{"id":"cdcfb496-7e72-4b13-aed3-1e9de3f03fff","arxiv_id":"2501.09094","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Pumping the upper polariton of liquid methane can transiently excite the IR-inactive symmetric bending mode, and the effect is strongest when the polariton is two-thirds photonic in character.","lead":"Simulations of liquid methane in an infrared cavity show that pumping the hybrid light-matter upper polariton can temporarily dump more energy into an infrared-inactive bending vibration than into the directly excited infrared-active one. The finding could matter for cavity-controlled photochemistry because it suggests that strong coupling can reach molecular states that plain infrared light cannot.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The v2-v4 coupling constants in Eq. (1b) are never evaluated from the simulated potentials, so the golden-rule agreement is not a quantitative validation; the two-potential consistency does not rule out a shared classical-potential artifact.","rationale":"The reader identified the force-field description of v2-v4 couplings as the weakest assumption. I agree and sharpen the concern: the paper never quantifies Xi24 and Z24 from the actual potentials, so the golden-rule agreement is only qualitative and could mask a model artifact. The two-potential consistency is a real piece of supporting evidence, as is the isotope control and the 2/3-rule for the turnover; these strongly suggest the effect is not a numerical accident. However, the central claim would be on firmer ground if the coupling constants were computed and the golden-rule rate were shown to match the simulated decay rate quantitatively. The proposed test is feasible with the existing potentials and would settle whether the v2 selectivity is a genuine consequence of the force-field anharmonicity or an artifact of the projection/analysis. Therefore the verdict should remain CONDITIONAL as issued by the reader, and no change to the verdict is needed.","tokens_in":28657,"tokens_out":7352,"duration_ms":85954,"concrete_test":"Extract the v2-v4 coupling constants directly from the COMPASS and GAP-ML potentials. Fit the potential energy surface along the v2 and v4 symmetry coordinates (or the normal modes from Ref. 68) around the equilibrium geometry to obtain the cubic anharmonic coupling xi24, and compute the Coriolis constant zeta24 from the rotational kinetic energy in the same coordinate frame. Then evaluate the golden-rule rate gamma_UP->v2 in Eq. (1b) and compare its magnitude and frequency dependence with the simulated UP decay-rate peak in Fig. 1d. If the computed golden-rule peak is much weaker than the simulated peak, the mechanism is not quantitatively supported by the potentials; if it agrees, the central claim gains a quantitative anchor. Additionally, recompute the symmetry-coordinate dynamics of Fig. 2 using the alternative normal coordinates of Ref.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim depends on the v2-v4 intramolecular couplings in the simulated classical potentials. The paper presents the Fermi golden rule rate in Eq. (1b) with parameters Xi24 (anharmonic) and Z24 (Coriolis), but these parameters are never extracted from COMPASS or GAP-ML. The comparison is only that the simulated decay rate has a peak when the UP frequency approaches approximately 1510 cm-1, and that the symmetry-coordinate dynamics show v2 exceeding v4 at large coupling. This leaves the quantitative link between the simulation and the analytical theory unverified: the same peak could in principle arise from other anharmonic terms not included in the golden-rule model, or from the way the symmetry coordinates mix v2 and v4 (Ref. 68, Wang and Carrington, warns about deficiencies of these coordinates). Both potentials are classical nuclear potentials; if both misrepresent the v2-v4 anharmonic/Couplings, the selectivity is a force-field artifact rather than a property of methane. The GAP-ML potential is trained on first-principles energies and therefore should reduce this risk, but the absence of a direct evaluation of Xi24 and Z24 from either potential means the qualitative agreement is not a quantitative check. The lack of error bars on the v2>v4 crossover in Figs. 2f-j and 2k-o further hampers assessment of the significance of the effect.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses classical cavity molecular dynamics (CavMD) simulations of liquid methane under vibrational strong coupling to show that pumping the upper polariton formed by the IR-active asymmetric bending mode (v4) can transiently deposit more energy into the IR-inactive symmetric bending mode (v2) than into the v4 mode itself. The central observation is supported by a battery of protocols: COMPASS force-field and GAP machine-learning potentials, CD4 isotope control, reduced molecular density, both cw molecular pumping and Gaussian cavity pumping with a lossy cavity, and N-scalability checks. The authors complement the simulations with a Fermi golden-rule model (Eq. (1)) and derive a condition for optimal v2 excitation efficiency, |X_c|^2 = 2/3 (Eq. (2)), which they test by varying the cavity frequency at fixed UP frequency. They argue that polariton formation, not direct IR excitation, is required for the selective v2 accumulation, and that this mechanism acts on timescales longer than the polariton lifetime.","tokens_in":28944,"tokens_out":3974,"duration_ms":44481,"significance":"If the central claim survives scrutiny, it is significant: it demonstrates a concrete mechanism whereby pumping a vibrational polariton can access a symmetry-forbidden dark mode, going beyond the 'optical filter' picture of vibrational strong coupling and providing a design principle for selectively driving IR-inactive vibrations. The strength of the paper is its multi-pronged validation: the v2 > v4 transient is directly observed in molecular dynamics, reproduced with a machine-learning potential trained on first-principles energies, confirmed in a CD4 isotope control, and robust to the excitation scheme (cw molecular pumping versus Gaussian cavity pumping with a lossy cavity). The analytical formula Eq. (2) gives a parameter-free prediction for the optimal light-matter mixing given a fixed UP frequency, and the numerical maximum in Fig. 4f is close to the predicted value. These strengths make the central observation credible despite the use of classical nuclear potentials. However, the quantitative connection between the simulated dynamics and the golden-rule parameters in Eq. (1) remains incomplete, and the statistical significance of the key crossover is not demonstrated.","major_comments":[{"comment":"The golden-rule rates in Eq. (1) contain the intramolecular couplings Xi24 and Z24 (and Xi44 and Delta_dd), but none of these parameters are extracted from the COMPASS or GAP-ML potentials used in the simulations. The agreement reported in Fig. 3 is only about the parametric dependence on Nsimu and density, not a quantitative validation of the rate itself. Because the central mechanistic claim is that v2-v4 anharmonic and Coriolis couplings mediate the selective transfer, the authors should evaluate these couplings from the simulated potentials (or from ab initio calculations for a single methane molecule) and compare the predicted gamma_UPv4->v2 with the simulated decay rates. Without this, the golden-rule analysis is illustrative rather than confirmatory, and the possibility remains that a different anharmonic term or the coordinate definitions, rather than the assumed Xi24/Z24 couplings, produce the simulated selectivity.","section":"Sec. II.D-II.E and SI Sec. I.B"},{"comment":"The central quantitative claim is that the v2 excitation 'exceeds' the v4 excitation within 5 ps at large coupling, but no error bars or confidence intervals are provided for the integrated peak intensities or the symmetry-coordinate energies. Since these are averages over 40 trajectories, the statistical significance of the crossover (e.g., in Fig. 2o) should be quantified with standard errors or bootstrap intervals. This is particularly important because the magnitude of the v2-v4 difference appears modest in some panels, and the reader cannot judge whether the effect is robust to trajectory-to-trajectory fluctuations.","section":"Sec. II.B, Figs. 2f-j and 2k-o"},{"comment":"The v2 and v4 populations are extracted from the bend symmetry coordinates of Eq. (S39), yet the authors themselves cite Ref. 68 (Wang and Carrington, J. Chem. Phys. 118, 6260 (2003)), which identifies deficiencies in exactly these bend symmetry coordinates for methane. If the coordinates mix v2 and v4 character, the quantitative energy partition between v2 and v4 could be biased. The time-resolved bending-angle spectra in Figs. 2a-e partially mitigate this concern because they show a distinct peak near 1510 cm-1 without relying on the symmetry coordinates, but the main quantitative v2 > v4 statement in Figs. 2k-o does rely on the deficient coordinates. The authors should either justify the adequacy of these coordinates for the present purpose, use an alternative coordinate set, or quantify the mixing and show that it does not affect the conclusion.","section":"Sec. II.C and SI Sec. III.E"}],"minor_comments":[{"comment":"The word 'Guassian' in the caption of Fig. S15 should be 'Gaussian'.","section":"SI Sec. IV.J, caption of Fig. S15"},{"comment":"The decay rates are plotted as points without error bars; since three pulse fluences are shown, including per-fluence uncertainties from the exponential fits would help the reader assess the double-peak structure.","section":"Fig. 1d"},{"comment":"The notation Nnn for the number of nearest neighbors within a degenerate mode is confusing; it resembles the molecular number N and should be renamed (e.g., n_nn or z) to avoid confusion.","section":"Eq. (S17) and related equations"},{"comment":"The proposed UPv4 + v2 -> v3 pathway is supported only by qualitative temporal arguments; a sentence noting that a quantitative rate estimate would require higher-order anharmonic coefficients would help calibrate the reader's confidence in this side channel.","section":"Sec. II.C, paragraph on v3 excitation"},{"comment":"The caption of Fig. 4f mentions 'v2+v3' in magenta, but the main-text discussion of the v2+v3 maximum energy gain would benefidd from a brief definition of what is summed, since v3 is a stretching mode and v2 is a bending mode.","section":"Sec. III.B, Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and makes a substantive contribution to the polariton-chemistry debate about whether polaritons act as optical filters. The central observation is not a model artifact: it is direct MD dynamics, reproduced by a first-principles-trained GAP potential, and supported by isotope and density controls. My main concern is the quantitative link to the golden-rule theory: the coupling parameters in Eq. (1) are never extracted from the potentials, so the 'qualitative agreement' claim is weaker than the text implies. The lack of error bars on the key crossover is a secondary but important issue for a quantitative journal. I would recommend major revision rather than rejection because the missing analysis is feasible within the scope of the manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the central observation is real and worth taking seriously. Pumping the upper polariton built from methane's IR-active v4 bend can transiently put more energy into the IR-inactive v2 bend than into v4 itself. That is a concrete counterexample to the optical-filter picture, and the authors support it with more controls than we usually see in CavMD papers.\n\nWhat's new: the specific v2 selectivity in liquid CH4, the assignment of anharmonic and Coriolis pathways, and a parameter-free 2/3 photonic-weight rule that predicts the turnover in Fig 4f and matches the simulated maximum. The N-invariance test in Fig 3a is a nice check showing the simulated rates don't depend on the number of molecules, which is what a macroscopic cavity should look like.\n\nWhat's done well: they use two independent potentials (COMPASS and a GAP machine-learning potential), an isotope control (CD4), density variation, and both cw molecular pumping and lossy-cavity Gaussian pumping. The main observation is directly from the MD trajectories, not manufactured by the analytical model. The SI derivation is careful.\n\nSoft spots: the analytical golden-rule coupling constants Xi24 and Z24 are never evaluated from the simulated potentials, so the agreement between Eq. (1) and the simulation is qualitative rather than a quantitative validation. The v2>v4 crossover in Figs 2f-j and 2k-o has no error bars. Both potentials are classical nuclear potentials; GAP reduces the risk of a shared force-field artifact in the v2-v4 couplings but doesn't eliminate it. The symmetry coordinates have known deficiencies (Ref 68), though the bending-angle spectra provide a cross-check.\n\nThese are real but addressable weaknesses. They don't overturn the central result, and the paper is honest about its limitations.\n\nBottom line: this is for the polariton chemistry and dynamics crowd. It deserves a serious referee. If I were handling it, I'd send it out and ask the authors to extract Xi24 and Z24 from the potentials and report ensemble statistics for the v2/v4 crossover. That would turn a suggestive result into a quantitative one.","headline":"A solid CavMD study showing that pumping a vibrational polariton can transiently put more energy into an IR-inactive mode than the IR-active mode that forms the polariton; the core observation holds despite a qualitative analytical link.","tokens_in":29481,"tokens_out":2349,"would_cite":true,"duration_ms":24070,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Cavity polaritons can transiently populate the symmetry-forbidden bending mode of methane, giving more energy to a mode IR light alone cannot excite.","keywords":["vibrational strong coupling","vibrational polaritons","IR-inactive modes","liquid methane","cavity molecular dynamics","Fermi's golden rule","intramolecular vibrational energy redistribution","upper polariton"],"falsifier":"Measure the polariton decay rate and the transient v2 population of liquid methane (or CH4 in an inert solvent) as the UP frequency is tuned across 1500 $cm^{-1}$ with a Gaussian pump. The paper predicts a second peak in the decay rate and a v2 transient exceeding the v4 transient within about 5 ps; if neither appears, or if a higher-level anharmonic surface calculation shows the v2-v4 couplings are negligible, the mechanism is refuted.","tokens_in":28445,"feed_emoji":"🔬","tokens_out":6905,"duration_ms":67714,"temperature":0.7,"pith_summary":"This paper uses classical cavity molecular dynamics simulations to argue that pumping the upper polariton (UP) formed by the infrared-active v4 asymmetric bending mode of liquid methane can transiently deposit more energy into the infrared-inactive v2 symmetric bending mode than into the v4 mode itself. Ordinary IR light cannot reach v2 because it is symmetry-forbidden, so the claim is that polariton formation opens a route to selectively drive a dark vibration. The authors find the effect in both an empirical force field and a machine-learning potential, and they reproduce the polariton decay rates with a Fermi's golden rule expression in which intramolecular anharmonic and Coriolis couplings carry the v2 channel. They further show that the v2 excitation is maximized when the UP is about two-thirds photonic and one-third molecular. If right, the mechanism implies that polariton pumping creates unusual vibrational population distributions on timescales longer than the polariton lifetime.","feed_headline":"Cavity light selectively heats a dark bending mode in methane","feed_subtitle":"Simulations show the IR-inactive symmetric bend of CH4 gains more energy than the IR-active mode that formed the polariton.","key_machinery":"The central object is the upper polariton (UP) formed when a cavity photon hybridizes with the collective bright mode of methane's triply degenerate, IR-active v4 asymmetric bending vibration. The transfer to the IR-inactive v2 mode is quantified by Fermi's golden rule rates, Eq. (1), in which the v2 channel is driven by the intramolecular anharmonic coupling Xi_24 and the Coriolis coupling Z_24 weighted by the molecular (bright-mode) weight |X_B|^2 and a spectral-overlap integral J_UP,v2. The simulations decompose molecular motion into Td symmetry coordinates to resolve the v1-v4 populations.","core_discovery":"The paper's central claim is that exciting the upper polariton formed by the v4 mode of methane can selectively energy-transfer into the IR-inactive v2 mode: in the simulated liquid at 110 K, once the UP frequency exceeds about 1500 $cm^{-1}$, the transient v2 vibrational energy per molecule exceeds that of the v4 mode within a few picoseconds of pumping. The effect appears after the polariton itself has decayed, through intramolecular energy redistribution mediated by the anharmonic coupling Xi_24 and the rovibrational Coriolis coupling Z_24 between the v2 and v4 transitions. The authors validate the result by repeating the simulations with a machine-learning potential and by deriving the UP decay rates from Fermi's golden rule, obtaining a second peak in the decay rate when the UP approaches the v2 frequency. They also show that pumping the bare v2 mode outside the cavity produces essentially no excitation, and that the strongest v2 accumulation occurs at a light-matter hybridization where the photonic weight is 2/3, consistent with their expression E_v2 is proportional to $E0^{2}$ |X_c|^4 |X_B|^2.","pith_inferences":["If the modulation of intramolecular couplings is the operative mechanism, similar IR-inactive-mode selectivity should be achievable in other tetrahedral or high-symmetry molecules (e.g., CD4, SiH4) by tuning a cavity to bring a polariton into resonance with the dark mode.","The 2/3 photonic-weight optimum is a design rule that could be tested in other polaritonic systems, such as exciton-polariton to triplet energy transfer, where a similar trade-off between absorption strength and transfer rate is expected.","A converged quantum-dynamics calculation (or an accurate anharmonic potential surface) that directly computes Xi_24 and Z_24 would provide a much stronger test of the mechanism than the two classical potentials used here.","Because the effect appears post-decay, time-resolved Raman or IR-pump/anti-Stokes-probe experiments on methane or CD4 under VSC could observe the predicted v2 population reversal within a few picoseconds."],"forward_implications":["UP pumping near 1500 cm^-1 in methane deposits more transient energy into v2 than v4, giving a control handle on a symmetry-forbidden mode.","Because the transfer happens after polariton decay, it provides a dark-state pathway that can act on timescales beyond the polariton lifetime, consistent with long-lived effects in strong-coupling experiments.","The v2 gain is largest at a specific hybridization (photonic weight 2/3), so cavity frequency and coupling can be tuned to optimize the selective excitation.","Lowering the molecular density prolongs the selective v2 excitation, suggesting gas-phase or dilute-solution implementations.","The qualitative agreement between two force fields and the golden-rule rates supports the robustness of the mechanism, though competing pathways such as UP + v2 to v3 also matter around resonance."],"supporting_citations":[{"why":"Introduces the CavMD scheme used for all simulations in the paper.","marker":"[23]"},{"why":"Supplies the CavMD pumping protocol and the earlier prediction of polariton-enhanced molecular nonlinear absorption that motivates the present nonequilibrium study.","marker":"[24]"},{"why":"Provides the Fermi's golden rule approach that yields the analytical UP decay rates in Eq. (1).","marker":"[26]"},{"why":"Supplies the COMPASS force field used to model liquid methane in the main simulations.","marker":"[53]"},{"why":"Supplies the machine-learning (GAP) potential used as the second, independent description of methane.","marker":"[54]"},{"why":"Establishes the Coriolis coupling between the v2 and v4 transitions of methane that appears in the golden-rule rate.","marker":"[61]"},{"why":"Experimental observation of long-timescale cavity-induced intramolecular redistribution that the paper's post-decay mechanism is meant to explain.","marker":"[6]"}],"fun_headline_variants":["Polariton pumping selectively excites methane's dark mode","How cavity light reaches vibrations infrared can't","Cavity polaritons switch on a silent bending mode in CH4","Simulations show polaritons can control IR-inactive vibrations","Dark mode activated by polariton decay in methane"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The simulated v2 selectivity is real only if the force fields (COMPASS and the machine-learning potential) reproduce the actual intramolecular anharmonic and Coriolis couplings between the v2 and v4 bending modes of methane; if those couplings are misrepresented, the effect is a simulation artifact.","fun_headline_variants_meta":{"raw":{"variants":["Polariton pumping selectively excites methane's dark mode","How cavity light reaches vibrations infrared can't","Cavity polaritons switch on a silent bending mode in CH4","Simulations show polaritons can control IR-inactive vibrations","Dark mode activated by polariton decay in methane"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001592,"raw_usage":{"total_tokens":6392,"prompt_tokens":1038,"completion_tokens":5354,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":654,"completion_tokens_details":{"reasoning_tokens":5273}},"tokens_in":654,"tokens_out":5354,"duration_ms":35949,"temperature":1.0,"reasoning_tokens":5273,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:11:01.616764+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the polariton decay rate and the transient v2 population of liquid methane (or CH4 in an inert solvent) as the UP frequency is tuned across 1500 $cm^{-1}$ with a Gaussian pump. The paper predicts a second peak in the decay rate and a v2 transient exceeding the v4 transient within about 5 ps; if neither appears, or if a higher-level anharmonic surface calculation shows the v2-v4 couplings are negligible, the mechanism is refuted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the CavMD scheme used for all simulations in the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the CavMD pumping protocol and the earlier prediction of polariton-enhanced molecular nonlinear absorption that motivates the present nonequilibrium study."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the Fermi's golden rule approach that yields the analytical UP decay rates in Eq. (1)."},{"cited_title":"Sun, COMPASS: An ab Initio Force-Field Optimized for Condensed-Phase Applications: Overview with De- tails on Alkane and Benzene Compounds, J","cited_arxiv_id":null,"evidence_quote":"Supplies the COMPASS force field used to model liquid methane in the main simulations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the machine-learning (GAP) potential used as the second, independent description of methane."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the Coriolis coupling between the v2 and v4 transitions of methane that appears in the golden-rule rate."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Experimental observation of long-timescale cavity-induced intramolecular redistribution that the paper's post-decay mechanism is meant to explain."}],"review_version":1}