Pith. sign in

REVIEW 4 major objections 6 minor 55 references

Deep Potential-Driven Molecular Dynamics of CO Ice Analogues: Investigating Desorption Following Vibrational Excitation

T0 review · 4 major / 6 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 A deep learning potential trained on ab initio data reproduces the desorption of CO from an ice aggregate and, with 11,000 trajectories, exposes a nearly linear rotational-translational energy coupling.

desk verdict Solid, reusable MLP study with a clever excited-molecule trick; the validation gap and yield mismatch matter, but the paper deserves review. read the letter →

arxiv 2506.10882 v1 pith:U5A3XNIP submitted 2025-06-12 physics.chem-ph

Maxime Infuso , Samuel Del Fré , Gilberto A. Alou , Mathieu Bertin , Jean-Hugues Fillion , Alejandro Rivero Santamaría , Maurice Monnerville This is my paper · ORCID
classification physics.chem-ph
keywords machinelearningpotentialCOicephotodesorptionvibrationalrelaxationdesorptiondynamicsinterstellaricesdeeptranslational-rotationalcoupling
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 claims that a machine-learned interatomic potential, trained on data from 100 ab initio molecular dynamics trajectories, can reproduce the full desorption of a carbon monoxide molecule from an aggregate of 50 CO molecules after vibrational excitation to $v=40$, matching prior theory and experiment. The key trick is to give the two atoms of the vibrationally excited CO their own dedicated neural networks, separate from the 49 ground-state molecules, so that the potential can represent the strongly anharmonic high-energy stretch without destabilizing the dynamics. With this potential the authors run 11,000 independent trajectories, more than 100 times the previous ab initio sample, and obtain smooth translational and rotational energy distributions of the desorbed molecules. The finer statistics reveal that both the most probable and the maximum rotational quantum number, $J_{\mathrm{peak}}$ and $J_{\mathrm{max}}$, rise nearly linearly with translational energy, evidence of coupled translational–rotational energy flow during desorption. If the claim holds, it makes systematically sampled simulations of interstellar ice photodesorption affordable.

What carries the argument

The central object is the machine-learned interatomic potential used to run the production dynamics. It is an atom-centered high-dimensional neural network potential in which the total energy is the sum of atomic energy contributions, each computed by a descriptor and a fitting network; the paper's innovation is to give the two atoms of the vibrationally excited CO molecule separate networks, X1 and X2, so that the strongly anharmonic potential-energy region of the stretched bond is fitted independently of the ground-state molecules. This separation is what allows the potential to stay accurate up to the $v=40$ energy (8.26 eV) for the hot molecule while the surrounding molecules are described by the standard carbon and oxygen networks. The potential reproduces the DFT dimer dissociation curve up to 8.26 eV for X1–X2 and up to about 6.5 eV for the ordinary CO networks, and the production trajectories are started from the same kind of thermalized aggregate ensemble and integrated with the same 0.1 fs time step as the prior AIMD study, making the statistical comparison direct.

What would settle it

Run a set of new AIMD trajectories from the same 11,000 initial conditions (or from a large random subset) and compare the desorption yield, desorption times, and the $J_{\mathrm{peak}}$ versus translational-energy trend with the MLP predictions; a statistically significant disagreement in any of these observables would falsify the claim that the potential generalizes. In addition, one could compute DFT forces directly for configurations sampled from the MLP trajectories in the late desorption phase, where the escaping molecule is far from the aggregate, and look for systematic force errors that would bias the desorption outcome.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes that a neural network potential of the atom-centered type can describe, at near-DFT accuracy, the relaxation and desorption dynamics that follow depositing 8.26 eV of vibrational energy into one CO molecule inside a 50-molecule CO ice aggregate. The authors train the potential on configurations sampled from 100 DFT-based AIMD trajectories, and they assign separate neural networks, labelled X1 and X2, to the carbon and oxygen atoms of the excited molecule so that the short, strongly anharmonic C–O bond of the hot molecule is not handled by the same parameters as the bonds of the cold surroundings. In 11,000 microcanonical trajectories started from a 15 K thermal ensemble, 77% of excitations lead to desorption, close to the 88% found in AIMD, with an average desorption time of about 2 ps and no desorption before 600 fs. The desorbed CO is predominantly $v=0$ (about 5% in $v=1$), and the translational distribution peaks near 25 meV and falls to zero above about 800 meV, reproducing the experimental time-of-flight data. With the improved sampling, rotational distributions can be resolved in narrow translational windows, and they show $J_{\mathrm{peak}}$ increasing from 4 to 8 and $J_{\mathrm{max}}$ from 25 to 44 as the translational energy interval moves from 0–50 meV to 150–200 meV; the paper interprets this near-linear rise as continuous, coupled rotational–translational energy redistribution during desorption.

Load-bearing premise

The neural network potential, trained on configurations sampled from 100 AIMD trajectories of the same desorption process, generalizes accurately to the full ensemble of 11,000 independent initial conditions, including the regions where the excited molecule and the desorbing molecule move far from the aggregate.

Editorial extensions

If this is right

  • The roughly 1500-fold lower cost per energy-and-force evaluation makes statistically converged trajectory ensembles for desorption studies feasible, something AIMD cannot deliver.
  • The method can be transferred to other localized excitations in condensed-phase molecular systems, as the X1/X2 separation is general.
  • The resolved rotational distributions provide a quantitative map of how vibrational relaxation energy is partitioned into translation and rotation of the desorbing molecule.
  • The near-linear $J_{\mathrm{peak}}$ and $J_{\mathrm{max}}$ versus translational energy relationship offers a testable signature of the indirect desorption mechanism in future experiments.

Reading between the lines

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

  • If the potential generalizes beyond the $v=40$ pathway it was trained on, the same architecture could scan other initial vibrational levels, ice temperatures, and aggregate sizes at essentially no additional ab initio cost; the paper does not demonstrate this.
  • The linear rotational-translational correlation found for CO may be a generic property of collision-mediated desorption from weakly bound molecular ices; testing on N2, O2, or CO2 ices would indicate whether the mechanism is universal.
  • The X1/X2 separation is effectively a 'tagged molecule' trick; it could be reused to follow energy flow out of any selected excited species in a multichromophore system, not just CO.
  • The 77% versus 88% desorption yield difference may be within statistical scatter, but it could also mark a small bias in the potential's rare-event regions; an independent MLP retrained from different random weights, or more AIMD trajectories, would settle this.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 presents a deep learning-based machine learning potential (MLP) for solid CO, constructed with DeePMD-kit, in which the vibrationally excited CO molecule is treated by two distinct atom types (X1 and X2). The MLP is trained on approximately 18,000 configurations drawn from 100 AIMD trajectories of vibrational relaxation from v=40 in a 50-molecule CO aggregate. Validation shows low energy RMSE (0.075 eV) and moderate force RMSE for the X1/X2 atoms, plus good agreement with a DFT dissociation curve. The authors then run 11,000 LAMMPS MD trajectories with the MLP, reporting a 77% desorption yield (compared to 88% in the prior AIMD study), smooth translational and rotational energy distributions that agree with experimental data, and a new result: a near-linear increase of J_peak and J_max with translational energy. The paper claims the MLP accurately describes the desorption process and provides a significant improvement in statistical sampling over AIMD.

Significance. If the validation concerns are resolved, this work offers a valuable methodological contribution: the selective X1/X2 treatment of a highly excited molecule is a promising strategy for extending MLPs to non-equilibrium, strongly anharmonic dynamics in condensed phases. The 11,000-trajectory ensemble provides a substantial gain in statistical power, enabling distribution comparisons with experiment that were previously impractical with AIMD. The agreement of the MLP rotational and translational distributions with experimental data is encouraging and supports the underlying desorption mechanism. However, the central claim of accuracy is currently weakened by the in-sample validation and by an unexplained 11 percentage-point discrepancy in the desorption yield relative to AIMD. The new linear-correlation finding also rests on only four points without statistical treatment for J_max. These issues need to be addressed before the results can be considered fully supported.

major comments (4)
  1. [III.B first paragraph] The MLP desorption yield (77%) differs from the AIMD yield (88%) by 11 percentage points. With 100 AIMD trajectories, the standard error of the AIMD yield is about 3.3% (sqrt(0.88*0.12/100)), and with 11,000 MLP trajectories the MLP yield has a standard error below 0.5%, so the difference is roughly 3 standard deviations. The text describes this as 'close' but provides no statistical comparison or discussion. Because the desorption yield is a primary observable of the process, this discrepancy may indicate a systematic bias in the MLP, a difference in initial-condition sampling, or a real physical difference. Please either add confidence intervals and a formal comparison, or run AIMD on a subset of the same initial conditions used in the MLP trajectories to discriminate between these possibilities.
  2. [III.A and II.A] All validation checks—energy parity (Fig. 2), force parity (Figs. 3-5), and the dissociation curve (Fig. 6)—are performed on configurations drawn from the same 100 AIMD trajectories that supplied the training set. The 11,000 production trajectories start from a new thermalization ensemble and evolve over 5 ps, generating configurations (including desorbed molecules far from the aggregate) that are not part of the validation set. The paper does not demonstrate that the MLP remains accurate on these out-of-training configurations. Please validate the MLP on a hold-out set of AIMD trajectories not used in training, or re-evaluate a sample of MLP-generated configurations with DFT and compare energies and forces, or at minimum discuss the in-sample nature of the validation and its implications for the reported uncertainties on desorption yields and energy distributions.
  3. [III.C, Figure 10] The claim that J_peak and J_max increase 'nearly linearly' with translational energy is based on only four data points for each quantity. J_max is defined by a 3%-of-maximum threshold that is sensitive to sampling noise, yet no uncertainty is given for J_max; the error bars shown for J_peak are not tied to a regression. Without a linear fit with confidence intervals, or bootstrap estimates for J_max, the linear trend is not quantitatively supported. Please provide uncertainties for J_max and perform a proper regression (including R² and parameter uncertainties), or soften the claim to a qualitative observation.
  4. [III.A, Figures 3-4] The X1/X2 force RMSE values (0.186 and 0.154 eV/Å) are roughly four times larger than the C/O values (0.045 and 0.038 eV/Å). The authors argue that this is acceptable because the absolute forces on X1/X2 are larger, but the excited molecule drives the entire energy-transfer process, so a systematic force error on these atoms could directly bias the desorption yield and the final energy distributions. The paper does not quantify how these force errors propagate into the observables. Please add a sensitivity test—for example, running a few MLP trajectories with a different MLP seed, or comparing a handful of MLP trajectories against AIMD from the same initial conditions—to estimate the impact of X1/X2 force errors on the reported results.
minor comments (6)
  1. [Throughout] There are several typographical errors: 'pannel' should be 'panel' in the captions of Figs. 8, 9, and 10; 'corelation' should be 'correlation' in Section III.A; 'mentionned' should be 'mentioned' in Section II.A; and 'the later' should be 'the latter' in the same section.
  2. [II.A] The dataset sampling intervals overlap: the first interval is 'until t=500 fs', the second is 'between 100 fs and 2500 fs', and the third is 'above 2500 fs'. Clarify the boundary times (e.g., whether 100 fs and 2500 fs belong to two intervals) to avoid ambiguity.
  3. [Figure 2] The colorbar label 'Absolute Error (eV)' appears truncated in the displayed figure; please check that all axis and colorbar labels are fully visible.
  4. [III.C, Figure 8 caption] The phrase 'High values distributions are computed for translational energies lower than 200 meV' is unclear; please rephrase to specify the energy window used for the high-translational-energy panel.
  5. [II.B] The procedure for setting the internal momentum to v=40 is not described in detail (e.g., whether the bond length and velocities are set deterministically or sampled from a distribution). Please specify the exact method for reproducibility.
  6. [Data availability] The paper does not state whether the trained MLP weights or the trajectory data will be made available. If the journal's policy requires data/code availability, please add a statement.

Circularity Check

2 steps flagged · score 4.0 of 10

MLP is validated largely against the same AIMD trajectories used to train it; the 77% vs 88% desorption-yield comparison is a self-consistency check, though the Jpeak/Jmax trend emerges from dynamics and retains partial independent content.

  1. fitted input called prediction [Section II.A (Neural network potential energy surface) and Section III.B (Molecular Dynamics Simulations)]
    "The dataset employed for training and validating the MLP was derived and expanded from previous research 25. ... a total of 100 trajectories, among which desorption was observed in 88 % of the cases. ... Among these 11,000 trajectories, 77 % resulted in the desorption of CO molecule(s). This result is close to that predicted by our previous AIMD calculations25 (88 %) , demonstrating that the new MLP provides an accurate and reliable representation of the system’s multidimensional potential energy landscape."

    The 88% desorption yield is measured on the same 100 AIMD trajectories whose configurations (18,000 sampled) form the training set. The MLP is a fit to those data; the 11,000-trajectory MD then samples the fitted potential, so the resulting 77% yield and the translational/rotational distributions are reproductions of the training distribution, not predictions of a disjoint process. The held-out 20% test set is drawn from the same trajectories, so no out-of-process validation exists. The agreement is a consistency check, not confirmation.

  2. self citation load bearing [Abstract and Section IV (Conclusion)]
    "In particular, the MLP is capable of accurately describing the desorption process of a single CO molecule within an aggregate of 50 CO molecules, in excellent agreement with both previous theoretical predictions and experimental measurements. ... The MLP MD simulations accurately reproduced the high desorption yield of CO molecules, in excellent agreement with previous AIMD predictions."

    The 'previous theoretical predictions' are refs. 25 and 26, both from the same laboratory group (Del Fré, Bertin, Fillion, Rivero Santamaría, Monnerville). Ref. 25 is the source of the training AIMD data; ref. 26 is the companion experimental/theoretical paper. Invoking agreement with these as evidence of accuracy is load-bearing self-citation: the previous theoretical predictions are the very simulations used to generate the training set. The experimental data provide an external anchor, but the theoretical part of the validation is self-referential.

full rationale

The paper's load-bearing validation of the MLP rests on agreement with prior AIMD simulations by the same group (ref. 25) whose configurations formed the training set. This is a genuine in-sample/reproduction issue: the 'prediction' of 77% desorption yield among 11,000 MLP trajectories is compared to 88% from the training-source AIMD, and the distributions in Figs. 7-8 are compared to the same reference. No independent process-level test is provided (e.g., different aggregate sizes, temperatures, excitation levels, or trajectories held out at the process level). However, the MLP is a full-dimensional PES fit to DFT energies and forces; the final desorption yield and energy distributions are not direct fit parameters, so the agreement is not forced by construction. The experimental data (even with overlapping authors) provide an external anchor, and the Jpeak/Jmax trend in Fig. 10 is a refined emergent quantity not present in the training labels. Therefore partial circularity: the central accuracy claim is substantially supported by agreement with its own training source, but independent content remains. Score 4.

Assumptions & free parameters 5 free parameters · 6 assumptions · 1 invented entities

The central claim depends on the accuracy of a MLP that is itself a fit to DFT data from the same authors' prior simulations. The analysis also depends on several hand-chosen thresholds and bin widths. The invented X1/X2 labels are a modeling device, not a new physical entity.

free parameters (5)
  • MLP neural network weights = not listed (tens of thousands of weights)
    Fitted to 18,000 DFT configurations from AIMD trajectories; this is the central data fit that defines the potential energy surface.
  • desorption distance criterion = 3 Å
    Molecule considered desorbed when its center of mass exceeds the aggregate surface by 3 Å; chosen by hand and affects the desorption yield.
  • Jmax tail definition threshold = 3% of KDE maximum
    The highest populated J is defined as where the KDE density falls to 3% of its peak; this ad hoc choice affects the reported Jmax values.
  • translational energy bin widths = 50 meV
    Rotational distributions are presented in 0-50, 50-100, 100-150, 150-200 meV bins; these bins determine the four points used for the linear trend claim.
  • training set sampling frequencies = 1 per 100 fs (0-500 fs), 1 per 10 fs (100-2500 fs), 1 per 100 fs (>2500 fs)
    Hand-chosen sampling densities for reducing the AIMD dataset; this reflects the authors' judgment about which regions matter.
assumptions (6)
  • domain assumption PBE-D3(BJ) DFT provides an accurate potential energy surface for CO aggregates
    All training data come from this level of theory; no benchmark against higher-level methods is provided.
  • domain assumption The prior AIMD dataset (100 trajectories) is representative of the desorption dynamics
    The MLP is trained on this dataset and validated against it, so it cannot fix systematic errors in the AIMD sampling.
  • ad hoc to paper Labeling the excited molecule with new atom types X1/X2 does not change the physics
    The distinction is a computational device; the model must still represent the same CO molecule with the same PES, which is only approximately enforced by training.
  • domain assumption A molecule is desorbed when its distance to the aggregate surface exceeds 3 Å
    This operational criterion may misclassify molecules that are still bound or that return to the cluster within 5 ps.
  • standard math Semiclassical rovibrational energy assignment is valid for CO under these conditions
    The Billing method is used without checking its accuracy for highly excited, strongly interacting molecules.
  • domain assumption Vibrational excitation at v=40 reproduces the final step of the experimentally observed DIET process
    This equivalence comes from prior work (refs 25, 26); the present paper does not test other excitation levels.
invented entities (1)
  • X1 and X2 atom types
    purpose: Distinct neural network parameters for the carbon and oxygen atoms of the vibrationally excited CO molecule
    These are computational labels without physical existence; their sole support is the in-paper comparison to DFT energies and forces, and the dissociation curve.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deep Potential-Driven Molecular Dynamics of CO Ice Analogues: Investigating Desorption Following Vibrational Excitation." pith.science (2026). https://pith.science/paper/U5A3XNIP

@misc{pith2026250610882,
  author       = {Pith},
  title        = {Pith review of: Deep Potential-Driven Molecular Dynamics of CO Ice Analogues: Investigating Desorption Following Vibrational Excitation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U5A3XNIP}},
  note         = {Machine review of arXiv:2506.10882}
}
read the original abstract

We present a new deep learning-based machine learning potential (MLP) for molecular dynamics simulations of solid carbon monoxide (CO), capable of accurately describing CO vibrations both in the fundamental state and in highly excited vibrational states, up to approximately v = 40. The MLP is based on the combination of high-dimensional neural network atomic potentials using the DeePMD-kit package, trained on prior ab initio molecular dynamics (AIMD) data, with selective treatment of the excited molecule allowing us to capture complex energy redistribution dynamics in condensed-phase environments. In particular, the MLP is capable of accurately describing the desorption process of a single CO molecule within an aggregate of 50 CO molecules, in excellent agreement with both previous theoretical predictions and experimental measurements. The MLP provides a much finer description of the translational and rotational energy distributions, capturing their character with high fidelity and allowing a more detailed comparison with experimental results. Furthermore, the analysis of the rotational energy, resolved over specific translational energies, revealed new insights into the coupling between translational and rotational degrees of freedom during the photodesorption process. This novel approach opens new perspectives for extensive statistical studies on desorption energies and detailed investigations of surface molecule excitations and the exploration of larger-scale models incorporating periodic boundary conditions to simulate more realistic CO aggregates.

Figures

Figures reproduced from arXiv: 2506.10882 by the authors.

Figure 1
Figure 1. FIG. 1. Representation of one CO aggregate. The core region, rep [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. Parity plot of the total energy values (top panel) and the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Parity plot of the absolute value of the forces obtained by the MLP prediction of the DFT set for each X [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: FIG. 4. Distribution of the absolute error of the forces obtained by the MLP prediction of the DFT set for each atom involved. The bin size [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Heatmap representing the relative force error and the angle between DFT and MLP predicted force vector. The highest density is [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. Comparison of the dimer dissociation computed at the DFT [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Translational energy distributions computed using MLP [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 9
Figure 9. Figure 9: FIG. 9. Theoretical rotational MLP energy distributions (blue) compared to experimental data [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: FIG. 10. Evolution of the most probable [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 27 canonical work pages

  1. [1]

    author author E. L. \ Gibb , author D. C. B. \ Whittet , author W. A. \ Schutte , author A. C. A. \ Boogert , author J. E. \ Chiar , author P. Ehrenfreund , author P. A. \ Gerakines , author J. V. \ Keane , author A. G. G. M. \ Tielens , author E. F. \ van Dishoeck , \ and\ author O. Kerkhof ,\ 10.1086/308940 journal journal The Astrophysical Journal \ vo...

    work page doi:10.1086/308940

  2. [2]

    author author A. A. \ Boogert , author P. A. \ Gerakines , \ and\ author D. C. \ Whittet ,\ 10.1146/annurev-astro-082214-122348 journal journal Annual Review of Astronomy and Astrophysics \ volume 53 ,\ pages 541 ( year 2015 ) NoStop

    doi:10.1146/annurev-astro-082214-122348 2015

  3. [3]

    author author M. K. \ McClure , author W. R. M. \ Rocha , author K. M. \ Pontoppidan , author N. Crouzet , author L. E. U. \ Chu , author E. Dartois , author T. Lamberts , author J. A. \ Noble , author Y. J. \ Pendleton , author G. Perotti , author D. Qasim , author M. G. \ Rachid , author Z. L. \ Smith , author F. Sun , author T. L. \ Beck , author A. C....

  4. [4]

    Caselli \ and\ author C

    author author P. Caselli \ and\ author C. Ceccarelli ,\ 10.1007/s00159-012-0056-x journal journal The Astronomy and Astrophysics Review \ volume 20 ,\ pages 56 ( year 2012 ) NoStop

    doi:10.1007/s00159-012-0056-x 2012

  5. [5]

    author author S. S. \ Prasad \ and\ author S. P. \ Tarafdar ,\ 10.1086/160896 journal journal The Astrophysical Journal \ volume 267 ,\ pages 603 ( year 1983 ) NoStop

    doi:10.1086/160896 1983

  6. [6]

    author author C. J. \ Shen , author J. M. \ Greenberg , author W. A. \ Schutte , \ and\ author E. F. \ Van Dishoeck ,\ 10.1051/0004-6361:20031669 journal journal Astronomy & Astrophysics \ volume 415 ,\ pages 203 ( year 2004 ) NoStop

    work page doi:10.1051/0004-6361:20031669 2004

  7. [7]

    Willacy \ and\ author W

    author author K. Willacy \ and\ author W. D. \ Langer ,\ @noop journal journal Astrophysical Journal \ volume 544 ,\ pages 903 ( year 2000 ) NoStop

    work page 2000

  8. [8]

    Pi \'e tu , author A

    author author V. Pi \'e tu , author A. Dutrey , \ and\ author S. Guilloteau ,\ 10.1051/0004-6361:20066537 journal journal Astronomy & Astrophysics \ volume 467 ,\ pages 163 ( year 2007 ) ,\ note number: 1 Publisher: EDP Sciences NoStop

    work page doi:10.1051/0004-6361:20066537 2007

Show all 55 references
  1. [9]

    Dartois , author A

    author author E. Dartois , author A. Dutrey , \ and\ author S. Guilloteau ,\ 10.1051/0004-6361:20021638 journal journal Astronomy & Astrophysics \ volume 399 ,\ pages 773 ( year 2003 ) NoStop

    2003 doi

  2. [10]

    author author E. F. \ Dishoeck \ and\ author J. H. \ Black ,\ in\ 10.1007/978-94-009-3945-5_18 booktitle Physical Processes in Interstellar Clouds ,\ editor edited by\ editor G. E. \ Morfill \ and\ editor M. Scholer \ ( publisher Springer Netherlands ,\ address Dordrecht ,\ ye...

    doi

  3. [11]

    author author K. I. \ \"O berg ,\ 10.1021/acs.chemrev.5b00694 journal journal Chemical Reviews \ volume 116 ,\ pages 9631 ( year 2016 ) NoStop

    2016 doi

  4. [12]

    author author K. I. \ \"Oberg , author G. W. \ Fuchs , author Z. Awad , author H. J. \ Fraser , author S. Schlemmer , author E. F. \ Van Dishoeck , \ and\ author H. Linnartz ,\ @noop journal journal Astrophysical Journal \ volume 662 ,\ pages L23 ( year 2007 ) NoStop

    2007

  5. [13]

    author author G. M. \ Mu\ noz Caro , author A. Jimenez-Escobar , author J. A. \ Martin-Gago , author C. Rogero , author C. Atienza , author S. Puertas , author J. M. \ Sobrado , \ and\ author J. Torres-Redondo ,\ @noop journal journal Astronomy & Astrophysics \ volume 522 ,\ p...

    2010

  6. [14]

    author author E. C. \ Fayolle , author M. Bertin , author C. Romanzin , author X. Michaut , author K. I. \ \"Oberg , author H. Linnartz , \ and\ author J.-H. \ Fillion ,\ @noop journal journal Astrophysical Journal Letters \ volume 739 ,\ pages L36 ( year 2011 ) NoStop

    2011

  7. [15]

    Bertin , author E

    author author M. Bertin , author E. C. \ Fayolle , author C. Romanzin , author K. I. \ \"Oberg , author X. Michaut , author A. Moudens , author L. Philippe , author P. Jeseck , author H. Linnartz , \ and\ author J.-H. \ Fillion ,\ @noop journal journal Physical Chemistry Chemi...

    2012

  8. [16]

    Bertin , author E

    author author M. Bertin , author E. C. \ Fayolle , author C. Romanzin , author H. A. M. \ Poderoso , author X. Michaut , author L. Philippe , author P. Jeseck , author K. I. \ Öberg , author H. Linnartz , \ and\ author J.-H. \ Fillion ,\ 10.1088/0004-637x/779/2/120 journal jou...

    doi

  9. [17]

    author author D. M. \ Paardekooper , author G. Fedoseev , author A. Riedo , \ and\ author H. Linnartz ,\ 10.1051/0004-6361/201629063 journal journal Astronomy & Astrophysics \ volume 596 ,\ pages A72 ( year 2016 ) NoStop

    2016 doi

  10. [18]

    Carrascosa , author G

    author author H. Carrascosa , author G. M. \ Mu \ n oz Caro , author C. Gonz \'a lez-d \'i az , author J. Suevos , \ and\ author Y.-J. \ Chen ,\ 10.3847/1538-4357/ac0a7a journal journal The Astrophysical Journal \ volume 916 ,\ pages 1 ( year 2021 ) NoStop

    doi

  11. [19]

    Gonz \'a lez D \'i az , author H

    author author C. Gonz \'a lez D \'i az , author H. Carrascosa de Lucas , author S. Aparicio , author G. M. \ Mu \ n oz Caro , author N.-E. \ Sie , author L.-C. \ Hsiao , author S. Cazaux , \ and\ author Y.-J. \ Chen ,\ 10.1093/mnras/stz1223 journal journal Monthly Notices of t...

    doi

  12. [20]

    author author K. I. \ \"Oberg , author E. F. \ van Dishoeck , \ and\ author H. Linnartz ,\ @noop journal journal Astronomy & Astrophysics \ volume 496 ,\ pages 281 ( year 2009 ) NoStop

    2009

  13. [21]

    \ Sie , author Y.-T

    author author N.-E. \ Sie , author Y.-T. \ Cho , author C.-H. \ Huang , author G. M. M. \ Caro , author L.-C. \ Hsiao , author H.-C. \ Lin , \ and\ author Y.-J. \ Chen ,\ 10.3847/1538-4357/ac922a journal journal The Astrophysical Journal \ volume 938 ,\ pages 48 ( year 2022 ) ...

    doi

  14. [22]

    \ Chen , author K.-J

    author author Y.-J. \ Chen , author K.-J. \ Chuang , author G. M. \ Mu \ n oz Caro , author M. Nuevo , author C.-C. \ Chu , author T.-S. \ Yih , author W.-H. \ Ip , \ and\ author C.-Y. R. \ Wu ,\ 10.1088/0004-637X/781/1/15 journal journal The Astrophysical Journal \ volume 781...

    doi

  15. [23]

    author author G. M. \ Mu \ n oz Caro , author Y.-J. \ Chen , author S. Aparicio , author A. Jim \'e nez-Escobar , author A. Rosu-Finsen , author J. Lasne , \ and\ author M. R. S. \ McCoustra ,\ 10.1051/0004-6361/201628121 journal journal Astronomy & Astrophysics \ volume 589 ,...

    doi

  16. [24]

    \ Lu , author H.-K

    author author H.-C. \ Lu , author H.-K. \ Chen , author B.-M. \ Cheng , author Y.-P. \ Kuo , \ and\ author J. F. \ Ogilvie ,\ 10.1088/0953-4075/38/20/006 journal journal Journal of Physics B: Atomic, Molecular and Optical Physics \ volume 38 ,\ pages 3693 ( year 2005 ) NoStop

    doi

  17. [25]

    Del Fr \'e , author A

    author author S. Del Fr \'e , author A. R. \ Santamar \'i a , author D. Duflot , author R. Basalg \`e te , author G. F \'e raud , author M. Bertin , author J.-H. \ Fillion , \ and\ author M. Monnerville ,\ 10.1103/PhysRevLett.131.238001 journal journal Physical Review Letters ...

    doi

  18. [26]

    author author A. B. \ Hacquard , author R. Basalg \`e te , author S. Del Fr \'e , author J. Rakovsk \'y , author A. Rivero Santamaria , author F. Benoit , author X. Michaut , author G. F \'e raud , author M. Bertin , author M. Monnerville , \ and\ author J.-H. \ Fillion ,\ 10....

    doi

  19. [27]

    author author A. M. \ Tokita \ and\ author J. Behler ,\ 10.1063/5.0160326 journal journal The Journal of Chemical Physics \ volume 159 ,\ pages 121501 ( year 2023 ) NoStop

    2023 doi

  20. [28]

    u tt , author A. Tkatchenko , \ and\ author K.-R. \ M \

    author author O. T. \ Unke , author S. Chmiela , author H. E. \ Sauceda , author M. Gastegger , author I. Poltavsky , author K. T. \ Sch \"u tt , author A. Tkatchenko , \ and\ author K.-R. \ M \"u ller ,\ 10.1021/acs.chemrev.0c01111 journal journal Chemical Reviews \ volume 12...

    doi

  21. [29]

    Wang , author C

    author author G. Wang , author C. Wang , author X. Zhang , author Z. Li , author J. Zhou , \ and\ author Z. Sun ,\ 10.1016/j.isci.2024.109673 journal journal iScience \ volume 27 ,\ pages 109673 ( year 2024 ) NoStop

    2024

  22. [30]

    Rivero Santamar \'i a , author M

    author author A. Rivero Santamar \'i a , author M. Ramos , author M. Alducin , author H. F. \ Busnengo , author R. D \'i ez Mui \ n o , \ and\ author J. I. \ Juaristi ,\ 10.1021/acs.jpca.1c00835 journal journal The Journal of Physical Chemistry A \ volume 125 ,\ pages 2588 ( y...

    doi

  23. [31]

    author author M. S. \ Chen , author T. Morawietz , author H. Mori , author T. E. \ Markland , \ and\ author N. Artrith ,\ 10.1063/5.0063880 journal journal The Journal of Chemical Physics \ volume 155 ,\ pages 074801 ( year 2021 ) NoStop

    doi

  24. [32]

    author author L. G. \ F. Dos Santos , author B. T. \ Nebgen , author A. E. A. \ Allen , author B. W. \ Hamilton , author S. Matin , author J. S. \ Smith , \ and\ author R. A. \ Messerly ,\ 10.1021/acs.jcim.4c01847 journal journal Journal of Chemical Information and Modeling \ ...

    doi

  25. [33]

    Wang , author L

    author author H. Wang , author L. Zhang , author J. Han , \ and\ author W. E ,\ 10.1016/j.cpc.2018.03.016 journal journal Computer Physics Communications \ volume 228 ,\ pages 178 ( year 2018 ) NoStop

    2018 doi

  26. [34]

    Zeng , author D

    author author J. Zeng , author D. Zhang , author D. Lu , author P. Mo , author Z. Li , author Y. Chen , author M. Rynik , author L. Huang , author Z. Li , author S. Shi , author Y. Wang , author H. Ye , author P. Tuo , author J. Yang , author Y. Ding , author Y. Li , author D....

  27. [36]

    Wen , author L

    author author T. Wen , author L. Zhang , author H. Wang , author W. E , \ and\ author D. J. \ Srolovitz ,\ 10.1088/2752-5724/ac681d journal journal Materials Futures \ volume 1 ,\ pages 022601 ( year 2022 ) NoStop

    2022 doi

  28. [37]

    author author D. P. \ Kingma ,\ @noop journal journal arXiv preprint arXiv:1412.6980 \ ( year 2014 ) NoStop

    2014 arXiv

  29. [38]

    Kresse \ and\ author J

    author author G. Kresse \ and\ author J. Furthm \"u ller ,\ 10.1016/0927-0256(96)00008-0 journal journal Comput. Mater. Sci. \ volume 6 ,\ pages 15 ( year 1996 a ) NoStop

    1996 doi

  30. [39]

    Kresse \ and\ author J

    author author G. Kresse \ and\ author J. Furthm \"u ller ,\ 10.1103/physrevb.54.11169 journal journal Physical Review B \ volume 54 ,\ pages 11169 ( year 1996 b ) NoStop

    1996 doi

  31. [40]

    author author J. P. \ Perdew , author K. Burke , \ and\ author M. Ernzerhof ,\ 10.1103/physrevlett.77.3865 journal journal Physical Review Letters \ volume 77 ,\ pages 3865 ( year 1996 ) NoStop

    1996 doi

  32. [41]

    Grimme , author J

    author author S. Grimme , author J. Antony , author S. Ehrlich , \ and\ author H. Krieg ,\ 10.1063/1.3382344 journal journal J. Chem. Phys. \ volume 132 ,\ pages 154104 ( year 2010 ) NoStop

    2010 doi

  33. [42]

    Grimme , author S

    author author S. Grimme , author S. Ehrlich , \ and\ author L. Goerigk ,\ 10.1002/jcc.21759 journal journal J. Comput. Chem. \ volume 32 ,\ pages 1456 ( year 2011 ) NoStop

    2011 doi

  34. [43]

    author author P. E. \ Bl \"o chl ,\ @noop journal journal Phys Rev B \ volume 50 ,\ pages 17953 ( year 1994 ) NoStop

    1994

  35. [44]

    Kresse \ and\ author D

    author author G. Kresse \ and\ author D. Joubert ,\ 10.1103/PhysRevB.59.1758 journal journal Physical Review B \ volume 59 ,\ pages 1758 ( year 1999 ) NoStop

    1999 doi

  36. [45]

    author author S. J. \ Plimpton , author A. Kohlmeyer , author A. P. \ Thompson , author S. G. \ Moore , \ and\ author R. Berger ,\ 10.5281/zenodo.10806852 title Lammps: Large-scale atomic/molecular massively parallel simulator , \ ( year 2023 ) NoStop

    doi

  37. [46]

    Nos \'e ,\ 10.1063/1.447334 journal journal The Journal of Chemical Physics \ volume 81 ,\ pages 511 ( year 1984 ) NoStop

    author author S. Nos \'e ,\ 10.1063/1.447334 journal journal The Journal of Chemical Physics \ volume 81 ,\ pages 511 ( year 1984 ) NoStop

    1984 doi

  38. [47]

    author author W. G. \ Hoover ,\ 10.1103/PhysRevA.31.1695 journal journal Physical Review A \ volume 31 ,\ pages 1695 ( year 1985 ) NoStop

    1985 doi

  39. [48]

    Verlet ,\ 10.1103/PhysRev.159.98 journal journal Phys

    author author L. Verlet ,\ 10.1103/PhysRev.159.98 journal journal Phys. Rev. \ volume 159 ,\ pages 98 ( year 1967 ) NoStop

    1967 doi

  40. [49]

    author author W. C. \ Swope , author H. C. \ Andersen , author P. H. \ Berens , \ and\ author K. R. \ Wilson ,\ 10.1063/1.442716 journal journal The Journal of Chemical Physics \ volume 76 ,\ pages 637 ( year 1982 ) ,\ http://arxiv.org/abs/https://pubs.aip.org/aip/jcp/article-...

    1982 doi

  41. [50]

    author author G. D. \ Billing ,\ @noop title Dynamics of Molecule Surface Interaction \ ( publisher John Wiley & Sons ,\ year 2000 ) NoStop

    2000

  42. [51]

    Behler \ and\ author M

    author author J. Behler \ and\ author M. Parrinello ,\ 10.1103/PhysRevLett.98.146401 journal journal Physical Review Letters \ volume 98 ,\ pages 146401 ( year 2007 b ) NoStop

    2007 doi

  43. [52]

    Behler ,\ 10.1063/1.3553717 journal journal The Journal of Chemical Physics \ volume 134 ,\ pages 074106 ( year 2011 ) NoStop

    author author J. Behler ,\ 10.1063/1.3553717 journal journal The Journal of Chemical Physics \ volume 134 ,\ pages 074106 ( year 2011 ) NoStop

    2011 doi

  44. [53]

    Hedman , author B

    author author D. Hedman , author B. McLean , author C. Bichara , author S. Maruyama , author J. A. \ Larsson , \ and\ author F. Ding ,\ 10.1038/s41467-024-47999-7 journal journal Nature Communications \ volume 15 ,\ pages 4076 ( year 2024 ) NoStop

    doi

  45. [54]

    \ Zhang , author C

    author author X.-W. \ Zhang , author C. Wang , author X. Liu , author Y. Fan , author T. Cao , \ and\ author D. Xiao ,\ 10.1038/s41467-024-48511-x journal journal Nature Communications \ volume 15 ,\ pages 4223 ( year 2024 ) NoStop

    2024 doi

  46. [55]

    Tosello Gardini , author U

    author author A. Tosello Gardini , author U. Raucci , \ and\ author M. Parrinello ,\ 10.1038/s41467-025-57688-8 journal journal Nature Communications \ volume 16 ,\ pages 2475 ( year 2025 ) NoStop

    2025 doi

  47. [56]

    Virtanen , author R

    author author P. Virtanen , author R. Gommers , author T. E. \ Oliphant , author M. Haberland , author T. Reddy , author D. Cournapeau , author E. Burovski , author P. Peterson , author W. Weckesser , author J. Bright , author S. J. \ van der Walt , author M. Brett , author J....

Pith tools

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