Pith. sign in

REVIEW 2 major objections 5 minor 1 cited by

Robust Binding Energy Distribution Sampling on Amorphous Solid Water Models. Method testing and validation with NH3, CO and CH4

T0 review · 2 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper infers converged binding-energy distributions for NH3, CO and CH4 on amorphous water ice, finding ammonia binds in two modes — a weak-binding subpopulation covering ~16% of sites — while CO and CH4 are single Gaussians.

desk verdict Careful, honest method paper whose NH3/CO distributions are likely useful, but the CH4 distribution carries a known and unpropagated low-level bias that needs addressing. read the letter →

arxiv 2504.18435 v1 pith:KJ67FLIC submitted 2025-04-25 astro-ph.IM

Maria Groyne , Benoît Champagne , Cedric Baijot , Michaël De Becker This is my paper · ORCID
classification astro-ph.IM
keywords bindingenergydistributionamorphoussolidwaterastrochemistryONIOMGFN2-xtbinterstellariceammoniadesorption
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 argues that the binding energy of an interstellar species on amorphous solid water (ASW) is not a single number but a distribution, and it develops a way to infer such distributions with converged statistics. A 2000-molecule ASW model is built by molecular-dynamics Heat & Quench, validated against experimental pair-distribution data, and sampled at 100 binding sites with several adsorbate orientations per site; binding energies are computed by a two-layer ONIOM method whose model-zone and real-system sizes are separately benchmarked to convergence. The results are converged binding-energy distributions for NH3, CO and CH4: ammonia's is bimodal, with a weak-binding subpopulation between roughly 15 and 30 kJ/mol — configurations that form no hydrogen bond with the surface — covering 16.3% of sites, while CO and CH4 each follow a single Gaussian. If these distributions are correct, astrochemical models that assume one desorption energy per species misrepresent when and how these molecules return to the gas phase, and the ammonia weak-binding tail offers a natural source of gas-phase NH3 in cold prestellar cores.

What carries the argument

The load-bearing machinery is the twofold binding-configuration sampling on a structurally validated amorphous ice model. The substrate is a 2000-molecule water box built by an MD Heat & Quench protocol with the TIP4P/2005 force field, checked against experimental O–O radial distribution functions, from which 100 regularly spaced hemispheric sub-clusters (grid spacing 4 Å, radius 16 Å) are cut. On each site, several starting adsorbate-to-substrate orientations are optimized — 3 for NH3, 3 for CO, 1 for CH4 — so that the local roughness of the potential energy surface contributes to the distribution alongside the variety of binding sites. Binding energies come from the two-layer ONIOM scheme ONIOM(B3LYP-D3(BJ)/6-311+G(d,p):GFN2-xtb), with an 8 Å model zone of about 20–25 water molecules inside a real system of about 235–250 water molecules, corrected for zero-point energy and basis-set superposition error; the real-system size is benchmarked up to $\Delta R_{\mathrm{LL}} = 11$ Å and retro-verified at 25 Å. Redundant configurations are removed with RMSD and binding-energy-difference cutoffs, and the fitted Gaussian (mixture) parameters are tested for statistical convergence by bootstrapping against a 5% tolerance.

What would settle it

Recompute the binding energies of the NH3 configurations with binding energies below about 30 kJ/mol, and of the full CH4 set, with the low-level layer upgraded from GFN2-xtb to a dispersion-aware DFT treatment with diffuse basis functions at a real-system radius of 19 Å ($\Delta R_{\mathrm{LL}} = 11$ Å), then re-fit the distributions: if the 16.3% weak-binding fraction of NH3 or the mean and width of the CH4 Gaussian shift beyond the paper's own 5% convergence tolerance, the reported distributions are artifacts of the low-level method rather than properties of the ice surface.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that ammonia's binding-energy distribution on amorphous solid water is bimodal and is best fitted by two Gaussian components: a dominant peak from configurations in which NH3 accepts a hydrogen bond from the surface, and a weaker low-energy component (roughly 15–30 kJ/mol) dominated by configurations that form no hydrogen bond with the surface, or at most one with NH3 as donor, covering 16.3% of the distribution. The same sampling yields single-Gaussian distributions for CO and CH4, and the paper shows the distributions' means and standard deviations are converged by bootstrapping once the two-level sampling is complete. It further establishes that one randomly chosen adsorbate orientation per site is not enough to converge the NH3 and CO statistics, that three starting orientations per site are required for both components of the NH3 double Gaussian, that two suffice for CO, and that one is sufficient for the symmetric CH4. Finally, the inferred distributions encompass all previously reported single values and dispersion ranges for these three species on water ice, while no prior individual study covers the full ranges reported here.

Load-bearing premise

Every reported binding energy inherits its accuracy from the semi-empirical GFN2-xtb method used as the low-level layer of the ONIOM scheme, and the paper's own tests (Appendix D) show that for the apolar adsorbate CH4 this method deviates from a DFT low-level treatment by 20–30% at real-system sizes beyond the chosen $\Delta R_{\mathrm{LL}} = 8$ Å shell, so a low-level-method bias in the CH4 distribution — and potentially in the tails of the others — cannot be excluded.

Editorial extensions

If this is right

  • Astrochemical kinetic models should carry a distribution rather than a single binding energy per species; for NH3 the paper supplies a two-component Gaussian whose weak subpopulation (16.3% of sites) can be fed directly into desorption and chemical-desorption schemes.
  • The weak-binding NH3 component (about 15–30 kJ/mol) provides a concrete desorption channel in cold prestellar cores: a chemical-desorption efficiency below the 1% upper limit inferred for L1544 could explain the observed gas-phase ammonia abundance.
  • Method transfer to new adsorbates should follow the paper's convergence rule: the number of starting orientations per site is whatever makes the bootstrapped statistics of the fitted distribution fall within tolerance — 3 for NH3, 2 for CO, 1 for CH4 — and one random orientation per site is insufficient for H-bonding adsorbates.
  • Binding energies on amorphous ice require a real system radius of order 16 Å; smaller shells produce artifacts (an NH3 site jumps from about 23 kJ/mol to over 80 kJ/mol at an 11 Å radius), so previously reported small-cluster values that fall outside the new ranges are attributed by the authors to insufficient system size.

Reading between the lines

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

  • If the bimodality is real, warm-up models of ice mantles will release ammonia in two episodes — an early weak-binding release followed by the main sublimation — changing the timing of gas-phase nitrogen chemistry; the paper only sketches the static prestellar-core case.
  • The paper's own Appendix D shows GFN2-xtb reproduces DFT low-level binding energies within about 5% for NH3 but deviates by 20–30% for apolar CH4 beyond the chosen shell; a conservative testable extension is to re-run the CH4 sampling, and any future apolar adsorbate, with a dispersion-accurate low-level method before adopting the distribution.
  • The 16.3% weak-binding fraction is a sharp prediction: rebuilding the ASW slab with a different quench rate, density, or ice composition (the authors announce a density-variation study) should move this fraction if the surface typology controls it, and leave it stable if it is intrinsic to the NH3–water interaction.
  • The bootstrapping-with-tolerance test itself is a transferable standard: any binding-energy distribution published without a convergence check of its fitted statistics — which includes most earlier work in this field — should be treated as provisional.
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

2 major / 5 minor

Summary. This manuscript presents a multiscale ONIOM-based workflow for inferring binding-energy (BE) distributions of NH3, CO, and CH4 on amorphous solid water (ASW). The ice model is built from TIP4P/2005 molecular-dynamics simulations with a Heat & Quench protocol and is validated against experimental radial distribution functions and hydrogen-bond statistics. Binding energies are computed on 100 hemispherical sub-clusters using ONIOM(B3LYP-D3(BJ)/6-311+G(d,p):GFN2-xtb), with a regular grid of starting positions and, for NH3 and CO, multiple starting adsorbate orientations per site. The authors benchmark the DFT high-level method against CCSD(T), investigate convergence with respect to the low-level shell size, retro-check the layer sizes on a 25 Å hemisphere, and use bootstrapping to assess statistical convergence of the final distributions. They report a double-Gaussian NH3 BE distribution with a low-energy component of about 15–30 kJ/mol covering 16.3% of the sample, and single-Gaussian CO and CH4 distributions, and they compare these distributions with previously reported single values and dispersions.

Significance. The paper is a valuable methodological contribution: the structural validation against experimental RDF data, the CCSD(T) benchmark for the high-level method, and the explicit bootstrapping analysis of sampling convergence go beyond most prior BE-distribution studies. The two-level sampling of binding sites and adsorbate orientations directly addresses a recognized source of BE dispersion and provides a concrete recipe for future adsorbates. If the reported distributions are accurate, they are astrochemically significant because desorption rates depend exponentially on BE and because the distribution shape, especially the NH3 weak-binding tail, matters for gas-phase abundance modeling. The main caveat is that the final distributions inherit systematic errors from the GFN2-xtb low-level treatment and from the truncated real-system radius for CH4; these are quantified in the manuscript but are not propagated into the reported means, widths, or literature comparisons, so the strength of the central quantitative claim is currently method-dependent.

major comments (2)
  1. [Sect. 2.2.2, Table 2, Appendix D (Fig. D.3)] The CH4 distribution is not computed at a converged real-system size, and the chosen low-level method is known to be the weakest for this adsorbate. Table 2 shows BE(CH4) = 9.8 kJ/mol at ΔRLL = 8 Å versus 10.3–10.5 kJ/mol at ΔRLL = 10–11 Å, and the text states that CH4 convergence appears only from ΔRLL ≈ 10 Å; nevertheless all 98 CH4 sites are computed at ΔRLL = 8 Å. Appendix D further reports that ONIOM(DFT:GFN2-xtb) deviates from ONIOM(DFT:DFT) by roughly 20% at ΔRLL = 8 Å and up to about 30% at larger real-system sizes for CH4. Because CH4 binding is dispersion-dominated, this systematic offset shifts both the mean and the width of the distribution in Fig. 4, and it weakens the claim that the reported CH4 distribution encompasses previously published experimental and computational values. The authors should either recompute the CH4 set at a converged ΔRLL (or with a DFT low level), or report the distribution with explicit systematic uncertainty and temper the 'encompassing' conclusion accordingly.
  2. [Sects. 3 and 4.3] The phrase 'converged statistics' is used for sampling convergence only, and the reported distributions carry no propagated uncertainty from the two largest identified error sources: the ~13.4% MARD of B3LYP-D3(BJ)/6-311+G(d,p) against CCSD(T) (Table 1) and the 20–30% low-level discrepancy for CH4 (Appendix D). The bootstrapping analysis in Sect. 4.3 convincingly shows that the empirical mean, standard deviation, and GMM components stabilize with sample size, but it does not constrain the systematic accuracy of the BE scale. Since desorption rates are exponential in BE, the absence of any systematic error budget makes the quantitative astrochemical statements, including the 16.3% NH3 tail fraction, overly precise. The authors should add a systematic-error discussion and state, at a minimum, how the main conclusions change when the upper end of the benchmark errors is assumed.
minor comments (5)
  1. [Sects. 2.2.2, Appendix D, Appendix E] The low-level method is repeatedly spelled 'GNF2-xtb'; it should be 'GFN2-xtb' throughout.
  2. [Appendix C, final paragraph] The final sentence says 'B3LYP-D3(BJ)/6-111+G(d,p)'; this should read '6-311+G(d,p)'.
  3. [Eq. (3) and Table 2] The notation for the BSSE correction (ΔE_BSSE*_CP versus ΔE_BSSE_CP) is confusing because the two quantities have opposite signs; please define the signs explicitly in the text and use one consistent symbol in Eq. (3), Table 2, and the discussion.
  4. [Appendix D (Figs. D.1–D.3)] The fonts, axis labels, and legends in these figures are small and difficult to read, and one legend again uses 'GNF2-xtb'; please enlarge and unify the formatting.
  5. [Sect. 4.3] There is a typo in 'convergence cu-off tolerance interval' (should be 'cutoff'), and the description of the bootstrapping procedure would be clearer if it stated explicitly that 100 resamples are drawn at each set size and that the three runs differ only in the number of orientations selected per cut.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity; the BE distributions are computed from an externally benchmarked multiscale model, not fitted to prior BE values.

full rationale

The paper's central deliverables (NH3, CO, and CH4 binding energy distributions) are not derived from the literature values they are later compared with. The method chain is self-contained: the ASW model is validated against experimental RDF data (Mariedahl et al. 2018), the high-level DFT functional is selected by benchmarking against CCSD(T) on tetrameric clusters (Table 1, Appendix B), the ONIOM layer sizes are tested by explicit convergence scans and a retro-verification with a 12-A model zone (Section 2.2.3), and the final distributions are assessed with bootstrapping convergence analyses rather than by forcing agreement with prior results. The only self-citation is the Beaujean et al. (2021) external-interface wrapper used to call GFN2-xtb through Gaussian; this is a code/tool citation and does not enter the physics or the derivation. The comparison showing that the new distributions 'encompass' earlier values is a consistency statement made after the computation, not an input to it. Appendix D does report that GFN2-xtb deviates from a DFT low-level treatment for CH4 by up to ~20% at the chosen DeltaRLL=8 A and ~30% at larger sizes; that is a quantified accuracy limitation of the low-level method, not a circular step, because the CH4 values are computed from the model rather than fitted to the compared references. No step in the derivation reduces to its own inputs by construction, and no load-bearing argument depends on an unverified self-citation.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

Central claim rests on a chain of validated approximations (force field, DFT functional, xtb low level, harmonic ZPE, sampling density). The main free parameters are the ONIOM layer radii and the redundancy cutoffs, all benchmarked to convergence but not error-propagated.

free parameters (6)
  • ONIOM high-level model zone radius RModel,HL = 8 Å
    Chosen from earlier benchmarks (Tinacci et al. 2022; Duflot et al. 2021) and retro-verified at 12 Å; not fitted to target BEs but is a hand-picked cutoff that the distributions depend on.
  • Low-level shell size ΔRLL = 8 Å
    Chosen from BE vs real system size convergence curves on two NH3 sites and one CO/CH4 site (Table 2); converged BE at ΔRLL=8-10 Å depending on species, and 8 Å is the declared compromise. The choice directly sets the amount of environment included.
  • Binding site grid spacing = 4 Å
    Chosen to avoid redundancy between adjacent sites; validated by RMSD analysis (Appendix F) but remains a sampling density assumption.
  • In-site redundancy cutoffs = ΔBE < 0.05 × <BE>; RMSD < 1 Å
    Ad hoc thresholds used to merge configurations from different starting orientations on the same cut; justified by matching integrated densities of RMSD and ΔBE histograms (Appendix F.1), but they shape the final distribution.
  • Number of starting orientations per site = 3 for NH3/CO, 1 for CH4
    Hand-picked by adsorbate symmetry; the convergence analysis in Sect. 4.3 shows that 3 orientations are necessary for NH3, 2 sufficient for CO, and 1 for CH4, so the final datasets follow these choices.
  • Universal ΔZPE scaling factor = 0.837 (global fit) or 0.841 (average of species)
    Proposed as a substitute for explicit ZPE computations in future applications; fitted to the paper's own BE/electronic-BE correlation (Fig. 7), not used to compute the reported distributions.
assumptions (7)
  • domain assumption TIP4P/2005 classical force field reproduces the structure of LDA amorphous solid water sufficiently for binding site sampling.
    Validated against experimental O-O, O-H, H-H RDFs and coordination numbers (Sect. 2.1, Fig. 1, Appendix A), but the model shows slight over-structuration and neglects ZPE.
  • domain assumption B3LYP-D3(BJ)/6-311+G(d,p) interaction energies for the model zone transfer the observed ~13% MARD vs CCSD(T) on tetramers to the full ONIOM model.
    Functional benchmark in Sect. 2.2.1 (Table 1) uses tetramer structures; the error is not propagated into the BE distributions.
  • domain assumption GFN2-xtb reliably describes long-range H-bond cooperativity and dispersion environment of the frozen shell at ΔRLL=8 Å for all three adsorbates.
    Appendix D shows deviations up to 20-30% vs DFT low-level for CH4; the paper argues the compromise is acceptable.
  • standard math Harmonic oscillator approximation for ZPE corrections.
    Frequency computations under harmonic assumption (Sect. 2.2, Eq. 3); standard approximation in this field.
  • domain assumption Counterpoise BSSE correction computed on the model zone only is adequate.
    The paper argues BSSE is non-zero only near the adsorption site (Sect. 2.2).
  • domain assumption 100 grid points on one replicated surface of a single 2000-molecule ASW box are representative of interstellar dust grain ice surfaces.
    Sampling design in Appendix E; the paper states this as the working hypothesis.
  • domain assumption Two-component Gaussian Mixture Model is the correct parametric description for NH3.
    Model selection by fit comparison in Sect. 3.1; bimodality is physically motivated by H-bond counting but remains a modeling choice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Robust Binding Energy Distribution Sampling on Amorphous Solid Water Models. Method testing and validation with NH3, CO and CH4." pith.science (2026). https://pith.science/paper/KJ67FLIC

@misc{pith2026250418435,
  author       = {Pith},
  title        = {Pith review of: Robust Binding Energy Distribution Sampling on Amorphous Solid Water Models. Method testing and validation with NH3, CO and CH4},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KJ67FLIC}},
  note         = {Machine review of arXiv:2504.18435}
}
read the original abstract

This work aims to develop a method based on a structurally reliable ice model and a statistically and physico-chemically robust approach for BE distribution inference, with the aim to be applicable to various relevant interstellar species. A multiscale computational approach is presented, with a Molecular Dynamics (MD) Heat & Quench protocol for the amorphous water ice model, and an ONIOM(B3LYP-D3(BJ)/6-311+G**:GFN2-xtb) scheme for the BE inference, with a prime emphasis onto the BE/real system size convergence. The sampling of the binding configurations is twofold, exploring both regularly spaced binding sites, as well as various adsorbate-to-substrate orientations on each locally distinct site. This second source of BE diversity accounts for the local roughness of the potential energy landscape of the substrate. Three different adsorbate test cases are considered, i.e. NH3, CO and CH4, owing to their significance in dust icy mantles, and their distinct binding behavior with water ices. The BE distributions for NH3, CO and CH4 have been inferred, with converged statistics. The distribution for NH3 is better represented by a double Gaussian component profile. Three starting adsorbate orientations per site are required to reach convergence for both Gaussian components of NH3, while 2 orientations are sufficient for CO, and one unique for CH4 (symmetric). Further geometrical and molecular surrounding insights have been provided. These results encompass previously reported results.

Figures

Figures reproduced from arXiv: 2504.18435 by the authors.

Figure 1
Figure 1. Comparison between the structure modeled through MD simu￾lation in this work (continuous line), and the experimental results from Mariedahl et al. (2018) (dashed line), with (a) RDF for the O-O distance and (b) running coordination number for ASW-LDA ices analogues. et al. (2018), the over-structuring effect may also arguably stem from the classical treatment of nuclear motion in the simulation methodology, which ne… view at source ↗
Figure 2
Figure 2. Normalized histogram of sampled NH3 BE (in kJ/mol), fitted with a Gaussian Mixture Model using scikit-learn algorithm (Pedregosa et al. 2011). Previously reported NH3 BE values onto water ice analogs with diverse simulation methods are given on top. Horizontal intervals report on published BE dispersions, while vertical dashed lines relate to works focusing on single BE values. Methods used in each work are discusse… view at source ↗
Figure 3
Figure 3. Normalized histogram of sampled binding energies (in kJ/mol) for CO, fitted with a single component gaussian profile. Previously re￾ported results of CO BE values onto water ice analogs with diverse sim￾ulation methods are given on top. See [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Statistical analysis of subgroup contributions using stacked his￾tograms normalized on the global BE distribution for NH3, each sub￾groups representing a type of adsorbate-ice H-bond configuration. BE and σ are the respective subgroup mean BE and standard deviation. we…
Figure 7
Figure 7. Figure 7: Linear correlation between the electronic binding energies and the ZPE corrected values (best fit - dashed line) for NH3 (blue), CO (green) and CH4 (orange) on ASW ice 4.2. Analysis of the contributions from ∆ZPE The computation of the ZPE corrections to the binding en…
Figure 6
Figure 6. Figure 6: Sub-group contributions of θ value ranges using stacked his￾tograms normalized on the global BE distribution for CO. BE and σ account for the respective subgroup mean BE and standard deviation. (c) CH4 weak-binding case. Although CH4 exhibits a rela￾tively high isotrop…
Figure 8
Figure 8. Figure 8: Bootstrapping analysis of GMM two components statistics of NH3 BE distribution, with increasing set size. The set size definition depends on the number of picked adsorbate orientation per cut, increas￾ing from 1 to 3 from the left to the right. The dashed line represen…
Figure 9
Figure 9. Figure 9: Bootstrapping analysis of mean BE & standard deviation of CO on LDA ice, with increasing set size. The set size definition depends on the number of picked adsorbate orientation per cut, increasing from 1 to 3 from the left to the right [PITH_FULL_IMAGE:figures/full_fi…
Figure 10
Figure 10. Figure 10: Bootstrapping analysis of mean BE & standard deviation of CH4 on LDA ice, with increasing set size. The set size is defined by the number of picked random hemispherical systems. 4.4. Astrochemical Implications The three adsorbate test cases were chosen for their disti…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Binding energy distributions of alcohols, thiols, and their precursors on interstellar water ice surfaces

    astro-ph.GA 2025-08 conditional novelty 6.0 of 10

    Quantum chemical binding energy distributions are reported for 16 O/S-bearing species on amorphous water ice, and using these distributions alters modeled ice-phase abundances.

Reference graph

Works this paper leans on

68 extracted references · 65 canonical work pages · cited by 1 Pith paper

  1. [1]

    Abascal , J. L. F. & Vega , C. 2005, J. Chem. Phys., 123, 234505

    work page 2005

  2. [2]

    2021, WIRES Comput

    Bannwarth , C., Caldeweyher , E., Ehlert , S., et al. 2021, WIRES Comput. Mol. Sci., 11, e1493

    work page 2021

  3. [3]

    Bannwarth , C., Ehlert , S., & Grimme , S. 2019, J. Chem. Theory Comput., 15, 1652

    work page 2019

  4. [4]

    Beaujean , P., Champagne , B., Grimme , S., & de Wergifosse , M. 2021, J. Phys. Chem. Lett., 12, 9684

    work page 2021

  5. [5]

    Boogert , A. C. A., Gerakines , P. A., & Whittet , D. C. 2015, , 53, 541

    work page 2015

  6. [6]

    2020, Mol

    Bovolenta , G., Bovino , S., Vöhringer-Martinez , E., et al. 2020, Mol. Astrophys., 21, 11pp

    work page 2020

  7. [7]

    W., Sameera , W

    Chung , L. W., Sameera , W. M. C., Ramozzi , R., et al. 2015, Chem. Rev., 115, 5678–5796

    work page 2015

  8. [8]

    P., Anderson , M

    Collings , M. P., Anderson , M. A., Chen , R., et al. 2004, , 364, 1133–1140

    work page 2004

Show all 68 references
  1. [9]

    2007, , 470, 221

    Crapsi , A., Caselli , P., Walmsley , M., & Tafalla , M. 2007, , 470, 221

    2007

  2. [10]

    S., Morokuma , K., & Frisch , M

    Dapprich , S., Komáromi , I., Bryun , K. S., Morokuma , K., & Frisch , M. J. 1999, J. Mol. Struct. THEOCHEM, 461-462, 1

    1999

  3. [11]

    K., & Loison , J.-C

    Das , A., Sil , M., Gorai , P., Chakrabarti , S. K., & Loison , J.-C. 2018, , 237, 13pp

    2018

  4. [12]

    2021, Front

    Duflot , D., Toubin , C., & Monnerville , M. 2021, Front. astron. space sci., 8, 16pp

    2021

  5. [13]

    2010, , 512, 5 pp

    Dulieu , F., Amiaud , L., Congiu , E., et al. 2010, , 512, 5 pp

    2010

  6. [14]

    2020, , 904, 20pp

    Ferrero , S., Zamirri , L., Ceccarelli , C., et al. 2020, , 904, 20pp

    2020

  7. [15]

    L., Hallbrucker , A., Kohl , I., Soper , A

    Finney , J. L., Hallbrucker , A., Kohl , I., Soper , A. K., & Bowron , D. T. 2002, Phys. Rev. Lett, 88, 225503

    2002

  8. [16]

    J., Trucks , G

    Frisch , M. J., Trucks , G. W., Schlegel , H. B., et al. 2016, Gaussian 16, Revision C.01, Gaussian, Inc., Wallingford CT

    2016

  9. [17]

    W., Cuppen , H

    Fuchs , G. W., Cuppen , H. M., Ioppolo , S., et al. 2009, , 505, 629

    2009

  10. [18]

    2024, , 974, 115

    Furuya , K. 2024, , 974, 115

    2024

  11. [19]

    Ganeshan , S., Ramírez , R., & Fernández-Serra , M. V. 2013, Phys. Rev. B Condens. Matter, 87, 134207

    2013

  12. [20]

    & Ugliengo , P

    Germain , A. & Ugliengo , P. 2020, in Gervasi, O., et al. -- ICCSA

    2020

  13. [21]

    Gould , R. J. & Salpeter , E. E. 1963, , 138, 393

    1963

  14. [22]

    2020, , 643, A115

    Grassi , T., Bovino , S., Caselli , P., et al. 2020, , 643, A115

    2020

  15. [23]

    Grimme , S., Antony , J., Ehrlich , S., & Krieg , H. 2010, J. Chem. Phys., 132, 154104

    2010

  16. [24]

    Grimme , S., Ehrlich , S., & Goerigk , L. 2011, J. Comput. Chem., 32, 1456

    2011

  17. [25]

    & Watanabe , N

    Hama , T. & Watanabe , N. 2013, Chem. Rev., 113, 8783

    2013

  18. [26]

    2016, , 825, 10pp

    He , J., Acharyya , K., & Vidali , G. 2016, , 825, 10pp

    2016

  19. [27]

    & Salpeter , E

    Hollenbach , D. & Salpeter , E. E. 1971, , 163, 155–164

    1971

  20. [28]

    Humphrey , W., Dalke , A., & Schulten , K. 1996, J. Mol. Graph., 14, 33

    1996

  21. [29]

    M., van Dishoeck , E

    Ioppolo , S., van Boheemen , Y., Cuppen , H. M., van Dishoeck , E. F., & Linnartz , H. 2011, , 413, 2281–2287

    2011

  22. [30]

    2017, , 607, 23pp

    Kamp , I., Thi , W.-F., Woitke , P., et al. 2017, , 607, 23pp

    2017

  23. [31]

    Karadakov , P. B. & Morokuma , K. 2000, Chem. Phys. Lett., 317, 589

    2000

  24. [32]

    J., Ioppolo , S., van Hemert , M

    Karssemeijer , L. J., Ioppolo , S., van Hemert , M. C., et al. 2014, , 781, 16

    2014

  25. [33]

    & Caselli , P

    Keto , E. & Caselli , P. 2010, , 402, 1625–1634

    2010

  26. [34]

    2021 a , , 505, 1530

    Kouchi , A., Tsuge , M., Hama , T., et al. 2021 a , , 505, 1530

    2021

  27. [35]

    2021 b , , 918, 20pp

    Kouchi , A., Tsuge , M., Hama , T., et al. 2021 b , , 918, 20pp

    2021

  28. [36]

    A., & Frank , I

    Kraus , P., Obenchain , D. A., & Frank , I. 2018, J. Phys. Chem. A, 122, 1077–1087

    2018

  29. [37]

    Lee , C., Yang , W., & Parr , R. G. 1988, Phys. Rev. B Condens. Matter, 37, 785

    1988

  30. [38]

    Mariedahl , D., Perakis , F., Späh , A., et al. 2018, J. Phys. Chem. B, 122, 7616

    2018

  31. [39]

    G., & Martínez , J

    Martinez , L., Andrade , R., Birgin , E. G., & Martínez , J. M. 2009, J. Comput. Chem., 30, 2157

    2009

  32. [40]

    McClure , M., Rocha , W. R. M., Pontoppidan , K. M., et al. 2023, Nat Astron, 7, 431–443

    2023

  33. [41]

    A., Simon , A., Mascetti , J., & Toubin , C

    Michoulier , G., Noble , J. A., Simon , A., Mascetti , J., & Toubin , C. 2018, PCCP, 20, 8753

    2018

  34. [42]

    2022, ACS Earth Space Chem., 6, 597–630

    Minissale , M., Aikawa , Y., Bergin , E., et al. 2022, ACS Earth Space Chem., 6, 597–630

    2022

  35. [43]

    2015, , 577, A2

    Minissale , M., Loison , J.-C., Baouche , S., et al. 2015, , 577, A2

    2015

  36. [44]

    2008, Chem

    Miyauchi , N., Hidaka , H., Chigai , T., et al. 2008, Chem. Phys. Lett., 456, 27

    2008

  37. [45]

    & Kästner , J

    Molpeceres , G. & Kästner , J. 2020, PCCP, 22, 7552

    2020

  38. [46]

    2018, , 482, 5389

    Molpeceres , G., Rimola , A., Ceccarelli , C., et al. 2018, , 482, 5389

    2018

  39. [47]

    2020, , 499, 1373

    Molpeceres , G., Zaverkin , V., & Kästner , J. 2020, , 499, 1373

    2020

  40. [48]

    H., Venkatesh , C

    Narten , A. H., Venkatesh , C. G., & Rice , S. A. 1976, J. Chem. Phys., 64, 1106–1121

    1976

  41. [49]

    A., Theule , P., Mispelaer , F., et al

    Noble , J. A., Theule , P., Mispelaer , F., et al. 2012, , 543, A5

    2012

  42. [50]

    2012, , 749, 67

    Oba , Y., Watanabe , N., Hama , T., et al. 2012, , 749, 67

    2012

  43. [51]

    2016, Chemical Review, 116, 9631

    Oberg, K. 2016, Chemical Review, 116, 9631

    2016

  44. [52]

    2011, JMLR, 12, 2825

    Pedregosa , F., Varoquaux , G., Gramfort , A., et al. 2011, JMLR, 12, 2825

    2011

  45. [53]

    M., Walsh , C., & Cuppen , H

    Penteado , E. M., Walsh , C., & Cuppen , H. M. 2017, , 844, 13pp

    2017

  46. [54]

    C., Hardy , D

    Phillips , J. C., Hardy , D. J., Maia , J. D. C., et al. 2020, J. Chem. Phys., 153, 044130

    2020

  47. [55]

    & Krim , L

    Pirim , C. & Krim , L. 2011, Chem. Phys., 380, 67

    2011

  48. [56]

    2018, , 855, 11pp

    Shimonishi , T., Nakatani , N., Furuya , K., & Hama , T. 2018, , 855, 11pp

    2018

  49. [57]

    2019, , 487, 1269–1282

    Sipilä , O., Caselli , P., Redaelli , E., Juvela , M., & Bizzocchi , L. 2019, , 487, 1269–1282

    2019

  50. [58]

    G., Sellgren , K., & Tokunaga , A

    Smith , R. G., Sellgren , K., & Tokunaga , A. T. 1989, , 344, 413

    1989

  51. [59]

    & Kästner , J

    Song , L. & Kästner , J. 2016, PCCP, 18, 29278

    2016

  52. [60]

    & Kästner , J

    Song , L. & Kästner , J. 2017, , 850, 9pp

    2017

  53. [61]

    & Grimme , S

    Sure , R. & Grimme , S. 2013, J. Comput. Chem., 34, 1672

    2013

  54. [62]

    Svensson , M., Humbel , S., Froese , R. D. J., et al. 1996, J. Phys. Chem., 100, 19357

    1996

  55. [63]

    Tielens , A. G. G. M. & Hagen , W. 1982, , 114, 245

    1982

  56. [64]

    2022, ACS Earth Space Chem., 6, 1514

    Tinacci , L., Germain , A., Pantaleone , S., et al. 2022, ACS Earth Space Chem., 6, 1514

    2022

  57. [65]

    S., Komáromi , I., et al

    Vreven , T., Bryun , K. S., Komáromi , I., et al. 2006, J. Chem. Theory Comput., 2, 815

    2006

  58. [66]

    2017, Mol

    Wakelam , V., Loison , J.-C., Mereau , R., & Ruaud , M. 2017, Mol. Astrophys., 6, 22

    2017

  59. [67]

    Watson , W. D. & Salpeter , E. E. 1972, , 174, 321

    1972

  60. [68]

    I., Boogert , A

    Öberg , K. I., Boogert , A. C. A., Pontoppidan , K. M., et al. 2011, , 740, 109

    2011

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.