Large-Scale Atomic Simulation via Machine Learning Potentials Constructed by Global Potential Energy Surface Exploration

Efficient Sampling for Machine Learning Electron Density and Its Response in Real Space

Electron density is a fundamental quantity that can in principle determine all ground state electronic properties of a given system. Although machine learning (ML) models for electron density based on either an atom-centered basis or a real-space grid have been proposed, the demand for a number of high-order basis functions or grid points is enormous. In this work, we propose an efficient grid-point sampling strategy that combines targeted sampling favoring a large density and a screening of grid points associated with linearly independent atomic features. This new sampling strategy is integrated with a field-induced recursively embedded atom neural network model to develop a real-space grid-based ML model for the electron density and its response to an electric field. This approach is applied to a QM9 molecular data set, a H2O/Pt(111) interfacial system, an Au(100) electrode, and an Au nanoparticle under an electric field. The number of training points is found to be much smaller than previous models, while yielding comparably accurate predictions for the electron density of the entire grid. The resultant machine-learned electron density model enables us to properly partition partial charge onto each atom and analyze the charge variation upon proton transfer in the H2O/Pt(111) system. The machine-learning electronic response model allows us to predict charge transfer and the electrostatic potential change induced by an electric field applied to an Au(100) electrode or an Au nanoparticle.

Introduction to machine learning potentials for atomistic simulations

Machine learning potentials have revolutionised the field of atomistic simulations in recent years and are becoming a mainstay in the toolbox of computational scientists. This paper aims to provide an overview and introduction into machine learning potentials and their practical application to scientific problems. We provide a systematic guide for developing machine learning potentials, reviewing chemical descriptors, regression models, data generation and validation approaches. We begin with an emphasis on the earlier generation of models, such as high-dimensional neural network potentials and Gaussian approximation potentials, to provide historical perspective and guide the reader towards the understanding of recent developments, which are discussed in detail thereafter. Furthermore, we refer to relevant expert reviews, open-source software, and practical examples-further lowering the barrier to exploring these methods. The paper ends with selected showcase examples, highlighting the capabilities of machine learning potentials and how they can be applied to push the boundaries in atomistic simulations.

Crumbling crystals: on the dissolution mechanism of NaCl in water

Dissolution of ionic salts in water is ubiquitous, particularly for NaCl. However, an atomistic scale understanding of the process remains elusive. Simulations lend themselves conveniently to studying dissolution since they provide the spatio-temporal resolution that can be difficult to obtain experimentally. Nevertheless, the complexity of various inter- and intra-molecular interactions require careful treatment and long time scale simulations, both of which are typically hindered by computational expense. Here, we use advances in machine learning potential methodology to resolve at an ab initio level of theory the dissolution mechanism of NaCl in water. The picture that emerges is that of a steady ion-wise unwrapping of the crystal preceding its rapid disintegration, reminiscent of crumbling. The onset of crumbling can be explained by a strong increase in the ratio of the surface area to volume of the crystal. Overall, dissolution comprises a series of highly dynamical microscopic sub-processes, resulting in an inherently stochastic mechanism. These atomistic level insights contribute to the general understanding of dissolution mechanisms in other crystals, and the methodology is primed for more complex systems of recent interest such as water/salt interfaces under flow and salt crystals under confinement.

The Transformation Mechanism of Graphite to Hexagonal Diamond under Shock Conditions

The formation of a hexagonal diamond represents one of the most intriguing questions in materials science. Under shock conditions, the graphite basal plane tends to slide and pucker to form diamond. However, how the shock strength determines the phase selectivity remains unclear. In this work, using a DFT-trained carbon global neural network model, we studied the shock-induced graphite transition. The poor sliding caused by scarce sliding time under high-strength shock leads to metastable hexagonal diamond with an orientation relationship of (001)G//(100)HD+[010]G//[010]HD, while under low-strength shock due to long sliding distance cubic diamond forms with the orientation (001)G//(111)CD+[100]G//[110]CD, unveiling the strength-dependent graphite transition mechanism. We for the first time provide computational evidence of the strength-dependent graphite transition from first-principles, clarifying the long-term unresolved shock-induced hexagonal diamond formation mechanism and the structural source of the strength-dependent trend, which facilitates the hexagonal diamond synthesis via controlled experiment.

Aqueous Titania Interfaces

Water-metal oxide interfaces are central to many phenomena and applications, ranging from material corrosion and dissolution to photoelectrochemistry and bioengineering. In particular, the discovery of photocatalytic water splitting on TiO2 has motivated intensive studies of water-TiO2 interfaces for decades. So far, a broad understanding of the interaction of water vapor with several TiO2 surfaces has been obtained. However, much less is known about liquid water-TiO2 interfaces, which are more relevant to many practical applications. Probing these complex systems at the molecular level is experimentally challenging and is sometimes possible only through computational studies. This review summarizes recent advances in the atomistic understanding, mostly through computational simulations, of the structure and dynamics of interfacial water on TiO2 surfaces. The main focus is on the nature, molecular or dissociated, of water in direct contact with low-index defect-free crystalline surfaces. The hydroxyls resulting from water dissociation are essential in the photooxidation of water and critically affect the surface chemistry of TiO2.

Accurate fundamental invariant-neural network representation of ab initio potential energy surfaces

Highly accurate potential energy surfaces are critically important for chemical reaction dynamics. The large number of degrees of freedom and the intricate symmetry adaption pose a big challenge to accurately representing potential energy surfaces (PESs) for polyatomic reactions. Recently, our group has made substantial progress in this direction by developing the fundamental invariant-neural network (FI-NN) approach. Here, we review these advances, demonstrating that the FI-NN approach can represent highly accurate, global, full-dimensional PESs for reactive systems with even more than 10 atoms. These multi-channel reactions typically involve many intermediates, transition states, and products. The complexity and ruggedness of this potential energy landscape present even greater challenges for full-dimensional PES representation. These PESs exhibit a high level of complexity, molecular size, and accuracy of fit. Dynamics simulations based on these PESs have unveiled intriguing and novel reaction mechanisms, providing deep insights into the intricate dynamics involved in combustion, atmospheric, and organic chemistry.

Global Neural Network Potential with Explicit Many-Body Functions for Improved Descriptions of Complex Potential Energy Surface

The high dimensional machine learning potential (MLP) that has developed rapidly in the past decade represents a giant step forward in large-scale atomic simulation for complex systems. The long-range interaction and the poor description of chemical reactions are typical problems of high dimensional MLP, which are mainly caused by the poor structure discrimination of the atom-centered ML model. Herein, we propose a low-cost neural-network-based MLP architecture for fitting global potential energy surface data, namely, G-MBNN, that can offer improved energy and force resolution on a complex potential energy surface. In G-MBNN, a set of many-body energy terms based on the local atomic environment are explicitly included in computing the total energy─the total energy of the system is written as the sum of atomic energy and many-body energy contributions. These extra many-body energy terms are computationally low-cost and, importantly, can provide easy access to delicate energy terms in complex systems such as very short repulsion, long-range attractions, and sensitive angular-dependent covalent interactions. We implement G-MBNN in the LASP code and demonstrate the improved accuracy of the new framework in representative systems, including ternary-element energy materials LiCoOx, TiO2 with defects, and a series of organic reactions.

Universal machine learning for the response of atomistic systems to external fields

Machine learned interatomic interaction potentials have enabled efficient and accurate molecular simulations of closed systems. However, external fields, which can greatly change the chemical structure and/or reactivity, have been seldom included in current machine learning models. This work proposes a universal field-induced recursively embedded atom neural network (FIREANN) model, which integrates a pseudo field vector-dependent feature into atomic descriptors to represent system-field interactions with rigorous rotational equivariance. This "all-in-one" approach correlates various response properties like dipole moment and polarizability with the field-dependent potential energy in a single model, very suitable for spectroscopic and dynamics simulations in molecular and periodic systems in the presence of electric fields. Especially for periodic systems, we find that FIREANN can overcome the intrinsic multiple-value issue of the polarization by training atomic forces only. These results validate the universality and capability of the FIREANN method for efficient first-principles modeling of complicated systems in strong external fields.

Combined molecular dynamics and coordinate driving method for automatically searching complicated reaction pathways

The combined molecular dynamics and coordinate driving (MD/CD) method is updated and generalized in this work to broaden its applications in automatically searching reaction pathways for complicated reactions. In this updated version, MD simulations are performed with the GFN's family of methods to systematically sample conformers for almost any systems with atomic numbers Z ≤ 86. The improved CD procedure is greatly accelerated by applying a pre-screening stage at the semiempirical GFN2-xTB level. An automatic module based on the Marcus theory and its improved version (the Wolynes theory) is designed to include single electron transfer (SET) processes into reaction pathways. The capabilities of this method are demonstrated by exploring the most possible reaction pathways of three experimentally reported reactions: the organophosphine-catalyzed trans phosphinoboration, the Fe(II) complex-mediated C(sp2)-H borylation reaction, and the SET-triggered deaminative radical cross-coupling reaction. Comprehensive reaction networks are obtained for all three reactions with reasonable computational costs. Detailed mechanisms for these reactions can account for the reported experimental facts.

ChecMatE: A workflow package to automatically generate machine learning potentials and phase diagrams for semiconductor alloys

Semiconductor alloy materials are highly versatile due to their adjustable properties; however, exploring their structural space is a challenging task that affects the control of their properties. Traditional methods rely on ad hoc design based on the understanding of known chemistry and crystallography, which have limitations in computational efficiency and search space. In this work, we present ChecMatE (Chemical Material Explorer), a software package that automatically generates machine learning potentials (MLPs) and uses global search algorithms to screen semiconductor alloy materials. Taking advantage of MLPs, ChecMatE enables a more efficient and cost-effective exploration of the structural space of materials and predicts their energy and relative stability with ab initio accuracy. We demonstrate the efficacy of ChecMatE through a case study of the InxGa1-xN system, where it accelerates structural exploration at reduced costs. Our automatic framework offers a promising solution to the challenging task of exploring the structural space of semiconductor alloy materials.

Open-Source Machine Learning in Computational Chemistry

The field of computational chemistry has seen a significant increase in the integration of machine learning concepts and algorithms. In this Perspective, we surveyed 179 open-source software projects, with corresponding peer-reviewed papers published within the last 5 years, to better understand the topics within the field being investigated by machine learning approaches. For each project, we provide a short description, the link to the code, the accompanying license type, and whether the training data and resulting models are made publicly available. Based on those deposited in GitHub repositories, the most popular employed Python libraries are identified. We hope that this survey will serve as a resource to learn about machine learning or specific architectures thereof by identifying accessible codes with accompanying papers on a topic basis. To this end, we also include computational chemistry open-source software for generating training data and fundamental Python libraries for machine learning. Based on our observations and considering the three pillars of collaborative machine learning work, open data, open source (code), and open models, we provide some suggestions to the community.

Parametrically Managed Activation Function for Fitting a Neural Network Potential with Physical Behavior Enforced by a Low-Dimensional Potential

Machine-learned representations of potential energy surfaces generated in the output layer of a feedforward neural network are becoming increasingly popular. One difficulty with neural network output is that it is often unreliable in regions where training data is missing or sparse. Human-designed potentials often build in proper extrapolation behavior by choice of functional form. Because machine learning is very efficient, it is desirable to learn how to add human intelligence to machine-learned potentials in a convenient way. One example is the well-understood feature of interaction potentials that they vanish when subsystems are too far separated to interact. In this article, we present a way to add a new kind of activation function to a neural network to enforce low-dimensional constraints. In particular, the activation function depends parametrically on all of the input variables. We illustrate the use of this step by showing how it can force an interaction potential to go to zero at large subsystem separations without either inputting a specific functional form for the potential or adding data to the training set in the asymptotic region of geometries where the subsystems are separated. In the process of illustrating this, we present an improved set of potential energy surfaces for the 14 lowest ³A' states of O₃. The method is more general than this example, and it may be used to add other low-dimensional knowledge or lower-level knowledge to machine-learned potentials. In addition to the O₃ example, we present a greater-generality method called parametrically managed diabatization by deep neural network (PM-DDNN) that is an improvement on our previously presented permutationally restrained diabatization by deep neural network (PR-DDNN).

In Situ Structure of a Mo-Doped Pt-Ni Catalyst during Electrochemical Oxygen Reduction Resolved from Machine Learning-Based Grand Canonical Global Optimization

Pt-Ni alloy is by far the most active cathode material for oxygen reduction reaction (ORR) in the proton-exchange membrane fuel cell, and the addition of a tiny amount of a third-metal Mo can significantly improve the catalyst durability and activity. Here, by developing machine learning-based grand canonical global optimization, we are able to resolve the in situ structures of this important three-element alloy system under ORR conditions and identify their correlations with the enhanced ORR performance. We disclose the bulk phase diagram of Pt-Ni-Mo alloys and determine the surface structures under the ORR reaction conditions by exploring millions of likely structure candidates. The pristine Pt-Ni-Mo alloy surfaces are shown to undergo significant structure reconstruction under ORR reaction conditions, where a surface-adsorbed MoO₄ monomer or Mo₂O _x dimers cover the Pt-skin surface above 0.9 V vs RHE and protect the surface from Ni leaching. The physical origins are revealed by analyzing the electronic structure of O atoms in MoO₄ and on the Pt surface. In viewing the role of high-valence transition metal oxide clusters, we propose a set of quantitative measures for designing better catalysts and predict that six elements in the periodic table, namely, Mo, Tc, Os, Ta, Re, and W, can be good candidates for alloying with PtNi to improve the ORR catalytic performance. We demonstrate that machine learning-based grand canonical global optimization is a powerful and generic tool to reveal the catalyst dynamics behavior in contact with a complex reaction environment.

Comprehensive Study of Oxygen Vacancies on the Catalytic Performance of ZnO for CO/H2 Activation Using Machine Learning-Accelerated First-Principles Simulations

Oxygen vacancies (OVs) play important roles on any oxide catalysts. In this work, using an investigation of the OV effects on ZnO(101̅0) for CO and H₂ activation as an example, we demonstrate, via machine learning potentials (MLPs), genetic algorithm (GA)-based global optimization, and density functional theory (DFT) validations, that the ZnO(101̅0) surface with 0.33 ML OVs is the most likely surface configuration under experimental conditions (673 K and 2.5 MPa syngas (H₂:CO = 1.5)). It is found that a surface reconstruction from the wurtzite structure to a body-centered-tetragonal one would occur in the presence of OVs. We show that the OVs create a Zn₃ cluster site, allowing H₂ homolysis and C-O bond cleavage to occur. Furthermore, the activity of intrinsic sites (Zn_3c and O_3c sites) is almost invariable, while the activity of the generated OV sites is strongly dependent on the concentration of the OVs. It is also found that OV distributions on the surface can considerably affect the reactions; the barrier of C-O bond dissociation is significantly reduced when the OVs are aligned along the [12̅10] direction. These findings may be general in the systems with metal oxides in heterogeneous catalysis and may have significant impacts on the field of catalyst design by regulating the concentration and distribution of the OVs.

Integrating Machine Learning in the Coarse-Grained Molecular Simulation of Polymers

Machine learning (ML) is having an increasing impact on the physical sciences, engineering, and technology and its integration into molecular simulation frameworks holds great potential to expand their scope of applicability to complex materials and facilitate fundamental knowledge and reliable property predictions, contributing to the development of efficient materials design routes. The application of ML in materials informatics in general, and polymer informatics in particular, has led to interesting results, however great untapped potential lies in the integration of ML techniques into the multiscale molecular simulation methods for the study of macromolecular systems, specifically in the context of Coarse Grained (CG) simulations. In this Perspective, we aim at presenting the pioneering recent research efforts in this direction and discussing how these new ML-based techniques can contribute to critical aspects of the development of multiscale molecular simulation methods for bulk complex chemical systems, especially polymers. Prerequisites for the implementation of such ML-integrated methods and open challenges that need to be met toward the development of general systematic ML-based coarse graining schemes for polymers are discussed.

Accurate and Interpretable Dipole Interaction Model-Based Machine Learning for Molecular Polarizability

Polarizabilities play significant roles in describing dispersive and inductive interactions of the atom and molecular systems. However, an accurate prediction of molecular polarizabilities from first principles is computationally prohibitive. Although physical models or statistical machine learning models have been proposed, either a lack of accurate description of local chemical environments or demanding a large number of samples for training has limited their practical applications. In this study, we combine a physically inspired dipole interaction model and an accurate neural network method for predicting the polarizability tensors of molecules. With the local chemical environment precisely described and the requirement of rotational covariance naturally fulfilled, this hybrid model is proven to give an accurate molecular polarizability prediction, essentially reducing the number of training samples. The atomic polarizabilities are physically interpretable and transferable to larger molecules unseen in the training set. This promising method may find its wide range of applications, such as spectroscopic simulations and the construction of polarizable force fields.

Coupled oxygen desorption and structural reconstruction accompanying reduction of copper oxide

Understanding structural transformation and phase transition accompanying reactions in a solid as a catalyst or oxygen carrier is important to the design and optimization of many catalytic or chemical looping reaction processes. Herein, we combined density functional theory calculation with the stochastic surface walking global optimization approach to track the structural transformation accompanying the reduction of CuO upon releasing oxygen. We then used machine learning (ML) methods to correlate the structural properties of CuO_x with varying x. By decomposing a reduction step into oxygen detachment and structural reconstruction, we identified two types of pathways: (1) uniform reduction with minimal structural changes; (2) segregated reduction with significant reconstruction. The results of ML analysis showed that the most important feature is the radial distribution functions of Cu-O at a percentage of oxygen vacancy [C(O_V)] < 50% and Cu-Cu at C(O_V) > 50% for CuO_x formation. These features reflect the underlying physicochemical origin, i.e., Cu-O breaking and Cu-Cu formation in the respective stage of reduction. Phase diagram analysis indicates that CuO will be reduced to Cu₂O under a typical oxygen uncoupling condition. This work demonstrates the complexity of solid structural transformation and the potential of ML methods in studying solid state materials involved in many chemical processes.

Tracking the Role of Defect Types in Co3O4 Structural Evolution and Active Motifs during Oxygen Evolution Reaction

Dynamic reconstruction of catalyst active sites is particularly important for metal oxide-catalyzed oxygen evolution reaction (OER). However, the mechanism of how vacancy-induced reconstruction aids OER remains ambiguous. Here, we use Co₃O₄ with Co or O vacancies to uncover the effects of different defects in the reconstruction process and the active motifs relevant to alkaline OER. Combining in situ characterization and theoretical calculations, we found that cobalt oxides are converted to an amorphous [Co(OH)₆] intermediate state, and then the mismatched rates of *OH adsorption and deprotonation lead to irreversible catalyst reconstruction. The stronger *OH adsorption but weaker deprotonation induced by O defects provides the driving force for reconstruction, while Co defects favor dehydrogenation and reduce the reconstruction rate. Importantly, both O and Co defects trigger highly OER-active bridge Co sites in reconstructed catalysts, of which Co defects induce a short Co-Co distance (3.38 Å) under compressive lattice stress and show the best OER activity (η₁₀ of 262 mV), superior to reconstructed oxygen-defected Co₃O₄-V_O (η₁₀ of 300 mV) and defect-free Co₃O₄ (η₁₀ of 320 mV). This work highlights that engineering defect-dependent reconstruction may provide a rational route for electrocatalyst design in energy-related applications.

Transition state search and geometry relaxation throughout chemical compound space with quantum machine learning

We use energies and forces predicted within response operator based quantum machine learning (OQML) to perform geometry optimization and transition state search calculations with legacy optimizers but without the need for subsequent re-optimization with quantum chemistry methods. For randomly sampled initial coordinates of small organic query molecules, we report systematic improvement of equilibrium and transition state geometry output as training set sizes increase. Out-of-sample S_N2 reactant complexes and transition state geometries have been predicted using the LBFGS and the QST2 algorithms with an root-mean-square deviation (RMSD) of 0.16 and 0.4 Å-after training on up to 200 reactant complex relaxations and transition state search trajectories from the QMrxn20 dataset, respectively. For geometry optimizations, we have also considered relaxation paths up to 5'595 constitutional isomers with sum formula C₇H₁₀O₂ from the QM9-database. Using the resulting OQML models with an LBFGS optimizer reproduces the minimum geometry with an RMSD of 0.14 Å, only using ∼6000 training points obtained from normal mode sampling along the optimization paths of the training compounds without the need for active learning. For converged equilibrium and transition state geometries, subsequent vibrational normal mode frequency analysis indicates deviation from MP2 reference results by on average 14 and 26 cm^-1, respectively. While the numerical cost for OQML predictions is negligible in comparison to density functional theory or MP2, the number of steps until convergence is typically larger in either case. The success rate for reaching convergence, however, improves systematically with training set size, underscoring OQML's potential for universal applicability.

Conical Intersection in Chemiluminescence of Cyclic Peroxides

Chemiluminescence (CL) utilizing chemiexcitation for energy transformation is one of the most highly sensitive and useful analytical techniques. The chemiexcitation is a chemical process of a ground-state reactant producing an excited-state product, in which a nonadiabatic event is facilitated by conical intersections (CIs), the specific molecular geometries where electronic states are degenerated. Cyclic peroxides, especially 1,2-dioxetane/dioxetanone derivatives, are the iconic chemiluminescent substances. In this Perspective, we concentrated on the CIs in the CL of cyclic peroxides. We first present a computational overview on the role of CIs between the ground (S₀) state and the lowest singlet excited (S₁) state in the thermolysis of cyclic peroxides. Subsequently, we discuss the role of the S₀/S₁ CI in the CL efficiency and point out misunderstandings in some theoretical studies on the singlet chemiexcitations of cyclic peroxides. Finally, we address the challenges and future prospects in theoretically calculating S₀/S₁ CIs and simulating the dynamics and chemiexcitation efficiency in the CL of cyclic peroxides.

Unveiling the Au Surface Reconstruction in a CO Environment by Surface Dynamics and Ab Initio Thermodynamics

Surface reconstruction changes the atomic configuration of the metal surface and thus alters its intrinsic physical and chemical properties. Recent in situ experiments have shown a variety of surface reconstructions under reaction conditions, but how to effectively predict and characterize these structures remains challenging. Herein, we combine a DFT-based kinetic Monte Carlo simulation method and ab initio thermodynamics to explore the low-energy configurations of metal surface reconstructions, which takes the surface dynamics under the reactive environment into account. We systematically simulate 13 Au surfaces ((100), (110), (111), (210), (211), (221), (310), (311), (320), (321), (322), (331), and (332)) in the CO environment and identify 19 candidate reconstruction patterns driven by CO adsorption. The breakup of the original surfaces is attributed to the lateral interactions among the nearest-neighboring adsorbates. This work provides an efficient approach to unveil the reconstructed metal surface structures in reactive environments for guiding the experiments.

Combined fragment-based machine learning force field with classical force field and its application in the NMR calculations of macromolecules in solutions

We have developed a combined fragment-based machine learning (ML) force field and molecular mechanics (MM) force field for simulating the structures of macromolecules in solutions, and then compute its NMR chemical shifts with the generalized energy-based fragmentation (GEBF) approach at the level of density functional theory (DFT). In this work, we first construct Gaussian approximation potential based on GEBF subsystems of macromolecules for MD simulations and then a GEBF-based neural network (GEBF-NN) with deep potential model for the studied macromolecule. Then, we develop a GEBF-NN/MM force field for macromolecules in solutions by combining the GEBF-NN force field for the solute molecule and ff14SB force field for solvent molecules. Using the GEBF-NN/MM MD simulation to generate snapshot structures of solute/solvent clusters, we then perform the NMR calculations with the GEBF approach at the DFT level to calculate NMR chemical shifts of the solute molecule. Taking a heptamer of oligopyridine-dicarboxamides in chloroform solution as an example, our results show that the GEBF-NN force field is quite accurate for this heptamer by comparing with the reference DFT results. For this heptamer in chloroform solution, both the GEBF-NN/MM and classical MD simulations could lead to helical structures from the same initial extended structure. The GEBF-DFT NMR results indicate that the GEBF-NN/MM force field could lead to more accurate NMR chemical shifts on hydrogen atoms by comparing with the experimental NMR results. Therefore, the GEBF-NN/MM force field could be employed for predicting more accurate dynamical behaviors than the classical force field for complex systems in solutions.

Toward High-level Machine Learning Potential for Water Based on Quantum Fragmentation and Neural Networks

Accurate and efficient simulation of liquids, such as water and salt solutions, using high-level wave function theories is still a formidable task for computational chemists owing to the high computational costs. In this study, we develop a deep machine learning potential based on fragment-based second-order Møller-Plesset perturbation theory (DP-MP2) for water through neural networks. We show that the DP-MP2 potential predicts the structural, dynamical, and thermodynamic properties of liquid water in better agreement with the experimental data than previous studies based on density functional theory (DFT). The nuclear quantum effects (NQEs) on the properties of liquid water are also examined, which are noticeable in affecting the structural and dynamical properties of liquid water under ambient conditions. This work provides a general framework for quantitative predictions of the properties of condensed-phase systems with the accuracy of high-level wave function theory while achieving significant computational savings compared to ab initio simulations.

Machine learning potential era of zeolite simulation

Zeolites, owing to their great variety and complexity in structure and wide applications in chemistry, have long been the hot topic in chemical research. This perspective first presents a short retrospect of theoretical investigations on zeolites using the tools from classical force fields to quantum mechanics calculations and to the latest machine learning (ML) potential simulations. ML potentials as the next-generation technique for atomic simulation open new avenues to simulate and interpret zeolite systems and thus hold great promise for finally predicting the structure-functionality relation of zeolites. Recent advances using ML potentials are then summarized from two main aspects: the origin of zeolite stability and the mechanism of zeolite-related catalytic reactions. We also discussed the possible scenarios of ML potential application aiming to provide instantaneous and easy access of zeolite properties. These advanced applications could now be accomplished by combining cloud-computing-based techniques with ML potential-based atomic simulations. The future development of ML potentials for zeolites in the respects of improving the calculation accuracy, expanding the application scope and constructing the zeolite-related datasets is finally outlooked.

Pressure tuned incommensurability and guest structure transition in compressed scandium from machine learning atomic simulation

Scandium (Sc) is the lightest non-main-group element and transforms to a host-guest (H-G) incommensurate structure under gigapascal (GPa) pressures. While the host structure is stable over a wide pressure range, the guest structure may exist in multiple forms, featuring different incommensurate ratios, and mixing up to generate long-range "disordered" guest structures. Here, we employed the recently developed global neural network (g-NN) potential and the stochastic surface walking (SSW) global optimization algorithm to explore the global potential energy surface of Sc under various pressures. We probe the global minima structure in a system made of hundreds of atoms and revealed that the solid-phase transition between Sc-I and H-G Sc-II phases is fully reconstructive in nature. Above 62.5 GPa, the pressure will further destabilize the face-centered tetragonal (fct, Sc-IIa) guest structure to a body-centered tetragonal phase (bct, Sc-IIb), while sustaining the host structure. The structural transition mechanism of this work will shed light on the nature of the complex H-G structural modifications in compressed metals.

Automated search for optimal surface phases (ASOPs) in grand canonical ensemble powered by machine learning

The surface of a material often undergoes dramatic structure evolution under a chemical environment, which, in turn, helps determine the different properties of the material. Here, we develop a general-purpose method for the automated search of optimal surface phases (ASOPs) in the grand canonical ensemble, which is facilitated by the stochastic surface walking (SSW) global optimization based on global neural network (G-NN) potential. The ASOP simulation starts by enumerating a series of composition grids, then utilizes SSW-NN to explore the configuration and composition spaces of surface phases, and relies on the Monte Carlo scheme to focus on energetically favorable compositions. The method is applied to silver surface oxide formation under the catalytic ethene epoxidation conditions. The known phases of surface oxides on Ag(111) are reproduced, and new phases on Ag(100) are revealed, which exhibit novel structure features that could be critical for understanding ethene epoxidation. Our results demonstrate that the ASOP method provides an automated and efficient way for probing complex surface structures that are beneficial for designing new functional materials under working conditions.

Modeling amide-I vibrations of alanine dipeptide in solution by using neural network protocol

Infrared spectroscopy is a powerful tool for the understanding of molecular structure and function of polypeptides. Theoretical interpretation of IR spectra relies on ab initio calculations may be very costly in computational resources. Herein, we developed a neural network (NN) modeling protocol to evaluate a model dipeptide's backbone amide-I spectra. DFT calculations were performed for the amide-I vibrational motions and structural parameters of alanine dipeptide (ALAD) conformers in different micro-environments ranging from polar to non-polar ones. The obtained backbone dihedrals, C = O bond lengths and amide-I frequencies of ALAD were gather together for NN architecture. The applications of built NN protocols for the prediction of amide-I frequencies of ALAD in other solvation conditions are quite satisfactory with much less computational cost comparing with electronic structure calculations. The results show that this cost-effective way enables us to decipher the polypeptide's dynamic secondary structures and biological functions with their backbone vibrational probes.

Dynamics of Heterogeneous Catalytic Processes at Operando Conditions

The rational design of high-performance catalysts is hindered by the lack of knowledge of the structures of active sites and the reaction pathways under reaction conditions, which can be ideally addressed by an in situ/operando characterization. Besides the experimental insights, a theoretical investigation that simulates reaction conditions-so-called operando modeling-is necessary for a plausible understanding of a working catalyst system at the atomic scale. However, there is still a huge gap between the current widely used computational model and the concept of operando modeling, which should be achieved through multiscale computational modeling. This Perspective describes various modeling approaches and machine learning techniques that step toward operando modeling, followed by selected experimental examples that present an operando understanding in the thermo- and electrocatalytic processes. At last, the remaining challenges in this area are outlined.

Building quantum mechanics quality force fields of proteins with the generalized energy-based fragmentation approach and machine learning

We combined our generalized energy-based fragmentation (GEBF) approach and machine learning (ML) technique to construct quantum mechanics (QM) quality force fields for proteins. In our scheme, the training sets for a protein are only constructed from its small subsystems, which capture all short-range interactions in the target system. The energy of a given protein is expressed as the summation of atomic contributions from QM calculations of various subsystems, corrected by long-range Coulomb and van der Waals interactions. With the Gaussian approximation potential (GAP) method, our protocol can automatically generate training sets with high efficiency. To facilitate the construction of training sets for proteins, we store all trained subsystem data in a library. If subsystems in the library are detected in a new protein, corresponding datasets can be directly reused as a part of the training set on this new protein. With two polypeptides, 4ZNN and 1XQ8 segment, as examples, the energies and forces predicted by GEBF-GAP are in good agreement with those from conventional QM calculations, and dihedral angle distributions from GEBF-GAP molecular dynamics (MD) simulations can also well reproduce those from ab initio MD simulations. In addition, with the training set generated from GEBF-GAP, we also demonstrate that GEBF-ML force fields constructed by neural network (NN) methods can also show QM quality. Therefore, the present work provides an efficient and systematic way to build QM quality force fields for biological systems.

Machine learning potentials for complex aqueous systems made simple

Understanding complex materials, in particular those with solid–liquid interfaces, such as water on surfaces or under confinement, is a key challenge for technological and scientific progress. Although established simulation approaches have been able to provide important atomistic insight, ab initio techniques struggle with the required time and length scales, while force field methods can often be limited in terms of their accuracy. Here we show how these limitations can be overcome in a simple and automated machine learning procedure to provide accurate models of interactions at the ab initio level, as illustrated for a variety of complex aqueous systems. These developments open up the prospect of the straightforward exploration of many technologically relevant systems by molecular simulations.

Simulation techniques based on accurate and efficient representations of potential energy surfaces are urgently needed for the understanding of complex systems such as solid–liquid interfaces. Here we present a machine learning framework that enables the efficient development and validation of models for complex aqueous systems. Instead of trying to deliver a globally optimal machine learning potential, we propose to develop models applicable to specific thermodynamic state points in a simple and user-friendly process. After an initial ab initio simulation, a machine learning potential is constructed with minimum human effort through a data-driven active learning protocol. Such models can afterward be applied in exhaustive simulations to provide reliable answers for the scientific question at hand or to systematically explore the thermal performance of ab initio methods. We showcase this methodology on a diverse set of aqueous systems comprising bulk water with different ions in solution, water on a titanium dioxide surface, and water confined in nanotubes and between molybdenum disulfide sheets. Highlighting the accuracy of our approach with respect to the underlying ab initio reference, the resulting models are evaluated in detail with an automated validation protocol that includes structural and dynamical properties and the precision of the force prediction of the models. Finally, we demonstrate the capabilities of our approach for the description of water on the rutile titanium dioxide (110) surface to analyze the structure and mobility of water on this surface. Such machine learning models provide a straightforward and uncomplicated but accurate extension of simulation time and length scales for complex systems.

Accelerating Metadynamics-Based Free-Energy Calculations with Adaptive Machine Learning Potentials

There is an increasing demand for free-energy calculations using ab initio molecular dynamics these days. Metadynamics (MetaD) is frequently utilized to reconstruct the free-energy surface, but it is often computationally intractable for the first-principles calculations. Machine learning potentials (MLPs) have become popular alternatives. However, the training could be a long and arduous process before using them in practical applications. To accelerate MetaD use with MLPs for the free-energy calculation in an easy manner, we propose the adaptive machine learning potential-accelerated metadynamics (AMLP-MetaD). In this method, the MLP in the form of a Gaussian approximation potential (GAP) can adapt itself based on its uncertainty estimation, which decides whether to accept the model prediction or recalculate it with a reference method (usually density functional theory) for further training during the MetaD simulation. We demonstrate that the free-energy landscape similar to the ab initio one can be obtained using AMLP-MetaD with a 10-time speedup. Moreover, the quality of the free-energy results can be deeply improved using Δ-MLP, which is the GAP-corrected density functional tight binding in our case. We exemplify this novel method with two model systems, CO adsorption on the Pt₁₃ cluster and the Pt(111) surface, which are of vital importance in heterogeneous catalysis. The successful application in these two tests highlights that our proposed method can be used in both cluster and periodic systems and for up to two collective variables.

A review on machine learning algorithms for the ionic liquid chemical space

There are thousands of papers published every year investigating the properties and possible applications of ionic liquids. Industrial use of these exceptional fluids requires adequate understanding of their physical properties, in order to create the ionic liquid that will optimally suit the application. Computational property prediction arose from the urgent need to minimise the time and cost that would be required to experimentally test different combinations of ions. This review discusses the use of machine learning algorithms as property prediction tools for ionic liquids (either as standalone methods or in conjunction with molecular dynamics simulations), presents common problems of training datasets and proposes ways that could lead to more accurate and efficient models.

Application of Reaction Path Search Calculations to Potential Energy Surface Fits

There has been significant progress in recent years in the use of machine learning techniques to model high-dimensional reactive potential energy surfaces using large-scale data obtained from ab initio electronic structure calculations. In these methods, the strategy used to gather data becomes a key issue as the molecular size increases. In this work, we examine the applicability of the reaction path search algorithm implemented in the Global Reaction Route Mapping (GRRM) code as a data-gathering approach. The electronic energies and gradients sampled by using the GRRM calculation are directly used in potential energy surface fitting to a permutationally invariant polynomial function. This simple approach was applied to the HNS and HCNO reaction systems, and we found that the fitted potential energy surfaces reasonably reproduce the features of the electronic structure calculations used in the GRRM calculations. This suggests that the GRRM sampling scheme can be used to construct an initial potential energy surface.

Searching Configurations in Uncertainty Space: Active Learning of High-Dimensional Neural Network Reactive Potentials

Neural network (NN) potential energy surfaces (PESs) have been widely used in atomistic simulations with ab initio accuracy. While constructing NN PESs, their training data points are often sampled by molecular dynamics trajectories. This strategy can be however inefficient for reactive systems involving rare events. Here, we develop an uncertainty-driven active learning strategy to automatically and efficiently generate high-dimensional NN-based reactive potentials, taking a gas-surface reaction as an example. The difference between two independent NN models is used as a simple and differentiable uncertainty metric, allowing us to quickly search in the uncertainty space and place new samples at which the PES is less reliable. By interfacing this algorithm with the first-principles simulation package, we demonstrate that a globally accurate NN potential of the H₂ + Ag(111) system can be constructed with merely ∼150 data points. This PES can be further refined to describe H₂ dissociation on Ag(100) by adding ∼130 more configurations on this facet. The entire process is completely automatic and self-terminated once the relative error criterion is fulfilled. Impressively, data points sampled by this uncertainty-driven strategy are substantially fewer than by the traditional trajectory-based sampling. The final NN PES not only converges well the quantum dissociation probability of the molecule but also well-reproduces the phonon properties of the substrate and is capable of describing surface temperature effects. These results show the potential of this active learning approach in developing high-dimensional NN reactive potentials in gas and condensed phases.

In Situ Surface Structures of PdAg Catalyst and Their Influence on Acetylene Semihydrogenation Revealed by Machine Learning and Experiment

PdAg alloy is an industrial catalyst for acetylene-selective hydrogenation in excess ethene. While significant efforts have been devoted to increase the selectivity, there has been little progress in the catalyst performance at low temperatures. Here by combining a machine-learning atomic simulation and catalysis experiment, we clarify the surface status of PdAg alloy catalyst under the reaction conditions and screen out a rutile-TiO₂ supported Pd₁Ag₃ catalyst with high performance: i.e., 85% selectivity at >96% acetylene conversion over a 100 h period in an experiment. The machine-learning global potential energy surface exploration determines the Pd-Ag-H bulk and surface phase diagrams under the reaction conditions, which reveals two key bulk compositions, Pd₁Ag₁ (R3̅m) and Pd₁Ag₃ (Pm3̅m), and quantifies the surface structures with varied Pd:Ag ratios under the reaction conditions. We show that the catalyst activity is controlled by the PdAg patterns on the (111) surface that are variable under reaction conditions, but the selectivity is largely determined by the amount of Pd exposure on the (100) surface. These insights provide the fundamental basis for the rational design of a better catalyst via three measures: (i) controlling the Pd:Ag ratio at 1:3, (ii) reducing the nanoparticle size to limit PdAg local patterns, (iii) searching for active supports to terminate the (100) facets.

Zirconia-Supported ZnO Single Layer for Syngas Conversion Revealed from Machine-Learning Atomic Simulation

ZnZrO ternary oxide represents a prominent catalytic system, identified recently for syngas conversion and CO₂ reduction via OX-ZEO technology. One intriguing observation of the ZnZrO catalyst is the very low amount of Zn required for achieving high activity, which challenges the current views on the active site of binary oxide catalysts. Herein, we demonstrate, via machine-learning-based atomic simulation, that the structure evolution of the ZnZrO system in synthesis can be traced from bulk to surface, which leads to the identification of the active site of the ZnZrO catalyst. Theory shows that an unprecedented single-layer Zn-O structure can adhere strongly to the monoclinic ZrO₂ minority (001) surface, forming a stable oxide-on-oxide interface Zn-O/M(001). The single-layer Zn-O can convert syngas to methanol with a high turnover frequency (7.38 s^-1) from microkinetics simulation. Electron structure analyses reveal that the pentahedron [ZnO₄] in Zn-O/M(001) enhances the surface electron donation to promote the catalytic activity.

The dome of gold nanolized for catalysis

Gold is noble in bulk but turns out to be a superior catalyst at the nanoscale when supported on oxides, in particular titania. The critical thickness for activity, namely two-layer gold particles on titania, observed two decades ago represents one of the most influential mysteries in the recent history of heterogeneous catalysis. By developing a Bayesian optimization controlled global potential energy surface exploration tool with machine learning potential, here we determine the atomic structures of gold particles within ∼2 nm on a TiO2 surface. We show that the smallest stable Au nanoparticle is Au24 which is pinned on the oxygen-rich TiO2 and exhibits an unprecedented dome architecture made by a single-layer Au sheet but with an apparent two-atomic-layer height. Importantly, this has the highest activity for CO oxidation at room temperature. The physical origin of the high activity is the outstanding electron storage ability of the nano-dome, which activates the lattice oxygen of the oxide. The determined CO oxidation mechanism, the simulated rate and the fitted apparent energy barrier are consistent with known experimental facts, providing key evidence for the presence of both the high-activity Au dome and the low-activity close-packed Au particles in real catalysts. The future direction for the preparation of active and stable Au-based catalysts is thus outlined.