W4‐17: A diverse and high‐confidence dataset of atomization energies for benchmarking high‐level electronic structure methods

Highly precise values for the energy ratios underlying the Lieb-Oxford bound and the convexity conjecture for the adiabatic connection

The response function Kohn-Sham (KS) inversion method is employed to a set of 67 atoms and molecules to access the kinetic and potential energy contributions to the correlation energy, as well as the correlation energy itself. We use these energy contributions to compute highly reliable and accurate reference values for the energy ratios underlying the Lieb-Oxford bound and the convexity conjecture for the adiabatic connection. Commonly used approximate exchange-correlation functionals that go beyond the local density approximation lead to values for the energy ratios that agree surprisingly well with the calculated reference data. The largest value for the energy ratio corresponding to the Lieb-Oxford bound observed for the considered systems is 1.4024, which is well below the estimate of 1.9554 ≤λLO≤2.1346 for the Lieb-Oxford bound. The convexity conjecture for the adiabatic connection is not violated for any of the considered systems. We show that the numerical errors of the employed response function KS inversion method using Gaussian basis sets can be kept almost negligibly small by choosing an appropriate computational setup. The KS inversion method, furthermore, requires only moderate computational effort and, therefore, is well-suited to calculate reference data for various quantities of interest in Kohn-Sham density-functional theory for large numbers of molecules.

Beyond CCSD(T) Accuracy at Lower Scaling with Auxiliary Field Quantum Monte Carlo

We introduce a black-box auxiliary field quantum Monte Carlo (AFQMC) approach to perform highly accurate electronic structure calculations using configuration interaction singles and doubles (CISD) trial states. This method consistently provides more accurate energy estimates than coupled cluster singles and doubles with perturbative triples (CCSD(T)), often regarded as the gold standard in quantum chemistry. This level of precision is achieved at a lower asymptotic computational cost, scaling as O(N6) compared to the O(N7) scaling of CCSD(T). We provide numerical evidence supporting these findings through results for challenging main group and transition metal-containing molecules.

Massively Parallel Tensor Network State Algorithms on Hybrid CPU-GPU Based Architectures

The interplay of quantum and classical simulation and the delicate divide between them is in the focus of massively parallelized tensor network state (TNS) algorithms designed for high performance computing (HPC). In this contribution, we present novel algorithmic solutions together with implementation details to extend current limits of TNS algorithms on HPC infrastructure building on state-of-the-art hardware and software technologies. Benchmark results obtained via large-scale density matrix renormalization group (DMRG) simulations on single node multiGPU NVIDIA A100 system are presented for selected strongly correlated molecular systems addressing problems on Hilbert space dimensions up to 4.17 × 1035.

Post-CCSD(T) Thermochemistry of Chlorine Fluorides as a Challenging Test Case for Evaluating Density Functional Theory and Composite Ab Initio Methods

Quantum chemistry plays a key role in exploring the chemical properties of highly reactive chlorine polyfluoride compounds (ClFn). Here, we investigate the thermochemical properties of ClFn species (n=2-6) by means of high-level thermochemical procedures approximating the CCSDT(Q) and CCSDTQ5 energies at the complete basis set limit. We consider total atomization energies (TAEs), Cl-F bond dissociation energies (BDEs), F2 elimination energies (F2 elim.), ionization potentials (IPs), and electron affinities (EAs). The TAEs have significant contributions from post-CCSD(T) correlation effects. The higher-order triple excitations, CCSDT-CCSD(T), are negative and amount to -0.338 (ClF2), -0.727 (ClF3), -0.903 (ClF4), -1.335 (ClF5), and -1.946 (ClF6) kcal/mol. However, the contributions from quadruple (and, where available, also quintuple) excitations are much larger and positive and amount to +1.335 (ClF2), +1.387 (ClF3), +2.367 (ClF4), +2.399 (ClF5), and +3.432 (ClF6) kcal/mol. Thus, the contributions from post-CCSD(T) excitations exceed the threshold of chemical accuracy in nearly all cases. Due to their increasing hyper-valency and multireference character, the ClFn series provides an interesting and challenging test case for both density functional theory and low-level composite ab initio procedures. Here, we highlight the limitations in achieving overall chemical accuracy across all DFT and most composite ab initio procedures.

Adapting hybrid density functionals with machine learning

Exact exchange contributions significantly affect electronic states, influencing covalent bond formation and breaking. Hybrid density functional approximations, which average exact exchange admixtures empirically, have achieved success but fall short of high-level quantum chemistry accuracy due to delocalization errors. We propose adaptive hybrid functionals, generating optimal exact exchange admixture ratios on the fly using data-efficient quantum machine learning models with negligible overhead. The adaptive Perdew-Burke-Ernzerhof hybrid density functional (aPBE0) improves energetics, electron densities, and HOMO-LUMO gaps in QM9, QM7b, and GMTKN55 benchmark datasets. A model uncertainty-based constraint reduces the method smoothly to PBE0 in extrapolative regimes, ensuring general applicability with limited training. By tuning exact exchange fractions for different spin states, aPBE0 effectively addresses the spin gap problem in open-shell systems such as carbenes. We also present a revised QM9 (revQM9) dataset with more accurate quantum properties, including stronger covalent binding, larger bandgaps, more localized electron densities, and larger dipole moments.

Data-Driven Improvement of Local Hybrid Functionals: Neural-Network-Based Local Mixing Functions and Power-Series Correlation Functionals

Local hybrid functionals (LHs) use a real-space position-dependent admixture of exact exchange (EXX), governed by a local mixing function (LMF). The systematic construction of LMFs has been hampered over the years by a lack of exact physical constraints on their valence behavior. Here, we exploit a data-driven approach and train a new type of "n-LMF" as a relatively shallow neural network. The input features are of meta-GGA character, while the W4-17 atomization-energy and BH76 reaction-barrier test sets have been used for training. Simply replacing the widely used "t-LMF" of the LH20t functional by the n-LMF provides the LH24n-B95 functional. Augmented by DFT-D4 dispersion corrections, LH24n-B95-D4 remarkably improves the WTMAD-2 value for the large GMTKN55 test suite of general main-group thermochemistry, kinetics, and noncovalent interactions (NCIs) from 4.55 to 3.49 kcal/mol. As we found the limited flexibility of the B95c correlation functional to disfavor much further improvement on NCIs, we proceeded to replace it by an optimized B97c-type power-series expansion. This gives the LH24n functional. LH24n-D4 gives a WTMAD-2 value of 3.10 kcal/mol, the so far lowest value of a rung 4 functional in self-consistent calculations. The new functionals perform moderately well for organometallic transition-metal energetics while leaving room for further data-driven improvements in that area. Compared to complete neural-network functionals like DM21, the present more tailored approach to train just the LMF in a flexible but well-defined human-designed LH functional retains the possibility of graphical LMF analyses to gain deeper understanding. We find that both the present n-LMF and the recent x-LMF suppress the so-called gauge problem of local hybrids without adding a calibration function as required for other LMFs. LMF plots show that this can be traced back to large LMF values in the small-density region between the interacting atoms in NCIs for n- and x-LMFs and low values for the t-LMF. We also find that the trained n-LMF has relatively large values in covalent bonds without deteriorating binding energies. The current approach enables fast and efficient routine self-consistent calculations using n-LMFs in Turbomole.

Top-Down versus Bottom-Up Approaches for σ-Functionals Based on the Approximate Exchange Kernel

Recently, we investigated a number of so-called σ- and τ-functionals based on the adiabatic-connection fluctuation-dissipation theorem (ACFDT); particularly, extensions of the random phase approximation (RPA) with inclusion of an exchange kernel in the form of an antisymmetrized Hartree kernel. One of these functionals, based upon the approximate exchange kernel (AXK) of Bates and Furche, leads to a nonlinear contribution of the spline function used within σ-functionals, which we previously avoided through the introduction of a simplified "top-down" approach in which the σ-functional modification is inserted a posteriori following the analytic coupling strength integration within the framework of the ACFDT and which was shown to provide excellent performance for the GMTKN55 database when using hybrid PBE0 reference orbitals. In this work, we examine the analytic "bottom-up" approach in which the spline function is inserted a priori, i.e., before evaluation of the analytic coupling strength integral. The new bottom-up functionals, denoted σ↑AXK, considerably improve upon their top-down counterparts for problems dominated by self-interaction and delocalization errors. Despite a small loss of accuracy for noncovalent interactions, the σ↑AXK@PBE0 functionals comprehensively outperform regular σ-functionals, scaled σ-functionals, and the previously derived σ+SOSEX- and τ-functionals in the WTMAD-1 and WTMAD-2 metrics of the GMTKN55 database.

Highly accurate real-space electron densities with neural networks

Variational ab initio methods in quantum chemistry stand out among other methods in providing direct access to the wave function. This allows, in principle, straightforward extraction of any other observable of interest, besides the energy, but, in practice, this extraction is often technically difficult and computationally impractical. Here, we consider the electron density as a central observable in quantum chemistry and introduce a novel method to obtain accurate densities from real-space many-electron wave functions by representing the density with a neural network that captures known asymptotic properties and is trained from the wave function by score matching and noise-contrastive estimation. We use variational quantum Monte Carlo with deep-learning Ansätze to obtain highly accurate wave functions free of basis set errors and from them, using our novel method, correspondingly accurate electron densities, which we demonstrate by calculating dipole moments, nuclear forces, contact densities, and other density-based properties.

Exploring the influence of (n - 1)d subvalence correlation and of spin-orbit coupling on chalcogen bonding

This article presents a comprehensive computational investigation into chalcogen bonding interactions, focusing specifically on elucidating the role of subvalence (n - 1)d and (n - 1)sp correlation. The incorporation of inner-shell (n - 1)d correlation leads to a decrease in interaction energies for chalcogen-bonded systems (at least those studied herein), contradicting the observations regarding halogen bonding documented by Kesharwani et al. in J. Phys. Chem. A, 2018, 122 (8), 2184-2197. The significance of (n - 1)sp subvalence correlation appears to be lower by an order of magnitude. Notably, among the various components of interaction energies computed at the PNO-LCCSD(T) or DF-CCSD levels, we identify the PNO-LMP2 or DF-MP2 component of the (n - 1)d correlation as predominant. Furthermore, we delve into the impact of second-order spin-orbit coupling (SOC2) on these interactions. The SOC2 effects appear to be less significant than the (n - 1)d correlation; however, they remain non-trivial, particularly for Te complexes. For the Se complexes, SOC2 is much less important. Generally, SOC2 stabilizes monomers more than dimers, resulting in reduced binding of the latter. Notably, at equilibrium and stretched geometries, SOC2 and (n - 1)d destabilize the complex; however, at compressed geometries, they exhibit opposing effects, with (n - 1)d becoming stabilizing.

Does Basis Set Superposition Error Significantly Affect Post-CCSD(T) Corrections?

We have investigated the title question for both a subset of the W4-11 total atomization energies benchmark, and for the A24x8 noncovalent interactions benchmark. Overall, counterpoise corrections to post-CCSD(T) contributions are about two orders of magnitude less important than those to the CCSD(T) interaction energy. Counterpoise corrections for connected quadruple substitutions (Q) are negligible, andQ Λ - Q $$ {(Q)}_{\Lambda}-(Q) $$ orT 4 - Q $$ {T}_4-(Q) $$ especially so. In contrast, for atomization energies, theT 3 - T $$ {T}_3-(T) $$ counterpoise correction can reach about 0.05 kcal/mol for small basis sets like cc-pVDZ, thought it rapidly tapers off with cc-pVTZ and especially aug-cc-pVTZ basis sets. It is reduced to insignificance by the extrapolation ofT 3 - T $$ {T}_3-(T) $$ applied in both W4 and HEAT thermochemistry protocols. In noncovalent dimers, the differential BSSE on post-CCSD(T) correlation contributions is negligible even in basis sets as small as the unpolarized split-valence cc-pVDZ(no d).

Assessing Computational Methods to Calculate the Binding Energies of Dimers of Five-Membered Heterocyclic Molecules

Computational electronic structure methods, including ab initio and density functional theory (DFT), have been assessed in calculating the binding energies of 14 five-membered heterocyclic dimers. The configurations were generated using classical molecular dynamics before optimization at the MP2/aug-cc-pVTZ. Benchmark binding energies are calculated at the CCSD(T)/CBS level of theory. Among the ab initio methods, the DLPNO-CCSD(T)/CBS method has the best performance, reproducing CCSD(T)/CBS with a mean absolute deviation (MAD) of 0.17 kcal/mol. In addition, a schematic CCSD(T)/CBS approach perfectly reproduces the canonical CCSD(T)/CBS with a mean absolute error of 0.08 kcal/mol. Regarding DFT functionals, it has been found that counterpoise corrections have negligible effects on the accuracy of the functionals. Furthermore, including the D3 empirical dispersion considerably enhances the accuracy of the DFT functionals. As a result, outstanding performance is noted for the double hybrid functional B2K-PLYP, with a mean absolute error of 0.25 kcal/mol. In addition to the B2K-PLYP double hybrid functional, M05-D3, B97D, M05-2X-D3, M05-2X, M06-HF, M08-HX, M11, TPSSh-D3, and RSX-0DH-D3(BJ) have MAD values lower than 0.5 kcal/mol. These functionals are recommended for further investigations of five-membered heterocyclic clusters.

Data Quality in the Fitting of Approximate Models: A Computational Chemistry Perspective

Empirical parametrization underpins many scientific methodologies including certain quantum-chemistry protocols [e.g., density functional theory (DFT), machine-learning (ML) models]. In some cases, the fitting requires a large amount of data, necessitating the use of data obtained using low-cost, and thus low-quality, means. Here we examine the effect of using low-quality data on the resulting method in the context of DFT methods. We use multiple G2/97 data sets of different qualities to fit the DFT-type methods. Encouragingly, this fitting can tolerate a relatively large proportion of low-quality fitting data, which may be attributed to the physical foundations of the DFT models and the use of a modest number of parameters. Further examination using "ML-quality" data shows that adding a large amount of low-quality data to a small number of high-quality ones may not offer tangible benefits. On the other hand, when the high-quality data is limited in scope, diversification by a modest amount of low-quality data improves the performance. Quantitatively, for parametrizing DFT (and perhaps also quantum-chemistry ML models), caution should be taken when more than 50% of the fitting set contains questionable data, and that the average error of the full set is more than 20 kJ mol-1. One may also follow the recently proposed transferability principles to ensure diversity in the fitting set.

Accurate Thermochemistry with Multireference Methods: A Stress Test for Internally Contracted Multireference Coupled-Cluster Theory

The internally contracted multireference coupled-cluster method with single, double and perturbative triple excitations, icMRCCSD(T), was tested for its performance in the context of computational high-accuracy thermochemistry. The results were gauged against the standard single-reference coupled-cluster hierarchy with up to 5-fold excitations. The test set comprised of a selection of first-row dinuclear compounds and the three 3d-transition metal compounds MnH, FeH, and CoH. The results revealed two problems with the current formulation of icMRCCSD(T). First, the choice of the Dyall Hamiltonian as the zeroth-order Hamiltonian, which leads to a biased description of the different orbital subspaces and particularly poor results for the atomic correlation energies, and second, the tendency to overestimate the perturbative correction for triply excited clusters, in particular in the presence of open shells and correspondingly low orbital-energy gaps. The two problems could be solved by resorting to the effective Fock operator as zeroth-order Hamiltonian and by adopting a modified amplitude equation that includes terms quadratic in the pair clusters. A similar modification was recently proposed by Masios et al. (Phys. Rev. Lett. 2023, 131, 186401) in the context of applying single-reference coupled-cluster theory to systems with small or vanishing band gaps and we chose the acronym '(cT*) correction' in analogy to that work. In contrast to the work of Masios et al., additional terms including single excitation clusters were omitted, as these again lead to an overestimation of correlation effects in more difficult cases. We also tested another alternative for the zeroth-order Hamiltonian and additional higher-order corrections for the correlation energy. These extensions did not significantly improve the results and were also computationally more demanding. The improved icMRCCSD(cT*)F method yields very accurate results with errors, relative to accurate benchmarks, better than 2 kJ/mol for total energies and atomization energies for the entire set of examples considered in this work.

Prospects for rank-reduced CCSD(T) in the context of high-accuracy thermochemistry

Obtaining sub-chemical accuracy (1 kJ mol-1) for reaction energies of medium-sized gas-phase molecules is a longstanding challenge in the field of thermochemical modeling. The perturbative triples correction to coupled-cluster single double triple [CCSD(T)] constitutes an important component of all high-accuracy composite model chemistries that obtain this accuracy but can be a roadblock in the calculation of medium to large systems due to its O(N7) scaling, particularly in HEAT-like model chemistries that eschew separation of core and valence correlation. This study extends the work of Lesiuk [J. Chem. Phys. 156, 064103 (2022)] with new approximate methods and assesses the accuracy of five different approximations of (T) in the context of a subset of molecules selected from the W4-17 dataset. It is demonstrated that all of these approximate methods can achieve sub-0.1 kJ mol-1 accuracy with respect to canonical, density-fitted (T) contributions with a modest number of projectors. The approximation labeled Z̃T appears to offer the best trade-off between cost and accuracy and shows significant promise in an order-of-magnitude reduction in the computational cost of the CCSD(T) component of high-accuracy model chemistries.

Factorized Quadruples and a Predictor of Higher-Level Correlation in Thermochemistry

Coupled cluster theory has had a momentous impact on the ab initio prediction of molecular properties, and remains a staple ingratiate in high-accuracy thermochemical model chemistries. However, these methods require inclusion of at least some connected quadruple excitations, which generally scale at best as O ( N 9 ) with the number of basis functions. It is very difficult to predict, a priori, the effect correlation of past CCSD(T) on a given reaction energy. The purpose of this work is to examine cost-effective quadruple corrections based on the factorization theorem of the many-body perturbation theory that may address these challenges. We show that the O ( N 7 ) factorized CCSD(TQf) method introduces minimal error to predicted correlation and reaction energies as compared to the O ( N 9 ) CCSD(TQ). Further, we examine the performance of Goodson's continued fraction method in the estimation of CCSDT(Q)Λ contributions to reaction energies as well as a "new" method related to %TAE[(T)] that we refer to as a scaled perturbation estimator. We find that the scaled perturbation estimator based upon CCSD(TQf)/cc-pVDZ is capable of predicting CCSDT(Q)Λ/cc-pVDZ contributions to reaction energies with an average error of 0.07 kcal mol-1 and an L2D of 0.52 kcal mol-1 when applied to a test-suite of nearly 3000 reactions. This offers a means by which to reliably "ballpark" how important post-CCSD(T) contributions are to reaction energies while incurring no more than CCSD(T) formal cost and a little mental math.

Basis Set Extrapolation from the Vanishing Counterpoise Correction Condition

Basis set extrapolations are typically rationalized either from analytical arguments involving the partial-wave or principal expansions of the correlation energy in helium-like systems or from fitting extrapolation parameters to reference energetics for a small(ish) training set. Seeking to avoid both, we explore a third alternative: extracting extrapolation parameters from the requirement that the BSSE (basis set superposition error) should vanish at the complete basis set limit. We find this to be a viable approach provided that the underlying basis sets are not too small and reasonably well balanced. For basis sets not augmented by diffuse functions, BSSE minimization and energy fitting yield quite similar parameters.

State-of-the-art local correlation methods enable affordable gold standard quantum chemistry for up to hundreds of atoms

In this feature, we review the current capabilities of local electron correlation methods up to the coupled cluster model with single, double, and perturbative triple excitations [CCSD(T)], which is a gold standard in quantum chemistry. The main computational aspects of the local method types are assessed from the perspective of applications, but the focus is kept on how to achieve chemical accuracy (i.e., <1 kcal mol-1 uncertainty), as well as on the broad scope of chemical problems made accessible. The performance of state-of-the-art methods is also compared, including the most employed DLPNO and, in particular, our local natural orbital (LNO) CCSD(T) approach. The high accuracy and efficiency of the LNO method makes chemically accurate CCSD(T) computations accessible for molecules of hundreds of atoms with resources affordable to a broad computational community (days on a single CPU and 10-100 GB of memory). Recent developments in LNO-CCSD(T) enable systematic convergence and robust error estimates even for systems of complicated electronic structure or larger size (up to 1000 atoms). The predictive power of current local CCSD(T) methods, usually at about 1-2 order of magnitude higher cost than hybrid density functional theory (DFT), has become outstanding on the palette of computational chemistry applicable for molecules of practical interest. We also review more than 50 LNO-based and other advanced local-CCSD(T) applications for realistic, large systems across molecular interactions as well as main group, transition metal, bio-, and surface chemistry. The examples show that properly executed local-CCSD(T) can contribute to binding, reaction equilibrium, rate constants, etc. which are able to match measurements within the error estimates. These applications demonstrate that modern, open-access, and broadly affordable local methods, such as LNO-CCSD(T), already enable predictive computations and atomistic insight for complicated, real-life molecular processes in realistic environments.

Density Functional Tight-Binding Models for Band Structures of Transition-Metal Alloys and Surfaces across the d-Block

First-principles electronic structure simulations are an invaluable tool for understanding chemical bonding and reactions. While machine-learning models such as interatomic potentials significantly accelerate the exploration of potential energy surfaces, electronic structure information is generally lost. Particularly in the field of heterogeneous catalysis, simulated electron band structures provide fundamental insights into catalytic reactivity. This ab initio knowledge is preserved in semiempirical methods such as density functional tight binding (DFTB), which extend the accessible computational length and time scales beyond first-principles approaches. In this paper we present Shell-Optimized Atomic Confinement (SOAC) DFTB electronic-part-only parametrizations for bulk and surface band structures of all d-block transition metals that enable efficient predictions of electronic descriptors for large structures or high-throughput studies on complex systems outside the computational reach of density functional theory.

Harmonic and anharmonic vibrational computations for biomolecular building blocks: Benchmarking DFT and basis sets by theoretical and experimental IR spectrum of glycine conformers

Advanced vibrational spectroscopic experiments have reached a level of sophistication that can only be matched by numerical simulations in order to provide an unequivocal analysis, a crucial step to understand the structure-function relationship of biomolecules. While density functional theory (DFT) has become the standard method when targeting medium-size or larger systems, the problem of its reliability and accuracy are well-known and have been abundantly documented. To establish a reliable computational protocol, especially when accuracy is critical, a tailored benchmark is usually required. This is generally done over a short list of known candidates, with the basis set often fixed a priori. In this work, we present a systematic study of the performance of DFT-based hybrid and double-hybrid functionals in the prediction of vibrational energies and infrared intensities at the harmonic level and beyond, considering anharmonic effects through vibrational perturbation theory at the second order. The study is performed for the six-lowest energy glycine conformers, utilizing available "state-of-the-art" accurate theoretical and experimental data as reference. Focusing on the most intense fundamental vibrations in the mid-infrared range of glycine conformers, the role of the basis sets is also investigated considering the balance between computational cost and accuracy. Targeting larger systems, a broad range of hybrid schemes with different computational costs is also tested.

Machine learning for accuracy in density functional approximations

Machine learning techniques have found their way into computational chemistry as indispensable tools to accelerate atomistic simulations and materials design. In addition, machine learning approaches hold the potential to boost the predictive power of computationally efficient electronic structure methods, such as density functional theory, to chemical accuracy and to correct for fundamental errors in density functional approaches. Here, recent progress in applying machine learning to improve the accuracy of density functional and related approximations is reviewed. Promises and challenges in devising machine learning models transferable between different chemistries and materials classes are discussed with the help of examples applying promising models to systems far outside their training sets.

Small Representative Databases for Testing and Validating Density Functionals and Other Electronic Structure Methods

Broad and diverse sets of accurate data provide useful metrics for assessing the performance of new theoretical methods. However, assessing methods against large databases can be an arduous task. Here, we present 17 representative energetic databases, defined as small databases whose errors and error spreads are representative of larger databases and which therefore can serve as efficient benchmarks for developing and testing electronic structure methods and density functionals. In 15 cases, the representative databases have 6 entries while being representative of larger databases with 14-107 entries, and in the other two cases, they have 14 entries while being representative of larger databases with 418-455 entries. The mean unsigned error (MUE) of 100 electronic structure methods on a given representative database is typically within about 8% of the MUE on its parent database, and the root-mean-square error (RMSE) is typically within about 11% of the RMSE on the parent database. Thus, the representative databases are quite successful in indicating accuracy while maintaining good diversity. The databases include both main-group and transition-metal compounds and reactions, and they include bond energies, reaction energies, barrier heights, noncovalent interactions, ionization potentials, and absolute energies.

Scalable Ab Initio Electronic Structure Methods with Near Chemical Accuracy for Main Group Chemistry

This study evaluates the precision of widely recognized quantum chemical methodologies, CCSD(T), DLPNO-CCSD(T), and localized ph-AFQMC, for determining the thermochemistry of main group elements. DLPNO-CCSD(T) and localized ph-AFQMC, which offer greater scalability compared to canonical CCSD(T), have emerged over the past decade as pivotal in producing precise benchmark chemical data. Our investigation includes closed-shell, neutral molecules, focusing on their heat of formation and atomization energy sourced from four specific small molecule data sets. First, we selected molecules from the G2 and G3 data sets, noted for their reliable experimental heat of formation data. Additionally, we incorporate molecules from the W4-11 and W4-17 sets, which provide high-level theoretical reference values for atomization energy at 0 K. Our findings reveal that both DLPNO-CCSD(T) and ph-AFQMC methods are capable of achieving a root-mean-square deviation of less than 1 kcal/mol across the combined data set, aligning with the threshold for chemical accuracy. Moreover, we make efforts to confine the maximum deviations within 2 kcal/mol, a degree of precision that significantly broadens the applicability of these methods in fields such as biology and materials science.

Softmax parameterization of the occupation numbers for natural orbital functionals based on electron pairing approaches

Within the framework of natural orbital functional theory, having a convenient representation of the occupation numbers and orbitals becomes critical for the computational performance of the calculations. Recognizing this, we propose an innovative parametrization of the occupation numbers that takes advantage of the electron-pairing approach used in Piris natural orbital functionals through the adoption of the softmax function, a pivotal component in modern deep-learning models. Our approach not only ensures adherence to the N-representability of the first-order reduced density matrix (1RDM) but also significantly enhances the computational efficiency of 1RDM functional theory calculations. The effectiveness of this alternative parameterization approach was assessed using the W4-17-MR molecular set, which demonstrated faster and more robust convergence compared to previous implementations.

Toward Accurate Quantum Chemical Methods for Molecules of Increasing Dimension: The New Family of Pisa Composite Schemes

The new versions of the Pisa composite scheme introduced in the present paper are based on the careful selection of different quantum chemical models for energies, geometries, and vibrational frequencies, with the aim of maximizing the accuracy of the overall description while retaining a reasonable cost for all the steps. In particular, the computation of accurate electronic energies has been further improved introducing more reliable complete basis set extrapolations and estimation of core-valence correlation, together with improved basis sets for third-row atoms. Furthermore, the reduced-cost frozen natural orbital (FNO) model has been introduced and validated for large molecules. Accurate molecular structures can be obtained avoiding complete basis set extrapolation and evaluating core-valence correlation at the MP2 level. Unfortunately, analytical gradients are not available for the FNO version of the model. Therefore, for large molecules, an accurate reduced-cost alternative is offered by evaluation of valence contributions with a double-hybrid functional in conjunction with the same MP2 contribution for core-valence correlation or by means of a one-parameter approximation. The same double-hybrid functional and basis set are employed to evaluate zero-point energies and partition functions. After the validation of the new models for small systems, a panel of molecular bricks of life has been used to analyze their performances for problems of current fundamental or technological interest. The fully black-box implementation of the computational workflow paves the way toward the accurate yet not prohibitively expensive study of medium- to large-sized molecules also by experimentally oriented researchers.

Big data benchmarking: how do DFT methods across the rungs of Jacob's ladder perform for a dataset of 122k CCSD(T) total atomization energies?

Total atomization energies (TAEs) are a central quantity in density functional theory (DFT) benchmark studies. However, so far TAE databases obtained from experiment or high-level ab initio wavefunction theory included up to hundreds of TAEs. Here, we use the GDB-9 database of 133k CCSD(T) TAEs generated by Curtiss and co-workers [B. Narayanan, P. C. Redfern, R. S. Assary and L. A. Curtiss, Chem. Sci., 2019, 10, 7449] to evaluate the performance of 14 representative DFT methods across the rungs of Jacob's ladder (namely, PBE, BLYP, B97-D, M06-L, τ-HCTH, PBE0, B3LYP, B3PW91, ωB97X-D, τ-HCTHh, PW6B95, M06, M06-2X, and MN15). We first use the A25[PBE] diagnostic for nondynamical correlation to eliminate systems that potentially include significant multireference effects, for which the CCSD(T) TAEs might not be sufficiently reliable. The resulting database (denoted by GDB9-nonMR) includes 122k species. Of the considered functionals, B3LYP attains the best performance relative to the G4(MP2) reference TAEs, with a mean absolute deviation (MAD) of 4.09 kcal mol-1. This first-generation hybrid functional, in which the three mixing coefficients were fitted against a small set of TAEs, is one of the few functionals that are not systematically biased towards overestimating the G4(MP2) TAEs, as demonstrated by a mean-signed deviation (MSD) of 0.45 kcal mol-1. The relatively good performance of B3LYP is followed by the heavily parameterized M06-L meta-GGA functional, which attains a MAD of 6.24 kcal mol-1. The PW6B95, M06, M06-2X, and MN15 functionals tend to systematically overestimate the G4(MP2) TAEs and attain MADs ranging between 18.69 (M06) and 28.54 (MN15) kcal mol-1. However, PW6B95 and M06-2X exhibit particularly narrow error distributions. Thus, scaling their TAEs by an empirical scaling factor reduces their MADs to merely 3.38 (PW6B95) and 2.85 (M06-2X) kcal mol-1. Empirical dispersion corrections (e.g., D3 and D4) are attractive, and therefore, their inclusion worsens the performance of methods that systematically overestimate the TAEs.

On the sensitivity of computed partial charges toward basis set and (exchange-)correlation treatment

Partial charges are a central concept in general chemistry and chemical biology, yet dozens of different computational definitions exist. In prior work [Cho et al., ChemPhysChem 21, 688-696 (2020)], we showed that these can be reduced to at most three 'principal components of ionicity'. The present study addressed the dependence of computed partial charges q on 1-particle basis set and (for WFT methods) n -particle correlation treatment or (for DFT methods) exchange-correlation functional, for several representative partial charge definitions such as QTAIM, Hirshfeld, Hirshfeld-I, HLY (electrostatic), NPA, and GAPT. Our findings show that semi-empirical double hybrids can closely approach the CCSD(T) 'gold standard' for this property. In fact, owing to an error compensation in MP2, CCSD partial charges are further away from CCSD(T) than is MP2. The nonlocal correlation is important, especially when there is a substantial amount of nonlocal exchange. Employing range separation proves to be "mostly" not advantageous, while global hybrids perform optimally for 20%-30% Hartree-Fock exchange across all charge types. Basis set convergence analysis shows that an augmented triple-zeta heavy-aug-cc-pV(T+d)Z basis set or a partially augmented jun-cc-pV(T+d)Z basis set is sufficient for Hirshfeld, Hirshfeld-I, HLY, and GAPT charges. In contrast, QTAIM and NPA display slower basis set convergence. It is noteworthy that for both NPA and QTAIM, HF exhibits markedly slower basis set convergence than the correlation components of MP2 and CCSD. Triples corrections in CCSD(T), denoted as CCSD(T)-CCSD, exhibit even faster basis set convergence.

Stepping-Stone CBH: Benchmark and Application of a Multilayered Isodesmic-Based Correction Scheme

We present a generalization of the connectivity-based hierarchy (CBH) of isodesmic-based correction schemes to a multilayered fragmentation platform for overall cost reduction while retaining high accuracy. The newly developed multilayered CBH approach, called stepping-stone CBH (SSCBH), is benchmarked on a diverse set of 959 medium-sized organic molecules. Applying SSCBH corrections to the PBEh-D3 density functional resulted in an average error of 0.76 kcal/mol for the full test set compared to accurate CCSD(T)-quality enthalpies and an even lower error of 0.44 kcal/mol on a subset containing only acyclic molecules. These results rival the traditional CBH-3 approach at a greatly reduced cost, allowing larger fragment corrections to be made at the MP2 level of theory rather than with G4. Our SSCBH approach will enable more widespread applications of CBH methods to a broader range of organic and biomolecular systems.

Basis Set Requirements of σ-Functionals for Gaussian- and Slater-Type Basis Functions and Comparison with Range-Separated Hybrid and Double Hybrid Functionals

σ-Functionals belong to the class of Kohn-Sham (KS) correlation functionals based on the adiabatic-connection fluctuation-dissipation theorem and are technically closely related to the random phase approximation (RPA). They have the same computational demand as the latter, with the computational effort of an energy evaluation for both methods being lower than that of a preceding hybrid DFT calculation for typical systems but yield much higher accuracy, reaching chemical accuracy of 1 kcal/mol for quantities such as reactions and transition energies in main group chemistry. In previous work on σ-functionals, rather large Gaussian basis sets have been used. Here, we investigate the actual basis set requirements of σ-functionals and present three setups that employ smaller Gaussian basis sets ranging from quadruple-ζ (QZ) to triple-ζ (TZ) quality and represent a good compromise between accuracy and computational efficiency. Furthermore, we introduce an implementation of σ-functionals based on Slater-type basis sets and present two setups of QZ and TZ quality for this implementation. We test the accuracy of these setups on a large database of various physical properties and types of reactions, as well as equilibrium geometries and vibrational frequencies. As expected, the accuracy of σ-functional calculations becomes somewhat lower with a decreasing basis set size. However, for all setups considered here, calculations with σ-functionals are clearly more accurate than those within the RPA and even more so than those of the conventional KS methods. For the smallest setup using Gaussian-type basis functions and Slater-type basis functions, we introduce a reparametrization that reduces the loss in accuracy due to the basis set error to some extent. A comparison with the range-separated hybrid ωB97X-V and the double hybrid DSD-BLYP-D3 shows that σ functionals outperform in accuracy both of these accurate and, for their class, representative functionals.

W4Λ: Leveraging Λ Coupled-Cluster for Accurate Computational Thermochemistry Approaches

High-accuracy composite wave function methods like Weizmann-4 (W4) theory, high-accuracy extrapolated ab initio thermochemistry (HEAT), and the Feller-Peterson-Dixon (FPD) approach enable sub-kJ/mol accuracy in gas-phase thermochemical properties. Their biggest computational bottleneck is the evaluation of the valence post-CCSD(T) correction term. We demonstrate here, for the W4-17 thermochemistry benchmark and subsets thereof, that the Λ coupled-cluster expansion converges more rapidly and smoothly than the regular coupled-cluster series. By means of CCSDT(Q)Λ and CCSDTQ(5)Λ, we can considerably (up to an order of magnitude) accelerate W4- and W4.3-type calculations without loss in accuracy, leading to the W4Λ and W4.3Λ computational thermochemistry protocols.

Quantum tunneling instability of the mythical hexazine and pentazine

Through computational analysis we found that pentazine and hexazine, two hypothetical high-energy density materials, exhibit inherent instability due to quantum tunneling effects. This instability remains even near the absolute zero, and therefore they can be deemed as unsynthesizable. We propose substituents that could potentially stabilize pentazine, especially dimethylamine.

Highly accurate σ- and τ-functionals for beyond-RPA methods with approximate exchange kernels

σ-Functionals are promising new developments for the Kohn-Sham correlation energy based upon the direct Random Phase Approximation (dRPA) within the adiabatic connection formalism, providing impressive improvements over dRPA for a broad range of benchmarks. However, σ-functionals exhibit a high amount of self-interaction inherited from the approximations made within dRPA. Inclusion of an exchange kernel in deriving the coupling-strength-dependent density-density response function leads to so-called τ-functionals, which - apart from a fourth-order Taylor series expansion - have only been realized in an approximate fashion so far to the best of our knowledge, most notably in the form of scaled σ-functionals. In this work, we derive, optimize, and benchmark three types of σ- and τ-functionals including approximate exchange effects in the form of an antisymmetrized Hartree kernel. These functionals, based on a second-order screened exchange type contribution in the adiabatic connection formalism, the electron-hole time-dependent Hartree-Fock kernel (eh-TDHF) otherwise known as RPA with exchange (RPAx), and an approximation thereof known as approximate exchange kernel (AXK), are optimized on the ASCDB database using two new parametrizations named A1 and A2. In addition, we report a first full evaluation of σ- and τ-functionals on the GMTKN55 database, revealing our exchange-including functionals to considerably outperform existing σ-functionals while being highly competitive with some of the best double-hybrid functionals of the original GMTKN55 publication. In particular, the σ-functionals based on AXK and τ-functionals based on RPAx with PBE0 reference stand out as highly accurate approaches for a wide variety of chemically relevant problems.

Corrected density functional theory and the random phase approximation: Improved accuracy at little extra cost

We recently introduced an efficient methodology to perform density-corrected Hartree-Fock density functional theory [DC(HF)-DFT] calculations and an extension to it we called "corrected" HF DFT [C(HF)-DFT] [Graf and Thom, J. Chem. Theory Comput. 19 5427-5438 (2023)]. In this work, we take a further step and combine C(HF)-DFT, augmented with a straightforward orbital energy correction, with the random phase approximation (RPA). We refer to the resulting methodology as corrected HF RPA [C(HF)-RPA]. We evaluate the proposed methodology across various RPA methods: direct RPA (dRPA), RPA with an approximate exchange kernel, and RPA with second-order screened exchange. C(HF)-dRPA demonstrates very promising performance; for RPA with exchange methods, on the other hand, we often find over-corrections.

A larger basis set describes atomization energy core-valence correction better than a higher-order coupled-cluster method

The accuracy of coupled-cluster methods for the computation of core-valence correction to atomization energy was assessed. Truncation levels up to CCSDTQP were considered together with (aug-)cc-pwCVnZ (n = D, T, Q, 5) basis sets and three different extrapolation techniques (canonical and flexible Helgaker formula and Riemann zeta function extrapolation). With the exception of CCSD, a more accurate correction can be obtained from a larger basis set using a lower-level coupled-cluster method, and not vice versa. For the CCSD(T) level, it also implies faster computations with modern codes. We also discussed the importance of moving to higher-order or all-electron methods for geometry optimizations. The present study provides the general knowledge needed for the most accurate state-of-the-art computations.

Automatic Generation of Accurate and Cost-Efficient Auxiliary Basis Sets

We have recently discussed an algorithm to automatically generate auxiliary basis sets (ABSs) of the standard form for density fitting (DF) or resolution-of-the-identity (RI) calculations in a given atomic orbital basis set (OBS) of any form, such as Gaussian-type orbitals, Slater-type orbitals, or numerical atomic orbitals [J. Chem. Theory Comput. 2021, 17, 6886]. In this work, we study two ways to reduce the cost of such automatically generated ABSs without sacrificing their accuracy. We contract the ABS with a singular value decomposition proposed by Kállay [J. Chem. Phys. 2014, 141, 244113], used here in a somewhat different setting. We also drop the high-angular momentum functions from the ABS, as they are unnecessary for global fitting methods. Studying the effect of these two types of truncations on Hartree-Fock (HF) and second-order Møller-Plesset perturbation theory (MP2) calculations on a chemically diverse set of first- and second-row molecules within the RI/DF approach, we show that accurate total and atomization energies can be achieved by a combination of the two approaches with significant reductions in the size of the ABS. While the original approach yields ABSs whose number of functions NbfABS scales with the number of functions in the OBS, NOBSbf, as NABSbf = γNOBSbf with the prefactor γ ≈ O ( 10 ) , the reduction schemes of this work afford results of essentially the same quality as the original unpruned and uncontracted ABS with γ ≈ 5-6, while an accuracy that may suffice for routine applications is achievable with a further reduced ABS with γ ≈ 3-4. The observed errors are similar at HF and MP2 levels of theory, suggesting that the generated ABSs are highly transferable and can also be applied to model challenging properties with high-level methods.

6-Endo-dig versus 5-exo-dig: Exploring Radical Cyclization Preference with First-, Second-, and Third-row Linkers using High-level Quantum Chemical Methods

As an expansion upon Baldwin rules, the cyclization reactions of hex-5-yn-1-yl radical systems with different first-, second-, and third-row linkers are explored at the CCSD(T) level via means of the SMD(benzene)-G4(MP2) thermochemical protocol. Unlike C, O, and N linkers, systems with B, Si, P, S, Ge, As, and Se linkers are shown to favor 6-endo-dig cyclization. This offers fundamental insights into the rational synthetic design of cyclic compounds. A thorough analysis of stereoelectronic effects, cyclization barriers, and intrinsic barriers illustrates that structural changes alter the cyclization preference by mainly impacting 5-exo-dig reaction barriers. Based on the high-level computational modeling, we proceed to develop a new tool for cyclization preference prediction from the correlation between cyclization barriers and radical structural parameters (e. g., linker bond length and bond angle). A strong correlation is found between the radical attack trajectory angle and the reaction barrier heights, i. e., cyclization preference. Finally, the influence of stereoelectronic effects on the two radical cyclization pathways is further investigated in stereoisomers of hypervalent silicon system, which provides novel insight into cyclization control.

A Systematic Exploration of B-F Bond Dissociation Enthalpies of Fluoroborane-Type Molecules at the CCSD(T)/CBS Level

Fluoroborane-type molecules (R1R2B-F) are of interest in synthetic chemistry, but to date, apart from a handful of small species (such as H2BF, HBF2, and BF3), little is known concerning the effect of substituents in governing the strength of the B-F bonds of such species toward homolytic dissociation in the gas phase. In this study, we have calculated the bond dissociation enthalpies (BDEs) of thirty unique B-F bonds at the CCSD(T)/CBS level using the high-level W1w thermochemical protocol. The B-F bonds in all species considered are very strong, ranging from 545.9 kJ mol-1 in (H2B)2B-F to 729.2 kJ mol-1 HBF2. Nevertheless, these BDEs still vary over a wide range of 183.3 kJ mol-1. The structural properties that affect the BDEs are examined in detail, and the homolytic BDEs are rationalized based on molecule stabilization enthalpies and radical stabilization enthalpies. Since polar B-F bonds may represent a challenging test case for density functional theory (DFT) methods, we proceed to examine the performance of a wide range of DFT methods across the rungs of Jacob's Ladder for their ability to compute B-F BDEs. We find that only a handful of DFT methods can reproduce the CCSD(T)/CBS BDEs with mean absolute deviations (MADs) below the threshold of chemical accuracy (i.e., with average deviations below 4.2 kJ mol-1). The only functionals capable of achieving this feat were (MADs given in parentheses): ωB97M-V (4.0), BMK (3.5), DSD-BLYP (3.8), and DSD-PBEB95 (1.8 kJ mol-1).

Intramolecular Proton-Coupled Hydride Transfers with Relatively Low Activation Barriers

We report that bifunctional molecules containing hydroxyl and carbonyl functional groups can undergo an effective transfer hydrogenation via an intramolecular proton-coupled hydride transfer (PCHT) mechanism. In this reaction mechanism, a hydride transfer between two carbon atoms is coupled with a proton transfer between two oxygen atoms via a cyclic bond rearrangement transition structure. The coupled transfer of the two hydrogens as H^δ+ and H^δ- is supported by atomic polar tensor charges. The activation energy for the PCHT reaction is strongly dependent on the length of the alkyl chain between the hydroxyl and carbonyl functional groups but relatively weakly dependent on the functional groups attached to the hydroxyl and carbonyl carbons. We investigate the PCHT reaction mechanism using the Gaussian-4 thermochemical protocol and obtain high activation energy barriers (ΔH^‡₂₉₈) of 210.5-228.3 kJ mol^-1 for chain lengths of one carbon atom and 160.2-163.9 kJ mol^-1 for chain lengths of two carbon atoms. However, for longer chain lengths containing 3-4 carbon atoms, we obtain ΔH^‡₂₉₈ values as low as 101.9 kJ mol^-1. Importantly, the hydride transfer between two carbon atoms proceeds without the need for a catalyst or hydride transfer activating agent. These results indicate that the intramolecular PCHT reaction provides an effective avenue for uncatalyzed, metal-free hydride transfers at ambient temperatures.

Quantum Alchemy Based Bonding Trends and Their Link to Hammett's Equation and Pauling's Electronegativity Model

We present an intuitive and general analytical approximation estimating the energy of covalent single and double bonds between participating atoms in terms of their respective nuclear charges with just three parameters, [E_AB ≈ a - bZ_AZ_B + c(Z_A^7/3 + Z_B^7/3) ]. The functional form of our expression models an alchemical atomic energy decomposition between participating atoms A and B. After calibration, reasonably accurate bond dissociation energy estimates are obtained for hydrogen-saturated diatomics composed of p-block elements coming from the same row 2 ≤ n ≤ 4 in the periodic table. Corresponding changes in bond dissociation energies due to substitution of atom B by C can be obtained via simple formulas. While being of different functional form and origin, our model is as simple and accurate as Pauling's well-known electronegativity model. Analysis indicates that the model's response in covalent bonding to variation in nuclear charge is near-linear, which is consistent with Hammett's equation.

Outstanding improvement in removing the delocalization error by global natural orbital functional

This work assesses the performance of the recently proposed global natural orbital functional (GNOF) against the charge delocalization error. GNOF provides a good balance between static and dynamic electronic correlations leading to accurate total energies while preserving spin, even for systems with a highly multi-configurational character. Several analyses were applied to the functional, namely, (i) how the charge is distributed in super-systems of two fragments, (ii) the stability of ionization potentials while increasing the system size, and (iii) potential energy curves of a neutral and charged diatomic system. GNOF was found to practically eliminate the charge delocalization error in many of the studied systems or greatly improve the results obtained previously with PNOF7.

An Imbalance in the Force: The Need for Standardized Benchmarks for Molecular Simulation

Force fields (FFs) for molecular simulation have been under development for more than half a century. As with any predictive model, rigorous testing and comparisons of models critically depends on the availability of standardized data sets and benchmarks. While such benchmarks are rather common in the fields of quantum chemistry, this is not the case for empirical FFs. That is, few benchmarks are reused to evaluate FFs, and development teams rather use their own training and test sets. Here we present an overview of currently available tests and benchmarks for computational chemistry, focusing on organic compounds, including halogens and common ions, as FFs for these are the most common ones. We argue that many of the benchmark data sets from quantum chemistry can in fact be reused for evaluating FFs, but new gas phase data is still needed for compounds containing phosphorus and sulfur in different valence states. In addition, more nonequilibrium interaction energies and forces, as well as molecular properties such as electrostatic potentials around compounds, would be beneficial. For the condensed phases there is a large body of experimental data available, and tools to utilize these data in an automated fashion are under development. If FF developers, as well as researchers in artificial intelligence, would adopt a number of these data sets, it would become easier to compare the relative strengths and weaknesses of different models and to, eventually, restore the balance in the force.

Sensitivity of coupled cluster electronic properties on the reference determinant: Can Kohn-Sham orbitals be more beneficial than Hartree-Fock orbitals?

Coupled cluster calculations are traditionally performed over Hartree-Fock reference orbitals (HF-CC methodology). However, in the literature it has been repeatedly raised whether the use of a Kohn-Sham reference (KS-CC methodology) might result in improved performance relative to HF-CC. In the present study, we re-examine the relation of HF-CC and KS-CC methods by comparing the results of widely applied truncated CC calculations (CCSD, CCSD(T), CCSDT) to the limit of full configuration interaction (FCI), which serves as an undebatable reference point of accuracy. Based on a series of CC calculations on diatoms and transition metal complexes, we demonstrate that no systematic improvement of coupled cluster electronic energies, densities and chemical reaction energies is expected when changing from HF to a KS reference. Nevertheless, fortuitous error cancellations might occasionally result in illusory improvement compared to HF-CC. Altogether, the application of KS-CC is not advantageous over HF-CC, but it is also not unreasonable as the choice of reference has negligible influence on the results at sufficiently high CC levels. In addition, KS-CC can be a particularly useful alternative if difficulties are encountered in HF or HF-CC convergence. It is also notable that KS-CC results are found to be practically independent of the chosen density functional, which implies that almost any KS-CC method can be used in place of HF-CC.

Scaled σ-functionals for the Kohn-Sham correlation energy with scaling functions from the homogeneous electron gas

The recently introduced σ-functionals constitute a new type of functionals for the Kohn-Sham (KS) correlation energy. σ-Functionals are based on the adiabatic-connection fluctuation-dissipation theorem, are computationally closely related to the well-known direct random phase approximation (dRPA), and are formally rooted in many-body perturbation theory along the adiabatic connection. In σ-functionals, the function of the eigenvalues σ of the Kohn-Sham response matrix that enters the coupling constant and frequency integration in the dRPA is replaced by another function optimized with the help of reference sets of atomization, reaction, transition state, and non-covalent interaction energies. σ-Functionals are highly accurate and yield chemical accuracy of 1 kcal/mol in reaction or transition state energies, in main group chemistry. A shortcoming of σ-functionals is their inability to accurately describe processes involving a change of the electron number, such as ionizations or electron attachments. This problem is attributed to unphysical self-interactions caused by the neglect of the exchange kernel in the dRPA and σ-functionals. Here, we tackle this problem by introducing a frequency- and σ-dependent scaling of the eigenvalues σ of the KS response function that models the effect of the exchange kernel. The scaling factors are determined with the help of the homogeneous electron gas. The resulting scaled σ-functionals retain the accuracy of their unscaled parent functionals but in addition yield very accurate ionization potentials and electron affinities. Moreover, atomization and total energies are found to be exceptionally accurate. Scaled σ-functionals are computationally highly efficient like their unscaled counterparts.

del Campo J. Charge delocalization error in Piris Natural Orbital Functionals

Piris Natural Orbital Functionals (PNOF) have been recognized as a low-scaling alternative to study strong correlated systems. In this work, we address the performance of the fifth functional (PNOF5) and the seventh functional (PNOF7) to deal with another common problem, the charge delocalization error. The effects of this problem can be observed in charged systems of repeated well-separated fragments, where the energy should be the sum of the charged and neutral fragments, regardless of how the charge is distributed. In practice, an energetic overstabilization of fractional charged fragments leads to a preference for having the charge delocalized throughout the system. To establish the performance of PNOF functionals regarding charge delocalization error, charged chains of helium atoms and the W4-17-MR set molecules were used as base fragments and their energy, charge distribution and correlation regime were studied. It was found that PNOF5 prefers localized charge distributions, while PNOF7 improves the treatment of interpair static correlation and tends to the correct energetic limit for several cases, although a preference for delocalized charge distributions may arise in highly strong correlation regimes. Overall, it is concluded that PNOF functionals can simultaneously deal with static correlation and charge delocalization errors, resulting in a promising choice to study charge-related problems.

Uniting Nonempirical and Empirical Density Functional Approximation Strategies Using Constraint-Based Regularization

In this work, we present a general framework that unites the two primary strategies for constructing density functional approximations (DFAs): nonempirical (NE) constraint satisfaction and empirical (E) data-driven optimization. The proposed method employs B-splines, bell-shaped spline functions with compact support, to construct each inhomogeneity correction factor (ICF). This choice offers several distinct advantages over traditional polynomial expansions by enabling explicit enforcement of linear and nonlinear constraints as well as ICF smoothness using Tikhonov and penalized B-splines (P-splines) regularization. As proof-of-concept, we use the so-called CASE (constrained and smoothed empirical) framework to construct a constraint-satisfying and data-driven global hybrid that exhibits enhanced performance across a diverse set of chemical properties. We argue that the CASE approach can be used to generate DFAs that maintain the physical rigor and transferability of NE-DFAs while leveraging high-quality quantum-mechanical data to remove the arbitrariness of ansatz selection and improve performance.

High-Level Quantum Chemistry Reference Heats of Formation for a Large Set of C, H, N, and O Species in the NIST Chemistry Webbook and the Identification and Validation of Reliable Protocols for Their Rapid Computation

A recent study has examined the accuracy of NIST heats of formation for a set of C, H, and O-containing species with a proposed low-cost quantum chemistry approach. In the present study, we have used high-level methods such as W1X-2 to obtain these data more rigorously, which we have then used to assess the NIST and the previously computed values. We find that many of these NIST data that are as suggested to be unreliable by the previous study are indeed inconsistent with our high-level reference values. However, we also find substantial deviations for the previously computed values from our benchmark. Thus, we have assessed the performance of alternative low-cost methods. In our assessment, we have additionally examined C, H, N, and O-containing species for which heats of formation are available from the NIST database. We find the ωB97M-V/ma-def2-TZVP, DSD-PBEP86/ma-def2-TZVP, and CCSD(T)-F12b/aug'-cc-pVDZ methods to be adequate for obtaining heats of formation with the atomization approach, once their atomic energies are optimized with our benchmark. Notably, the low-cost ωB97M-V method yields values that agree to be within 10 kJ mol^-1 for more than 90% of the (∼1500) species. A higher 20 kJ mol^-1 threshold captures 98% of the data. The outlier species typically contain many electron-withdrawing (nitro) groups. In these cases, the use of isodesmic-type reactions rather than the atomization approach is more reliable. Our assessment has also identified significant outliers from the NIST database, for which experimental re-determination of the heats of formation would be desirable.

MP2-F12 Basis Set Convergence near the Complete Basis Set Limit: Are h Functions Sufficient?

We have investigated the title question for the W4-08 thermochemical benchmark using l-saturated truncations of a large reference (REF) basis set, as well as for standard F12-optimized basis sets. With the REF basis set, the root-mean-square (RMS) contribution of i functions to the MP2-F12 total atomization energies (TAEs) is about 0.01 kcal/mol, the largest individual contributions being 0.04 kcal/mol for P₂ and P₄. However, even for these cases, basis set extrapolation from {g,h} basis sets adequately addresses the problem. Using basis sets insufficiently saturated in the spdfgh angular momenta may lead to exaggerated i function contributions. For extrapolation from spdfg and spdfgh basis sets, basis set convergence appears to be quite close to the theoretical asymptotic ∝ L^–7 behavior. We hence conclude that h functions are sufficient even for highly demanding F12 applications. With one-parameter extrapolation, spdf and spdfg basis sets are adequate, aug-cc-pV{T,Q}Z-F12 yielding a RMSD = 0.03 kcal/mol. A limited exploration of CCSD(F12*) and CCSD-F12b suggests our conclusions are applicable to higher-level F12 methods as well.

ATOMIC-2 Protocol for Thermochemistry

ATOMIC is a midlevel thermochemistry protocol that uses Pople's concept of bond separation reactions (BSRs) as a theoretical framework to reduce computational demands in the evaluation of atomization energies and enthalpies of formation. Various composite models are available that approximate bond separation energies at the complete-basis-set limit of all-electron CCSD(T), each balancing computational cost with achievable accuracy. Evaluated energies are then combined with very high-level, precomputed atomization energies of all auxiliary molecules appearing in the BSR to obtain the atomization energy of the molecule under study. ATOMIC-2 is a new version of the protocol that retains the overall concept and all previously defined composite models but improves on ATOMIC-1 in various other ways: Geometry optimization and zero-point-energy evaluation are performed at the density functional level (PBE0-D3/6-311G(d)), which shows significant computational savings and better accuracy than the previously employed RI-MP2/cc-pVTZ. The BSR framework is improved, using more accurate complete-basis-set (CBS) extrapolations toward the Full CI limit for the atomization energies of all auxiliary molecules. Finally, and most importantly, an error and uncertainty model termed ATOMIC-2_um is added that estimates average bias and uncertainty for each of the atomization energy contributions that arise from the simplified treatment of some contributions to bond separation energies (CCSD(T)) and the neglect of others (such as higher order, scalar relativistic, or diagonal Born-Oppenheimer corrections) or from residual error in the energies of auxiliary molecules. Large and diverse benchmarks including up to 1179 molecules are used to evaluate necessary reference data and to correlate the observed error for each of the contributions with appropriate proxies that are available without additional quantum-chemical calculations for a particular molecule and represent its size and type. The implementation of ATOMIC-2 considers neutral, closed-shell molecules containing H, C, N, O, and F atoms; compared to ATOMIC-1, the framework has been extended to cover a few challenging but rare bond topologies. In comparison to highly accurate reference data for 184 molecules taken from the ATcT database (V. 1.122r), regular ATOMIC-2 shows noticeable underbinding, but the bias-corrected protocol ATOMIC-2_um is found to be more accurate than either ATOMIC-1 or standard Gaussian-4 theory, and the uncertainty model is consistent with statistics of actually observed errors. Problems arising from ambiguous or challenging Lewis-valence structures defining BSRs are discussed, and computational efficiency is demonstrated. Computer code is made available to perform ATOMIC-2_um analyses.

Tackling an accurate description of molecular reactivity with double-hybrid density functionals

In this Communication, we assess a panel of 18 double-hybrid density functionals for the modeling of the thermochemical and kinetic properties of an extended dataset of 449 organic chemistry reactions belonging to the BH9 database. We show that most of DHs provide a statistically robust performance to model barrier height and reaction energies in reaching the “chemical accuracy.” In particular, we show that nonempirical DHs, such as PBE0-DH and PBE-QIDH, or minimally parameterized alternatives, such as ωB2PLYP and B2K-PLYP, succeed to accurately model both properties in a balanced fashion. We demonstrate, however, that parameterized approaches, such as ωB97X-2 or DSD-like DHs, are more biased to only one of both properties.

Small-Basis Set Density-Functional Theory Methods Corrected with Atom-Centered Potentials

Density functional theory (DFT) is currently the most popular method for modeling noncovalent interactions and thermochemistry. The accurate calculation of noncovalent interaction energies, reaction energies, and barrier heights requires choosing an appropriate functional and, typically, a relatively large basis set. Deficiencies of the density-functional approximation and the use of a limited basis set are the leading sources of error in the calculation of noncovalent and thermochemical properties in molecular systems. In this article, we present three new DFT methods based on the BLYP, M06-2X, and CAM-B3LYP functionals in combination with the 6-31G* basis set and corrected with atom-centered potentials (ACPs). ACPs are one-electron potentials that have the same form as effective-core potentials, except they do not replace any electrons. The ACPs developed in this work are used to generate energy corrections to the underlying DFT/basis-set method such that the errors in predicted chemical properties are minimized while maintaining the low computational cost of the parent methods. ACPs were developed for the elements H, B, C, N, O, F, Si, P, S, and Cl. The ACP parameters were determined using an extensive training set of 118655 data points, mostly of complete basis set coupled-cluster level quality. The target molecular properties for the ACP-corrected methods include noncovalent interaction energies, molecular conformational energies, reaction energies, barrier heights, and bond separation energies. The ACPs were tested first on the training set and then on a validation set of 42567 additional data points. We show that the ACP-corrected methods can predict the target molecular properties with accuracy close to complete basis set wavefunction theory methods, but at a computational cost of double-ζ DFT methods. This makes the new BLYP/6-31G*-ACP, M06-2X/6-31G*-ACP, and CAM-B3LYP/6-31G*-ACP methods uniquely suited to the calculation of noncovalent, thermochemical, and kinetic properties in large molecular systems.