{"id":"70632d53-c93a-4806-b2f7-fb2a1088aed6","arxiv_id":"2506.11577","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Scattering resonances in inelastic ND3-H2 and ND3-HD collisions were resolved experimentally and matched by theory only when CCSDT(Q)-level corrections were added to the potential energy surface.","lead":"Researchers observed quantum scattering resonances in collisions between ND3 ammonia and H2 or HD at collision energies below 25 cm-1. The result extends resonance experiments from four-atom to six-atom systems and introduces a low-recoil VUV detection scheme for ND3.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claim that the data 'could only be reproduced' with the CCSDT(Q)-corrected PES is undercut by the paper's own SI showing a 2% scaled Maret PES recovers almost exactly the same cross sections; the experimental resonance observation itself remains solid.","rationale":"The paper has substantial independent support in its experimental core: three mutually consistent energy calibration methods, state-pure beams, near-recoil-free VUV ionization, fully converged close-coupling scattering calculations, a detailed partial-wave assignment, and deposited data. The stress-test concern is not about the reality or resolution of the resonances; it is about the theoretical uniqueness statement in the abstract. The authors' own SI Note 4.6 states that the 1.02-scaled Maret PES reproduces almost exactly the CCSD(T)+ET(Q) results. That internal admission means the experimental data do not uniquely require the CCSDT(Q) correction as implemented. The orientation-averaged radial correction is also a genuine uncertainty: only a handful of orientations were used to define a single exponential correction, and its transferability to the full angular expansion is untested. If the correction has hidden orientation dependence, resonance positions could shift by more than the quoted 0.1-2.5 cm^-1 energy resolution. However, the reader's weakest assumption (transferability of the orientation-averaged correction) and the scaled-PES degeneracy are closely related: both attack the same 'only CCSDT(Q)' claim. The reader already issued CONDITIONAL, and this stress-test confirms that conditional status rather than changing it. The requested concrete test--comparing the 1.02-scaled PES against the CCSDT(Q)-corrected PES with a real fit metric--would settle whether the uniqueness claim can be retained or must be softened.","tokens_in":39777,"tokens_out":4247,"duration_ms":45192,"concrete_test":"Using the deposited experimental ICS and DCS data (DANS DOI in the paper), convolute the Maret PES globally scaled by 1.02 with the experimental energy spread and the flux-to-density Monte Carlo simulation, and compute a quantitative goodness-of-fit (reduced chi-square of the 20 angular distributions plus position and magnitude residuals of the three resonance features) against the CCSD(T)+ET(Q) PES. If the scaled PES fits within the 95% confidence intervals as well as the CCSDT(Q)-corrected PES, the 'could only be reproduced' claim must be revised; if the scaled PES is statistically rejected by the imaging or ICS data, the uniqueness claim is restored.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest wording in the paper is the abstract's claim that the experimental data 'could only be reproduced' using the CCSD(T)+CCSDT(Q)-corrected PES. This uniqueness claim is directly weakened by the authors' own SI. In Suppl. Note 4.6 they write that with the CCSDT(Q)-corrected PES 'we recover almost exactly the results obtained with the reference PES of Ref. [36] scaled by 2%,' and Suppl. Fig. 14 shows that a 1.02 global scaling of the Maret PES substantially improves the resonance positions compared to the unscaled PES. Thus the experimental ICS and DCS data do not uniquely select the CCSDT(Q) level: an overall 2% well-depth scaling of a lower-level PES appears statistically indistinguishable. This matters because the 'only' wording carries the central theoretical conclusion. A second, related issue is the construction of the correction itself: E_CCSDT(Q)-CCSD(T) = 360 exp(-1.43R) cm^-1 is a single radial function fitted to CCSDT(Q)-CCSD(T) differences at the aVDZ level for 'a few fixed angular coordinates' and then applied uniformly to the full 5D PES. No test is reported for how a full orientation-resolved correction would alter the anisotropic terms of the PES expansion, which are what drive the resonance positions. The experimental advance is not in doubt: the energy calibration is checked by three independent methods, the partial-wave analysis is detailed, and the VMI images show clear resonance-induced angular changes. What is not established is the theoretical-uniqueness claim. The paper should either soften 'only' or quantify with a fit metric that the 1.02-scaled PES is actually rejected by the data.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports a joint experimental and theoretical study of state-to-state inelastic collisions ND3(1-1 -> 1+1) + H2/HD in the collision-energy range 0.5-25 cm-1. Using Stark-decelerated ND3, a cryogenic hydrogen beam, and a new near-recoil-free VUV 1+1' REMPI detection scheme, the authors resolve resonance features in integral cross sections and energy-dependent differential cross sections, and characterize the resonances through close-coupling calculations and partial-wave analysis. The central claim is that the experimental data could only be reproduced with a new CCSD(T)/aVTZ+mb PES augmented by an isotropic CCSDT(Q)-CCSD(T) radial correction.","tokens_in":40072,"tokens_out":7051,"duration_ms":73513,"significance":"The experimental advance is substantial: this appears to be the first resolved scattering-resonance study for a six-atom system, and it extends high-resolution VMI imaging beyond NO-containing benchmark systems. The measurement is supported by careful calibration via three independent methods, state-purity checks, and Monte Carlo flux-to-density corrections, and the data are deposited for reuse. The theoretical interpretation is more fragile: the paper's own SI shows that a uniform 2% scaling of an earlier lower-level PES reproduces almost exactly the same cross sections as the new CCSDT(Q)-corrected PES, so the abstract's uniqueness claim is not established. The resonance assignments and partial-wave characterizations remain valuable and are largely independent of that uniqueness claim.","major_comments":[{"comment":"The abstract's claim that the experimental data could only be reproduced with the CCSDT(Q)-corrected PES is not supported by the evidence presented in the manuscript and its SI. The main text (Results, discussion of Fig. 1) concedes that an intensity mismatch across the sampled collision energies remained even for this PES, and SI Note 4.6 states that with this PES the authors recover almost exactly the results obtained with the reference PES of Ref. [36] scaled by 2%. Because Suppl. Fig. 14 shows that a constant 1.02 scaling already brings the Maret PES resonance positions substantially closer to experiment, the ICS data do not uniquely select the CCSDT(Q) level. The authors should temper the uniqueness language in the abstract, Results, and Discussion, and should quantify the difference between the ET(Q)-corrected and 1.02-scaled Maret predictions, for example by reporting residuals or a chi-square comparison of the convoluted ICS.","section":"Abstract; Results (Fig. 1); SI Note 4.6"},{"comment":"The correction E_CCSDT(Q)-CCSD(T) = 360 exp(-1.43R) cm-1 is derived from CCSDT(Q)-CCSD(T) differences at the aVDZ level for a few fixed angular coordinates and then added as a purely radial, isotropic term to the full 5D PES. The SI reports no test of how an orientation-resolved correction would change the angular expansion coefficients in Eq. S12, which are the terms that drive resonance positions and partial-wave mixing. The correction is also computed in aVDZ while being applied to an aVTZ+mb PES. The authors should explicitly identify this as an approximation and limitation in the main text, and should avoid claiming that the data require the CCSDT(Q) level without addressing the transferability of the radial-only correction.","section":"SI Note 4.6"},{"comment":"The statement in Results that only the CCSD(T)+ET(Q) PES gave good agreement is supported only by visual comparison of experimental and simulated images; no quantitative metric is given for the DCS agreement, and the text itself reports a deviation in the backscattered region at 3.4 and 4.1 cm-1 attributed to a small resonance-position shift. Since the 1.02-scaled Maret PES was apparently not used to generate simulated DCS images, the visual comparison cannot establish that the ET(Q)-corrected PES is uniquely required. A quantitative comparison of DCS residuals for the candidate PESs would strengthen the theoretical conclusion.","section":"Results, DCS comparison (Figs. 3-4)"}],"minor_comments":[{"comment":"There are several typos in the main text: accross should be across, possiblities should be possibilities, unprecendented should be unprecedented, exquisit should be exquisite, and obervations should be observations.","section":"Throughout"},{"comment":"Many references list the journal as J. Comp. Phys. where J. Chem. Phys. is clearly intended (e.g., refs. 1-4, 22-26, 47-48, 57, 63-64, and 79); the bibliography should be corrected.","section":"References"},{"comment":"The SI should specify how many fixed angular coordinates were used for the CCSDT(Q)-CCSD(T) comparison and how the radial dependence was averaged; the statement that the results were almost identical for all orientations is qualitative and should be supported by showing the spread of the corrections.","section":"SI Note 4.6"},{"comment":"The main text states that CBS extrapolation to AVTZ and AVQZ gave incorrect results without explaining what is meant by incorrect; this is only clarified in SI Note 4.5 and would benefit from a cross-reference or a one-sentence summary in the main text.","section":"Methods; SI Note 4.5"}],"recommendation":"major_revision","confidential_remarks":"The experimental core of this paper is strong and likely suitable for a high-impact journal. The main risk is the overstrong theoretical uniqueness claim in the abstract, which the authors themselves undermine in SI Note 4.6. I would encourage the editor to require a revised and quantitative statement of what the data distinguish, while allowing the experimental resonance observation to stand as the paper's primary result."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is the first experiment to resolve scattering resonances in state-to-state inelastic collisions for a six-atom system, and the experimental work is careful enough that I trust the resonance positions. The low-recoil VUV REMPI detection for ND3 is a genuine advance, and the DCS imaging across the resonances shows the angular structure changing as theory predicts. The partial-wave analysis, including the assignment of ten resonances as Feshbach or shape resonances, is detailed and convincing. The scattering basis sets (j<=6, j2<=4, J<=16) and the convergence checks look adequate, and the energy calibration was cross-checked three ways. The citation pattern is appropriate. The soft spot is the abstract's claim that the data 'could only be reproduced' with the CCSD(T)+CCSDT(Q) PES. As the authors' own SI Note 4.6 shows, a global 1.02 scaling of the older Maret PES recovers almost exactly the same cross sections as the CCSDT(Q)-corrected PES. That means the data do not uniquely select the CCSDT(Q) level; they are consistent with a slightly deeper well of either origin. The correction itself is a single radial exponential fitted to CCSDT(Q)-CCSD(T) differences at aVDZ for a few orientations, applied uniformly to the full 5D PES, so there is real uncertainty in how the anisotropic terms shift. The main text also admits an intensity mismatch, and the flux-to-density correction depends on a Monte Carlo model. None of this casts doubt on the experimental observation, which is solid, but it does mean the theoretical-uniqueness conclusion is not established. I would send this to peer review. The right referee will ask the authors to soften the 'only' language or supply a quantitative fit metric showing the 1.02-scaled PES is actually rejected by the data. I would also ask them to deposit the PES itself; data are on DANS, but the potential is the thing people will want.","headline":"A genuinely new six-atom resonance experiment with careful data, but the CCSDT(Q) uniqueness claim is overreaching and the paper's own SI gives the reason.","tokens_in":40707,"tokens_out":2873,"would_cite":true,"duration_ms":28741,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["34.50.-s","34.50.Ez"],"model":"deepseek-v4-flash","headline":"This paper reports resolved scattering resonances in a six-atom collision system, ND3 with H2 and HD — previously resolved only for systems of up to four atoms — and shows the data are reproduced only by a potential energy surface…","keywords":["scattering resonances","inelastic collisions","ND3-H2","potential energy surface","Feshbach resonance","Stark deceleration","velocity map imaging","CCSDT(Q)"],"falsifier":"Recompute the integral cross sections with a CCSDT(Q) correction evaluated on the full grid of orientations rather than the orientation-averaged radial fit $360 e^{-1.43R}$ cm$^{-1}$: if the predicted resonance positions near 1 cm$^{-1}$ move by more than the low-energy experimental resolution of about 0.1 cm$^{-1}$, the attribution of the agreement specifically to the CCSDT(Q) level is weakened. Independently, applying a static electric field in the crossing region and measuring the Stark shift of the $E = 1.23$ cm$^{-1}$ Feshbach resonance would test its assigned $2^{-}_{1}$, $\\ell = 4$ quasi-bound character directly.","tokens_in":39568,"feed_emoji":"⚛️","tokens_out":31942,"duration_ms":256826,"temperature":0.7,"pith_summary":"This paper reports the first experimentally resolved scattering resonances in a six-atom collision system: state-to-state inelastic collisions of ND$_3$ with H$_2$ and HD over collision energies from 0.5 to 25 cm$^{-1}$, where a resonance is a brief quantum 'orbiting' state in which the colliding molecules temporarily trap each other and the collision probability spikes at specific energies. Until now, fully resolved resonances of this kind had been seen only in systems of up to four atoms, almost all built around the NO radical, so the result moves resonance physics into the realm of polyatomic symmetric-top molecules with strong electric dipole moments. The authors show that matching the measured resonance positions requires a newly computed potential energy surface at the CCSD(T) level of theory with an additional correction from CCSDT(Q) calculations, beyond the usual 'gold standard' of quantum chemistry; a previously available surface and several cheaper modifications all fail. If correct, this gives polyatomic quantum dynamics a new benchmark and a route toward experiments that steer collisions with external electric fields, for which ND$_3$'s near-linear Stark effect makes the resonances near 1 cm$^{-1}$ strongly field-responsive.","feed_headline":"First resonances resolved in a six-atom collision pair","feed_subtitle":"The six-atom ND3-H2 and ND3-HD data match theory only when the potential surface goes beyond the CCSD(T) gold standard.","key_machinery":"The argument is carried by a corrected potential energy surface plus a partial-wave analysis. The surface is five-dimensional, treating NH$_3$ and H$_2$ as rigid rotors, computed at the CCSD(T)/aVTZ level with midbond functions on a grid of 29,568 unique geometries, and corrected by adding a radial-only term $360 e^{-1.43R}$ cm$^{-1}$, obtained by averaging the CCSDT(Q)/aVDZ minus CCSD(T)/aVDZ difference over a few fixed orientations; this deepens the surface by about 2% and shifts the resonances downward into agreement with experiment. The scattering calculations solve the close-coupling equations in the body-fixed frame with a rotational basis up to $j = 6$ for ND$_3$ and $j_2 = 4$ for H$_2$/HD. The resonances are assigned by decomposing the scattering wave function in the conserved total angular momentum $J$ and parity $P$, reading off the entrance partial wave, the resonant partial wave $\\ell_{res}$ of the quasi-bound state, and the exit partial wave; for the $1^{-}_{1} \\to 1^{+}_{1}$ transition $\\ell$ must change from even to odd or vice versa. Experimentally, the enabling mechanism is a 1+1′ REMPI scheme using vacuum-ultraviolet light that ionizes ND$_3$ with near-zero recoil, keeping the velocity-map images sharp enough to resolve each resonance's angular fingerprint.","core_discovery":"The paper claims that scattering resonances in the integral cross sections of the inversion-deexcitation transition ND$_3$($1^{-}_{1} \\to 1^{+}_{1}$), induced by collisions with para-H$_2$ ($j_2 = 0$) or HD, are experimentally resolved across 0.5–25 cm$^{-1}$: three resonance features for each system, and a rapidly evolving angular structure in the differential cross sections, imaged with a newly developed near-recoil-free VUV 1+1′ REMPI scheme. The measured data can be reproduced only by close-coupling calculations on a new five-dimensional potential energy surface built at the CCSD(T)/aVTZ+mb level and augmented by a radial correction derived from CCSDT(Q)/aVDZ calculations; the previously available Maret surface, and all cheaper modifications tried (global scaling, shorter N–H bond length, explicit umbrella coordinate, different basis sets), fail to place the resonances correctly. A partial-wave analysis of the scattering wave functions characterizes the ten most prominent ND$_3$–H$_2$ resonances: nearly all are Feshbach resonances in which the pair temporarily occupies the asymptotically closed $2^{-}_{1}$ inversion-rotation level of ND$_3$, the two at 7.87 and 7.97 cm$^{-1}$ are shape resonances (centrifugal-barrier trapping), and the one at 14.47 cm$^{-1}$ is a combined Feshbach-shape resonance.","pith_inferences":["The paper notes that a uniform 2% scaling of the older potential nearly reproduces the CCSDT(Q) results, which suggests the main effect of the higher-order correction is an overall well-depth shift; a testable consequence is that a full-surface CCSDT(Q) calculation should preserve the resonance pattern and move positions only slightly, whereas a reshuffled pattern would point to the orientation-av","The same apparatus could map the reverse $1^{+}_{1} \\to 1^{-}_{1}$ transition or the isotopologues ND$_3$–D$_2$ and NH$_3$–H$_2$: since each resonance here is assigned to a specific closed rotational channel of ND$_3$, comparing isotope-shifted resonance patterns would separate mass and rotational-constant effects from electronic-structure error in the potential.","The Feshbach assignments carry a direct, testable prediction for field control: the closed $2^{-}_{1}$ level and the open $1^{\\pm}_{1}$ channels have different Stark shifts, so an applied electric field should move the 1.2 cm$^{-1}$ resonance in a calculable way — a first step toward engineering polyatomic resonances with external fields.","Because the older surface is the one used in astrochemical models of ammonia inversion-line thermometry, the corrected resonance positions imply that low-temperature collisional excitation rates for ammonia in cold interstellar clouds may shift by a few percent, and the same crossed-beam method could measure those temperature-relevant rates directly."],"forward_implications":["Resonance-resolved collision studies now reach six-atom systems: the combination of a Stark decelerator, a cryogenic secondary beam at low crossing angle, and near-recoil-free VUV ionization worked for ND$_3$, a strongly polar symmetric top, and this recipe is portable to other polar molecules that previously lacked imaging-compatible detection schemes.","Predicting low-energy resonance positions for polyatomic complexes requires going beyond the CCSD(T) gold standard: only the surface with the CCSDT(Q) radial correction reproduced the measured integral cross sections, and simulated images based on the older surface did not capture the energy evolution of the differential cross sections.","Because ND$_3$ has a strong, near-linear Stark effect, the resonances near 1 cm$^{-1}$ are expected to respond sensitively to experimentally attainable electric fields of about 75 kV/cm, enabling experiments that modify the collision dynamics with external fields.","The demonstration of a recoil-free imaging scheme for ND$_3$ removes a long-standing detection bottleneck, so differential cross sections through resonances can now be probed for molecules beyond NO, not just integral cross sections."],"supporting_citations":[{"why":"The crossed-beam apparatus with Stark deceleration and velocity map imaging on which these experiments are built; the NO-He study where CCSDT(Q)-level corrections were first required.","marker":"[24]"},{"why":"The previously published CCSD(T)-level potential of Maret et al.; its predicted resonances sit at higher energies than measured, defining the baseline the new surface must beat.","marker":"[66]"},{"why":"The companion NO study in which only CCSDT(Q)-corrected theory reproduced resolved resonances, the direct precedent for the correction level claimed here.","marker":"[26]"},{"why":"The near-recoil-free VUV 1+1' REMPI detection scheme for ND3 that made high-resolution velocity-map images of the differential cross sections possible.","marker":"[67]"},{"why":"The close-coupling body-fixed scattering methodology for NH3/ND3 with H2/D2 used to compute the integral and differential cross sections.","marker":"[68]"}],"fun_headline_variants":["Imaging scattering resonances in six-atom collisions","First resolved resonances in a six-atom pair","Quantum resonances imaged in polyatomic scattering","Six-atom collision resonances brought into focus"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that only the CCSDT(Q)-corrected surface works rests on the assumption that a single orientation-averaged radial correction, $E = 360 e^{-1.43R}$ cm$^{-1}$ fitted from CCSDT(Q)/aVDZ calculations at a few fixed orientations, remains representative of the higher-order correlation effect across the entire five-dimensional potential surface.","fun_headline_variants_meta":{"raw":{"variants":["Imaging scattering resonances in six-atom collisions","First resolved resonances in a six-atom pair","Quantum resonances imaged in polyatomic scattering","Six-atom collision resonances brought into focus"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000327,"raw_usage":{"total_tokens":1875,"prompt_tokens":1037,"completion_tokens":838,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":653,"completion_tokens_details":{"reasoning_tokens":777}},"tokens_in":653,"tokens_out":838,"duration_ms":9127,"temperature":1.0,"reasoning_tokens":777,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:04:30.482604+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the integral cross sections with a CCSDT(Q) correction evaluated on the full grid of orientations rather than the orientation-averaged radial fit $360 e^{-1.43R}$ cm$^{-1}$: if the predicted resonance positions near 1 cm$^{-1}$ move by more than the low-energy experimental resolution of about 0.1 cm$^{-1}$, the attribution of the agreement specifically to the CCSDT(Q) level is weakened. Independently, applying a static electric field in the crossing region and measuring the Stark shift of the $E = 1.23$ cm$^{-1}$ Feshbach resonance would test its assigned $2^{-}_{1}$, $\\ell = 4$ quasi-bound character directly.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The crossed-beam apparatus with Stark deceleration and velocity map imaging on which these experiments are built; the NO-He study where CCSDT(Q)-level corrections were first required."}],"review_version":1}