{"id":"eb3d0f9a-4ef0-435d-84a5-1d076270780f","arxiv_id":"1908.04669","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Higher flame power in solution precursor HVOF spraying of Nb-doped TiO2 increased anatase content and reduced porosity, contrary to the usual high-temperature anatase-to-rutile trend.","lead":"This paper deposits niobium-doped titanium dioxide coatings using a liquid-precursor thermal spray method and finds that a hotter flame produces denser coatings with more of the anatase crystal form. The result is a step toward making doped ceramic coatings with controllable microstructure and phase for sensors and transparent conductors.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 25 kW vs 75 kW comparison confounds flame power with a ~3x change in total gas flow and droplet dynamics; with one run per condition and no error bars, the causal 'power increases anatase' claim is not established.","rationale":"The reader and I identify the same weakest point: the central causal claim rests on a two-point comparison in which the 'flame power' setting was changed by changing both fuel and oxidizer flows, so the control variable is not isolated. The paper's own Section 2.1 shows total gas flow increases from 260 to 762 l/min; the simulated temperatures of 1880 and 2480 degrees Celsius support a melting mechanism, but they do not identify which process variable is responsible. This is a correctness risk rather than an internal inconsistency: the direction of the effect may be reproducible, and the single-splat study and TGA/DSC give credible qualitative support for the proposed in-flight transformation model. However, without replication, intermediate power points, or a gas-flow-matched control, the phrase 'three calculation methods confirmed' overstates the evidence. The three methods are also correlated, and the systematic uncertainties from undoped reference structures and unquantified amorphous content are acknowledged in the text. A conditional verdict is therefore the most honest assessment; no change to the reader's verdict is needed. If the proposed matrix test were run and the trend held across intermediate powers and replicates, the causal claim would be much stronger.","tokens_in":14492,"tokens_out":5276,"duration_ms":57245,"concrete_test":"Run a spray matrix with at least five flame-power settings (25, 35, 50, 60, and 75 kW) at fixed O2/H2 ratio, stand-off distance, and precursor feed, with two or three replicate coatings per condition; quantify anatase by Rietveld refinement including an internal standard, such as 10 wt% corundum, and an amorphous contribution. If the anatase fraction does not increase monotonically with power, or the 25 versus 75 kW difference is within replicate scatter, the causal claim fails. To separate flame enthalpy from gas momentum, additionally compare a 75 kW run against a 25 kW run whose total gas flow is raised to match the 75 kW condition by adding nitrogen to the 78/182 l/min H2/O2 mixture; if anatase content tracks total gas flow rather than flame power, the attribution in the conclusions is wrong.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing assumption is that the anatase increase (5.1% to 18.7% by Rietveld; same direction by peak height and peak area) is caused by flame power alone. Section 2.1 specifies only two runs: 25 kW with 78 l/min H2 and 182 l/min O2, and 75 kW with 229 l/min H2 and 533 l/min O2. Total gas flow increases by a factor of about 3, so nozzle exit velocity, gas momentum, droplet Weber and Reynolds numbers, in-flight residence time, and convective heat transfer all change together with the combustion enthalpy. The O2/H2 ratio is nearly unchanged, so stoichiometry may not be the main confound, but the two runs are not a single-variable comparison. The paper's proposed mechanism, complete melting at the higher flame temperature followed by rapid cooling to nucleate anatase, is plausible, yet the data cannot exclude altered precursor fragmentation or solvent evaporation caused by the higher gas flow as the actual driver. Each condition was sprayed only once and no replicate runs or error bars are reported, so run-to-run scatter is not ruled out. The 'three methods' are not fully independent: the peak-height and peak-area formulas use the same two reflections with constants calibrated for undoped titania, and the Rietveld refinement uses undoped ICSD structures while not quantifying the amorphous content that the text itself mentions. Thus the conclusion that flame power 'equated to an increase in anatase phase content' overstates a two-point correlation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the first deposition of niobium-doped TiO2 coatings by solution precursor high-velocity oxy-fuel (SP-HVOF) spraying at two nominal flame powers, 25 kW and 75 kW. The coatings were characterized by SEM/EDS and XRD, with anatase content quantified by Rietveld refinement, peak-height, and peak-area methods. Single-splat swipe tests at stand-off distances of 65, 85, and 105 mm, together with TGA/DSC of the precursor and dried powder, are used to propose a qualitative model of precursor breakup, solvent evaporation, precipitation, pyrolysis, sintering, melting, and splat formation. The central claim is that increasing flame power increases anatase content from 5.1% to 18.7% (Rietveld) and reduces coating porosity, and that this trend is confirmed by all three XRD quantification methods.","tokens_in":14822,"tokens_out":3038,"duration_ms":33235,"significance":"If the central claim holds, the paper would establish flame power as a practical control handle for phase content in SP-HVOF of Nb-doped TiO2, which matters for photocatalytic, sensing, and transparent-conductive applications. The work is also useful as a first demonstration of SP-HVOF for this material system, and the swipe-test series at three stand-off distances is a well-designed qualitative probe of in-flight particle evolution. Credit is due for explicitly acknowledging several limitations of the XRD quantification, such as the use of undoped ICSD structures and the presence of unquantified amorphous content. However, the causal attribution to flame power rests on a single pair of spray runs that differ in multiple gas-flow parameters, and the three 'independent' quantification methods share the same reflections and calibration assumptions, so the headline claim is not yet established at the level of confidence implied by the conclusions.","major_comments":[{"comment":"The causal attribution of the observed phase and porosity differences to flame power is not supported by the experimental design. The 25 kW run used 78 l/min H2 and 182 l/min O2, while the 75 kW run used 229 l/min H2 and 533 l/min O2, so the total gas flow increases by roughly a factor of three. Nozzle exit velocity, gas momentum, droplet Weber and Reynolds numbers, secondary fragmentation, in-flight residence time, and convective heat transfer all co-vary with the combustion enthalpy. With one spray per condition and no replicate runs or error bars, the statement in the conclusions that 'an increase in flame power equated to an increase in the anatase phase content' overstates a two-point correlation. The authors should either add a single-variable or factorial spray matrix, or clearly reframe the claim as a correlation between the full 25 kW/75 kW parameter sets rather than a causal effect of flame power alone.","section":"Section 2.1 and Section 3.1"},{"comment":"The claim that 'three calculation methods confirmed' the anatase increase is not fully justified because the three methods are not independent. The peak-height method (Eq. 1) and the peak-area method (Eq. 2) are both applied to the same anatase (101) and rutile (110) reflections, and both use calibration constants determined for undoped titania in other spray processes. The Rietveld refinement also used undoped ICSD structures and did not quantify the amorphous content that the text itself mentions as a possible cause of peak broadening. The agreement among the three methods therefore reflects shared systematic biases as well as any real phase-content difference. The authors should report the refined lattice parameters and goodness-of-fit values, quantify or bound the amorphous content, and provide uncertainties on the reported 5.1% and 18.7% values.","section":"Section 3.1.1, Equations (1) and (2), Figure 4"},{"comment":"The porosity reduction at higher flame power is presented as a central microstructural result, but it is supported only by qualitative inspection of BSE-SEM cross-sections in Figure 1. No image-analysis-based porosity measurements, standard deviations, or replicate cross-sections are provided. A quantitative porosity estimate would strengthen the manuscript and is needed to support the statement that flame power is an effective way to modify coating microstructure.","section":"Section 3.1 and Figure 1"}],"minor_comments":[{"comment":"The heading 'Results and Conclusions' is unusual; if the conclusions are collected in Section 4, the third section should be titled 'Results and Discussion'.","section":"Section 3 title"},{"comment":"The caption states 'All images were taken at 20 kW'; this should read '20 kV' to refer to the SEM accelerating voltage.","section":"Figure 8 caption"},{"comment":"The sentence 'In both cases the stand-off distance was 85 mm' should clarify that this applies to the coating deposition runs, since the swipe tests in the same section use stand-off distances of 65, 85, and 105 mm.","section":"Section 2.1"},{"comment":"The statement that solid features are absent at larger stand-off distances is too strong, since the 85 mm swipe sample clearly contains solid particulates mixed with molten splats in Figure 8; the text should distinguish between craters from solid-particle impacts and the presence of solid particles generally.","section":"Section 3.2.1"},{"comment":"The numerical constants in Equations (1) and (2) are cited to prior literature, but the manuscript should state explicitly that these constants were derived for undoped titania and may not transfer directly to Nb-doped SP-HVOF coatings with different crystallite sizes.","section":"Equations (1) and (2)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a reasonable fit for a thermal-spray or coatings journal, but the phase-content claim needs stronger experimental backing. The main gap is not a mathematical error but an experimental-design issue: one run per condition with multiple co-varying parameters. I would be willing to reconsider after the authors add more spray conditions, replicate runs, or an explicit uncertainty analysis, and after they soften the causal language to match what the data can support."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a legitimate first demonstration, not a breakthrough, and the central causal claim is stronger than the data support. The novel bit is the first solution precursor HVOF deposition of Nb-doped TiO2 and a phase trend that says higher flame power gives more anatase, which goes against the usual anatase-to-rutile story. That is worth knowing.\n\nWhat it does well: the deposition itself is a sensible extension of the group's earlier SHVOF work, the single-splat collection at three stand-off distances is a useful way to visualize in-flight transformation, and the TGA/DSC data anchor the proposed model in real thermal events. The XRD work uses three quantification methods, and the authors are upfront that the calibration constants and structure models come from undoped titania, which is the right kind of honesty for a first report.\n\nThe soft spots are real but not disqualifying. The 25 kW and 75 kW runs differ in both hydrogen and oxygen flow rates, so gas velocity, momentum, droplet Weber/Reynolds numbers, residence time, and convective heat transfer change along with the flame power. With one run per condition and no error bars, \"flame power increases anatase\" is a two-point correlation, not a controlled result. The three XRD methods also share the same two reflections and constants, so they are not independent confirmation. The refinement does not quantify amorphous content the text itself says is present. That said, the confound weakens the attribution, not the observed trend; the coatings genuinely differ and the proposed rapid-melt-and-quench mechanism is plausible.\n\nFor a first report, this is fine. The overreach is in the conclusions, where \"proven\" and \"equated to\" go beyond what two runs can support. The recommendation is to send to peer review, with a request for the authors to add replicate runs and preferably an intermediate condition, and to soften the causal language if they cannot.\n\nWho it is for: thermal spray and functional coating researchers, particularly those working on liquid-feedstock titania. I would not cite it in my near-term work, but I would bring it to a reading group as an example of a solid first demonstration that overstates causality.","headline":"A solid first demonstration of SP-HVOF Nb-doped TiO2 with a plausible but under-supported flame-power/anatase trend.","tokens_in":15333,"tokens_out":2152,"would_cite":false,"duration_ms":22404,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Higher flame power raises the anatase phase content of solution-precursor sprayed Nb-doped titanium oxide coatings, contrary to the usual high-temperature expectation.","keywords":["solution precursor thermal spray","HVOF","niobium-doped titanium dioxide","anatase","rutile","single splat","phase content","porosity"],"falsifier":"Spray the same Nb-doped precursor at 25, 50, and 75 kW while holding the hydrogen-to-oxygen ratio fixed, and repeat each condition several times; if the anatase content does not rise monotonically with flame power, or if run-to-run scatter is as large as the 5.1% to 18.7% gap, the claimed causal role of flame power would be refuted.","tokens_in":14269,"feed_emoji":"🔥","tokens_out":7596,"duration_ms":74414,"temperature":0.7,"pith_summary":"This paper presents a counterintuitive result in solution precursor high-velocity oxy-fuel spraying: niobium-doped titanium dioxide coatings made with a hotter flame contain more of the anatase phase, normally the less stable form at high temperature, not more rutile. Three independent x-ray diffraction analyses put the anatase content at about 5% for a 25 kW flame and about 19% for a 75 kW flame, with porosity dropping at the higher power. The proposed cause is that the hotter flame fully melts the in-flight material, and its rapid cooling on impact favours anatase nucleation from the melt, while niobium doping additionally slows the usual anatase-to-rutile conversion. If this is right, flame power becomes a practical control dial for phase content and density in coatings deposited from solution precursors.","feed_headline":"Hotter flame boosts anatase phase in sprayed titania","feed_subtitle":"In solution-precursor HVOF coatings, anatase rises from about 5 to 19 percent as flame power goes from 25 to 75 kW.","key_machinery":"The analysis hinges on three phase-quantification routes applied to the same x-ray diffraction data: Rietveld refinement, a whole-pattern fitting method; a peak-height formula based on the anatase (101) and rutile (110) reflections; and a peak-area formula for the same reflections. Together they establish the direction of the phase change rather than relying on one method. The physical explanation is carried by a proposed in-flight transformation model: primary and secondary droplet fragmentation, solvent evaporation, precipitation or shell formation, pyrolysis and sintering, melting, and splat impact. The model is anchored by swipe-test single splats collected at stand-off distances of 65, 85, and 105 mm, and by TGA-DSC measurements that identify the pyrolysis, crystallisation, and anatase-to-rutile transformation temperatures of the precursor.","core_discovery":"The discovery claimed is that flame power controls phase selection in solution precursor HVOF of Nb-doped TiO2: raising the flame power from 25 kW to 75 kW raises the anatase content from 5.1% to 18.7% as measured by Rietveld refinement, with peak-height and peak-area methods showing the same direction, while coating porosity falls sharply. The paper interprets this through the melting state of the material on impact. At 25 kW the simulated flame temperature, about 1880 °C, is barely above the melting point of TiO2, so solid, sintered particles arrive only partially molten and build a porous coating; at 75 kW, about 2480 °C, particles melt completely and the rapid cooling of splats on the substrate favours anatase nucleation directly from the melt. Niobium doping is treated as a contributing factor because it inhibits the anatase-to-rutile transformation, but not as the primary cause. Single-splat collections at three stand-off distances show the progression from droplets to precipitated, pyrolysed, sintered, and finally molten material, with niobium-rich outer layers on the splats.","pith_inferences":["If the melt-nucleation interpretation is right, the same higher-power/higher-anatase trend should appear in other dopant-titania combinations sprayed from solution precursors, provided the dopant does not suppress anatase nucleation; that is a testable prediction the paper does not make.","Because the 25 kW and 75 kW runs co-varied fuel and oxygen flows, a cleaner version of this study would separate flame enthalpy from gas velocity and residence time, for example by varying total gas flow at fixed hydrogen-to-oxygen ratio while tracking droplet momentum.","The paper's constant niobium signal across flame powers suggests that doping level and phase content can be set somewhat independently, which points toward tuning conductivity and photocatalysis separately in Nb-doped titania coatings; the paper does not pursue those functional consequences."],"forward_implications":["Flame power can be used as a process parameter to tune coating density: low power yields porous, loosely bonded coatings, and high power yields dense coatings with fully molten splats.","Phase content in solution-precursor titania is not monotonic in the usual thermodynamic direction; hotter spraying can increase the fraction of anatase, so process design should not assume that high temperature means rutile.","The proposed in-flight transformation model gives a concrete route from a liquid precursor to a final coating, predicting that stand-off distance controls the balance between solid particulates and molten splats.","Quantifying titania phases by three independent XRD methods provides a template for checking phase claims in thermal spray coatings where peak broadening and small grain sizes complicate single-method results."],"supporting_citations":[{"why":"Supplies the peak-height formula used to quantify anatase content from the (101) and (110) reflections.","marker":"Ref 40"},{"why":"Supplies the peak-area formula used as the second independent phase quantification method.","marker":"Ref 42"},{"why":"Gives the simulated flame temperatures, about 1880 °C at 25 kW and 2480 °C at 75 kW, that anchor the melting argument.","marker":"Ref 43"},{"why":"Documents how niobium doping retards the anatase-to-rutile transformation through reduced oxygen vacancies.","marker":"Ref 46"},{"why":"Provides the thermodynamic basis for anatase nucleation from molten TiO2 upon rapid cooling.","marker":"Ref 48"},{"why":"Supplies the TGA/DSC assignment of pyrolysis and crystallisation peaks of dried titania precursor.","marker":"Ref 33"}],"fun_headline_variants":["Hotter flame, more anatase in sprayed titania","Flame power controls phase in Nb-doped TiO2 coating","Surprise: Higher HVOF flame favors anatase over rutile","Flame temperature steers anatase fraction in TiO2 spray"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central claim assumes that the two spray runs differed only in flame power, but the lower-power run used 78 l/min hydrogen with 182 l/min oxygen and the higher-power run used 229 l/min hydrogen with 533 l/min oxygen, so gas velocity, momentum, stoichiometry, and residence time also changed; if one of those co-varying factors drives the phase or porosity difference, the flame-power claim does not stand alone.","fun_headline_variants_meta":{"raw":{"variants":["Hotter flame, more anatase in sprayed titania","Flame power controls phase in Nb-doped TiO2 coating","Surprise: Higher HVOF flame favors anatase over rutile","Flame temperature steers anatase fraction in TiO2 spray"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000226,"raw_usage":{"total_tokens":1490,"prompt_tokens":992,"completion_tokens":498,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":608,"completion_tokens_details":{"reasoning_tokens":428}},"tokens_in":608,"tokens_out":498,"duration_ms":5399,"temperature":1.0,"reasoning_tokens":428,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:35:20.567406+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Spray the same Nb-doped precursor at 25, 50, and 75 kW while holding the hydrogen-to-oxygen ratio fixed, and repeat each condition several times; if the anatase content does not rise monotonically with flame power, or if run-to-run scatter is as large as the 5.1% to 18.7% gap, the claimed causal role of flame power would be refuted.","supporting_citations":[],"review_version":1}