Stability Prediction of Gold Nanoclusters with Different Ligands and Doped Metals: Deep Learning and Experimental Tests

Yuming Gu,Shisi Tang,Xu Liu,Xinyi Liang,Qin Zhu,Hongfeng Wu,Xiao Yang,Weihao Jin,Hongwei Chen,Chunyan Liu,Yan Zhu,Jing Ma

DOI: https://doi.org/10.1039/d3ta06892g

IF: 11.9

2024-01-01

Journal of Materials Chemistry A

Abstract:The formation energy of gold nanoclusters could be predicted quickly by deep learning.

What problem does this paper attempt to address?

New Insights Into The Stability And Structural Evolution Of Some Gold Nanoclusters

Liren Liu,Jinyun Yuan,Longjiu Cheng,Jinlong Yang

DOI: https://doi.org/10.1039/c6nr07878h

IF: 6.7

2017-01-01

Nanoscale

Abstract:Revealing the stability and structural patterns is important for precisely synthesizing or assembling ligand protected nanoclusters, and even their applications as functional nanomaterials. Investigations on structural evolutional patterns and structural stability are very challenging, because structures change with the nanocluster size and the structural stability depends on both the electron structures of cores and ligand type. Herein, we propose a hybrid superatom network (hSAN) model to understand the stability of some gold nanoclusters with different kinds of ligands. In this model, 4c-2e superatom Au-4 can form conjugated superatom networks by vertex sharing, and ligands further connect the conjugated superatom networks together to form a bigger complex network, i.e. a hSAN. The stability of the clusters, including [Au-24(Cu CPh)(14)(PPh3)(4)](2+), Au-28(S-c-C6H11)(20), Au-36(SCH2Ph-tBu)(8)Cl-20, Au-40(O-MBT)(24) and Au-52(TBBT)(32) can be explained uniformly by the hSAN model. Beyond that, a new heuristic structural pattern named the Au-13 topological rule is proposed. In the light of this heuristic rule, every Au-7 bi-tetrahedral kernel is included in an Au-13 structure with quasi-O-h symmetry, i.e. as long as the Au-7 bi-tetrahedral kernel is formed, it will be surrounded by six Au atoms to form an Au-13 structure topologically. According to this understanding, a new nanocluster [Au-44(Cu=CH3)(26)(PCH3)(4)](2+) and a new nanowire with the structural evolutional formula [Au((20n+4))(Cu CH3)((12n+2))(PCH3)(4)](2+) (n = 1, 2, 3, 4,.) are predicted. Both the understanding of the stability and the structure rule are free from the type of ligand, and will be useful for the structural predictions and determinations of ligand protected gold nanoclusters.
(Meta-)stability and Core-Shell Dynamics of Gold Nanoclusters at Finite Temperature

Diego Guedes-Sobrinho,Weiqi Wang,Ian Hamilton,Juarez L. F. Da Silva,Luca M. Ghiringhelli

DOI: https://doi.org/10.48550/arXiv.1811.04438

2018-11-12

Abstract:Gold nanoclusters have been the focus of numerous computational studies but an atomistic understanding of their structural and dynamical properties at finite temperature is far from satisfactory. To address this deficiency, we investigate gold nanoclusters via ab initio molecular dynamics, in a range of sizes where a core-shell morphology is observed. We analyze their structure and dynamics using of state-of-the-art techniques, including unsupervised machine-learning nonlinear dimensionality reduction (sketch-map) for describing the similarities and differences among the range of sampled configurations. We find that, whereas the gold nanoclusters exhibit continuous structural rearrangement, they clearly show persistent motifs: a cationic core of one to five atoms is loosely bound to a shell which typically displays a substructure resulting from the competition between locally spherical vs planar fragments. Besides illuminating the properties of core-shell gold nanoclusters, the present study proposes a set of useful tools for understanding their nature in operando.

Mesoscale and Nanoscale Physics,Atomic and Molecular Clusters
Deep Learning Accelerated Gold Nanocluster Synthesis

Jiali Li,Tiankai Chen,Kaizhuo Lim,Lingtong Chen,Saif A. Khan,Jianping Xie,Xiaonan Wang

DOI: https://doi.org/10.1002/aisy.201900029

IF: 7.298

2019-07-01

Advanced Intelligent Systems

Abstract:The understanding of inorganic reactions, especially those far from the equilibrium state, is relatively limited due to the inherent complexity. Poor understanding of the underlying synthetic chemistry constrains the design of efficient synthesis routes toward the desired final products, especially those at atomic precision. Using the synthesis of atomically precise gold nanoclusters as a demonstration platform, a deep learning framework for guiding material synthesis is successfully developed to accelerate the workflow. With only 54 examples, the graph convolutional neural networks (GCNN) plus siamese neural networks (SNN) classification model is trained. The prediction capability is demonstrated with the successful prediction of literature‐reported protocols. In addition, understanding of the synthesis process can be acquired from a decision tree trained by plentiful generated data from a well‐trained classification model. This study not only provides a data‐driven method accelerating gold nanocluster synthesis, but also sheds light on understanding complex inorganic material synthesis with low data. The understanding of inorganic reactions, especially those far from the equilibrium state, is relatively limited due to the inherent complexity. Using the synthesis of atomically precise gold nanoclusters as a demonstration platform, a deep learning framework is successfully developed to accelerate the workflow that sheds light on understanding complex inorganic material synthesis with low data.
Accelerated prediction of atomically precise cluster structures using on-the-fly machine learning

Yunzhe Wang,Shanping Liu,Peter Lile,Sam Norwood,Alberto Hernandez,Sukriti Manna,Tim Mueller

DOI: https://doi.org/10.1038/s41524-022-00856-x

IF: 12.256

2022-08-20

npj Computational Materials

Abstract:The chemical and structural properties of atomically precise nanoclusters are of great interest in numerous applications, but predicting the stable structures of clusters can be computationally expensive. In this work, we present a procedure for rapidly predicting low-energy structures of nanoclusters by combining a genetic algorithm with interatomic potentials actively learned on-the-fly. Applying this approach to aluminum clusters with 21 to 55 atoms, we have identified structures with lower energy than any reported in the literature for 25 out of the 35 sizes. Our benchmarks indicate that the active learning procedure accelerated the average search speed by about an order of magnitude relative to genetic algorithm searches using only density functional calculations. This work demonstrates a feasible way to systematically discover stable structures for large nanoclusters and provides insights into the transferability of machine-learned interatomic potentials for nanoclusters.

materials science, multidisciplinary,chemistry, physical
Quantum Chemical Calculations and Machine Learning Predictions Innovate Synthesis for High-Performance Optical Gold Nanorods

Jinchang Yin,Haonan Wu,Jintao Zhang,Shuangshuang Wu,Hongting Zheng,Fuli Zhao,Yuanzhi Shao

DOI: https://doi.org/10.1021/acs.chemmater.2c00839

IF: 10.508

2022-06-28

Chemistry of Materials

Abstract:Understanding the optical properties of gold nanorods (GNRs) in the colloidal state is crucial to engineering them for versatile applications in many fields. Concomitant gold nanospheres (GNSs) are easily involved in GNR synthesis, incurring a negative effect on the GNR performance. To unravel the underlying mechanism, we constructed a GNR–GNS heterodimer to imitate their colloidal state and calculated the relevant optical and electronic properties through a quantum chemical approach. The calculations reveal that GNSs prevent certain charge-transfer excitations of adjacent GNRs by affecting the electronic structure and thereby the excitation behavior of the GNR. We synthesized 310 sets of GNR–GNS colloidal solutions with a seed-mediated growth method and then measured their absorption spectra to extract the datasets available for 11 machine learning algorithms. Among them, XGBoost had the best prediction accuracy of over 94%. A direct relevance from the initial synthesis parameters to the final optical properties of GNR–GNS colloids has been successfully identified by the machine learning approach, which could skip the cumbersome step-by-step procedure used for the conventional nanostructure characterization as well as optimize the batch GNR synthesis process with improved GNR performance simultaneously. Methodologically, such a three-in-one approach combining chemical synthesis, quantum chemical calculations, and machine learning predictions can be extended to other chemical synthetic studies, with methodological guidance to chemistry and materials science researchers.

materials science, multidisciplinary,chemistry, physical
Theoretical Investigation on the Thermal Stability of Hollow Gold Nanoparticles

Luyun Jiang,Xing Yin,Jianwei Zhao,Hongmei Liu,Yunhong Liu,Fenying Wang,Junjie Zhu,Freddy Boey,Hua Zhang

DOI: https://doi.org/10.1021/jp905280g

2009-01-01

Abstract:The thermal stability of hollow gold nanoparticles has been studied by molecular dynamics (MD) simulations. On the basis of the MD Simulation results, the hollow nanoparticles have been classified into three categories, namely, stable, half-stable, and unstable systems. The stability of hollow nanoparticles strongly depends on the wall thickness and the aspect ratio, defined as the outer radius over the wall thickness. These features can be further presented in a two-dimensional phase space according to a large number of simulations. By using this stability diagram, the collapsing mechanism has been proposed.
Exploration of the Doping Effect in the Thiolate-protected Gold Nanoclusters: DFT Simulations of H2S-nanoalloy Complexes

Hui jia,Changlong Liu,Yonghui Li

DOI: https://doi.org/10.48550/arXiv.2112.15365

2021-12-31

Abstract:The atomically precise method has become an important technique to adjust the core of thiolate-protected gold nanoclusters to improve physical and chemical properties. But the doping effect on the structural stability has not been systematically summarized. In this work, the H2S-nanoalloy molecules with different doping metal atoms has been investigated to elucidate the impact of the dopant on the structures. With DFT simulation results, the zinc group atoms as dopants may be influenced by surrounded gold atoms and the binding of the thiolate units are enhanced. The simulated zinc group data when combined to the gold group and plantinum group data can be summarized in the perspective of balance between the ligand-core binding and core cohesive energies. Most of dopants drive the modeled nanoclusters away from the balance especially when the metal atom replaced the gold atom in gold-sulfur bindings. But when cores of the nanoclusters are dominated by gold atoms, the dopants may achieve "saturation" such that the balance in the doped clusters may be corrected. This work provide a simple profile to understand the internal shift of the structure introduced by the atomically precise method.

Mesoscale and Nanoscale Physics
Discovering Syntheses of Atomically Precise Metal Nanoclusters by Applying Machine Learning to a High‐Throughput Platform

Li Tang,Li Wang,Bin Wang,Yong Pei,Shuxin Wang

DOI: https://doi.org/10.1002/chem.202302602

2024-05-25

Chemistry - A European Journal

Abstract:Understanding the complete structure of noble metal nanoclusters is both academically and practically significant. However, progress has been hindered by the low synthetic efficiency of many nanocluster syntheses. In this study, we present the first high‐throughput syntheses of homo‐gold, homo‐copper, and gold‐copper alloy nanoclusters in dichloromethane at room temperature. Through high‐throughput screening, we successfully obtained three nanoclusters in a single reaction: Au18(SC6H11)14, [Au41Cu66(SC6H11)44](SbF6)3, and an unidentified copper cluster (referred to as Au18, Au41Cu66, and Cu‐NC). The optimized synthesis route was achieved with the assistance of machine learning for experimental data analysis, which also guided the synthesis of other metal nanoclusters such as Au40Cu34(4‐S‐PhF)40 (Au40Cu34), [Au6Cu6(SPh)12]n ([Au6Cu6]n), and Au18Cu32(3,5‐C8H9S)36 (Au18Cu32)). This research demonstrates that high‐throughput screening can be a valuable tool in accelerating the development of nanocluster syntheses.

chemistry, multidisciplinary
Deep Neural Networks for Accurate Predictions of Garnet Stability

Weike Ye,Chi Chen,Zhenbin Wang,Iek-Heng Chu,Shyue Ping Ong

DOI: https://doi.org/10.1038/s41467-018-06322-x

2017-12-06

Abstract:Predicting the stability of crystals is one of the central problems in materials science. Today, density functional theory (DFT) calculations are the computational tool of choice to obtain energies of crystals with quantitative accuracy. Despite algorithmic and computing advances, DFT calculations remain comparatively expensive and scale poorly with system size. Here we show that deep neural networks utilizing just two descriptors - the Pauling electronegativity and ionic radii - can predict the DFT formation energies of C3A2D3O12 garnets with extremely low mean absolute errors of 7-8 meV/atom, an order of magnitude improvement over previous machine learning models and well within the limits of DFT accuracy. Further extension to mixed garnets with little loss in accuracy can be achieved using a binary encoding scheme that introduces minimal increase in descriptor dimensionality. Our results demonstrate that generalizable deep-learning models for quantitative crystal stability prediction can be built on a small set of chemically-intuitive descriptors. Such models provide the means to rapidly transverse vast chemical spaces to accurately identify stable compositions, accelerating the discovery of novel materials with potentially superior properties.

Materials Science
Computational Exploration of the Nanogold Energy Landscape

Michael McGuigan,J.W. Davenport

DOI: https://doi.org/10.48550/arXiv.0912.5316

2009-12-29

Abstract:We use density functional theory to quantify finite size and shape effects for gold nanoclusters. We concentrate on the computation of binding energy as a function of bond length for icosahedral and cuboctohedral clusters. We find that the cuboctoheral gold clusters have lower energy for 13 atoms. For 55 atoms we find that the icosahedral gold clusters have lower binding energy. We also introduce a one parameter family of geometries that interpolate between the icosahedral and cuboctohedral clusters that is parametrized by an angle variable. We determine the binding energy dependence on shape as a function of the angle variable for 13 and 55 atom clusters with a minimum at the cuboctohedral point and icosahedral point respectively. We also compute the binding energy for the 147 atom gold nanocluster and show that the binding energy of the icosahedral cluster is lower than the 147 atom cuboctohedral gold cluster. We also compute the binding energy of the $Au_{55}O_2$ molecule with possible applications to catalysis.

Mesoscale and Nanoscale Physics,Materials Science,Atomic and Molecular Clusters
Gold-Thiolate Nanocluster Dynamics and Intercluster Reactions Enabled by a Machine Learned Interatomic Potential

Caitlin McCandler,Antti Pihlajamäki,Sami Malola,Hannu Häkkinen,Kristin Persson

DOI: https://doi.org/10.26434/chemrxiv-2024-rz914

2024-03-07

Abstract:Mono-layer protected metal clusters comprise a rich class of molecular systems, and are promising candidate materials for a variety of applications. While a growing number of protected nanoclusters have been synthe- sized and characterized in crystalline forms, their dynamical behavior in solution, including pre-nucleation cluster formation, is not well understood due to limitations both in characterization and first-principles mod- eling techniques. Recent advancements in machine-learned interatomic potentials are rapidly enabling the study of complex interactions such as dynamical behavior and reactivity at the nanoscale. Here, we develop an Au-S-C-H Atomic Cluster Expansion (ACE) interatomic potential for efficient and accurate molecular dynamics simulations of thiolate-protected gold nanoclusters (Aun (SCH3)m ). Trained on more than 30,000 density functional theory calculations of gold nanoclusters, the interatomic potential exhibits ab initio level accuracy in energies and forces, and replicates nanocluster dynamics including thermal vibration and chiral inversion. Long dynamics simulations (up to 0.1 μs time scale) reveal a novel mechanism explaining the ther- mal instability of neutral Au25(SR)18 clusters. Specifically, we observe multiple stages of isomerization of the Au25(SR)18 cluster, including a novel chiral isomer. Additionally we simulate coalescence of two Au25(SR)18 clusters and observe series of new clusters where the formation mechanisms are critically mediated by ligand exchange in the form of [Au–S]n rings.

Chemistry
A Comprehensive Study Of Catalytic, Morphological And Electronic Properties Of Ligand-Protected Gold Nanoclusters Using Xps, Stm, Xafs, And Tpd Techniques

Qiyuan Wu,Jiajie Cen,Yue Zhao,Xiao Tong,Yuanyuan Li,Anatoly I. Frenkel,Shen Zhao,Alexander Orlov

DOI: https://doi.org/10.1039/c7cp06376h

IF: 3.3

2018-01-01

Physical Chemistry Chemical Physics

Abstract:Ultra-small gold nanoclusters were synthesized via a ligand exchange method and deposited onto different TiO2 supports to study their properties. STM imaging revealed that the as-synthesized gold nanoclusters had 2-D morphology consisting of monolayers of gold atoms. Subsequent XPS, XAFS, and CO oxidation TPD results indicated that heat treatments of gold clusters at different temperatures significantly altered their electronic and catalytic properties due to ligand deprotection and cluster agglomeration.
Artificial intelligence real-time prediction and physical interpretation of atomic binding energies in nano-scale metal clusters

Philomena Schlexer Lamoureux,Tej S. Choksi,Verena Streibel,Frank Abild-Pedersen

DOI: https://doi.org/10.48550/arXiv.2005.02572

2020-05-06

Abstract:Single atomic sites often determine the functionality and performance of materials, such as catalysts, semi-conductors or enzymes. Computing and understanding the properties of such sites is therefore a crucial component of the rational materials design process. Because of complex electronic effects at the atomic level, atomic site properties are conventionally derived from computationally expensive first-principle calculations, as this level of theory is required to achieve relevant accuracy. In this study, we present a widely applicable machine learning (ML) approach to compute atomic site properties with high accuracy in real time. The approach works well for complex non-crystalline atomic structures and therefore opens up the possibility for high-throughput screenings of nano-materials, amorphous systems and materials interfaces. Our approach includes a robust featurization scheme to transform atomic structures into features which can be used by common machine learning models. Performing a genetic algorithm (GA) based feature selection, we show how to establish an intuitive physical interpretation of the structure-property relations implied by the ML models. With this approach, we compute atomic site stabilities of metal nanoparticles ranging from 3-55 atoms with mean absolute errors in the range of 0.11-0.14 eV in real time. We also establish the chemical identity of the site as most important factor in determining atomic site stabilities, followed by structural features like bond distances and angles. Both, the featurization and GA feature selection functionality are published in open-source python modules. With this method, we enable the efficient rational design of highly specialized real-world nano-catalysts through data-driven materials screening.

Chemical Physics,Computational Physics
Evaluation of Machine Learning Interatomic Potentials for the Properties of Gold Nanoparticles

Marco Fronzi,Rika Kobayashi,Naoki Matsumura,Kenta Watanabe,Roger D. Amos,Rafael K. Morizawa

DOI: https://doi.org/10.3390/nano12213891

IF: 5.3

2022-11-03

Nanomaterials

Abstract:We have investigated Machine Learning Interatomic Potentials in application to the properties of gold nanoparticles through the DeePMD package, using data generated with the ab-initio VASP program. Benchmarking was carried out on Au20 nanoclusters against ab-initio molecular dynamics simulations and show we can achieve similar accuracy with the machine learned potential at far reduced cost using LAMMPS. We have been able to reproduce structures and heat capacities of several isomeric forms. Comparison of our workflow with similar ML-IP studies is discussed and has identified areas for future improvement.

materials science, multidisciplinary,physics, applied,nanoscience & nanotechnology,chemistry
Reactivity and Lability Modulated by a Valence Electron Moving in and out of 25‐atom Gold Nanoclusters

Guangjun Li,Weigang Hu,Yongnan Sun,Jiayu Xu,Xiao Cai,Xinglian Cheng,Yuying Zhang,Ancheng Tang,Xu Liu,Mingyang Chen,Weiping Ding,Yan Zhu

DOI: https://doi.org/10.1002/anie.202009278

2020-01-01

Angewandte Chemie

Abstract:The emergence of atomically precise metal nanoclusters with unique electronic structures provides access to currently inaccessible catalytic challenges at the single-electron level. We investigate the catalytic behavior of gold Au-25(SR)(18)nanoclusters by monitoring an incoming and outgoing free valence electron of Au 6s(1). Distinct performances are revealed: Au-25(SR)(18)(-)is generated upon donation of an electron to neutral Au-25(SR)(18)(0)and this is associated with a loss in reactivity, whereas Au-25(SR)(18)(+)is generated from dislodgment of an electron from neutral Au-25(SR)(18)(0)with a loss in stability. The reactivity diversity of the three Au-25(SR)(18)clusters stems from different affinities with reactants and the extent of intramolecular charge migration during the reactions, which are closely associated with the valence occupancies of the clusters varied by one electron. The stability difference in the three clusters is attributed to their different equilibria, which are established between the AuSR dissociation and polymerization influenced by one electron.
Machine Learning Prediction of Hydrogen Adsorption Energy on Platinum Nanoclusters: A Comparative Study of SOAP Descriptors

Zhiheng Yu,Yanli Li,Yanwei Wen,Bin Shan,Jiaqiang Yang

DOI: https://doi.org/10.1016/j.comptc.2024.114923

IF: 2.292

2024-01-01

Computational and Theoretical Chemistry

Abstract:Hydrogen binding energy in metal materials is of high significance in the hydrogen storage as well as the hydrogen evolution reaction of electrocatalysis. In this work, the datasets (more than 9000 data) of hydrogen adsorbed on Pt nanoclusters with different sizes are obtained by first-principles calculations. Data analysis shows that the binding strength of hydrogen with Pt is closely relevant to the local structures of the adsorption sites. The local features of the distance between the platinum and hydrogen and the size of the nanoclusters are supplemented to the Smooth Overlap of Atomic Positions descriptors to fit and predict the adsorption energies of hydrogen on different Pt nano-structures by performing the machine learning method. Gaussian Process Regression (GPR) and Random Forest Regressor (RFR) are used to construct the prediction model of hydrogen binding energies and it is found the R2 of test set is improved from 0.63 to 0.78 with modified descriptors. By applying it into other nanoclusters, the MAE of the prediction model is 0.08 eV, which exhibits high accuracy of the hydrogen adsorption energy. Our model can be easily extended to the prediction of hydrogen adsorption energy of other materials with affordable computational cost and accuracy, which would be helpful for the structural design of high-performance catalysts.
Systematically Tuning the Electronic Structure of Gold Nanoclusters through Ligand Derivatization

Anthony Cirri,Hanna Morales Hernández,Christina Kmiotek,Christopher J. Johnson,Hanna Morales Hernández

DOI: https://doi.org/10.1002/anie.201907586

2019-08-19

Angewandte Chemie International Edition

Abstract:While the ability to crystallize metal nanoclusters has revealed their geometric structure, the lack of a similarly precise measure of their electronic structure has hampered the development of synthetic design rules to precisely engineer their electronic properties. We track the evolution of highly‐resolved electronic absorption spectra of gold nanoclusters with precisely mass‐selected chemical composition in a controlled environment. Simple derivatization of the ligands yields larger spectral changes than varying the overall atomic composition of the cluster for two clusters with similar symmetry and size. The nominally metal‐localized HOMO‐LUMO transition of these nanoclusters lowers in energy linearly with increasing electron donation from the exterior of the ligand shell for both cluster sizes. Very weak surface interactions, such as binding of He or N2, yield significant state‐dependent shifts, identifying states with significant interfacial character. These observations demonstrate a pathway for deliberate tuning of interfacial chemistry for chemical and technological applications.

chemistry, multidisciplinary
Data-driven simulation and characterisation of gold nanoparticle melting

Claudio Zeni,Kevin Rossi,Theodore Pavloudis,Joseph Kioseoglou,Stefano de Gironcoli,Richard Palmer,Francesca Baletto

DOI: https://doi.org/10.1038/s41467-021-26199-7

2022-04-22

Abstract:The simulation and analysis of the thermal stability of nanoparticles, a stepping stone towards their application in technological devices, require fast and accurate force fields, in conjunction with effective characterisation methods. In this work, we develop efficient, transferable, and interpretable machine learning force fields for gold nanoparticles based on data gathered from Density Functional Theory calculations. We use them to investigate the thermodynamic stability of gold nanoparticles of different sizes (1 to 6 nm), containing up to 6266 atoms, concerning a solid-liquid phase change through molecular dynamics simulations. We predict nanoparticle melting temperatures in good agreement with available experimental data. Furthermore, we characterize the solid-liquid phase change mechanism employing an unsupervised learning scheme to categorize local atomic environments. We thus provide a data-driven definition of liquid atomic arrangements in the inner and surface regions of a nanoparticle and employ it to show that melting initiates at the outer layers.

Computational Physics,Materials Science
Metastability of A Gold Nanoring: Density-Functional Calculations

Q Sun,Q Wang,P Jena,R Note,JZ Yu,Y Kawazoe

DOI: https://doi.org/10.1103/physrevb.70.245411

2004-01-01

Abstract:First-principles calculations based on gradient corrected density functional theory show that a cluster of as few as 90 gold atoms can be stabilized in a ring structure having fcc (111) motif with the binding energy/atom and interatomic distance approaching 91% and 96% of the bulk values, respectively. Although the ring structure lies 0.139 eV/atom higher in energy than a polyicosahedral structure, the calculated frequencies are real. Thus under appropriate experimental conditions it may still be possible to synthesize a metastable form of gold nanoring, as found in the recent experiment.
Deep-learning-based prediction of nanoparticle phase transitions during in situ transmission electron microscopy

Wenkai Fu,Steven R. Spurgeon,Chongmin Wang,Yuyan Shao,Wei Wang,Amra Peles

DOI: https://doi.org/10.48550/arXiv.2205.11407

2022-05-23

Materials Science

Abstract:We develop the machine learning capability to predict a time sequence of in-situ transmission electron microscopy (TEM) video frames based on the combined long-short-term-memory (LSTM) algorithm and the features de-entanglement method. We train deep learning models to predict a sequence of future video frames based on the input of a sequence of previous frames. This unique capability provides insight into size dependent structural changes in Au nanoparticles under dynamic reaction condition using in-situ environmental TEM data, informing models of morphological evolution and catalytic properties. The model performance and achieved accuracy of predictions are desirable based on, for scientific data characteristic, based on limited size of training data sets. The model convergence and values for the loss function mean square error show dependence on the training strategy, and structural similarity measure between predicted structure images and ground truth reaches the value of about 0.7. This computed structural similarity is smaller than values obtained when the deep learning architecture is trained using much larger benchmark data sets, it is sufficient to show the structural transition of Au nanoparticles. While performance parameters of our model applied to scientific data fall short of those achieved for the non-scientific big data sets, we demonstrate model ability to predict the evolution, even including the particle structural phase transformation, of Au nano particles as catalyst for CO oxidation under the chemical reaction conditions. Using this approach, it may be possible to anticipate the next steps of a chemical reaction for emerging automated experimentation platforms.

Stability Prediction of Gold Nanoclusters with Different Ligands and Doped Metals: Deep Learning and Experimental Tests

New Insights Into The Stability And Structural Evolution Of Some Gold Nanoclusters

(Meta-)stability and Core-Shell Dynamics of Gold Nanoclusters at Finite Temperature

Deep Learning Accelerated Gold Nanocluster Synthesis

Accelerated prediction of atomically precise cluster structures using on-the-fly machine learning

Quantum Chemical Calculations and Machine Learning Predictions Innovate Synthesis for High-Performance Optical Gold Nanorods

Theoretical Investigation on the Thermal Stability of Hollow Gold Nanoparticles

Exploration of the Doping Effect in the Thiolate-protected Gold Nanoclusters: DFT Simulations of H2S-nanoalloy Complexes

Discovering Syntheses of Atomically Precise Metal Nanoclusters by Applying Machine Learning to a High‐Throughput Platform

Deep Neural Networks for Accurate Predictions of Garnet Stability

Computational Exploration of the Nanogold Energy Landscape

Gold-Thiolate Nanocluster Dynamics and Intercluster Reactions Enabled by a Machine Learned Interatomic Potential

A Comprehensive Study Of Catalytic, Morphological And Electronic Properties Of Ligand-Protected Gold Nanoclusters Using Xps, Stm, Xafs, And Tpd Techniques

Artificial intelligence real-time prediction and physical interpretation of atomic binding energies in nano-scale metal clusters

Evaluation of Machine Learning Interatomic Potentials for the Properties of Gold Nanoparticles

Reactivity and Lability Modulated by a Valence Electron Moving in and out of 25‐atom Gold Nanoclusters

Machine Learning Prediction of Hydrogen Adsorption Energy on Platinum Nanoclusters: A Comparative Study of SOAP Descriptors

Systematically Tuning the Electronic Structure of Gold Nanoclusters through Ligand Derivatization

Data-driven simulation and characterisation of gold nanoparticle melting

Metastability of A Gold Nanoring: Density-Functional Calculations

Deep-learning-based prediction of nanoparticle phase transitions during in situ transmission electron microscopy