{"id":"ebd13dd8-024d-404c-bbee-5855d6dc7cae","arxiv_id":"1908.05432","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":5,"one_line_summary":"A full-dimensional ab initio potential energy surface and variational vibrational states for the CH4-Ar van der Waals complex are reported, quantifying the error of reduced-dimensionality models.","lead":"This paper builds a complete 12-dimensional quantum mechanical model of a methane molecule bound to an argon atom, including all vibrational motions. It uses this model to show that simplifying assumptions about the methane fragment, often used in studies of weakly bound complexes, miss some important quantum effects.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The highest 12D state(s), #44–46 at 99.32 cm^-1, have no convergence data and sit ~0.8 cm^-1 below the b=3 dissociation limit, inside the stated 1–2 cm^-1 ZPVE uncertainty; the 'all bound states' claim is therefore not yet established.","rationale":"Strengths: the PES fit is independent of the bound-state targets, the GM dissociation energy is reproduced within 0.3 cm^-1, and the lower 12D excitation energies show excellent b=2 to b=3 stability (Delta2 <= 0.03 cm^-1 for states 1–43), so the abstract's rms comparisons over the tabulated lower states are credible. The reader's weakest_assumption already identified the unproven ZPVE convergence; I agree with that root but think the decisive failure mode is not a uniform shift in the rms values but the near-threshold state #44–46. This state is the only one without convergence data, it sits inside the paper's own ZPVE uncertainty, and it is used to count bound states. A missing or spurious near-threshold state would weaken the 'all bound states' and dissociation claims while leaving the lower-state rms conclusions intact. The concrete radial-basis and corrected-threshold check would settle whether #44–46 is genuinely bound and converged. Because the paper's headline numbers are otherwise well supported, the right verdict is CONDITIONAL rather than REJECT or a full change to the reader's ACCEPT: authors should either supply the missing convergence/test data for the highest state or explicitly qualify the bound-state-count and dissociation statements.","tokens_in":15864,"tokens_out":19362,"duration_ms":184238,"concrete_test":"Recompute the 12D b=3 calculation for the highest states with 20 Morse radial functions instead of 13 (with a correspondingly enlarged Smolyak R grid), and separately form the corrected dissociation limit using the b=10 CH4 ZPVE and the paper's stated 1–2 cm^-1 complex-ZPVE correction. If #44–46 shifts by more than 0.05 cm^-1 with the larger radial basis, or if it lies above the corrected dissociation limit (or within its uncertainty band without a documented convergence study), then Table V is not a converged list of all bound states and the bound-state-count comparison with 3D(⟨B⟩0) is not established.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The reader's weakest assumption (the b=3 ZPVE is not demonstrably converged) has a sharper consequence than an rms shift: Table V's last row, #44–46 at 99.32 cm^-1, is the highest bound-state row and has no Delta0/Delta1/Delta2 entries and no 3D counterpart, so its convergence is undocumented. It lies roughly 0.8 cm^-1 below the b=3 dissociation limit (De(GM) = 153.13 cm^-1 plus CH4 ZPVE(b=3) = 9692.41 cm^-1 minus complex ZPVE(b=3) = 9745.40 cm^-1 gives 100.14 cm^-1). Section III.D states that the b=3 12D ZPVE is estimated to be 1–2 cm^-1 too high; with the b=10 CH4 ZPVE the corrected D0 is about 98.35 cm^-1 plus the 1–2 cm^-1 complex-ZPVE correction, i.e. 99.35–100.35 cm^-1, so the state at 99.32 cm^-1 is borderline bound. The paper claims Table V lists all bound states and uses the bound-state count to say that 3D(⟨B⟩0) reproduces the correct number of bound states. If #44–46 is an artifact of the finite methane or radial basis, the bound-state count and the dissociation/dynamics conclusions change, even though the rms values for lower states may be unaffected. This is a concrete, load-bearing gap in the 'converged 12D reference' claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the first full-dimensional (12D) variational bound-state computation for the CH4–Ar van der Waals complex. The authors construct a new permutationally invariant, full-dimensional ab initio PES (FullD-2019) from 15,995 CCSD(T)-F12b/aug-cc-pVTZ points, benchmark its dissociation energies at the global and secondary minima against composite CCSD(T)-F12b/CCSDT(Q)/core-correlation estimates, and then solve the 12D vibrational Schrödinger equation with the GENIUSH–Smolyak approach. Convergence of the intermolecular basis is reported at the 0.01 cm−1 level, and the methane-fragment basis is varied from pruning parameter b=0 to b=3. The resulting b=3 excitation energies are used as the reference to assess two rigid-monomer 3D models, finding rms deviations of 0.93 cm−1 (rigid-geometry model) and 0.32 cm−1 (rotationally adjusted model). The paper also shows that a 12D computation with only a single methane harmonic-oscillator function (b=0) already outperforms the purely rigid 3D model, with rms 0.40 cm−1, and that including ten methane functions (b=1) reduces the rms to 0.07 cm−1. The central claims are that the FullD-2019 PES is of near-spectroscopic quality and that the 12D calculations provide a valid benchmark for reduced-dimensionality models of this complex.","tokens_in":16217,"tokens_out":12287,"duration_ms":101290,"significance":"If the results hold, this is a significant methodological and systems-oriented contribution. It provides a new full-dimensional PES for a weakly bound complex of a polyatomic molecule and a rare-gas atom, and it demonstrates the feasibility of genuinely 12D variational bound-state calculations with the GENIUSH–Smolyak machinery. The PES fitting quality (rms below 1 cm−1 up to 55,000 cm−1) and the benchmark dissociation energies are carefully established, and the comparison of 3D rigid-monomer models against full-dimensional results is a direct, useful assessment of a widely used approximation. The manuscript is generally clear, with extensive convergence data for the lower states, and the reported energies and rms values are internally consistent with the tables. The main open issue concerns the highest bound state(s), whose convergence and even bound-state character are not yet fully established; this does not undermine the lower-state conclusions but does affect the claim that the table lists all bound states and the associated bound-state-count comparisons.","major_comments":[{"comment":"The highest 12D state(s), rows #44–46 at 99.32 cm−1, are the only rows without any Δ0, Δ1, or Δ2 convergence entries, and they have no counterparts in either 3D column. The text claims that Table V lists all bound states, and later uses the bound-state count to argue that the 3D(⟨B⟩0) model reproduces the correct number of bound states. For these top states, however, there is no evidence that the b=3 energy is converged with respect to the methane-fragment basis. The state also lies close to the dissociation limit: using the b=3 numbers in the text, De(GM)=153.13 cm−1, CH4 ZPVE(b=3)=9692.41 cm−1, and complex ZPVE(b=3)=9745.40 cm−1 give D0≈100.14 cm−1, while Section III.D states that the b=3 complex ZPVE is 1–2 cm−1 too high; the correspondingly corrected D0 is 99.35–100.35 cm−1, placing the 99.32 cm−1 state inside the uncertainty of the dissociation threshold. The manuscript should provide convergence data for these states, or at least a quantitative discussion of why they are bound, before asserting that the table contains all bound states and before using the bound-state count in the 3D assessment.","section":"Section III.D, Table V"},{"comment":"The statement that the 3D(⟨B⟩0) model 'reproduces the correct number of bound states' is not supported by Table V as printed. Table V shows no entries for #44–46 in either 3D column, so the 3D(⟨B⟩0) model appears to have only 44 bound states, while the 12D b=3 column lists 47. If the 3D model actually has the three extra states, the table is incomplete; if it does not, the 'correct number' claim is incorrect. This discrepancy affects the central comparison between the full-dimensional and reduced-dimensional models and should be resolved in revision.","section":"Section III.D, text vs. Table V"},{"comment":"The sentence claiming that the 3D(⟨r⟩0) model 'erroneously predicts an additional, triply degenerate, bound state below the dissociation asymptote' is difficult to reconcile with Table V, in which the 3D(⟨r⟩0) column contains no state beyond the 12D list and actually shows fewer bound states than the 12D column. The reader cannot identify the additional state or the comparison basis. Please clarify which state is meant and how it relates to the 12D results, or correct the statement so that the table and text agree.","section":"Section III.D, text"}],"minor_comments":[{"comment":"The row for states #44–46 lacks any assignment (j, nR, Γ). Even if the states are genuine, providing at least a symmetry or angular-momentum label would help the reader assess their character; if no assignment is possible, a brief explanation is needed.","section":"Table V, row #44–46"},{"comment":"Footnote b states that the intermolecular radial and angular representations are 'sufficient for converging the figures shown in the table,' but the highest-state row has no convergence figures. Please qualify the statement to indicate which rows are covered.","section":"Table V, footnote b"},{"comment":"The estimate that the b=3 ZPVE is 1–2 cm−1 too high is stated without showing the underlying comparison for the complex itself. Since this estimate drives the uncertainty in D0 and thus the bound-state interpretation of the top state, a more explicit derivation or a small convergence table for the complex ZPVE would be valuable.","section":"Section III.D, ZPVE discussion"},{"comment":"The phrase 'single harmonic oscillator basis function (9D)' is slightly confusing because a single 9D function is an entire 9D Hartree-product function, not one oscillator in one dimension. A wording such as 'a single 9D harmonic-oscillator basis function (b=0)' would be clearer.","section":"Abstract and Section I"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong contribution and the main lower-state results are convincing. The referee's disagreement with the reader's accept recommendation is confined to the treatment of the highest bound states and the bound-state count, which are load-bearing for the 'all bound states' and 3D-assessment claims. These issues are fixable either by adding convergence data for the top states or by softening the claims appropriately; they do not require redoing the entire study. I recommend major revision rather than outright rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline is right: this is the first full-dimensional (12D) PES for CH4-Ar and the first variational computation on it. The PES fit is impressive—rms under 1 cm^-1 up to 55,000 cm^-1, and the benchmark dissociation energies are reproduced within 0.3 and 2.6 cm^-1. The 12D computation using GENIUSH-Smolyak is a real technical step forward, and the comparison with 3D rigid-monomer models gives the field a clean numerical benchmark. The convergence tests for the lower states (b=0 to b=3) are careful and back the 0.01 cm^-1 claim for excitation energies up to state #43. No circularity: the PES is fit to ab initio energies, not to the target bound states.\n\nThe soft spot is exactly what the stress-test flags. States #44–46 at 99.32 cm^-1 are listed as bound states with no convergence data and no 3D counterparts. They sit ~0.8 cm^-1 below the b=3 dissociation limit, inside the stated 1–2 cm^-1 ZPVE uncertainty. The paper's claim that 3D(⟨B⟩0) 'reproduces the correct number of bound states' cannot be checked against Table V, since that column is blank for those three states. If they are artifacts of the finite methane basis, the bound-state count changes and the 3D comparison needs a rewording. The lower-state rms values are unaffected, so this is not a fatal flaw for the main results, but it is a real gap in the 'all bound states' statement.\n\nMinor points: 'spectroscopic quality' is a bit strong when the isolated methane vibrational rms is 2.88 cm^-1; good for ab initio, but not what that phrase usually means. The ESI has the PES coefficients, but the code is not released, which makes reproduction of the method itself harder.\n\nWho gets value: method developers in quantum dynamics, and anyone using reduced-dimensional models for weakly bound complexes. It deserves a serious referee. I'd recommend acceptance after the authors address the highest states—either show they are converged or exclude them from the bound-state count and adjust the 3D(B0) sentence.","headline":"A solid, first-of-its-kind 12D study with one genuine loose end: the highest 'bound' states need convergence evidence before the all-bound-states claim can stand.","tokens_in":16836,"tokens_out":9564,"would_cite":true,"duration_ms":82456,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Twelve-dimensional vibrational states of CH4-Ar set the benchmark for 3D rigid-monomer models.","keywords":["full-dimensional potential energy surface","CH4-Ar van der Waals complex","variational vibrational bound states","Smolyak sparse grid","rigid-monomer approximation","dispersion interactions","monomer flexibility effects","predissociative states"],"falsifier":"Recompute the 12D bound states with a methane basis larger than $b=3$ ($b=4$ or more) and check whether any intermolecular excitation energy shifts by more than roughly $0.05~\\mathrm{cm}^{-1}$; a larger shift would mean the claimed convergence, and hence the derived 3D-model errors, is not stable.","tokens_in":15601,"feed_emoji":"⚛️","tokens_out":8455,"duration_ms":78534,"temperature":0.7,"pith_summary":"The paper tries to establish that the weakly bound methane–argon complex can now be treated with all twelve vibrational dimensions on a single ab initio potential surface, and that the resulting full-dimensional energies give a direct benchmark for the rigid-monomer models commonly used for such van der Waals complexes. On the new FullD-2019 surface, a 12D calculation that keeps only one nine-dimensional harmonic-oscillator basis function for the methane fragment has an rms error of $0.40~\\text{cm}^{-1}$ against the converged 12D excitation energies, less than half the $0.93~\\text{cm}^{-1}$ error of a rigorously constrained 3D rigid-monomer model. Adding ten methane basis functions lowers the error to $0.07~\\text{cm}^{-1}$, clearly outperforming the best adjusted 3D model. A reader should care because dispersion-bound complexes are natural testbeds for how intermolecular forces distort and exchange energy with a floppy molecular fragment, and this paper supplies the quantum-dynamical reference against which reduced models can be checked.","feed_headline":"Full 12D CH4-Ar bound states expose 3D model errors","feed_subtitle":"A single-basis 12D run matches converged levels to 0.40 cm-1, beating rigid-monomer fits.","key_machinery":"The load-bearing machinery is the FullD-2019 potential energy surface combined with the GENIUSH\\textendash Smolyak variational procedure. The PES is a permutationally invariant polynomial of Morse-type variables $y_{ij}=\\exp(-r_{ij}/a)$ fit to 15,995 CCSD(T)-F12b/aug-cc-pVTZ points, with terms up to degree 7; the Morse variables give smooth dissociation, and the high-order terms reproduce the intermediate-range $R^{-6}$ dispersion character without a switching function. The variational procedure builds the kinetic energy operator automatically from user-defined internal coordinates and uses Smolyak sparse grids to prune the direct product basis and grid, controlling the exponential cost of 12 dimensions. A second key ingredient is the choice of generalized normal coordinates referenced to the effective ground-state methane structure, which builds in the dominant static anharmonicity and lets a tiny methane basis capture monomer-flexibility effects. Convergence of the intermolecular excitation energies relies on error cancellation between the zero-point-dominated reference and excited states.","core_discovery":"The central claim is that the full-dimensional bound-state vibrational energies of CH$_4$\\cdot Ar computed on the new FullD-2019 PES are converged enough to serve as a reference, and that relative to this reference the standard rigid-monomer approximations are accurate to about $1~\\text{cm}^{-1}$ or worse unless adjusted. The 12D computation with the smallest methane basis ($b=0$, a single 9D harmonic-oscillator function) gives intermolecular excitation energies within $0.40~\\text{cm}^{-1}$ rms of the $b=3$ reference; the $b=1$ computation with ten methane functions gives $0.07~\\text{cm}^{-1}$ rms. The rigorously constrained 3D model built at the ground-state average methane structure has $0.93~\\text{cm}^{-1}$ rms and incorrectly predicts an extra bound state, while an adjusted 3D model that tunes the methane C\\textendash H distance in the kinetic energy operator to reproduce the effective rotational constant recovers the correct count and $0.32~\\text{cm}^{-1}$ rms. The PES itself matches benchmark dissociation energies at the global and secondary minima within $0.3$ and $2.6~\\text{cm}^{-1}$, and its isolated-methane levels reproduce experimental values through the pentad with $2.88~\\text{cm}^{-1}$ rms.","pith_inferences":["The paper does not state this, but the $b=0$ result implies that a rigid-monomer model using the ground-state effective structure already captures most monomer-flexibility effects, which could be tested quickly on other atom\\textendash molecule complexes without running full-dimensional calculations.","The convergence tests are for relative excitation energies; the zero-point energy itself is estimated to be $1$\\textendash$2~\\text{cm}^{-1}$ too high at $b=3$, so absolute dissociation thresholds carry that uncertainty even though relative levels are stable.","Following the paper's stated route, repeating the 12D computation with a methane basis that includes excited normal-mode combinations ($b=6$ or $7$) and comparing predissociative line positions with measured spectra would be a sharper test of the surface and of the reduced-model assessment.","Because the degree-7 Morse-variable fit reproduces the $R^{-6}$ tail without an explicit dispersion term, the same fitting strategy may transfer to other dispersion-bound complexes; a direct test would be fitting a known long-range potential and checking whether the recovered asymptotic falloff matches $1/R^6$."],"forward_implications":["The $b=0$ 12D computation, with a single methane basis function, gives an rms of $0.40~\\text{cm}^{-1}$ against the converged $b=3$ results, making it a cheap yet accurate alternative to 3D models.","The $b=1$ 12D computation (ten methane basis functions) reaches $0.07~\\text{cm}^{-1}$ rms, much better than both 3D models, while taking only about 20 hours on 20 cores.","The rigorously constrained 3D model built at the ground-state average structure has $0.93~\\text{cm}^{-1}$ rms and predicts a spurious triply degenerate bound state; the full-dimensional calculation corrects that.","The PES reproduces benchmark dissociation energies at the global and secondary minima within $0.3$ and $2.6~\\text{cm}^{-1}$, respectively, and its isolated-methane vibrational levels match experiment to $2.88~\\text{cm}^{-1}$ rms through the pentad.","Because the surface dissociates correctly, the method can be extended to methane-excited predissociative states, where intramolecular-to-intermolecular vibrational energy transfer becomes observable."],"supporting_citations":[{"why":"Supplies the GENIUSH\\textendash Smolyak variational method used for all 12D vibrational computations.","marker":"[40]"},{"why":"Provides the permutationally invariant polynomial basis in Morse variables used to fit the FullD-2019 PES.","marker":"[49]"},{"why":"Defines the CCSD(T)-F12b electronic-structure method used to generate the 15,995 ab initio energy points.","marker":"[41]"},{"why":"Introduces the Smolyak sparse-grid scheme that prunes the direct product basis and grid in the high-dimensional computations.","marker":"[24, 25]"},{"why":"Earlier rigid-monomer 3D potential and bound states that the present full-dimensional results are compared with.","marker":"[11]"},{"why":"Experimental methane vibrational energies used to benchmark the isolated-methane levels of the PES.","marker":"[56]"}],"fun_headline_variants":["12D CH4-Ar PES: single-basis 0.40 cm-1, ten-basis 0.07","Full 12D CH4-Ar PES: 3D model misses a bound state","12D methane-Ar PES: 0.07 cm-1 vs 0.93 for 3D","CH4-Ar full-dim PES: 3D error cut to 0.07 cm-1","Methane-Ar 12D PES: 3D model off by ~1 cm-1"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the $b=3$ 12D excitation energies are converged, which relies on error cancellation between the zero-point and excited intramolecular states; if that cancellation fails, the reported rms values and the comparison with 3D models lose their reference.","fun_headline_variants_meta":{"raw":{"variants":["12D CH4-Ar PES: single-basis 0.40 cm-1, ten-basis 0.07","Full 12D CH4-Ar PES: 3D model misses a bound state","12D methane-Ar PES: 0.07 cm-1 vs 0.93 for 3D","CH4-Ar full-dim PES: 3D error cut to 0.07 cm-1","Methane-Ar 12D PES: 3D model off by ~1 cm-1"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001031,"raw_usage":{"total_tokens":4448,"prompt_tokens":1153,"completion_tokens":3295,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":769,"completion_tokens_details":{"reasoning_tokens":3161}},"tokens_in":769,"tokens_out":3295,"duration_ms":23642,"temperature":1.0,"reasoning_tokens":3161,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:14:04.293537+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the 12D bound states with a methane basis larger than $b=3$ ($b=4$ or more) and check whether any intermolecular excitation energy shifts by more than roughly $0.05~\\mathrm{cm}^{-1}$; a larger shift would mean the claimed convergence, and hence the derived 3D-model errors, is not stable.","supporting_citations":[{"cited_title":"Leforestier, Phil","cited_arxiv_id":null,"evidence_quote":"Supplies the GENIUSH\\textendash Smolyak variational method used for all 12D vibrational computations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the permutationally invariant polynomial basis in Morse variables used to fit the FullD-2019 PES."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the CCSD(T)-F12b electronic-structure method used to generate the 15,995 ab initio energy points."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier rigid-monomer 3D potential and bound states that the present full-dimensional results are compared with."},{"cited_title":"F´ abri, E","cited_arxiv_id":null,"evidence_quote":"Experimental methane vibrational energies used to benchmark the isolated-methane levels of the PES."}],"review_version":1}