{"id":"c76f84f8-be45-4688-b10b-bf272b03a140","arxiv_id":"1908.08008","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Surface polarons on rutile TiO2(110) suppress water dissociation at polaronic sites by about 240 meV, slightly enhance it at some non-polaronic sites, and strengthen hydrogen bonds between adsorbed water and hydroxide.","lead":"Using computer simulations, this paper maps how small polarons, localized electrons coupled to lattice distortions, on a titania surface change whether water molecules stick and split apart. It finds that polarons block water splitting on the sites where they sit, can help it on neighboring sites, and strengthen hydrogen bonds, offering a new way to explain long-running experimental disagreements.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Finite-size and polaron-pinning effects are untested, so the quantitative site-resolved energetics (especially the ~30 meV non-polaronic shifts) are not yet secure.","rationale":"I read the paper as making a qualitative claim that is well supported: a small polaronic Ti5c site strongly suppresses water dissociation at that site, with corroboration from PBE+U, HSE, and oxygen-vacancy-created polarons, and from a consistent NEB barrier. The reader's weakest assumption correctly identifies DFT accuracy and missing error estimates as the main risk. My stress-test focuses that concern more sharply: the site-resolved heat maps in a 2x4 cell with a single excess electron and a uniform background charge require that the polaron remain pinned to the same Ti site for every adsorption configuration, and that the cell be large enough to avoid image interactions. Neither condition is explicitly verified in the text. This matters most for the secondary claim of a ~30 meV enhancement at non-polaronic sites, which is smaller than the likely error from finite-size and charged-cell artifacts; it could also shift the quoted 240 meV suppression by tens of meV. I do not see a reason to reject the paper, because the central suppression is robust across functionals and polaron-creation methods, and the concern is testable with additional calculations. Since the reader already judged the paper CONDITIONAL and my concern is a more specific version of the same issue, I do not change the verdict.","tokens_in":9646,"tokens_out":5955,"duration_ms":67470,"concrete_test":"Recompute the Fig. 2 heat maps and the Fig. 3a dissociation energies for the 2x4, 4x4, and 4x6 supercells using PBE+U, tracking the polaron center via spin-density or Bader analysis before and after each optimization. In addition, rerun one non-polaronic adsorption site with the polaron artificially pinned to the original Ti5c site (by constraining a Ti-O bond length) and compare with the unconstrained relaxation, and apply a monopole-quadrupole correction to the charged-slab energies instead of relying only on the uniform background. If the 240 meV suppression is stable to within about 50 meV and the sign of the non-polaronic ~30 meV shifts does not change, the concern is resolved; if the polaron migrates or the energies shift substantially, the site-resolved Janus picture requires revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that a small polaron suppresses water dissociation by ca. 240 meV at the polaronic Ti5c site, and the secondary claims of up to ~30 meV enhancement at non-polaronic sites and of hydrogen-bond strengthening, all rest on total-energy differences computed in a single 2x4 supercell with one excess electron and a compensating uniform background charge. The most load-bearing untested assumption is that this cell is large enough that the polaron's elastic and electrostatic field, as well as the adsorbed water or hydroxide, do not interact significantly with their periodic images. The paper reports that Ti-O bonds were sometimes pre-elongated to direct the excess electron toward a chosen Ti atom, but it does not state that the polaron position was constrained or tracked during geometry optimization for each adsorption configuration used to build the heat maps in Fig. 2. If the polaron hops to a different Ti5c site when water or OH is placed at non-polaronic sites, the reported site-resolved energies mix polaron migration energetics with adsorption energetics, making the heat maps ill-defined. No supercell-size convergence test, no charged-cell correction (e.g., Makov-Payne), and no explicit check that the polaron center remains fixed are reported. The 240 meV suppression is likely robust because it is reproduced with HSE and with oxygen-vacancy-created polarons, but the quantitative value and especially the ~30 meV non-polaronic effects are below the plausible error bar of this setup. This is an addressable concern rather than a demonstration that the effect is absent, but it currently leaves the Janus picture quantitatively underdetermined.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports DFT calculations of water adsorption, dissociation, and hydrogen bonding on the rutile TiO2(110) surface in the presence of a small polaron. Using PBE+U (U = 4.2 eV) as the primary method, with HSE and oxygen-vacancy-created polarons as checks, the authors find that a polaron localized on a Ti5c site suppresses water dissociation on that site by about 240 meV (with the NEB barrier rising from 0.14 eV to 0.22 eV), while dissociation on some non-polaronic sites is slightly enhanced by up to about 30 meV. They further report that polarons strengthen hydrogen bonds, especially the O-H...O bonds involving hydroxide groups, which modifies dissociation energetics at higher coverages. The authors connect these findings to ongoing debates about water dissociation on rutile(110) and suggest that polaronic effects are generally important for oxide-water interfaces.","tokens_in":9934,"tokens_out":1981,"duration_ms":20844,"significance":"If the central claim holds, this is an important contribution: it establishes a concrete, site-resolved mechanism by which a small surface polaron — a common defect in titania and other oxides — can act as a local suppressor of water dissociation while slightly promoting dissociation elsewhere. The robustness checks are a genuine strength: the suppression is reproduced with HSE and with polarons created by oxygen vacancies, and the NEB barrier confirms the trend. The paper also makes a falsifiable connection to prior photoemission experiments on polaron segregation. The claim is not circular: the Hubbard U is taken from literature and varied in the SI, and the excess-electron model is a standard way to create a polaron. The main caveat is quantitative: the ~30 meV non-polaronic enhancements and the reported hydrogen-bond strengthening are within the expected error of the DFT+U electronic structure and supercell approximations, and the paper does not yet provide the convergence evidence needed to secure those finer numbers.","major_comments":[{"comment":"The site-resolved heat maps in Fig. 2 rest on the assumption that the polaron remains centered on the chosen Ti5c site during geometry optimization for every adsorption configuration. The main text says only that specific Ti-O bonds were sometimes elongated before optimization 'to direct the excess electron'; it does not state that the polaron position was monitored or constrained in each of the adsorption calculations used to build the maps. If the polaron hops when water or OH is placed on a non-polaronic site, the reported energy differences mix polaron migration energy with adsorption energy, and the heat maps become ill-defined. Please specify, for each site in Fig. 2, whether the gap-state charge density remained localized on the intended Ti atom after relaxation, and list any cases where it did not.","section":"Fig. 2 and Methods"},{"comment":"All quantitative claims, including the ~30 meV non-polaronic enhancements and the ~240 meV suppression, are based on total-energy differences in a single 2x4 supercell with one excess electron and a compensating background charge. No supercell-size convergence test, no charged-cell correction, and no estimate of polaron-image interactions is reported in the main text. Since the polaron is a localized change in both charge and lattice strain, its elastic and electrostatic fields could interact with periodic images at this cell size; the magnitude of this effect is precisely the meV scale of the non-polaronic site dependence. Please provide a convergence study with larger cells (e.g., 2x6 or 2x8) at least for the key configurations that give the largest non-polaronic enhancement and the polaronic-site suppression, or explicitly justify why image interactions are negligible for the 30 meV numbers.","section":"Fig. 2 and computational details"},{"comment":"The HSE and oxygen-vacancy checks in Fig. 3a convincingly show that the qualitative suppression at the polaronic site is not an artifact of the PBE+U model. However, the same checks do not yet establish the quantitative value of the suppression (ca. 240 meV) or the sign and size of the non-polaronic effects, because those numbers are reported only for PBE+U. Given that the non-polaronic enhancement is up to about 30 meV — comparable to zero-point energy differences and to the typical accuracy of DFT+U for such adsorption energies — the paper should either report the site-resolved dissociation energies from HSE for at least the polaronic and most-enhanced non-polaronic sites, or explicitly discuss why the HSE and PBE+U site-resolved patterns are expected to match at this level. Without that, the quantitative 'Janus' claim (enhancement on non-polaronic sites) remains a PBE+U-only result.","section":"Fig. 3a"},{"comment":"The NEB calculation reports a single barrier for the pristine surface and a single barrier for the polaronic surface, but the text does not say whether the product and reactant states on the polaronic surface both have the polaron on the same Ti5c site, or whether the polaron moves during the climb. If the polaron migrates along the reaction coordinate, the barrier of 0.22 eV is not simply the dissociation barrier at a fixed polaronic site but includes polaron migration energetics. Please clarify the polaron position along the NEB path and, if it moves, provide a path with the polaron constrained to the reactive Ti5c site to separate the electronic suppression from polaron mobility.","section":"Fig. 3b"}],"minor_comments":[{"comment":"The heat maps in Fig. 2 would be easier to interpret if the color scale and the exact quantity plotted (change in adsorption energy relative to pristine surface, with positive values suppressing adsorption) were printed directly in the figure rather than only in the caption.","section":"Fig. 2"},{"comment":"The sentence 'water adsorption on non-polaronic sites can be significantly destabilized, with the largest suppression up to ca. 170 meV' is followed by a statement that adsorption at the polaronic site is favored. Please reconcile these two sentences explicitly: the polaronic site is the most stable, but the map shows that some non-polaronic sites are strongly destabilized, which is consistent but could be stated with more precision.","section":"Fig. 2a"},{"comment":"When reporting the hydrogen-bond shortening (1.76 Å to 1.71 Å and 2.02 Å to 1.62 Å), please state whether these are O-H...O distances or H...O distances, and whether they are from PBE+U or another functional. This will help readers compare with the literature.","section":"Page 8"},{"comment":"The phrase 'Janus nature' in the title is evocative but the main text does not define it; consider adding one sentence in the introduction that explains the two faces (suppression at polaronic sites, slight enhancement at non-polaronic sites) to orient the reader.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The central physics is likely correct and the paper is well suited to the journal. My main concern is that the quantitative site-resolved picture, especially the ~30 meV non-polaronic enhancement, rests on assumptions about polaron localization and supercell size that are not yet demonstrated. These are fixable within the scope of the paper (track the polaron in each geometry, run larger cells for key configurations, and report HSE checks for the non-polaronic sites), so I recommend major revision rather than rejection. I would also gently suggest that the authors include the polaron-position tracking information in the main text, since it is the load-bearing point for the heat maps."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is a good paper, worth a serious referee, with one clearly supported central claim and a couple of secondary numbers that are softer than the text implies.\n\nWhat's actually new is the site-resolved picture. They compute adsorption energy maps for water and hydroxide on rutile TiO2(110) in the presence of a small polaron and show that the polaron breaks site equivalence. On the polaronic Ti5c site, water dissociation is suppressed by about 240 meV; the dissociation barrier rises from 0.14 to 0.22 eV. On non-polaronic sites, dissociation can be slightly enhanced (up to ~30 meV) or suppressed, and the polaron strengthens hydrogen bonds between water and hydroxide, mainly because the terminal oxygen becomes a better HB acceptor. That site-resolved map is not in the earlier literature, and the hydrogen-bond mechanism is a genuine addition.\n\nWhat the paper does well: the central suppression is robust. It shows up with PBE+U, with HSE, with polarons created by oxygen vacancies, and in the NEB barrier. The PBE and optB88-vdW results act as negative controls, since those functionals fail to localize the polaron. The citation pattern is solid, and the paper connects its findings to the experimental debate on water dissociation on rutile(110) without overclaiming.\n\nSoft spots, in proportion: the 240 meV number is probably solid, but the smaller effects—the ~30 meV non-polaronic shifts and the hydrogen-bond strengthening—are within the expected error bar of this computational setup. No zero-point energy, no supercell-size convergence test, no charged-cell correction, and the text never explicitly states that the polaron position was tracked or constrained during every geometry optimization. If the polaron hops to a different Ti5c site when water or hydroxide sits at a non-polaronic site, the heat maps mix polaron migration energetics with adsorption energetics and the site labels become ambiguous. That said, this is an addressable problem, not a fatal one. I read the paper as the authors intending the polaron to stay put at the chosen site—they mention pre-elongating Ti-O bonds to direct the electron—but the reported results do not include a check that this is what happened in every configuration.\n\nWho gets value: surface scientists and photocatalysis people who care about water on titania, and DFT practitioners who model polaronic oxides. It deserves a serious referee. Send it to review, but require the authors to add a finite-size check and to state whether the polaron center remained fixed in each calculation.","headline":"A solid DFT study showing surface polarons suppress water dissociation at polaronic Ti5c sites; the secondary energetics need error bars but the core effect holds.","tokens_in":10502,"tokens_out":1914,"would_cite":true,"duration_ms":17914,"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":"On rutile TiO2(110), a single small surface polaron suppresses water dissociation at the polaronic Ti5c site by roughly 240 meV while slightly assisting it at some non-polaronic sites.","keywords":["small polarons","rutile TiO2(110)","water dissociation","hydrogen bonding","surface adsorption","density functional theory","PBE+U","HSE hybrid functional"],"falsifier":"Measure water or D2O dissociation on reduced rutile(110) as a function of controlled polaron population, for example by varying vacuum-reduction level or UV illumination, using scanning tunneling microscopy and photoemission. If the claim is right, water sitting on a polaronic Ti5c site should almost never remain dissociated (reverse barrier near zero), while water on neighboring non-polaronic sites should show a small but measurable increase in dissociation; seeing no site-selective suppression would falsify the mechanism.","tokens_in":9417,"feed_emoji":"💧","tokens_out":6616,"duration_ms":61438,"temperature":0.7,"pith_summary":"The paper sets out to show that small surface polarons—electrons tied to a local lattice distortion—are not passive spectators in the chemistry of rutile TiO2(110). It argues that a polaron changes water adsorption in a site-specific way: water binds slightly more strongly to the polaronic Ti5c site, but the dissociated hydroxide form is electrostatically disfavored there, so the energy cost of splitting water rises by about 240 meV and the forward barrier grows from 0.14 to 0.22 eV. At the same time, some non-polaronic sites become marginally more favorable for dissociation (up to about 30 meV), and polarons strengthen hydrogen bonds in water overlayers. A sympathetic reader would care because this site-dependent, two-faced behavior offers a mechanism that could reconcile long-standing experimental disagreement about whether water dissociates on this surface, and because the same physics should apply to other polaronic oxides.","feed_headline":"Polaron on TiO2(110) blocks water dissociation by 240 meV","feed_subtitle":"The same electron makes neighboring sites slightly more reactive and strengthens hydrogen bonds, a split that may explain the TiO2 water…","key_machinery":"The central object is the small surface polaron: an excess electron localized on a Ti 3d orbital of a five-fold-coordinated titanium atom, stabilized by a local distortion of the surrounding lattice. It does two kinds of work in the argument. First, it acts as a local negative charge that electrostatically repels the negatively charged terminal hydroxide (OtH) produced by water dissociation, which explains why dissociation is suppressed specifically at the polaronic site while molecular water, which is neutral, binds slightly more strongly there. Second, it increases the valence charge on the terminal oxygen of adsorbed water and hydroxide, turning those groups into better hydrogen-bond acceptors and shortening the hydrogen bonds to them; this partially offsets the suppression when water dissociation products are connected by hydrogen bonds. The site-dependence is captured by comparing adsorption energies at the polaronic site and at neighboring non-polaronic sites in a 2x4 supercell.","core_discovery":"The central claim is that on the (110) surface of rutile titania, a small polaron localized on a five-fold-coordinated titanium site has a Janus effect on water chemistry. Water molecules bind favorably to the polaronic site, yet the terminal hydroxide left by dissociation is negatively charged and is repelled by the polaron's excess electron; the net effect is a suppression of water dissociation by about 240 meV on that site, with the reverse dissociation barrier dropping from about 0.18 eV to nearly zero (about 0.01 eV). On non-polaronic Ti5c sites the dissociation energy can shift by up to about 30 meV in either direction, so a small enhancement is possible. The same suppression is found with PBE+U and HSE functionals and when polarons are created by oxygen vacancies rather than by an excess electron. In hydrogen-bonded water structures the polaron strengthens hydrogen bonds, mainly because the terminal oxygen of an adsorbed water or hydroxide becomes a better hydrogen-bond acceptor when the polaron is nearby.","pith_inferences":["If the mechanism is electrostatic site selection, then other anionic adsorbates such as methoxide, halides, or carboxylates should also be repelled from polaron sites while neutral lone-pair donors are attracted; this could be tested with a series of probe molecules on reduced rutile.","Because polaron mobility depends on temperature and photoexcitation, measured dissociation rates on a real TiO2 surface should correlate with polaron population at the adsorption site; a controlled experiment varying electron dose or photon flux could directly probe the 240 meV suppression.","The strengthening of hydrogen bonds by polarons implies polaron formation could shift O-H vibrational frequencies and alter proton-transfer dynamics of interfacial water, a prediction accessible to infrared or sum-frequency spectroscopy.","The quantitative energies rely on a small supercell with one polaron; finite-size and zero-point corrections could shift the 240 meV value, so the most robust prediction is the sign and site-selectivity of the effect rather than the exact energy."],"forward_implications":["Water placed on a polaronic Ti5c site will overwhelmingly stay molecular: the barrier to dissociation rises to 0.22 eV and the barrier for the reverse reaction is essentially zero (about 0.01 eV).","Polaronic substrates present a patchwork of adsorption sites; water on some non-polaronic sites can dissociate up to about 30 meV more easily than on the pristine surface.","At higher water coverages, including monolayers and liquid films, the polaronic suppression persists, though hydrogen-bond strengthening partially offsets it.","The key conclusions do not depend on how the polaron is made: the same suppression appears when the polaron comes from an oxygen vacancy or from an excess electron, and with both PBE+U and HSE treatments.","Polarons should be considered when interpreting experiments on wet titania: samples with different reduction states or polaron populations could give different water-dissociation answers."],"supporting_citations":[{"why":"Experimental observation that water induces polaron segregation to the rutile (110) surface, motivating the water-polaron coupling studied here.","marker":"[23]"},{"why":"First-principles prediction that excess electrons in anatase localize toward surface hydroxyls, supplying the comparative background for polaron-adsorbate interactions.","marker":"[22]"},{"why":"Supplies the PBE+U method used to localize the small polaron in the main simulations.","marker":"[54]"},{"why":"Provides the HSE hybrid functional used to confirm that the suppression at polaronic sites is not an artifact of PBE+U.","marker":"[51, 52]"},{"why":"Supplies the climbing-image nudged elastic band method used to compute dissociation and recombination barriers.","marker":"[70]"},{"why":"Experiments cited as evidence for a dissociation channel on Ti5c sites of pristine rutile, providing the experimental claim the polaronic suppression can explain.","marker":"[44, 45]"},{"why":"Microscopic surface experiments cited as evidence that water does not dissociate on non-defective rutile (110), the counterpart claim the polaron effect can reconcile.","marker":"[46]"},{"why":"Review of the debated water dissociation state on rutile (110), framing the problem the paper addresses.","marker":"[20]"}],"fun_headline_variants":["TiO2 polaron's Janus effect: blocks water split, boosts neighbors","One polaron, two faces: stalls H2O dissociation on TiO2(110)","Polaron on TiO2(110) is a Janus: hinders and helps water split","TiO2(110) polaron: blocks water split at 240 meV, aids neighbors"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole picture rests on the assumption that adding one excess electron to a 2x4 supercell and treating it with PBE+U or HSE gives the right polaron shape, and that the computed energy differences among molecular, dissociated, and hydrogen-bonded states are accurate to within a few tens of millielectronvolts, with no reported correction for zero-point energy, temperature, or supercell size.","fun_headline_variants_meta":{"raw":{"variants":["TiO2 polaron's Janus effect: blocks water split, boosts neighbors","One polaron, two faces: stalls H2O dissociation on TiO2(110)","Polaron on TiO2(110) is a Janus: hinders and helps water split","TiO2(110) polaron: blocks water split at 240 meV, aids neighbors"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000865,"raw_usage":{"total_tokens":3752,"prompt_tokens":951,"completion_tokens":2801,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":567,"completion_tokens_details":{"reasoning_tokens":2707}},"tokens_in":567,"tokens_out":2801,"duration_ms":19319,"temperature":1.0,"reasoning_tokens":2707,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:51:06.586704+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure water or D2O dissociation on reduced rutile(110) as a function of controlled polaron population, for example by varying vacuum-reduction level or UV illumination, using scanning tunneling microscopy and photoemission. If the claim is right, water sitting on a polaronic Ti5c site should almost never remain dissociated (reverse barrier near zero), while water on neighboring non-polaronic sites should show a small but measurable increase in dissociation; seeing no site-selective suppression would falsify the mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Experimental observation that water induces polaron segregation to the rutile (110) surface, motivating the water-polaron coupling studied here."},{"cited_title":"Selcuk and A","cited_arxiv_id":null,"evidence_quote":"First-principles prediction that excess electrons in anatase localize toward surface hydroxyls, supplying the comparative background for polaron-adsorbate interactions."},{"cited_title":"Wang, Y.-G","cited_arxiv_id":null,"evidence_quote":"Microscopic surface experiments cited as evidence that water does not dissociate on non-defective rutile (110), the counterpart claim the polaron effect can reconcile."},{"cited_title":"Diebold, J","cited_arxiv_id":null,"evidence_quote":"Review of the debated water dissociation state on rutile (110), framing the problem the paper addresses."}],"review_version":1}