How accurate are approximate quantum chemical methods at modelling solute-solvent interactions in solvated clusters?

High-Level Quantum Chemical Prediction of C-F Bond Dissociation Energies of Perfluoroalkyl Substances

In this study, 550 C-F bond dissociation energies (BDEs) of a variety of per- and polyfluoroalkyl substances (PFASs) obtained from high-level DLPNO-CCSD(T)/CBS calculations were used to assess the accuracy of contemporary density functional theory (DFT) and semiempirical methods. DLPNO-CCSD(T)/CBS gas phase C-F BDEs fall between 404.9-550.7 kJ mol-1 and M06-2X and ωB97M-V in conjunction with the aug-cc-pVTZ basis set predicted BDEs closest to the benchmark level with a mean absolute deviation (MAD) of 7.3 and 8.3 kJ mol-1, respectively. It was observed that DFT prediction errors increase with the degree of fluorination and system size. As such, previous model chemistry recommendations based on smaller nonfluorinated systems may not be carried over to modeling the energetics of PFASs and related systems.

Phenomenological description of the acidity of the citric acid and its deprotonated species: informational-theoretical study

CONTEXT

In spite of the fact that molecular acidity is a fundamental physicochemical property of molecular systems, the vast majority of theoretical studies have focused attention on monoprotic acids and on the prediction of pKa's. Polyprotic acids, represent a challenge for electronic structure calculations since the multiple acidic sites result in a vast group of species with different conformations and reactivities. In this work, Information-theoretic (IT) concepts of localizability, order and uniformity are applied to the Citric Acid and its deprotonated species through the one-electron density functionals: Shannon entropy (S), Fisher information (I) and Disequilibrium (D), respectively. We pursue the goal of characterizing the acidity of the aforementioned species with the aim to associate the IT concepts to chemical features such as the polarizability of the protonated/deprotonated species, the liability of the acidic sites, atomic electrostatic potentials, covalent bonding. IT analyses looks very promising for future studies on the acidity of specific deprotonation-sites of polyprotic acids.

METHODS

Density functional theory (DFT) calculations were performed with Gaussian 09 program. A sensitivity analysis of the IT-measures was performed for the citric acid and the citrate using B3LYP, B3PW91, BPW91, M05-2X, M06-2X and PBEPBE functionals with the 6-311++g(3df,2p), 6-311++g(d,p), 6.311+g(d,p) and aug-cc-pVDZ basis sets. The rest of the analysis was performed with the M05-2X/6-311+G(d,p) level of theory. Additionally, aqueous media was considered by use of the SMD solvent model. The IT-measures were calculated using a suite of programs developed in our laboratory jointly with the DGRID software package.

Structural Sampling and Solvation Models for the Simulation of Electronic Spectra: Pyrazine as a Case Study

The impact of sampling methods on spectral broadening in the gas phase and on the convergence of spectra in aqueous solution when using microsolvation, continuum solvation, and hybrid models is studied using pyrazine as a test case. For the sake of comparing classical Maxwell-Boltzmann and Wigner samplings in the gas phase, static and time-resolved X-ray absorption spectra after photoexcitation to the lowest ¹B_2u(ππ*) state, as well as the static UV-vis absorption spectrum, are considered. In addition, the UV-vis absorption spectrum of pyrazine in aqueous solution is also computed in order to systematically investigate its convergence with the number of explicitly included solvent shells with and without taking bulk solvation effects into account with the conductor-like screening model to represent implicit water beyond such explicit solute complexes. Concerning the static and time-resolved X-ray absorption spectra of pyrazine at the carbon K-edge as well as its UV-vis absorption spectrum in the gas phase, we find that these spectra obtained with Wigner and Maxwell-Boltzmann samplings substantially agree. For the UV-vis absorption spectrum in the aqueous solution, only the first two energetically low-lying bands converge quickly with the size of the explicitly included solvation shells, either with or without an additional continuum solvation taken into account. In stark contrast, calculations of the higher-lying excitations relying on finite microsolvated clusters without additional continuum solvation severely suffer from unphysical charge-transfer excitations into Rydberg-like orbitals at the cluster/vacuum interface. This finding indicates that computational UV-vis absorption spectra covering sufficiently high-lying states converge only if continuum solvation of the explicitly microsolvated solutes is included in the models.

Assessment of DLPNO-CCSD(T)-F12 and its use for the formulation of the low-cost and reliable L-W1X composite method

In the present study, we have investigated the performance of RIJCOSX DLPNO-CCSD(T)-F12 methods for a wide range of systems. Calculations with a high-accuracy option ["DefGrid3 RIJCOSX DLPNO-CCSD(T₁ )-F12"] extrapolated to the complete-basis-set limit using the maug-cc-pV[D+d,T+d]Z basis sets provides fairly good agreements with the canonical CCSD(T)/CBS reference for a diverse set of thermochemical and kinetic properties [with mean absolute deviations (MADs) of ~1-2 kJ mol^-1 except for atomization energies]. On the other hand, the low-cost "RIJCOSX DLPNO-CCSD(T)-F12D" option leads to substantial deviations for certain properties, notably atomization energies (MADs of up to tens of kJ mol^-1 ). With the high-accuracy CBS approach, we have formulated the L-W1X method, which further includes a low-cost core-valence plus scalar-relativistic term. It shows generally good accuracy. For improved accuracies in specific cases, we advise replacing maug-cc-pV(n+d)Z with jun-cc-pV(n+d)Z for the calculation of electron affinities, and using well-constructed isodesmic-type reactions to obtain atomization energies. For medium-sized systems, DefGrid3 RIJCOSX DLPNO-CCSD(T₁ )-F12 calculations are several times faster than the corresponding canonical computation; the use of the local approximations (RIJCOSX and DLPNO) leads to a better scaling than that for the canonical calculation (from ~6-7 down to ~2-4 for our test systems). Thus, the DefGrid3 RIJCOSX DLPNO-CCSD(T₁ )-F12 method, and the L-W1X protocol that based on it, represent a useful means for obtaining accurate thermochemical quantities for larger systems.

Searching for a Reliable Density Functional for Molecule-Environment Interactions, Found B97M-V/def2-mTZVP

In the present study, we have examined density functional theory methods for the calculation of the interaction energy between a small molecule and its environment. For simple systems such as a neutral solute in a neutral solvent, good accuracy can be attained using low-cost "3c" methods, in particular r²SCAN-3c. When part(s) of the system is charged, the accurate computation of the interactions is more challenging. In these cases, we find the B97M-V/def2-mTZVP method to agree well with reference values; it also shows good accuracy for the more straightforward neutral systems. Thus, B97M-V/def2-mTZVP provides a means for accurate and low-cost computation of interaction energies, notably the binding between a substrate or a drug molecule and an enzyme, which may facilitate rational drug design.

Modeling the Conformational Preference of the Lignocellulose Interface and Its Interaction with Weak Acids

We have examined the conformational space of model systems for the hydrogen-bonded and covalent linkages between the sugar and lignin components of lignocellulose. Specifically, glucose and paracoumaryl alcohol moieties are used in our models. Multistage screening protocols are used to identify and validate a set of lowest-energy isomers. We found a cost-effective screening process involving an initial screening with DFTB3/3ob using a 20 kJ mol^-1 threshold, a refinement with the SCANh/6-31+G(2d,p) method with a 10 kJ mol^-1 cutoff, a third step at the DSD-PBEP86/ma-def2-TZVP//MS1-D3/6-31+G(2d,p) level with the same 10 kJ mol^-1 threshold, and a last step at the CCSD(T)/CBS//B3LYP/cc-pVTZ level with a tighter 5 kJ mol^-1 threshold. The use of machine learning (with the London and Axilrod-Teller-Muto potential) can further accelerate the screening process. In general, all low-energy conformers are characterized by hydrogen bonding between the sugar and lignin moieties. We examined the interactions of covalently bonded sugar-lignin models with weak acids (HSO₃^-, H₂PO₃^-, HSeO₃^-, H₂citrate^-, etc.) and found that they interacted strongly with the oxygen of the sugar-O-lignin linkage. Our results suggest that acids such as dihydrogen citrate may be attractive alternatives to the commonly used HSO₃^- for lignocellulose processing.

Ionization Energies and Redox Potentials of Hydrated Transition Metal Ions: Evaluation of Domain-Based Local Pair Natural Orbital Coupled Cluster Approaches

Hydrated transition metal ions are prototypical systems that can be used to model properties of transition metals in complex chemical environments. These seemingly simple systems present challenges for computational chemistry and are thus crucial in evaluations of quantum chemical methods for spin-state and redox energetics. In this work, we explore the applicability of the domain-based pair natural orbital implementation of coupled cluster (DLPNO-CC) theory to the calculation of ionization energies and redox potentials for hydrated ions of all first transition row (3d) metals in the 2+/3+ oxidation states, in connection with various solvation approaches. In terms of model definition, we investigate the construction of a minimally explicitly hydrated quantum cluster with a first and second hydration layer. We report on the convergence with respect to the coupled cluster expansion and the PNO space, as well as on the role of perturbative triple excitations. A recent implementation of the conductor-like polarizable continuum model (CPCM) for the DLPNO-CC approach is employed to determine self-consistent redox potentials at the coupled cluster level. Our results establish conditions for the convergence of DLPNO-CCSD(T) energetics and stress the absolute necessity to explicitly consider the second solvation sphere even when CPCM is used. The achievable accuracy for redox potentials of a practical DLPNO-based approach is, on average, 0.13 V. Furthermore, multilayer approaches that combine a higher-level DLPNO-CCSD(T) description of the first solvation sphere with a lower-level description of the second solvation layer are investigated. The present work establishes optimal and transferable methodological choices for employing DLPNO-based coupled cluster theory, the associated CPCM implementation, and cost-efficient multilayer derivatives of the approach for open-shell transition metal systems in complex environments.

The quantum chemical solvation of indole: accounting for strong solute-solvent interactions using implicit/explicit models

In this paper, we investigate the efficacy of different quantum chemical solvent modelling methods of indole in both water and methylcyclohexane solutions. The goal is to show that one can yield good photophysical properties in strongly coupled solute-solvent systems using standard DFT methods. We use standard and linearly-corrected Polarisable Continuum Models (PCM), as well as explicit solvation models, and compare the different model parameters, including the choice of density functional, basis set, and number of explicit solvent molecules. We demonstrate that implicit models overestimate energies and oscillator strengths. In particular, for indole-water, no level inversion is observed, suggesting a dielectric medium on its own is insufficient. In contrast, energies are seen to converge fairly rapidly with respect to cluster size towards experimentally measured properties in the explicit models. We find that the use of B3LYP with a diffuse basis set can adequately represent the photophysics of the system with a cluster size of between 9-12 explicit water molecules. Sampling of configurations from a molecular dynamics simulation suggests that the single point results are suitably representative of the solvated ensemble. For indole-water, we show that solvent reorganisation plays a significant role in stabilisation of the excited state energies. It is hoped that the findings and observations of this study will aid in the choice of solvation model parameters in future studies.

Assessment of DFT approaches in noble gas clathrate-like clusters: stability and thermodynamics

We have assessed the performance and accuracy of different wavefunction-based electronic structure methods, such as DFMP2 and domain-based local pair-natural orbital (DLPNO-CCSD(T)), as well as a variety of density functional theory (DFT) approaches on He@(H₂O)_N cage systems. We have selected representative clathrate-like structures corresponding to the building blocks present in each of the sI, sII and sH natural gas clathrate hydrates, and we have carefully studied the interaction between a He atom with each of their individual cages. We reported well-converged DFMP2 and DLPNO-CCSD(T) reference data, together with interaction and cohesive energies of four different density functionals (two GGA, revPBE and PW86PBE, and two hybrids, B3LYP and PBE0), including diverse dispersion correction schemes (D3(0), D3(BJ), D4 and XDM) for both He-filled and empty clathrate-like cages. After the analysis of the results, we came to the conclusion that the PW86PBE functional, with both XDM and D4 corrections, and the PBE0-D4 functional present reasonably adequate approaches to describe the guest-host noncovalent interactions that take place in such He clathrate hydrates. Taking into account that the He@sII is the only helium clathrate that scientists have been able to synthesize recently, we have performed a thermodynamic study on the individual 5¹² and 5¹²6⁴ cages present in the sII crystal. We determined the change in enthalpy, ΔH, and in Gibbs free energy, ΔG, at various temperatures and pressures, and we found out that in the range of experimental conditions the reactions associated with the encapsulation of the He atom inside the cages are exothermic and spontaneous. Finally, we highlighted the importance of an accurate description of the interaction in He@water mixtures, as a crucial component in construction of reliable data-driven models.

Establishing the accuracy of density functional approaches for the description of noncovalent interactions in ionic liquids

We test a number of dispersion corrected versatile Generalized Gradient Approximation (GGA) and meta-GGA functionals for their ability to predict the interactions of ionic liquids, and show that most can achieve energies within 1 kcal mol^-1 of benchmarks. This compares favorably with an accurate dispersion corrected hybrid, ωB97X-V. Our tests also reveal that PBE (Perdew-Burke-Ernzerhof GGA) calculations using the plane-wave projector augmented wave method and Gaussian Type Orbitals (GTOs) differ by less than 0.6 kJ mol^-1 for ionic liquids, despite ions being difficult to evaluate in periodic cells - thus revealing that GTO benchmarks may be used also for plane-wave codes. Finally, the relatively high success of explicit van der Waals density functionals, compared to elemental and ionic dispersion models, suggests that improvements are required for low-cost dispersion correction models of ions.

Accurate Quantum Chemical Prediction of Gas-Phase Anion Binding Affinities and Their Structure-Binding Relationships

This paper systematically examines the performance of contemporary wavefunction and density functional theory methods to identify robust and cost-efficient methods for predicting gas-phase anion binding energies. This includes the local coupled cluster LNO-CCSD(T) and DLPNO-CCSD(T), as well as double-hybrid DSD-PBEP86-D3(BJ) and various hybrid functionals M06-2X, B3LYP-D3(BJ), ωB97M-V, and ωB97X-V. The focus is on dual-hydrogen-bonding anion receptors that are commonly found in supramolecular chemistry and organocatalysis, namely, (thio)ureas, deltamides, (thio)squaramides, and croconamides as well as the yet-to-be-explored rhodizonamides. Of the methods examined, M06-2X emerged as the overall best performing method as the other functionals including DSD-PBEP86-D3(BJ) and the local coupled cluster DLPNO-CCSD(T) method displayed systematic errors that increase with the degree of carbonylation of the receptors. Hybrid ONIOM models that employed semiempirical methods (PM7, GFN1-xTB, and GFN2-xTB) and "threefold"-corrected small-basis set potentials (HF-3c, B97-3c, and PBEh-3c) were explored, and the best models resulted in 50- to 500-fold reduction in CPU time compared to W1-local. These calculations provide important insight into the structure-binding relationships where there is a direct correlation between Brønsted acidity and anion binding affinity, though the strength of the correlation also depends on other factors such as hydrogen-bonding geometry and the geometrical distortion that the receptor needs to undergo to bind the anion.

Accurate Thermochemistry for Main-Group Elements up to Xenon with the Wn-P34 Series of Composite Methods

In the present study, we introduce the accurate Wn-P34 quantum chemistry composite methods with applicability to heavy p-block elements up to xenon. For a set of thermochemical properties for prototypical third- and fourth-row species and for a diverse set of small light-main-group species, they show accuracies of ∼3 kJ mol^-1 or better. Overall, the Wn-P34 methods are comparable in accuracy to Wn, with a widened applicability to heavier elements. We have used Wn-P34 to compile the P34 set of accurate thermochemical values for heavy p-block species, and we have applied this set to assess a wide range of lower-cost methods. The results of our assessment show that the G4(MP2)-XK composite method provides adequate treatments for these species, but several widely used double-hybrid density functional theory (DH-DFT) methods show uncharacteristically large deviations. In contrast, we find it presently surprising that some pure and hybrid DFT methods such as TPSS and SCANh perform quite well. We hope that our findings and new tools would facilitate the application of computational chemistry for heavy elements, of which the properties are yet to be broadly explored.

Connecting Gas-Phase Computational Chemistry to Condensed Phase Kinetic Modeling: The State-of-the-Art

In recent decades, quantum chemical calculations (QCC) have increased in accuracy, not only providing the ranking of chemical reactivities and energy barriers (e.g., for optimal selectivities) but also delivering more reliable equilibrium and (intrinsic/chemical) rate coefficients. This increased reliability of kinetic parameters is relevant to support the predictive character of kinetic modeling studies that are addressing actual concentration changes during chemical processes, taking into account competitive reactions and mixing heterogeneities. In the present contribution, guidelines are formulated on how to bridge the fields of computational chemistry and chemical kinetics. It is explained how condensed phase systems can be described based on conventional gas phase computational chemistry calculations. Case studies are included on polymerization kinetics, considering free and controlled radical polymerization, ionic polymerization, and polymer degradation. It is also illustrated how QCC can be directly linked to material properties.

Polycyclic aromatic hydrocarbons: from small molecules through nano-sized species towards bulk graphene

We have examined the use of systematic bond-separation reactions and purposely constructed chemistry-preserving isodesmic reactions for the thermochemical calculation of aromatic hydrocarbon species. The bond-separation approach yields somewhat disappointing accuracy even when the reaction energies are obtained with generally robust composite and double-hybrid (DH) density functional theory (DFT) methods. In contrast, for the purposely constructed reactions, we find a dramatic improvement in the accuracy for energies calculated with all methods examined. Notably, for medium-sized aromatic hydrocarbons, we find that an effective approach for formulating a well-balanced reaction is to split the target species into two halves with an aromatic overlapping region. Overall, the G4(MP2)-XK, MPW2PLYP, MN15, PBE, and DC-DFTB3 methods are reasonable within their respective classes of methods for the calculation of bond-separation as well as chemistry-preserving isodesmic reactions. We have further computed per-carbon atomization energy (AE) for a series of D_6h benzene-type molecules, and thus obtained a formula for extrapolation to the graphene limit [AE_n = 711.5 × (1 - 1/n^0.640) kJ mol^-1, where n = number of carbons]. It suggests that nano-graphene with a length larger than 10 nm would resemble properties of bulk graphene, and conversely, downsizing a nano-graphene beyond this point may lead to considerably altered properties from the bulk.

Quantum Chemical Microsolvation by Automated Water Placement

We developed a quantitative approach to quantum chemical microsolvation. Key in our methodology is the automatic placement of individual solvent molecules based on the free energy solvation thermodynamics derived from molecular dynamics (MD) simulations and grid inhomogeneous solvation theory (GIST). This protocol enabled us to rigorously define the number, position, and orientation of individual solvent molecules and to determine their interaction with the solute based on physical quantities. The generated solute-solvent clusters served as an input for subsequent quantum chemical investigations. We showcased the applicability, scope, and limitations of this computational approach for a number of small molecules, including urea, 2-aminobenzothiazole, (+)-syn-benzotriborneol, benzoic acid, and helicene. Our results show excellent agreement with the available ab initio molecular dynamics data and experimental results.

Accuracy of DLPNO-CCSD(T): Effect of Basis Set and System Size

The DLPNO-CCSD(T) method is designed to study large molecular systems at significantly reduced cost relative to its canonical counterpart. However, the error in this approach is also size-extensive and relies on cancellation of errors for the calculation of relative energies. This work provides a direct comparison of canonical CCSD(T) and TightPNO DLPNO-CCSD(T) calculations of reaction energies and barriers of a broad range of chemical reactions. The dataset includes acidities, anion binding affinities, enolization, Diels-Alder, nucleophilic substitution, and atom transfer reactions and complements existing theoretical datasets in terms of system size as well as new reaction types (e.g., anion binding affinities and chlorine atom transfer reactions). The performance of DLPNO-CCSD(T) was further examined with respect to systematic variation of basis set and system size and amounts of nonbonded interaction present in the system. The errors in the DLPNO-CCSD(T) were found to be relatively insensitive to the choice of basis set for small systems but increase monotonically with system size. Additionally, calculations of barriers appear to be more challenging than reaction energies with errors exceeding 5 kJ mol^-1 for many Diels-Alder reactions. Further tests on three realistic organic reactions reveal the impact of the DLPNO approximation in calculating absolute and relative barriers that are important for predictions such as stereoselectivity.

Data-Driven Strategies for Accelerated Materials Design

The ongoing revolution of the natural sciences by the advent of machine learning and artificial intelligence sparked significant interest in the material science community in recent years. The intrinsically high dimensionality of the space of realizable materials makes traditional approaches ineffective for large-scale explorations. Modern data science and machine learning tools developed for increasingly complicated problems are an attractive alternative. An imminent climate catastrophe calls for a clean energy transformation by overhauling current technologies within only several years of possible action available. Tackling this crisis requires the development of new materials at an unprecedented pace and scale. For example, organic photovoltaics have the potential to replace existing silicon-based materials to a large extent and open up new fields of application. In recent years, organic light-emitting diodes have emerged as state-of-the-art technology for digital screens and portable devices and are enabling new applications with flexible displays. Reticular frameworks allow the atom-precise synthesis of nanomaterials and promise to revolutionize the field by the potential to realize multifunctional nanoparticles with applications from gas storage, gas separation, and electrochemical energy storage to nanomedicine. In the recent decade, significant advances in all these fields have been facilitated by the comprehensive application of simulation and machine learning for property prediction, property optimization, and chemical space exploration enabled by considerable advances in computing power and algorithmic efficiency.In this Account, we review the most recent contributions of our group in this thriving field of machine learning for material science. We start with a summary of the most important material classes our group has been involved in, focusing on small molecules as organic electronic materials and crystalline materials. Specifically, we highlight the data-driven approaches we employed to speed up discovery and derive material design strategies. Subsequently, our focus lies on the data-driven methodologies our group has developed and employed, elaborating on high-throughput virtual screening, inverse molecular design, Bayesian optimization, and supervised learning. We discuss the general ideas, their working principles, and their use cases with examples of successful implementations in data-driven material discovery and design efforts. Furthermore, we elaborate on potential pitfalls and remaining challenges of these methods. Finally, we provide a brief outlook for the field as we foresee increasing adaptation and implementation of large scale data-driven approaches in material discovery and design campaigns.

Anion Binding Affinity: Acidity versus Conformational Effects

High-level quantum chemical calculations were used to elucidate the gas- and solution-phase conformational equilibria for a series of symmetrically substituted (thio)ureas, (thio)squaramides, and croconamides. Gas-phase calculations predict that the thermodynamic conformer of many of these anion receptors is not the dual-hydrogen-bond-facilitating anti-anti conformer as is commonly assumed. For N,N'-diaryl thiosquaramides and croconamides, the syn-syn conformer is typically the predominant conformer. Solution-phase calculations show that the anti-anti conformer is increasingly stabilized as the polarity of the solvent increases. However, the syn-syn conformer remains the lowest energy conformation for croconamides. These predictions are used to explain the acidity versus chloride binding affinity correlations recently reported for some of these compounds. The chloride binding constants for thioureas and croconamides are significantly lower than expected on the basis of their pK_a values, and this may be due in part to the need for these receptors to reorganize into the anti-anti conformer. Experimental NMR nuclear Overhauser effect (NOE) measurements of an asymmetrically substituted squaramide and its thio analogue are consistent with the syn-syn conformation being predominant at 298 K. The conformational equilibria should therefore be an important consideration for the design and development of future anion receptors and organocatalysts.