{"id":"d7d88b4a-c835-4c92-83d9-59e8a42dd941","arxiv_id":"2506.09908","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Graphite-water interfaces under potential can exist in three states: transient clean water, aged hydrocarbon-covered water, and a field-stabilized clean water state with diverse hydrogen-bond geometries.","lead":"This paper combines two angstrom-scale probes, 3D atomic force microscopy and shell-isolated Raman spectroscopy, to watch water at graphite surfaces in real time. It finds that the interface can switch among three states, including a hydrocarbon-dominated state that negative voltages can push back to clean water.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The d12-to-identity mapping (3 Å = water, 4–5 Å = hydrocarbons) is the load-bearing step; it correlates non-co-located AFM and SHINERS data, and the paper concedes the 4–5 Å species' exact nature is unknown.","rationale":"The reader's weakest_assumption identifies the d12-to-molecular-identity mapping as the key fragility; I agree with that identification and sharpen it: the non-co-located nature of the AFM and SHINERS measurements, combined with the paper's own admission that the 4–5 Å species' exact nature is unknown, makes this mapping genuinely load-bearing. The paper's evidence is otherwise strong: results come from two independent laboratories over four years, with 16 3D-AFM and 4 SHINERS data sets; the silica control addresses a major artifact channel; and the potential-driven transitions are reversible and consistent across methods. I do not see an internal inconsistency in the three-state model, and the OH-peak assignment ambiguity, while acknowledged, is not the most fragile step because the appearance/disappearance of OH-3 to OH-5 is a qualitative spectral change rather than a subtle fitting detail. The remaining concern is that the chemical identity of the layers observed by AFM is inferred from separate, non-co-located SHINERS experiments; the same-sample covary test described above would settle whether that inference is correct. If the test passes, the central claim is well supported; if it fails, the model's central identification would need revision. This is consistent with the reader's CONDITIONAL verdict, so no verdict change is needed.","tokens_in":17609,"tokens_out":5997,"duration_ms":83024,"concrete_test":"Perform 3D-AFM and SHINERS on the same HOPG electrode in the same electrochemical cell: deposit a sparse sub-monolayer of Au/SiO2 particles, record 3D-AFM force maps in particle-free regions immediately before and after each SHINERS acquisition at matched ageing times and potentials, and compare d12 with νCH/FP/νOH areas over the same time course. If d12 and C–H intensity do not covary, or if d12 differs between particle-free and particle-decorated regions, the d12-to-identity mapping fails. A complementary check is tip-enhanced Raman or AFM-IR on the same HOPG spot to directly identify C–H modes at the 4–5 Å layered regions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The three-state model requires d12 to be a reliable chemical descriptor. In the section 'Interfacial configurations at OCP', d12 ≈ 3 Å is assigned to pristine water layers and d12 ≈ 4–5 Å to adventitious hydrocarbon layers, relying on prior publications (refs 6, 25, 37), some from the same group. The paper itself states the exact nature of the 4–5 Å species 'remains unknown to date', and the fingerprint-mode assignments are described as 'still under debate' (ref 50). The AFM and SHINERS measurements are not simultaneous or co-located: AFM probes bare HOPG, whereas SHINERS probes Au/SiO2 nanoparticle–HOPG hotspots, a confined and geometrically different environment. The central claim therefore depends on the assumption that the time evolution of d12 in one set of experiments and of νCH/FP intensities in another set reflects the same molecular exchange. If 4–5 Å spacings can also arise from strongly depleted or tip-perturbed water layers, or if the nanoparticle hotspots systematically favor hydrocarbon accumulation relative to bare HOPG, then State 2 is not established as hydrocarbon-dominated, and the State 2 → State 3 transition is not established as water replacing hydrocarbons. This is a correctness risk in the argument's external validity, not merely a disagreement with consensus.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript combines three-dimensional atomic force microscopy (3D-AFM) and shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS) to study the HOPG/water and HOPG/aqueous-electrolyte interface in situ. The authors report a three-state model: State 1, a transient pristine water interface with strong hydrogen-bond breaking and ~3 Å layer spacing; State 2, a steady hydrocarbon-dominated interface with 4–5 Å layer spacing and weak hydrogen-bond breaking of the remaining water; and State 3, a stable pristine water interface at negative potentials with a broader distribution of hydrogen-bond configurations. The paper argues that prior contradictory microscopy and spectroscopy results on solid–water interfaces can be reconciled because different studies sampled different interfacial states. The work is based on 16 independent 3D-AFM data sets and 4 SHINERS data sets acquired in two laboratories over four years, with a graphite-free SiO2 control that rules out a direct silica contribution to the emergent OH peaks.","tokens_in":17900,"tokens_out":8426,"duration_ms":84805,"significance":"If the three-state model is correct, it provides a unified explanation for longstanding disagreements in the interfacial-water literature and highlights the importance of controlling sample age and electrochemical history. The multi-year, cross-laboratory dataset is a notable strength, as is the explicit silica control and the careful attention to initial sample state and potential protocol. The paper includes a large amount of reproducibility metadata: raw force maps and spectra are documented in the supplementary figures, and key time-evolution and potential-dependent trends are shown for multiple independent runs. However, the central chemical identification of the 4–5 Å AFM layer as hydrocarbon rests on non-co-located AFM and SHINERS measurements and on prior same-group assignments, which limits the external validity of the current evidence for State 2 and for the State 2→3 transition.","major_comments":[{"comment":"The assignment of d12 ≈ 3 Å to pristine water layers and d12 ≈ 4–5 Å to adventitious hydrocarbon layers is load-bearing for State 2 and for the State 2→State 3 transition, yet the paper itself states that the exact nature of the 4–5 Å species 'remains unknown to date' and that the fingerprint-mode assignments are 'still under debate' (ref 50). Moreover, the AFM and SHINERS measurements are not co-located: AFM probes bare HOPG, whereas SHINERS probes the Au/SiO2 nanoparticle–HOPG gap, a confined and geometrically different environment with potentially different hydrocarbon accumulation kinetics. The claim that the time evolution of d12 and of the νCH/FP intensities reflects the same molecular exchange is thus an assumption that should be explicitly defended—for example, by co-located chemical identification on bare HOPG, by a control showing that the nanoparticle hotspots do not alter accumulation kinetics, or by an explicit estimate of how the geometric difference affects the inferred identity. Please also reconcile the SHINERS State 2→3 transition potential (around −2 V in Fig. 4c) with the AFM d12 transition (mostly −1 to −1.5 V in Fig. 4a,b) before using the two observables as a single state descriptor.","section":"Interfacial configurations at OCP"},{"comment":"The peak positions, FWHMs, and area fractions of OH-1 to OH-5 reported in Figs. 3e–h and 4e–g are presented without error bars or uncertainty estimates, even though they are derived from a five-Voigt decomposition with free positions, widths, and areas. The quantitative claims that support the 'broader distribution of HB configurations' in State 3 (increased OH-1, OH-4, OH-5; decreased OH-2) and the emergence of OH-3 to OH-5 at −2 V in the non-pristine case should be accompanied by the variance across independent scans (the paper mentions seven scans for the OH-5 Stark shift, but does not provide statistics for the other metrics) and by a sensitivity analysis to the number of Voigt peaks, baseline treatment, and initial-guess dependence. Without this, the reader cannot assess whether the observed trends are robust or artifacts of the fitting procedure.","section":"Electrified graphite–water interfaces: pristine response; Electrified graphite–water interfaces: non-pristine response"},{"comment":"The identification of OH-5 as the antisymmetric OH stretch of non-donor water is based solely on the absence of a Stark shift. While the paper states this was consistently observed in seven potential scans over three independent sets, the argument relies on the assumption that the only configuration with zero Stark shift is the proposed ND orientation. The authors do not provide an independent test (e.g., isotopic dilution or polarization-resolved Raman) to discriminate among alternative configurations with a dipole difference perpendicular to the field. Because the presence of 'monomer-like' ND water in State 3 is highlighted in the conclusion, this assignment should be labeled as a tentative hypothesis or supported by additional evidence.","section":"Electrified graphite–water interfaces: pristine response"}],"minor_comments":[{"comment":"Please qualify the word 'correlative' in the abstract: the 3D-AFM and SHINERS measurements are not performed on the same physical location or simultaneously, so 'correlative' should be phrased to indicate that the two techniques were applied to the same sample type and protocol rather than co-located.","section":"Abstract"},{"comment":"In the sentence defining d12, please specify exactly how the first and second layer positions are measured in the force–distance curves (e.g., peak maxima) and how many independent force curves contribute to each reported d12 value; this information is currently only in the Methods section and would improve the transparency of the main-text descriptor.","section":"Interfacial configurations at OCP"},{"comment":"The potential-dependence plots of d12 would be clearer if the direction of the potential scan (forward vs. backward) was indicated for each symbol, particularly for Fig. 4b where the return-to-OCP point is included.","section":"Fig. 3b and Fig. 4b"},{"comment":"The first use of the state labels State 1, State 2, and State 3 occurs in the abstract; please define these terms explicitly at their first mention in the main text (or state where they are defined) and add a sentence in the Fig. 2 caption summarizing the three-state scheme so that readers can track the notation.","section":"Introduction / Fig. 2"},{"comment":"Reference 50 (Larkin, chapter 7) is a general spectral interpretation guide; for the statement that precise assignments of the hydrocarbon fingerprint modes are 'still under debate', please cite a more specific source or sources that discuss the ambiguity of C–O/C–F/aromatic ring assignments in surface-enhanced Raman spectra.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's central correlative claim would be materially strengthened by co-located AFM and SHINERS, but the authors may reasonably argue this is not feasible with current techniques. The editor may wish to consider whether the reliance on refs 6, 25, and 37 for the d12-to-identity mapping—all from the same groups—should be disclosed explicitly in the text as a limitation. The paper fits the scope of the journal and the data are extensive; the main question is whether the interpretive leap about hydrocarbon identity is presented with sufficient caution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is the first paper I've seen that gives a credible, unified explanation for why the graphite–water interface has looked so different across labs. The three-state model — transient pristine water, hydrocarbon-dominated aged interface, and potential-stabilized pristine water — is presented with enough independent evidence that I think the qualitative picture is right.\n\nWhat's actually new: the correlative combination of 3D-AFM and SHINERS with deliberate control of the initial state. Sixteen 3D-AFM sets and four SHINERS sets, run over four years in two labs, converge on the same sequence. The silica control effectively rules out the obvious artifact that the SHINERS hotspot itself produces the high-frequency OH peaks. And the key observation — that an aged, hydrocarbon-covered interface and a fresh, pristine interface both land in the same State 3 at negative potentials — is a strong piece of evidence in favor of the model.\n\nSoft spots: the paper reports no error bars or fit-uncertainty analysis for the five-Voigt decomposition, and the OH assignments to specific hydrogen-bond numbers are admitted to be contested. That's a moderate issue; the qualitative trends don't hinge on exact assignments, but the quantitative claims (e.g., area fractions changing with potential) should have been accompanied by uncertainty estimates. Minor.\n\nThe more substantive soft spot is the d12-to-identity mapping. The paper relies on prior same-group work for the 3 Å = water, 4–5 Å = hydrocarbon assignment, and concedes the exact nature of the 4–5 Å species is unknown. The SHINERS data is the independent check, and the time-correlation between νCH/FP growth and d12 increase is supportive. But the two measurements are not co-located: AFM probes bare HOPG, whereas SHINERS probes Au/SiO2–HOPG gaps. Nanoconfinement in those hotspots could plausibly accelerate hydrocarbon accumulation compared to a bare surface. That means 'State 2 is hydrocarbon-dominated' is well supported for the gap environment, but the extrapolation to the bare surface is partly inferred.\n\nIs that fatal? I don't think so. The stress-test worry that the 4–5 Å spacing could be tip-perturbed water collapses when you see the SHINERS spectra evolve in the same direction over the same timescale, and the potential-driven transition is observed in both techniques. The non-co-location issue is a real caveat, but not a refutation.\n\nVerdict: this paper deserves a serious referee. It's a substantial dataset, the method pairing is novel, and the central claim is falsifiable. I'd send it out. I'd bring it to a reading group, and if I worked on interfacial electrochemistry, I'd cite it. The authors should be pushed to add error bars and more details on the peak fitting stability, but the core is solid.","headline":"Three-state graphite–water model is credible and well supported, but the load-bearing d12-to-hydrocarbon mapping relies on non-co-located measurements and an unidentified species.","tokens_in":18480,"tokens_out":4213,"would_cite":true,"duration_ms":44420,"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":"Three graphite-water interface states explain a decade of conflicting results","keywords":["interfacial water","three-dimensional atomic force microscopy","SHINERS","graphite-water interface","hydrogen bond network","adventitious hydrocarbons","electrified interfaces","Raman spectroscopy"],"falsifier":"A co-located measurement would settle it: if simultaneous 3D-AFM and interface-enhanced Raman on one spot ever showed strong C–H bands while $d_{12}$ stayed at 3 Å, or showed 4–5 Å spacing with no hydrocarbon signal, the $d_{12}$-to-identity mapping would fail.","tokens_in":17409,"feed_emoji":"💧","tokens_out":3695,"duration_ms":41106,"temperature":0.7,"pith_summary":"The paper argues that contradictory reports on water structure at graphite and other solid surfaces arise because the interface is not a single structure but switches among three states. At open circuit, a freshly cleaned interface is transient pristine water with strongly broken hydrogen bonds (State 1); over about an hour it is replaced by adventitious hydrocarbons with 4–5 Å spacing (State 2). Applying sufficiently negative potentials converts either state into a stable pristine-water state (State 3) whose water has a broader spread of hydrogen-bond configurations. If this picture is right, it reconciles microscopy and spectroscopy datasets that previously seemed incompatible.","feed_headline":"Three interface states reconcile graphite-water disputes","feed_subtitle":"Correlative AFM and Raman show water gives way to hydrocarbons, then returns pristine under negative bias.","key_machinery":"The central descriptor is $d_{12}$, the first-to-second-layer spacing in 3D-AFM force–distance curves, which acts as a molecular-identity proxy: ~3 Å indicates pristine hydration layers and 4–5 Å indicates adventitious hydrocarbon layers. The complementary machinery is SHINERS, whose O–H stretching band is deconvoluted into five Voigt peaks (OH-1 through OH-5) assigned to decreasing hydrogen-bond coordination numbers; the peak positions, widths, and area fractions report hydrogen-bond breaking and diversification, while C–H and fingerprint peaks report hydrocarbon presence. Correlating these two signals on the same interface within ~2 nm of the surface is what lets the paper distinguish the three states.","core_discovery":"The authors claim, based on correlative 3D atomic force microscopy and shell-isolated nanoparticle-enhanced Raman spectroscopy (SHINERS), that the graphite–water interface at open circuit can be in either a transient pristine-water state with $d_{12}\\sim3$ Å spacing and strong hydrogen-bond breaking (State 1), or a steady hydrocarbon-dominated state with $d_{12}\\sim4$ to 5 Å and weak hydrogen-bond breaking of the surrounding water (State 2). At sufficiently negative potentials, both states converge to a stable pristine-water state (State 3) with unchanged $d_{12}\\sim3$ Å, constant interfacial water amount, and a broader distribution of hydrogen-bond configurations that includes a low-coordinated non-donor water species. The transition from State 2 to State 3 occurs mostly between −1 V and −1.5 V vs Ag/AgCl, with hydrocarbons displaced by water; returning to open circuit restores State 1, which then ages back to State 2. This three-state model is offered as the resolution of long-standing disputes about whether solid–water interfaces are pristine water layers or hydrocarbon-dominated regions.","pith_inferences":["A direct consequence the authors leave implicit: electrochemical studies of carbon and other electrodes should report the ageing state and history of the interface, because measurements taken at open circuit without this control can silently switch between State 1 and State 2 within an hour.","The hydrocarbon-to-water transition at negative potentials suggests a practical in-situ cleaning strategy for hydrophobic electrodes: a short negative polarization sweep could remove adventitious organic layers before sensitive interfacial measurements.","The zero Stark shift of OH-5 implies a nearly fixed molecular orientation with its transition dipole parallel to the surface; this is a testable prediction that polarization-resolved Raman or phase-sensitive sum-frequency spectroscopy could verify directly.","A natural extension would be to check whether the same three-state behavior appears on metal electrodes such as gold or platinum, where hydrocarbon contamination is also common but the surface chemistry is different; the authors only propose this, they do not demonstrate it."],"forward_implications":["If the three-state model is correct, prior studies that reported either pristine water layers or hydrocarbon layers at graphite–water and other solid–water interfaces were not necessarily wrong; they were observing different states of the same system.","The $d_{12}$ spacing is potential-independent for pristine interfaces, meaning that applied electric fields reorient interfacial water and change its hydrogen-bond configurations without changing the layer spacing or total water density.","Applying sufficiently negative potentials converts a hydrocarbon-contaminated interface into a clean pristine-water interface, and the effect is reversible upon returning to open circuit followed by hours of ageing.","The hydrogen-bond configuration at an interface is controlled by three separable factors: the solid surface, any intervening hydrocarbon layer, and the interfacial electric field, each of which can now be studied independently.","The same state diagram is expected to apply to other hydrophobic and air-exposed solid–water interfaces, not just graphite, which would make the result broadly relevant to electrochemistry, biosensing, and photocatalysis."],"supporting_citations":[{"why":"Establishes that 4–5 Å spacing in 3D-AFM corresponds to hydrophobic/adventitious layers on graphite-like surfaces in water.","marker":"[6]"},{"why":"Provides the ~3 Å hydration-layer spacing in electrolyte solutions used to interpret State 1.","marker":"[9]"},{"why":"Supplies the in-situ SHINERS approach to electrified interfacial water and the five-component νOH decomposition used here.","marker":"[19]"},{"why":"Introduces shell-isolated nanoparticle-enhanced Raman spectroscopy, the spectroscopic method on which the chemical assignment rests.","marker":"[29]"},{"why":"Prior X-ray absorption study of interfacial water on gold electrodes that provides the non-donor water assignment used for OH-5.","marker":"[18]"},{"why":"Reports interfacial layering of hydrocarbons on pristine graphite surfaces in water, supporting the State 2 interpretation of 4–5 Å spacing.","marker":"[37]"}],"fun_headline_variants":["Correlative AFM-Raman reveals three graphite-water states","Three states, not two: graphite-water interface decoded","Water or hydrocarbons? Bias decides graphite interface","Graphite-water: transient, dirty, then pristine under bias"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire state assignment rests on reading a 3 Å interlayer spacing as pristine water and a 4–5 Å spacing as adventitious hydrocarbons, but the AFM and Raman measurements are not made on the same spot at the same time, so this identity mapping is assumed rather than directly proven.","fun_headline_variants_meta":{"raw":{"variants":["Correlative AFM-Raman reveals three graphite-water states","Three states, not two: graphite-water interface decoded","Water or hydrocarbons? Bias decides graphite interface","Graphite-water: transient, dirty, then pristine under bias"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000778,"raw_usage":{"total_tokens":3424,"prompt_tokens":916,"completion_tokens":2508,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":532,"completion_tokens_details":{"reasoning_tokens":2444}},"tokens_in":532,"tokens_out":2508,"duration_ms":21288,"temperature":1.0,"reasoning_tokens":2444,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:37:43.986038+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A co-located measurement would settle it: if simultaneous 3D-AFM and interface-enhanced Raman on one spot ever showed strong C–H bands while $d_{12}$ stayed at 3 Å, or showed 4–5 Å spacing with no hydrocarbon signal, the $d_{12}$-to-identity mapping would fail.","supporting_citations":[{"cited_title":"R., Martin-Jimenez, D","cited_arxiv_id":null,"evidence_quote":"Establishes that 4–5 Å spacing in 3D-AFM corresponds to hydrophobic/adventitious layers on graphite-like surfaces in water."},{"cited_title":"& Garcia, R","cited_arxiv_id":null,"evidence_quote":"Provides the ~3 Å hydration-layer spacing in electrolyte solutions used to interpret State 1."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the in-situ SHINERS approach to electrified interfacial water and the five-component νOH decomposition used here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces shell-isolated nanoparticle-enhanced Raman spectroscopy, the spectroscopic method on which the chemical assignment rests."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior X-ray absorption study of interfacial water on gold electrodes that provides the non-donor water assignment used for OH-5."},{"cited_title":"Arvelo, D., R","cited_arxiv_id":null,"evidence_quote":"Reports interfacial layering of hydrocarbons on pristine graphite surfaces in water, supporting the State 2 interpretation of 4–5 Å spacing."}],"review_version":1}