{"id":"0700cd4d-e9e1-4668-81ae-257ab1092b94","arxiv_id":"2607.24239","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":2,"one_line_summary":"State-to-state ammonia collision cross sections show a local maximum whose position scales with parity splitting, evidencing suppression of dipole-dipole coupling at low energy.","lead":"Experiments on cold ammonia collisions found a local maximum in cross sections, showing effective molecular dipoles can switch off at low energy. This changes how cold polar-molecule experiments and evaporative cooling should be designed.","discovery_kind":"new_method","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"Central claim survives scrutiny: the LM, its Δ-ordering, and the low-energy dipole–dipole suppression are direct experimental observations. The least secure point is the one the reader flagged—the unconverged j=1-truncated CC basis—but it bears on quantitative details, not the headline result.","rationale":"Good-faith reading: for the central claim to hold, one needs (a) genuine collision-induced |−⟩→|+⟩ signal uncontaminated by field-induced parity mixing, (b) a real turnover in the ICS within the calibrated energy range, and (c) recoil signatures that distinguish partner-flip from partner-elastic channels. On (a): the ionization-region field is 20.6 V/cm, giving a Stark mixing energy ~10⁻³ cm⁻¹ versus Δ ≥ 0.05 cm⁻¹—comfortably small—and the guide-exit geometry is designed to keep collisions out of strong-field regions, with background cycles subtracted. On (b): continuous energy scans with statistical error bars across three isotopologue pairs plus NO–NH3, with Δ spanning a factor of 16, is strong evidence for the Δ-ordering. On (c): the dot-vs-halo contrast between ND3–NH3 and NO–NH3 at matched low energies is a clean, self-checking observable. Residual concerns—relative ICS scaled to theory, peak-position scaling derived from theory curves only, unconverged basis—are real but disclosed and quantitatively scoped. No ad hominem issues, no circularity beyond the stated normalization. I considered whether the field-mixing or energy-calibration issues were more load-bearing than the basis truncation and concluded they are not, given the numbers above. Verdict stays ACCEPT; the proposed convergence check converts the one open caveat into a settled one.","tokens_in":13018,"tokens_out":3284,"duration_ms":61864,"concrete_test":"For one system (ND3–NH3, intermediate Δ, where both the LM and the sub-peak plateau are in range), rerun the CC scattering with the rotational basis extended to include the j_k = 2 manifold (if full extension is prohibitive, use a truncated partial-wave set or the DWBA machinery already used for the secondary-LM analysis to estimate closed-channel contributions perturbatively). Compute at ~5 energies bracketing and below the LM. Compare E_peak and the d-d/d-q branching against the j = 1 results: if E_peak shifts by more than ~10% or the d-q onset moves appreciably, repeat for the other systems and refit the Fig. 3 power law; if shifts are within that band, the truncation concern is retired and the ACCEPT verdict is fully secure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I read the central claim as having two tiers. Tier 1 is experimental: a local maximum exists in the ICS for all four systems, its position moves up with Δ across a factor of ~16 in Δ (0.1 → 1.6 cm⁻¹), and below the maximum the correlated-recoil VMI images show the detected product carries little or no recoil energy (dot for ND3–NH3 at lowest energy; halo for NO–NH3), i.e., mutual parity flips are suppressed. These observations do not depend on the scattering calculations: experimental ICS are relative and scaled to theory, so absolute agreement is partly by construction, but peak positions and image sizes are shape features the normalization cannot fix. Tier 2 is theory-dependent: (i) the quantitative theory–experiment match in Fig. 2, (ii) the revised scaling law E_peak/E_dd ∝ (Δ/E_dd)^{4/3}, which is extracted purely from CC curves (Fig. 3) rather than fit to measured peak positions, and (iii) the attribution of the ND3–NH3 sub-1 cm⁻¹ plateau to a dipole-quadrupole-driven secondary LM. All three rest on a basis truncated to j_k = 1_1⁻ and 1_1⁺, with convergence explicitly not reached. The authors' defense—long-range dominance—is plausible at LM energies since higher-j channels are closed there, but closed channels still enter at second order via dipole coupling (renormalizing the effective C3), and the paper itself shows rotational excitation rings above ~20 cm⁻¹ that the truncated basis cannot describe (handled with an ad hoc isotropic DCS). Because the three systems differ in μ and d, a non-uniform basis-induced shift of computed E_peak could tilt the fitted 4/3 exponent. This weakens the precision of the scaling law and the plateau attribution if it lands, but not the existence of the LM or the d-d suppression.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript reports a joint experimental–theoretical study of state-to-state inelastic |−⟩→|+⟩ (parity-flip) collisions between ammonia isotopologues (ND3–ND3, ND3–NH3, NH3–NH3) and NO–NH3 over collision energies of 0.3–100 cm⁻¹, enabled by a novel merged guide that circumvents the impossibility of merging two beams of strongly polar molecules. The authors observe a local maximum (LM) in the integral cross section for every system studied, with the peak position increasing with the total parity splitting Δ across a factor of ~16 (0.1 to 1.6 cm⁻¹). This constitutes the first experimental confirmation of the self-polarization/switch-off mechanism predicted by Tang et al. (Ref. 24). Velocity-map-imaging measurements of correlated recoil show that below the LM the mutual parity-flip (dipole–dipole) channel is suppressed relative to single-partner flip (dipole–quadrupole) channels — a dot for ND3–NH3 at the lowest energy, a dot-plus-halo for NO–NH3. Coupled-channels (CC) calculations on ab initio PESs reproduce the observations, and re-scaling the theoretical ICSs in dipole-dipole units yields a revised universal scaling law E_peak/E_dd ∝ (Δ/E_dd)^{4/3}, replacing the earlier linear Langevin estimate.","tokens_in":13446,"tokens_out":2776,"duration_ms":43134,"significance":"If the results hold, this is a landmark measurement for cold molecular collisions: it provides the first experimental validation of a genuinely quantum, counterintuitive scattering phenomenon (effective dipole switch-off below the LM), demonstrates its universality across four systems and a factor of ~16 in Δ, and delivers a direct channel-resolved fingerprint (correlated-recoil VMI) separating dipole–dipole from dipole–quadrupole dynamics. The experimental strengths are substantial: continuous energy scans with statistical (95% CI) error bars, Monte-Carlo flux-to-density corrections, isotope-labeled beam separation, recoil-free VUV REMPI detection, and independent cross-validation on the NO–NH3 system. The consequences the authors draw — for evaporative cooling strategies, for the interpretation of previous crossed-beam results, and for field control of dipolar collisions near 1 kV/cm — are well founded and will be widely cited. The revised 4/3 scaling law is a falsifiable, parameter-free prediction extractable by other groups. Publication is warranted subject to the points below.","major_comments":[{"comment":"The CC calculations truncate the rotational basis to j_k = 1_1^- and 1_1^+ for both molecules, with the statement that 'convergence with respect to the rotational basis was not necessarily reached'. The authors justify this by long-range dominance, which is plausible at LM energies since higher-j channels are closed; however, closed channels still contribute at second order through dipole coupling, renormalizing the effective C3 and hence the LM position itself. This truncation is load-bearing for three quantitative claims: (i) the quality of the theory–experiment match in Fig. 2, (ii) the E_peak values entering the revised scaling law E_peak/E_dd ∝ (Δ/E_dd)^{4/3} in Fig. 3, and (iii) the attribution of the sub-1 cm⁻¹ ND3–NH3 plateau to a dipole-quadrupole-driven secondary LM. The manuscript itself shows rotational excitation rings above ~20 cm⁻¹ that the basis cannot describe (patched w","section":"Results and Discussion; Methods (CC basis truncation)"},{"comment":"The revised power law E_peak ∝ Δ^{4/3} μ C3^{2/3} is extracted purely from the theoretical CC curves (Fig. 3). Yet the experiment independently measures LM positions for four systems spanning Δ = 0.1–1.6 cm⁻¹, and the abstract claims the calculations 'explained the observed scaling'. The manuscript never quantitatively confronts the measured peak positions with the 4/3 law. Since experimental ICSs are scaled to theory only in overall magnitude, the peak positions are shape features independent of that normalization, and this comparison is available from data the authors already have. I ask the authors to tabulate the experimental LM positions (with collision-energy uncertainties from the horizontal error bars in Fig. 2 and from the finite energy grid) and show explicitly whether they obey E_peak ∝ Δ^{4/3} μ C3^{2/3}, or at least are consistent with it. If the experimental peak positions","section":"Fig. 3 and associated text (revised scaling law)"}],"minor_comments":[{"comment":"Fig. 2 caption: 'Experimental cross sections are scaled to the theoretical ones.' Please specify the procedure — is a single global scale factor used per system over the whole energy range, and how is it determined (least squares on a restricted energy window?)? This matters for assessing which features of the agreement are and are not by construction.","section":"Fig. 2 caption"},{"comment":"Fig. 4 caption: the d-d images for ND3–NH3 'were scaled for better visibility'. Please give the scaling factors, since the relative d-d vs d-q image intensities are themselves part of the argument for channel suppression.","section":"Fig. 4"},{"comment":"Fig. 3: the abscissa label appears garbled in the provided manuscript ('Δ/Eddσ/a2dd' overlapping); please check the figure rendering. Also indicate the numerical uncertainty with which the maxima of the computed ICS curves are located, as this propagates into the fitted exponent.","section":"Fig. 3"},{"comment":"Above ~20 cm⁻¹ an isotropic DCS is assumed for the inner-ring angular distributions because the truncated basis cannot describe rotational excitation. This is reasonable as a display device, but please state explicitly in the main text (not only implicitly) that the simulated images at the highest energies are not quantitative predictions of the ring structure.","section":"Results and Discussion (VMI simulations)"},{"comment":"Typos/formatting: 'unkown' (introduction); 'Theoreticalab initiomethods' (missing spaces, introduction); 'complimentary DCS measurements' should read 'complementary'; Ref. [46] contains a capitalized 'Https://doi.org'; author list shows 'Y .T.' spacing artifacts throughout. The energy-level inset in Fig. 1 would benefit from numerical Δ values for each species next to the diagram.","section":"Throughout"},{"comment":"The data availability statement promises a Zenodo deposit; please consider also depositing the CC input/channel basis definitions and the 2×2 minimal model (SM) so that the 4/3 scaling and the secondary-LM DWBA analysis are independently reproducible.","section":"Data availability"}],"recommendation":"minor_revision","confidential_remarks":"This is a strong experimental paper from the group that has essentially defined this subfield (the merged guide, VUV REMPI detection, and the prior NO–ND3 work are all theirs). The headline claims — existence of the LM, its ordering with Δ, and the low-energy suppression of mutual parity flips — are direct observations and do not hinge on the scattering theory. The two requests in my major comments (a convergence diagnostic for the truncated basis, and an explicit experimental test of the 4/3 law) require only analysis of data the authors already possess or limited additional calculations, so I view this as minor revision rather than major. I verified there is no circularity in the experiment–theory comparison beyond overall normalization. Suitable for the journal once the two quantitative points are addressed."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The headline result is experimental and solid: they finally see the local maximum in state-to-state ICS for ND3–ND3, ND3–NH3, NH3–NH3 and NO–NH3, with the peak moving up as total parity splitting Δ grows, and VMI recoil that directly shows mutual parity flips dying out below the peak. That is the thing people in cold molecules have been waiting for since the Tang et al. prediction, and the novel quadrupole-to-hexapole merger is what made the similar-dipole systems accessible.\n\nWhat they do well: continuous energy scans, isotope labeling, flux-to-density Monte Carlo, and channel-resolved images that fingerprint d–d versus d–q without needing the theory to invent the signal. The NO–NH3 revisit is a smart control. Relative ICS scaled to theory is ordinary practice here; peak positions and image sizes are shape features the scale factor cannot fake. Circularity is low—the PESs are prior literature.\n\nSoft spot, in proportion: the ammonia–ammonia CC basis is truncated to j=1 and they say convergence was not reached. That weakens the quantitative match, the revised E_peak/E_dd ∝ (Δ/E_dd)^{4/3} law (extracted from theory curves, not measured peaks), and the secondary-LM attribution for the ND3–NH3 plateau. Closed-channel renormalization of C3 is a legitimate worry. It does not erase the existence of the LM or the recoil evidence for dipole suppression. High-energy rings are handled with an ad hoc isotropic DCS—fine for illustration, not for angular detail.\n\nThis is for people who run traps, merged beams, or evaporative cooling of OH/NH3/H2CO-class molecules, and for anyone quoting Langevin rates at cold energies. Math and citation pattern look normal for the field. I would bring it to reading group, cite the experimental LM and the VMI channel separation, and treat the 4/3 exponent as provisional until a larger basis is shown. Send it to referees; it deserves the time.","headline":"First clean observation of the predicted local maximum and dipole switch-off in parity-doublet collisions; experiment carries the paper, truncated CC basis is a real but secondary caveat.","tokens_in":14355,"tokens_out":533,"would_cite":true,"duration_ms":17555,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["34.50.Ez","37.10.Mn","82.20.Xr"],"model":"grok-4.5","headline":"Cold ammonia collisions show a local maximum in cross sections because effective dipole moments switch off at low energy.","keywords":["cold polar molecules","dipole-dipole collisions","parity doublet","local maximum","ammonia isotopologues","merged molecular beams","velocity map imaging","state-to-state cross sections"],"falsifier":"Recompute the integral cross sections with a substantially larger rotational basis (or measure an independent system with known higher-j channels open) and check whether the local-maximum position and the low-energy d–d versus d–q branching still match the reported experiment and the Δ^{4/3} scaling.","tokens_in":14036,"feed_emoji":"❄️","tokens_out":1035,"duration_ms":22528,"temperature":0.7,"pith_summary":"This paper shows that collisions between polar molecules with parity doublets do not keep getting stronger as energy falls, as classical Langevin capture would predict. Using a novel beam merger, the authors measure state-to-state cross sections for ammonia isotopologues (and NO–NH3) from about 0.3 to 100 cm−1 and find a clear local maximum whose position tracks the molecules’ parity-splitting energies. Below that maximum, velocity-map images of correlated recoil show that mutual parity flips (driven by dipole–dipole coupling) are suppressed in favor of single-partner flips (dipole–quadrupole). The result matters because many cold-molecule experiments and theories assume dipoles stay “on”; if they can switch off, inelastic rates, evaporative cooling prospects, and field-control strategies all change in a concrete way.","feed_headline":"Cold ammonia dipoles switch off below a cross-section peak","feed_subtitle":"State-to-state data show mutual parity flips fade at low energy, reshaping cold-molecule collision rates","key_machinery":"The local maximum (LM) from self-polarization of parity doublets: when the dipole–dipole energy at the classical turning point exceeds the parity splittings, opposite-parity states mix and Langevin capture is recovered; when it does not, the dipoles switch off. A merged-guide beam geometry plus recoil-resolved VMI separates mutual-flip (d–d) from single-flip (d–q) channels.","core_discovery":"State-to-state integral cross sections for ND3–ND3, ND3–NH3, NH3–NH3, and NO–NH3 exhibit a local maximum whose peak energy scales with total parity splitting Δ as E_peak ∝ Δ^{4/3} μ C_3^{2/3}. At energies above and near the peak, dipole–dipole coupling drives mutual parity flips in both partners; below the peak the effective dipoles switch off, dipole–dipole channels collapse, and dipole–quadrupole single-partner flips can dominate—directly confirmed by the recoil energy left on the detected molecule in velocity-map images.","pith_inferences":["Systems without near-degenerate parity doublets (ordinary closed-shell rotors) should lack this LM and keep closer to classical dipolar capture until true ultracold Wigner thresholds.","Time-varying fields timed to the collision could toggle d–d versus d–q dominance shot-by-shot, offering a switchable inelastic channel without changing collision energy.","Secondary LMs from dipole–quadrupole coupling may set the practical floor for inelastic loss in mixed-parity-splitting pairs used for sympathetic cooling."],"forward_implications":["Inelastic rates for parity-doublet polar molecules can fall by large factors below the local maximum, contrary to Langevin E^{-2/3} scaling.","Evaporative cooling and trap-loss estimates that assume always-on dipoles need revision in the energy window around the LM.","Crossed- or merged-beam studies of two similar dipoles become harder exactly where the LM suppresses signal.","Modest external electric or magnetic fields that mix or tune the parity doublets can switch the interaction between dipole–dipole and dipole–quadrupole character at accessible energies.","The revised peak scaling E_peak ∝ Δ^{4/3} μ C_3^{2/3} gives a practical map for choosing which molecule pairs will show the LM in a given apparatus."],"fun_headline_variants":["Ammonia dipoles switch off below cross-section peak","Parity-split scaling sets cold NH3 collision peak","Dipole-dipole channels collapse at low energy","Mutual parity flips fade below E peak in NH3","State-to-state data show effective dipoles turning off"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The ammonia–ammonia scattering calculations keep only the lowest inversion-doublet levels for each molecule and still claim reliable long-range cross sections even though the rotational basis is not fully converged.","fun_headline_variants_meta":{"raw":{"variants":["Ammonia dipoles switch off below cross-section peak","Parity-split scaling sets cold NH3 collision peak","Dipole-dipole channels collapse at low energy","Mutual parity flips fade below E peak in NH3","State-to-state data show effective dipoles turning off"]},"model":"grok-4.5","effort":"low","cost_usd":0.003454,"raw_usage":{"total_tokens":1162,"prompt_tokens":825,"num_sources_used":0,"completion_tokens":61,"cost_in_usd_ticks":34544000,"prompt_tokens_details":{"text_tokens":825,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":276,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":825,"tokens_out":61,"duration_ms":5285,"temperature":1.0,"reasoning_tokens":276,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T20:13:09.918476+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Recompute the integral cross sections with a substantially larger rotational basis (or measure an independent system with known higher-j channels open) and check whether the local-maximum position and the low-energy d–d versus d–q branching still match the reported experiment and the Δ^{4/3} scaling.","supporting_citations":[],"review_version":1}