Pith. sign in

REVIEW 2 major objections 3 minor 115 references

MC-PDFT Nuclear Gradients and L-PDFT Energies with Meta and Hybrid Meta On-Top Functionals for Ground- and Excited-State Geometry Optimization and Vertical Excitation Energies

T0 review · 2 major / 3 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read This paper derives analytic nuclear gradients for meta on-top functionals in MC-PDFT, enabling L-PDFT meta-GA energies, and finds MC23 matches tPBE0 and NEVPT2 for vertical excitations while beating the best TD-DFT functional.

desk verdict Solid gradient derivation and L-PDFT extension, but the headline benchmark claim that MC23 is 'best' only holds on a filtered subset and needs qualification or independent validation of the filter. read the letter →

arxiv 2506.03304 v3 pith:FDDY7HMH submitted 2025-06-03 physics.chem-ph quant-ph

classification physics.chem-phquant-ph
keywords MC-PDFTon-topfunctionalsmeta-GAanalyticnucleargradientsL-PDFTverticalexcitationenergiesMC23QUESTbenchmark
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

MC-PDFT corrects a multiconfigurational reference wave function with an on-top density functional, so the accuracy of the method depends on the quality of that functional. The newest class of on-top functionals, the meta-gradient-approximation (meta-GA) family that includes MC23, could previously be used only for energies, not for geometry optimizations or dynamics, because no analytic nuclear gradients existed. This paper derives those gradients, together with the underlying derivatives of the meta-GA on-top functional with respect to the density matrices, and the same derivatives unlock L-PDFT multi-state energy calculations with meta-GA functionals. Benchmarked on 441 vertical excitations from the QUEST database, MC23 is the most accurate of nine meta and hybrid meta on-top functionals and is comparable to the established tPBE0 functional and to NEVPT2, while a direct comparison shows MC-PDFT outperforming the best Kohn-Sham density functional. The result matters because it puts the most flexible class of on-top functionals within reach of the full pipeline of excited-state modeling, from geometry optimization through spectroscopy.

What carries the argument

The load-bearing object is the Jacobian $\mathbf{J}^{\tilde{\rho}}_{\rho}$ of the meta-GA translation scheme: the matrix of derivatives of the seven effective spin-density variables (two spin densities, three gradient-magnitude squares, and two kinetic-energy densities) with respect to the four MC-PDFT argument functions (density $\rho$, on-top pair density $\Pi$, density gradient $\nabla\rho$, and kinetic-energy density $\tau$). The translation is controlled by the ratio $R = 4\Pi/\rho^2$ through the factor $\zeta_t = \sqrt{1-R}$ (set to zero when $R \ge 1$), and the two rows for $\tau_\uparrow$ and $\tau_\downarrow$ are the new pieces that extend the earlier gradient-approximation Jacobian to meta functionals. The one- and two-electron on-top potential terms $V^q_p$ and $v^{qs}_{pr}$ that enter both the Lagrange-multiplier response equations for the nuclear gradients and the L-PDFT Hamiltonian are computed as the chain-rule product $\mathbf{v}_{\mathrm{ot}} = \mathbf{v}_{xc} \cdot \mathbf{J}^{\tilde{\rho}}_{\rho}$.

What would settle it

Re-examine the 99 excluded excitations with larger active spaces, or simply rank the nine functionals on all 540 QUEST excitations without the 0.55 eV filter; if the excluded cases are dominated by genuine functional error rather than active-space inadequacy, MC23's rank as the best meta functional and the claimed parity with NEVPT2 would not survive the unfiltered comparison.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is twofold. Methodologically, the authors derive the Jacobian of the meta-GA translation that maps the four MC-PDFT argument functions to the seven effective spin-density variables of a meta-GGA exchange-correlation functional, and they use it to obtain analytic nuclear gradients for state-specific and state-averaged MC-PDFT, with and without density fitting, verified against numerical gradients on LiH to about ${\sim}10^{-5}$ hartree/bohr. The same one- and two-electron on-top potential terms are exactly what builds the linearized multi-state L-PDFT Hamiltonian, so meta-GA functionals become usable there as well. Scientifically, on 441 QUEST vertical excitations selected for active-space adequacy, the MC23 functional posts a mean unsigned error of 0.17 eV, the best of the nine meta and hybrid meta functionals tested, against 0.15 eV for both tPBE0 and NEVPT2. On the 359-excitation subset shared with a prior TD-DFT study, the best on-top functional (tPBE0, 0.15 eV) beats the best Kohn-Sham functional (M06-SX, 0.20 eV) and roughly halves the error of its own Kohn-Sham counterpart PBE0 (0.29 eV).

Load-bearing premise

The benchmark removes every excitation for which the tPBE0 functional's error exceeds 0.55 eV on the assumption that such large errors mean the active space is inadequate, and this filter is load-bearing because on the unfiltered 540-excitation set the ranking changes, with tTPSSh (0.46 eV) below MC23 (0.48 eV).

Editorial extensions

If this is right

  • Ground- and excited-state geometry optimization with MC23 and the other meta-GA on-top functionals becomes routine; the paper demonstrates it for s-trans-butadiene and benzophenone.
  • L-PDFT, the multi-state formulation used near conical intersections and avoided crossings, can now run with any meta-GA on-top functional, and the paper reports the first such L-PDFT energies.
  • MC23 transfers to excited states even though it was optimized on ground-state data, giving the best vertical excitation energies among the nine meta and hybrid meta functionals tested.
  • MC-PDFT reaches NEVPT2-level accuracy for vertical excitation energies at a fraction of the perturbation theory's cost, and it beats the best TD-DFT functional in a controlled subset comparison.
  • L-PDFT and MC-PDFT agree within about 0.02 eV across functionals, so the cheaper linearized route preserves single-state accuracy wherever state interaction is weak.

Reading between the lines

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

  • The 'best meta functional' ranking is conditional on the 0.55 eV tPBE0 filter: on the unfiltered 540-excitation set the order shifts, with tTPSSh (0.46 eV) edging MC23 (0.48 eV), so the headline accuracy claims should be read as applying to excitations for which the reference active spaces are deemed adequate.
  • Because the translation factor $\zeta_t$ is stepwise with a kink at $R = 1$, the on-top potential is only piecewise smooth; trajectories that cross into the $R \ge 1$ region could encounter a derivative discontinuity, an effect the paper does not probe.
  • The near-parity of L-PDFT and MC-PDFT errors suggests the linearized method could carry meta-GA functionals into nonadiabatic dynamics work once the paper's stated next step, analytic L-PDFT gradients for meta-GA, is completed.
  • Since MC23 was fitted to ground-state properties and a few spin splittings, a functional optimized on excitation energies as well could plausibly push the 0.17 eV mean unsigned error below the NEVPT2 level it currently matches.
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

2 major / 3 minor

Summary. This paper derives and implements analytic nuclear gradients for multiconfiguration pair-density functional theory (MC-PDFT) with meta-GA and hybrid meta-GA on-top functionals, including the MC23 functional, and uses the same derivative machinery to enable linearized PDFT (L-PDFT) energies with meta-GA functionals. The gradients are validated against numerical gradients for LiH with both SS- and SA-CASSCF references, and the new capabilities are applied to ground- and excited-state geometry optimizations of s-trans-butadiene and benzophenone. The paper also benchmarks nine meta-GA and hybrid meta-GA on-top functionals on 441 QUEST vertical excitation energies and compares MC-PDFT results with TD-DFT on a 359-excitation intersection. The authors report that MC23 is the best of the nine meta-GA functionals, comparable to tPBE0 and NEVPT2, and that MC-PDFT outperforms the best KS-DFT functional.

Significance. The methodological contribution is significant and appears sound: the analytic-gradient implementation for meta-GA on-top functionals, with agreement to numerical gradients at the 10^-6 hartree/bohr level, is an enabling step for geometry optimization and dynamics with the MC23 functional, and the L-PDFT extension is a natural and useful consequence. The benchmark results are interesting, but the headline performance claims are contingent on a tPBE0-error-based filter, and the unfiltered data do not support the unqualified statement that MC23 is the best meta-GA functional. The paper's strengths include a clear derivation, a reproducible implementation in PySCF-forge, and numerical validation.

major comments (2)
  1. [Section IV.D, Fig. S5, abstract, and conclusion] The claim that MC23 performs best among the nine meta and hybrid meta functionals for vertical excitation energies is established only on the 441-excitation subset obtained by removing all excitations for which the tPBE0 error exceeds 0.55 eV. On the unfiltered 540-excitation set shown in Fig. S5, MC23 has an MUE of 0.48 eV, while tTPSSh has 0.46 eV and tSCAN also has 0.48 eV, so MC23 is not the best meta-GA functional without the filter. The filter is motivated in Section IV.D by the statement that large tPBE0 errors 'may be an indication that the active spaces of those reference wave functions are inadequate,' but no independent evidence is provided that tPBE0 error is a valid diagnostic of active-space inadequacy rather than simply a measure of functional error. Because tPBE0 is itself one of the ranked methods, the selection is not neutral. The abstract and conclusion state the 'best' claim without this caveat. Please either validate the filter with an independent criterion (for example, a wave-function diagnostic such as the approximate pair coefficient or a comparison with larger active spaces) or qualify all headline claims to the filtered subset, including in the abstract and conclusion.
  2. [Section IV.E and Fig. 3] The direct comparison to TD-DFT is performed on a 359-excitation intersection that inherits the tPBE0 filter, so the statement that 'MC-PDFT outperforms even the best performing Kohn-Sham density functional' is likewise filter-dependent. Furthermore, the Liang et al. subset used for the TD-DFT comparison has its own selection criteria, so the comparison mixes two different filters. Please report the unfiltered comparison if the underlying data are available, or explicitly state that this conclusion holds only for the tPBE0-filtered intersection. Without this qualification, the comparison overstates the general superiority of MC-PDFT over TD-DFT.
minor comments (3)
  1. [Section IV.D] The text states that excitations are removed from 'the original set of 520 excitations,' but Fig. S5 labels the unfiltered set as 'All data: 540 Excitations.' Please reconcile these numbers.
  2. [Section III, functional list] In the list of tested functionals, 'tMN5-L' appears to be a typo for 'tMN15-L'; please correct it.
  3. [Supporting Information heading] The SI section header contains 'benzophenol'; the molecule studied is benzophenone. Please correct the typo.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the gradient derivation is self-contained and the benchmark uses external QUEST references.

full rationale

The paper's central methodological derivation is self-contained. The meta-GA on-top potential terms (Eqs. 19-44) and the analytic nuclear-gradient expressions (Eqs. 50-63) are obtained by direct chain-rule differentiation of the translated functional definition (Eq. 10) and are numerically validated on LiH (Table I), so no central result is imported as a prediction or forced by construction. The L-PDFT potentials are the same derivatives used as on-top potentials, not a separate fitted quantity. The benchmark uses external QUEST theoretical best estimates as references and evaluates fixed functionals from prior work (MC23, tTPSS, tTPSSh, tSCAN, etc.) with no parameter fitted to the 441 or 540 excitation energies reported here. The tPBE0-error-based exclusion of excitations with unsigned error above 0.55 eV is a disclosed sample-selection rule rather than a fitted constant; although it affects which comparisons are drawn (the unfiltered ranking differs, with tTPSSh slightly ahead of MC23 in Fig. S5), that is a representativeness or validity concern, not a circular reduction of the derivation to its inputs. Self-citations supply computational inputs and prior functional definitions, but the gradient result is independently checked against numerical gradients and the benchmark is anchored to external QUEST data, so no load-bearing self-citation chain or definitional equivalence is present.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central new contribution is a mathematical derivation, so the ledger contains no invented entities. The two free parameters are the benchmark filter threshold and the MC23 mixing coefficient from prior work. The main domain assumption is that the filtered QUEST subset is representative.

free parameters (2)
  • tPBE0 error cutoff = 0.55 eV
    Hand-chosen threshold to remove 79 of 520 QUEST excitations before benchmarking; directly affects the reported MUEs and ranking.
  • MC23 hybrid mixing parameter lambda = 0.2856
    Adopted from ref 20, fitted to ground-state and spin-splitting data; used here as a fixed constant in all MC23 calculations.
assumptions (3)
  • domain assumption The meta-GA translation Jacobian (eqs. 25-44) is correct, including the piecewise definition of zeta_t and its derivatives at R=1.
    The gradient and L-PDFT derivations rest on these chain-rule derivatives; they are validated only indirectly via LiH numerical gradients.
  • domain assumption The SA-CASSCF reference wave functions from ref 16, selected via the approximate pair coefficient scheme, are adequate for all retained QUEST excitations.
    The benchmark assumes the reference active spaces are of sufficient quality; this is the justification for removing high-error excitations.
  • standard math Standard calculus and Lagrange multiplier response theory as used in refs 30 and 31 apply to MC-PDFT with meta-GA functionals.
    The derivation follows the established response formalism, with no new mathematical machinery introduced.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MC-PDFT Nuclear Gradients and L-PDFT Energies with Meta and Hybrid Meta On-Top Functionals for Ground- and Excited-State Geometry Optimization and Vertical Excitation Energies." pith.science (2026). https://pith.science/paper/FDDY7HMH

@misc{pith2026250603304,
  author       = {Pith},
  title        = {Pith review of: MC-PDFT Nuclear Gradients and L-PDFT Energies with Meta and Hybrid Meta On-Top Functionals for Ground- and Excited-State Geometry Optimization and Vertical Excitation Energies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FDDY7HMH}},
  note         = {Machine review of arXiv:2506.03304}
}
read the original abstract

Multiconfiguration pair-density functional theory (MC-PDFT) is a post-MCSCF multireference electronic-structure method that explicitly models strong electron correlation, and linearized pair-density functional theory (L-PDFT) is a recently developed multi-state extension that can accurately model conical intersections and locally-avoided crossings. Because MC-PDFT and L-PDFT rely on an on-top energy functional, their accuracy depends on the quality of the on-top functional used. Recent work has introduced translated meta-gradient-approximation (meta-GA) on-top functionals, and specifically the MC23 hybrid meta-GA on-top functional, which is the first on-top functional specifically optimized for MC-PDFT. Here we report the derivation and implementation of analytic nuclear gradients for MC-PDFT calculations using meta-GA and hybrid meta-GA on-top functionals. This development also enables analytic nuclear gradients for the widely successful tPBE0 hybrid on-top functional. Because MC-PDFT nuclear-gradient calculations involve the derivative of the on-top functional, this development also enables the use of meta-GA on-top functionals in L-PDFT single-point energy calculations. We use the new capabilities to test MC23 for ground-state geometries, excited-state geometries, and vertical excitation energies of s-trans-butadiene and benzophenone as well as to test MC23, another hybrid meta-GA, and seven other meta-GA on-top functionals for 441 vertical excitation energies. We find MC23 performs the best of all nine meta and hybrid meta functionals for vertical excitation energies and is comparable in accuracy to tPBE0 and to the NEVPT2 multireference wave function method. Additionally, we directly compare our MC-PDFT vertical excitation results to previously computed TD-DFT values and find that MC-PDFT outperforms even the best performing Kohn-Sham density functional.

Figures

Figures reproduced from arXiv: 2506.03304 by the authors.

Figure 1
Figure 1. FIG. 1. Ground-state geometry of benzophenone optimized [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Error statistics across the 441 excitations for SA [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Error statistics comparison between Liang et al. [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

115 extracted references · 80 canonical work pages

  1. [1]

    Kohn, W.; Sham, L. J. Self-consistent equations includ- ing exchange and correlation effects. Phys. Rev. 1965, 140, A1133–A1138

  2. [2]

    O.; Taylor, P

    Roos, B. O.; Taylor, P. R.; Siegbahn, P. E. M. A complete active space SCF method (CASSCF) using a density ma- trix formulated super-CI approach. Chem. Phys. 1980, 48, 157–173

  3. [3]

    Roos, B. O. In Ab Initio Methods in Quantum Chem- istry Part II ; Lawley, K. P., Ed.; Advances in Chemical Physics; John Wiley & Sons, Ltd: Great Britain, 1987; Vol. 69; pp 399–445

  4. [4]

    M.; Elbert, S

    Ruedenberg, K.; Cheung, L. M.; Elbert, S. T. MCSCF optimization through combined use of natural orbitals and the Brillouin–Levy–Berthier theorem. Int. J. Quan- tum Chem. 1979, 16, 1069–1101

  5. [5]

    R.; Gagliardi, L

    Hermes, M. R.; Gagliardi, L. Multiconfigurational self- consistent field theory with density matrix embedding: the localized active space self-consistent field method. J. Chem. Theory Comput. 2019, 15, 972–986

  6. [6]

    R.; Pandharkar, R.; Gagliardi, L

    Hermes, M. R.; Pandharkar, R.; Gagliardi, L. Varia- tional localized active space self-consistent field method. J. Chem. Theory Comput. 2020, 16, 4923–4937

  7. [7]

    O.; Sadlej, A

    Andersson, K.; Malmqvist, P.- ˚ A.; Roos, B. O.; Sadlej, A. J.; Wolinski, Krzysztof. Second-order pertur- bation theory with a CASSCF reference function. J. Phys. Chem. 1990, 94, 5483–5488

  8. [8]

    Andersson, K.; Malmqvist, P.- ˚ A.; Roos, B. O. Second- order perturbation theory with a complete active space self-consistent field reference function. J. Chem. Phys. 1992, 96, 1218–1226

Show all 115 references
  1. [9]

    Introduction of N -electron valence states for multireference perturbation theory

    Angeli, C.; Cimiraglia, R.; Evangelisti, S.; Leininger, T .; Malrieu, J.-P. Introduction of N -electron valence states for multireference perturbation theory. J. Chem. Phys. 2001, 114, 10252–10264

  2. [10]

    K.; Luo, S.; Ma, D.; Olsen, J.; Truhlar, D

    Li Manni, G.; Carlson, R. K.; Luo, S.; Ma, D.; Olsen, J.; Truhlar, D. G.; Gagliardi, L. Multiconfiguration pair- density functional theory. J. Chem. Theory Comput. 2014, 10, 3669–3680

  3. [11]

    Malmqvist, P.- ˚ A.; Rendell, A.; Roos, B. O. The re- stricted active space self-consistent-field method, imple- mented with a split graph unitary group approach. J. Phys. Chem. 1990, 94, 5477–5482

  4. [12]

    Zhou, C.; Gagliardi, L.; Truhlar, D. G. Multiconfigu- ration pair-density functional theory for iron porphyrin with CAS, RAS, and DMRG active spaces. J. Phys. Chem. A 2019, 123, 3389–3394. 13

  5. [13]

    K.; Truh- lar, D

    Odoh, S.; Li Manni, G.; Carlson, R. K.; Truh- lar, D. G.; Gagliardi, L. Separated-Pair Approxima- tion and Separated-Pair Pair-Density Functional Theory. Chem. Sci. 2016, 7, 3669–3680

  6. [14]

    R.; Cramer, C

    Pandharkar, R.; Hermes, M. R.; Cramer, C. J.; Truh- lar, D. G.; Gagliardi, L. Localized active space pair- density functional theory. J. Chem. Theory Comput. 2021, 17, 2843–2851

  7. [15]

    E.; Ghosh, S.; Truhlar, D

    Hoyer, C. E.; Ghosh, S.; Truhlar, D. G.; Gagliardi, L. Multiconfiguration pair-density functional theory is as accurate as CASPT2 for electronic excitation. J. Phys. Chem. Lett. 2016, 7, 586–591

  8. [16]

    S.; Hermes, M

    King, D. S.; Hermes, M. R.; Truhlar, D. G.; Gagliardi, L. Large-scale benchmarking of multireference vertical- excitation calculations via automated active-space selec - tion. J. Chem. Theory Comput. 2022, 18, 6065–6076

  9. [17]

    B.; Truhlar, D

    Calio, P. B.; Truhlar, D. G.; Gagliardi, L. Nonadiabatic molecular dynamics by multiconfiguration pair-density functional theory. J. Chem. Theory Comput. 2022, 18, 614–622

  10. [18]

    R.; Truhlar, D

    Hennefarth, M. R.; Truhlar, D. G.; Gagliardi, L. Semi- classical nonadiabatic molecular dynamics using lin- earized pair-density functional theory. J. Chem. Theory Comput. 2024, 20, 8741–8748

  11. [19]

    K.; Truhlar, D

    Carlson, R. K.; Truhlar, D. G.; Gagliardi, L. Multiconfig- uration pair-density functional theory: A fully translated gradient approximation and its performance for transi- tion metal dimers and the spectroscopy of Re 2Cl8 2–. J. Chem. Theory Comput. 2015, 11, 4077–4085

  12. [20]

    J.; Zhang, D.; Zhang, S.; Gagliardi, L.; Truh- lar, D

    Bao, J. J.; Zhang, D.; Zhang, S.; Gagliardi, L.; Truh- lar, D. G. A hybrid meta on-top functional for multi- configuration pair-density functional theory. Proc. Natl. Acad. Sci. 2025, 122, e2419413121

  13. [21]

    R.; Truhlar, D

    Pandharkar, R.; Hermes, M. R.; Truhlar, D. G.; Gagliardi, L. A new mixing of nonlocal exchange and nonlocal correlation with multiconfiguration pair-density functional theory. J. Phys. Chem. Lett. 2020, 11, 10158– 10163

  14. [22]

    R.; Hermes, M

    Hennefarth, M. R.; Hermes, M. R.; Truhlar, D. G.; Gagliardi, L. Linearized pair-density functional theory. J. Chem. Theory Comput. 2023, 19, 3172–3183

  15. [23]

    Reference energies for double excitations

    Loos, P.-F.; Boggio-Pasqua, M.; Scemama, A.; Caf- farel, M.; Jacquemin, D. Reference energies for double excitations. J. Chem. Theory Comput. 2019, 15, 1939– 1956

  16. [24]

    A mountaineering strategy to excited states: Highly accurate reference energies and benchmarks

    Loos, P.-F.; Scemama, A.; Blondel, A.; Garniron, Y.; Caffarel, M.; Jacquemin, D. A mountaineering strategy to excited states: Highly accurate reference energies and benchmarks. J. Chem. Theory Comput. 2018, 14, 4360– 4379

  17. [25]

    A mountaineering strategy to excited states: Highly accurate energies and benchmarks for medium sized molecules

    Loos, P.-F.; Lipparini, F.; Boggio-Pasqua, M.; Sce- mama, A.; Jacquemin, D. A mountaineering strategy to excited states: Highly accurate energies and benchmarks for medium sized molecules. J. Chem. Theory Comput. 2020, 16, 1711–1741

  18. [26]

    Mountaineering strategy to excited states: Highly accurate energies and benchmarks for exotic molecules and radicals

    Loos, P.-F.; Scemama, A.; Boggio-Pasqua, M.; Jacquemin, D. Mountaineering strategy to excited states: Highly accurate energies and benchmarks for exotic molecules and radicals. J. Chem. Theory Comput. 2020, 16, 3720–3736

  19. [27]

    QUESTDB: A database of highly accurate excitation energies for the electronic structure community

    V´ eril, M.; Scemama, A.; Caffarel, M.; Lipparini, F.; Boggio-Pasqua, M.; Jacquemin, D.; Loos, P.-F. QUESTDB: A database of highly accurate excitation energies for the electronic structure community. WIREs Comput. Mol. Sci. 2021, 11, e1517

  20. [28]

    The quest for highly accurate excitation energies: A computational perspective

    Loos, P.-F.; Scemama, A.; Jacquemin, D. The quest for highly accurate excitation energies: A computational perspective. J. Phys. Chem. Lett. 2020, 11, 2374–2383

  21. [29]

    Molecular Electronic-Strcuture Theory; John Wiley & Sons: Chich- ester, UK, 2000, pp

    Helgaker, T.; Jørgensen, P.; Olsen, J. Molecular Electronic-Strcuture Theory; John Wiley & Sons: Chich- ester, UK, 2000, pp. 36-37

  22. [30]

    M.; Hoyer, C

    Sand, A. M.; Hoyer, C. E.; Sharkas, K.; Kidder, K. M.; Lindh, R.; Truhlar, D. G.; Gagliardi, L. Analytic gra- dients for complete active space pair-density functional theory. J. Chem. Theory Comput. 2018, 14, 126–138

  23. [31]

    R.; Hermes, M

    Scott, T. R.; Hermes, M. R.; Sand, A. M.; Oakley, M. S.; Truhlar, D. G.; Gagliardi, L. Analytic gradients for state- averaged multiconfiguration pair-density functional the- ory. J. Chem. Phys. 2020, 153, 014106

  24. [32]

    H.; Teukolsky, S

    Press, W. H.; Teukolsky, S. A.; Vetterling, W. T.; Flan- nery, B. P. Numerical Recipes in Fortran 77: The Art of Scientific Computing , 2nd ed.; Cambridge University Press: Cambridge, UK, 1992

  25. [33]

    A direct implementation of the second-order derivatives of multiconfigurational SCF energies and an analysis of the preconditioning in the associated response equation

    Bernhardsson, A.; Lindh, R.; Olsen, J.; Fulscher, M. A direct implementation of the second-order derivatives of multiconfigurational SCF energies and an analysis of the preconditioning in the associated response equation. Mol. Phys. 1999, 96, 617–628

  26. [34]

    W.; Thorne, K

    Misner, C. W.; Thorne, K. S.; Wheeler, J. A. Gravitation; Macmillan: San Francisco, CA, USA, 1973

  27. [35]

    Beebe, N. H. F.; Linderberg, J. Simplifications in the generation and transformation of two-electron integrals in molecular calculations. Int. J. Quantum Chem. 1977, 12, 683–705

  28. [36]

    R.; Oakley, M

    Scott, T. R.; Oakley, M. S.; Hermes, M. R.; Sand, A. M.; Lindh, R.; Truhlar, D. G.; Gagliardi, L. Analytic gradi- ents for multiconfiguration pair-density functional theory with density fitting: development and application to ge- ometry optimization in the ground and excited sta...

  29. [37]

    C.; Blunt, N

    Sun, Q.; Berkelbach, T. C.; Blunt, N. S.; Booth, G. H.; Guo, S.; Li, Z.; Liu, J.; McClain, J. D.; Sayfut- yarova, E. R.; Sharma, S.; Wouters, S.; Chan, G. K.- L. PySCF: the Python-based simulations of chemistry framework. WIREs Comput. Mol. Sci. 2018, 8, e1340

  30. [38]

    Sun, Q. et al. Recent developments in the PySCF pro- gram package. J. Chem. Phys. 2020, 153, 024109

  31. [39]

    2025; https://github.com/pyscf/ pyscf-forge

    PySCF-forge. 2025; https://github.com/pyscf/ pyscf-forge

  32. [40]

    Libcint: An efficient general integral library for Gaussian basis functions

    Sun, Q. Libcint: An efficient general integral library for Gaussian basis functions. J. Comput. Chem. 2015, 36, 1664–1671

  33. [41]

    Marques, M. A. L.; Oliveira, M. J. T.; Burnus, T. Libxc: A library of exchange and correlation functionals for den- sity functional theory. Comput. Phys. Commun. 2012, 183, 2272–2281

  34. [42]

    Lehtola, S.; Steigemann, C.; Oliveira, M. J. T.; Mar- ques, M. A. L. Recent developments in libxc — A com- prehensive library of functionals for density functional theory. SoftwareX 2018, 7, 1–5

  35. [43]

    Dunning Jr., T. H. Gaussian basis sets for use in corre- lated molecular calculations. I. The atoms boron through neon and hydrogen. J. Chem. Phys. 1989, 90, 1007–1023

  36. [44]

    A.; Dunning Jr., T

    Kendall, R. A.; Dunning Jr., T. H.; Harrison, R. J. Elec- tron affinities of the first-row atoms revisited. Systematic basis sets and wave functions. J. Chem. Phys. 1992, 96, 6796–6806. 14

  37. [45]

    Hermes, M. R. mrh. 2025; https://github.com/ MatthewRHermes/mrh

  38. [46]

    Geometry optimization made sim- ple with translation and rotation coordinates

    Wang, L.-P.; Song, C. Geometry optimization made sim- ple with translation and rotation coordinates. J. Chem. Phys. 2016, 144, 214108

  39. [47]

    The role of databases in support of computa- tional chemistry calculations

    Feller, D. The role of databases in support of computa- tional chemistry calculations. J. Comput. Chem. 1996, 17, 1571–1586

  40. [48]

    L.; Didier, B

    Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L.; Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. Basis set exchange: A community database for computational sciences. J. Chem. Inf. Model. 2007, 47, 1045–1052

  41. [49]

    Papajak, E.; Truhlar, D. G. Convergent partially aug- mented basis sets for post-hartree-fock calculations of molecular properties and reaction barrier heights. J. Chem. Theory Comput. 2010, 7, 10–18

  42. [50]

    P.; Staroverov, V

    Tao, J.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Climbing the density functional ladder: Nonempirical meta–generalized gradient approximation designed for molecules and solids. Phys. Rev. Lett. 2003, 91, 146401

  43. [51]

    P.; Tao, J.; Staroverov, V

    Perdew, J. P.; Tao, J.; Staroverov, V. N.; Scuseria, G. E. Meta-generalized gradient approximation: Explanation of a realistic nonempirical density functional. J. Chem. Phys. 2004, 120, 6898–6911

  44. [52]

    N.; Scuseria, G

    Staroverov, V. N.; Scuseria, G. E.; Tao, J.; Perdew, J. P. Comparative assessment of a new nonempirical density functional: Molecules and hydrogen-bonded complexes. J. Chem. Phys. 2003, 119, 12129–12137

  45. [53]

    Sun, J.; Ruzsinszky, A.; Perdew, J. P. Strongly con- strained and appropriately normed semilocal density functional. Phys. Rev. Lett. 2015, 115, 036402

  46. [54]

    W.; Kaplan, A

    Furness, J. W.; Kaplan, A. D.; Ning, J.; Perdew, J. P.; Sun, J. Accurate and numerically efficient r2SCAN meta- generalized gradient approximation. J. Phys. Chem. Lett. 2020, 11, 8208–8215

  47. [55]

    accurate and numerically efficient r2SCAN meta-generalized gradient approximation

    Furness, J. W.; Kaplan, A. D.; Ning, J.; Perdew, J. P.; Sun, J. Correction to “accurate and numerically efficient r2SCAN meta-generalized gradient approximation”. J. Phys. Chem. Lett. 2020, 11, 9248

  48. [56]

    Zhao, Y.; Truhlar, D. G. A new local density functional for main-group thermochemistry, transition metal bond- ing, thermochemical kinetics, and noncovalent interac- tions. J. Chem. Phys. 2006, 125, 194101

  49. [57]

    S.; Truhlar, D

    Wang, Y.; Jin, X.; Yu, H. S.; Truhlar, D. G.; He, X. Re- vised M06-L functional for improved accuracy on chem- ical reaction barrier heights, noncovalent interactions, and solid-state physics. Proc. Natl. Acad. Sci. 2017, 114, 8487–8492

  50. [58]

    S.; He, X.; Truhlar, D

    Yu, H. S.; He, X.; Truhlar, D. G. MN15-L: A new lo- cal exchange-correlation functional for Kohn–Sham den- sity functional theory with broad accuracy for atoms, molecules, and solids. J. Chem. Theory Comput. 2016, 12, 1280–1293

  51. [59]

    D.; Handy, N

    Boese, A. D.; Handy, N. C. New exchange-correlation density functionals: The role of the kinetic-energy den- sity. J. Chem. Phys. 2002, 116, 9559–9569

  52. [60]

    S.; Gagliardi, L

    King, D. S.; Gagliardi, L. A ranked-orbital approach to select active spaces for high-throughput multireference computation. J. Chem. Theory Comput. 2021, 17, 2817– 2831

  53. [61]

    J.; Hermes, M

    Bao, J. J.; Hermes, M. R.; Scott, T. R.; Sand, A. M.; Lindh, R.; Gagliardi, L.; Truhlar, D. G. Analytic gra- dients for compressed multistate pair-density functional theory. Mol. Phys. 2022, 120, e2110534

  54. [62]

    R.; Hermes, M

    Hennefarth, M. R.; Hermes, M. R.; Truhlar, D. G.; Gagliardi, L. Analytic nuclear gradients for complete ac- tive space linearized pair-density functional theory. J. Chem. Theory Comput. 2024, 20, 3637–3658

  55. [63]

    Shu, Y.; Truhlar, D. G. Doubly excited character or static correlation of the reference state in the controversial 2 1Ag state of trans-butadiene? J. Am. Chem. Soc. 2017, 139, 13770–13778

  56. [64]

    H.; Bunnenberg, E.; Djerassi, C.; Records, R

    Haugen, W.; Trætteberg, M.; Kaufmann, F.; Motzfeldt, K.; Williams, D. H.; Bunnenberg, E.; Djerassi, C.; Records, R. The molecular structure of 1,3-butadiene and 1,3,5-trans-hexatriene. Acta Chem. Scand. 1966, 20, 1726–1728

  57. [65]

    A.; Chan, G

    Watson, M. A.; Chan, G. K.-L. Excited states of butadi- ene to chemical accuracy: Reconciling theory and exper- iment. J. Chem. Theory Comput. 2012, 8, 4013–4018

  58. [66]

    The benzophenone S 1(n,π*) → T1(n,π*) states intersystem crossing reinvestigated by ultrafast absorption spectroscopy and multivariate curve resolution

    Alo ¨ ıse, S.; Ruckebusch, C.; Blanchet, L.; R´ ehault, J .; Buntinx, G.; Huvenne, J.-P. The benzophenone S 1(n,π*) → T1(n,π*) states intersystem crossing reinvestigated by ultrafast absorption spectroscopy and multivariate curve resolution. J. Phys. Chem. A 2008, 112, 224–231

  59. [67]

    Triplet exciton formation in a benzophenone single crystal stud- ied by picosecond time-resolved absorption spectroscopy

    Katoh, R.; Kotani, M.; Hirata, Y.; Okada, T. Triplet exciton formation in a benzophenone single crystal stud- ied by picosecond time-resolved absorption spectroscopy. Chem. Phys. Lett. 1997, 264, 631–635

  60. [68]

    Why does intersystem crossing occur in isolated molecules of benzaldehyde, ace- tophenone, and benzophenone? J

    Ohmori, N.; Suzuki, T.; Ito, M. Why does intersystem crossing occur in isolated molecules of benzaldehyde, ace- tophenone, and benzophenone? J. Phys. Chem. 1988, 92, 1086–1093

  61. [69]

    Hoffmann, R.; Swenson, J. R. Ground- and excited-state geometries of benzophenone. J. Phys. Chem. 1970, 74, 415–420

  62. [70]

    Sergentu, D.-C.; Maurice, R.; Havenith, R. W. A.; Broer, R.; Roca-Sanju´ an, D. Computational determina- tion of the dominant triplet population mechanism in photoexcited benzophenone. Phys. Chem. Chem. Phys. 2014, 16, 25393–25403

  63. [71]

    Nuclear energy gradient s for internally contracted complete active space second- order perturbation theory: Multistate extensions

    Vlaisavljevich, B.; Shiozaki, T. Nuclear energy gradient s for internally contracted complete active space second- order perturbation theory: Multistate extensions. J. Chem. Theory Comput. 2016, 12, 3781–3787

  64. [72]

    B.; Sung, N.; Hawkinson, S

    Fleischer, E. B.; Sung, N.; Hawkinson, S. Crystal struc- ture of benzophenone. J. Phys. Chem. 1968, 72, 4311– 4312

  65. [73]

    Emission characteristics of benzophenone vapor at low pressure

    Itoh, T. Emission characteristics of benzophenone vapor at low pressure. J. Phys. Chem. 1985, 89, 3949–3951

  66. [74]

    R.; King, D

    Hennefarth, M. R.; King, D. S.; Gagliardi, L. Linearized pair-density functional theory for vertical excitation en- ergies. J. Chem. Theory Comput. 2023, 19, 7983–7988

  67. [75]

    Liang, J.; Feng, X.; Hait, D.; Head-Gordon, M. Revisit- ing the performance of time-dependent density functional theory for electronic excitations: Assessment of 43 popu- lar and recently developed functionals from rungs one to four. J. Chem. Theory Comput. 2022, 18, 3460–3473

  68. [76]

    G.; He, X

    Wang, Y.; Verma, P.; Zhang, L.; Li, Y.; Liu, Z.; Truh- lar, D. G.; He, X. M06-SX screened-exchange density functional for chemistry and solid-state physics. Proc. Natl. Acad. Sci. 2020, 117, 2294–2301. 15 TOC GRAPHIC Supporting Information: MC-PDFT Nuclear Gradients and L-PDFT...

  69. [77]

    cyanoformaldehyde-x1

  70. [78]

    cyanoformaldehyde-x2

  71. [79]

    cyanoformaldehyde-x3

  72. [80]

    cyanoformaldehyde-x4

  73. [82]

    cyclopentadienethione-x1

  74. [83]

    cyclopentadienethione-x2

  75. [84]

    cyclopentadienethione-x4

  76. [85]

    cyclopentadienethione-x6

  77. [86]

    cyclopentadienethione-x7

  78. [87]

    cyclopentadienethione-x8

  79. [88]

    cyclopentadienone-x1

  80. [89]

    cyclopentadienone-x2

  81. [90]

    cyclopentadienone-x6

  82. [91]

    cyclopentadienone-x7

  83. [92]

    cyclopentadienone-x8

  84. [96]

    cyclopropenethione-x1

  85. [97]

    cyclopropenethione-x2

  86. [98]

    cyclopropenethione-x3

  87. [99]

    cyclopropenethione-x4

  88. [100]

    cyclopropenethione-x5

  89. [101]

    cyclopropenethione-x6

  90. [102]

    cyclopropenethione-x7

  91. [103]

    cyclopropenethione-x8

  92. [104]

    cyclopropenethione-x9

  93. [105]

    cyclopropenethione-x10

  94. [110]

    cyclopropenone-x5 S9

  95. [207]

    hydrogen-chloride-x1

  96. [232]

    methylenecyclopropene-x1

  97. [233]

    methylenecyclopropene-x2

  98. [234]

    methylenecyclopropene-x3

  99. [235]

    methylenecyclopropene-x5

  100. [236]

    methylenecyclopropene-x6

  101. [302]

    streptocyanine-c1-x1

  102. [303]

    streptocyanine-c1-x2

  103. [304]

    streptocyanine-c3-x1

  104. [305]

    streptocyanine-c3-x2

  105. [306]

    streptocyanine-c5-x1

  106. [307]

    streptocyanine-c5-x2

  107. [359]

    Mean unsigned error (MUE) on the 359- excitation subset for MC-PDFT, L-PDFT, SA-CASSCF, and NEVPT2

    water-x6 TABLE S1. Mean unsigned error (MUE) on the 359- excitation subset for MC-PDFT, L-PDFT, SA-CASSCF, and NEVPT2. Method MUE SA-CASSCF 0 .47 NEVPT2 0 .15 tPBE 0 .15 L-tPBE 0 .16 MC23 0 .17 L-MC23 0 .18 tTPSS 0 .18 L-tTPSS 0 .20 tTPSSh 0 .18 L-tTPSSh 0 .19 tτ -HCTH 0 .20 L...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.