{"id":"0ec0d8aa-2198-4425-8aa7-a8206c582726","arxiv_id":"2608.05590","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Under intense X-rays, water-covered silica nanoparticles produce H3+ ions, and the self-induced surface electric field controls how much H+, H2+, and H3+ forms.","lead":"Scientists fired intense X-ray pulses at tiny water-coated glass beads and detected H3+ ions flying off them. The result suggests X-ray-irradiated dust in space may make this key interstellar molecule on its surface, not just in empty space.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The per-shot surface field is calibrated by electron counting only up to 2 V/nm on 500 nm particles, then extrapolated to 7 V/nm and to smaller sizes without an independent check; a size-dependent eTOF detection efficiency could make the field-axis collapse in Fig. 3(b) an artifact.","rationale":"The reader's weakest assumption is that the eTOF electron count is a constant-efficiency proxy for the net escaped charge, calibrated once against SIMION over 0–2 V/nm. My reading of SI A.8 confirms this is the pivotal assumption: the per-shot field axis, and hence the central collapse, is built on that calibration. The concern is load-bearing because the field is not merely a control parameter but is derived from the same measured signal that is binned; if the calibration constant is size-dependent, the comparison across particle sizes loses its meaning. I agree with the reader that this is the weakest link. The paper otherwise presents a strong single-particle multimodal dataset, with independent fluence monitors, a CNN-based morphology classifier, and simulation support, and the field-dependence within a single particle size is probably robust. The issue is specifically the extrapolation beyond the calibrated range. My concrete test would settle it directly: re-calibrating the smaller particles with the same independent SIMION route, or measuring the size-dependence of the electron capture fraction. If the calibration holds, the central claim survives; if not, the collapse and the dominance claim would need to be revisited. The reader's CONDITIONAL verdict remains appropriate; the concern reinforces the need for the calibration check rather than changing the verdict. The paper should not be rejected on this basis, but it should not be accepted until the field calibration across sizes is verified.","tokens_in":21831,"tokens_out":9238,"duration_ms":102940,"concrete_test":"Perform SIMION trajectory simulations calibrated to the Si+ kinetic-energy distributions for the 300 nm and 100 nm SiO2 and 150 nm Au@SiO2 particles over their full field range (up to ~7 V/nm), exactly as was done for the 500 nm particles in Fig. S8, and compare the resulting surface fields with those derived from electron counting via Eq. 2. If the two disagree by more than the stated 25% at matched fields, the common curves in Fig. 3(b) are not established. As a complementary check, compute the fraction of the electron TOF signal inside the 100–200 ns window per particle size and field bin; if the fraction at matched E differs significantly with R, the single calibration constant is invalid.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that surface electric field is the dominant parameter relies on assigning a per-shot field from the eTOF electron count via Eq. 2, E = Ne/(4πε0R^2), as described in §Data Analysis and SI A.8. The absolute scale of this assignment is fixed once by matching the electron-counting field to SIMION-based Si+ kinetic energies for 500 nm particles over 0–2 V/nm. All higher fields (up to ~7 V/nm for 100 nm particles) and all other particle sizes (300 nm, 150 nm Au@SiO2) are then assigned by electron counting alone, using the same calibration constant. The problem is that the eTOF integration window (100–200 ns) captures a fraction of the emitted electrons that depends on Coulomb retardation of the escaping electrons. Retardation is set by the surface potential V = E·R, not by E alone. At a matched nominal field of, say, 2 V/nm, a 500 nm particle (R ≈ 250 nm) has V ≈ 500 V, whereas a 100 nm particle (R ≈ 50 nm) has V ≈ 100 V. The electron arrival-time distribution, and hence the fraction inside the fixed window, can therefore differ between sizes at the same E. The SI acknowledges a monotonic reduction from ≈90% to ≈75% detection with intensity, but this is not shown to be independent of R. If the detection efficiency is not a common function of E across sizes, then fields computed from Eq. 2 with one global constant are misassigned for the smaller particles, and the collapse of H+, H2+, and H3+ yields onto common curves in Fig. 3(b) could be a consequence of the field miscalibration rather than of a universal dependence on E. This directly threatens the claim that field, not particle size, composition, or aggregation, governs the yields. The internal consistency checks (matching SIMION over 0–2 V/nm) do not validate the extrapolation, and the stated 25% systematic is small only against the two-order-of-magnitude field range, not against the size-dependent variation at matched field.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a single-particle X-ray free-electron laser experiment on hydrated silica nanoparticles (500, 300, and 100 nm SiO2 and 150 nm Au@SiO2) at 1.88 keV, combining ion velocity-map imaging, electron time-of-flight spectroscopy, and coherent diffractive imaging. The authors assign a per-shot surface electric field from the eTOF electron count via E = Ne/(4πε0R^2), bin the relative yields of H+, H2+, and H3+ by this field, and report that the yields collapse onto common curves across particle size, composition, and monomer/dimer morphology. Density functional theory and nonadiabatic quantum molecular dynamics simulations are used to argue that the field drives interfacial hole transfer and water fragmentation, supplying H2 and H2+ for the canonical reaction H2+ + H2 -> H3+ + H. The paper concludes that the self-induced surface field is the dominant parameter governing the hydrogen ion yields and that the canonical trihydrogen reaction proceeds on hydrated silica under radiation-driven ionization.","tokens_in":22199,"tokens_out":13442,"duration_ms":135364,"significance":"If the field-control and common-curve results hold, this is a significant advance: it connects radiation-driven charging of nanograins to field-driven interfacial chemistry, proposes a plausible new route to H3+ on interstellar dust, and extends the analogy to band bending in photoelectrocatalysis. The experimental analysis has notable strengths: shot-level bootstrap uncertainties with validation against permutation and jackknife, multiple intensity windows including strictly sub-transition selections, a field-insensitive Ar3+ reference, CNN-based monomer/dimer classification with F1 > 0.99, and open data and code availability. The main risk is the field calibration: the per-shot field axis is not independently verified for the smaller particles and higher fields, so the central cross-size collapse needs a size-resolved check before the dominance claim can be accepted.","major_comments":[{"comment":"The per-shot surface field is obtained from the eTOF electron count with a single global calibration constant that is established on 500 nm particles over 0–2 V/nm and then applied to all sizes and up to roughly 7 V/nm. The eTOF integration window (100–200 ns) captures a fraction of the emitted electrons that depends on the retarding surface potential V = E·R, which differs by a factor of five between 500 nm and 100 nm particles at the same nominal E. SI A.8 acknowledges that the captured fraction drops from about 90% to about 75% with intensity, but it does not demonstrate that this fraction is a common function of E independent of R. If the collection efficiency is not the same function of E for every particle size, the fields for the 100 nm and 300 nm particles are systematically misassigned, and the collapse in Fig. 3(b) could be an artifact of the field axis rather than evidence for field control. This is load-bearing for the cross-size claim, so I request a size-resolved cross-check (e.g., SIMION Si+ kinetic-energy fields for each diameter) or an explicit demonstration that the eTOF detection efficiency depends only on E and not on R.","section":"Data Analysis; Eq. (2); SI A.8"},{"comment":"The manuscript overstates the mechanistic conclusion. The data show field-dependent H3+ yields and the presence of H2+, but they do not directly establish that H3+ is formed by the canonical reaction H2+ + H2 -> H3+ + H. Neutral H2 is inferred from water-fragmentation chemistry rather than detected, and alternative pathways (for example H2+ + H2O, or H3+ from a transient H3O+ intermediate) are not excluded by the present measurements. The abstract and Outlook state that the canonical reaction proceeds on the surface; that claim should be softened to 'consistent with' or supported by additional evidence such as deuterated-water experiments, coincidence detection of H2, or kinetic modeling that rules out competing channels.","section":"Field-Dependent Trihydrogen Formation; Outlook"}],"minor_comments":[{"comment":"The dimer surface-field superposition should specify the exact evaluation point and formula (surface point, center, or interparticle midpoint). The matched-field monomer/dimer comparison is sensitive to this choice, especially for contact pairs.","section":"SI A.12"},{"comment":"The main analysis window of 500–2000 ion counts includes data above the WDM transition for the 100 nm particles, for which the transition is estimated at about 1500 counts. The strictly sub-transition windows in Fig. S10 are reassuring, but the main Fig. 3(b) should either use a window that is sub-transition for all particle types or state the overlap clearly in the caption.","section":"Methods; SI A.7"},{"comment":"The statement that 'the two independent measures agree over their common range' is only qualitative; please provide a quantitative comparison (for example, a residual plot or deviation statistics) for the electron-counting field versus the SIMION Si+ kinetic-energy field over 0–2 V/nm.","section":"SI A.8"},{"comment":"The absence of an H3O+ signal is explained by dissociative electron recombination, but this is an assumption rather than a directly measured channel. The text should label this as a hypothesis, not a demonstrated step.","section":"Unified Field-Driven Mechanism"},{"comment":"The CNN classifier is trained with synthetic dimer diffraction patterns generated by the coherent sum of two isolated spheres. The high validation F1-score is encouraging, but the authors should state that the classifier may not generalize to fused or necked aggregates, which would not present as two-sphere diffraction patterns.","section":"SI A.10"},{"comment":"The field range is described as 'roughly two orders of magnitude,' but the quoted fields (0.57–7.3 V/nm in SI A.8, with a 0–2 V/nm range for the 500 nm particles) span closer to one and a half orders of magnitude; please reconcile the quoted range or define it precisely.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the journal's scope and the data/code availability is a strength. The main technical risk is the unverified size independence of the eTOF-based field calibration; I would not accept the paper until that is addressed with a size-resolved cross-check or an explicit argument that the detection efficiency is a function of E alone. The mechanistic overclaim about the canonical reaction should also be tempered. These issues are fixable within the manuscript's scope, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nQuick take: this is a genuinely interesting experiment with a real new observation — H3+ from hydrated silica under 1.88 keV X-rays — but the paper's central claim that the self-induced surface field is the single controlling parameter rests on a field calibration that may not transfer across particle sizes. That is the thing I'd want checked before building on it.\n\nThe new bit is real: previous nanoparticle H3+ work used strong-field lasers; this is the first X-ray-driven version, and it is relevant to X-ray dominated regions. The multi-modal detection (VMI, eTOF, CDI on the same shot) is a legitimate technical step forward, and the data analysis is unusually careful for a first report: shot-level bootstrap errors, multiple intensity windows, the field-insensitive Ar3+ reference, the monomer/dimer comparison, and the carbon exclusion control are all good. They also provide data and code, which is more than most.\n\nThe soft spot is the field axis. They derive E from the eTOF electron count via a sphere formula, calibrate the absolute scale once against SIMION on 500 nm particles over 0–2 V/nm, and then apply that same constant to 100 nm and 150 nm core-shell particles at fields up to ~7 V/nm. The stress-test note points to a real problem: the eTOF integration window captures a fraction of the emitted electrons that depends on the Coulomb retarding potential, and that potential is E·R, not E alone. At the same nominal E, a 100 nm particle has one-fifth of the surface potential of a 500 nm particle, so the arrival-time distribution and the captured fraction can differ. The paper acknowledges the fraction drops from about 90% to 75% with intensity, but it does not show that this fraction is a common function of E across sizes. If it is not, the smaller particles' assigned fields are biased, and the collapse in Fig. 3(b) could be an artifact of the miscalibration rather than a universal dependence on E. That directly threatens the \"dominant parameter\" claim. The monomer/dimer match is at one size, so it does not help. I'd want a size-dependent calibration, e.g., SIMION-based field determination for each particle type, or a direct measurement of the electron detection efficiency as a function of E and R.\n\nTwo smaller issues: the canonical reaction H2+ + H2 → H3+ + H on the surface is inferred from the presence of both precursors and the barrier-free Langevin rate, not directly demonstrated; isotopic labeling with D2O would settle it. And the \"unifying mechanism\" language for astrochemistry and catalysis goes a bit beyond what a single-particle experiment can show.\n\nWho this is for: X-ray/matter people, astrochemists interested in H3+ formation routes, and anyone doing single-particle mass spec. It deserves a serious referee — the experiment is hard, the observation is new, and the field-calibration question is answerable. If the calibration survives scrutiny, the paper becomes much stronger. I'd cite it for the observation and the method, but I'd hold off on citing the field-dominance mechanism until the calibration issue is settled.\n\nMy recommendation: send it to review, and ask the authors to address the size-dependent field calibration head-on.","headline":"Real new observation — X-ray-driven H3+ on hydrated silica — but the field-dominance claim rests on a calibration that may not transfer across particle sizes.","tokens_in":22972,"tokens_out":5068,"would_cite":true,"duration_ms":49798,"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":"X-ray charging of hydrated silica nanoparticles creates V/nm surface fields that set the relative yields of H+, H2+, and H3+, enabling the canonical interstellar reaction H2+ + H2 → H3+ + H to proceed on the surface.","keywords":["trihydrogen cation","interstellar chemistry","surface electric field","silica nanoparticles","X-ray free-electron laser","water fragmentation","interfacial charge transfer","single-particle imaging"],"falsifier":"Measure the absolute charge of individual X-ray-irradiated nanoparticles independently—for example by charge detection mass spectrometry of the ejected ion plume or by laser-induced electron detachment from the charged particle—and compare the derived surface field to $E = Ne/(4\\pi\\epsilon_0 R^2)$; a discrepancy that grows systematically with field would show the collapse is an artifact of the electron-counting proxy.","tokens_in":98,"feed_emoji":"⚛️","tokens_out":15826,"duration_ms":201380,"temperature":0.7,"pith_summary":"The paper sets out to answer whether the canonical interstellar trihydrogen formation reaction, $\\mathrm{H_2^+} + \\mathrm{H_2} \\rightarrow \\mathrm{H_3^+} + \\mathrm{H}$, can proceed on an inorganic surface under radiation-driven ionization—the conditions that actually initiate this chemistry in X-ray-dominated regions of space. It shows that intense 1.88 keV X-ray pulses charge hydrated silica nanoparticles until the self-induced surface electric field reaches the V/nm scale, and that this field is the single dominant parameter setting the relative yields of $\\mathrm{H^+}$, $\\mathrm{H_2^+}$, and $\\mathrm{H_3^+}$ across particle size, composition, and aggregation. When single-shot data are binned by the field $E = Ne/(4\\pi\\epsilon_0 R^2)$ rather than by fluence, the ion yields collapse onto common curves: $\\mathrm{H^+}$ and $\\mathrm{H_3^+}$ rise with field while $\\mathrm{H_2^+}$ falls. If this is right, a reaction central to astrochemistry has a working surface route on interstellar dust grains, and the same field-driven charge-transfer mechanism connects radiation-dominated environments with electrode-free surface catalysis.","feed_headline":"X-ray fields, not size or aggregation, set H3+ yields on silica","feed_subtitle":"Measuring one number per particle—the surface field—predicts how much trihydrogen forms on silica dust.","key_machinery":"The load-bearing object is the per-shot surface electric field $E = Ne/(4\\pi\\epsilon_0 R^2)$, computed by counting the electrons $N$ ejected from a particle of measured radius $R$ and treating the particle as a uniformly charged sphere. It is the independent variable that collapses the data: binning the hydrogen ion yields by $E$ rather than by fluence, size, composition, or aggregation brings the relative yields of $\\mathrm{H^+}$, $\\mathrm{H_2^+}$, and $\\mathrm{H_3^+}$ onto common curves. The mechanism carrying the chemistry is field-driven interfacial charge transfer: at fields of a few V/nm, the water-derived O 2p states are tilted into the gap region of the silica surface, increasing the driving force and hybridization for hole localization in the water layer, so that holes transfer from silica into adsorbed water within about 60 fs, fragmenting water and generating the $\\mathrm{H_2}$ and $\\mathrm{H_2^+}$ precursors of the canonical reaction.","core_discovery":"The paper's central claim is that the self-induced surface electric field on X-ray-charged hydrated silica nanoparticles is the dominant parameter governing the relative yields of $\\mathrm{H^+}$, $\\mathrm{H_2^+}$, and $\\mathrm{H_3^+}$, independent of particle size (500, 300, and 100 nm $\\mathrm{SiO_2}$ and 150 nm Au@$\\mathrm{SiO_2}$ core-shell), composition, and aggregation. Using coincidence measurements of ion velocity-map imaging, electron time-of-flight, and coherent diffractive imaging on individual particles, the authors assign every shot a field $E = Ne/(4\\pi\\epsilon_0 R^2)$ and find that the yields of $\\mathrm{H^+}$ and $\\mathrm{H_3^+}$ rise while $\\mathrm{H_2^+}$ falls along a common curve. The field acts upstream of the barrier-free proton-hop reaction: it drives interfacial charge transfer from silica valence states into the adsorbed water layer, fragments water, and sets the balance of the $\\mathrm{H_2}$ and $\\mathrm{H_2^+}$ precursors. The canonical reaction then proceeds on the intact surface, and the same field dependence appears for the water-ion yield while the silica-framework ions show no monotonic field dependence and the gas-phase argon reference stays flat. The conclusion is that field-driven charge transfer at V/nm fields is a unifying mechanism between radiation-dominated astrophysical environments and field-driven surface catalysis.","pith_inferences":["Editorial inference: if the field is truly the controlling parameter, astrochemical models of $\\mathrm{H_3^+}$ production in X-ray-dominated regions should use the grain-surface field (which scales as $N/R^2$ per charging event) as the rate-determining variable, rather than fluence or absorbed dose alone.","Editorial inference: the common-curve collapse implies a testable universal scaling—any two grains with the same $E$ should give the same relative $\\mathrm{H^+}$/$\\mathrm{H_2^+}$/$\\mathrm{H_3^+}$ yields regardless of material or aggregation; extending the measurement to other ice mantles or oxide substrates would probe whether the universality holds beyond hydrated silica.","Editorial inference: by analogy with band bending at semiconductor photoelectrodes, the field-driven hole-transfer picture may be a general design principle for radiation-driven catalysis on wide-bandgap insulators, so other hydrated oxide surfaces under X-ray irradiation should show similar field-controlled ion yields."],"forward_implications":["Interstellar grains in X-ray-dominated regions gain a concrete route to $\\mathrm{H_3^+}$ through the canonical reaction on water-covered silicate surfaces, with the yield set by the local surface field rather than by grain size alone.","Radiation-driven surface chemistry on nanoparticles can be parameterized by a single intensive quantity—the self-induced surface field—so laboratory results on one particle size can be transferred to another at matched field.","The field ordering of the $\\mathrm{H^+}$, $\\mathrm{H_2^+}$, and $\\mathrm{H_3^+}$ yields persists above the warm-dense-matter transition, indicating that the field remains the controlling parameter even as the particle begins to enter a plasma-like regime.","The multimodal single-particle approach ties an individual particle's chemistry to its charge state and structure shot-by-shot, resolving heterogeneity that ensemble measurements would obscure.","The V/nm fields produced by X-ray charging offer an electrode-free driving voltage for surface catalysis, with the same mechanism potentially steering selectivity in reactions such as hydrogen evolution and $\\mathrm{CO_2}$ reduction."],"supporting_citations":[{"why":"Supplies the density-functional and nonadiabatic quantum molecular dynamics model of field-driven water fragmentation and silanol dissociation on strongly ionized silica surfaces.","marker":"[38]"},{"why":"Provides earlier single-particle tracking and simulations showing surface-charge-driven weakening of O–H bonds on silica nanoparticles, the direct predecessor of the field-driven picture.","marker":"[26]"},{"why":"Demonstrated anomalous H3+ formation from water on nanoparticles in strong laser fields, establishing water-covered nanoparticles as a source of H3+ that this work extends to X-ray ionization.","marker":"[25]"},{"why":"Ion-optics trajectory simulations used to convert measured Si+ kinetic energies into surface fields and to calibrate the electron-counting field.","marker":"[36]"},{"why":"The single-photon-counting pnCCD detector that provides the coherent diffractive imaging used to classify monomers and dimers shot-by-shot.","marker":"[35]"},{"why":"Provides the dissociative recombination channel H3O+ + e- → OH + H + H that converts confined H3O+ into H atoms, supplying the neutral H2 precursor.","marker":"[48]"},{"why":"Shows H2+ is a product of water-cluster ionization, supporting the direct observation of the H2+ precursor on the hydrated surface.","marker":"[49]"}],"fun_headline_variants":["Surface field, not particle size, dictates H3+ formation on dust","One field per grain predicts trihydrogen yield on silica","X-ray-charged silica's electric field rules H3+ production","Field on nanoparticles, not size, drives interstellar H3+ route","V/nm field on silica sets H3+ yields, not dust properties"],"cache_read_input_tokens":24704,"weakest_assumption_plain":"The central load-bearing premise is that the per-shot electron count is a reliable, order-preserving measure of the net charge escaping each particle, so that the surface field $E = Ne/(4\\pi\\epsilon_0 R^2)$ can be assigned to every shot and used to collapse the data; if the electron signal stops tracking the charge at high fields, the apparent common curves could be a binning artifact.","fun_headline_variants_meta":{"raw":{"variants":["Surface field, not particle size, dictates H3+ formation on dust","One field per grain predicts trihydrogen yield on silica","X-ray-charged silica's electric field rules H3+ production","Field on nanoparticles, not size, drives interstellar H3+ route","V/nm field on silica sets H3+ yields, not dust properties"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000523,"raw_usage":{"total_tokens":2583,"prompt_tokens":1053,"completion_tokens":1530,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":669,"completion_tokens_details":{"reasoning_tokens":1439}},"tokens_in":669,"tokens_out":1530,"duration_ms":11832,"temperature":1.0,"reasoning_tokens":1439,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T05:57:33.632801+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the absolute charge of individual X-ray-irradiated nanoparticles independently—for example by charge detection mass spectrometry of the ejected ion plume or by laser-induced electron detachment from the charged particle—and compare the derived surface field to $E = Ne/(4\\pi\\epsilon_0 R^2)$; a discrepancy that grows systematically with field would show the collapse is an artifact of the electron-counting proxy.","supporting_citations":[{"cited_title":"M.et al.Catalysis in Extreme Field Environments: A Case Study of Strongly Ionized SiO2 Nanoparticle Surfaces.Journal of the American Chemical Society146, 27563–27570 (2024)","cited_arxiv_id":null,"evidence_quote":"Supplies the density-functional and nonadiabatic quantum molecular dynamics model of field-driven water fragmentation and silanol dissociation on strongly ionized silica surfaces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides earlier single-particle tracking and simulations showing surface-charge-driven weakening of O–H bonds on silica nanoparticles, the direct predecessor of the field-driven picture."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrated anomalous H3+ formation from water on nanoparticles in strong laser fields, establishing water-covered nanoparticles as a source of H3+ that this work extends to X-ray ionization."},{"cited_title":"simionforthepersonalcomputerinreflection.InternationalJournalofMassSpectrometry 200, 3–25 (2000)","cited_arxiv_id":null,"evidence_quote":"Ion-optics trajectory simulations used to convert measured Si+ kinetic energies into surface fields and to calibrate the electron-counting field."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The single-photon-counting pnCCD detector that provides the coherent diffractive imaging used to classify monomers and dimers shot-by-shot."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the dissociative recombination channel H3O+ + e- → OH + H + H that converts confined H3O+ into H atoms, supplying the neutral H2 precursor."},{"cited_title":"& Hirao, K","cited_arxiv_id":null,"evidence_quote":"Shows H2+ is a product of water-cluster ionization, supporting the direct observation of the H2+ precursor on the hydrated surface."}],"review_version":1}