{"id":"ab71cf33-c75b-4eba-8fb7-c8ee2cb3a2d1","arxiv_id":"1908.10619","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Exposure of Fe3O4(001) to liquid or ambient-pressure water converts the reconstructed surface into a partially hydroxylated (1x1) surface with a self-limited, chain-like iron oxyhydroxide phase that saturates at roughly 40% coverage.","lead":"This paper shows that exposing the (001) surface of magnetite, an iron oxide mineral, to liquid or high-pressure water grows a thin, ordered oxyhydroxide layer that stops growing once it covers about 40% of the surface. The result matters because water-driven reactions on iron oxide surfaces are central to corrosion, mineral weathering, and electrocatalysis.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The chain features are assigned to an Fe-extracting oxyhydroxide phase, but no atomic model is given and the spatially averaged XPS cannot distinguish the chains from the surrounding hydroxylated surface, leaving the central phase claim underdetermined.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the structural assignment of the STM chain features as an Fe-extracting oxyhydroxide phase rests on indirect evidence, and the paper itself provides no concrete atomic model. My review of the full text confirms this is the most critical unresolved point. The experimental observations of reconstruction lifting, hydroxylation, and passivation are well documented and mutually consistent, but the specific claim that OH groups coordinate Fe cations extracted from subsurface layers is not uniquely determined by the spatially averaged XPS and STM topography. The atomic-H control is a valuable test of passivation, but it does not isolate the chain composition. Given that the paper is otherwise careful and the data are reproducible in two different UHV systems, the appropriate verdict remains CONDITIONAL: the structural model should be either supported by an independent probe (computational STM simulation, surface X-ray diffraction, or chemical fingerprinting) or explicitly presented as a hypothesis rather than a conclusion. My recommendation is UNCHANGED because the concern does not overturn the verdict but reinforces it.","tokens_in":12569,"tokens_out":3945,"duration_ms":46383,"concrete_test":"Perform DFT-based STM image simulations for candidate chain models, including (i) the proposed Fe-extracted oxyhydroxide motif and (ii) simple hydroxyl chains on the unreconstructed (1x1) surface, and compare quantitatively (apparent height, corrugation, chain-end features, coverage-dependent contrast) with the experimental STM images in Figs. 2c and 3a. If a hydroxyl-only chain model reproduces the observed contrast and 40% saturation behavior equally well, the central oxyhydroxide claim would be weakened; if only the Fe-extraction model matches, it would be supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the bright chain features observed in STM are OwaterH- groups coordinated to Fe cations extracted from subsurface layers. The supporting evidence is strong for hydroxylation and reconstruction lifting, but weak for the specific Fe-extraction mechanism. The 531.6 eV O 1s component (21% of total area) is spatially averaged: the surrounding (1x1) hydroxylated surface covers roughly 60% of the area and can account for most or all of this component. Similarly, the Fe2+ satellite increase is expected from reconstruction lifting and H adsorption alone, independent of chain formation. No spectroscopic or microscopic fingerprint isolates the chains' chemical composition. The authors explicitly state in the Discussion that they 'refrain from putting forward any concrete atomic model,' underscoring that the Fe-extraction picture is an interpretation of STM contrast, not a determined structure. The atomic-H control shows that pre-saturating the surface with H prevents chain formation, but this is also consistent with an alternative in which the lifted reconstruction no longer provides the subsurface Fe needed for extraction, rather than H blocking dissociation per se. Thus the load-bearing assumption is the identification of the chain phase as an oxyhydroxide with subsurface Fe extraction, and this remains unverified by independent structural or chemical probes.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a combined LEED, XPS, and STM study of the Fe3O4(001) surface after exposure to liquid water and near-ambient-pressure water vapor. The authors observe that water exposure lifts the subsurface cation vacancy reconstruction, producing a (1x1) terminated surface, and that bright chain-like features grow on the surface and saturate at approximately 40% coverage. They propose that dissociative water adsorption produces protons that hydroxylate the surface oxygen lattice and abstract subsurface Fe cations that, together with OH groups, form a self-limited iron oxyhydroxide phase. The paper also shows that pre-saturating the surface with atomic hydrogen prevents chain formation, and that similar chain features appear in previous EC-STM images at pH 7, suggesting relevance for electrochemical interfaces.","tokens_in":12788,"tokens_out":2539,"duration_ms":29063,"significance":"If the interpretation is correct, the paper provides the first atomic-scale images of a self-limited oxyhydroxide phase forming on a bulk oxide under water exposure, with implications for corrosion, dissolution, and electrocatalysis. The experimental approach—UHV-compatible liquid dosing combined with STM, LEED, and XPS—is novel and yields several robust observations, including the independence of the saturation coverage from exposure time and temperature, the atomic-hydrogen control experiment, and the direct comparison with EC-STM. These strengths make the paper valuable even if the detailed structural model is not fully proven. However, the central claim that the chains are an Fe-extracting oxyhydroxide phase rests on indirect evidence; the authors themselves refrain from proposing a concrete atomic model, and the spatially averaged XPS cannot unambiguously isolate the chemical state of the chains.","major_comments":[{"comment":"The assignment of the bright chain features to a surface iron oxyhydroxide phase with Fe extracted from the subsurface is not directly supported by the data. The XPS O 1s component at 531.6 eV (21% of total area) is spatially averaged over the entire surface, which includes the surrounding (1x1) hydroxylated regions estimated to cover ~60% of the area. The observed OH component could therefore be accounted for by hydroxylation of the surrounding surface alone, and no spectroscopic or microscopic fingerprint isolates the chains' chemical composition. The authors state in the Discussion that they 'refrain from putting forward any concrete atomic model,' which underscores that the Fe-extraction picture is an interpretation of STM contrast rather than a determined structure. To make the central claim load-bearing, the manuscript should provide additional evidence that the chains contain Fe cations and OH groups, for example through chain-selective spectroscopy, DFT-based STM image simulation, or a quantitative comparison of the XPS OH intensity with the chain coverage.","section":"Discussion, Fig. 2e"},{"comment":"The increase in the Fe2+ satellite and the shoulder at ~708 eV after water exposure is attributed to the formation of the oxyhydroxide phase, but this spectral change is also expected from the lifting of the SCV reconstruction and H adsorption alone, independent of chain formation. The manuscript does not quantify the expected Fe2+ increase from reconstruction lifting, so the Fe 2p data cannot discriminate between the proposed Fe extraction and a simple hydroxylation/reconstruction-lifting scenario. The authors should either show that the Fe2+ increase is larger than what can be explained by H adsorption alone, or temper the claim that the Fe 2p data support the oxyhydroxide phase.","section":"Results, Fig. 2d"},{"comment":"The atomic-hydrogen pre-saturation experiment is presented as evidence that surface passivation occurs because H saturation prevents further water dissociation. However, an equally consistent interpretation is that lifting the SCV reconstruction removes the subsurface Fe supply needed for Fe extraction, thereby preventing chain growth regardless of water dissociation kinetics. The manuscript should explicitly discuss this alternative and, if possible, provide an experimental or computational test that separates H-blocking from Fe-unavailability, for example by pre-hydroxylating the surface without lifting the reconstruction or by performing the experiment on a surface where the reconstruction has been lifted by another means.","section":"Results, Fig. 3b-c"}],"minor_comments":[{"comment":"The caption contains a typographical error: 'in-situ EC-STM image of of Fe3O4(001)' should read 'of Fe3O4(001)'. The rest of the caption is clear.","section":"Fig. 5 caption"},{"comment":"The coverage estimate of '≈40%' depends on the assumption that a fully packed (1x2) chain phase corresponds to 100% coverage. This normalization should be stated more explicitly in the main text rather than only in the figure caption, and the authors could provide the counting statistics or error bars for the coverage values.","section":"Results, Fig. 2c and Discussion"},{"comment":"The notation 'Fetet' (tetrahedral Fe in the subsurface) is used throughout but is not defined in a single place. A brief definition, perhaps in Figure 1 caption, would improve readability for non-specialists.","section":"Discussion"},{"comment":"The observation of isolated OwaterH- species and groups of four protrusions is interesting, but the text does not provide a quantitative measure of their density or a clear statement of how they correlate with chain length or coverage. A short statistical analysis would strengthen the precursor-state interpretation.","section":"Results, Fig. 3a"}],"recommendation":"major_revision","confidential_remarks":"The manuscript reports a technically impressive set of experiments and the observed phenomenology—reconstruction lifting, chain growth, and 40% saturation that is independent of time and temperature—is likely to be of broad interest. The main concern is that the central chemical conclusion (formation of an Fe-extracting oxyhydroxide phase) is underdetermined by the available evidence, and the authors themselves concede the absence of an atomic model. The paper would be publishable after the authors either provide additional evidence for the chain composition or explicitly reframe the conclusions as a hypothesis that is supported but not proven. I see no concern about novelty or scope for a chemistry/physics interface journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe experimental achievement alone makes this paper worth reading. The TU Wien group’s UHV-compatible liquid dosing setup lets them put a droplet of ultra-pure water on a clean Fe3O4(001) surface and then get back to STM, XPS, and LEED without ever seeing air. What they find is genuinely new: both liquid water and 20 mbar water lift the subsurface cation vacancy reconstruction and, over roughly 10–20 minutes, grow an ordered chain-like phase that saturates near 40% coverage, independent of dose time and temperature. The atomic-H control is a good experiment—pre-saturating the surface with hydrogen prevents chain formation, supporting a self-limiting mechanism tied to saturation of the surface oxygen lattice.\n\nThe soft spot is the structural assignment. The chains are presumed to be an oxyhydroxide phase built from OH groups of dissociated water coordinating to Fe cations extracted from subsurface layers. But the authors explicitly refrain from giving a concrete atomic model, and the spatially averaged XPS cannot separate the chains from the surrounding hydroxylated (1×1) surface. The 531.6 eV O 1s component and the increased Fe2+ satellite are consistent with hydroxylation and reconstruction lifting alone. So the “oxyhydroxide” identification is an interpretation of STM contrast, not a demonstrated structure. The stress-test worry that the atomic-H experiment may mean reconstruction lifting has consumed the subsurface Fe reservoir, rather than H blocking dissociation per se, is not fully resolved. Indeed, the paper shows water still dissociates on the H-saturated surface (the OH signal increases to 29%), so the passivation is more precisely of chain growth, not of all dissociation. The authors could tighten the wording here.\n\nNone of this undermines the core experimental facts. The saturation coverage is robust, the controls are thoughtful, and the SI provides supporting data. The uncertainty is the chemistry of the chains—a real uncertainty, but one the authors flag clearly. A serious referee can ask pointed questions, and the authors have already shown they can design mechanistic experiments. This deserves peer review, not desk rejection. It would fit well in a surface-science or physical-chemistry venue, and I would probably cite it for the experimental setup and the self-limited 40% phase even if the fine structure remains unresolved.","headline":"Liquid and high-pressure water exposure of Fe3O4(001) gives a self-limiting ordered chain phase at ~40% coverage—the experimental work is convincing, but the oxyhydroxide assignment with subsurface Fe extraction is an interpretation awaiting a structural probe or a concrete atomic model.","tokens_in":13422,"tokens_out":4673,"would_cite":true,"duration_ms":51612,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["68.47.Gh","68.37.Ef"],"model":"deepseek-v4-flash","headline":"Liquid water makes magnetite's (001) surface grow an ordered, self-limited iron oxyhydroxide phase.","keywords":["Fe3O4(001)","magnetite","water dissociation","iron oxyhydroxide","subsurface cation vacancy reconstruction","scanning tunnelling microscopy","X-ray photoelectron spectroscopy","self-limited growth"],"falsifier":"Expose the saturated surface to isotopically labelled water (D2O or $H2^{18}$O) and track the 531.6 eV O 1s component: if that component does not follow the label, or if high-resolution STM resolves the chains as single protrusions not sitting between the iron rows, the two-OH-per-unit-cell model and the self-limited oxyhydroxide mechanism would be contradicted.","tokens_in":12346,"feed_emoji":"💧","tokens_out":8211,"duration_ms":79001,"temperature":0.7,"pith_summary":"The paper tries to establish what happens to the clean (001) surface of magnetite when it is exposed to liquid or near-ambient-pressure water: the surface, which in vacuum is stabilized by a subsurface cation vacancy reconstruction, becomes hydroxylated, loses that reconstruction, and slowly grows an ordered iron oxyhydroxide phase. The central discovery is that this growth is self-limited, saturating at roughly 40% coverage, because the water-splitting reaction stops once the surface oxygen lattice is saturated with hydrogen. According to the proposed mechanism, protons from dissociated water bind to surface oxygen, lift the reconstruction, and reduce surface Fe3+ to Fe2+, while hydroxide groups pair up and pull iron cations out of subsurface layers to form chain-like oxyhydroxide structures. This matters because such water-induced phases are the first atomic-scale steps of passivation and dissolution of an abundant environmental mineral, and because the same chain features appear in electrochemical scanning tunnelling microscopy at neutral pH, suggesting the phase also forms at a solid–electrolyte interface.","feed_headline":"Water grows a self-limiting oxyhydroxide on magnetite","feed_subtitle":"Water lifts magnetite's surface reconstruction and forms an ordered oxyhydroxide that stops at about 40% coverage.","key_machinery":"The central object is the (√2 × √2)R45° subsurface cation vacancy (SCV) reconstruction of Fe3O4(001), in which an interstitial tetrahedral iron in the second layer replaces two octahedral iron atoms in the third layer; the paper's story is about how water removes this reconstruction and what replaces it. The load-bearing experimental machinery is the STM contrast rule, developed on the hydrogen-covered surface, that hydroxyl groups on the surface make nearby Fe cations appear bright in empty-state images, together with LEED's observation that the (√2 × √2) spots vanish after saturation and XPS's 531.6 eV OH component. The mechanistic machinery is dissociative water adsorption at high water chemical potential, which the paper links to the ~$10^{-5}$ mbar pressure threshold seen in earlier near-ambient-pressure XPS: above this threshold, water forms partially dissociated agglomerates on the surface at room temperature, and when the sample is re-evacuated the molecular water desorbs but the OH groups remain. Those OH groups then extract subsurface iron cations to form the oxyhydroxide chains, and the chain morphology is interpreted as autocatalytic growth with a rate-limiting step of moving subsurface iron into the chains.","core_discovery":"Exposure of the UHV-prepared Fe3O4(001)-(√2 × √2)R45° surface to ultrapure liquid water or 20 mbar water vapour makes the subsurface cation vacancy reconstruction disappear, leaving a (1 × 1) surface in LEED, and produces a roughly 40% coverage of bright chain-like protrusions in STM that are aligned along the [110] iron rows and show (1 × 2) spacing in dense regions. The paper assigns these protrusions to hydroxyl groups (OwaterH-) from dissociated water that are bound to iron cations extracted from subsurface layers, forming a surface iron oxyhydroxide phase; XPS shows an OH component at 531.6 eV and an increase in Fe2+ character, while the C 1s signal rules out contamination as the source of the chains. The mechanism proposed is that the protons adsorb on surface oxygen atoms, lift the reconstruction and reduce surface Fe3+, while pairs of OH- species coordinate subsurface tetrahedral iron and flip it up to the surface, nucleating autocatalytic chains. Growth terminates because once the surface oxygen lattice is saturated with hydrogen, no further water can dissociate; the saturation coverage comes out near the 50% expected from two OH groups per unit cell, with the observed 40% attributed to pre-existing hydroxylation or an additional minority OH site. The same chain structures are identified retrospectively in earlier EC-STM images at pH 7, which the paper takes as evidence that the oxyhydroxide phase exists at the solid–liquid interface under neutral conditions.","pith_inferences":["If hydrogen saturation of the surface oxygen lattice is what stops growth, then dosing with D2O or H2^18O should move only the OH-related XPS component and the vibrational signature of the chains, a direct isotopic test of the assignment that the paper does not carry out.","The same mechanism predicts that other spinel (001) surfaces carrying a subsurface cation vacancy reconstruction would show similar self-limited chain growth in liquid water; testing Fe3O4(111), which exposes undercoordinated iron, would specifically probe the proposed extraction of subsurface tetrahedral iron.","Because the paper explicitly refrains from proposing a concrete atomic model, a computational search constrained by the observed (1 × 2) chain spacing, the 40% saturation coverage and the two-OH-per-unit-cell estimate could turn the proposed oxyhydroxide into a falsifiable structural prediction.","If the oxyhydroxide is indeed the stable interfacial phase at neutral pH, then the common assumption that electrochemical magnetite electrodes present the bulk-truncated oxide to solution would need revision, which could affect kinetic models of corrosion and of iron-based water-splitting catalysts."],"forward_implications":["The Fe3O4(001) surface is not inert under humid conditions: in liquid water or at 20 mbar, the well-defined UHV surface transforms to a hydroxylated, oxyhydroxide-covered surface at room temperature.","The ~40% saturated oxyhydroxide provides a reproducible model system for the first stages of magnetite passivation and dissolution; the spatial separation of protonated areas and OH-covered chains mirrors what the paper predicts for alkaline and acidic solution, respectively.","Because the same chain features are seen by EC-STM at pH 7, electrochemical studies of Fe3O4(001) in neutral electrolyte should be interpreted with this oxyhydroxide phase in mind, not as a pristine oxide surface.","In alkaline conditions, where OH- can adsorb directly from solution, the paper predicts that iron extraction would not need water dissociation and could form a complete monolayer of oxyhydroxide.","This is offered as the first atomic-scale imaging of an ambient-pressure-water-driven conversion to an (oxy)hydroxide on a bulk oxide, making the surface a candidate model for the active phases invoked in iron-containing oxygen-evolution electrocatalysts."],"supporting_citations":[{"why":"Defines the subsurface cation vacancy reconstruction that is the starting surface and the reference for the (1 × 1) transition.","marker":"[24]"},{"why":"Shows that in UHV water desorbs below 250 K with no surface change, establishing the contrast with liquid/ambient-pressure behaviour.","marker":"[25]"},{"why":"Reports the near-ambient-pressure XPS threshold and mixed H2O/OH adsorption that the liquid-exposure results are connected to.","marker":"[26]"},{"why":"Supplies the earlier EC-STM images whose chain features are now identified as the same oxyhydroxide phase.","marker":"[22]"},{"why":"Shows atomic H lifts the reconstruction, producing bright OH-related features and Fe2+ that anchor the STM/XPS interpretation.","marker":"[32]"},{"why":"Provides the formate saturation coverage used to calibrate the OH surface density from XPS.","marker":"[31]"},{"why":"Gives the simulation-based structural motif of Fe cations coordinated by water-derived OH that supports the oxyhydroxide assignment.","marker":"[40]"},{"why":"Documents the analogous ambient-pressure-water conversion of a cobalt oxide film to an oxyhydroxide, the comparison system for a bulk-oxide first.","marker":"[15]"}],"fun_headline_variants":["Magnetite's watery makeover: a self-limiting oxyhydroxide film","Self-limiting oxyhydroxide forms on magnetite in water","Water triggers a 40% oxyhydroxide layer on magnetite","Magnetite gets a passivating oxyhydroxide in water","Self-limited oxide growth: water on magnetite"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The structural conclusion rests on reading the bright chain-like protrusions in the microscope images as hydroxyl groups from dissociated water bonded to iron atoms pulled up from just beneath the surface; if those protrusions are instead adsorbate clusters, substrate defects, or contaminant-derived species, the proposed oxyhydroxide phase would not be established.","fun_headline_variants_meta":{"raw":{"variants":["Magnetite's watery makeover: a self-limiting oxyhydroxide film","Self-limiting oxyhydroxide forms on magnetite in water","Water triggers a 40% oxyhydroxide layer on magnetite","Magnetite gets a passivating oxyhydroxide in water","Self-limited oxide growth: water on magnetite"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000398,"raw_usage":{"total_tokens":2161,"prompt_tokens":1106,"completion_tokens":1055,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":722,"completion_tokens_details":{"reasoning_tokens":961}},"tokens_in":722,"tokens_out":1055,"duration_ms":8571,"temperature":1.0,"reasoning_tokens":961,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:38:07.374622+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Expose the saturated surface to isotopically labelled water (D2O or $H2^{18}$O) and track the 531.6 eV O 1s component: if that component does not follow the label, or if high-resolution STM resolves the chains as single protrusions not sitting between the iron rows, the two-OH-per-unit-cell model and the self-limited oxyhydroxide mechanism would be contradicted.","supporting_citations":[],"review_version":1}