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.

Recommendations for Velocity Adjustment in Surface Hopping.

This study investigates velocity adjustment directions after hopping in surface hopping dynamics. Using fulvene and a protonated Schiff base (PSB4) as case studies, we investigate the population decay and reaction yields of different sets of dynamics with the velocity adjusted in either the nonadiabatic coupling, gradient difference, or momentum directions. For the latter, in addition to the conventional algorithm, we investigated the performance of a reduced kinetic energy reservoir approach recently proposed. Our evaluation also considered velocity adjustment in the directions of approximate nonadiabatic coupling vectors. While results for fulvene are susceptible to the adjustment approach, PSB4 is not. We correlated this dependence to the topography near the conical intersections. When nonadiabatic coupling vectors are unavailable, the gradient difference direction is the best adjustment option. If the gradient difference is also unavailable, a semiempirical vector direction or the momentum direction with a reduced kinetic energy reservoir becomes an excellent option to prevent an artificial excess of back hoppings. The precise velocity adjustment direction is less crucial for describing the nonadiabatic dynamics than the kinetic energy reservoir's size.

A multi-descriptor analysis of substituent effects on the structure and aromaticity of benzene derivatives: π-Conjugation versus charge effects.

This work provides a detailed multi-component analysis of aromaticity in monosubstituted (X = CH3, C H 2 - , C H 2 + , NH2, NH-, NH+, OH, O-, and O+) and para-homodisubstituted (X = CH3, CH2, NH2, NH, OH, and O) benzene derivatives. We investigate the effects of substituents using single-reference (B3LYP/DFT) and multireference (CASSCF/MRCI) methods, focusing on structural (HOMA), vibrational (AI(vib)), topological (ELFπ), electronic (MCI), magnetic (NICS), and stability (S0-T1 splitting) properties. The findings reveal that appropriate π-electron-donating and π-electron-accepting substituents with suitable size and symmetry can interact with the π-system of the ring, significantly influencing π-electron delocalization. While the charge factor has a minimal impact on π-electron delocalization, the presence of a pz orbital capable of interacting with the π-electron delocalization is the primary factor leading to a deviation from the typical aromaticity characteristics observed in benzene.

Temperature effects on the internal conversion of excited adenine and adenosine.

This work aims to elucidate the dependence of the excited-state lifetime of adenine and adenosine on temperature. So far, it has been experimentally shown that while adenine's lifetime is unaffected by temperature, adenosine's lifetime strongly depends on it. However, the non-Arrhenius temperature dependence has posed a challenge in explaining this phenomenon. We used surface hopping to simulate the dynamics of adenine and adenosine in the gas phase at 0 and 400 K. The temperature effects were observed under the initial conditions via Wigner sampling with thermal corrections. Our results confirm that adenine's excited-state lifetime does not depend on temperature, while adenosine's lifetime does. Adenosine's dependency is due to intramolecular vibrational energy transfer from adenine to the ribose group. At 0 K, this transfer reduced the mean kinetic energy of adenine's moiety so much that internal conversion is inhibited, and the lifetime elongated by a factor of 2.3 compared to that at 400 K. The modeling also definitively ruled out the influence of viscosity, which was proposed as an alternative explanation previously.

WS22 database, Wigner Sampling and geometry interpolation for configurationally diverse molecular datasets.

Multidimensional surfaces of quantum chemical properties, such as potential energies and dipole moments, are common targets for machine learning, requiring the development of robust and diverse databases extensively exploring molecular configurational spaces. Here we composed the WS22 database covering several quantum mechanical (QM) properties (including potential energies, forces, dipole moments, polarizabilities, HOMO, and LUMO energies) for ten flexible organic molecules of increasing complexity and with up to 22 atoms. This database consists of 1.18 million equilibrium and non-equilibrium geometries carefully sampled from Wigner distributions centered at different equilibrium conformations (either at the ground or excited electronic states) and further augmented with interpolated structures. The diversity of our datasets is demonstrated by visualizing the geometries distribution with dimensionality reduction as well as via comparison of statistical features of the QM properties with those available in existing datasets. Our sampling targets broader quantum mechanical distribution of the configurational space than provided by commonly used sampling through classical molecular dynamics, upping the challenge for machine learning models.

Structural stability and the low-lying singlet and triplet states of BN-n-acenes, n = 1-7.

The chemical stability and the low-lying singlet and triplet excited states of BN-n-acenes (n = 1-7) were studied using single reference and multireference methodologies. From the calculations, descriptors such as the singlet-triplet splitting, the natural orbital (NO) occupations and aromaticity indexes are used to provide structural and energetic analysis. The boron and nitrogen atoms form an isoelectronic pair of two carbon atoms, which was used for the complete substitution of these units in the acene series. The structural analysis confirms the effects originated from the insertion of a uniform pattern of electronegativity difference within the molecular systems. The covalent bonds tend to be strongly polarized which does not happen in the case of a carbon-only framework. This effect leads to a charge transfer between neighbor atoms resulting in a more strengthened structure, keeping the aromaticity roughly constant along the chain. The singlet-triplet splitting also agrees with this stability trend, maintaining a consistent gap value for all molecules. The BN-n-acenes molecules possess a ground state with monoconfigurational character indicating their electronic stability. The low-lying singlet excited states have charge transfer character, which proceeds from nitrogen to boron.

Newton-X Platform: New Software Developments for Surface Hopping and Nuclear Ensembles.

Newton-X is an open-source computational platform to perform nonadiabatic molecular dynamics based on surface hopping and spectrum simulations using the nuclear ensemble approach. Both are among the most common methodologies in computational chemistry for photophysical and photochemical investigations. This paper describes the main features of these methods and how they are implemented in Newton-X. It emphasizes the newest developments, including zero-point-energy leakage correction, dynamics on complex-valued potential energy surfaces, dynamics induced by incoherent light, dynamics based on machine-learning potentials, exciton dynamics of multiple chromophores, and supervised and unsupervised machine learning techniques. Newton-X is interfaced with several third-party quantum-chemistry programs, spanning a broad spectrum of electronic structure methods.

Correction: Modeling the heating and cooling of a chromophore after photoexcitation.

Correction for 'Modeling the heating and cooling of a chromophore after photoexcitation' by Elizete Ventura et al., Phys. Chem. Chem. Phys., 2022, 24, 9403-9410, https://doi.org/10.1039/D2CP00686C.

Pre-Dewar structure modulates protonated azaindole photodynamics.

Recent experimental work revealed that the lifetime of the S3 state of protonated 7-azaindole is about ten times longer than that of protonated 6-azaindole. We simulated the nonradiative decay pathways of these molecules using trajectory surface hopping dynamics after photoexcitation into S3 to elucidate the reason for this difference. Both isomers mainly follow a common ππ* relaxation pathway involving multiple state crossings while coming down from S3 to S1 in the subpicosecond time scale. However, the simulations reveal that the excited-state topographies are such that while the 6-isomer can easily access the region of nonadiabatic transitions, the internal conversion of the 7-isomer is delayed by a pre-Dewar bond formation with a boat conformation.

Modeling the heating and cooling of a chromophore after photoexcitation.

The heating of a chromophore due to internal conversion and its cooling down due to energy dissipation to the solvent are crucial phenomena to characterize molecular photoprocesses. In this work, we simulated the ab initio nonadiabatic dynamics of cytosine, a prototypical chromophore undergoing ultrafast internal conversion, in three solvents-argon matrix, benzene, and water-spanning an extensive range of interactions. We implemented an analytical energy-transfer model to analyze these data and extract heating and cooling times. The model accounts for nonadiabatic effects, and excited- and ground-state energy transfer, and can analyze data from any dataset containing kinetic energy as a function of time. Cytosine heats up in the subpicosecond scale and cools down within 25, 4, and 1.3 ps in argon, benzene, and water, respectively. The time constants reveal that a significant fraction of the benzene and water heating occurs while cytosine is still electronically excited.

Simulations of molecular photodynamics in long timescales.

Nonadiabatic dynamics simulations in the long timescale (much longer than 10 ps) are the next challenge in computational photochemistry. This paper delimits the scope of what we expect from methods to run such simulations: they should work in full nuclear dimensionality, be general enough to tackle any type of molecule and not require unrealistic computational resources. We examine the main methodological challenges we should venture to advance the field, including the computational costs of the electronic structure calculations, stability of the integration methods, accuracy of the nonadiabatic dynamics algorithms and software optimization. Based on simulations designed to shed light on each of these issues, we show how machine learning may be a crucial element for long time-scale dynamics, either as a surrogate for electronic structure calculations or aiding the parameterization of model Hamiltonians. We show that conventional methods for integrating classical equations should be adequate to extended simulations up to 1 ns and that surface hopping agrees semiquantitatively with wave packet propagation in the weak-coupling regime. We also describe our optimization of the Newton-X program to reduce computational overheads in data processing and storage. This article is part of the theme issue 'Chemistry without the Born-Oppenheimer approximation'.

Choosing the right molecular machine learning potential.

Quantum-chemistry simulations based on potential energy surfaces of molecules provide invaluable insight into the physicochemical processes at the atomistic level and yield such important observables as reaction rates and spectra. Machine learning potentials promise to significantly reduce the computational cost and hence enable otherwise unfeasible simulations. However, the surging number of such potentials begs the question of which one to choose or whether we still need to develop yet another one. Here, we address this question by evaluating the performance of popular machine learning potentials in terms of accuracy and computational cost. In addition, we deliver structured information for non-specialists in machine learning to guide them through the maze of acronyms, recognize each potential's main features, and judge what they could expect from each one.

MLatom 2: An Integrative Platform for Atomistic Machine Learning.

Atomistic machine learning (AML) simulations are used in chemistry at an ever-increasing pace. A large number of AML models has been developed, but their implementations are scattered among different packages, each with its own conventions for input and output. Thus, here we give an overview of our MLatom 2 software package, which provides an integrative platform for a wide variety of AML simulations by implementing from scratch and interfacing existing software for a range of state-of-the-art models. These include kernel method-based model types such as KREG (native implementation), sGDML, and GAP-SOAP as well as neural-network-based model types such as ANI, DeepPot-SE, and PhysNet. The theoretical foundations behind these methods are overviewed too. The modular structure of MLatom allows for easy extension to more AML model types. MLatom 2 also has many other capabilities useful for AML simulations, such as the support of custom descriptors, farthest-point and structure-based sampling, hyperparameter optimization, model evaluation, and automatic learning curve generation. It can also be used for such multi-step tasks as Δ-learning, self-correction approaches, and absorption spectrum simulation within the machine-learning nuclear-ensemble approach. Several of these MLatom 2 capabilities are showcased in application examples.

Fewest switches surface hopping with Baeck-An couplings.

In the Baeck-An (BA) approximation, first-order nonadiabatic coupling vectors are given in terms of adiabatic energy gaps and the second derivative of the gaps with respect to the coupling coordinate. In this paper, a time-dependent (TD) BA approximation is derived, where the couplings are computed from the energy gaps and their second time-derivatives. TD-BA couplings can be directly used in fewest switches surface hopping, enabling nonadiabatic dynamics with any electronic structure methods able to provide excitation energies and energy gradients. Test results of surface hopping with TD-BA couplings for ethylene and fulvene show that the TD-BA approximation delivers a qualitatively correct picture of the dynamics and a semiquantitative agreement with reference data computed with exact couplings. Nevertheless, TD-BA does not perform well in situations conjugating strong couplings and small velocities. Considered the uncertainties in the method, TD-BA couplings could be a competitive approach for inexpensive, exploratory dynamics with a small trajectories ensemble. We also assessed the potential use of TD-BA couplings for surface hopping dynamics with time-dependent density functional theory (TDDFT), but the results are not encouraging due to singlet instabilities near the crossing seam with the ground state.

Interplay between Aromaticity and Radicaloid Character in Nitrogen-Doped Oligoacenes Revealed by High-Level Multireference Methods.

Aromaticity is a multivariable concept in organic chemistry that plays a central role for understanding the structure, stability, and reactivity of polycyclic aromatic hydrocarbons (PAHs). Several types of PAHs are characterized as singlet biradicaloid species and their chemical stability is intimately linked to the degree of aromatic character. In this study, theoretically designed routes to tune the biradical character (and thereby its chemical stability) of nitrogen-substituted octacenes have been investigated on the basis of the high-level multireference averaged quadratic coupled-cluster MR-AQCC method necessary for the appropriate description of polyradicaloid systems. The influence of nitrogen centers on the aromaticity of octacene is probed through structural (HOMA) and electron localization (ELF) indices by comparing the N- against NH-doping cases. These analyses reveal that the aromaticity and biradical character of octacene is only slightly affected by replacing one pair of CH groups with N atoms, i.e., by N-doping. However, a significant aromatic stabilization can be obtained when NH-doping is applied at the inner octacene rings; this is also accompanied by an overall decrease of the open-shell character, as evidenced by the gradual quenching of the unpaired electrons and increase in the singlet-triplet splittings when the NH doping groups are moved toward the center of the octacene molecule. Our findings aid in the rational design of new PAH compounds with balanced biradicaloid character and chemical stability which is important, e.g., for practical applications in organic solar cells based on the singlet-fission mechanism.

Tuning the Biradicaloid Nature of Polycyclic Aromatic Hydrocarbons: The Effect of Graphitic Nitrogen Doping in Zethrenes.

Zethrenes are interesting polycyclic aromatic hydrocarbons (PAHs), which possess unique optoelectronic and magnetic properties because of their singlet open-shell biradicaloid character, making them promising candidates for application in organic electronics. Tuning their properties is a key task in order to develop efficient compounds for practical use by balancing the desired biradicaloid character against its chemical instability. In this work, high-level theoretical multireference methods appropriate for the correct description of polyradicaloid systems are used to develop rules for doping of zethrenes by means of nitrogen taking heptazethrene (HZ) as a benchmark example. The results of the quantum chemical calculations have been concentrated on a series of quantitative descriptors such as unpaired densities and singlet-triplet (S-T) splittings. They clearly indicate different regions in the HZ where N-doping can either lead to strong enhancement of the biradicaloid character or to strong quenching towards a closed shell state. A wide scale of varying open-shell character is accessible from the different doping positions. It is shown that the S-T splittings correlate well with the total number of unpaired electrons in the medium range of biradicaloid character. For pronounced biradical character the S-T splitting decays to about zero with a margin of ±0.15 eV. In the opposite closed-shell limit, much larger S-T splittings of up to 3 eV are computed.

On the importance of non-covalent interactions for porous membranes: unraveling the role of pore size.

Membrane-based gas separation technology is of crucial importance in the current economy and nanoporous graphene, given its single-atomic layer, is an essential building-block material to achieve efficiency towards permeability and selectivity for such processes. Classically, pore size is the main feature that governs the diffusion energy barrier. Its nature, nevertheless, is also affected by other non-negligible physical mechanisms not yet discussed. Here we propose a theoretical study on the role of non-covalent interactions towards H2 diffusion through two graphene-based membranes. Symmetry-Adapted Perturbation Theory (SAPT) was used to investigate the total interaction energy and its physically meaningful components (electrostatics, exchange, induction and dispersion). The study reveals the importance of quantum effects such as polarization and electron delocalization in order to counterbalance the abiding idea of pore size being the dominant factor accounting for the energy barrier. These results have important implications for the rational design of efficient nanoporous devices for separation applications.

Singlet La and Lb Bands for N-Acenes (N = 2-7): A CASSCF/CASPT2 Study.

In this work CASPT2 calculations of polyacenes (from naphthalene to heptacene) were performed to find a methodology suitable for calculations of the absorption spectra, in particular of the La (B2u state) and Lb (B3u state) bands, of more extended systems. The effect of the extension of the active space and of freezing σ orbitals was investigated. The MCSCF excitation energy of the B2u state is not sensitive to the size of the active space used. However, the CASPT2 results depend strongly on the amount of σ orbitals frozen reflecting the ionic character of the B2u state. On the other hand, the excitation energies of the B3u state are much more sensitive to the size of the active space used in the calculations reflecting its multiconfigurational character. We found a good agreement with experimental data for both bands by including 14 electrons in 14 π orbitals in the active space followed by the CASPT2(14,14) perturbation scheme in which both σ and π orbitals are included.

How to efficiently tune the biradicaloid nature of acenes by chemical doping with boron and nitrogen.

Acenes are fascinating polyaromatic compounds that combine impressive semiconductor properties with an open-shell character by varying their molecular sizes. However, the increasing chemical instabilities related to their biradicaloid structures pose a great challenge for synthetic chemistry. Modifying the π-bond topology through chemical doping allows modulation of the electronic properties of graphene-related materials. In spite of the practical importance of these techniques, remarkably little is known about the basic question - the extent of the radical character created or quenched thereby. In this work, we report a high-level computational study on two acene oligomers doubly-doped with boron and nitrogen, respectively. These calculations demonstrate precisely which the chemical route is in order to either quench or enhance the radical character. Moving the dopants from the terminal rings to the central ones leads to a remarkable variation in the biradicaloid character (and thereby also in the chemical stability). This effect is related to a π-charge transfer involving the dopants and the radical carbon centers at the zigzag edges. This study also provides specific guidelines for a rational design of large polyaromatic compounds with enhanced chemical stability.