadly, our work shows that simple yet interpretable machine learning models could pave the way to describe and understand the nature and origin of disorder in complex materials, therefore improving functional materials design.
- Open Access
- 22 July 2023
- Published by Oxford University Press (OUP) in Microscopy and Microanalysis
- Vol. 29 (Supplement) , 1894
Exploring the Emergence of Complex Grain Boundary Structures via Hybrid Probabilistic Generative Model
It is a persistent challenge to identify important structural motifs and their assembly rules in atomic resolution scanning transmission electron microscopy (STEM) images. However, with the help of machine learning, these motifs can now be quickly located and analysed without human intervention [ 1]. Despite this, it is still difficult to comprehend the relationships between motifs, especially when these relationships appear random.
- 22 July 2023
- Published by Oxford University Press (OUP) in Microscopy and Microanalysis
- Vol. 29 (Supplement) , 1894
Open AccessExploring the Emergence of Complex Grain Boundary Structures via Hybrid Probabilistic Generative Model
It is a persistent challenge to identify important structural motifs and their assembly rules in atomic resolution scanning transmission electron microscopy (STEM) images. However, with the help of machine learning, these motifs can now be quickly located and analysed without human intervention [ 1]. Despite this, it is still difficult to comprehend the relationships between motifs, especially when these relationships appear random.
- Open Access
- 15 April 2022
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 8 (15) , eabk1005
Learning motifs and their hierarchies in atomic resolution microscopy
Characterizing materials to atomic resolution and first-principles structure-property prediction are two pillars for accelerating functional materials discovery. However, we are still lacking a rapid, noise-robust framework to extract multilevel atomic structural motifs from complex materials to complement, inform, and guide our first-principles models. Here, we present a machine learning framework that rapidly extracts a hierarchy of complex structural motifs from atomically resolved images. We demonstrate how such motif hierarchies can rapidly reconstruct specimens with various defects. Abstracting complex specimens with simplified motifs enabled us to discover a previously unidentified structure in a Mo─V─Te─Nb polyoxometalate (POM) and quantify the relative disorder in a twisted bilayer MoS2. In addition, these motif hierarchies provide statistically grounded clues about the favored and frustrated pathways during self-assembly. The motifs and their hierarchies in our framework coarse-grain disorder in a manner that allows us to understand a much broader range of multiscale samples with functional imperfections and nontrivial topological phases.
- 15 April 2022
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 8 (15) , eabk1005
Open AccessLearning motifs and their hierarchies in atomic resolution microscopy
Characterizing materials to atomic resolution and first-principles structure-property prediction are two pillars for accelerating functional materials discovery. However, we are still lacking a rapid, noise-robust framework to extract multilevel atomic structural motifs from complex materials to complement, inform, and guide our first-principles models. Here, we present a machine learning framework that rapidly extracts a hierarchy of complex structural motifs from atomically resolved images. We demonstrate how such motif hierarchies can rapidly reconstruct specimens with various defects. Abstracting complex specimens with simplified motifs enabled us to discover a previously unidentified structure in a Mo─V─Te─Nb polyoxometalate (POM) and quantify the relative disorder in a twisted bilayer MoS2. In addition, these motif hierarchies provide statistically grounded clues about the favored and frustrated pathways during self-assembly. The motifs and their hierarchies in our framework coarse-grain disorder in a manner that allows us to understand a much broader range of multiscale samples with functional imperfections and nontrivial topological phases.
- Open Access
Improving Photoelectrochemical Activity of ZnO/TiO2 Core–Shell Nanostructure through Ag Nanoparticle Integration
In solar energy harvesting using solar cells and photocatalysts, the photoexcitation of electrons and holes in semiconductors is the first major step in the solar energy conversion. The lifetime of carriers, a key factor determining the energy conversion and photocatalysis efficiency, is shortened mainly by the recombination of photoexcited carriers. We prepared and tested a series of ZnO/TiO2-based heterostructures in search of designs which can extend the carrier lifetime. Time-resolved photoluminescence tests revealed that, in ZnO/TiO2 core–shell structure the carrier lifetime is extended by over 20 times comparing with the pure ZnO nanorods. The performance improved further when Ag nanoparticles were integrated at the ZnO/TiO2 interface to construct a Z-scheme structure. We utilized these samples as photoanodes in a photoelectrochemical (PEC) cell and analyzed their solar water splitting performances. Our data showed that these modifications significantly enhanced the PEC performance. Especially, under visible light, the Z-scheme structure generated a photocurrent density 100 times higher than from the original ZnO samples. These results reveal the potential of ZnO-Ag-TiO2 nanorod arrays as a long-carrier-lifetime structure for future solar energy harvesting applications.
Open AccessImproving Photoelectrochemical Activity of ZnO/TiO2 Core–Shell Nanostructure through Ag Nanoparticle Integration
In solar energy harvesting using solar cells and photocatalysts, the photoexcitation of electrons and holes in semiconductors is the first major step in the solar energy conversion. The lifetime of carriers, a key factor determining the energy conversion and photocatalysis efficiency, is shortened mainly by the recombination of photoexcited carriers. We prepared and tested a series of ZnO/TiO2-based heterostructures in search of designs which can extend the carrier lifetime. Time-resolved photoluminescence tests revealed that, in ZnO/TiO2 core–shell structure the carrier lifetime is extended by over 20 times comparing with the pure ZnO nanorods. The performance improved further when Ag nanoparticles were integrated at the ZnO/TiO2 interface to construct a Z-scheme structure. We utilized these samples as photoanodes in a photoelectrochemical (PEC) cell and analyzed their solar water splitting performances. Our data showed that these modifications significantly enhanced the PEC performance. Especially, under visible light, the Z-scheme structure generated a photocurrent density 100 times higher than from the original ZnO samples. These results reveal the potential of ZnO-Ag-TiO2 nanorod arrays as a long-carrier-lifetime structure for future solar energy harvesting applications.
- Open Access
A machine perspective of atomic defects in scanning transmission electron microscopy
Enabled by the advances in aberration‐corrected scanning transmission electron microscopy (STEM), atomic‐resolution real space imaging of materials has allowed a direct structure‐property investigation. Traditional ways of quantitative data analysis suffer from low yield and poor accuracy. New ideas in the field of computer vision and machine learning have provided more momentum to harness the wealth of big data and sophisticated information in STEM data analytics, which has transformed STEM from a localized characterization technique to a macroscopic tool with intelligence. In this review article, we discuss the prime significance of defect topology and density in two‐dimensional (2D) materials, which have proved to be a powerful means to tune a wide range of properties. Subsequently, we systematically review advanced data analysis methods that have demonstrated promising prospects in analyzing STEM data, particularly for identifying structural defects, with high throughput and veracity. A unified framework for atomic structure identification is also summarized.
Open AccessA machine perspective of atomic defects in scanning transmission electron microscopy
Enabled by the advances in aberration‐corrected scanning transmission electron microscopy (STEM), atomic‐resolution real space imaging of materials has allowed a direct structure‐property investigation. Traditional ways of quantitative data analysis suffer from low yield and poor accuracy. New ideas in the field of computer vision and machine learning have provided more momentum to harness the wealth of big data and sophisticated information in STEM data analytics, which has transformed STEM from a localized characterization technique to a macroscopic tool with intelligence. In this review article, we discuss the prime significance of defect topology and density in two‐dimensional (2D) materials, which have proved to be a powerful means to tune a wide range of properties. Subsequently, we systematically review advanced data analysis methods that have demonstrated promising prospects in analyzing STEM data, particularly for identifying structural defects, with high throughput and veracity. A unified framework for atomic structure identification is also summarized.