Projected Wave Function Extrapolation Scheme to Accelerate Plane-Wave Hybrid Functional-Based Born-Oppenheimer Molecular Dynamics Simulations

Born-Oppenheimer molecular dynamics (BOMD) simulations are of great interest for the dynamic properties of molecular and solid systems. However, BOMD simulations necessitate not only an extensive period of dynamical evolution but also costly self-consistent-field (SCF) electronic structure calculations, especially for hybrid functional-based BOMD (H-BOMD) simulations within plane-wave basis sets. Here, we propose an improved always stable predictor-corrector (ASPC) method for the wave function extrapolation to accelerate the plane-wave H-BOMD simulations, named projected ASPC (PASPC), yielding a wave function closer to the actual solution space and efficiently reducing the number of SCF iterations at each MD step. We investigated the convergence properties of different extrapolation schemes for molecular and solid systems. Numerical results demonstrate that plane-wave H-BOMD simulations can be significantly faster than conventional cases by combining the accelerated algorithms with the PASPC method. The energy drift is also evaluated, showing that PASPC produces energy drift with smaller oscillations and can simulate a larger time step for systems containing heavy atoms, demonstrating the accuracy of the extrapolation schemes. Furthermore, H-BOMD simulations showcase more accurate power and infrared spectra of silicon dioxide and liquid water that are comparable to those of experimental measurements.

PiNN: Equivariant Neural Network Suite for Modeling Electrochemical Systems

Electrochemical energy storage and conversion play increasingly important roles in electrification and sustainable development across the globe. A key challenge therein is to understand, control, and design electrochemical energy materials with atomistic precision. This requires inputs from molecular modeling powered by machine learning (ML) techniques. In this work, we have upgraded our pairwise interaction neural network Python package PiNN via introducing equivariant features to the PiNet2 architecture for fitting potential energy surfaces along with PiNet2-dipole for dipole and charge predictions as well as PiNet2-χ for generating atom-condensed charge response kernels. By benchmarking publicly accessible data sets of small molecules, crystalline materials, and liquid electrolytes, we found that the equivariant PiNet2 shows significant improvements over the original PiNet architecture and provides a state-of-the-art overall performance. Furthermore, leveraging on plug-ins such as PiNNAcLe for an adaptive learn-on-the-fly workflow in generating ML potentials and PiNNwall for modeling heterogeneous electrodes under external bias, we expect PiNN to serve as a versatile and high-performing ML-accelerated platform for molecular modeling of electrochemical systems.

Generalized bond polarizability model for more accurate atomistic modeling of Raman spectra

Raman spectroscopy is an important tool for studying molecules, liquids and solids. While Raman spectra can be obtained theoretically from molecular dynamics (MD) simulations, this requires the calculation of electronic polarizability along the simulation trajectory. First-principles calculations of electronic polarizability are computationally expensive, motivating the development of atomistic models for the evaluation of the changes in the electronic polarizability with the changes in the atomic coordinates of the system. The bond polarizability model (BPM) is one of the oldest and simplest such atomistic models but cannot reproduce the effects of angular vibrations, leading to inaccurate modeling of Raman spectra. Here, we demonstrate that the generalization of BPM through the inclusion of terms for atom pairs that are traditionally considered to be not involved in bonding dramatically improves the accuracy of polarizability modeling and Raman spectra calculations. The generalized BPM (GBPM) reproduces the ab initio polarizability and Raman spectra for a range of tested molecules (SO2, H2S, H2O, NH3, CH4, CH3OH, and CH3CH2OH) with high accuracy and also shows significantly improved agreement with ab initio results for the more complex ferroelectric BaTiO3 systems. For liquid water, the anisotropic Raman spectrum derived from atomistic MD simulations using the GBPM evaluation of polarizability shows significantly improved agreement with the experimental spectrum compared to the spectrum derived using the BPM. Thus, the GBPM can be used for the modeling of Raman spectra using large-scale molecular dynamics and provides a good basis for the further development of atomistic polarizability models.

Probing London Dispersion in Proton-Bound Onium Ions: Are Alkyl-Alkyl Steric Interactions Reliably Modeled?

We report spectroscopic and spectrometric experiments that probe the London dispersion interaction between tert-butyl substituents in three series of covalently linked, protonated bis-pyridines in the gas phase. Molecular ions in the three test series, along with several reference molecules for control, were electrosprayed from solution into the gas phase and then probed by infrared multiphoton dissociation spectroscopy and trapped ion mobility spectrometry. The observed N-H stretching frequencies provided an experimental readout diagnostic of the ground-state geometry of each ion, which could be furthermore compared to a second, independent structural readout via the collision cross section. In each of the three series, the strength of a London dispersion interaction could be modulated systematically by a progressive increase in the size of substituents from H to Me to tert-Bu. Parallel to the experimental study, extensive dispersion-corrected density functional theory (DFT-D3BJ) calculations were performed with a range of exchange correlation functionals. A full analysis of the conformational space for the flexible members of the series, and an analysis of the vibrational spectra in the context of a general double-well potential, finds that DFT-D3BJ appears to significantly overbind alkyl-alkyl interactions, specifically interactions between tert-Bu groups, even failing to predict the minimum energy structures reliably in the case of molecules in which London dispersion competes with other noncovalent interactions such as hydrogen bonding.

Study on Microscopic Properties of Molten NaF-AlF3-CaF2/LiF/KF Using First-Principles Molecular Dynamics

This study employs first-principles molecular dynamics (FPMD) simulations combined with the Voronoi tessellation method to explore the microstructure, transport properties, electronic properties, and Raman spectra of the NaF-AlF3-CaF2/LiF/KF systems with varying cryolite ratios, additive types, and concentrations. The results indicate that Na+, Ca2+, Li+, and K+ exist in a free state in the molten salts, while Al3+ forms complex ion groups in the form of [AlFx]3-x with F-, and free F- also exists in the molten salts. In the NaF-AlF3-CaF2 system, the average Al-F distance is slightly shorter than that in the other two systems, while the Al-F coordination number is higher in NaF-AlF3-LiF. Cryolite ratios and additive concentrations have little effect on the average Al-F distance. The diffusion abilities of different ions follow the order: Li+ > Na+ > F- > Al3+, with the diffusion ability of K+ being close to that of Li+. In Al-F chemical bonds, both ionic and covalent bonds coexist and double bonds can be observed in certain transient structures. The Al-F complex ion groups mainly consist of [AlF4]-, [AlF5]2-, and [AlF6]3-, with the concentration of [AlF6]3- being higher in the NaF-AlF3-LiF system compared to the other two systems. This study establishes the relationship between the microscopic properties and the composition of aluminum electrolytes, demonstrating the suitability of FPMD combined with Voronoi tessellation for probing the microstructural properties of aluminum electrolytes.

The Analysis of Vibrational Spectra: Past, Present and Future

Vibrational spectroscopy can be said to have started with the seminal work of Coblentz in the 1900s, who recorded the first recognisable infrared spectra. Today, vibrational spectroscopy is ubiquitous and there are many ways to measure a vibrational spectrum. But this is usually only the first step, almost always there is a need to assign the resulting spectra: "what property of the system results in a feature at this energy"? How this question has been answered has changed over the last century, as our understanding of the fundamental physics of matter has evolved. In this Perspective, I will present my view of how the analysis of vibrational spectra has evolved over time. The article is divided into three sections: past, present and future. The "past" section consists of a very brief history of vibrational spectroscopy. The "present" is centered around ab initio studies, particularly with density functional theory (DFT) and I will describe how this has become almost routine. For the "future", I will extrapolate current trends and also speculate as to what might come next.

Assignment of IR spectra of ethanol at Brønsted sites of H-ZSM-5 to monomer adsorption using a Fermi resonance model

Understanding how alcohol molecules interact with the Brønsted acid sites (BAS) of zeolites is a prerequisite to the design of zeolite catalysts and catalytic processes. Here, we report IR spectra for the adsorption of ethanol on a highly crystalline sample of H-ZSM-5 zeolites exposed to ethanol gas at increasing pressure. We use density functional theory in combination with a FERMI resonance model to assign the measured spectra to a single adsorbed ethanol molecule per BAS. Specifically, we assign the bands at 2450 cm-1 and 1670 cm-1 to a FERMI resonance between the fundamental (Z)O-H stretching band of a single-ethanol-loaded BAS and the first overtone of the (Z)O-H out-of-plane bending. We conclude that adsorbed dimers do not contribute in a noticeable way up to a concentration of almost one ethanol molecule per BAS site. We further show that hybrid functionals (B3LYP) are required to get a close match between the predicted and experimental spectra, whereas commonly used generalized gradient type functionals such as PBE incorrectly describe the potential energy surface. They overestimate the redshift of the OH stretching band on hydrogen bond formation which results in an erroneous assignment of the IR bands.

Deep Learning for Generating Phase-Conditioned Infrared Spectra

Infrared (IR) spectroscopy is an efficient method for identifying unknown chemical compounds. To accelerate IR spectrum analysis, various calculation and machine learning methods for simulating IR spectra of molecules have been studied in chemical science. However, existing calculation and machine learning methods assumed a rigid constraint that all molecules are in the gas phase, i.e., they overlooked the phase dependency of the IR spectra. In this paper, we propose an efficient phase-aware machine learning method to generate phase-conditioned IR spectra from 2D molecular structures. To this end, we devised a phase-aware graph neural network and combined it with a transformer decoder. To the best of our knowledge, the proposed method is the first IR spectrum generator that can generate the phase-conditioned IR spectra of real-world complex molecules. The proposed method outperformed state-of-the-art methods in the tasks of generating IR spectra on a benchmark dataset containing experimentally measured 11,546 IR spectra of 10,288 unique molecules. All implementations of the proposed method are publicly available at https://github.com/ngs00/PASGeN.

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.

Leveraging infrared spectroscopy for automated structure elucidation

The application of machine learning models in chemistry has made remarkable strides in recent years. While analytical chemistry has received considerable interest from machine learning practitioners, its adoption into everyday use remains limited. Among the available analytical methods, Infrared (IR) spectroscopy stands out in terms of affordability, simplicity, and accessibility. However, its use has been limited to the identification of a selected few functional groups, as most peaks lie beyond human interpretation. We present a transformer model that enables chemists to leverage the complete information contained within an IR spectrum to directly predict the molecular structure. To cover a large chemical space, we pretrain the model using 634,585 simulated IR spectra and fine-tune it on 3,453 experimental spectra. Our approach achieves a top-1 accuracy of 44.4% and top-10 accuracy of 69.8% on compounds containing 6 to 13 heavy atoms. When solely predicting scaffolds, the model accurately predicts the top-1 scaffold in 84.5% and among the top-10 in 93.0% of cases.

Assessing the Partial Hessian Approximation in QM/MM-Based Vibrational Analysis

The partial Hessian approximation is often used in vibrational analysis of quantum mechanics/molecular mechanics (QM/MM) systems because calculating the full Hessian matrix is computationally impractical. This approach aligns with the core concept of QM/MM, which focuses on the QM subsystem. Thus, using the partial Hessian approximation implies that the main interest is in the local vibrational modes of the QM subsystem. Here, we investigate the accuracy and applicability of the partial Hessian vibrational analysis (PHVA) approach as it is typically used within QM/MM, i.e., only the Hessian belonging to the QM subsystem is computed. We focus on solute-solvent systems with small, rigid solutes. To separate two of the major sources of errors, we perform two separate analyses. First, we study the effects of the partial Hessian approximation on local normal modes, harmonic frequencies, and harmonic IR and Raman intensities by comparing them to those obtained using full Hessians, where both partial and full Hessians are calculated at the QM level. Then, we quantify the errors introduced by QM/MM used with the PHVA by comparing normal modes, frequencies, and intensities obtained using partial Hessians calculated using a QM/MM-type embedding approach to those obtained using partial Hessians calculated at the QM level. Another aspect of the PHVA is the appearance of normal modes resembling the translation and rotation of the QM subsystem. These pseudotranslational and pseudorotational modes should be removed as they are collective vibrations of the atoms in the QM subsystem relative to a frozen MM subsystem and, thus, not well-described. We show that projecting out translation and rotation, usually done for systems in isolation, can adversely affect other normal modes. Instead, the pseudotranslational and pseudorotational modes can be identified and removed.

Robust Computation and Analysis of Vibrational Spectra of Layered Framework Materials Including Host-Guest Interactions

Layered framework materials, a rapidly advancing class of porous materials, are composed of molecular components stitched together via covalent bonds and are usually synthesized through wet-chemical methods. Computational infrared (IR) and Raman spectra are among the most important characterization tools for this material class. Besides the a priori known spectra of the molecular building blocks and the solvent, they allow for in situ monitoring of the framework formation during synthesis. Therefore, they need to capture the additional peaks from host-guest interactions and the bands from emerging bonds between the molecular building blocks, verifying the successful synthesis of the desired material. In this work, we propose a robust computational framework based on ab initio molecular dynamics (AIMD), where we compute IR and Raman spectra from the time-correlation functions of dipole moments and polarizability tensors, respectively. As a case study, we apply our methodology to a covalent organic framework (COF) material, COF-1, and present its AIMD-computed IR and Raman spectra with and without 1,4-dioxane solvent molecules in its pores. To determine robust settings, we meticulously validate our model and explore how stacking disorder and different methods for computing dipole moments and polarizabilities affect IR and Raman intensities. Using our robust computational protocol, we achieve excellent agreement with experimental data. Furthermore, we illustrate how the computed spectra can be dissected into individual contributions from the solvent molecules, the molecular building blocks of COF-1, and the bonds connecting them.

Isomer-Specific, Cryogenic Ion Vibrational Spectroscopy Investigation of D2- and N2-Tagged, Protonated Formic Acid Complexes Using Two-Color, IR-IR Photobleaching

Here we analyze cryogenic ion vibrational spectra of tagged protonated formic acid (PFA) with electronic structure and anharmonic vibrational calculations to establish the isomers generated by electrospray ionization (ESI) followed by buffer gas cooling to ∼25 K. Two isomers are identified (the trans form (E,Z) and the cis form (E,E)) and generated in comparable abundance despite the fact that the calculated E,E structure lies 6.40 kJ mol-1 above the E,Z form. A large (∼60 kJ mol-1) barrier separates them such that the E,E form can be kinetically trapped upon cooling in the ion trap. The anticooperativity between the H-bonds of the OH groups is explored by measuring the shift in the D2-bound OH fundamental when a second D2 is attached. Both isomers are observed in the N2-tagged counterparts, displaying the expected red-shifted OH bands. These results indicate that ESI generates both isomers and both must be considered when analyzing cluster spectra based on the PFA core ion.

Second-Order Mass-Weighting Scheme for Atom-Centered Density Matrix Propagation Molecular Dynamics

The atom-centered density matrix propagation (ADMP) method is an extended Lagrangian approach to ab initio molecular dynamics, which includes the density matrix in an orthonormalized atom-centered Gaussian basis as additional, fictitious, electronic degrees of freedom, classically propagated along with the nuclear ones. A high adiabaticity between the nuclear and electronic subsystems is mandatory in order to keep the trajectory close to the Born-Oppenheimer (BO) surface. In this regard, the fictitious electronic mass μ, being a symmetric, nondiagonal matrix in its most general form, represents a free parameter, exploitable to optimize the propagation of the electronic density. Although mass-weighting schemes in ADMP exist, a systematic procedure to define an optimal value of the fictitious masses is not available yet. In this work, in order to rationally evaluate the electronic mass, fictitious electronic normal modes are defined through the diagonalization of the Hessian of the electronic density matrix. If the same frequency is imposed on all such modes (compatible with the chosen integration time step), then the corresponding μ matrix can be calculated and then employed for the following propagation. Analysis of several ADMP test simulations reveals that such Hessian-based mass-weighting approach is able to ensure, together with a 0.1/0.2 fs time steps, a high separation between the (real) nuclear and the (fictitious) electronic frequencies, which determines a high adiabaticity. This high, unprecedented, accuracy in the propagation leads, in turn, to low errors in the estimated nuclear vibrational frequencies, making the ADMP method totally comparable to a fully converged BO molecular dynamics simulation but more computationally efficient. This work, therefore, contributes to a further development of the ADMP ab initio molecular dynamics method, aimed at improving its accuracy through a more rational evaluation of the fictitious electronic mass parameter.

Structure and Dynamics of ATP and the ATP-Zn2+ Complex in Solution

Despite the crucial role of ATP in life and artificial life-like applications, fundamental aspects relevant to its function, such as its conformational properties and its interaction with water and ions, remain unclear. Here, by integrating linear and two-dimensional infrared spectroscopy with ab initio molecular dynamics, we provide a detailed characterization of the vibrational spectra of the phosphate groups in ATP and in its complex with Zn2+ in water. Our study highlights the role of conformational disorder and solvation dynamics, beyond the harmonic normal-mode analysis, and reveals a complex scenario in which electronic and environmental effects tune the coupling between phosphate vibrations. We identify βγ-bidentate and αβγ-tridentate modes as the preferential coordination modes of Zn2+, as was proposed in the literature for Mg2+, although this conclusion is reached by a different spectral interpretation.

Rapid Characterization of Point Defects in Solid-State Ion Conductors Using Raman Spectroscopy, Machine-Learning Force Fields, and Atomic Raman Tensors

The successful design of solid-state photo- and electrochemical devices depends on the careful engineering of point defects in solid-state ion conductors. Characterization of point defects is critical to these efforts, but the best-developed techniques are difficult and time-consuming. Raman spectroscopy─with its exceptional speed, flexibility, and accessibility─is a promising alternative. Raman signatures arise from point defects due to local symmetry breaking and structural distortions. Unfortunately, the assignment of these signatures is often hampered by a shortage of reference compounds and corresponding reference spectra. This issue can be circumvented by calculation of defect-induced Raman signatures from first principles, but this is computationally demanding. Here, we introduce an efficient computational procedure for the prediction of point defect Raman signatures in solid-state ion conductors. Our method leverages machine-learning force fields and "atomic Raman tensors", i.e., polarizability fluctuations due to motions of individual atoms. We find that our procedure reduces computational cost by up to 80% compared to existing first-principles frozen-phonon approaches. These efficiency gains enable synergistic computational-experimental investigations, in our case allowing us to precisely interpret the Raman spectra of Sr(Ti0.94Ni0.06)O3-δ, a model oxygen ion conductor. By predicting Raman signatures of specific point defects, we determine the nature of dominant defects and unravel impacts of temperature and quenching on in situ and ex situ Raman spectra. Specifically, our findings reveal the temperature-dependent distribution and association behavior of oxygen vacancies and nickel substitutional defects. Overall, our approach enables rapid Raman-based characterization of point defects to support defect engineering in novel solid-state ion conductors.

Machine-learning to predict anharmonic frequencies: a study of models and transferability

With more and more accurate electronic structure methods at hand, the inclusion of anharmonic effects in the post-processing of such data towards thermochemical properties is the next step. In this context, the description of anharmonicity has been an important topic of physical chemistry and chemical physics for a long time. In this study, anharmonic frequencies of various hydrogen-halides and halogenated hydrocarbon molecular clusters are calculated using harmonic as well as explicitly anharmonic methods, i.e., normal mode analysis and vibrational self-consistent field. Simple harmonic model based descriptors were used to predict anharmonic frequencies via multilinear regression and gradient boosting regression. Gradient boosting regression is capable of predicting reliable anharmonic data and even the simple multilinear regression model yields reasonable predictions that can account for mode-to-mode couplings. Moreover, the transferability to unseen chemical systems is assessed and it is confirmed that the machine-learned models can be applied to larger, unseen molecules.

After 50 Years of Vibrational Circular Dichroism Spectroscopy: Challenges and Opportunities of Increasingly Accurate and Complex Experiments and Computations

VCD research continues to thrive, driven by ongoing experimental and theoretical advances. Modern studies deal with increasingly complex samples featuring weak intermolecular interactions and shallow potential energy surfaces. Likewise, the combination of VCD measurements with, for instance, cryo-spectroscopic techniques has significantly increased their sensitivity. The extent to which such modern measurements enhance the informative value of VCD depends significantly on the quality of the theoretical models, which must adequately account for anharmonicity, solvation and molecular dynamics. We herein discuss how experimental advancements engage in a stimulating interplay with recent theoretical developments, pursuing either the static or the dynamic computational route. Both paths have their own strengths and limitations, each addressing fundamentally different problems. We give an outlook on future challenges of VCD research, including the possibility to combine static and dynamic approaches to obtain a full picture of the sample.

Influence of terahertz waves on the binding of choline to choline acetyltransferase: insights from molecular dynamics simulations

Acetylcholine (Ach) is a common neurotransmitter in the central nervous system (CNS) and peripheral nervous system (PNS). It is one of the neurotransmitters in the autonomic nervous system and the main neurotransmitter in all autonomic ganglia. Experiments have confirmed that electromagnetic waves can affect the synthesis of animal neurotransmitters, but the microscopic effects of electromagnetic waves in the terahertz (THz) frequency band are still unclear. Based on density functional theory (DFT) and molecular dynamics (MD) simulation methods, this paper studies the effect of THz electromagnetic waves on the binding of choline to choline acetyltransferase (ChAT). By emitting THz waves that resonate with the characteristic vibration mode of choline near the active site, it was found that THz waves with a frequency of 45.3 THz affected the binding of choline to ChAT, especially the binding of the active site histidine His324 to choline. The main evidence is that under the action of THz waves, the binding free energy of choline to histidine His324 and ChAT at the active site was significantly reduced compared to noE, which may have a potential impact on the enzymatic synthesis of Ach. It is expected to achieve the purpose of regulating the synthesis of the neurotransmitter Ach under the action of THz waves and treat certain nervous system diseases.

Boosting the Modeling of Infrared and Raman Spectra of Bulk Phase Chromophores with Machine Learning

In the field of vibrational spectroscopy simulation, hybrid approximations to Kohn-Sham density-functional theory (KS-DFT) are often considered computationally prohibitive due to the significant effort required to evaluate the exchange-correlation potential in planewave codes. In this Letter, we show that by leveraging the porting of KS-DFT on GPU and incorporating machine-learning techniques, simulating IR and Raman spectra of real-life chromophores in bulk aqueous solution becomes a routine application at this level of theory.

Accuracy and limitations of the bond polarizability model in modeling of Raman scattering from molecular dynamics simulations

Calculation of Raman scattering from molecular dynamics (MD) simulations requires accurate modeling of the evolution of the electronic polarizability of the system along its MD trajectory. For large systems, this necessitates the use of atomistic models to represent the dependence of electronic polarizability on atomic coordinates. The bond polarizability model (BPM) is the simplest such model and has been used for modeling the Raman spectra of molecular systems but has not been applied to solid-state systems. Here, we systematically investigate the accuracy and limitations of the BPM parameterized from the density functional theory results for a series of simple molecules, such as CO2, SO2, H2S, H2O, NH3, and CH4; the more complex CH2O, CH3OH, CH3CH2OH, and thiophene molecules; and the BaTiO3 and CsPbBr3 perovskite solids. We find that BPM can reliably reproduce the overall features of the Raman spectra, such as shifts of peak positions. However, with the exception of highly symmetric systems, the assumption of non-interacting bonds limits the quantitative accuracy of the BPM; this assumption also leads to qualitatively inaccurate polarizability evolution and Raman spectra for systems where large deviations from the ground state structure are present.

Non-invasive detection of hazardous materials with a thermal-to-epithermal neutron station: a feasibility study towards practical application

The growing scale of the devastation that even a single terrorist attack can cause requires more effective methods for the detection of hazardous materials. In particular, there are no solutions for effectively monitoring threats at sea, both for the off-shore infrastructure and ports. Currently, state-of-the-art detection methods determine the density distribution and the shapes of tested subjects but only allow for a limited degree of substance identification. This work aims to present a feasibility study of the possible usage of several methods available on the thermal-to-epithermal neutron station, VESUVIO, at the ISIS neutron and muon spallation source, UK, for the detection of hazardous materials. To this end, we present the results of a series of experiments performed concurrently employing neutron transmission and Compton scattering using melamine, a commonly used explosive surrogate, in order to determine its signal characteristics and limits of detection and quantitation. The experiments are supported by first-principles modelling, providing detailed scrutiny of the material structure and the nuclear dynamics behind the neutron scattering observables.

Unravelling Mn4Ca cluster vibrations in the S1, S2 and S3 states of the Kok-Joliot cycle of photosystem II

Vibrational spectroscopy serves as a powerful tool for characterizing intermediate states within the Kok-Joliot cycle. In this study, we employ a QM/MM molecular dynamics framework to calculate the room temperature infrared absorption spectra of the S1, S2, and S3 states via the Fourier transform of the dipole time auto-correlation function. To better analyze the computational data and assign spectral peaks, we introduce an approach based on dipole-dipole correlation function of cluster moieties of the reaction center. Our analysis reveals variation in the infrared signature of the Mn4Ca cluster along the Kok-Joliot cycle, attributed to its increasing symmetry and rigidity resulting from the rising oxidation state of the Mn ions. Furthermore, we successfully assign the debated contributions in the frequency range around 600 cm-1. This computational methodology provides valuable insights for deciphering experimental infrared spectra and understanding the water oxidation process in both biological and artificial systems.

Vibrational Spectra Simulations in Amino Acid-Based Imidazolium Ionic Liquids

We present maximally localized Wannier functions and Voronoi tessellation to obtain dipole moment distributions for vibrational spectra in several important ionic liquids calculated by using ab initio molecular dynamics simulations. IR and Raman spectra of various imidazolium-based ionic liquids (ILs) paired with six amino acid anions are shown herein. For IR spectra, two approaches (Wannier and Voronoi) are in agreement with respect to the relative intensities and the overall shapes for the main peaks. Under Raman spectra, the polarizability of the covalent bonds is shown to affect the strength of the Raman scattering signal. The advantage of the Voronoi tessellation method, being that it does not have strong spikes in its time development, is demonstrated by the comparison of two theoretical methods (Wannier and Voronoi) with experimental data. We analyze the errors between theoretical and experimental spectroscopic data, with the Voronoi method shown to accurately reproduce experimental values. In addition, theoretical spectroscopy shows the ability to accurately separate components of a mixture. The combination of theoretical and experimental methods is utilized to understand the spectroscopic properties of amino acid-based imidazolium ILs.

Is Ortho-Terphenyl a Rigid Glass Former?

Ortho-terphenyl (OTP) has long been used as a model system to study the glass transition due to its apparent simplicity and a widespread assumption that it is a rigid molecule. Here, we employ terahertz time-domain spectroscopy and low-frequency Raman spectroscopy to investigate the rigidity of OTP by direct observation of the low-frequency vibrational dynamics. These terahertz phonons involve complex large-amplitude atomic motions where intramolecular and intermolecular displacements are often mixed. Comparison of experimental results with density functional theory and ab initio molecular dynamics simulations shows that the assumption of rigidity neglects important implications for the glass transition and must be revisited. These results highlight the significance of terahertz modes on elasticity, which will be even more critical in more complex systems such as biomolecules.

Molecular dynamics simulations reliably identify vibrational modes in far-IR spectra of phospholipids

The properties of self-assembled phospholipid membranes are of essential importance in biochemistry and physical chemistry, providing a platform for many cellular life functions. Far-infrared (far-IR) vibrational spectroscopy, on the other hand, is a highly information-rich method to characterize intermolecular interactions and collective behaviour of lipids that can help explain, e.g., chain packing, thermodynamic phase behaviour, and sequestration. However, reliable interpretation of the far-IR spectra is still lacking. Here we present a molecular dynamics (MD) based approach to simulate vibrational modes of individual lipids and in an ensemble. The results are a good match to synchrotron far-IR measurements and enable identification of the molecular motions corresponding to each vibrational mode, thus allowing the correct interpretation of membrane spectra with high accuracy and resolving the longstanding ambiguities in the literature in this regard. Our results demonstrate the feasibility of using MD simulations for interpreting far-IR spectra broadly, opening new avenues for practical use of this powerful method.

Weakly Hydrated Solute of Mixed Hydrophobic-Hydrophilic Nature

Infrared (IR) spectroscopy is a commonly used and invaluable tool in studies of solvation phenomena in aqueous solutions. Concurrently, density functional theory calculations and ab initio molecular dynamics simulations deliver the solvation shell picture at the molecular detail level. The mentioned techniques allowed us to gain insights into the structure and energy of the hydrogen bonding network of water molecules around methylsulfonylmethane (MSM). In the hydration sphere of MSM, there are two types of populations of water molecules: a significant share of water molecules weakly bonded to the sulfone group and a smaller share of water molecules strongly bonded to each other around the methyl groups of MSM. The very weak hydrogen bond of water molecules with the hydrophilic group causes the extended network of water hydrogen bonds to be not "anchored" on the sulfone group, and consequently, the MSM hydration shell is labile.

Exploring the Structural Stability of 1T-PdO2 and the Interface Properties of the 1T-PdO2/Graphene Heterojunction

Motivated by a recent study on the air stability of PdSe2, which also reports the metastability of the PdO2 monolayer [Hoffman A. N.. npj 2D Mater. Appl.2019, 3( (1), ), 50.], in this work, we use density functional theory (DFT) to further explore the thermal, dynamic, and mechanical stability of monolayer PdO2 and study its structural and electronic properties. We further studied its vertical heterojunction composed of 1T-PdO2 and graphene monolayers. We show that both the monolayer and the heterojunction are energetically and dynamically stable with no negative frequencies in the phonon spectrum and belong to the vdW-type. 1T-PdO2 is an indirect-band-gap semiconductor with band-gap values of 0.5 eV (GGA) and 1.54 eV (HSE06). The interface properties of the heterojunction show that the n-type Schottky barrier height (SBH) becomes negative at the vertical interface in the PdO2/graphene contact, forming an Ohmic contact and mainly suggesting the potential of graphene for efficient electrical contact with the PdO2 monolayer. However, at the same time, a negative band bending occurs at the lateral interface based on the current-in-plane model. Moreover, the optical absorption of the PdO2/graphene heterojunction under visible-light irradiation is significantly enhanced compared to the situation in their free-standing monolayers.

Raman Spectra of Amino Acids and Peptides from Machine Learning Polarizabilities

Raman spectroscopy is an important tool in the study of vibrational properties and composition of molecules, peptides, and even proteins. Raman spectra can be simulated based on the change of the electronic polarizability with vibrations, which can nowadays be efficiently obtained via machine learning models trained on first-principles data. However, the transferability of the models trained on small molecules to larger structures is unclear, and direct training on large structures is prohibitively expensive. In this work, we first train two machine learning models to predict the polarizabilities of all 20 amino acids. Both models are carefully benchmarked and compared to density functional theory (DFT) calculations, with the neural network method being found to offer better transferability. By combination of machine learning models with classical force field molecular dynamics, Raman spectra of all amino acids are also obtained and investigated, showing good agreement with experiments. The models are further extended to small peptides. We find that adding structures containing peptide bonds to the training set greatly improves predictions, even for peptides not included in training sets.

Accelerating Molecular Vibrational Spectra Simulations with a Physically Informed Deep Learning Model

In recent years, machine learning (ML) surrogate models have emerged as an indispensable tool to accelerate simulations of physical and chemical processes. However, there is still a lack of ML models that can accurately predict molecular vibrational spectra. Here, we present a highly efficient multitask ML surrogate model termed Vibrational Spectra Neural Network (VSpecNN), to accurately calculate infrared (IR) and Raman spectra based on dipole moments and polarizabilities obtained on-the-fly via ML-enhanced molecular dynamics simulations. The methodology is applied to pyrazine, a prototypical polyatomic chromophore. The VSpecNN-predicted energies are well within the chemical accuracy (1 kcal/mol), and the errors for VSpecNN-predicted forces are only half of those obtained from a popular high-performance ML model. Compared to the ab initio reference, the VSpecNN-predicted frequencies of IR and Raman spectra differ only by less than 5.87 cm-1, and the intensities of IR spectra and the depolarization ratios of Raman spectra are well reproduced. The VSpecNN model developed in this work highlights the importance of constructing highly accurate neural network potentials for predicting molecular vibrational spectra.

ZIF-8 Vibrational Spectra: Peak Assignments and Defect Signals

Zeolitic imidazolate framework (ZIF-8) is a promising material for gas separation applications. It also serves as a prototype for numerous ZIFs, including amorphous ones, with a broader range of possible applications, including sensors, catalysis, and lithography. It consists of zinc coordinated with 2-methylimidazolate (2mIm) and has been synthesized with methods ranging from liquid-phase to solvent-free synthesis, which aim to control its crystal size and shape, film thickness and microstructure, and incorporation into nanocomposites. Depending on the synthesis method and postsynthesis treatments, ZIF-8 materials may deviate from the nominal defect-free ZIF-8 crystal structure due to defects like missing 2mIm, missing zinc, and physically adsorbed 2mIm trapped in the ZIF-8 pores, which may alter its performance and stability. Infrared (IR) spectroscopy has been used to assess the presence of defects in ZIF-8 and related materials. However, conflicting interpretations by various authors persist in the literature. Here, we systematically investigate ZIF-8 vibrational spectra by combining experimental IR spectroscopy and first-principles molecular dynamics simulations, focusing on assigning peaks and elucidating the spectroscopic signals of putative defects present in the ZIF-8 material. We attempt to resolve conflicting assignments from the literature and to provide a comprehensive understanding of the vibrational spectra of ZIF-8 and its defect-induced variations, aiming toward more precise quality control and design of ZIF-8-based materials for emerging applications.

Hydration of N-Hydroxyurea from Ab Initio Molecular Dynamics Simulations

N-Hydroxyurea (HU) is an important chemotherapeutic agent used as a first-line treatment in conditions such as sickle cell disease and β-thalassemia, among others. To date, its properties as a hydrated molecule in the blood plasma or cytoplasm are dramatically understudied, although they may be crucial to the binding of HU to the radical catalytic site of ribonucleotide reductase, its molecular target. The purpose of this work is the comprehensive exploration of HU hydration. The topic is studied using ab initio molecular dynamic (AIMD) simulations that apply a first principles representation of the electron density of the system. This allows for the calculation of infrared spectra, which may be decomposed spatially to better capture the spectral signatures of solute-solvent interactions. The studied molecule is found to be strongly hydrated and tightly bound to the first shell water molecules. The analysis of the distance-dependent spectra of HU shows that the E and Z conformers spectrally affect, on average, 3.4 and 2.5 of the closest H2O molecules, respectively, in spheres of radii of 3.7 Å and 3.5 Å, respectively. The distance-dependent spectra corresponding to these cutoff radii show increased absorbance in the red-shifted part of the water OH stretching vibration band, indicating local enhancement of the solvent's hydrogen bond network. The radially resolved IR spectra also demonstrate that HU effortlessly incorporates into the hydrogen bond network of water and has an enhancing effect on this network. Metadynamics simulations based on AIMD methodology provide a picture of the conformational equilibria of HU in solution. Contrary to previous investigations of an isolated HU molecule in the gas phase, the Z conformer of HU is found here to be more stable by 17.4 kJ·mol-1 than the E conformer, pointing at the crucial role that hydration plays in determining the conformational stability of solutes. The potential energy surface for the OH group rotation in HU indicates that there is no intramolecular hydrogen bond in Z-HU in water, in stark contrast to the isolated solute in the gas phase. Instead, the preferred orientation of the hydroxyl group is perpendicular to the molecular plane of the solute. In view of the known chaotropic effect of urea and its N-alkyl-substituted derivatives, N-hydroxyurea emerges as a unique urea derivative that exhibits a kosmotropic ordering of nearby water. This property may be of crucial importance for its binding to the catalytic site of ribonucleotide reductase with a concomitant displacement of a water molecule.

Deep Learning-Assisted Spectrum-Structure Correlation: State-of-the-Art and Perspectives

Spectrum-structure correlation is playing an increasingly crucial role in spectral analysis and has undergone significant development in recent decades. With the advancement of spectrometers, the high-throughput detection triggers the explosive growth of spectral data, and the research extension from small molecules to biomolecules accompanies massive chemical space. Facing the evolving landscape of spectrum-structure correlation, conventional chemometrics becomes ill-equipped, and deep learning assisted chemometrics rapidly emerges as a flourishing approach with superior ability of extracting latent features and making precise predictions. In this review, the molecular and spectral representations and fundamental knowledge of deep learning are first introduced. We then summarize the development of how deep learning assist to establish the correlation between spectrum and molecular structure in the recent 5 years, by empowering spectral prediction (i.e., forward structure-spectrum correlation) and further enabling library matching and de novo molecular generation (i.e., inverse spectrum-structure correlation). Finally, we highlight the most important open issues persisted with corresponding potential solutions. With the fast development of deep learning, it is expected to see ultimate solution of establishing spectrum-structure correlation soon, which would trigger substantial development of various disciplines.

Disentangling the effects of structure and lone-pair electrons in the lattice dynamics of halide perovskites

Halide perovskites show great optoelectronic performance, but their favorable properties are paired with unusually strong anharmonicity. It was proposed that this combination derives from the ns2 electron configuration of octahedral cations and associated pseudo-Jahn-Teller effect. We show that such cations are not a prerequisite for the strong anharmonicity and low-energy lattice dynamics encountered in these materials. We combine X-ray diffraction, infrared and Raman spectroscopies, and molecular dynamics to contrast the lattice dynamics of CsSrBr3 with those of CsPbBr3, two compounds that are structurally similar but with the former lacking ns2 cations with the propensity to form electron lone pairs. We exploit low-frequency diffusive Raman scattering, nominally symmetry-forbidden in the cubic phase, as a fingerprint of anharmonicity and reveal that low-frequency tilting occurs irrespective of octahedral cation electron configuration. This highlights the role of structure in perovskite lattice dynamics, providing design rules for the emerging class of soft perovskite semiconductors.

Accelerated Bayesian Inference for Molecular Simulations using Local Gaussian Process Surrogate Models

While Bayesian inference is the gold standard for uncertainty quantification and propagation, its use within physical chemistry encounters formidable computational barriers. These bottlenecks are magnified for modeling data with many independent variables, such as X-ray/neutron scattering patterns and electromagnetic spectra. To address this challenge, we employ local Gaussian process (LGP) surrogate models to accelerate Bayesian optimization over these complex thermophysical properties. The time-complexity of the LGPs scales linearly in the number of independent variables, in stark contrast to the computationally expensive cubic scaling of conventional Gaussian processes. To illustrate the method, we trained a LGP surrogate model on the radial distribution function of liquid neon and observed a 1,760,000-fold speed-up compared to molecular dynamics simulation, beating a conventional GP by three orders-of-magnitude. We conclude that LGPs are robust and efficient surrogate models poised to expand the application of Bayesian inference in molecular simulations to a broad spectrum of experimental data.

Tensorial Properties via the Neuroevolution Potential Framework: Fast Simulation of Infrared and Raman Spectra

Infrared and Raman spectroscopy are widely used for the characterization of gases, liquids, and solids, as the spectra contain a wealth of information concerning, in particular, the dynamics of these systems. Atomic scale simulations can be used to predict such spectra but are often severely limited due to high computational cost or the need for strong approximations that limit the application range and reliability. Here, we introduce a machine learning (ML) accelerated approach that addresses these shortcomings and provides a significant performance boost in terms of data and computational efficiency compared with earlier ML schemes. To this end, we generalize the neuroevolution potential approach to enable the prediction of rank one and two tensors to obtain the tensorial neuroevolution potential (TNEP) scheme. We apply the resulting framework to construct models for the dipole moment, polarizability, and susceptibility of molecules, liquids, and solids and show that our approach compares favorably with several ML models from the literature with respect to accuracy and computational efficiency. Finally, we demonstrate the application of the TNEP approach to the prediction of infrared and Raman spectra of liquid water, a molecule (PTAF-), and a prototypical perovskite with strong anharmonicity (BaZrO3). The TNEP approach is implemented in the free and open source software package gpumd, which makes this methodology readily available to the scientific community.

High-pressure structures of solid hydrogen: Insights from ab initio molecular dynamics simulations

Understanding the structural behavior of solid hydrogen under high pressures is crucial for uncovering its unique properties and potential applications. In this study, starting from the phase I of solid hydrogen-free-rotator hcp structure, we conduct extensive ab initio molecular dynamics calculations to simulate the cooling, heating, and equilibrium processes within a pressure range of 80-260 GPa. Without relying on any structure previously predicted, we identify the high-pressure phase structures of solid hydrogen as P21/c for phase II, P6522 for phase III, and BG1BG2BG3 six-layer structure for phase IV, which are different from those proposed previously using the structure-search method. The reasonability of these structures are validated by Raman spectra and x-ray diffraction patterns by comparison with the experimental results. Our results actually show pronounced changes in the c/a ratio between phases I, III, and IV, which hold no brief for the experimental interpretation of an isostructural hcp transformations for phases I-III-IV.

The Coiling Effect in Ether Ionic Liquids: Exploiting Acetate as a Probe for Transport Properties and Microenvironment Analysis

The inherently high viscosity of ionic liquids (ILs) can limit their potential applications. One approach to address this drawback is to modify the cation side chain with ether groups. Herein, we assessed the structure-property relationship by focusing on acetate (OAc), a strongly coordinating anion, with 1,3-dialkylimidazolium cations with different side chains, including alkyl, ether, and hydroxyl functionalized, as well as their combinations. We evaluated their viscosity, thermal stabilities, and microstructure using Raman and infrared (IR) spectroscopies, allied to density functional theory (DFT) and ab initio molecular dynamics (AIMD) simulations. The viscosity data showed that the ether insertion significantly enhances the fluidity of the ILs, consistent with the coiling effect of the cation chain. Through a combined experimental and theoretical approach, we analyzed how the OAc anion interacts with ether ILs, revealing a characteristic bidentate coordination, particularly in hydroxyl functionalized ILs due to specific hydrogen bonding with the OH group. IR spectroscopy showed subtle shifts in the acidic hydrogens of imidazolium ring C(2)-H and C(4,5)-H, suggesting weaker interactions between OAc and the imidazolium ring in ether-functionalized ILs. Additionally, spatial distribution functions (SDF) and dihedral angle distribution obtained via AIMD confirmed the intramolecular hydrogen bonding due to the coiling effect of the ether side chain.

Biomolecular dynamics with machine-learned quantum-mechanical force fields trained on diverse chemical fragments

The GEMS method enables molecular dynamics simulations of large heterogeneous systems at ab initio quality.

The effect of machine learning predicted anharmonic frequencies on thermodynamic properties of fluid hydrogen fluoride

Anharmonic effects play a crucial role in determining thermochemical properties of liquids and gases. For such extended phases, the inclusion of anharmonicity in reliable electronic structure methods is computationally extremely demanding, and hence, anharmonic effects are often lacking in thermochemical calculations. In this study, we apply the quantum cluster equilibrium method to transfer density functional theory calculations at the cluster level to the macroscopic, liquid, and gaseous phase of hydrogen fluoride. This allows us to include anharmonicity, either via vibrational self-consistent field calculations for smaller clusters or using a regression model for larger clusters. We obtain the structural composition of the fluid phases in terms of the population of different clusters as well as isobaric heat capacities as an example for thermodynamic properties. We study the role of anharmonicities for these analyses and observe that, in particular, the dominating structural motifs are rather sensitive to the anharmonicity in vibrational frequencies. The regression model proves to be a promising way to get access to anharmonic features, and the extension to more sophisticated machine-learning models is promising.

Evidence of Excited-State Vibrational Mode Governing the Photorelaxation of a Charge-Transfer Complex

Modern, nonlinear, time-resolved spectroscopic techniques have opened new doors for investigating the intriguing but complex world of photoinduced ultrafast out-of-equilibrium phenomena and charge dynamics. The interaction between light and matter introduces an additional dimension, where the complex interplay between electronic and vibrational dynamics needs the most advanced theoretical-computational protocols to be fully understood on the molecular scale. In this study, we showcase the capabilities of ab initio molecular dynamics simulation integrated with a multiresolution wavelet protocol to carefully investigate the excited-state relaxation dynamics in a noncovalent complex involving tetramethylbenzene (TMB) and tetracyanoquinodimethane (TCNQ) undergoing charge transfer (CT) upon photoexcitation. Our protocol provides an accurate description that facilitates a direct comparison between transient vibrational analysis and time-resolved spectroscopic signals. This molecular level perspective enhances our understanding of photorelaxation processes confined in the adiabatic regime and offers an improved interpretation of vibrational spectra. Furthermore, it enables the quantification of anharmonic vibrational couplings between high- and low-frequency modes, specifically the TCNQ "rocking" and "bending" modes. Additionally, it identifies the primary vibrational mode that governs the adiabaticity between the ground state and the CT state. This comprehensive understanding of photorelaxation processes holds significant importance in the rational design and precise control of more efficient photovoltaic and sensor devices.

Complex-Valued K-Means Clustering of Interpolative Separable Density Fitting Algorithm for Large-Scale Hybrid Functional Enabled Ab Initio Molecular Dynamics Simulations within Plane Waves

K-means clustering, as a classic unsupervised machine learning algorithm, is the key step to select the interpolation sampling points in interpolative separable density fitting (ISDF) decomposition for hybrid functional electronic structure calculations. Real-valued K-means clustering for accelerating the ISDF decomposition has been demonstrated for large-scale hybrid functional enabled ab initio molecular dynamics (hybrid AIMD) simulations within plane-wave basis sets where the Kohn-Sham orbitals are real-valued. However, it is unclear whether such K-means clustering works for complex-valued Kohn-Sham orbitals. Here, we propose an improved weight function defined as the sum of the square modulus of complex-valued Kohn-Sham orbitals in K-means clustering for hybrid AIMD simulations. Numerical results demonstrate that the K-means algorithm with a new weight function yields smoother and more delocalized interpolation sampling points, resulting in smoother energy potential, smaller energy drift, and longer time steps for hybrid AIMD simulations compared to the previous weight function used in the real-valued K-means algorithm. In particular, we find that this improved algorithm can obtain more accurate oxygen-oxygen radial distribution functions in liquid water molecules and a more accurate power spectrum in crystal silicon dioxide compared to the previous K-means algorithm. Finally, we describe a massively parallel implementation of this ISDF decomposition to accelerate large-scale complex-valued hybrid AIMD simulations containing thousands of atoms (2,744 atoms), which can scale up to 5,504 CPU cores on modern supercomputers.

Transitioning from Methanol to Olefins (MTO) toward a Tandem CO2 Hydrogenation Process: On the Role and Fate of Heteroatoms (Mg, Si) in MAPO-18 Zeotypes

The tandem CO2 hydrogenation to hydrocarbons over mixed metal oxide/zeolite catalysts (OXZEO) is an efficient way of producing value-added hydrocarbons (platform chemicals and fuels) directly from CO2via methanol intermediate in a single reactor. In this contribution, two MAPO-18 zeotypes (M = Mg, Si) were tested and their performance was compared under methanol-to-olefins (MTO) conditions (350 °C, PCH3OH = 0.04 bar, 6.5 gCH3OH h-1 g-1), methanol/CO/H2 cofeed conditions (350 °C, PCH3OH/PCO/PH2 = 1:7.3:21.7 bar, 2.5 gCH3OH h-1 g-1), and tandem CO2 hydrogenation-to-olefin conditions (350 °C, PCO2/PH2 = 7.5:22.5 bar, 1.4-12.0 gMAPO-18 h molCO2-1). In the latter case, the zeotypes were mixed with a fixed amount of ZnO:ZrO2 catalyst, well-known for the conversion of CO2/H2 to methanol. Focus was set on the methanol conversion activity, product selectivity, and performance stability with time-on-stream. In situ and ex situ Fourier transform infrared spectroscopy (FT-IR), X-ray diffraction (XRD), solid-state nuclear magnetic resonance (NMR), sorption experiments, and ab initio molecular dynamics (AIMD) calculations were performed to correlate material performance with material characteristics. The catalytic tests demonstrated the better performance of MgAPO-18 versus SAPO-18 at MTO conditions, the much superior performance of MgAPO-18 under methanol/CO/H2 cofeeds, and yet the increasingly similar performance of the two materials under tandem conditions upon increasing the zeotype-to-oxide ratio in the tandem catalyst bed. In situ FT-IR measurements coupled with AIMD calculations revealed differences in the MTO initiation mechanism between the two materials. SAPO-18 promoted initial CO2 formation, indicative of a formaldehyde-based decarboxylation mechanism, while CO and ketene were the main constituents of the initiation pool in MgAPO-18, suggesting a decarbonylation mechanism. Under tandem CO2 hydrogenation conditions, the presence of high water concentrations and low methanol partial pressure in the reaction medium led to lower, and increasingly similar, methanol turnover frequencies for the zeotypes. Despite both MAPO-18 zeotypes showing signs of activity loss upon storage due to the interaction of the sites with ambient humidity, they presented a remarkable stability after reaching steady state under tandem reaction conditions and after steaming and regeneration cycles at high temperatures. Water adsorption experiments at room temperature confirmed this observation. The faster activity loss observed in the Mg version is assigned to its harder Mg2+-ion character and the higher concentration of CHA defects in the AEI structure, identified by solid-state NMR and XRD. The low stability of a MgAPO-34 zeotype (CHA structure) upon storage corroborated the relationship between CHA defects and instability.

Accelerating the Reaction Kinetics of CO2 Reduction to Multi-Carbon Products by Synergistic Effect between Cation and Aprotic Solvent on Copper Electrodes

Improving the selectivity of electrochemical CO2 reduction to multi-carbon products (C2+ ) is an important and highly challenging topic. In this work, we propose and validate an effective strategy to improve C2+ selectivity on Cu electrodes, by introducing a synergistic effect between cation (Na+ ) and aprotic solvent (DMSO) to the electrolyte. Based on constant potential ab initio molecular dynamics simulations, we first revealed that Na+ facilitates C-C coupling while inhibits CH3 OH/CH4 products via reducing the water network connectivity near the electrode. Furthermore, the water network connectivity was further decreased by introducing an aprotic solvent DMSO, leading to suppression of both C1 production and hydrogen evolution reaction with minimal effect on *OCCO* hydrogenation. The synergistic effect enhancing C2 selectivity was also experimentally verified through electrochemical measurements. The results showed that the Faradaic efficiency of C2 increases from 9.3 % to 57 % at 50 mA/cm2 under a mixed electrolyte of NaHCO3 and DMSO compared to a pure NaHCO3 , which can significantly enhance the selectivity of the C2 product. Therefore, our discovery provides an effective electrolyte-based strategy for tuning CO2 RR selectivity through modulating the microenvironment at the electrode-electrolyte interface.

Prediction and Control in DNA Nanotechnology

DNA nanotechnology is a rapidly developing field that uses DNA as a building material for nanoscale structures. Key to the field's development has been the ability to accurately describe the behavior of DNA nanostructures using simulations and other modeling techniques. In this Review, we present various aspects of prediction and control in DNA nanotechnology, including the various scales of molecular simulation, statistical mechanics, kinetic modeling, continuum mechanics, and other prediction methods. We also address the current uses of artificial intelligence and machine learning in DNA nanotechnology. We discuss how experiments and modeling are synergistically combined to provide control over device behavior, allowing scientists to design molecular structures and dynamic devices with confidence that they will function as intended. Finally, we identify processes and scenarios where DNA nanotechnology lacks sufficient prediction ability and suggest possible solutions to these weak areas.

MLatom 3: A Platform for Machine Learning-Enhanced Computational Chemistry Simulations and Workflows

Machine learning (ML) is increasingly becoming a common tool in computational chemistry. At the same time, the rapid development of ML methods requires a flexible software framework for designing custom workflows. MLatom 3 is a program package designed to leverage the power of ML to enhance typical computational chemistry simulations and to create complex workflows. This open-source package provides plenty of choice to the users who can run simulations with the command-line options, input files, or with scripts using MLatom as a Python package, both on their computers and on the online XACS cloud computing service at XACScloud.com. Computational chemists can calculate energies and thermochemical properties, optimize geometries, run molecular and quantum dynamics, and simulate (ro)vibrational, one-photon UV/vis absorption, and two-photon absorption spectra with ML, quantum mechanical, and combined models. The users can choose from an extensive library of methods containing pretrained ML models and quantum mechanical approximations such as AIQM1 approaching coupled-cluster accuracy. The developers can build their own models using various ML algorithms. The great flexibility of MLatom is largely due to the extensive use of the interfaces to many state-of-the-art software packages and libraries.

Pure and Sc-doped diopside (CaMgSi2O6) vibrational spectra: modelling and experiments

We investigated the structure of pure and Sc-doped synthetic diopside (a monoclinic single-chain silicate nominally CaMgSi2O6); in Sc-doped diopside, Sc3+ substitutes Mg2+ in the structure and, to achieve charge balance, vacancies form at the expense of Ca2+. We compared the structure obtained from ab initio modelling techniques at the density functional theory (DFT) level with the structure solved by employing single crystal X-ray diffraction. Furthermore, we compared IR and Raman spectroscopy experiments with vibrational density of states (VDOS) calculated from the Fourier transform of the velocity autocorrelation function obtained using ab initio (DFT) molecular dynamics simulations. In this framework, we developed a computational tool to assign the vibrational mode associated with a specific frequency. This method consists of projecting velocities along a specific set of internal coordinates such as stretching or bending, in cases involving defects or vacancies, to calculate a partial VDOS (pVDOS) that takes into account only the vibrations associated with selected internal modes, aiding the interpretation of the total VDOS and the experimental spectra in a relevant way. The computed data were validated with the experiments and we observed that doping the diopside structure with Sc produces peak broadening and the occurrence of new peaks in the Raman spectra and that site vacancies are associated with the nearby Sc site. The present work constitutes an interesting starting point to exploit the calculated VDOS/pVDOS to characterize experimental vibrational spectra of complex systems containing local vacancies, substitutions or defects as the Sc-doped diopside.

A Critical Assessment on Calculating Vibrational Spectra in Nanostructured Materials

Vibrational spectroscopy is an omnipresent spectroscopic technique to characterize functional nanostructured materials such as zeolites, metal-organic frameworks (MOFs), and metal-halide perovskites (MHPs). The resulting experimental spectra are usually complex, with both low-frequency framework modes and high-frequency functional group vibrations. Therefore, theoretically calculated spectra are often an essential element to elucidate the vibrational fingerprint. In principle, there are two possible approaches to calculate vibrational spectra: (i) a static approach that approximates the potential energy surface (PES) as a set of independent harmonic oscillators and (ii) a dynamic approach that explicitly samples the PES around equilibrium by integrating Newton's equations of motions. The dynamic approach considers anharmonic and temperature effects and provides a more genuine representation of materials at true operating conditions; however, such simulations come at a substantially increased computational cost. This is certainly true when forces and energy evaluations are performed at the quantum mechanical level. Molecular dynamics (MD) techniques have become more established within the field of computational chemistry. Yet, for the prediction of infrared (IR) and Raman spectra of nanostructured materials, their usage has been less explored and remain restricted to some isolated successes. Therefore, it is currently not a priori clear which methodology should be used to accurately predict vibrational spectra for a given system. A comprehensive comparative study between various theoretical methods and experimental spectra for a broad set of nanostructured materials is so far lacking. To fill this gap, we herein present a concise overview on which methodology is suited to accurately predict vibrational spectra for a broad range of nanostructured materials and formulate a series of theoretical guidelines to this purpose. To this end, four different case studies are considered, each treating a particular material aspect, namely breathing in flexible MOFs, characterization of defects in the rigid MOF UiO-66, anharmonic vibrations in the metal-halide perovskite CsPbBr3, and guest adsorption on the pores of the zeolite H-SSZ-13. For all four materials, in their guest- and defect-free state and at sufficiently low temperatures, both the static and dynamic approach yield qualitatively similar spectra in agreement with experimental results. When the temperature is increased, the harmonic approximation starts to fail for CsPbBr3 due to the presence of anharmonic phonon modes. Also, the spectroscopic fingerprints of defects and guest species are insufficiently well predicted by a simple harmonic model. Both phenomena flatten the potential energy surface (PES), which facilitates the transitions between metastable states, necessitating dynamic sampling. On the basis of the four case studies treated in this Review, we can propose the following theoretical guidelines to simulate accurate vibrational spectra of functional solid-state materials: (i) For nanostructured crystalline framework materials at low temperature, insights into the lattice dynamics can be obtained using a static approach relying on a few points on the PES and an independent set of harmonic oscillators. (ii) When the material is evaluated at higher temperatures or when additional complexity enters the system, e.g., strong anharmonicity, defects, or guest species, the harmonic regime breaks down and dynamic sampling is required for a correct prediction of the phonon spectrum. These guidelines and their illustrations for prototype material classes can help experimental and theoretical researchers to enhance the knowledge obtained from a lattice dynamics study.

How Aqueous Solvation Impacts the Frequencies and Intensities of Infrared Absorption Bands in Flavin: The Quest for a Suitable Solvent Model

FTIR spectroscopy accompanied by quantum chemical simulations can reveal important information about molecular structure and intermolecular interactions in the condensed phase. Simulations typically account for the solvent either through cluster quantum mechanical (QM) models, polarizable continuum models (PCM), or hybrid quantum mechanical/molecular mechanical (QM/MM) models. Recently, we studied the effect of aqueous solvent interactions on the vibrational frequencies of lumiflavin, a minimal flavin model, using cluster QM and PCM models. Those models successfully reproduced the relative frequencies of four prominent stretching modes of flavin's isoalloxazine ring in the diagnostic 1450-1750 cm-1 range but poorly reproduced the relative band intensities. Here, we extend our studies on this system and account for solvation through a series of increasingly sophisticated models. Only by combining elements of QM clusters, QM/MM, and PCM approaches do we obtain an improved agreement with the experiment. The study sheds light more generally on factors that can impact the computed frequencies and intensities of IR bands in solution.

Dynamical Control of Topology in Polar Skyrmions via Twisted Light

Twisted light carries a nonzero orbital angular momentum, that can be transferred from light to electrons and particles ranging from nanometers to micrometers. Up to now, the interplay between twisted light with dipolar systems has scarcely been explored, though the latter bear abundant forms of topologies such as skyrmions and embrace strong light-matter coupling. Here, using first-principles-based simulations, we show that twisted light can excite and drive dynamical polar skyrmions and transfer its nonzero winding number to ferroelectric ultrathin films. The skyrmion is successively created and annihilated alternately at the two interfaces, and experiences a periodic transition from a markedly "Bloch" to "Néel" character, accompanied with the emergence of a "Bloch point" topological defect with vanishing polarization. The dynamical evolution of skyrmions is connected to a constant jump of topological number between "0" and "1" over time. These intriguing phenomena are found to have an electrostatic origin. Our study thus demonstrates that, and explains why this unique light-matter interaction can be very powerful in creating and manipulating topological solitons in functional materials.