Pith. sign in

REVIEW 4 major objections 6 minor 25 references

The vDZP Basis Set Is Effective For Many Density Functionals

T0 review · 4 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The vDZP basis set, originally developed for a single composite method, works across many density functionals without reparameterization, giving near-composite accuracy in thermochemistry, barrier heights, geometries, and torsional…

desk verdict Solid transferability benchmark for vDZP, but the headline GMTKN55 numbers rest on 50/55 subsets — the five omitted are the heavy-element/large-system cases most likely to expose trouble. read the letter →

arxiv 2411.13253 v1 pith:F4OMZMR5 submitted 2024-11-20 physics.chem-ph

Corin C. Wagen , Jonathon E. Vandezande This is my paper · ORCID
classification physics.chem-ph PACS 31.15.Ew
keywords vDZPbasissetdensityfunctionaltheorycompositemethodssuperpositionerrorGMTKN55effectivecorepotentialstransitionmetalbarrierheightstorsionalenergyprofiles
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Computational chemists have long assumed that accurate density functional theory needs triple-zeta basis sets, and that small-basis methods only work if the functional, basis, and empirical corrections are reparameterized together into composite schemes. This paper claims that vDZP, a polarized double-zeta basis set with effective core potentials and deeply contracted functions originally built for the composite method ωB97X-3c, is not overfit to that one method. When paired with four additional functionals (B97-D3BJ, r2SCAN-D4, B3LYP-D4, and M06-2X), vDZP yields main-group thermochemistry errors close to those of the huge (aug)-def2-QZVP basis and roughly half the errors of conventional double-zeta sets, with no reparameterization beyond the standard D4 dispersion correction. On transition-metal barriers, rotational constants, and drug-like torsional profiles, vDZP methods match or beat fine-tuned composite methods. If right, vDZP gives a general low-cost basis that breaks the speed-accuracy tradeoff that motivated bespoke composite schemes.

What carries the argument

vDZP is a polarized valence double-zeta basis set whose defining features are large-core effective core potentials that remove core electrons, deeply contracted valence basis functions, and parameters optimized on molecular systems rather than free atoms. These features suppress the two classic failures of small basis sets, basis-set incompleteness error (the density is too rigid) and basis-set superposition error (fragments borrow each other's basis functions), down to near triple-zeta levels. The basis thus acts as a drop-in replacement: any functional can be combined with it, and the only extra ingredient needed is the now-standard D4 empirical dispersion correction for functionals that lack dispersion.

What would settle it

Run the five omitted GMTKN55 subsets with vDZP and all five functionals in a quantum chemistry program that implements effective core potentials correctly, then compare the weighted mean absolute deviations back to the (aug)-def2-QZVP references: if any subset's error is dramatically larger than the trends in Tables 1–3, the paper's general-applicability claim fails.

Watch

Extended reading notes

Core claim

At the paper's center is the demonstration that vDZP is a general-purpose basis rather than a bespoke component. The authors combine vDZP with five functionals spanning GGA, meta-GGA, hybrid, and range-separated hybrid classes, and evaluate them on GMTKN55 (minus five subsets that the software's effective-core-potential implementation cannot handle), revMOBH35, ROT34, and TorsionNet206. The result: weighted mean absolute deviations for vDZP are only moderately worse than those of (aug)-def2-QZVP, clearly better than 6-31G(d), def2-SVP, and pcseg-1, and comparable to the purpose-built composite methods B97-3c, r2SCAN-3c, and ωB97X-3c. For geometry predictions, vDZP-based methods actually produce the lowest mean deviations among the low-cost methods tested, and for torsions they are within 0.02–0.06 kcal/mol of triple-zeta hybrids. The paper argues this breaks the assumption that the tradeoff between speed and accuracy can only be resolved by tight coupling of methods, basis sets, and empirical corrections.

Load-bearing premise

The conclusion that vDZP is generally applicable depends on the five omitted GMTKN55 subsets (NBPRC, FH51, DC13, C60ISO, and HEAVY28) not hiding a systematic failure, particularly for heavy elements and large systems where an ECP-heavy, deeply contracted basis might be most fragile.

Editorial extensions

If this is right

  • vDZP can replace conventional double-zeta bases such as 6-31G(d), def2-SVP, and pcseg-1 for routine DFT, roughly halving weighted mean absolute errors on GMTKN55 without increasing runtime.
  • For drug-design workflows, torsional scans with B97-D3BJ/vDZP or r2SCAN-D4/vDZP approach triple-zeta accuracy at a fraction of the cost, making broad conformer screening feasible.
  • The five functionals tested all retain most of their large-basis accuracy with vDZP, so users are not locked into a single bespoke composite method.
  • For heavy-element-rich systems, vDZP's extensive use of effective core potentials can make it substantially faster than composite methods, with a 3.4-fold speedup observed for perbromo-n-pentane.
  • Basis-set design that optimizes molecular performance directly, rather than atomic energies, is a promising route to further Pareto improvements in quantum chemistry.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's omission of five GMTKN55 subsets means the claim of general applicability is directly tested only for the main-group organic subset; a user targeting heavy-element or large-system chemistry should look for a dedicated test before relying on vDZP.
  • A natural next experiment is to run vDZP with a functional not in the test set, such as a double-hybrid, to see if the 'no reparameterization' pattern holds when the functional itself has a different error profile.
  • The reported timing data suggest vDZP's real speed advantage will emerge only after integral codes are optimized for deeply contracted, low-angular-momentum basis functions; until then, practical speed parity with composite methods is partly hardware- and software-dependent.
  • If the omitted heavy-element subsets do degrade, the paper's stronger claim of high generality would fall back to 'highly effective for organic main-group chemistry,' a useful but narrower result.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper reports that the vDZP basis set, originally developed for the composite method ωB97X-3c, can be combined with several unmodified density functionals (B97-D3BJ, r2SCAN-D4, B3LYP-D4, M06-2X, and ωB97X-D4) to yield accuracies that approach those of the much larger (aug)-def2-QZVP basis set and of purpose-built composite methods such as B97-3c and r2SCAN-3c. The evidence consists of GMTKN55 weighted total mean absolute deviations (WTMAD2, with five subsets omitted), revMOBH35 transition-metal barrier heights, ROT34 rotational constants, TorsionNet206 torsional profiles, and n-alkane/perbromo-n-alkane timing studies. The central claim is that vDZP is a generally applicable low-cost basis set that does not require functional- or correction-specific reparameterization.

Significance. If the claim holds, this is a practically important result: it would overturn the usual assumption that double-zeta basis sets are inadequate except inside tightly coupled composite schemes, and it would give computational chemists a cheap, general-purpose basis for main-group thermochemistry, geometries, and conformational energies. The paper's strengths are its use of standard external benchmark sets, its reliance on previously optimized external parameters (vDZP, D4, and the functionals themselves) rather than any fit performed in this work, and the clear head-to-head comparisons against conventional double-zeta basis sets. The main quantitative finding—that vDZP substantially outperforms 6-31G(d), def2-SVP, and pcseg-1 while approaching composite-method accuracy—is useful even if the general-applicability claim is later bounded by the omitted benchmark subsets.

major comments (4)
  1. [Methodology and Table 1] The general-applicability conclusion rests on GMTKN55 WTMAD2 values in Table 1, but five subsets (NBPRC, FH51, DC13, C60ISO, HEAVY28) are omitted because of "documented errors in Psi4's effective-core-potential implementation." These subsets are not random: they contain heavy main-group elements, halogen-containing systems, and large conjugated molecules, exactly the regimes in which a basis built from large-core ECPs and deep contraction could be most fragile. As written, the WTMAD2 numbers are not the full GMTKN55 WTMAD2, and the statement that vDZP is "generally applicable" is therefore conditional. The authors should either compute these subsets with a code having a correct ECP implementation (e.g., ORCA, Q-Chem, or a newer Psi4 version) or report per-subset errors/bounds for the omitted sets to demonstrate that no systematic failure is hidden.
  2. [Methodology] The paper states that "a custom basis-set file was used which adds the missing basis functions for fluorine" because of a documented absence of fluorine in Psi4's internal vDZP implementation. No content or provenance of this custom file is given. Fluorine appears in many GMTKN55 subsets and in the other benchmarks; if the added functions are not exactly the published vDZP fluorine functions, the reported numbers are not vDZP results for those species. The custom basis file, its source, and a verification against an independent implementation should be provided, and the affected subset results should be identified.
  3. [Table 1, footnote a] The abstract and conclusion claim that vDZP works "without any method- or correction-specific reparameterization," but the ωB97X-D4 row in Table 1 carries the footnote "utilizes a refit D4 correction." If this is the D4 parametrization optimized for ωB97X-3c/vDZP, then that row is not a test of an unmodified functional and should not be cited as evidence of transferability. The claim should be restricted to the four truly unmodified functionals, or the refit should be justified as a standard parameter set rather than a bespoke correction.
  4. [Supporting Information and Methodology] The paper provides only summary spreadsheets in the Supporting Information and no input files, scripts, or machine-readable protocol. Given that the central claim is a numerical benchmark comparison, the absence of input files for the GMTKN55, revMOBH35, ROT34, and TorsionNet206 calculations—especially the custom fluorine basis file and the exact D4 version/parameters—prevents independent verification and limits the utility of the results as a community resource. The authors should archive the complete input set.
minor comments (6)
  1. [Results and discussion, Table 1] The text says reference values were obtained with "(aug)-def2-QZVP," but the table column header reads "def2-QZVP"; this inconsistency should be resolved.
  2. [Methodology] There is a typo in the list of omitted subsets: "HEA VY28" should be "HEAVY28," and a footnote explaining that Table 1's WTMAD2 excludes these subsets should be added directly to the table.
  3. [Table 1, footnote a] The footnote "ωB97X-3x utilizes a refit D4 correction" appears to contain a typo; it should presumably refer to ωB97X-D4 or ωB97X-3c.
  4. [Introduction] The phrase "remove core elections" should be "remove core electrons."
  5. [Results and discussion, timing] The timing section reports that vDZP-based methods were on average 40% slower than the corresponding composite methods, yet Figure 1 and the discussion conclude "comparable efficiency." The 40% slowdown and the hardware-specific nature of the result should be stated more prominently so that the Pareto-efficiency claim is not overstated.
  6. [Results and discussion, TorsionNet206] For the TorsionNet206 comparison, the reference method used to score the benchmark (CCSD(T)/def2-TZVP) is mentioned, but the protocol for obtaining the MAE values (e.g., single-point energies at a fixed geometry versus relaxed scans) is not fully specified; a sentence clarifying this would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: vDZP and D4 are external, previously determined inputs, and the paper tests their transfer to new functionals against standard external benchmarks.

full rationale

The paper reports benchmark results rather than a derivation. The load-bearing inputs — the vDZP basis set and the D4 dispersion correction — come from prior work by other groups (Refs. 13 and the cited composite-method literature), and the present authors do not reparameterize either. The central claim is that vDZP, originally designed for ωB97X-3c, transfers to four other functionals. This is an empirical transfer test, not a quantity defined in terms of the conclusion: success is measured by WTMAD2, MAE, and rotational/torsional errors against external reference sets (GMTKN55, revMOBH35, ROT34, TorsionNet206), not by any fit performed here. No equation in the paper reduces to its own input; no fitted parameter is relabeled as a prediction; and the references to the ωB97X-3c and B97-3c/r2SCAN-3c work are external citations, not same-author self-citations that carry the argument. The omission of five GMTKN55 subsets (NBPRC, FH51, DC13, C60ISO, HEAVY28) due to Psi4 ECP errors is a genuine generality and completeness concern that may affect the strength of the headline claim, but it is a correctness or scope issue, not circularity: the reported numbers still come from the paper's own independent computations against standard references. The footnote that ωB97X-D4 uses a refit D4 correction from Ref. 13 is an external parameterization detail and does not make the reported transfer test circular. Therefore no specific circular step can be exhibited, and the score is 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claim depends entirely on external, previously optimized inputs: the vDZP basis set, the D4 dispersion model, and the empirical density functionals. No new entity is postulated and no parameter is fitted in this paper. The main assumptions are computational implementation correctness and the representativeness of the benchmark sets, with the omitted Psi4 ECP subsets being the largest concern.

free parameters (3)
  • vDZP basis set parameters (exponents, contraction coefficients, ECPs) = As optimized in Müller et al. 2023 (ωB97X-3c development)
    The central claim rests on these externally optimized basis functions; they are not re-fit here.
  • D4 dispersion parameters = As published in the Grimme group D4 model
    Used without reparameterization for D4-corrected functionals; central accuracy comparisons depend on them.
  • Functional parameters for B97-D3BJ, r2SCAN-D4, B3LYP-D4, M06-2X, ωB97X-D4 = Standard published parametrizations
    The claim that vDZP works with 'many functionals' inherits the quality of these empirical functionals.
assumptions (4)
  • domain assumption The custom Psi4 basis file correctly adds the missing fluorine functions to vDZP
    Psi4's internal vDZP lacks fluorine; the paper uses a custom file but does not validate it against the original vDZP definition. If the F functions are wrong, F-containing benchmark results are compromised.
  • domain assumption Omitted GMTKN55 subsets do not bias the general-applicability conclusion
    NBPRC, FH51, DC13, C60ISO, and HEAVY28 are excluded; the paper assumes the remaining subsets are representative of the full benchmark.
  • domain assumption Reference values in GMTKN55, revMOBH35, ROT34, and TorsionNet206 are accurate and appropriate for assessing DFT errors
    The entire accuracy assessment compares against these external benchmark references.
  • domain assumption Standard DFT/ECP machinery in Psi4 is numerically reliable for the computed energies and gradients
    No validation against other codes is provided; one documented ECP issue already forced subset omissions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of The vDZP Basis Set Is Effective For Many Density Functionals." pith.science (2026). https://pith.science/paper/F4OMZMR5

@misc{pith2026241113253,
  author       = {Pith},
  title        = {Pith review of: The vDZP Basis Set Is Effective For Many Density Functionals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F4OMZMR5}},
  note         = {Machine review of arXiv:2411.13253}
}
read the original abstract

In recent years, "composite" density-functional-theory-based methods comprising specially optimized combinations of functionals, basis sets, and empirical corrections have become widely used owing to their robustness and computational efficiency, but the bespoke nature of these methods makes them challenging to develop. Here, we report that the recently reported vDZP basis set can be used in combination with a wide variety of density functionals to produce efficient and accurate results comparable to those obtained with composite methods, but without any method- or correction-specific reparameterization. This result enables rapid quantum chemical calculations to be run with a variety of density functionals without the typical errors incurred by small basis sets.

Figures

Figures reproduced from arXiv: 2411.13253 by the authors.

Figure 1
Figure 1. Timings for single-point energies of n-alkanes. We also note that the extensive use of ECPs in vDZP can lead to substantial rate acceler￾ations for systems with large numbers of heavy elements, since substantially fewer electrons will be modeled. For the particularly dramatic case of perbromo-n-pentane, B97-3c is 3.4x slower than B97-D3BJ/vDZP, and r2SCAN-3c is 2.7x slower than r2SCAN-D4/vDZP. Over￾all, vDZP-based m… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 24 canonical work pages

  1. [1]

    Reviews in Computational Chemistry; John Wiley & Sons, Ltd, 2017; Chapter 3, pp 93--149

    Nagy, B.; Jensen, F. Reviews in Computational Chemistry; John Wiley & Sons, Ltd, 2017; Chapter 3, pp 93--149

    work page 2017

  2. [2]

    Basis sets for molecular calculations

    Huzinaga, S. Basis sets for molecular calculations. Computer Physics Reports 1985, 2, 281--339

    work page 1985

  3. [3]

    R.; Truhlar, D

    Papajak, E.; Zheng, J.; Xu, X.; Leverentz, H. R.; Truhlar, D. G. Perspectives on Basis Sets Beautiful: Seasonal Plantings of Diffuse Basis Functions. Journal of Chemical Theory and Computation 2011, 7, 3027--3034, PMID: 26598144

    work page 2011

  4. [4]

    D.; Martin, J

    Boese, A. D.; Martin, J. M. L.; Handy, N. C. The role of the basis set: Assessing density functional theory. The Journal of Chemical Physics 2003, 119, 3005–3014

    work page 2003

  5. [5]

    Why the Standard B3LYP/6-31G* Model Chemistry Should Not Be Used in DFT Calculations of Molecular Thermochemistry: Understanding and Correcting the Problem

    Kruse, H.; Goerigk, L.; Grimme, S. Why the Standard B3LYP/6-31G* Model Chemistry Should Not Be Used in DFT Calculations of Molecular Thermochemistry: Understanding and Correcting the Problem. The Journal of Organic Chemistry 2012, 77, 10824--10834, PMID: 23153035

    work page 2012

  6. [6]

    Best-Practice DFT Protocols for Basic Molecular Computational Chemistry

    Bursch, M.; Mewes, J.-M.; Hansen, A.; Grimme, S. Best-Practice DFT Protocols for Basic Molecular Computational Chemistry. Angewandte Chemie International Edition 2022, 61, e202205735

    work page 2022

  7. [7]

    Assessing conformer energies using electronic structure and machine learning methods

    Folmsbee, D.; Hutchison, G. Assessing conformer energies using electronic structure and machine learning methods. International Journal of Quantum Chemistry 2021, 121, e26381

    work page 2021

  8. [8]

    Corrected small basis set Hartree-Fock method for large systems

    Sure, R.; Grimme, S. Corrected small basis set Hartree-Fock method for large systems. Journal of Computational Chemistry 2013, 34, 1672--1685

    work page 2013

Show all 25 references
  1. [9]

    G.; Bannwarth, C.; Hansen, A

    Grimme, S.; Brandenburg, J. G.; Bannwarth, C.; Hansen, A. Consistent structures and interactions by density functional theory with small atomic orbital basis sets. The Journal of Chemical Physics 2015, 143, 054107

    2015

  2. [10]

    G.; Hochheim, M.; Bredow, T.; Grimme, S

    Brandenburg, J. G.; Hochheim, M.; Bredow, T.; Grimme, S. Low-Cost Quantum Chemical Methods for Noncovalent Interactions. The Journal of Physical Chemistry Letters 2014, 5, 4275--4284, PMID: 26273974

    2014

  3. [11]

    G.; Bannwarth, C.; Hansen, A.; Grimme, S

    Brandenburg, J. G.; Bannwarth, C.; Hansen, A.; Grimme, S. B97-3c: A revised low-cost variant of the B97-D density functional method. The Journal of Chemical Physics 2018, 148, 064104

    2018

  4. [12]

    Swiss army knife

    Grimme, S.; Hansen, A.; Ehlert, S.; Mewes, J.-M. r ^2 SCAN-3c: A “Swiss army knife” composite electronic-structure method. The Journal of Chemical Physics 2021, 154, 064103

    2021

  5. [13]

    B97X-3c: A composite range-separated hybrid DFT method with a molecule-optimized polarized valence double- basis set

    Müller, M.; Hansen, A.; Grimme, S. B97X-3c: A composite range-separated hybrid DFT method with a molecule-optimized polarized valence double- basis set. The Journal of Chemical Physics 2023, 158, 014103

    2023

  6. [14]

    Optimal Small Basis Set and Geometric Counterpoise Correction for DFT Computations

    Chan, B. Optimal Small Basis Set and Geometric Counterpoise Correction for DFT Computations. Journal of Chemical Theory and Computation 2023, 19, 3958--3965, PMID: 37288982

    2023

  7. [15]

    Turney, J. M. et al. Psi4: an open-source ab initio electronic structure program. WIREs Computational Molecular Science 2012, 2, 556--565

    2012

  8. [16]

    E.; Frisch, M

    Stratmann, R.; Scuseria, G. E.; Frisch, M. J. Achieving linear scaling in exchange-correlation density functional quadratures. Chemical Physics Letters 1996, 257, 213--223

    1996

  9. [17]

    Geometry optimization made simple with translation and rotation coordinates

    Wang, L.-P.; Song, C. Geometry optimization made simple with translation and rotation coordinates. The Journal of Chemical Physics 2016, 144, 214108

    2016

  10. [18]

    A look at the density functional theory zoo with the advanced GMTKN55 database for general main group thermochemistry , kinetics and noncovalent interactions

    Goerigk, L.; Hansen, A.; Bauer, C.; Ehrlich, S.; Najibi, A.; Grimme, S. A look at the density functional theory zoo with the advanced GMTKN55 database for general main group thermochemistry , kinetics and noncovalent interactions. Phys. Chem. Chem. Phys. 2017, 19, 32184--32215

    2017

  11. [19]

    A.; Janes, T

    Iron, M. A.; Janes, T. Evaluating Transition Metal Barrier Heights with the Latest Density Functional Theory Exchange–Correlation Functionals: The MOBH35 Benchmark Database. The Journal of Physical Chemistry A 2019, 123, 3761--3781, PMID: 30973722

    2019

  12. [20]

    Semidalas, E.; Martin, J. M. The MOBH35 Metal–Organic Barrier Heights Reconsidered: Performance of Local-Orbital Coupled Cluster Approaches in Different Static Correlation Regimes. Journal of Chemical Theory and Computation 2022, 18, 883--898, PMID: 35045709

    2022

  13. [21]

    Implementation of nuclear gradients of range-separated hybrid density functionals and benchmarking on rotational constants for organic molecules

    Risthaus, T.; Steinmetz, M.; Grimme, S. Implementation of nuclear gradients of range-separated hybrid density functionals and benchmarking on rotational constants for organic molecules. Journal of Computational Chemistry 2014, 35, 1509--1516

    2014

  14. [22]

    K.; Jang, H.; Horton, J

    Behara, P. K.; Jang, H.; Horton, J. T.; Gokey, T.; Dotson, D. L.; Boothroyd, S.; Bayly, C. I.; Cole, D. J.; Wang, L.-P.; Mobley, D. L. Benchmarking Quantum Mechanical Levels of Theory for Valence Parametrization in Force Fields. The Journal of Physical Chemistry B 2024, 128, 7...

    2024

  15. [23]

    A machine learning-based high-precision density functional method for drug-like molecules

    Xiao, J.; Chen, Y.; Zhang, L.; Wang, H.; Zhu, T. A machine learning-based high-precision density functional method for drug-like molecules. Artificial Intelligence Chemistry 2024, 2, 100037

    2024

  16. [24]

    A.; Hehre, W

    Pople, J. A.; Hehre, W. J. Computation of electron repulsion integrals involving contracted Gaussian basis functions. Journal of Computational Physics 1978, 27, 161--168

    1978

  17. [25]

    low-cost

    Gill, P. M. In Molecular integrals Over Gaussian Basis Functions; Sabin, J. R., Zerner, M. C., Eds.; Advances in Quantum Chemistry; Academic Press, 1994; Vol. 25; pp 141--205 mcitethebibliography vDZP.bib0000664000000000000000000011504214717350372011073 0ustar rootroot @inbook...

    1994

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.