Pith. sign in

REVIEW 3 major objections 4 minor 93 references

Exact quantum dynamics of methanol: full-dimensional ab initio potential energy surface of spectroscopic quality and variational vibrational states

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper reports a full-dimensional ab initio methanol potential energy surface on which variational vibrational band origins agree with gas-phase experiment within 5 cm-1 up to 2500 cm-1, including torsional tunnelling splittings.

desk verdict A serious, high-quality methanol PES study whose headline agreement with experiment is real but partly propped up by an 11.8 cm-1 torsional barrier discrepancy that should be explained. read the letter →

arxiv 2505.16262 v2 pith:Z3CHUIK4 submitted 2025-05-22 physics.chem-ph quant-ph

Ayaki Sunaga , Tibor Győri , Gábor Czakó , Edit Matyus This is my paper · ORCID
classification physics.chem-phquant-ph
keywords methanolpotentialenergysurfacevariationalvibrationalcomputationinternalrotationtunnellingsplittingspermutationallyinvariantpolynomialsactivelearningcoupled-cluster
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

The paper seeks to show that methanol's quantum vibrational dynamics can now be computed from first principles to near-spectroscopic accuracy. It builds a new full-dimensional potential energy surface, PES25, from about 39,400 explicitly correlated coupled-cluster energies selected by an active-learning procedure and fitted with permutationally invariant polynomials. Solving the 12-dimensional vibrational Schrödinger equation on this surface, the authors obtain vibrational band origins that agree with gas-phase experiment within 5 cm$^{-1}$ up to about 2500 cm$^{-1}$, with mean absolute errors around 1–1.5 cm$^{-1}$. This makes methanol a molecule in which large-amplitude internal rotation, tunnelling splittings, and small-amplitude vibrations are all described by a single ab initio surface, which is directly relevant to astronomical methanol spectra and precision tests of fundamental physics.

What carries the argument

The load-bearing construct is PES25, a permutationally invariant polynomial (PIP) representation of the potential written in Morse variables, fitted to 39,401 explicitly correlated coupled-cluster energies chosen by an iterative active-learning loop that adds geometries where the current surface is most uncertain. The vibrational calculation uses a coordinate system whose large-amplitude part is the torsional angle $\tau$, expressed as a symmetric combination of the three H-C-O-H dihedral angles, with 11 small-amplitude curvilinear normal coordinates built along the minimum-energy path in $\tau$. A harmonic-oscillator basis for the small-amplitude modes times a Fourier basis for $\tau$, combined with a Smolyak non-product grid and basis/grid pruning, keeps the 12D Hamiltonian matrix tractable. This machinery is what lets the torsional tunnelling splittings and anharmonic couplings be computed exactly rather than modeled.

What would settle it

Build a second full-dimensional surface at a higher electronic-structure level (for example quadruple-zeta basis or all-electron correlation) using the same active-learning and coordinate pipeline, and recompute the same band origins. If the mean absolute deviation from experiment does not fall below the current 0.9–1.5 cm$^{-1}$, or if the individual deviations scatter rather than shrink, the present agreement would be shown to rely partly on error cancellation.

Watch

Extended reading notes

Core claim

On the paper's own terms, PES25, a full-dimensional permutationally invariant polynomial fit to 39,401 CCSD(T)-F12b/cc-pVTZ-F12 energies, together with an exact variational solution of the 12D vibrational Schrödinger equation, reproduces the vibrational band origins of CH$_3$OH within 5 cm$^{-1}$ of gas-phase experiment. The improvement is concentrated in the vibration-torsion coupling and combination-band ranges, where the earlier PES13 deviated by 10–20 cm$^{-1}$; the maximum deviations with PES25 are 2.1, 4.0, and 3.2 cm$^{-1}$ in the torsional, coupling, and combination ranges, respectively. The paper argues that this makes PES25 the first full-dimensional ab initio surface for methanol to reach spectroscopic quality in variational vibrational computations, covering not only the small-amplitude fundamentals but also states in which torsional excitation is combined with CO-stretch overtones.

Load-bearing premise

The result depends on the assumption that the electronic-structure method used to generate the surface is accurate to roughly a few cm$^{-1}$ in the energy regions sampled by the vibrations; if errors in the torsional barrier or anharmonic coupling regions are larger and cancel against fitting errors, the agreement with experiment would be partly accidental.

Editorial extensions

If this is right

  • Methanol's unmeasured vibrational states up to about 2500 cm$^{-1}$ become concrete predictions with estimated accuracy of a few cm$^{-1}$, ready to be tested by high-resolution infrared spectroscopy.
  • Tunnelling splittings of torsion-vibration states can now be obtained from the PES wavefunctions rather than from effective models, so the splitting pattern as a function of vibrational excitation is a direct observable test of the surface.
  • The same surface can be combined with electric dipole and polarizability surfaces to simulate infrared and Raman spectra line by line for methanol and its isotopologues.
  • Because methanol's torsion-rotation levels are used as astrophysical probes of the proton-to-electron mass ratio, a validated ab initio surface gives those searches a firmer theoretical anchor.
  • The residual few-cm$^{-1}$ errors identify where electronic structure theory must improve next: core correlation, larger basis sets, and relativistic or quantum-electrodynamic corrections.

Reading between the lines

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

  • The agreement is not a fully blind test: the paper's supplement reports that the active-learning sampling was continued after an early comparison with experiment showed deviations, so experimental information influenced which fitting geometries were added. A strictly out-of-sample test would fix the sampling protocol before any experimental comparison.
  • The reported torsional barrier on PES25 (363.1 cm$^{-1}$) is about 12 cm$^{-1}$ above the direct ab initio value (351.3 cm$^{-1}$); since tunnelling splittings are exponentially sensitive to the barrier, a precise measurement of $A/E$ splittings in highly excited torsional states could help decide which value is correct.
  • The method should transfer to other molecules with internal rotation, such as acetaldehyde or methylamine, where a large-amplitude torsional mode also couples to small-amplitude vibrations; the methanol result suggests the active-learning plus polynomial-fitting pipeline can reach spectroscopic quality there too.
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

3 major / 4 minor

Summary. The paper reports a full-dimensional (12D) ab initio potential energy surface for methanol, PES25, obtained by fitting 39,401 CCSD(T)-F12b/cc-pVTZ-F12 energies with permutationally invariant polynomials using an active-learning (Robosurfer) procedure. The authors then solve the vibrational Schrödinger equation variationally with the GENIUSH-Smolyak approach and compare the computed vibrational band origins with gas-phase experimental data up to about 2500 cm−1. Table V reports mean absolute errors of 0.9, 1.5, and 1.5 cm−1 for the torsional, SAM-torsion, and combination-band ranges, respectively, with maximum errors below 4.1 cm−1. The paper also provides detailed state assignments, convergence tests, and a discussion of remaining sources of error.

Significance. If the reported accuracy is robust, this is a significant milestone: it would be the first full-dimensional ab initio surface for methanol that reproduces variational vibrational band origins to within a few cm−1 over a range including torsion-vibration couplings and combination bands. The work also demonstrates a carefully documented active-learning pipeline and a scalable variational approach for a 12-dimensional large-amplitude system, which is of broader methodological value. Strengths include the explicit convergence tests, the extensive supplementary state tables, and the fact that the PES coefficients are fitted to electronic-structure energies rather than directly to experimental band origins. However, the validation is weakened by two issues discussed below: a relatively large discrepancy between the fitted and ab initio torsional barrier, and the use of experimental data as a stopping criterion during PES development.

major comments (3)
  1. [Supplementary Material, Section S3 and Fig. S13] The final PES25 has a relaxed torsional barrier of 363.1 cm−1, whereas the CCSD(T)-F12b/cc-pVTZ-F12 reference value is 351.3 cm−1, a deviation of 11.8 cm−1. This is more than twice the largest MAX error reported in Table V (4.0 cm−1), and it occurs in the coordinate that controls the torsional and tunnelling motions emphasized in the paper. The paper acknowledges the deviation as relatively large but does not quantify its effect on the computed vibrational levels. As a result, the excellent agreement with experiment in Table V could arise from cancellation between the fitting error in the barrier and the electronic-structure error of the barrier, rather than from a faithful ab initio surface in this region. Please add a quantitative test, for example by recomputing the torsional and vibration-torsion states with the barrier shifted to the ab initio value, or by comparing the levels obtained with the intermediate fitting sets shown in Fig. S13, and report the resulting changes in the band origins. This is necessary to support the claim that PES25 is a reliable ab initio surface in the dynamically relevant region.
  2. [Supplementary Material, Section S1.3 and Table S2] The active-learning procedure used experimental band origins as a stopping criterion. At the 30,401-point stage, the paper states that the vibrational levels deviated more from experiment than the levels computed with PES13, which prompted continued sampling; the final 39,401-point set was then validated against the same experimental data in Table V. Thus the reported agreement is not a fully independent predictive test of the ab initio model. This usage of experimental data as a development checkpoint should be disclosed in the main text, along with the number of points added after the experimental comparison was first consulted and a discussion of whether the final agreement persists if the experimental comparison is withheld until the end. This does not invalidate the PES, but it changes the strength of the claim that the surface is a purely first-principles prediction.
  3. [Section II, Table II] The total RMS fitting error of PES25 over the fitting set is 27.3 cm−1 (weighted RMS 17.3 cm−1), which is an order of magnitude larger than the mean absolute errors reported in Table V. The paper argues convincingly that fitting-set RMSE is not a direct measure of spectroscopic accuracy, but it does not provide a propagation-of-error estimate for the final band origins. A practical test would be to compare vibrational levels obtained with the DGELS and DGELSY variants given in Table S4, or with a refit to a random subset of the fitting points, so that the reader can see how much of the sub-cm−1 VBO agreement is robust to the fitting ambiguity. Without such a test, the relation between the large fitting RMSE and the claimed spectroscopic accuracy remains qualitative.
minor comments (4)
  1. [Abstract and Table V] The abstract says the computed band origins agree with experiment within 5 cm−1. This is technically consistent with Table V, but since the maximum deviations in the table are 2.1, 4.0, and 3.2 cm−1, the phrase 'within a few cm−1' or quoting the actual maxima would be more informative.
  2. [Throughout text and Supplementary Material] The corrupted symbol '⁄tildelow' appears in several places, for example in Section II.B and Section S1.3. It appears to be a LaTeX rendering error and should be replaced by the intended symbol, presumably 'approximately'.
  3. [Section III.A] The basis and grid parameters b = 7 and H = 21 are introduced without definitions in the main text; the authors refer to earlier papers, but a brief parenthetical explanation would improve readability, especially since these parameters are central to the reported convergence of 0.5–1.5 cm−1.
  4. [Figure 2] The caption of Figure 2 does not explain the colored bands or the legend entries 'TW' and 'Ref.' in sufficient detail; please add a sentence describing what the shaded region represents and how the reference values were obtained.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the PES is fitted to ab initio electronic energies, and the experimental band origins are not fitting inputs.

full rationale

The derivation chain is self-contained: CCSD(T)-F12b/cc-pVTZ-F12 electronic energies are collected by Robosurfer; the PES is obtained by weighted least-squares fitting of permutationally invariant polynomials to those energies (Eqs. S1 and S3, Table II); and the vibrational band origins are computed by direct variational solution of the vibrational Schroedinger equation (Eq. 3, main text). The experimental VBOs appear only as a final comparison (Table V, Table VI), not as data in the least-squares fit. The only target-informed element is described in Section S1.3: at an intermediate fitting set of 30,401 geometries, a preliminary variational computation showed larger deviations from experiment than PES13, and this prompted continued Robosurfer sampling. That is model selection using the experimental test set, which can bias the reported agreement, but it is not a parametric fit of the PES to the VBOs and does not make the predicted band origins equal to the input data by construction. The torsional-barrier discrepancy highlighted in Fig. S13 (PES25 363.1 cm-1 vs ab initio 351.3 cm-1) is a fidelity or cancellation issue, not circularity, because the PES coefficient values are still determined by the ab initio fitting set rather than by the experimental levels. Self-citations concern the group's own variational machinery (GENIUSH-Smolyak, Robosurfer) and prior methanol/complex applications; they are methodological references, not load-bearing justifications that replace independent evidence for the physical result. No uniqueness theorem or ansatz is imported to force the conclusion. Accordingly, no formal circular step is present.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim (VBOs within 5 cm-1 of experiment) depends on the PES25 fit and the variational solver. The free parameters are the 9652 PIP coefficients and the hand-chosen fitting hyperparameters. The axioms are the standard approximations of electronic structure theory and the numerical assumptions of the surrogate model. No new physical entities are introduced.

free parameters (4)
  • PIP polynomial coefficients (9652) = not listed
    Coefficients of the permutationally invariant polynomial fit, determined by weighted linear least squares against 39,401 CCSD(T)-F12b/cc-pVTZ-F12 energies. They define PES25 and control the vibrational eigenvalues.
  • Distance transformation parameters a = a=1.9, a_2b=2.5, a_3b=1.7, a_4b=3.0 bohr
    Chosen via test fits to minimize fitting error and improve numerical conditioning; they shape the basis functions of the PES fit.
  • Weight function parameters = E_slopeStart=13,011 cm-1, E_slopeEnd=15,984 cm-1, w_end=0.75, E_dwt=0.1 hartree
    Hand-selected weighting for the LLS fit and for deciding which geometries to add to the fitting set.
  • Polynomial degrees of main and extra q-body terms = main degree 7; extra 2-body degree 2, 3-body degree 3, 4-body degree 4
    Chosen by the authors to balance flexibility and numerical rank.
assumptions (6)
  • standard math Born-Oppenheimer separation of electronic and nuclear motion
    The PES concept assumes nuclei move on a single electronic potential energy surface. Standard for molecular quantum dynamics.
  • domain assumption CCSD(T)-F12b/cc-pVTZ-F12 with frozen core predicts electronic energies accurately enough in the sampled region
    The entire PES is built from this electronic structure level. The paper estimates basis set errors of ~1.7 cm-1 from literature (Ref. 85) and notes core correlation is missing.
  • ad hoc to paper The PIP functional form can represent the true PES in the sampled 12D region
    The chosen polynomial degrees and distance transformation parameters are specific to this work; the fit RMSE of 27.3 cm-1 shows the representational error is not negligible.
  • ad hoc to paper The fitted PES interpolates reliably between the 39,401 fitting points
    Validation uses ~174,000 spare geometries that were not in the fitting set, but these spares were used in the active-learning selection procedure, so they are not fully independent (stated in Section II.B).
  • standard math The vibrational kinetic energy operator in curvilinear coordinates (Eq. 4) is exact for the chosen coordinates
    Derived from the metric tensor; standard result in the GENIUSH framework.
  • domain assumption The basis and grid truncation (n_tau=32, M_tau=54, b=7, H=21) converge the variational eigenvalues
    Convergence to 0.5 cm-1 up to 2000 cm-1 and 1.5 cm-1 up to 2200 cm-1 is claimed based on prior and present tests; states above 2200 cm-1 may be less converged.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exact quantum dynamics of methanol: full-dimensional ab initio potential energy surface of spectroscopic quality and variational vibrational states." pith.science (2026). https://pith.science/paper/Z3CHUIK4

@misc{pith2026250516262,
  author       = {Pith},
  title        = {Pith review of: Exact quantum dynamics of methanol: full-dimensional ab initio potential energy surface of spectroscopic quality and variational vibrational states},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Z3CHUIK4}},
  note         = {Machine review of arXiv:2505.16262}
}
abstract

The methanol molecule is a sensitive probe of astrochemistry, astrophysics, and fundamental physics. The first-principles elucidation and prediction of its rotation-torsional-vibrational motions are enabled in this work by the computation of a full-dimensional, \emph{ab initio} potential energy surface (PES) and numerically exact quantum dynamics. An active-learning approach is used to sample explicitly correlated coupled-cluster electronic energies, and the datapoints are fitted with permutationally invariant polynomials to obtain a spectroscopic-quality PES representation. Variational vibrational energies and corresponding tunnelling splittings are computed up to the first overtone of the C-O stretching mode by direct numerical solution of the vibrational Schr\"odinger equation with optimal internal coordinates and efficient basis and grid truncation techniques. As a result, the computed vibrational band origins finally agree with experiment within 5 cm$^{-1}$, allowing for the exploration of the large-amplitude quantum mechanical motion and tunnelling splittings coupled with the small-amplitude vibrational dynamics. These developments open the route towards simulating rovibrational spectra used to probe methanol in outer space and in precision science laboratories, as well as for probing interactions with external magnetic fields.

Figures

Figures reproduced from arXiv: 2505.16262 by the authors.

Figure 1
Figure 1. FIG. 1. Definition of the primitive internal coordinates of the methanol molecule and the body-fixed [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Variational vibrational band origins computed in this work, with the newly developed PES25, at [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

93 extracted references · 69 canonical work pages

  1. [1]

    Brief description of the operation of Robosurfer The PES development was performed using a development version of theRobosurferprogram system68. Without going into the details of the operation ofRobosurfer, it automates the development of PESs through an iterative, active-learning-like process, which gradually expands 18 the fitting set with additional sa...

    work page

  2. [2]

    The most important heuristics are the geometrical and energetic distances of new geometries from the union of the fitting and spare sets

    A set of heuristics is used to estimate which of the new geometries are the most likely to have large fitting errors on the current PES. The most important heuristics are the geometrical and energetic distances of new geometries from the union of the fitting and spare sets. These heuristics select new geometries that are sufficiently different from points...

    work page

  3. [3]

    While fitting the 14 200 point fitting set with DGELSY did reveal an outlier (Fig

    Process of PES development II: fitting error reduction First, we attempted to improve the numerical rank of the LLS matrix by continuing PES devel- opment with the DGELSY LLS solver, the hypothesis being that discarding the weakly determined monomials could result in higher fitting errors at geometries that would make those monomials more strongly determi...

    work page

  4. [4]

    microiterations

    The fitting error (deviation between theab initioenergy and the value of the fitted PES) is evaluated for every geometry in the spares set. If the geometry with the highest weighted fitting error has an error larger than the configured error target, it is moved to the fitting set, a new PES is fitted, and the errors are reevaluated. These “microiterations...

    work page

  5. [23]

    R. M. Lees, L.-H. Xu, J. W. C. Johns, Z.-F. Lu, B. P. Winnewisser, M. Lock, and R. L. Sams, Fourier transform spectroscopy of CH3OH: rotation–torsion–vibration structure for the CH 3-rocking and OH-bending modes, J. Mol. Spectrosc.228, 528 (2004)

    2004

  6. [24]

    Abbouti Temsamani, L.-H

    M. Abbouti Temsamani, L.-H. Xu, and R. M. Lees, A rotation–torsion–vibration treatment with three-dimensional internal coordinate approach and additional FTIR spectral assignments for the CH3-bending fundamentals of methanol, J. Mol. Spectrosc.218, 220 (2003)

    2003

  7. [25]

    L.-H. Xu, J. Fisher, R. M. Lees, H. Y. Shi, J. T. Hougen, J. C. Pearson, B. J. Drouin, G. A. Blake, and R. Braakman, Torsion–rotation global analysis of the first three torsional states (ν t = 0,1,2) and terahertz database for methanol, J. Mol. Spectrosc.251, 305 (2008)

    2008

  8. [26]

    Santagata, D

    R. Santagata, D. B. A. Tran, B. Argence, O. Lopez, S. K. Tokunaga, F. Wiotte, H. Mouhamad, A. Goncharov, M. Abgrall, Y. Le Coq, H. Alvarez-Martinez, R. Le Targat, W. K. Lee, D. Xu, P.-E. Pottie, B. Darqui´ e, and A. Amy-Klein, High-precision methanol spectroscopy with a widely tunable SI-traceable frequency-comb-based mid-infrared QCL, Optica6, 411 (2019)

    2019

Show all 93 references
  1. [27]

    P. R. Bunker and P. Jensen,Molecular Symmetry and Spectroscopy, 2nd Ed(NRC Research Press, 2006)

    2006

  2. [28]

    Quack, Detailed symmetry selection rules for reactive collisions, Mol

    M. Quack, Detailed symmetry selection rules for reactive collisions, Mol. Phys.34, 477 (1977)

    1977

  3. [29]

    Quack and F

    M. Quack and F. Merkt,Handbook of High-resolution Spectroscopy(John Wiley & Sons, 2011)

    2011

  4. [30]

    J. P. Perchard, The torsion-vibration spectrum of methanol trapped in neon matrix, Chem. Phys. 332, 86 (2007)

    2007

  5. [31]

    J. P. Perchard, F. Romain, and Y. Bouteiller, Determination of vibrational parameters of methanol from matrix-isolation infrared spectroscopy and ab initio calculations. part 1 – spectral analysis in 13 the domain 11000–200 cm −1, Chem. Phys.343, 35 (2008)

    2008

  6. [32]

    D. F. Dinu, K. Oenen, J. Schlagin, M. Podewitz, H. Grothe, T. Loerting, and K. R. Liedl, How vibrational notations can spoil infrared spectroscopy: A case study on isolated methanol, ACS Phys. Chem. Au4, 679 (2024)

    2024

  7. [33]

    J. M. Bowman, Self-consistent field energies and wavefunctions for coupled oscillators, J. Chem. Phys. 68, 608 (1978)

    1978

  8. [34]

    K. Yagi, K. Hirao, T. Taketsugu, M. W. Schmidt, and M. S. Gordon, Ab initio vibrational state calculations with a quartic force field: applications to H 2CO, C 2H4, CH 3OH, CH 3CCH, and C 6H6, J. Chem. Phys.121, 1383 (2004)

    2004

  9. [35]

    Partal Ure˜ na, J

    F. Partal Ure˜ na, J. J. L´ opez Gonz´ alez, and F. M´ arquez, Anharmonic spectra of methanol and silanol: A comparative study, J. Mol. Spectrosc.233, 203 (2005)

    2005

  10. [36]

    Scribano, D

    Y. Scribano, D. M. Lauvergnat, and D. M. Benoit, Fast vibrational configuration interaction using generalized curvilinear coordinates and self-consistent basis, J. Chem. Phys.133, 094103 (2010)

    2010

  11. [37]

    Schr¨ oder and G

    B. Schr¨ oder and G. Rauhut, Vibrational configuration interaction theory, inVibrational Dynamics of Molecules(WORLD SCIENTIFIC, 2022) pp. 1–40

    2022

  12. [38]

    D. F. Dinu, M. Podewitz, H. Grothe, T. Loerting, and K. R. Liedl, On the synergy of matrix-isolation infrared spectroscopy and vibrational configuration interaction computations, Theor. Chem. Acc. 139, 174 (2020)

    2020

  13. [39]

    H. R. Larsson, Benchmarking vibrational spectra: 5000 accurate eigenstates of acetonitrile using tree tensor network states, J. Phys. Chem. Lett.16, 3991 (2025)

    2025

  14. [40]

    A. G. Cs´ asz´ ar, C. F´ abri, T. Szidarovszky, E. M´ atyus, T. Furtenbacher, and G. Czak´ o, The fourth age of quantum chemistry: molecules in motion, Phys. Chem. Chem. Phys.14, 1085 (2012)

    2012

  15. [41]

    Tennyson, Perspective: Accurate ro-vibrational calculations on small molecules, J

    J. Tennyson, Perspective: Accurate ro-vibrational calculations on small molecules, J. Chem. Phys. 145, 120901 (2016)

    2016

  16. [42]

    Carrington, Jr, Perspective: Computing (ro-)vibrational spectra of molecules with more than four atoms, J

    T. Carrington, Jr, Perspective: Computing (ro-)vibrational spectra of molecules with more than four atoms, J. Chem. Phys.146, 120902 (2017)

    2017

  17. [43]

    A. Chen, A. Nauts, and D. Lauvergnat, ElVibRot-MPI: parallel quantum dynamics with smolyak algorithm for general molecular simulation, arXiv [physics.comp-ph] (2021), arXiv:2111.13655 [physics.comp-ph]

    2021 arXiv

  18. [44]

    M´ atyus, A

    E. M´ atyus, A. Mart ´ ın Santa Dar ´ ıa, and G. Avila, Exact quantum dynamics developments for floppy molecular systems and complexes, Chem. Commun.59, 366 (2023)

    2023

  19. [45]

    Simmons and T

    J. Simmons and T. Carrington, Jr, Computing vibrational spectra using a new collocation method with a pruned basis and more points than basis functions: Avoiding quadrature, J. Chem. Phys.158, 144115 (2023)

    2023

  20. [46]

    Wodraszka and T

    R. Wodraszka and T. Carrington, Jr, Using a pruned basis and a sparse collocation grid with more points than basis functions to do efficient and accurate MCTDH calculations with general potential energy surfaces, J. Chem. Phys.160, 214121 (2024)

    2024

  21. [47]

    S. D. Kallullathil and T. Carrington, Computing vibrational energy levels using a canonical polyadic tensor method with a fixed rank and a contraction tree, J. Chem. Phys.158, 214102 (2023)

    2023

  22. [48]

    Simk´ o, P

    I. Simk´ o, P. M. Felker, and Z. Baˇ ci´ c, HCl trimer: HCl-stretch excited intramolecular and intermolec- ular vibrational states from 12D fully coupled quantum calculations employing contracted intra- and inter-molecular bases, J. Chem. Phys.160, 164304 (2024)

    2024

  23. [49]

    Simk´ o, P

    I. Simk´ o, P. Felker, and Z. Baˇ ci´ c, H2O trimer: Rigorous 12D quantum calculations of intermolecular vibrational states, tunneling splittings, and low-frequency spectrum, J. Chem. Phys.162, 034301 (2025)

    2025

  24. [50]

    Lauvergnat and A

    D. Lauvergnat and A. Nauts, Quantum dynamics with sparse grids: a combination of smolyak scheme and cubature. application to methanol in full dimensionality, Spectrochim. Acta A Mol. Biomol. Spectrosc.119, 18 (2014)

    2014

  25. [51]

    Nauts and D

    A. Nauts and D. Lauvergnat, Numerical on-the-fly implementation of the action of the kinetic energy operator on a vibrational wave function: application to methanol, Mol. Phys.116, 3701 (2018)

    2018

  26. [52]

    Avila and E

    G. Avila and E. M´ atyus, Toward breaking the curse of dimensionality in (ro)vibrational computations of molecular systems with multiple large-amplitude motions, J. Chem. Phys.150, 174107 (2019)

    2019

  27. [53]

    Avila and E

    G. Avila and E. Matyus, Full-dimensional (12d) variational vibrational states of CH 4F−: Interplay of anharmonicity and tunneling, J. Chem. Phys.151, 154301 (2019)

    2019

  28. [54]

    Avila, D

    G. Avila, D. Papp, G. Czak´ o, and E. M´ atyus, Exact quantum dynamics background of dispersion interactions: case study for CH 4·Ar in full (12) dimensions, Phys. Chem. Chem. Phys.22, 2792 (2020)

    2020

  29. [55]

    D. Papp, V. Tajti, G. Avila, E. M´ atyus, and G. Czak´ o, CH4·F− revisited: full-dimensional ab initio potential energy surface and variational vibrational states, Mol. Phys.121, e2113565 (2023)

    2023

  30. [56]

    Mart ´ ın Santa Dar ´ ıa, G

    A. Mart ´ ın Santa Dar ´ ıa, G. Avila, and E. M´ atyus, Variational vibrational states of HCOOH, J. Mol. Spectrosc.385, 111617 (2022)

    2022

  31. [57]

    Avila, A

    G. Avila, A. Mart ´ ın Santa Dar ´ ıa, and E. M´ atyus, Vibrational infrared and raman spectra of HCOOH from variational computations, Phys. Chem. Chem. Phys.25, 15183 (2023)

    2023

  32. [58]

    Sunaga, G

    A. Sunaga, G. Avila, and E. M´ atyus, Variational vibrational states of methanol (12D), J. Chem. 14 Theory Comput.20, 8100 (2024)

    2024

  33. [59]

    Qu and J

    C. Qu and J. M. Bowman, Full-dimensional, ab initio potential energy surface for CH 3OH→ CH3+OH, Mol. Phys.111, 1964 (2013)

    2013

  34. [60]

    Huang, B

    X. Huang, B. J. Braams, and J. M. Bowman, Ab initio potential energy and dipole moment surfaces for H5O+ 2 , J. Chem. Phys.122, 44308 (2005)

    2005

  35. [61]

    B. J. Braams and J. M. Bowman, Permutationally invariant potential energy surfaces in high dimen- sionality, Int. Rev. Phys. Chem.28, 577 (2009)

    2009

  36. [62]

    C. Qu, Q. Yu, and J. M. Bowman, Permutationally invariant potential energy surfaces, Annual Review of Physical Chemistry69, 151 (2018)

    2018

  37. [63]

    C. Qu, Q. Yu, B. L. Van Hoozen, Jr, J. M. Bowman, and R. A. Vargas-Hern´ andez, Assessing gaussian process regression and permutationally invariant polynomial approaches to represent high- dimensional potential energy surfaces, J. Chem. Theory Comput.14, 3381 (2018)

    2018

  38. [64]

    D. R. Moberg and A. W. Jasper, Permutationally invariant polynomial expansions with unrestricted complexity, J. Chem. Theory Comput.17, 5440 (2021)

    2021

  39. [65]

    P. L. Houston, C. Qu, Q. Yu, R. Conte, A. Nandi, J. K. Li, and J. M. Bowman, PESPIP: Software to fit complex molecular and many-body potential energy surfaces with permutationally invariant polynomials, J. Chem. Phys.158, 044109 (2023)

    2023

  40. [66]

    M. S. Drehwald, A. Jamali, and R. A. Vargas-Hern´ andez, MOLPIPx: An end-to-end differentiable package for permutationally invariant polynomials in python and rust, J. Chem. Phys.162, 084115 (2025), 2411.17011

    2025 arXiv

  41. [67]

    Bowman, C

    J. Bowman, C. Qu, R. Conte, A. Nandi, P. Houston, and Q. Yu, A perspective marking 20 years of using permutationally invariant polynomials for molecular potentials, ChemRxiv; doi:10.26434/chemrxiv-2025-v14p7 10.26434/chemrxiv-2025-v14p7 (2025)

    2025 doi

  42. [68]

    Gy˝ ori and G

    T. Gy˝ ori and G. Czak´ o, Automating the development of high-dimensional reactive potential energy surfaces with theRobosurferprogram system, J. Chem. Theory Comput.16, 51 (2020)

    2020

  43. [69]

    T. B. Adler, G. Knizia, and H.-J. Werner, A simple and efficient CCSD(T)-F12 approximation, J. Chem. Phys.127, 221106 (2007)

    2007

  44. [70]

    K. A. Peterson, T. B. Adler, and H.-J. Werner, Systematically convergent basis sets for explicitly correlated wavefunctions: the atoms H, He, B-Ne, and Al-Ar, J. Chem. Phys.128, 084102 (2008)

    2008

  45. [71]

    Lindh, U

    R. Lindh, U. Ryu, and B. Liu, The reduced multiplication scheme of the rys quadrature and new recurrence relations for auxiliary function based two-electron integral evaluation, J. Chem. Phys.95, 5889 (1991)

    1991

  46. [72]

    Werner, P

    H.-J. Werner, P. J. Knowles, G. Knizia, F. R. Manby, and M. Sch¨ utz, Molpro: a general-purpose quantum chemistry program package, Wiley Interdiscip. Rev. Comput. Mol. Sci.2, 242 (2012)

    2012

  47. [73]

    Werner, P

    H.-J. Werner, P. J. Knowles, F. R. Manby, J. A. Black, K. Doll, A. Heßelmann, D. Kats, A. K¨ ohn, T. Korona, D. A. Kreplin, Q. Ma, T. F. Miller, 3rd, A. Mitrushchenkov, K. A. Peterson, I. Polyak, G. Rauhut, and M. Sibaev, The molpro quantum chemistry package, J. Chem. Phys.152...

    2020

  48. [74]

    Werner, P

    H.-J. Werner, P. J. Knowles, P. Celani, W. Gy¨ orffy, A. Hesselmann, D. Kats, G. Knizia, A. K¨ ohn, T. Korona, D. Kreplin, R. Lindh, Q. Ma, F. R. Manby, A. Mitrushenkov, G. Rauhut, M. Sch¨ utz, K. R. Shamasundar, T. B. Adler, R. D. Amos, S. J. Bennie, A. Bernhardsson, A. Berni...

    2023

  49. [75]

    Anderson, Z

    E. Anderson, Z. Bai, C. Bischof, S. Blackford, J. Demmel, J. Dongarra, J. Du Croz, A. Green- baum, S. Hammarling, A. McKenney, and D. Sorensen,LAPACK Users’ Guide, 3rd ed. (Society for Industrial and Applied Mathematics, Philadelphia, PA, 1999)

    1999

  50. [76]

    Meyer and H

    R. Meyer and H. H. G¨ unthard, Internal rotation and vibration in CH2=CCl–CH2D, J. Chem. Phys. 50, 353 (1969)

    1969

  51. [77]

    Bell, Variation of geometry, vibrational frequencies and zero-point energies with internal rotation, J

    S. Bell, Variation of geometry, vibrational frequencies and zero-point energies with internal rotation, J. Mol. Struct.320, 125 (1994)

    1994

  52. [78]

    Avila and T

    G. Avila and T. Carrington, Jr, Nonproduct quadrature grids for solving the vibrational schrodinger equation, J. Chem. Phys.131, 174103 (2009)

    2009

  53. [79]

    Avila and T

    G. Avila and T. Carrington, Jr, Using nonproduct quadrature grids to solve the vibrational schrodinger equation in 12D, J. Chem. Phys.134, 054126 (2011)

    2011

  54. [80]

    M´ atyus, G

    E. M´ atyus, G. Czak´ o, and A. G. Cs´ asz´ ar, Toward black-box-type full- and reduced-dimensional variational (ro)vibrational computations, J. Chem. Phys.130, 134112 (2009)

    2009

  55. [81]

    M. H. Cortez, N. R. Brinkmann, W. F. Polik, P. R. Taylor, Y. J. Bomble, and J. F. Stanton, Factors contributing to the accuracy of harmonic force field calculations for water, J. Chem. Theory Comput. 15 3, 1267 (2007)

    2007

  56. [82]

    D. P. Tew, W. Klopper, M. Heckert, and J. Gauss, Basis set limit CCSD(T) harmonic vibrational frequencies, J. Phys. Chem. A111, 11242 (2007)

    2007

  57. [83]

    Puzzarini, M

    C. Puzzarini, M. Heckert, and J. Gauss, The accuracy of rotational constants predicted by high-level quantum-chemical calculations. I. molecules containing first-row atoms, J. Chem. Phys.128, 194108 (2008)

    2008

  58. [84]

    Das and G

    S. Das and G. Rauhut, Rovibrational calculations without model hamiltonians: The infrared and microwave spectra of thiopropynal, Int. J. Quantum Chem.124, e27378 (2024)

    2024

  59. [85]

    Rauhut, G

    G. Rauhut, G. Knizia, and H.-J. Werner, Accurate calculation of vibrational frequencies using ex- plicitly correlated coupled-cluster theory, J. Chem. Phys.130, 054105 (2009)

    2009

  60. [86]

    Pyykk¨ o, K

    P. Pyykk¨ o, K. G. Dyall, A. G. Cs´ asz´ ar, G. Tarczay, O. L. Polyansky, and J. Tennyson, Estimation of Lamb-shift effects for molecules: Application to the rotation-vibration spectra of water, Phys. Rev. A63, 024502 (2001)

    2001

  61. [87]

    Polyansky, A

    O. Polyansky, A. Cs´ asz´ ar, S. Shirin, N. Zobov, P. Barletta, J. Tennyson, D. Schwenke, and P. Knowles, High-accuracyab initiorotation-vibration transitions for water, Science299, 539 (2003)

    2003

  62. [88]

    Temelso, E

    B. Temelso, E. F. Valeev, and C. David Sherrill, A comparison of one-particle basis set complete- ness, higher-order electron correlation, relativistic effects, and adiabatic corrections for spectroscopic constants of BH, CH +, and NH, J. Phys. Chem. A108, 3068 (2004)

    2004

  63. [89]

    Piszczatowski, G

    K. Piszczatowski, G. Lach, M. Przybytek, J. Komasa, K. Pachucki, and B. Jeziorski, Theoretical determination of the dissociation energy of molecular hydrogen, J. Comput. Theo. Chem.5, 3039 (2009)

    2009

  64. [90]

    Si lkowski, K

    M. Si lkowski, K. Pachucki, J. Komasa, and M. Puchalski, Leading-order QED effects in the ground electronic state of molecular hydrogen, Phys. Rev. A107, 032807 (2023)

    2023

  65. [91]

    Pachucki and J

    K. Pachucki and J. Komasa, Relativistic correction from the four-body nonadiabatic ex- ponential wave function, J. Chem. Theory. Comnput.20, 8644 (2024), pMID: 39327784, https://doi.org/10.1021/acs.jctc.4c00861

    2024 doi

  66. [92]

    Ferenc, V

    D. Ferenc, V. I. Korobov, and E. M´ atyus, Nonadiabatic, relativistic, and leading-order QED correc- tions for rovibrational intervals of 4He + 2 (x 2Σ + u ), Phys. Rev. Lett.125, 213001 (2020)

    2020

  67. [93]

    Marg´ ocsy, B

    ´A. Marg´ ocsy, B. R´ acsai, P. Jeszenszki, and E. M´ atyus, Rovibrational computations for the He2 a 3Σ+ u state including non-adiabatic, relativistic, and QED corrections, (under review) arXiv:2506.19116 ((2025))

    2025

  68. [94]

    Moruzzi, B

    G. Moruzzi, B. P. Winnewisser, M. Winnewisser, I. Mukhopadhyay, and F. Strumia,Microwave, Infrared, and Laser Transitions of Methanol Atlas of Assigned Lines from 0 to 1258 cm −1 (CRC New York, 1995)

    1995

  69. [95]

    Fehrensen, D

    B. Fehrensen, D. Luckhaus, M. Quack, M. Willeke, and T. R. Rizzo, Ab initio calculations of mode selective tunneling dynamics in C 12H3OH and C 13H3OH, J. Chem. Phys.119, 5534 (2003)

    2003

  70. [96]

    Weigend, A fully direct RI-HF algorithm: Implementation, optimised auxiliary basis sets, demon- stration of accuracy and efficiency, Phys

    F. Weigend, A fully direct RI-HF algorithm: Implementation, optimised auxiliary basis sets, demon- stration of accuracy and efficiency, Phys. Chem. Chem. Phys.4, 4285 (2002)

    2002

  71. [97]

    The aug-cc-pVnZ-JKfit series of auxiliary basis sets are unpublished, but included in the basis set library of Molpro 2023. According to their brief description (found in weigend jkfit aug.libmol), these augmented JKfit basis sets were generated by them from cc-pVnZ-JKfit by a...

    2023

  72. [98]

    K. E. Yousaf and K. A. Peterson, Optimized auxiliary basis sets for explicitly correlated methods, J. Chem. Phys.129, 184108 (2008)

    2008

  73. [99]

    Weigend, A

    F. Weigend, A. K¨ ohn, and C. H¨ attig, Efficient use of the correlation consistent basis sets in resolution of the identity MP2 calculations, J. Chem. Phys.116, 3175 (2002)

    2002

  74. [100]

    Eckert, P

    F. Eckert, P. Pulay, and H. Werner, Ab initio geometry optimization for large molecules, J. Comput. Chem.18, 1473 (1997)

    1997

  75. [101]

    This is detected based on the energy of the PES at the geometry in question, if the PES returns an energy lower thanE slopeStart a hole is assumed

    Geometries with anab initioenergy higher than the semihard limit are not rejected outright, but they are only considered for moving into the fitting set if they are required for fixing deep unphysical minima (holes) on the fitted PES. This is detected based on the energy of th...

  76. [102]

    Available at https://www.intel.com/content/www/us/en/developer/tools/oneapi/onemkl.html

  77. [103]

    Xianyi, W

    Z. Xianyi, W. Qian, and Z. Yunquan, Model-driven level 3 blas performance optimization on loongson 3a processor, in2012 IEEE 18th International Conference on Parallel and Distributed Systems(2012) pp. 684–691

    2012

  78. [104]

    Q. Wang, X. Zhang, Y. Zhang, and Q. Yi, Augem: automatically generate high performance dense linear algebra kernels on x86 cpus, inProceedings of the International Conference on High Perfor- mance Computing, Networking, Storage and Analysis, SC ’13 (Association for Computing M...

    2013

  79. [105]

    Available at https://github.com/OpenMathLib/OpenBLAS. 16

  80. [106]

    Momentum

    T. Gy˝ ori and G. Czak´ o, ManyHF: A pragmatic automated method of finding lower-energy Hartree- Fock solutions for potential energy surface development, J. Chem. Phys.156, 071101 (2022). 17 Supplemental Material Exact quantum dynamics of methanol: full-dimensionalab initio po...

    2022

  81. [107]

    by running quasi-classical trajectories (QCT) and geometry optimizations

    The PES is fitted based on the geometries in the fitting set, and large numbers of new geometries are rapidly generated using the PES, e.g. by running quasi-classical trajectories (QCT) and geometry optimizations

  82. [108]

    range separation

    GeneralROBOSURFERconfiguration and fitting methodology The PES is fitted using the Braams–Bowman implementation of permutationally invariant poly- nomials61. Briefly, the polynomials are expanded in terms of Morse variables,y ij = exp(−rij/a), wherer ij is the distance between...

  83. [109]

    Preparation of the initial fitting set Robosurferrequires an initial fitting set to begin PES development, this was generated using a version of the geometry set generation code first described in Ref. 106. As this program was originally intended for generating datasets of two...

  84. [110]

    absolute

    Process of PES development I: initial phase Over the course of PES development we have taken the approach of starting with a small fitting set, low degree polynomials, and a high fitting error target, then gradually increasing both the flexibility of the fitting function and t...

  85. [111]

    Here, we noticed an outlier in the lower energy region (Fig

    Process of PES development III: finalization With the final fitting function and parameters established,Robosurferiterations were con- tinued from the 26 401 geometry fitting set withE targ = 0.075 kcal/mol (26.2 cm −1) and in 17 iterations 2 000 points were added to the fitti...

Pith tools

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