4.7 Article

Electronic Exchange and Correlation in van der Waals Systems: Balancing Semilocal and Nonlocal Energy Contributions

期刊

JOURNAL OF CHEMICAL THEORY AND COMPUTATION
卷 14, 期 3, 页码 1361-1369

出版社

AMER CHEMICAL SOC
DOI: 10.1021/acs.jctc.7b01172

关键词

-

资金

  1. European Research Council (ERC), Consolidator Grant BeStMo

向作者/读者索取更多资源

Short-range correlations in motion of electrons in matter are captured well by semilocal exchange-correlation (XC) functionals in density functional theory (DFT), but long-range correlations are neglected in such models and must be treated by van der Waals (vdW) dispersion methods. Whereas the effective range of distances at which fluctuations are correlated is usually explicit in the vdW models, the complementary range of semilocal functionals can be observed only implicitly, requiring an introduction of empirical damping functions to couple the semilocal and nonlocal contributions to the XC energy. We present a comprehensive study of the interplay between these short-range and long-range energy contributions in eight semilocal functionals (LDA, PBE, TPSS, SCAN, PBE0, B3LYP, SCAN0, M06-L) and three vdW models (MBD, D3, VV10) on noncovalently bonded organic dimers (S66x8), molecular crystals (X23), and supramolecular complexes (S12L), as well as on a series of graphene-flake dimers, covering a range of intermolecular distances and binding energies (0.5-130 kcal/mol). The binding-energy profiles of many of the DFT+vdW combinations differ both quantitatively and qualitatively, and some of the qualitative differences are independent of the choice of the vdW model, establishing them as intrinsic properties of the respective semilocal functionals. We find that while the SCAN+vdW method yields a narrow range of binding-energy errors, the effective range of SCAN depends on system size, and we link this behavior to the specific dependence of SCAN on the electron localization function alpha around alpha = 1. Our study provides a systematic procedure to evaluate the consistency of semilocal XC functionals when paired with nonlocal vdW models and leads us to conclude that nonempirical generalized-gradient and hybrid functionals are currently among the most balanced semilocal choices for vdW systems.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.7
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

Article Physics, Multidisciplinary

Casimir Self-Interaction Energy Density of Quantum Electrodynamic Fields

Alexandre Tkatchenko, Dmitry V. Fedorov

Summary: Quantum electrodynamic fields exhibit fluctuations in the form of particle-antiparticle dipoles, characterized by a nonvanishing polarizability density. In this study, a quantum scaling law is extended to describe the volumetric and radial polarizability density of a quantum field associated with electrons and positrons, and the Casimir self-interaction energy density (E over bar SIE) of the field is derived in terms of the fine-structure constant. The proposed model satisfies the cosmological equation of state w = -1 and the calculated E over bar SIE falls within the range of the recent measurements of the cosmological constant ? obtained by the Planck Mission and the Hubble Space Telescope.

PHYSICAL REVIEW LETTERS (2023)

Article Multidisciplinary Sciences

Electronic excited states in deep variational Monte Carlo

M. T. Entwistle, Z. Schaetzle, P. A. Erdman, J. Hermann, F. Noe

Summary: Deep neural networks can accurately learn and represent electronic ground and excited states, enabling the modeling of various excited-state processes for atoms and molecules.

NATURE COMMUNICATIONS (2023)

Article Multidisciplinary Sciences

Accurate global machine learning force fields for molecules with hundreds of atoms

Stefan Chmiela, Valentin Vassilev-Galindo, Oliver T. Unke, Adil Kabylda, Huziel E. Sauceda, Alexandre Tkatchenko, Klaus-Robert Mueller

Summary: We have developed an exact iterative approach to train global symmetric gradient domain machine learning (sGDML) force fields, which can accurately describe complex molecular systems and materials. We evaluated the accuracy and efficiency of sGDML on a newly developed MD22 benchmark dataset containing molecules from 42 to 370 atoms.

SCIENCE ADVANCES (2023)

Article Multidisciplinary Sciences

Efficient interatomic descriptors for accurate machine learning force fields of extended molecules

Adil Kabylda, Valentin Vassilev-Galindo, Stefan Chmiela, Igor Poltavsky, Alexandre Tkatchenko

Summary: Machine learning force fields (MLFFs) are being optimized to enable molecular dynamics simulations with ab initio accuracy but at a fraction of the computational cost. Challenges remain in developing efficient descriptors for non-local interatomic interactions and reducing dimensionality of descriptors for enhanced applicability and interpretability. An automatized approach is proposed to reduce interatomic descriptor features while maintaining accuracy and efficiency of MLFFs. The results show the importance of non-local features in preserving overall accuracy and reducing the required features to a comparable number with local interatomic features.

NATURE COMMUNICATIONS (2023)

Article Chemistry, Physical

DeepQMC: An open-source software suite for variational optimization of deep-learning molecular wave functions

Z. Schaetzle, P. B. Szabo, M. Mezera, J. Hermann, F. Noe

Summary: Computing accurate and efficient approximations to solve the Schrödinger equation in computational chemistry has been a challenge for decades. Quantum Monte Carlo methods, with their highly parallel and scalable algorithm, show promise in achieving high accuracy in a variety of molecular systems. The use of machine-learned parametrizations, relying on neural networks as universal function approximators, has further improved the accuracy of these methods. The development of software libraries like DEEPQMC aims to provide a common framework for future investigations and make this field accessible to practitioners from both the quantum chemistry and machine learning communities.

JOURNAL OF CHEMICAL PHYSICS (2023)

Article Chemistry, Physical

Universal Pairwise Interatomic van der Waals Potentials Based on Quantum Drude Oscillators

Almaz Khabibrakhmanov, Dmitry V. Fedorov, Alexandre Tkatchenko

Summary: This study presents a universal parameterization method for quantum-mechanical van der Waals (vdW) potentials based on two free-atom properties, namely the static dipole polarizability and the dipole-dipole C-6 dispersion coefficient. The derived vdW-QDO potential accurately predicts vdW binding energy curves for noble-gas dimers and exhibits correct asymptotic behavior. It is also shown to accurately describe vdW interactions in dimers consisting of group II elements. The applicability of the atom-in-molecule vdW-QDO model for predicting dispersion energies for molecular systems is demonstrated. This work is an important step toward constructing universal vdW potentials for (bio)molecular computational studies.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2023)

Article Chemistry, Physical

Optimized Quantum Drude Oscillators for Atomic and Molecular Response Properties

Szabolcs Goger, Almaz Khabibrakhmanov, Ornella Vaccarelli, Dmitry V. Fedorov, Alexandre Tkatchenko

Summary: The quantum Drude oscillator (QDO) is an efficient coarse-grained approach used to model electronic and optical response properties of atoms and molecules. An optimized parametrization (OQDO) is presented in this study, where the parameters are fixed using only dipolar properties. The OQDO accurately reproduces atomic polarization potentials and multipolar dispersion coefficients for the periodic table of elements and small molecules, showing great potential in the development of next-generation quantum-mechanical force fields for (bio)molecular simulations.

JOURNAL OF PHYSICAL CHEMISTRY LETTERS (2023)

Correction Multidisciplinary Sciences

Efficient interatomic descriptors for accurate machine learning force fields of extended molecules (vol 14, 3562, 2023)

Adil Kabylda, Valentin Vassilev-Galindo, Stefan Chmiela, Igor Poltavsky, Alexandre Tkatchenko

NATURE COMMUNICATIONS (2023)

Review Chemistry, Multidisciplinary

Ab initio quantum chemistry with neural-network wavefunctions

Jan Hermann, James Spencer, Kenny Choo, Antonio Mezzacapo, W. M. C. Foulkes, David Pfau, Giuseppe Carleo, Frank Noe

Summary: Deep learning methods have surpassed human capabilities in pattern recognition and data processing, and have become increasingly important in scientific discovery. In molecular science, a key application of machine learning is to learn potential energy surfaces or force fields from ab initio solutions of the electronic Schrodinger equation obtained with various quantum chemistry methods. This review discusses a complementary approach that uses machine learning to directly solve quantum chemistry problems from first principles, focusing on quantum Monte Carlo methods with neural-network ansatzes to solve the electronic Schrodinger equation.

NATURE REVIEWS CHEMISTRY (2023)

Article Chemistry, Multidisciplinary

Accurate quantification of the stability of the perylene-tetracarboxylic dianhydride on Au(111) molecule-surface interface

Victor G. Ruiz, Christian Wagner, Friedrich Maass, Hadi H. Arefi, Stephan Stremlau, Petra Tegeder, F. Stefan Tautz, Alexandre Tkatchenko

Summary: The authors accurately characterized the adsorption energy of perylene-tetracarboxylic dianhydride molecules on Au(111) using temperature-programmed desorption, single-molecule atomic force microscopy, and non-local density-functional theory. Studying inorganic/organic hybrid systems is crucial for designing complex interfaces.

COMMUNICATIONS CHEMISTRY (2023)

Article Chemistry, Multidisciplinary

Freedom of design in chemical compound space: towards rational in silico design of molecules with targeted quantum-mechanical properties

Leonardo Medrano Sandonas, Johannes Hoja, Brian G. Ernst, Alvaro Vazquez-Mayagoitia, Robert A. Distasio, Alexandre Tkatchenko

Summary: The rational design of molecules with specific quantum-mechanical properties requires an understanding of the relationships between structure-property and property-property in chemical compound space. In this study, the analysis of a comprehensive dataset revealed a high degree of flexibility in designing molecules with desired properties or a set of distinct molecules with various properties. Using multi-objective optimization, non-trivial paths through chemical compound space were identified that lead to molecules with optimal combinations of polarizabilities and HOMO-LUMO gaps.

CHEMICAL SCIENCE (2023)

Article Chemistry, Physical

Data-driven tailoring of molecular dipole polarizability and frontier orbital energies in chemical compound space

Szabolcs Goger, Leonardo Medrano Sandonas, Carolin Mueller, Alexandre Tkatchenko

Summary: Understanding the correlations - or lack thereof - between molecular properties is crucial for efficient molecular design. This study explores the relationship between electronic structure and chemical properties in molecular systems, specifically the energy gap and dipole polarizability. Through analysis of a comprehensive dataset and molecular composition, it is demonstrated that there is no correlation between polarizability and HOMO-LUMO gap for sufficiently diverse chemical compounds. The lack of correlation allows for the design of novel materials, exemplified by the case of organic photodetector candidates.

PHYSICAL CHEMISTRY CHEMICAL PHYSICS (2023)

Article Physics, Multidisciplinary

Anisotropic van der Waals dispersion forces in polymers: Structural symmetry breaking leads to enhanced conformational search

Mario Galante, Alexandre Tkatchenko

Summary: The modeling of conformations and dynamics is crucial for understanding the properties of (bio)polymers in soft matter. Noncovalent interactions, rather than short-range interactions, play a dominant role in determining the global conformations. Considering many-body effects in van der Waals (vdW) dispersion has been shown to guide the conformation towards globally optimized structures by reducing the roughness of the energy landscape and promoting global spatial symmetries.

PHYSICAL REVIEW RESEARCH (2023)

Article Physics, Multidisciplinary

Modeling noncovalent interatomic interactions on a photonic quantum computer

Matthieu Sarkis, Alessio Fallani, Alexandre Tkatchenko

Summary: Noncovalent interactions play a crucial role in determining the structure, stability, and dynamics of materials, molecules, and biological complexes. However, accurately modeling these interactions on classical computers is challenging. In this study, we demonstrate the potential of the Coulomb-coupled quantum Drude oscillator (cQDO) model for simulating noncovalent interactions on a photonic quantum computer. We calculate the binding energy curve of diatomic systems using Xanadu's STRAWBERRY FIELDS photonics library. Our findings significantly expand the application of quantum computing to atomistic modeling, beyond the standard electronic-structure problem of small molecules. We also propose efficient functional forms for cQDO wave functions that can be optimized on classical computers and capture the bonded-to-noncovalent transition with increasing interatomic distances.

PHYSICAL REVIEW RESEARCH (2023)

Article Chemistry, Multidisciplinary

Modeling molecular ensembles with gradient-domain machine learning force fields

Alex M. Maldonado, Igor Poltavsky, Valentin Vassilev-Galindo, Alexandre Tkatchenko, John A. Keith

Summary: Gradient-domain machine learning (GDML) force fields show excellent accuracy, data efficiency, and applicability for molecules, and a many-body approach opens the possibility of increased transferability to molecular ensembles.

DIGITAL DISCOVERY (2023)

暂无数据