Ab Initio Level Research Articles

A Bond-Based Machine Learning Model for Molecular Polarizabilities and A Priori Raman Spectra.

The use of machine learning (ML) algorithms in molecular simulations has become commonplace in recent years. There now exists, for instance, a multitude of ML force field algorithms that have enabled simulations approaching ab initio level accuracy at time scales and system sizes that significantly exceed what is otherwise possible with traditional methods. Far fewer algorithms exist for predicting rotationally equivariant, tensorial properties such as the electric polarizability. Here, we introduce a kernel ridge regression algorithm for machine learning of the polarizability tensor. This algorithm is based on the bond polarizability model and allows prediction of the tensor components at the cost similar to that of scalar quantities. We subsequently show the utility of this algorithm by simulating gas phase Raman spectra of biphenyl and malonaldehyde using classical molecular dynamics simulations of these systems performed with the recently developed MACE-OFF23 potential. The calculated spectra are shown to agree very well with the experiments and thus confirm the expediency of our algorithm as well as the accuracy of the used force field. More generally, this work demonstrates the potential of physics-informed approaches to yield simple yet effective machine learning algorithms for molecular properties.

A simplified method for theoretical sum frequency generation spectroscopy calculation and interpretation: The "pop model".

Existing methods to compute theoretical spectra are restricted to the use of time-correlation functions evaluated from accurate atomistic molecular dynamics simulations, often at the abinitio level. The molecular interpretation of the computed spectra requires additional steps to deconvolve the spectroscopic contributions from local water and surface structural populations at the interface. The lack of a standard procedure to do this often hampers rationalization. To overcome these challenges, we rewrite the equations for spectra calculation into a sum of partial contributions from interfacial populations, weighted by their abundance at the interface. We show that SFG signatures from each population can be parameterized into a minimum dataset of reference partial spectra. Accurate spectra can then be predicted by just evaluating the statistics of interfacial populations, which can be done even with force field simulations as well as with analytic models. This approach broadens the range of simulation techniques from which theoretical spectra can be calculated, opening toward non-atomistic and Monte Carlo simulation approaches. Most notably, it allows constructing accurate theoretical spectra for interfacial conditions that cannot even be simulated, as we demonstrate for the pH-dependent SFG spectra of silica/water interfaces.

Modeling conformational changes in alginic acid oligomers induced by external forces

In this study, the mechanism and nature of mechanical force-induced conformational transitions of alginate oligomers with different ratios of β-D-mannuronic acid (M unit) and α-L-guluronic acid (G unit) units were investigated. The influence of the type of glycosidic linkage in either homo- or heterooligomers on the nature of conformational transitions was also considered. For this purpose, two different theoretical methods were used: quantum mechanics (QM) at the DFT level with the EGO (Enforced Geometry Optimization) approach previously tested also for other saccharide systems, and molecular dynamics (MD) simulations within hybrid interaction potentials, which take into account both the ab initio (QM) level of theory and classical molecular mechanics (MM) force fields. This allowed to characterize in detail the structural and energetic properties of the conformational transition occurring upon the influence of external, mechanical forces (e.g. ring conformations at the path of ring-inversion process as well as the energies corresponding to initial, final and intermediate states). The results indicate qualitatively various responses against the applied force, depending on the G:M ratio, which have their source in differing topologies of glycosidic linkage in either G or M units. This is of potential relevance for determining the content of naturally heterogeneous alginate chains by the AFM experimental studies. The effects of explicit solvent and non-zero temperature are not of primarily importance in the context of determined stretching properties.

Unveiling the Role of Solvent in Solution Phase Chemical Reactions using Deep Potential-Based Enhanced Sampling Simulations.

We have used a deep learning-based active learning strategy to develop ab initio level accurate machine-learned (ML) potential for a solution-phase reactive system. Using this ML potential, we carried out enhanced sampling simulations to sample the reaction process efficiently. Multiple transitions between the reactant and product states allowed us to calculate the converged free energy surface for the reaction. As a prototypical example, we have investigated the Menshutkin reaction, a classic bimolecular nucleophilic substitution reaction (SN2) in aqueous medium. Our analyses revealed that water stabilizes the ionic product state by enhanced solvation, facilitating the reaction and making it more spontaneous. Our approach expands the scope of studying the chemical reaction under realistic conditions, such as explicit solvents at finite temperatures, closely mimicking experiments.

On the Simulation of Photoreactions Using Restricted Open-Shell Kohn-Sham Theory.

It is a well-established standard to describe ground-state chemical reactions at an ab initio level of multi-electron theory. Fast reactions can be directly simulated. The most widely used approach is density functional theory for the electronic structure in combination with molecular dynamics for the nuclear motion. This approach is known as ab initio molecular dynamics. In contrast, the simulation of excited-state reactions at this level of theory is significantly more difficult. It turns out that the self-consistent solution of the Kohn-Sham equations is not easily reached in excited-state simulations. The first program that solved this problem was the Car-Parrinello molecular dynamics code, using restricted open-shell Kohn-Sham theory. Meanwhile, there are alternatives, most prominently the Q-Chem code, which widens the range of applications. The present study investigates the suitability of both codes for the molecular dynamics simulation of excited-state motion and presents applications to photoreactions.

Toward Gaussian Process Regression Modeling of a Urea Force Field.

FFLUX is a next-generation, machine-learnt force field built on three cornerstones: quantum chemical topology, Gaussian process regression, and (high-rank) multipolar electrostatics. It is capable of performing molecular dynamics with near-quantum accuracy at a lower computational cost than standard ab initio molecular dynamics. Previous work with FFLUX was concerned with water and formamide. In this study, we go one step further and challenge FFLUX to model urea, a larger and more flexible system. In result, we have trained urea models at the B3LYP/aug-cc-pVTZ level of theory, with a mean absolute error of 0.4 kJ mol-1 and a maximum prediction error below 7.0 kJ mol-1. To test their performance in molecular dynamics simulations, two sets of FFLUX geometry optimizations were carried out: 5 dimers corresponding to energy minima and 75 random dimers. The 5 dimers were recovered with a root-mean-square deviation below 0.1 Å with respect to their ab initio references. Out of the 75 random dimers, 68% converged to the qualitatively same dimer as those obtained at the ab initio level. Furthermore, we have ranked the 5 FFLUX-optimized dimers in the order of their relative FFLUX single-point energies and compared them with the ab initio method. The energy ranking fully agreed but for one crossover between two successive minima. Finally, we have demonstrated the importance of geometry-dependent (i.e., flexible) multipole moments, showing that the lack of multipole moment flexibility can lead to average errors in the total intermolecular electrostatic energy of more than 2 orders of magnitude.

Computational study on the bifurcation mechanism in the H2CO− + CH3Cl →CH3CH2O + Cl−/H2CO + CH3 + Cl− reaction: The importance of intramolecular vibrational redistributions

We have investigated the bifurcation dynamics of the H2CO− + CH3Cl →CH3CH2O + Cl− (C-substitution channel)/H2CO + CH3 + Cl− (electron transfer channel) reaction using three different molecular dynamics (MD) methods: quasi-classical trajectory (QCT), classical MD, and ring-polymer MD simulations. All MD calculations were performed using the on-the-fly approach at the ab initio molecular orbital theory level. We found that the C-substitution channel is dominant in the classical and ring-polymer MD calculations, while the electron transfer channel is dominant in the QCT calculations. The different bifurcation mechanisms in the types of MD simulations were investigated using supervised machine-learning classification. This difference is attributed to the varying efficiency of the intramolecular vibrational redistribution process among the different MD methods. Thus, in the bifurcation mechanism of the H2CO− + CH3Cl reaction, the bifurcation fractions are primarily determined by the vibrational dynamics occurring in the later stages of the reaction.

Formation and coarsening of epitaxially-supported metal nanoclusters

This mini-review describes developments over the last ∼30 years in characterizing the nucleation & growth of epitaxially-supported metal nanoclusters (NCs) or islands during vapor deposition, as well as their post-deposition coarsening. A beyond-mean-field treatment for homogeneous nucleation & growth corrects the deficiencies of traditional treatments in describing, e.g., the island size distribution, but also necessitates consideration of the spatial distribution of islands and their capture zones. We discuss advances in modeling capabilities, including those based upon on an ab-initio level treatment of periphery diffusion kinetics, for description of the non-equilibrium growth shapes of these NCs, focusing on 2D NCs. For post-deposition coarsening of arrays of NCs, there is generally a competition between Ostwald Ripening (OR) and Smoluchowski Ripening (SR). SR is also known as Particle Migration & Coalescence. For 2D NCs in homoepitaxial systems, conventional OR is observed on pristine fcc(111) surfaces, dramatically enhanced OR in the presence of even trace amounts of chalcogens for Cu(111) and Ag(111), and anomalous OR on anisotropic fcc(110) surfaces. The unexpected discovery of SR for fcc(100) homoepitaxial systems prompted extensive analysis of the underlying diffusivities of 2D NCs as a function of size, as well as of NC coalescence dynamics. A comprehensive understanding of these processes is now available. Self-assembly of 3D NCs during deposition, issues related to heterogeneous nucleation, directed assembly, NC growth structure selection, and coarsening are addressed. For SR of 3D epitaxial NCs, recent insights into the size-dependence of diffusivity are described.

Real-Time KMC Simulation of Vacancy-Mediated Intermixing in Au@Ag Octahedral Core-Cubic Shell Nanocrystals with Ab Initio-Guided Kinetics.

Utilization of core-shell rather than monometallic nanocrystals (NCs) facilitates fine-tuning of NC properties for applications. However, compositional evolution via intermixing can degrade these properties prompting recent experimental studies. We develop an atomistic-level stochastic model for vacancy-mediated intermixing exploiting a formalism which allows incorporation at an ab initio density functional theory level of not just the thermodynamics of vacancy formation, but also relevant diffusion barriers for a vast number of possible local environments (in the core and in the shell, at the interface, and in the intermixed phase). This facilitates a predictive treatment and comprehensive understanding of intermixing on the relevant time scale (e.g., 101-103 s). In contrast, previous modeling at the atomistic level utilized only unrealistic generic prescriptions of barriers or employed simplified continuum treatments. For Au@Ag octahedral core-cubic shell NCs, our modeling not only captures the experimentally observed rate or time scale for intermixing of ∼100 s at 450 °C for 60 nm NCs, but also elucidates the underlying rate controlling processes and the effective intermixing barrier.

Tipping point to explain melamine’s non-planarity or otherwise. Can quasi-molecules be “seen” in the parent molecule?

The structure of melamine is optimized both at the DFT level with B3LYP and M06-2X functionals using AVTZ basis sets, and at ab initio level using DF-CCSD(T)/VDZ. Predicted planar with DFT when adding the zero-point-energy correction, it is a planar but distorted hexagonal triazine ring with three pyramidal amino groups at the coupled-cluster level. The potential energy surface of melanine is then expected to be rather flat near equilibrium. Both its planarity at the saddle point and non-planarity at equilibrium are deciphered, and pinpointed the role played by the embedded CNH2 quasi-molecules (tiles).

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

Monolayer 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 synthesized and characterized in crystalline forms, their dynamical behavior in solution, including prenucleation cluster formation, is not well understood due to limitations both in characterization and first-principles modeling techniques. Recent advancements in machine-learned interatomic potentials are rapidly enabling the study of complex interactions such as dynamical behavior and reactivity on 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 mechanism explaining the thermal instability of neutral Au25(SR)18 clusters. Specifically, we observe multiple stages of isomerization of the Au25(SR)18 cluster, including a chiral isomer. Additionally, we simulate coalescence of two Au25(SR)18 clusters and observe series of clusters where the formation mechanisms are critically mediated by ligand exchange in the form of [Au-S]n rings.

Static versus dynamically polarizable environments within the many-body GW formalism.

Continuum- or discrete-polarizable models for the study of optoelectronic processes in embedded subsystems rely mostly on the restriction of the surrounding electronic dielectric response to its low frequency limit. Such a description hinges on the assumption that the electrons in the surrounding medium react instantaneously to any excitation in the central subsystem, thus treating the environment in the adiabatic limit. Exploiting a recently developed embedded GW formalism with an environment described at the fully abinitio level, we assess the merits of the adiabatic limit with respect to an environment where the full dynamics of the dielectric response are considered. Furthermore, we show how to properly take the static limit of the environment's susceptibility by introducing the so-called Coulomb-hole and screened-exchange contributions to the reaction field. As a first application, we consider a C60 molecule at the surface of a C60 crystal, namely, a case where the dynamics of the embedded and embedding subsystems are similar. The common adiabatic assumption, when properly treated, generates errors below 10% on the polarization energy associated with frontier energy levels and associated energy gaps. Finally, we consider a water molecule inside a metallic nanotube, the worst case for the environment's adiabatic limit. The error on the gap polarization energy remains below 10%, even though the error on the frontier orbital polarization energies can reach a few tenths of an electronvolt.

Building an ab initio solvated DNA model using Euclidean neural networks.

Accurately modeling large biomolecules such as DNA from first principles is fundamentally challenging due to the steep computational scaling of ab initio quantum chemistry methods. This limitation becomes even more prominent when modeling biomolecules in solution due to the need to include large numbers of solvent molecules. We present a machine-learned electron density model based on a Euclidean neural network framework that includes a built-in understanding of equivariance to model explicitly solvated double-stranded DNA. By training the machine learning model using molecular fragments that sample the key DNA and solvent interactions, we show that the model predicts electron densities of arbitrary systems of solvated DNA accurately, resolves polarization effects that are neglected by classical force fields, and captures the physics of the DNA-solvent interaction at the ab initio level.

Highly parallelizable path sampling with minimal rejections using asynchronous replica exchange and infinite swaps

Capturing rare yet pivotal events poses a significant challenge for molecular simulations. Path sampling provides a unique approach to tackle this issue without altering the potential energy landscape or dynamics, enabling recovery of both thermodynamic and kinetic information. However, despite its exponential acceleration compared to standard molecular dynamics, generating numerous trajectories can still require a long time. By harnessing our recent algorithmic innovations-particularly subtrajectory moves with high acceptance, coupled with asynchronous replica exchange featuring infinite swaps-we establish a highly parallelizable and rapidly converging path sampling protocol, compatible with diverse high-performance computing architectures. We demonstrate our approach on the liquid-vapor phase transition in superheated water, the unfolding of the chignolin protein, and water dissociation. The latter, performed at the ab initio level, achieves comparable statistical accuracy within days, in contrast to a previous study requiring over a year.

3D atomic structure from a single X-ray free electron laser pulse

X-ray Free Electron Lasers (XFEL) are cutting-edge pulsed x-ray sources, whose extraordinary pulse parameters promise to unlock unique applications. Several new methods have been developed at XFELs; however, no methods are known, which allow ab initio atomic level structure determination using only a single XFEL pulse. Here, we present experimental results, demonstrating the determination of the 3D atomic structure from data obtained during a single 25 fs XFEL pulse. Parallel measurement of hundreds of Bragg reflections was done by collecting Kossel line patterns of GaAs and GaP. To the best of our knowledge with these measurements, we reached the ultimate temporal limit of the x-ray structure solution possible today. These measurements open the way for obtaining crystalline structures during non-repeatable fast processes, such as structural transformations. For example, the atomic structure of matter at extremely non-ambient conditions or transient structures formed in irreversible physical, chemical, or biological processes may be captured in a single shot measurement during the transformation. It would also facilitate time resolved pump-probe structural studies making them significantly shorter than traditional serial crystallography.

Theoretical study of copper corrosion prevention by terpenoid green inhibitors of thymol, carvacrol, and thymohydroquinone

Electronic parameters have been used to predict the corrosion inhibition performance of the green inhibitors of thymol, carvacrol, and thymohydroquinone against copper. Molecular modeling was carried out at the density functional and ab initio levels and Monte Carlo molecular dynamics. Tymohydroquinone can prevent copper corrosion better than the other two monoterpenoids. Molecular dynamics studies show that corrosion inhibitors are adsorbed on the metal surface, thereby inhibiting corrosion from occurring on the metal surface.

Dependence of lactose adsorption on the exposed crystal facets of metals: a comparative study of gold, silver and copper.

In this theoretical work, we investigated the adsorption of a lactose molecule on metal-based surfaces, with a focus on the influence of the nature of the metal and of the type of exposed crystal facet on the adsorbed structures and energetics. More precisely, we considered three flat crystallographic facets of three face-centered cubic metals (gold, silver, and copper). For the global exploration of the energy landscape, we employed a multi-stage procedure where high-throughput searches, using a stochastic method that performs global optimization by iterating local searches, are followed by a refinement of the most probable adsorption conformations of the molecule at the ab initio level. We predicted the optimal conformation of lactose on each of the nine metal-surface combinations, classified the many low-energy minima into possible adsorption modes, and analyzed the structural, electronic and energetic aspects of the lactose molecule on the surface, as well as their dependence on the type of metal and exposed crystal facet. We observed structural similarities between the various minimum-energy conformations of lactose in vacuum and on the surface, a rough correlation between adsorption and interaction energies of the molecule, and a small charge transfer between molecule and surface whose direction is metal-dependent. During adsorption, an electronic reorganization occurs at the metal-molecule interface only, without affecting the vacuum-pointing atoms of the lactose molecule. For all types of surfaces, lactose exhibits the weakest adsorption on silver substrates, while for each coinage metal the adsorption is strongest on the (110) crystal facet. This study demonstrates that the control of exposed facets can allow to modulate the interaction between metals and small saccharides.

Photoisomerization pathways of trans-resveratrol.

Resveratrol is well-known for promoting health benefits due to its antioxidant, anti-aging, anti-carcinogenic, and other beneficial activities. Understanding the photophysics of resveratrol is essential for determining its applicability to pharmaceutical innovations. In the present work, we used an explore-then-assess strategy to map the internal conversion pathways of trans-resveratrol. This strategy consists of exploring the multidimensional configurational space with nonadiabatic dynamics simulations based on a semiempirical multireference method, followed by a feasibility assessment of reduced-dimensionality pathways at a high ab initio theoretical level. The exploration step revealed that internal conversion to the ground state may occur near five distinct conical intersections. The assessment step showed that the main photoisomerization pathway involves a twisted-pyramidalized S1/S0 conical intersection, yielding either trans or cis isomers. However, a secondary path was identified, where cis-trans isomerization happens in the excited state and internal conversion occurs at a cyclic conical intersection, yielding a closed-ring resveratrol derivative. This derivative, which can be formed through this direct path or an indirect photoexcitation, may be connected to the production of oxygen-reactive species previously reported and have implications in photodynamic therapy.

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 Raman spectrum of florencite‐(REE) [REEAl3(PO4)2(OH)6]: An integrated experimental and computational approach

AbstractFlorencite is a hydrous light rare‐earth elements (LREE) aluminium phosphate [REEAl3(PO4)2(OH)6], that amongst the REE‐rich minerals is quite common. The main end‐members are Ce‐, La‐ and Nd‐rich terms that were found in several genetic environments. Despite the large occurrence worldwide, to the authors' knowledge, florencite has attracted very few studies, particularly concerning the characterization of its Raman spectrum. We present a detailed study of the Raman spectrum of florencite, combining experimental measurements and theoretical calculations. Experimental Raman spectra (in the 100–1300 cm−1 spectral range) are measured on four florencite samples characterized by different chemical composition, that is, different REE abundance. The results highlight a remarkable coincidence between different Raman spectra measured on each sample, despite the significantly different chemical compositions in terms of their REE content. The same similarities were also observed in the computed spectra at the ab initio level; moreover, the calculations allowed the attributions of the different Raman signals to specific vibrational modes.