{"id":"a283a55c-9f44-4e5d-a7df-5eae89e326a7","arxiv_id":"2607.23409","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Y co-doping suppresses B-induced soft phonon modes in α-Fe₂O₃ and combines reduced bandgap (~1.58 eV) with a more balanced dielectric and optical response.","lead":"DFT calculations show boron doping destabilizes hematite phonons, while yttrium co-doping restores stability and improves low-energy optical absorption. The work links lattice dynamics to frequency-dependent optics for photoelectrode design.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The \"Y heals B's soft modes\" mechanism is inferred from a single co-dopant configuration and one supercell size; phonon (in)stability is configuration- and cell-size-sensitive, so the stability half of the central claim may not be robust.","rationale":"I partially agree with the reader: the RPA-without-GW/BSE concern is real and correctly identifies the softest part of the optical ranking (Eg = 1.58/1.65 eV values and the \"improved, smoother response\" language are not quantitatively trustworthy for a correlated, excitonic oxide like hematite, as the cited Piccinin GW-BSE work demonstrates). However, I judge the single-configuration, single-supercell basis of the stability-restoration claim to be at least as load-bearing, because (a) the paper presents the Y-heals-B mechanism as its central novel insight linking dynamics to optics, and (b) it is cheaply testable with the authors' existing setup, unlike a full GW-BSE treatment. Both concerns point the same direction: CONDITIONAL remains the right verdict. The phonon result is plausibly correct — the physics (large Y³⁺ relieving local strain from small B) is reasonable and the co-doped dispersion shown is clean — but \"plausibly correct for one configuration\" is not \"demonstrated.\" The Fvib-for-unstable-structure issue is a smaller red flag in the same neighborhood. I do not recommend REJECT: the work is standard practice, internally documented, and the fixes (more configurations, convergence check, RPA caveats, drop or qualify B-doped optics) are all within reach. Verdict stays CONDITIONAL, with the condition list expanded to include configuration/supercell robustness of the phonon result.","tokens_in":16547,"tokens_out":1951,"duration_ms":14478,"concrete_test":"Enumerate at least three distinct B–Y relative arrangements in the co-doped supercell (nearest-neighbor, next-nearest, separated), relax each, and compute phonon dispersions for each; additionally recompute the B-doped and co-doped phonons in a 2× larger supercell. If any co-doped configuration shows imaginary modes, or if the B-doped soft modes disappear with supercell size, the stabilization mechanism and the claimed stability ranking must be qualified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest claim has a stability half and an optics half. The reader correctly flags the optics half (independent-particle RPA on DFT+U, no GW/BSE, and spectra reported for the dynamically unstable B-doped cell). I focus on the stability half, which the paper itself calls its \"noteworthy finding\" (§3.1): B-doped α-Fe₂O₃ has imaginary modes; (B,Y)-co-doped does not; therefore Y \"suppresses soft modes... through lattice relaxation.\"\n\nThis causal mechanism rests on comparing one B-doped cell against one (B,Y)-co-doped cell. Three specific weaknesses: (1) The paper states B's soft modes persist \"across various doping arrangements,\" showing configuration sensitivity was explored for B — but no analogous statement exists for the co-doped system, whose stability may depend on the B–Y relative placement (nearest-neighbor vs. separated). If stability holds only for the tested arrangement, the general claim \"Y co-doping restores stability\" overreaches. (2) No supercell-size convergence is reported for the phonons; imaginary modes in doped cells can be artifacts of commensuration with the chosen cell, and conversely stability can appear spuriously if the soft-mode wavelength exceeds the cell. (3) Related internal inconsistency: §3.1 reports a harmonic Fvib = 126.4 kJ/mol at 300 K for B-doped α-Fe₂O₃, but vibrational free energy is ill-defined for a structure with imaginary frequencies — the quoted number is not meaningful and slightly undercuts the thermodynamic ranking built on it. None of this is outside-consensus speculation; it is about whether the paper's own comparison supports its causal story.","agreement_with_reader":"partial"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript reports spin-polarized DFT+U (PBE, U_eff = 4.3 eV, D3) calculations of pristine, B-doped, Y-doped, and (B,Y)-co-doped α-Fe₂O₃ at one substitutional concentration per dopant. Phonon dispersions from Phonopy finite displacements (0.01 Å) show imaginary branches for B-doped hematite but entirely positive spectra for pristine, Y-doped, and co-doped cells; the authors conclude that Y co-doping suppresses B-induced soft modes via lattice relaxation, and they support this with temperature-dependent Fvib and Svib. The optical half computes the complex dielectric function in the independent-particle RPA and derives n, k, χE, σop, Rop, and δop via Eqs. (1a)–(1f). Reported optical gaps are 2.30 eV (pristine), 1.65 eV (B), 2.25 eV (Y), and 1.58 eV (B,Y); B introduces a sub-2 eV absorption tail, and the co-doped system is presented as combining low-energy absorption with a smoother dielectric response and restored stability. The central claim is thus two-part: (i) Y heals B's dynamical instability, and (ii) co-doping yields an improved, broadened optical response suitable for photoactive applications.","tokens_in":20610,"tokens_out":2630,"duration_ms":1262037,"significance":"If the results hold, the paper usefully couples two properties usually treated separately — dynamical stability and frequency-dependent optics — for the same set of doped hematite systems, and the finding that co-doping can stabilize a lattice destabilized by one dopant while retaining its sub-bandgap absorption is a design-relevant result for hematite photoanodes. Strengths that deserve explicit credit: a standard, largely reproducible workflow (QE/PAW, literature-derived U = 4.3 eV, Phonopy finite displacements, a complete suite of optical functions derived consistently from ε(ω)); honest reporting of the B-doped instability rather than hiding it; and concrete, falsifiable numerical predictions (Eg values, εre,0, Epeak in Table 1; stability dichotomy). The significance is tempered by the IPA-level optics and by the fragility of the stability evidence, which limit how much of the device-oriented framing can be taken at face value.","major_comments":[{"comment":"The paper's self-described 'noteworthy finding' — that Y suppresses B's soft modes and restores dynamical stability — rests on an asymmetric comparison. For B-doped α-Fe₂O₃ the text states soft modes persist 'across various doping arrangements,' but no analogous statement exists for the co-doped system, whose stability may depend on B–Y relative placement. Furthermore, §2 never states the phonon supercell size, and no supercell-size convergence is reported; imaginary modes can appear or vanish with cell commensuration. Please report phonons for at least two inequivalent B–Y arrangements and demonstrate convergence with supercell size. The mechanism ('lattice relaxation and improved interatomic forces') is currently asserted without evidence; bond-length/strain distributions or mode-resolved force-constant analysis would substantiate it.","section":"§2 (Computational Methodology) and §3.1, Figs. 1–2"},{"comment":"The value Fvib = 126.4 kJ/mol at 300 K is quoted for B-doped α-Fe₂O₃, a structure with imaginary phonon branches. Harmonic vibrational free energy is ill-defined in the presence of imaginary modes (how were they treated in the sum?), so this number — and the four-way Fvib ranking built partly on it ('the (B, Y)-co-doped structure consistently exhibits the lowest Fvib') — is not meaningful as written. Either remove the B-doped entry, compute it with a justified treatment (e.g., excluding or renormalizing unstable modes, with the procedure stated), or restrict the thermodynamic ranking to the three dynamically stable systems.","section":"§3.1, Fig. 3(a)"},{"comment":"All optical quantities derive from an independent-particle RPA dielectric function on DFT+U eigenvalues, with no quasiparticle (GW) or excitonic (BSE) corrections — despite the manuscript citing Piccinin's GW-BSE study [10], which shows both effects are substantial in hematite. The load-bearing optical claims are quantitative: Eg = 2.30/1.65/2.25/1.58 eV, the sub-2 eV absorption tails, and the 'improved/optimal' ranking of the co-doped system. As a concrete test: benchmark the pristine RPA ε(ω) against ref. [10] and/or experimental ellipsometry, state the expected error (e.g., a rigid shift), and carry that uncertainty into the doped results. Relatedly, the B-doped optical spectrum is computed on a dynamically unstable structure; its relevance should be stated as conditional on kinetic stabilization, since the practical absorption claim rests on the co-doped (stable) cell.","section":"§2, Eqs. (1a)–(1f); §3.2.1, Fig. 4, Table 1"},{"comment":"The text reports |σop| maxima 'near 0.8 × 10¹⁵ Hz and 4.5 × 10¹⁵ Hz.' Since 1 eV corresponds to 2.42 × 10¹⁴ Hz, 4.5 × 10¹⁵ Hz is ~18.6 eV — far outside the 0–8 eV window in which ε(ω) (Figs. 4–6) was computed, and from which σop follows via Eq. (1d). The Fig. 7 axis extending to 7.5 (presumably ×10¹⁵ Hz, ~31 eV) compounds the inconsistency. Either the axis units are mislabeled or σop was generated beyond the computed ε data; in both cases the reported σop spectra — and the conclusion that co-doping gives 'enhanced conductivity over a wider frequency range' — cannot be verified. Please reconcile the units and replot within the actual energy window.","section":"§3.2.3, Fig. 7"}],"minor_comments":[{"comment":"§3.1, paragraph on Fig. 2: 'there is no photonic band gap' should read 'phononic band gap' (a photonic band gap would be a very different claim).","section":"§3.1"},{"comment":"The method for extracting the optical Eg values in Table 1 is never stated (onset of εim? Tauc analysis? KS gap?). Please specify, especially since the values (e.g., 1.58 eV) are quoted to three significant figures.","section":"§3.2.1, Table 1"},{"comment":"§2: 'We examined all possible substitutional positions in the host lattice' is too vague to be reproducible — state how many configurations were tested for each system and how the reported one was selected. Also state the magnetic ordering used (α-Fe₂O₃ is antiferromagnetic; this is never mentioned).","section":"§2"},{"comment":"Caption of Fig. 7: 'pristine pristine α-Fe₂O₃' (duplicated word). §3.2.4: 'Beyond λ < 600 nm' should presumably be 'Beyond λ = 600 nm' or 'λ > 600 nm.'","section":"§3.2.3–3.2.4"},{"comment":"The Y-doped δop ≈ 28 µm at λ = 2500 nm (0.5 eV) implies an exponentially small k in the IPA spectrum, where absorption below the gap is numerical noise rather than physics (no phonon-assisted or free-carrier terms are included). The long-wavelength δop discussion should be truncated to energies within the computed interband window or explicitly caveated.","section":"§3.2.4, Fig. 8(b)"},{"comment":"Figures 1, 2, 4–7 appear to lack axis labels/units in places (e.g., Fig. 1 axes show only numbers; the PDOS and THz labels are illegible), and several panels are low-resolution. Please regenerate with labeled axes and units throughout.","section":"Figs. 1–8"},{"comment":"§3.2.2 refers to 'the material's indirect band gap' for pristine α-Fe₂O₃; the direct/indirect character of hematite's gap is debated (cf. ref. [10]) — either justify with the computed band structure or soften the statement.","section":"§3.2.2"}],"recommendation":"major_revision","confidential_remarks":"This is a competent but conventional computational follow-up to the authors' two 2025/2026 papers (refs [12, 31]) on the same B/Y/Nb-doped hematite systems; the marginal advance over those works is the phonon stability analysis, which is also the part with the most fragile support (single co-dopant configuration, unstated supercell size). The manuscript is within scope for a materials/journal-of-alloys-type venue, though its device-level framing overstates what an IPA-level optical study can deliver. I do not see overlap or integrity concerns; the citation pattern is normal self-citation of prior defect-energetics work."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing worth knowing is that they actually compute phonons and full frequency-dependent optics for B-, Y-, and (B,Y)-doped α-Fe₂O₃, and the stability contrast is real in the cells they ran: pristine and Y-doped are stable, B-doped has imaginary modes, and the co-doped cell comes back positive. That pairing is new relative to their own prior electronic-structure papers and to most of the doped-hematite literature they cite.\n\nWhat they do well is the workflow and the presentation of the lattice side. Standard QE DFT+U (U=4.3), Phonopy finite displacement, RPA dielectric, BFGS relaxations — all internally consistent. The dispersions and PDOS make the soft-mode claim checkable. The optical suite (ε, n, k, χ, σ_op, R, δ) is thorough for a design-oriented materials paper, and the co-doped system’s lower Eg plus smoother dielectric profile is a usable design pointer if you already care about B/Y in hematite photoelectrodes.\n\nSoft spots, in proportion. The optics are independent-particle RPA on DFT+U eigenvalues. Hematite is correlated and excitonic; they cite Piccinin’s GW-BSE work and then do not use it. So the low-energy tails, tabulated Eg values, and the ranking of “improved optical response” are qualitative at best. They also report full optical spectra for the dynamically unstable B-doped cell, which is awkward. On the stability half — their “noteworthy finding” — the causal story “Y suppresses B’s soft modes via lattice relaxation” is inferred from one co-doped arrangement and one supercell size. They say B’s soft modes persist across arrangements, but they do not show the same for B–Y relative placement, and there is no phonon supercell convergence. Minor but real: they quote a harmonic F_vib for the B-doped structure that has imaginary modes; that number is not meaningful and slightly muddies the thermodynamic ranking.\n\nNone of this sinks the paper. It is incremental doped-oxide design work, not a fundamental breakthrough. It is for people already optimizing hematite photoanodes or oxide optics who want a stability filter before chasing B doping. Methods are mainstream and reproducible enough for a referee to dig in; no code/data dump, which is normal for this venue class.\n\nI would send it to peer review. Ask them to qualify the RPA limits, justify or drop optics on the unstable B cell, and either show another co-dopant geometry / larger cell or tone down the general “Y restores stability” claim. Worth a serious referee, not a desk reject.","headline":"Clean phonon story on B/Y hematite with a real stability result, but the optics half is RPA-on-DFT+U and the “Y heals B” claim rests on thin configuration evidence.","tokens_in":17614,"tokens_out":664,"would_cite":false,"duration_ms":20462,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Y co-doping with boron restores dynamical stability to hematite and pairs low-energy light absorption with a smoother optical response.","keywords":["hematite","α-Fe2O3","doping","co-doping","phonon dispersion","dynamical stability","dielectric function","optical properties"],"falsifier":"Measure phonon spectra or inelastic scattering on well-characterized B-doped versus (B, Y)-co-doped hematite: if B-doped samples show no soft modes, or if co-doping fails to eliminate them, the stability claim fails; likewise, if measured absorption onsets and low-energy tails disagree with the calculated 1.58–1.65 eV gaps and dielectric peaks, the optical ranking collapses.","tokens_in":17397,"feed_emoji":"☀️","tokens_out":973,"duration_ms":22802,"temperature":0.7,"pith_summary":"This paper argues that doping hematite with boron, yttrium, or both changes both how stable its crystal lattice is and how it absorbs and redirects light across frequencies. First-principles phonon calculations show that boron alone softens the lattice enough to produce imaginary vibration modes, while yttrium alone keeps the structure stable and yttrium-plus-boron suppresses those soft modes through lattice relaxation. On the optical side, boron pulls absorption to lower energies by adding valence states, yttrium reshapes orbital hybridization, and the co-doped material inherits the low-energy absorption while keeping a more balanced dielectric and refractive response. The authors present this co-doping route as a practical way to make hematite more useful for photoelectrodes, optoelectronics, and photonic devices that need both structural durability and stronger visible-light interaction.","feed_headline":"Y plus boron stabilizes hematite and red-shifts its light response","feed_subtitle":"Co-doping kills soft phonon modes while keeping low-energy absorption useful for solar and photonic devices","key_machinery":"Simultaneous phonon-dispersion analysis (finite-displacement Phonopy on DFT+U supercells) and independent-particle RPA dielectric function, from which refractive index, extinction coefficient, optical conductivity, reflectivity, and penetration depth are derived. The machinery links dopant-driven changes in interatomic force constants to the presence or absence of soft modes and to the shape of the optical spectra.","core_discovery":"Pristine and Y-doped α-Fe₂O₃ are dynamically stable with entirely positive phonon modes, whereas B-doped α-Fe₂O₃ develops imaginary modes from Fe–O framework distortions; adding Y together with B suppresses those soft modes and restores stability. At the same time, B lowers the optical gap to about 1.65 eV (1.58 eV when co-doped) and opens low-energy absorption, while Y and especially (B, Y) co-doping improve static dielectric constant and smooth the frequency-dependent optical functions relative to the mono-doped and pristine cases.","pith_inferences":["If soft-mode suppression is mainly ionic-radius and bond-strength driven, other large trivalent cations paired with small p-block dopants may stabilize similarly without needing yttrium specifically.","Device-level tests of co-doped films should check whether the calculated smoother dielectric loss actually reduces recombination or trapping under operating bias and illumination.","Temperature-dependent Raman or neutron data on the co-doped phase would test whether the harmonic free-energy advantage survives anharmonic effects near device temperatures."],"forward_implications":["B-only doping is a poor practical route because lattice instability accompanies the desired low-energy absorption.","(B, Y) co-doping offers a single composition that is both dynamically stable and optically red-shifted into the visible/near-IR.","Static dielectric constant and polarization response improve under Y and co-doping, aiding light confinement and screening.","Penetration depth and optical conductivity become dopant-tunable knobs for photoelectrode thickness and photon-to-charge design.","The same mono-/co-doping logic can be used to screen other hematite dopant pairs for joint vibrational–optical performance."],"fun_headline_variants":["B doping destabilizes hematite phonons; Y co-doping restores stability","Y with boron kills soft modes and red-shifts hematite absorption","Co-doping hematite with B and Y stabilizes lattice and optical gap","B-Y co-doping suppresses imaginary phonons in α-Fe2O3","Y co-dopant relaxes B-distorted Fe-O framework and tunes optics"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The optical spectra and reported band gaps rest on independent-particle random-phase approximation built on DFT+U eigenvalues, without quasiparticle or excitonic corrections that are known to matter for hematite.","fun_headline_variants_meta":{"raw":{"variants":["B doping destabilizes hematite phonons; Y co-doping restores stability","Y with boron kills soft modes and red-shifts hematite absorption","Co-doping hematite with B and Y stabilizes lattice and optical gap","B-Y co-doping suppresses imaginary phonons in α-Fe2O3","Y co-dopant relaxes B-distorted Fe-O framework and tunes optics"]},"model":"grok-4.5","effort":"low","cost_usd":0.005274,"raw_usage":{"total_tokens":1536,"prompt_tokens":931,"num_sources_used":0,"completion_tokens":91,"cost_in_usd_ticks":52744000,"prompt_tokens_details":{"text_tokens":931,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":514,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":931,"tokens_out":91,"duration_ms":10186,"temperature":1.0,"reasoning_tokens":514,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T23:00:16.531119+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Measure phonon spectra or inelastic scattering on well-characterized B-doped versus (B, Y)-co-doped hematite: if B-doped samples show no soft modes, or if co-doping fails to eliminate them, the stability claim fails; likewise, if measured absorption onsets and low-energy tails disagree with the calculated 1.58–1.65 eV gaps and dielectric peaks, the optical ranking collapses.","supporting_citations":[],"review_version":1}