arXiv subjects

Oliver T. Unke

Publications and source records attributed to Oliver T. Unke.

26 records · Page 2Linked to original sources

arXiv2021

Machine Learning Force Fields

In recent years, the use of Machine Learning (ML) in computational chemistry has enabled numerous advances previously out of reach due to the computational complexity of traditional electronic-structure methods. One of the most promising applications is the construction of ML-based force fields (FFs), with the aim to narrow the gap between the accuracy of ab initio methods and the efficiency of classical FFs. The key idea is to learn the statistical relation between chemical structure and potential energy without relying on a preconceived notion of fixed chemical bonds or knowledge about the relevant interactions. Such universal ML approximations are in principle only limited by the quality and quantity of the reference data used to train them. This review gives an overview of applications of ML-FFs and the chemical insights that can be obtained from them. The core concepts underlying ML-FFs are described in detail and a step-by-step guide for constructing and testing them from scratch is given. The text concludes with a discussion of the challenges that remain to be overcome by the next generation of ML-FFs.

physics.chem-ph↗

arXiv2020

Isomerization and Decomposition Reactions of Acetaldehyde Relevant to Atmospheric Processes from Dynamics Simulations on Neural Network-Based Potential Energy Surfaces

Acetaldehyde (AA) isomerization (to vinylalcohol, VA) and decomposition (into either CO+CH$_4$ and H$_2$+H$_2$CCO) is studied using a fully dimensional, reactive potential energy surface represented as a neural network (NN). The NN, trained on 432'399 reference structures from MP2/aug-cc-pVTZ calculations has a MAE of 0.0453 kcal/mol and an RMSE of 1.186 kcal/mol for a test set of 27'399 structures. For the isomerization process AA $\rightarrow$ VA the minimum dynamical path implies that the C-H vibration, and the C-C-H (with H being the transferring H-atom) and the C-C-O angles are involved to surmount the 68.2 kcal/mol barrier. Using an excess energy of 93.6 kcal/mol - the energy available in the solar spectrum and sufficient to excite to the first electronically excited state - to initialize the molecular dynamics, no isomerization to VA is observed on the 500 ns time scale. Only with excess energies of $\sim$ 127.6 kcal/mol (including the zero point energy of the AA molecule), isomerization occurs on the nanosecond time scale. Given that collisional de-excitation at atmospheric conditions in the stratosphere occurs on the 100 ns time scale, it is concluded that formation of VA following photoexcitation of AA from actinic photons is unlikely. This also limits the relevance of this reaction pathway to be a source for formic acid.

physics.chem-ph↗

arXiv2020

Thermal Activation of Methane by MgO$^+$: Temperature Dependent Kinetics, Reactive Molecular Dynamics Simulations and Statistical Modeling

The kinetics of MgO$^+$ + CH$_4$ was studied experimentally using the variable ion source, temperature adjustable selected ion flow tube (VISTA-SIFT) apparatus from 300 $-$ 600 K and computationally by running and analyzing reactive atomistic simulations. Rates and product branching fractions were determined as a function of temperature. The reaction proceeded with a rate of $k = 5.9 \pm 1.5 10^{-10}(T/300 $ K$)^{-0.5 \pm 0.2}$ cm$^3$ s$^{-1}$. MgOH$^+$ was the dominant product at all temperatures, but Mg$^+$, the co-product of oxygen-atom transfer to form methanol, was observed with a product branching fraction of $0.08 \pm 0.03 (T / 300 $ K$)^{-0.8 \pm 0.7}$. Reactive molecular dynamics simulations using a reactive force field, as well as a neural network yield rate coefficients about one order of magnitude lower. This underestimation of the rates is traced back to the multireference character of the transition state [MgOCH$_4$]$^+$. Statistical modeling of the temperature-dependent kinetics provides further insight into the reactive potential surface. The rate limiting step was found to be consistent with a four-centered activation of the C-H bond, consistent with previous calculations. The product branching was modeled as a competition between dissociation of an insertion intermediate directly after the rate-limiting transition state, and traversing a transition state corresponding to a methyl migration leading to a Mg-CH$_3$OH$^+$ complex, though only if this transition state is stabilized significantly relative to the dissociated MgOH$^+$ + CH$_3$ product channel. An alternative non-statistical mechanism is discussed, whereby a post-transition state bifurcation in the potential surface could allow the reaction to proceed directly from the four-centered TS to the Mg-CH$_3$OH$^+$ complex thereby allowing a more robust competition between the product channels.

physics.chem-ph↗

arXiv2019

Reactive Dynamics and Spectroscopy of Hydrogen Transfer from Neural Network-Based Reactive Potential Energy Surfaces

The in silico exploration of chemical, physical and biological systems requires accurate and efficient energy functions to follow their nuclear dynamics at a molecular and atomistic level. Recently, machine learning tools gained a lot of attention in the field of molecular sciences and simulations and are increasingly used to investigate the dynamics of such systems. Among the various approaches, artificial neural networks (NNs) are one promising tool to learn a representation of potential energy surfaces. This is done by formulating the problem as a mapping from a set of atomic positions $\mathbf{x}$ and nuclear charges $Z_i$ to a potential energy $V(\mathbf{x})$. Here, a fully-dimensional, reactive neural network representation for malonaldehyde (MA), acetoacetaldehyde (AAA) and acetylacetone (AcAc) is learned. It is used to run finite-temperature molecular dynamics simulations, and to determine the infrared spectra and the hydrogen transfer rates for the three molecules. The finite-temperature infrared spectrum for MA based on the NN learned on MP2 reference data provides a realistic representation of the low-frequency modes and the H-transfer band whereas the CH vibrations are somewhat too high in frequency. For AAA it is demonstrated that the IR spectroscopy is sensitive to the position of the transferring hydrogen at either the OCH- or OCCH$_3$ end of the molecule. For the hydrogen transfer rates it is demonstrated that the O-O vibration is a gating mode and largely determines the rate at which the hydrogen is transferred between the donor and acceptor. Finally, possibilities to further improve such NN-based potential energy surfaces are explored. They include the transferability of an NN-learned energy function across chemical species (here methylation) and transfer learning from a lower level of reference data (MP2) to a higher level of theory (pair natural orbital-LCCSD(T)).

physics.chem-ph↗

arXiv2019

High-Dimensional Potential Energy Surfaces for Molecular Simulations

An overview of computational methods to describe high-dimensional potential energy surfaces suitable for atomistic simulations is given. Particular emphasis is put on accuracy, computability, transferability and extensibility of the methods discussed. They include empirical force fields, representations based on reproducing kernels, using permutationally invariant polynomials, and neural network-learned representations and combinations thereof. Future directions and potential improvements are discussed primarily from a practical, application-oriented perspective.

physics.chem-ph↗

arXiv2019

Machine Learning Potential Energy Surfaces

Machine Learning techniques can be used to represent high-dimensional potential energy surfaces for reactive chemical systems. Two such methods are based on a reproducing kernel Hilbert space representation or on deep neural networks. They can achieve a sub-1 kcal/mol accuracy with respect to reference data and can be used in studies of chemical dynamics. Their construction and a few typical examples are briefly summarized in the present contribution.

physics.chem-ph↗

arXiv2019

Reactive Atomistic Simulations of Diels-Alder Reactions: the Importance of Molecular Rotations

The Diels-Alder reaction between 2,3-dibromo-1,3-butadiene and maleic anhydride has been studied by means of multisurface adiabatic reactive molecular dynamics and the PhysNet neural network architecture. This system is used as a prototype to explore the concertedness, synchronicity and possible ways of promotion of Diels-Alder reactions. Analysis of the minimum dynamic path indicates that rotational energy is crucial ($\sim$ 65 %) to drive the system towards the transition state in addition to collision energy ($\sim$ 20 %). Comparison with the reaction of butadiene and maleic anhydride shows that the presence of bromine substituents in the diene accentuates the importance of rotational excitation to promote the reaction. At the high total energies at which reactive events are recorded, the reaction is found to be direct and mostly synchronous.

physics.chem-ph↗

arXiv2019

PhysNet: A Neural Network for Predicting Energies, Forces, Dipole Moments and Partial Charges

In recent years, machine learning (ML) methods have become increasingly popular in computational chemistry. After being trained on appropriate ab initio reference data, these methods allow to accurately predict the properties of chemical systems, circumventing the need for explicitly solving the electronic Schrödinger equation. Because of their computational efficiency and scalability to large datasets, deep neural networks (DNNs) are a particularly promising ML algorithm for chemical applications. This work introduces PhysNet, a DNN architecture designed for predicting energies, forces and dipole moments of chemical systems. PhysNet achieves state-of-the-art performance on the QM9, MD17 and ISO17 benchmarks. Further, two new datasets are generated in order to probe the performance of ML models for describing chemical reactions, long-range interactions, and condensed phase systems. It is shown that explicitly including electrostatics in energy predictions is crucial for a qualitatively correct description of the asymptotic regions of a potential energy surface (PES). PhysNet models trained on a systematically constructed set of small peptide fragments (at most eight heavy atoms) are able to generalize to considerably larger proteins like deca-alanine (Ala$_{10}$): The optimized geometry of helical Ala$_{10}$ predicted by PhysNet is virtually identical to ab initio results (RMSD = 0.21 Å). By running unbiased molecular dynamics (MD) simulations of Ala$_{10}$ on the PhysNet-PES in gas phase, it is found that instead of a helical structure, Ala$_{10}$ folds into a wreath-shaped configuration, which is more stable than the helical form by 0.46 kcal mol$^{-1}$ according to the reference ab initio calculations.

physics.chem-ph↗