{"id":"b0999e75-9f10-4702-9a3f-7c22e04ede0a","arxiv_id":"2506.23945","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Liquid-phase high-harmonic spectra show a weak second plateau, attributed to electrons recombining at the second solvation shell via off-site recombination.","lead":"Scientists observed a second plateau in high-harmonic generation from several liquids and attribute it to electrons recombining on neighboring molecules rather than at their original site. The result explains why the higher-order nonlinear response of liquids was previously hidden and points toward new attosecond probes of electron motion in solutions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The NNN recombination mechanism requires a delocalized hole at the second shell, but static MD shows a localized HOMO; dynamic hole delocalization is asserted, not demonstrated.","rationale":"The reader's weakest_assumption identified two fragile premises: cluster fidelity for the weak second plateau and dynamic hole delocalization given the localized static HOMO. I focus on the latter because it is the load-bearing link between the observed second plateau and the specific NNN mechanism. The central claim is not merely that a second plateau exists, but that it is dominated by recombination at the second solvation shell. Establishing dominance requires a quantitative account of the recombination amplitude at that shell, which in turn requires a hole there. The static HOMO being localized is an internal inconsistency with the asserted mechanism unless the laser dynamics delocalize it; that is never computed. The semi-classical trajectory model and the extended Lewenstein model both assume or parametrize the needed hole weight, so the ellipticity side peaks are not an independent verification. Still, the paper's experimental observation of a second plateau and its qualitative agreement with several observables are real and valuable; the concern is about over-interpretation of the mechanism, which the authors themselves hedge with 'most likely'. Thus the conditional verdict is appropriate and should not be changed, but the paper should be required to provide direct evidence of hole delocalization or soften the dominance claim.","tokens_in":29929,"tokens_out":4432,"duration_ms":57065,"concrete_test":"Run the liquid-water TDDFT simulation and compute the time-resolved hole density, e.g., ρ_h(r,t) = ρ_gs(r) − ρ_t(r), projected onto molecular Voronoi cells, at the return times predicted for NNN trajectories. If the second-shell hole population at those times is negligible compared to the parent-molecule population, then NNN recombination cannot dominate the second plateau, directly falsifying the central claim. If the hole is found to delocalize to the second shell within the laser pulse, the concern is resolved and the mechanism is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that the second plateau is dominated by electrons recombining at the second solvation shell. That mechanism requires a hole on a second-shell molecule at the moment of recombination, with enough amplitude to give an observable dipole. The static MD results in SM S1 (Table I) show the HOMO, i.e., the hole state, is localized with a spread of about 2 Å, while only the LUMO is delocalized (~9.5 Å). In the main text, the authors state that 'due to delocalization of the hole across neighboring molecules', off-site recombination is allowed, and later invoke 'wide hole delocalization' induced by laser driving and ionization. However, no calculation of the time-dependent hole is presented; the claim of dynamic hole delocalization is an assertion. Without it, recombination at a second-shell site is a strongly suppressed process, and the interpretation of the second plateau as NNN-dominated loses its microscopic basis. The extended Lewenstein model in SM S3 does not resolve this: it simply assumes multiple Gaussian centers with a relative weight of β=0.01, placing a small but fixed NNN amplitude into the model by hand. The resulting ellipticity side peaks therefore follow from the assumption rather than independently confirming it. The ab initio time-frequency agreement is also explicitly acknowledged as non-quantitative, with about 3 eV missing for water and an artificial time-axis shift of ~0.1T to match the numerics. For these reasons, the statement that NNN recombination 'dominates' is stronger than the evidence supports; the evidence shows consistency, not dominance.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the experimental observation of a second plateau in high-harmonic generation (HHG) from liquid water (H2O and D2O), ethanol, and isopropanol at 1500-1800 nm, with a yield roughly two to three orders of magnitude below the first plateau. This second plateau is also reproduced by ab initio TDDFT cluster calculations for liquid water, ammonia, and methane. The authors attribute the second plateau to electrons that recombine at the second solvation shell (next-nearest-neighbor, NNN) rather than at the ionization site, based on semiclassical trajectory analysis, time-frequency maps, and ellipticity-dependent measurements. An extended Lewenstein model with a delocalized hole is used to support the ellipticity side peaks, and a third plateau is predicted by the ab initio simulations.","tokens_in":30264,"tokens_out":6392,"duration_ms":66842,"significance":"If the mechanism claim holds, this would establish off-site (NNN) recombination as a distinct and observable HHG channel in liquids, opening a route to attosecond probing of electron dynamics and electronic delocalization in solutions and disordered media. The paper has substantial strengths: the experimental observation of the second plateau is documented in multiple liquids, it is independently reproduced by ab initio TDDFT across several systems, and the predicted weak wavelength and intensity scaling is confirmed in both experiment and theory. The ellipticity side peaks are a falsifiable signature that is reproduced by the ab initio data. However, the mechanistic attribution to NNN recombination relies on semiclassical trajectories with a manually adjusted time shift and an assumed delocalized hole, and the ab initio spectra are not decomposed into on-site versus off-site recombination channels. These gaps make the central 'dominated by NNN recombination' claim stronger than the present evidence.","major_comments":[{"comment":"The assignment of the second plateau to NNN recombination is largely based on semiclassical trajectories whose agreement with the ab initio time-frequency maps is obtained with a manually applied time-axis shift of ~0.1T and still leaves a ~3 eV discrepancy for water (main text; Methods). Because the recombination distance (5.3 Å) is chosen from the edge of the second peak of the radial distribution and the time shift is fitted to the numerics, the energy-range agreement is not an independent confirmation. The authors should vary the effective mass and NNN distance over physically plausible ranges and show that the NNN energy window and the conclusions are stable; otherwise the identification of the dominant trajectory family is underdetermined.","section":"II (Fig. 3) and Methods, 'Semi-classical trajectory simulations'"},{"comment":"The proposed mechanism requires a delocalized hole with significant amplitude at the second solvation shell, but the molecular-dynamics snapshots in SM S1 show that the HOMO (the hole state) is localized with a spread of about 2 Å, while only the LUMO is delocalized (~9.5 Å). The main text invokes 'wide hole delocalization' due to laser driving and ionization, but no time-dependent hole calculation is presented. The extended Lewenstein model in SM S3 does not fill this gap: it simply assumes Gaussian centers at 5.5 Å with relative weight β = 0.01, thereby placing the NNN amplitude into the model by hand. The ellipticity side peaks are therefore a consequence of this assumption, not an independent verification. The authors should provide direct evidence of laser-induced hole delocalization (e.g., time-resolved hole density from TDDFT) or explicitly state that the NNN recombination amplitude is a free parameter.","section":"SM S1, Table I; main text, 'Lastly, we consider...' and Fig. 4"},{"comment":"The statement that the cluster TDDFT calculation 'allows us to rule out macroscopic effects' as the origin of the second plateau is too strong. The finite-cluster simulation demonstrates that a purely microscopic mechanism can produce a second plateau, which is an important existence proof, but it does not exclude a macroscopic contribution (e.g., phase matching, reabsorption, or propagation) to the experimentally measured 10^-3-level yield. Ruling out macroscopic effects would require a propagation calculation or a thickness-dependent measurement; the manuscript should either provide such evidence or weaken this claim.","section":"II, after Fig. 1(d)"},{"comment":"The central claim that the second plateau is 'dominated' by NNN recombination is not directly extracted from the ab initio data. The TDDFT spectra are total harmonic yields and are not decomposed into on-site, NN, and NNN recombination contributions. The extended Lewenstein model (SM S3) is not a decomposition of the TDDFT result; it reproduces side-peak shapes only after choosing the NNN weight, spread, and number of centers. To support 'dominated', the authors should provide a quantitative site-decomposed analysis (e.g., projecting the time-dependent dipole onto recombinations at different solvation shells) or, failing that, soften the claim to 'consistent with' rather than 'dominated by'.","section":"II and SM S3"}],"minor_comments":[{"comment":"The word 'propranol' should read 'isopropanol' (or '2-propanol'); the experimental liquid is isopropanol, as shown in Extended Data Fig. 1.","section":"Abstract"},{"comment":"The caption contains a duplicated phrase: 'showing the showing the temporal dependence'.","section":"Fig. 3 caption"},{"comment":"The axis label 'Ellip � city' is garbled and should be 'Ellipticity'.","section":"SM S3, Figs. S5 and S6"},{"comment":"There is a typo in 'for the the elliptically-driven case' in the description of the Lewenstein-like SFA model.","section":"Methods, last paragraph"}],"recommendation":"major_revision","confidential_remarks":"The experimental observation of the second plateau appears credible and is strengthened by the ab initio reproduction across several liquids. The main weakness is the mechanistic attribution: the NNN recombination picture is plausible but is supported by a semiclassical model with fitted parameters and an assumed delocalized hole, while the static MD results show a localized HOMO. I would ask the authors to either provide a direct decomposition of the ab initio response by recombination site or soften the dominant-claim language. The 'rule out macroscopic effects' statement is also logically overreaching and should be revised."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The real news here is the experimental observation: a second HHG plateau in several liquids, about two to three orders of magnitude down from the first, reproducible across wavelengths and liquids, and matched qualitatively by ab initio TDDFT cluster calculations. That is a genuine new result and it answers the old conundrum of the missing higher-order response in liquid HHG. I think the observation stands, and the paper is careful about its limits.\n\nThe proposed mechanism is off-site recombination at the second solvation shell. The evidence is plausible but not as strong as the observation. The semiclassical trajectories need a fitted NNN distance (5.3 Å), a manual time-axis shift (~0.1T), and still miss the water energy by ~3 eV. The ellipticity side peaks look like a real fingerprint of off-site recombination, and the extended Lewenstein model gives the qualitative effect, but that model puts the NNN amplitude in by hand (β=0.01). The MD snapshots show a localized HOMO, so the 'wide hole delocalization' invoked for NNN dominance is more asserted than demonstrated. The authors themselves hedge: 'most likely culprit', 'not quantitative'. So I would not call the mechanism proven. I would call it a well-supported hypothesis, and the paper is honest about that.\n\nReal soft spots: no independent benchmark for the cluster TDDFT at the 10^-3 yield level, so ruling out macroscopic effects is an extrapolation from first-plateau validation; the dynamic hole delocalization lacks a time-dependent hole calculation; and the trajectory model has several free parameters. None of this is fatal to the main claim, because the existence and genericity of the second plateau do not depend on the NNN attribution.\n\nThe citation pattern is fine, including the use of the authors' own earlier cluster method. The paper is for strong-field and attosecond people working on liquids and disordered media. It deserves a serious referee. My recommendation: send it to peer review, expect heavy revision, and push the authors to either soften the mechanism claims or provide direct evidence for dynamic hole delocalization.","headline":"Solid experimental discovery of a second plateau in liquid HHG, with a plausible but unproven NNN recombination mechanism; the observation warrants peer review, the mechanism needs more evidence.","tokens_in":30818,"tokens_out":1813,"would_cite":true,"duration_ms":20351,"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":"Liquid high-harmonic generation shows a second plateau, and the paper argues it comes from electrons that recombine with molecules at the second solvation shell rather than at the molecule that released them.","keywords":["high-harmonic generation","liquid phase","second plateau","off-site recombination","solvation shell","ellipticity dependence","time-dependent density functional theory","semiclassical trajectories"],"falsifier":"Measure the second-plateau cutoff while changing the liquid's density, temperature, or solvent so that the second-solvation-shell distance moves: the analytical law predicts $E_c$ shifts linearly with $d_{\\mathrm{NNN}}$ (about $2 q E d_{\\mathrm{NNN}}$ to first order in the field) and stays essentially independent of wavelength, so a quadratic wavelength dependence or a cutoff that fails to track the measured radial distribution function would rule out next-nearest-neighbor recombination. A complementary check is the ellipticity profile: if second-plateau harmonics ever show a single Gaussian with no side peaks, the off-site channel is not the mechanism.","tokens_in":29773,"feed_emoji":"💧","tokens_out":14295,"duration_ms":135931,"temperature":0.7,"pith_summary":"Driving liquids with intense mid-infrared laser pulses produces high harmonics — photons at multiples of the laser frequency. Earlier work found that, unlike gases and solids, liquids show a first harmonic plateau whose cutoff barely moves when the laser intensity or wavelength changes, leaving it unclear where the additional absorbed energy goes. This paper reports the answer: a second, much weaker plateau appears about 5 eV beyond the first cutoff, roughly two to three orders of magnitude down in yield, in water, heavy water, ethanol, and propanol. Combining experiments, ab initio cluster simulations, and semiclassical electron-trajectory models, the authors show that this second plateau is dominated by electrons that recombine not at the molecule that released them but at a neighboring molecule — predominantly in the second solvation shell — and the same mechanism predicts yet higher plateaus. If this picture holds, the harmonic spectrum becomes a probe of how widely electrons delocalize in liquids and a ruler for intermolecular distances.","feed_headline":"500x weaker second plateau in liquids traced to distant recombination","feed_subtitle":"The missing high-order response in liquid harmonics comes from electrons rejoining at the second solvation shell.","key_machinery":"The load-bearing machinery is a real-space semiclassical trajectory model extended to off-site recombination, calibrated against ab initio cluster TDDFT simulations and time-frequency analysis. In this picture, electrons follow the three-step cycle — tunnel ionization, laser acceleration, recombination — but the return condition is imposed at the distance of the second solvation shell rather than at the parent molecule: $x(t_f) = x_{\\mathrm{NNN}}$, with $x_{\\mathrm{NNN}} \\approx 5.3$ Å in water, while a mean free path of about 3.2 Å truncates the on-site trajectories that make the first plateau. This single modification produces the second plateau's energy range, its weak scaling of the cutoff with wavelength and intensity, and an approximate analytical cutoff law $E_c \\approx 2 q E d_{\\mathrm{NNN}} - \\frac{m^2 \\omega^4 d_{\\mathrm{NNN}}^3}{q E}$. The complementary piece of machinery is the ellipticity response: the authors generalize the strong-field-approximation dipole model to a hole delocalized over several sites — a central Gaussian plus satellite Gaussians at the solvation-shell distance — which reproduces the multi-Gaussian ellipticity profiles and side peaks that experimentally distinguish second-plateau harmonics from first-plateau ones. Together, these tools single out next-nearest-neighbor recombination as the dominant second-plateau channel and tie that microscopic mechanism to the measured spectrum, which lets the authors rule out macroscopic phase-matching or propagation effects as the plateau's origin.","core_discovery":"On the paper's own terms, the discovery is that liquid-phase HHG contains a second plateau beyond the first, and that this plateau has a distinct microscopic origin. The authors argue — from the concurrence of experiment, cluster-based ab initio TDDFT simulation, time-frequency analysis, and an extended trajectory model — that the second plateau arises from electrons that tunnel-ionize, propagate a mean-free-path-limited distance, and coherently recombine not at the parent molecule but at a neighboring molecule, predominantly at the second solvation shell (the next-nearest-neighbor site, at roughly 5.3 Å in water). The first plateau is governed by on-site recombination of mean-free-path-limited short trajectories; the inter-plateau decay region is fed by recombination at the first solvation shell and by suppressed longer on-site trajectories; back-scattered and free gas-like trajectories are inconsistent with the measured energy range, scaling, and time-frequency data. The signature that singles out next-nearest-neighbor recombination is the anomalous ellipticity dependence of second-plateau harmonics — a multi-Gaussian profile with side peaks that grow with harmonic energy, reproduced by an extended strong-field-approximation model with a delocalized hole — together with a cutoff that scales weakly with wavelength and intensity. The paper also derives an approximate analytical cutoff law for off-site recombination, linear in the field and wavelength-independent to first order, and reports ab initio predictions of a third plateau that the experiment cannot yet resolve.","pith_inferences":["Inference: the analytical cutoff law implies a direct test the paper does not perform — varying density, temperature, or solvent to shift the second-solvation-shell distance should move the second cutoff linearly in field strength while leaving it wavelength-independent.","Inference: the requirement of wide hole delocalization suggests solutes that strongly hybridize with the solvent (ions, hydrogen-bonded chromophores) should display enhanced or displaced second-plateau emission, an expectation the authors state only qualitatively.","Inference: the predicted third plateau could be sought with higher dynamic-range detection or by spectrally blocking the intense lower harmonics; its measured cutoff would test the multi-shell ladder beyond the second shell.","Inference: the tension between localized equilibrium HOMO states and the invoked dynamic hole delocalization could be settled by time-resolved hole-migration simulations or pump-probe experiments that track whether off-site recombination follows the hole's spreading after ionization."],"forward_implications":["The missing higher-order nonlinear response in liquid HHG is not absent: it appears as an exponentially suppressed second plateau, about 500 times weaker, whose cutoff grows only weakly with laser intensity and wavelength.","The second-plateau cutoff and ellipticity side peaks encode the second-solvation-shell distance, so liquid HHG spectra become a way to extract effective intermolecular separations.","Because the mechanism depends on holes being delocalized across neighboring molecules, liquid HHG can act as a probe of electronic wavefunction delocalization and of hybridization between solute and solvent states.","The same trajectory picture predicts even higher plateaus from recombination at further solvation shells, implying a ladder of exponentially suppressed nonlinear responses.","Since the second plateau is a microscopic effect, similar plateaus should appear in other disordered phases with short-range order, such as amorphous solids."],"supporting_citations":[{"why":"Supplies the ab initio cluster TDDFT approach (finite clusters, frozen nuclei) that reproduces the second plateau microscopically and rules out macroscopic origins.","marker":"[73]"},{"why":"Established the mean-free-path-limited, on-site-recombination mechanism for the first plateau that this paper extends to off-site channels.","marker":"[9]"},{"why":"Prior measurement and mechanism for liquid HHG that defined the first plateau and the experimental baseline the second plateau adds to.","marker":"[8]"},{"why":"Provides the thin flat-liquid-jet technique whose signal-to-noise ratio makes the roughly 500-times-weaker second plateau observable.","marker":"[74]"},{"why":"Real-space-grid TDDFT code with which all liquid simulations were run.","marker":"[82]"},{"why":"The three-step ionize-accelerate-recombine model whose real-space trajectories the paper generalizes to off-site recombination.","marker":"[20]"},{"why":"Strong-field-approximation dipole model the paper extends to a delocalized multi-site hole to reproduce the ellipticity side peaks.","marker":"[21]"},{"why":"Links liquid HHG to the radial distribution function, supporting the use of solvation-shell distances as trajectory endpoints.","marker":"[70]"},{"why":"Gas-phase ellipticity baseline (single Gaussian) against which the multi-Gaussian liquid response is compared.","marker":"[83]"}],"fun_headline_variants":["Liquid harmonics reveal hidden second plateau from faraway recombinations","Off-site recombination explains extra plateau in liquid high harmonics","Second plateau in liquid HHG traced to second solvation shell electrons","Distant electron reunions spark higher-order plateaus in liquid harmonics","Liquid HHG's second plateau: electrons return to neighbors, not origin"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central claim stands on the finite-cluster TDDFT model reproducing the bulk liquid's harmonic response down to the roughly one-thousandth yield level of the second plateau (frozen nuclei, no independent benchmark at that weakness), and on laser-driven holes being delocalized widely enough to recombine at the second solvation shell even though equilibrium snapshots show the hole-carrying orbital localized on one molecule.","fun_headline_variants_meta":{"raw":{"variants":["Liquid harmonics reveal hidden second plateau from faraway recombinations","Off-site recombination explains extra plateau in liquid high harmonics","Second plateau in liquid HHG traced to second solvation shell electrons","Distant electron reunions spark higher-order plateaus in liquid harmonics","Liquid HHG's second plateau: electrons return to neighbors, not origin"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000277,"raw_usage":{"total_tokens":1740,"prompt_tokens":1124,"completion_tokens":616,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":740,"completion_tokens_details":{"reasoning_tokens":526}},"tokens_in":740,"tokens_out":616,"duration_ms":6737,"temperature":1.0,"reasoning_tokens":526,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:27:16.039396+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the second-plateau cutoff while changing the liquid's density, temperature, or solvent so that the second-solvation-shell distance moves: the analytical law predicts $E_c$ shifts linearly with $d_{\\mathrm{NNN}}$ (about $2 q E d_{\\mathrm{NNN}}$ to first order in the field) and stays essentially independent of wavelength, so a quadratic wavelength dependence or a cutoff that fails to track the measured radial distribution function would rule out next-nearest-neighbor recombination. A complementary check is the ellipticity profile: if second-plateau harmonics ever show a single Gaussian with no side peaks, the off-site channel is not the mechanism.","supporting_citations":[{"cited_title":"Neufeld, Z","cited_arxiv_id":null,"evidence_quote":"Supplies the ab initio cluster TDDFT approach (finite clusters, frozen nuclei) that reproduces the second plateau microscopically and rules out macroscopic origins."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the thin flat-liquid-jet technique whose signal-to-noise ratio makes the roughly 500-times-weaker second plateau observable."},{"cited_title":"Tancogne-Dejean, M","cited_arxiv_id":null,"evidence_quote":"Real-space-grid TDDFT code with which all liquid simulations were run."},{"cited_title":"Li, J.-X","cited_arxiv_id":null,"evidence_quote":"Links liquid HHG to the radial distribution function, supporting the use of solvation-shell distances as trajectory endpoints."},{"cited_title":"M¨ oller, Y","cited_arxiv_id":null,"evidence_quote":"Gas-phase ellipticity baseline (single Gaussian) against which the multi-Gaussian liquid response is compared."}],"review_version":1}