arXiv subjects

Jan M. L. Martin

Publications and source records attributed to Jan M. L. Martin.

At least 55 records · Page 3Linked to original sources

arXiv2021

Exploring Avenues Beyond Revised DSD Functionals: II. Random-Phase Approximation and scaled MP3 corrections

For revDSD double hybrids, the Görling-Levy second-order perturbation theory component is an Achilles' Heel when applied to systems with significant near-degeneracy ("static") correlation. We have explored its replacement by the direct random phase approximation (dRPA), inspired by the SCS-dRPA75 functional of Kállay and coworkers. The addition to the final energy of both a D4 empirical dispersion correction, and of a semilocal correlation component lead, to significant improvements, with DSD-PBEdRPA75-D4 approaching the performance of revDSD-PBEP86-D4 and the Berkeley $ω$B97M(2). This form appears to be fairly insensitive to the choice of semilocal functional, but does exhibit stronger basis set sensitivity than the PT2-based double hybrids (due to much larger prefactors for the nonlocal correlation). As an alternative, we explored adding an MP3-like correction term (in a medium-sized basis sets) to a range-separated $ω$DSD-PBEP86-D4 double hybrid, and found it to have significantly lower WTMAD2 (weighted mean absolute deviation) for the large and chemically diverse GMTKN55 benchmark suite; the added computational cost can be mitigated through density fitting techniques.

physics.chem-ph↗

arXiv2021

Prototypical pi-pi dimers re-examined by means of high-level CCSDT(Q) composite ab inito methods

The benzene...ethene and parallel-displaced (PD) benzene...benzene dimers are the most fundamental systems involving p-p stacking interactions. Several high-level ab initio investigations calculated the binding energies of these dimers at the CCSD(T)/CBS level of theory using various approaches such as reduced virtual orbital spaces and/or MP2-based basis set corrections. Here we obtain CCSDT(Q) binding energies using a Weizmann-3-type approach. In particular, we extrapolate the SCF, CCSD, and (T) components using large heavy-atom augmented Gaussian basis sets (namely, SCF/jul-cc-pV{5,6}Z, CCSD/jul-cc-pV{Q,5}Z, and (T)/jul-cc-pV{T,Q}Z). We consider post-CCSD(T) contributions up to CCSDT(Q), inner-shell, scalar-relativistic, and Born-Oppenheimer corrections. Overall, our best relativistic, all-electron CCSDT(Q) binding energies are Delta Ee,all,rel = 1.234 (benzene...ethene) and 2.550 (benzene...benzene PD), Delta H0 = 0.949 (benzene...ethene) and 2.310 (benzene...benzene PD), and Delta H298 = 0.130 (benzene...ethene) and 1.461 (benzene...benzene PD) kcal/mol. Important conclusions are reached regarding the basis set convergence of the SCF, CCSD, (T), and post-CCSD(T) components. Explicitly correlated calculations are used as a sanity check on the conventional binding energies. Overall, post-CCSD(T) contributions are destabilizing by 0.028 (benzene...ethene) and 0.058(benzene...benzene) kcal/mol, thus they cannot be neglected if 0.1 kcal/mol accuracy is sought.

physics.chem-ph↗

arXiv2021

What types of chemical problems benefit from density-corrected DFT? A probe using an extensive and chemically diverse test suite

For the large and chemically diverse GMTKN55 benchmark suite, we have studied the performance of density-corrected density functional theory (HF-DFT), compared to self-consistent DFT, for several pure and hybrid GGA and meta-GGA exchange-correlation (XC) functionals (PBE, BLYP, TPSS, SCAN) as a function of the percentage of HF exchange in the hybrid. The D4 empirical dispersion correction has been added throughout. For subsets dominated by dynamical correlation -- particularly noncovalent interaction subsets -- HF-DFT is highly beneficial, particularly at low HF exchange percentages. For subsets with significant static correlation (i.e., where a Hartree-Fock determinant is not a good zero-order wavefunction), HF-DFT may do more harm than good. While the self-consistent series show optima at or near 37.5% (i.e., 3/8) for all four XC functionals -- consistent with Grimme's proposal of the PBE38 functional -- HF-BnLYP-D4, HF-PBEn-D4, and HF-TPSSn-D4 all exhibit minima nearer 25% (i.e., 1/4). Intriguingly, for HF-SCANn-D4, the minimum is near 10%, but the weighted mean absolute error (WTMAD2) for GMTKN55 is only barely lower than that of HF-SCAN-D4 (i.e., where the post-HF step is a pure meta-GGA). The latter becomes an attractive option, only slightly more costly than pure Hartree-Fock, and devoid of adjustable parameters other than the three in the dispersion correction. Moreover, its WTMAD2 is only surpassed by the highly empirical M06-2X and by the combinatorically optimized empirical range-separated hybrids wB97X-V and wB97M-V.

physics.chem-ph↗

arXiv2020

Canonical and DLPNO-based composite wavefunction methods parametrized against large and chemically diverse training sets. 2. Correlation consistent basis sets, core-valence correlation, and F12 alternatives

A hierarchy of wavefunction composite methods (cWFT), based on G4- type cWFT methods available for elements H through Rn, was recently reported by Semidalas and Martin [J. Chem. Theor. Comput. 2020, 16, 4238]. We extend this hierarchy by considering the inner-shell correlation energy in the second-order Moller-Plesset correction and replacing the Weigend-Ahlrichs def2-mZVPP(D) basis sets used in the aforementioned paper with complete basis set extrapolation from augmented correlation consistent core-valence triple-zeta, aug-cc-pwCVTZ(-PP), and quadruple-zeta, aug-cc-pwCVQZ(-PP), basis sets, thus creating cc-G4- type methods. For the large and chemically diverse GMTKN55 benchmark suite, they represent a substantial further improvement and bring WTMAD2 (weighted mean absolute deviation) down below 1 kcal/mol. Intriguingly, the lion's share of the improvement comes from better capture of valence correlation; the inclusion of core-valence correlation is almost an order of magnitude less important. These robust correlation consistent cWFT methods approach the CCSD(T) complete basis limit with just one or a few fitted parameters. Particularly the DLPNO variants such as cc-G4-T-DLPNO are applicable to fairly large molecules at modest computational cost, as is (for a reduced range of elements) a different variant using MP2-F12/cc-pVTZ-F12 for the MP2 component.

physics.chem-ph↗

arXiv2020

Energetics of (H2O)20 Isomers by means of F12 Canonical and Localized Coupled Cluster Methods

We consider the performance of combined PNO-F12 approaches for the dissociation energy of water clusters as large as (H2O)20 by comparison to canonical CCSD(T)/CBS reference values obtained through n-body decomposition of post-MP2 corrections. We find that PNO-LCCSD(T)-F12b approaches with "Tight" cutoffs are generally capable of reproducing canonical CCSD(T) interaction energies to within ~0.25% and isomerization energies to ~1.5%, while requiring only a fraction of the canonical computational cost. However, basis set convergence patterns and effect of counterpoise corrections are more erratic than for canonical calculations, highlighting the need for canonical benchmarks on closely related systems.

physics.chem-ph↗

arXiv2020

Canonical and DLPNO-based G4(MP2)XK-inspired composite wavefunction methods parametrized against large and chemically diverse training sets: Are they more accurate and/or robust than double hybrid DFT?

The large and chemically diverse GMTKN55 benchmark was used as a training set for parametrizing composite wave function thermochemistry protocols akin to G4(MP2)XK theory (Chan et al, JCTC 2019, 15, 4478-4484). Even after reparametrization, the GMTKN55 WTMAD2 (weighted mean absolute deviation, type 2) for G4(MP2)-XK is actually inferior to that of the best rung-4 DFT functional, wB97M-V. By increasing the basis set for the MP2 part to def2-QZVPPD, we were able to substantially improve performance at modest cost (if an RI-MP2 approximation is made), with WTMAD2 for this G4(MP2)-XK-D method now comparable to the better rung-5 functionals (albeit at greater cost). A three-tier approach with a scaled MP3/def2-TZVPP intermediate step, however, leads to a G4(MP3)-D method that is markedly superior to even the best double hybrids wB97M(2) and revDSD-PBEP86-D4. Evaluating the CCSD(T) component with a triple-zeta, rather than split-valence, basis set yields only a modest further improvement that is incommensurate with the drastic increase in computational cost. G4(MP3)-D and G4(MP2)- XK-D have about 40% better WTMAD2, at similar or lower computational cost, than their counterparts G4 and G4(MP2), respectively: detailed comparison reveals that the difference lies in larger molecules due to basis set incompleteness error. An E2/ {T,Q} extrapolation and a CCSD(T)/def2-TZVP step provided the G4-T method of high accuracy and with just three fitted parameters. Using KS orbitals in MP2 leads to the G4(MP3|KS)-D method, which entirely eliminates the CCSD(T) step and has no steps costlier than scaled MP3; this shows a path forward to further improvements in double-hybrid density functional methods. G4-T-DLPNO, a variant in which post-MP2 corrections are evaluated at the DLPNO- CCSD(T) level, achieves nearly the accuracy of G4-T but is applicable to much larger systems.

physics.chem-ph↗

arXiv2020

Performance of Localized Coupled Cluster Methods in a Moderately Strong Correlation Regime: Hückel-Möbius Interconversions in Expanded Porphyrins

Localized orbital coupled cluster theory has recently emerged as an nonempirical alternative to DFT for large systems. Intuitively, one might expect such methods to perform less well for highly delocalized systems. In the present work, we apply both canonical CCSD(T) and a variety of localized approximations thereto to a set of expanded porphyrins -- systems that can switch between Hückel, figure-eight, and Möbius topologies under external stimuli. Both minima and isomerization transition states are considered. We find that Möbius(-like) structures have much stronger static correlation character than the remaining structures, and that this causes significant errors in DLPNO-CCSD(T) and even DLPNO-CCSD(T1) approaches, unless TightPNO cutoffs are employed. If sub-kcal/mol reproduction of canonical relative energies is required even for Möbius-type systems (or other systems plagued by strong static correlation), then Nagy and Kallay's LNO-CCSD(T) method with "tight" settings can provide that, at much greater computational expense than either the PNO-LCCSD(T) or DLPNO-LCCSD(T) approaches but with still a much gentler CPU time scaling than canonical approaches. We would propose the present POLYPYR21 dataset as a benchmark for localized orbital methods, or more broadly, for the ability of lower-level methods to handle energetics with strongly varying degrees of static correlation.

physics.chem-ph↗

arXiv2019

Empirical Double-Hybrid Density Functional Theory: A 'Third Way' In Between WFT and DFT

Double hybrid density functional theory arguably sits on the seamline between wavefunction methods and DFT: it represents a special case of Rung 5 on the "Jacobs Ladder" of John P. Perdew. For large and chemically diverse benchmarks such as GMTKN55, empirical double hybrid functionals with dispersion corrections can achieve accuracies approaching wavefunction methods at a cost not greatly dissimilar to hybrid DFT approaches, provided RI-MP2 and/or another MP2 acceleration techniques are available in the electronic structure code. Only a half-dozen or fewer empirical parameters are required. For vibrational frequencies, accuracies intermediate between CCSD and CCSD(T) can be achieved, and performance for other properties is encouraging as well. Organometallic reactions can likewise be treated well, provided static correlation is not too strong. Further prospects are discussed, including range-separated and RPA-based approaches.

physics.chem-ph↗

arXiv2019

Some Observations on the Performance of the Most Recent Exchange-Correlation Functionals for the Large and Chemically Diverse GMTKN55 Benchmark

Benchmarks that span a broad swath of chemical space, such as GMTKN55, are very useful for assessing progress in the quest for more universal DFT functionals. We find that the WTMAD2 metrics for a great number of functionals show a clear "Jacob's Ladder hierarchy"; that the "combinatorial" development strategy of Head-Gordon and coworkers generates "best on rung" performers; that the quality of the nonlocal dispersion correction becomes more important as functionals become more accurate for nondispersion properties; that fitting against small, unrepresentative benchmark sets leads to underperforming functionals; and that ωB97M(2) is currently the best DFT functional of any kind, but that revDSD-D4 functionals are able to reach similar performance using fewer parameters, and that revDOD-D4 in addition permits reduced-scaling algorithms. If one seeks a range-separated hybrid (RSH) GGA that also performs well for optical excitation energies, CAM-QTP-01 may be a viable option. The D4 dispersion model, with its partial charge dependence, appears to be clearly superior to D3BJ and even possibly NL. Should one require a double hybrid without dispersion model, noDispSD-SCAN is a viable option. Performance for the MOBH35 transition metal benchmark is different: the best double hybrids are competitive but not superior to ωB97M-V, which offers the best performance compromise for mixed main group-transition metal problems.

physics.chem-ph↗

arXiv2019

Coupled Cluster Benchmark of New Density Functionals and Domain Pair Natural Orbital Methods: Mechanisms of Hydroarylation and Oxidative Coupling Catalyzed by Ru(II) Chloride Carbonyls

In the present work, we tested the performance of several new functionals for studying the mechanisms of concurrent reaction of hydroarylation and oxidative coupling catalyzed by Ru(II) chloride carbonyls. We find that DLPNO-CCSD(T) is an acceptable substitute for full canonical CCSD(T) calculations; that the recent ωB97X-V and ωB97M-V functionals exhibit superior performance to commonly used DFT functionals; and that the revised DSD-PBEP86 double hybrid represents an improvement over the original, even though transition metals were not involved in its parametrization.

physics.chem-ph↗

arXiv2018

Do CCSD and approximate CCSD-F12 variants converge to the same basis set limits? The case of atomization energies

While the title question is a clear 'yes' from purely theoretical arguments, the case is less clear for practical calculations with finite (one-particle) basis sets. To shed further light on this issue, the basis set limits of CCSD (coupled cluster theory with all single and double excitations) and of different approximate implementations of CCSD-F12 (explicitly correlated CCSD) have been investigated on detail for the W4-17 thermochemical benchmark. Near the CBS ([1-particle] complete basis set) limit, CCSD and CCSD(F12*) agree to within their respective uncertainties (about \pm0.04 kcal/mol) due to residual basis set incompleteness error, but a nontrivial difference remains between CCSD-F12b and CCSD(F12*), which is roughly proportional to the degree of static correlation. The observed basis set convergence behavior results from the superposition of a rapidly converging, attractive, CCSD[F12]--CCSD-F12b difference (consisting mostly of third-order terms), and a more slowly converging, repulsive, fourth-order difference between CCSD(F12*) and CCSD[F12]. For accurate thermochemistry, we recommend CCSD(F12*) over CCSD-F12b if at all possible. There are some indications that the nZaPa family of basis sets exhibits somewhat smoother convergence than the correlation consistent family.

physics.chem-ph↗

arXiv2018

A Simple Model For Scalar Relativistic Corrections To Molecular Total Atomization Energies

Scalar relativistic corrections to atomization energies of 1st-and 2nd-row molecules can be rationalized in terms of a simple additive model, linear in changes in atomic s populations. In a sample of 200 first-and second-row molecules, such a model can account for over 98% of the variance (99% for the first-row subset). The remaining error can be halved again by adding a term involving the change in atomic p populations: those coefficients need not be fitted but can be fixed from atomic electron affinity calculations. This model allows a fairly accurate a priori estimate for the importance of scalar relativistic corrections on a reaction energy, at essentially zero computational cost. While this is not a substitute for explicit calculation of Douglas-Kroll-Hess (DKH) or exact two-component (X2C) relativistic corrections, the model offers an interpretative tool for the chemical analysis of scalar relativistic contributions to reaction energies.

physics.chem-ph↗

arXiv2018

Prediction of electronic couplings for molecular charge transfer using optimally-tuned range-separated hybrid functionals

Electronic coupling matrix elements are important to the theoretical description of electron transfer processes. However, they are notoriously difficult to obtain accurately from time- dependent density functional theory (TDDFT). Here, we use the HAB11 benchmark dataset of coupling matrix elements to assess whether TDDFT using optimally-tuned range-separated hybrid functionals, already known to be successful for the description of charge transfer excitation energies, also allows for an improved accuracy in the prediction of coupling matrix elements. We find that this approach outperforms all previous TDDFT calculations, based on semi-local, hybrid, or non-tuned range-separated hybrid functionals, with a remaining average deviation as low as ~12%. We discuss potential sources for the remaining error.

physics.chem-ph↗

arXiv2018

Probing the basis set limit for thermochemical contributions of inner-shell correlation: Balance of core-core and core-valence contributions

The inner-shell correlation contributions to the total atomization energies (TAEs) of the W4-17 computational thermochemistry benchmark have been determined at the CCSD(T) level near the basis set limit using several families of core correlation basis sets, such as aug-cc-pCVnZ (n=3-6), aug-cc-pwCVnZ (n=3-5), and nZaPa-CV (n=3-5). The three families of basis sets agree very well with each other (0.01 kcal/mol RMS) when extrapolating from the two largest available basis sets: however, there are considerable differences in convergence behavior for the smaller basis sets. nZaPa-CV is superior for the core-core term and awCVnZ for the core-valence term. While the aug-cc-pwCV(T+d)Z basis set of Yockel and Wilson is superior to aug-cc-pwCVTZ, further extension of this family proved unproductive. The best compromise between accuracy and computational cost, in the context of high-accuracy computational thermochemistry methods such as W4 theory, is CCSD(T)/awCV{T,Q}Z, where the {T,Q} notation stands for extrapolation from the awCVTZ and awCVQZ basis set pair. For lower-cost calculations, a previously proposed combination of CCSD-F12b/cc-pCVTZ-F12 and CCSD(T)/pwCVTZ(no f) appears to 'give the best bang for the buck'. While core-valence correlation accounts for the lion's share of the inner shell contribution in first-row molecules, for second-row molecules core-core contributions may become important, particularly in systems like P4 and S4 with multiple adjacent second-row atoms. The average absolute core-core/core-valence ratio is 0.08 for the first-row species in W4-17, but 0.47 for the second-row subset.

physics.chem-ph↗

arXiv2018

A Simple 'Range Extender' for Basis Set Extrapolation Methods for MP2 and Coupled Cluster Correlation Energies

We discuss the interrelations between various basis set extrapolation formulas and show that for the nZaPa and aug-cc-pVnZ basis set formulas, for n=4--6 their behavior closely resembles the Petersson (L+a)^{-3} formula with a shift a specific to the basis set family and level of theory. This is functionally equivalent to the Pansini-Varandas extrapolation for large L. This naturally leads to a simple way to extend these extrapolations to n=7 and higher. The formula is validated by comparison with newly optimized extrapolation factors for the AV{6,7}Z basis set pairs and literature values for {6,7}ZaPa. For L\geq5, the CCSD extrapolations of both the Schwenke and Varandas type are functionally equivalent to E(L)=E_\infty+A.(L-0.30)^{-3}, i.e., E(\infty)=E(L)+[E(L)-E(L-1)]/([(L-0.30)/(L-1.30)]^3-1)

physics.chem-ph↗

arXiv2017

The S66 noncovalent interactions benchmark reconsidered using explicitly correlated methods near the basis set limit

The S66 benchmark for noncovalent interactions has been re-evaluated using explicitly correlated methods with basis sets near the one-particle basis set limit. It is found that post-MP2 "high-level corrections" are treated adequately well using a combination of CCSD(F12*) with (aug-)cc-pVTZ-F12 basis sets on the one hand, and (T) extrapolated from conventional CCSD(T)/heavy-aug-cc-pV{D,T}Z on the other hand. Implications for earlier benchmarks on the larger S66x8 problem set in particular, and for accurate calculations on noncovalent interactions in general, are discussed. At a slight cost in accuracy, (T) can be considerably accelerated by using sano-V{D,T}Z+ basis sets, while half-counterpoise CCSD(F12*)(T)/cc-pVDZ-F12 offers the best compromise between accuracy and computational cost.

physics.chem-ph↗

arXiv2017

The aug-cc-pVnZ-F12 Basis Set Family: Correlation Consistent Basis Sets for Explicitly Correlated Benchmark Calculations on Anions and Noncovalent Complexes

We have developed a new basis set family, denoted aug-cc-pVnZ-F12 (or aVnZ-F12 for short), for explicitly correlated calculations. The sets included in this family were constructed by supplementing the corresponding cc-pVnZ-F12 sets with additional diffuse functions on the higher angular momenta (i.e., additional d-h functions on non-hydrogen atoms, and p-g on hydrogen), optimized for the MP2-F12 energy of the relevant atomic anions. The new basis sets have been benchmarked against electron affinities of the first- and second-row atoms, the W4-17 dataset of total atomization energies, the S66 dataset of noncovalent interactions, the BEGDB water clusters subset, and the WATER23 subset of the GMTKN24 and GMTKN30 benchmark suites. The aVnZ-F12 basis sets displayed excellent performance, not just for electron affinities but also for noncovalent interaction energies of neutral and anionic species. Appropriate CABS (complementary auxiliary basis sets) were explored for the S66 noncovalent interactions benchmark: between similar-sized basis sets, CABS were found to be more transferable than generally assumed.

physics.chem-ph↗

arXiv2017

MP2-F12 Basis Set Convergence for the S66 Noncovalent Interactions Benchmark: Transferability of the Complementary Auxiliary Basis Set (CABS)

Complementary auxiliary basis sets for F12 explicitly correlated calculations appear to be more transferable between orbital basis sets than has been generally assumed. We also find that aVnZ-F12 basis sets, originally developed with anionic systems in mind, appear to be superior for noncovalent interactions as well, and propose a suitable CABS sequence for them.

physics.chem-ph↗