{"id":"c8e14993-e20f-4781-987e-8d508ef3511e","arxiv_id":"2505.14095","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Gold nanoislands on a sodium aluminophosphosilicate glass incorporate sodium and phosphorus from the glass at 550 °C in air, forming stable multi-element islands with altered lattice and electronic properties.","lead":"Gold nanoislands on a specially formulated phosphate glass absorb sodium and phosphorus from the glass when heated to 550 °C in air, forming stable multi-element islands. The finding offers a simple, additive-free route to anchoring and chemically modifying gold nanostructures for catalysis, sensing, and optoelectronics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The solid-solution claim is built on contradictory d-spacing evidence: HRTEM/XRD show Au(111) dilation while NBD reports ~7% contraction, and XPS labels P as phosphate, so Na/P incorporation into the Au lattice is not established.","rationale":"The paper makes a striking, potentially important observation: after thermal dewetting on NAPS glass, P/Na/Al are found on extracted GNIs, the islands are embedded strongly enough to leave cavities after ultrasonication, and NAPS-30 controls behave differently. These observations are internally consistent and support a genuine interfacial reaction/anchoring phenomenon, and the paper deserves credit for the control experiments (SiO2-Au) and multi-technique characterization. My concern is narrower but load-bearing: the specific claim that Na and P dissolve into the Au lattice as an Au-rich solid solution—the premise for lattice distortion, d-band modification, and hot-carrier conclusions—is supported only by NBD d-spacing values that contradict the HRTEM/XRD dilation reported in the same paper, and by DFT models that actually predict dilation rather than contraction. The XPS table labeling P as PO4^3- adds a further inconsistency, since atomic P incorporation would require a reduced P species. Because the strongest claim in the abstract and conclusions depends on this solid-solution interpretation, the paper should not be accepted without resolving the sign and chemical state of the lattice modification. The proposed cross-sectional STEM-EELS/EDS experiment would settle the issue directly. This is not a rejection of the underlying phenomenon; it is a request to replace an internally contradictory structural inference with direct, calibrated evidence. The reader's conditional verdict therefore remains appropriate, with the condition made more explicit: the solid-solution and electronic-modification claims should be substantiated or removed.","tokens_in":30801,"tokens_out":8321,"duration_ms":80341,"concrete_test":"Prepare a FIB cross-section through an intact NAPS-15-Au island and acquire aberration-corrected HAADF-STEM images with simultaneous EELS/EDS line scans across the Au/glass interface; measure the Au(111) d-spacing as a function of depth into the island using the bulk Au lattice parameter as an internal calibration, and determine the P L2,3 or P 2p chemical state. If the d-spacing shows no sustained inward gradient and P is in a phosphate-like environment, the Au-rich solid-solution claim is not supported; if a gradient and reduced P are detected, the RGMI lattice-incorporation mechanism is confirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that RGMI produces Au-rich solid solutions with Na/P incorporated in the Au lattice rests on the structural analysis in the 'Atomic-Scale Architecture' section, but that evidence is internally contradictory. HRTEM/IFFT (Figs 2C–E, S13–S14) and ex-situ/in-situ XRD (Fig S24A–B) show an increase in Au(111) d-spacing (lower 2θ), i.e. lattice dilation, which the authors attribute to Na incorporation. In the same extracted particles, NBD (Figs 2G–H, S19) is quoted as giving d(1-11)=0.218 nm and d(220)=0.131 nm, a ~7% contraction relative to pure Au (0.235 nm and 0.144 nm); the implied lattice constant (~0.378 nm) contradicts the dilation. The DFT model (NaP-Au(111), Fig 4B/S25) predicts increased interplanar spacing, matching the HRTEM/XRD result but not the NBD contraction. It is not possible for the same nanoislands to be both dilated and contracted by ~7% unless the two measurements probe different phases, which is not established. Moreover, angle-resolved XPS Table S2 explicitly assigns P as P(PO4^3-), and no P 2p chemical-state analysis is given for the extracted GNIs; the proposed reaction Au + yBOx → AuBy + xyO2 requires phosphate reduction to elemental P, for which no evidence is presented. Thus the load-bearing premise—dissolved Na/P atoms in the Au lattice—is not secured. What is secured is that glass-derived elements are associated with the islands and that the islands are mechanically anchored.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports that annealing 50 nm DC-sputtered gold films on sodium aluminophosphosilicate (NAPS-X) glasses at 550 °C in air causes the films to dewet into gold nano-islands (GNIs) that become enriched in Na, P, Al, and some Si, while being mechanically anchored in the glass. The authors introduce the term reactive glass-metal interaction (RGMI) to describe this process, arguing that it is a new class of metal-support interaction distinct from classical RMSI because it uses a non-reducible glass under oxidizing conditions. Supporting experiments include EDS/STEM mapping of extracted islands, XPS depth profiles and angle-resolved XPS, HRTEM/IFFT and nano-beam diffraction, ex-situ/in-situ XRD, AFM morphology and cavity-depth analysis, and transient absorption spectroscopy. These are complemented by molecular-dynamics simulations of the NAPS-X glasses and DFT models of Na and P adsorption on Au(111). The paper concludes that RGMI yields multi-element gold nano-islands with modified electronic structure, lattice distortion, and hot-carrier dynamics, with potential applications in catalysis, sensing, and optoelectronics.","tokens_in":31149,"tokens_out":6542,"duration_ms":70514,"significance":"The proposed RGMI concept is potentially significant because it would extend metal-support interaction chemistry to non-reducible glass supports under ambient, oxidizing conditions, in contrast to RMSI on reducible oxides in hydrogen-rich environments. The experimental effort is broad and largely well documented: several glass compositions are compared, the mechanical-anchoring evidence based on post-ultrasonication cavities is convincing, the XPS/EDS association of glass elements with the islands is reproducible, and the MD and DFT calculations are described in sufficient detail to be checked. The authors also state limitations of the in-situ measurements, which is commendable. However, the strongest structural claim—that Na and P dissolve in the Au lattice and distort the Au(111) planes—is currently built on internally contradictory diffraction evidence, and the chemical-state evidence for elemental P or Au-P bonding is missing. These points must be resolved before the central mechanistic and electronic-structure conclusions can be accepted.","major_comments":[{"comment":"The manuscript presents mutually incompatible lattice-parameter evidence for the same extracted NAPS-15 GNIs. HRTEM/IFFT (Figs 2C–E, S13–S14) and ex-situ/in-situ XRD (Fig S24A–B) are interpreted as showing an increased Au(111) d-spacing (lattice dilation), whereas NBD (Figs 2G–H, S19) yields d(1-11)=0.218 nm and d(220)=0.131 nm, approximately 7% contracted relative to pure Au and corresponding to a≈0.37–0.38 nm. These two results cannot both describe a single Au-rich phase. The text needs a quantitative reconciliation, including calibration and error analysis for the IFFT and NBD measurements and a statement of whether core and periphery are different phases, or the solid-solution/lattice-distortion claim should be withdrawn. As written, the conclusion that Na/P incorporation distorts the Au lattice is not secured.","section":"Atomic-Scale Architecture (Figs 2C–H, S13–S14, S19, S24)"},{"comment":"The proposed dominant reaction Au + yBOx → AuBy + xyO2, with B = P, Si, Al, requires reduction of phosphate to elemental P and its solution in Au. However, the angle-resolved XPS in Table S2 assigns the P 2p signal to P(PO4^3-), and no P 2p chemical-state spectrum is provided for the extracted GNIs; the EDS maps are equally consistent with phosphate-containing glass or a phosphate-based surface layer attached to the islands. Without direct chemical-state evidence on the extracted particles, or a control that removes residual glass, the claim of Au-rich solid-solution/alloy formation is not established. The data as presented support glass-element association with the islands and mechanical anchoring, which are the robust parts of the RGMI observation.","section":"Mechanistic Insights (reaction pathway; Table S2)"},{"comment":"The DFT modeling is presented as 'further validation' of the TEM-observed interlayer expansion, but the NaP-Au(111) model was selected because it best reproduces that same disputed TEM result (Section S12; Fig S25). This comparison cannot independently confirm the lattice-distortion claim. Please state the predictive content of the DFT, for example quantitative agreement with measured interlayer changes and bond lengths, and, if possible, compare the models against the NBD contraction or against a measurement that does not depend on the HRTEM IFFT analysis.","section":"DFT section (Fig 4; S12, S25)"},{"comment":"The main text states that the electron-phonon coupling constant g 'increases systematically with silica concentration up to 25 mol% before decreasing for NAPS-30-Au,' while Section S13 states the opposite: 'The g value for MEGNIs decreases with increasing SiO2 concentration in NAPS-X-Au glasses up to 25 mol% and then increases for NAPS-30-Au glass.' Since the g-versus-composition trend is the central evidence for composition-controlled hot-carrier dynamics, this contradiction must be resolved. In addition, attributing the g changes to density-of-states variations near the Fermi level is speculative without direct electronic-structure measurements on the MEGNIs.","section":"Ultrafast Carrier Dynamics (Fig 5; S13)"}],"minor_comments":[{"comment":"The text refers to 'elemental mapping in HAADF-STEM mode (Figure 1F, S15-S16),' but the relevant main-text panel appears to be Figure 2F, not Figure 1F; please correct the cross-reference.","section":"Atomic-Scale Architecture"},{"comment":"The acronyms RGMI and RMGI are used interchangeably, for example 'our RMGI phenomenon' in the Mechanistic section; please use RGMI consistently.","section":"Throughout"},{"comment":"Section S13 refers to 'Figure 5G' for the zero-power extrapolation, but Figure 5 in the main text has no panel G; the corresponding result appears to be in Figure S29C. Please fix the reference.","section":"S13 (Transient Absorption Spectroscopy)"},{"comment":"Equation (1) uses x in the denominator of the Fermi function exponent, while Table S1 lists x with units of nm^-1; for dimensional consistency x should be a length in nm. Please define x unambiguously.","section":"Eq (1) and Table S1"},{"comment":"The fitted d1/2 values (123 nm and 84 nm) are much larger than the deposited 50 nm Au film; this may reflect island topography and sputter-induced roughening, so the absolute depth scale should be presented with appropriate caveats.","section":"XPS depth profiling"}],"recommendation":"major_revision","confidential_remarks":"The experimental work is broad enough that I see a solid basis for a revised manuscript centered on surface enrichment of Au nano-islands by glass elements and strong mechanical anchoring. However, the current version overstates the lattice-incorporation and electronic-modification claims, and the reported g trend is self-contradictory. I recommend major revision rather than rejection because the contradictions are localizable and the core phenomenon of glass-element migration to the islands appears reproducible. The authors should be asked to either provide reconciled diffraction evidence with error analysis or explicitly soften the solid-solution claims in the abstract and conclusions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Things to know up front: the headline claim that Na and P dissolve into the Au lattice to form an Au-rich solid solution is not supported by the paper's own data. HRTEM/IFFT and XRD show lattice dilation (larger d-spacing), NBD shows ~7% contraction, and XPS identifies P as phosphate, not the elemental P the proposed reaction requires. That inconsistency is load-bearing.\n\nWhat is solid: the systematic NAPS-X composition series, the multi-technique characterization (EDS, XPS depth/angle, TEM, TA), the five-model DFT comparison, and the hot-carrier data. The observation that glass-derived elements decorate and anchor the gold islands is convincing, and the cavity-depth trend with SiO2 content is a nice compositional handle. Credit also for the explicit acknowledgment in Section S2 that prior H-glass work already reported chemically interacting glass–gold hybrids; this paper extends that to a mechanistic picture. The authors honestly note the difficulty of in-situ characterization in oxidizing atmospheres.\n\nSoft spots, in order of severity. First, the contradiction above. The same extracted particles cannot be both dilated and contracted by ~7% unless the measurements probe different phases or regions, and that is not established. The DFT model (NaP-Au(111)) predicts dilation, matching HRTEM/XRD but not NBD; that model is also an adsorption structure, not a solid solution, so it cannot rescue the ‘incorporation’ story. Second, the XPS data label P as phosphate; no P 2p chemical-state data are given for the extracted GNIs, so the reductive pathway Au + yBOx → AuBy + xyO2 has no spectroscopic support. Third, the hot-carrier interpretation attributes composition-dependent g values to DFT-derived DOS, but the DOS is for an idealized adsorption geometry; size, embedment, and thermal conductivity effects are acknowledged as competing factors, making the assignment speculative. The sputter-depth calibration is a minor concern, not a fatal one.\n\nNet: the phenomenon called RGMI is probably real, and the dataset is valuable, but the solid-solution/lattice-distortion argument needs major revision or a softened claim. The right outcome is peer review with a request for reconciliation of the structural data, chemical-state analysis of P in the islands, and either direct cross-sectional proof of subsurface incorporation or a revised conclusion that stops at surface adsorption.\n\nWho gets value: anyone working on metal–support interactions, glass-embedded nanoparticles, or plasmonic hot-carrier engineering. I'd bring it to a reading group to discuss the structural inconsistency, but I would not cite the solid-solution conclusion in its current form.\n\nRecommendation: send to serious referees, expect heavy revision.","headline":"A valuable dataset undermined by an internally contradictory structural argument: HRTEM/XRD show lattice dilation while NBD shows contraction, so the solid-solution claim is not secured, though the glass–gold interaction itself is probably real.","tokens_in":31789,"tokens_out":2419,"would_cite":false,"duration_ms":25989,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-07T15:39:48.444440+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}