The bond capacity electronegativity equilibration charge model (EEQBC) for the elements Z = 1–103 DOI
Thomas Froitzheim, Marcel Müller, Andreas Hansen

и другие.

The Journal of Chemical Physics, Год журнала: 2025, Номер 162(21)

Опубликована: Июнь 3, 2025

The accurate and efficient assignment of atomic partial charges is crucial for many applications in theoretical computational chemistry, including polarizable force fields, dispersion corrections, charge-dependent basis sets. Classical charge models struggle to distinguish between neutral zwitterionic fragments because, unlike quantum mechanical methods, there are no discrete electronic states. This limitation can lead either reduced or additional artificial transfer (CT) at different interfragment distances. To address this issue, we propose a new version bond capacity electronegativity equilibration (EEQBC) model, which limits CT distant the simple EEQ framework. EEQBC offers excellent agreement with DFT-based reference elements up lawrencium (Z = 103) mean absolute errors as low 0.02 0.07 e- random PubChem molecules "mindless" (MLMs), respectively. Thanks its efficiency both their analytical nuclear gradients, highly suitable an initial guess next-generation tight-binding methods. For seamless accessibility, implemented upcoming 0.5.0 release freely available multicharge program github.com/grimme-lab/multicharge.

Язык: Английский

Accurate and Affordable Simulation of Molecular Infrared Spectra with AIQM Models DOI

Yi-Fan Hou,

Cheng Wang, Pavlo O. Dral

и другие.

The Journal of Physical Chemistry A, Год журнала: 2025, Номер unknown

Опубликована: Апрель 14, 2025

Infrared (IR) spectroscopy is a potent tool for identifying molecular structures and studying the chemical properties of compounds, hence, various theoretical approaches have been developed to simulate predict IR spectra. However, based on quantum calculations suffer from high computational cost (e.g., density functional theory, DFT) or insufficient accuracy semiempirical methods orders magnitude faster than DFT). Here, we introduce new approach, universal machine learning (ML) models AIQM series targeting CCSD(T)/CBS level, that can deliver spectra with close DFT (compared experiment) speed GFN2-xTB method. This approach harmonic oscillator approximation frequency scaling factors fitted experimental data. While benchmarks reported here are focused spectra, our implementation supports anharmonic simulations via dynamics VPT2. These implementations available in MLatom as described https://github.com/dralgroup/mlatom be performed online web browser.

Язык: Английский

Процитировано

0
Beyond Numerical Hessians: Higher-Order Derivatives for Machine Learning Interatomic Potentials via Automatic Differentiation DOI

Nils Gönnheimer,

Karsten Reuter, Johannes T. Margraf

и другие.

Journal of Chemical Theory and Computation, Год журнала: 2025, Номер unknown

Опубликована: Апрель 24, 2025

The development of machine learning interatomic potentials (MLIPs) has revolutionized computational chemistry by enhancing the accuracy empirical force fields while retaining a large speed-up compared to first-principles calculations. Despite these advancements, calculation Hessian matrices for systems remains challenging, in particular because analytical second-order derivatives are often not implemented. This necessitates use computationally expensive finite-difference methods, which can furthermore display low precision some cases. Automatic differentiation (AD) offers promising alternative reduce this effort and makes more efficient accurate. Here, we present implementation AD-based popular MACE equivariant graph neural network architecture. benefits method showcased via high-throughput prediction heat capacities porous materials with MACE-MP-0 foundation model. is essential precisely describing gas adsorption was previously possible only bespoke ML models or We find that availability accurate comparable zero-shot manner additionally allows investigation finite-size rounding errors data.

Язык: Английский

Процитировано

0
Unlocking the Potential of Machine Learning in Enhancing Quantum Chemical Calculations for Infrared Spectral Prediction DOI Creative Commons

Adithya Ranjith Kartha,

Dhanush P. Ajayakumar,

Muhammad Zaffwan Idris

и другие.

ACS Omega, Год журнала: 2025, Номер unknown

Опубликована: Апрель 28, 2025

Infrared (IR) spectroscopy is a fundamental tool for analyzing molecular structures and chemical interactions by identifying the vibrational modes of molecules. Traditional quantum mechanical methods, such as density functional theory, are highly accurate but computationally expensive impractical large-scale systems. This project investigates integration machine learning (ML) techniques to predict IR spectra, offering promising alternative that significantly reduces computational costs while maintaining high accuracy. Additionally, explores utilization spectra identification classification into families, enhancing practical utility spectral data in various scientific applications. Using TensorFlow-based ML frameworks, models were developed trained on set derived from high-quality chemistry analyzers. These sets, sourced optimized geometry spectrum Gaussian 16 Program Suite, include extensive data, bond lengths, modes, other properties. The aim key features, frequencies intensities, interpretability linking principles predictions. with provides scalable well accelerated solution complex approach holds potential fields drug discovery, materials science, engineering, where rapid predictions critical. perspective highlights advancements achieved, current challenges, future context spectroscopy, providing solid foundation further exploration at intersection science.

Язык: Английский

Процитировано

0
Computing Bulk Phase IR Spectra from Finite Cluster Data via Equivariant Neural Networks DOI
Aman Jindal, Philipp Schienbein, Banshi Das

и другие.

Journal of Chemical Theory and Computation, Год журнала: 2025, Номер unknown

Опубликована: Май 17, 2025

Calculating accurate IR spectra from molecular dynamics simulations is crucial for understanding structural and benchmarking simulations. While machine learning has accelerated such calculations, leveraging finite-cluster data to compute condensed-phase remains unexplored. In this work, we address a fundamental question: Can model trained exclusively on electronic structure calculations of finite-size clusters reproduce the bulk spectrum? Using atomic polar tensor as target training property, demonstrate that corresponding equivariant neural network accurately recovers spectrum liquid water, establishing key link between properties.

Язык: Английский

Процитировано

0
The bond capacity electronegativity equilibration charge model (EEQBC) for the elements Z = 1–103 DOI
Thomas Froitzheim, Marcel Müller, Andreas Hansen

и другие.

The Journal of Chemical Physics, Год журнала: 2025, Номер 162(21)

Опубликована: Июнь 3, 2025

The accurate and efficient assignment of atomic partial charges is crucial for many applications in theoretical computational chemistry, including polarizable force fields, dispersion corrections, charge-dependent basis sets. Classical charge models struggle to distinguish between neutral zwitterionic fragments because, unlike quantum mechanical methods, there are no discrete electronic states. This limitation can lead either reduced or additional artificial transfer (CT) at different interfragment distances. To address this issue, we propose a new version bond capacity electronegativity equilibration (EEQBC) model, which limits CT distant the simple EEQ framework. EEQBC offers excellent agreement with DFT-based reference elements up lawrencium (Z = 103) mean absolute errors as low 0.02 0.07 e- random PubChem molecules "mindless" (MLMs), respectively. Thanks its efficiency both their analytical nuclear gradients, highly suitable an initial guess next-generation tight-binding methods. For seamless accessibility, implemented upcoming 0.5.0 release freely available multicharge program github.com/grimme-lab/multicharge.

Язык: Английский

Процитировано

0