{"id":"0118eb91-1aa6-440e-b6e5-8f628dd14911","arxiv_id":"2411.18638","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Three-dimensional phase-field simulations show that non-planar temperature isotherm patterns significantly alter cellular spacing, mushy zone length, and microsegregation during additive-manufacturing-like solidification.","lead":"This paper uses 3D phase-field simulations to show that the shape of the temperature isotherm, not just its gradient and speed, changes the cellular microstructure that forms when a laser melt pool solidifies. It matters because it suggests a route to control microstructure and defects in metal 3D printing by shaping the heat input pattern, not only changing power or speed.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Pulsed-isotherm case is a non-steady-state snapshot with unreported duty cycle, yet it carries much of the 'finer cells and reduced microsegregation' conclusion.","rationale":"The reader's weakest assumption concerns transfer to real melt pools via the frozen-temperature approximation and absence of convection. That is a legitimate limitation, but the paper states it clearly and it does not affect the internal validity of the simulation comparison. The more actionable weakness is that the pulsed case, a principal source of the finer-cells and reduced-microsegregation claim, is not a steady-state solidification run: the thermal field switches between G and 0, and the resulting two-zone structure in Fig. 4f is history-dependent. Since the paper reports no tp and gives no cycle-convergence test, the quantitative magnitudes in Figs. 5-13 for the pulsed case are not yet established. The other non-planar cases are steady-state in principle, but their amplitudes are reported only as bare numbers with no units, so their physical magnitude is also ambiguous. The Echebarria model, constant G/V control, and qualitative agreement with pulsed-laser experiments are genuine supporting evidence, so I would not reject the paper. The existing conditional verdict already requires missing replication and parameter disclosure; my recommendation is therefore to keep the verdict unchanged and to add the specific pulsed-cycle test as a condition.","tokens_in":20391,"tokens_out":11973,"duration_ms":127014,"concrete_test":"Rerun the pulsed case for at least five complete on/off cycles, reporting tp and the number of cycles, and recompute λc from the χ maximum and kv using only the portion of the solid formed after cycle 3, where cycle-to-cycle changes are small. Also run one additional tp value, for example 2x or 0.5x the baseline, to test duty-cycle sensitivity. If the cycle-converged differences versus the planar case fall below about 5% or change sign, the headline conclusion loses its pulsed support; if they remain near 15%, the transient-history concern is resolved and the remaining issue is parameter disclosure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is most vulnerable at the pulsed isotherm, not at the frozen-temperature approximation. Section 3.1 says steady-state signatures are extracted, but Eq. (13) alternates between a moving planar field and a uniform T0, and Fig. 4f shows two morphologically distinct zones: deep cells in the laser-on portion and shallow cells in the laser-off portion. The final snapshot is therefore a transient that depends on the on/off history. The abstract's claim that non-planar isotherms give finer cells and reduced microsegregation leans heavily on the pulsed case in Figs. 7 and 12, and the residence time tp is not reported anywhere, so the duty cycle and number of pulses cannot be reconstructed. The authors' own Appendix A concedes that pulsed-G runs may not reach steady state and that results depend on residence time. Until a periodic steady state or a controlled sweep over tp is demonstrated, the reported 14-15% differences in kv and λc cannot be cleanly attributed to isotherm geometry; they may be pulse-history transients. This is testable independently of the physical-realism debate.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports three-dimensional phase-field simulations of directional cellular solidification in a Ni-Nb alloy under a frozen-temperature approximation, comparing six prescribed isotherm patterns (planar, Gaussian-noise, sinusoidal, transverse, curved, and pulsed) that all move at constant velocity V with the same nominal thermal gradient G. The central claim is that small variations in the isotherm shape, even at fixed G and V, measurably alter solidification outcomes: cell spacing, mushy-zone extent, solid-fraction steepness, microsegregation (kv), and solute-rich droplet fraction. The authors characterize these outcomes using solid-fraction profiles, Euler-characteristic percolation analysis, bridging-plane statistics, and line concentration profiles, and they report up to roughly 60% differences in steepness, 15% in cell spacing, 14% in microsegregation, and 40% in droplet fraction between patterns. The work is positioned as a first qualitative step toward incorporating melt-pool isotherm geometry into microstructure models.","tokens_in":20540,"tokens_out":3869,"duration_ms":44788,"significance":"If the reported differences are robust, the paper would provide a useful qualitative demonstration that isotherm geometry alone, decoupled from G and V, can shift cellular microstructural descriptors in a regime relevant to laser powder-bed fusion. The study has several genuine strengths: it uses a standard, previously validated quantitative phase-field model; it keeps all material, numerical, and cooling parameters fixed while varying only the isotherm pattern; it is genuinely three-dimensional; and it employs nontrivial morphological statistics (Euler characteristic, bridging plane, percolation of solid/liquid phases) rather than relying on visual inspection. The qualitative agreement with published experiments on sinusoidal and pulsed laser strategies is also encouraging. However, the central quantitative claims are currently supported by single simulations per pattern on a small lateral domain, and one of the two patterns most responsible for the headline 'finer cells, reduced microsegregation' conclusion is a transient, non-steady-state protocol whose control parameters are not fully reported.","major_comments":[{"comment":"The pulsed-isotherm case is not a steady-state protocol, yet it carries much of the 'non-planar isotherms produce finer cells and reduced microsegregation' conclusion. Equation (13) alternates between a moving planar field and a uniform T0, and Fig. 4f shows two morphologically distinct zones (deep cells during laser-on, shallow cells during laser-off). The residence time tp is nowhere reported, so the duty cycle and number of pulses cannot be reconstructed from the manuscript. Appendix A concedes that pulsed-G runs 'may not reach a steady state' and that the results depend on residence time. The conclusions in Section 5 and the values in Figs. 7 and 12 therefore conflate geometry effects with pulse-history transients. The authors should either demonstrate a periodic steady state, provide a controlled sweep over tp and pulse count, or remove the pulsed case from the headline quantitative comparisons.","section":"§3.1, Eq. (13); Fig. 4f; Appendix A"},{"comment":"The quantitative claims of 15% differences in cell spacing and 14% differences in microsegregation are presented with 'confidence intervals' described as the standard deviation around the mean obtained by averaging the data, but each isotherm pattern is represented by a single simulation. There is no ensemble averaging over initial noise realizations, no spatial subsampling protocol, and no statistical test. Given the lateral domain of 1.024 µm contains only roughly 8-17 cells for the reported λc values, these differences may be within the natural spread of a single small-domain simulation. The authors should specify exactly how the mean and standard deviation were computed, how many independent cell-spacing or concentration samples contributed, and ideally provide multiple realizations (e.g., different random seeds for the initial noise) to demonstrate that the between-pattern differences exceed within-pattern variability.","section":"§3.2, Figs. 7 and 12"},{"comment":"The 'Gaussian noise' isotherm as written does not appear to create a spatially varying isotherm. In Eq. (9), δ is described as 'the random number' drawn from [-1,1], which, if it is a single scalar per timestep, only adds a global shift to the temperature field and leaves the isotherm planar. The authors report identical results for planar and noise cases across all metrics; if δ is spatially uniform, this identity is by construction and does not constitute a finding about noise robustness. The manuscript should define a genuine spatial noise field with an amplitude, correlation length, and random seed, and then show how the results depend on those parameters; alternatively, the noise case should be presented as a null check with the specification made explicit.","section":"Eq. (9) and Figs. 5-6, 11-12"},{"comment":"The isotherm amplitudes and frequencies are introduced as 'just reference values' with no calibration to the experimental or process conditions they purport to represent. In particular, An = 0.5, As = 0.5, Ac = 0.0005, and tp are not derived from thermal simulations or experimental data, and no parameter sweep is performed except for the tilt angle in the transverse case. The abstract's phrasing that 'small variations in the isotherm can considerably impact' the microstructure is therefore not quantitatively established: the calculations show that certain chosen finite-amplitude distortions change the outcome, but they do not show that variations small compared to realistic melt-pool disturbances cause those changes. At minimum, a sensitivity study over the isotherm amplitudes, or a calibration to published thermal-field data, is needed before the 'small variations' claim can be supported.","section":"§2.1, Eqs. (9)-(13)"}],"minor_comments":[{"comment":"The pulsed case is defined with a formatting error ('t>t p') and the phrase 'before setting it off for a time tp' is ambiguous: it should be stated explicitly whether the laser-off interval has the same duration tp, and the total number of on/off cycles within the 80,000 Δt runtime should be reported.","section":"Eq. (13)"},{"comment":"Panel (d) is labeled 'Angular' in the figure caption but 'transverse' everywhere else in the text; this inconsistent terminology should be corrected.","section":"Fig. 8 caption"},{"comment":"There is a typo, 'An preliminary analysis' should read 'A preliminary analysis'.","section":"Fig. 10 caption"},{"comment":"The authors state that 'we could not make a quantitative comparison of our results with the literature'; given that the paper reports quantitative percentage differences in the conclusions, it would be helpful to state explicitly which aspects are intended to be qualitative and which are intended to be quantitative, so readers do not over-interpret the listed percentages.","section":"§4, Discussion"},{"comment":"The conclusion states that solid percolation occurs for fs between 0.6 and 0.8, but the mechanism connecting the bridging plane of χ = 0 to this fs interval is not explained in the text; a sentence describing how the solid fraction at the bridging plane is obtained and why this range matters would improve clarity.","section":"Section 5, percolation statement"}],"recommendation":"major_revision","confidential_remarks":"The paper is not circular: the isotherm shapes are prescribed and the cellular structures emerge from the phase-field equations, and the frozen-temperature approximation is openly acknowledged as a simplification. The main gatekeeping issue is statistical and protocol-based rather than conceptual: the pulsed case is transient, the noise case may be trivially planar, and single runs on a small domain cannot support 14-15% effect sizes. I believe these issues are fixable within the manuscript's scope by removing or re-characterizing the pulsed case, properly defining the noise field, and adding replication or a formal uncertainty analysis. The topic fits the journal's materials-science readership, but the authors should also consider whether 'first three-dimensional phase-field simulations' is an overclaim in light of existing 3D phase-field studies of cellular growth under AM-like conditions; this is a novelty-scope point the editor may wish to weigh."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid, clearly written 3D phase-field study that does something genuinely new—it treats the shape of the moving isotherm as an independent variable and surveys five non-planar patterns against the usual planar reference. The Euler-characteristic analysis of the bridging plane is a nice addition, and the tilted-growth result from an inclined isotherm is a real observation. It deserves a serious referee.\n\nWhat the paper does well: the model is standard (Echebarria et al.) and previously validated; the material parameters are mostly from published sources; and the frozen-temperature and no-convection assumptions are stated plainly and argued for. The qualitative claim—that isotherm geometry measurably affects cell spacing, mushy zone extent, and microsegregation—is supported by the simulations and is consistent with the cited experiments on sinusoidal and pulsed laser strategies. Credit is due for the honest Discussion and for the appendix's admission that the pulsed simulations may not reach steady state.\n\nWhere the soft spots are: the headline percentages (up to 15% in cell spacing, 14% in microsegregation, 60% in solid-fraction steepness) come from one simulation per isotherm pattern. The confidence intervals in Figs. 7 and 12 are spatial averages, not independent-run statistics, so they do not quantify run-to-run variability. The sinusoidal frequency and pulse residence time are never reported, which makes the study hard to reproduce and means the pulsed duty cycle is unknown. More importantly, the pulsed case is a transient: Eq. (13) alternates between a moving planar field and a uniform T0, and Fig. 4f shows two morphologically distinct zones. Appendix A concedes that such runs may not reach steady state. Hanging quantitative differences on that case is shaky. That said, the stress-test note overstates the damage: the sinusoidal isotherm, a genuinely steady-state pattern, also gives finer cells and lower microsegregation than planar, so the abstract's qualitative claim does not collapse if the pulsed numbers are set aside. The pulsed case should be either reported with a proper duty cycle and a demonstrated periodic steady state, or demoted to qualitative observation.\n\nBottom line: worth publishing after revision. A referee should ask for replication or at least error bars from multiple runs, full reporting of the missing parameters, and a clear handling of the pulsed transient. I would not desk-reject.","headline":"Solid first 3D survey of isotherm-shape effects on cellular solidification; the qualitative story holds, but the pulsed case and missing parameters keep the quantitative claims from being fully trustworthy.","tokens_in":21137,"tokens_out":2488,"would_cite":false,"duration_ms":25268,"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":"Small variations in isotherm shape, at fixed gradient and velocity, change cellular solidification microstructure in a melt pool.","keywords":["directional solidification","temperature gradient patterns","isotherm shape","cellular growth","phase-field simulation","mushy zone","microsegregation","additive manufacturing"],"falsifier":"Run a directional-solidification experiment, or a coupled thermal-fluid solidification simulation, in which only the curvature of the temperature isotherm is changed while $G$ and $V$ stay fixed; if cell spacing, mushy-zone depth, and microsegregation do not shift by roughly the reported amounts (~15%, ~60%, and ~14%), the central claim fails. A practical version would compare two laser beam shapes or scan patterns that produce different melt-pool boundary curvatures but the same measured $G$ and $V$, and measure cell spacing and segregation in the solidified track.","tokens_in":20088,"feed_emoji":"🔬","tokens_out":13209,"duration_ms":121528,"temperature":0.7,"pith_summary":"The paper asks whether the shape of the moving temperature isotherm itself, not just the magnitude of the thermal gradient and the solidification speed, controls the cellular microstructures that form when a laser melt pool solidifies. To isolate this, the authors run three-dimensional phase-field simulations of a Ni-Nb alloy (a quasi-binary stand-in for Inconel 718) at fixed gradient $G = 10^7$ K/m and velocity $V = 0.1$ m/s, comparing a planar isotherm against noise-perturbed, sinusoidal, transverse, curved, and pulsed variants. The central claim is that small isotherm distortions substantially change the outcome: non-planar isotherms give finer cell spacing, a longer semisolid mushy zone, reduced microsegregation, and different amounts of solute-rich droplets, with differences up to about 60% in mushy-zone steepness, 15% in cell spacing, 14% in microsegregation, and 40% in droplet fraction. A tilted isotherm tilts the whole cellular array as well. The paper concludes that isotherm geometry should be treated as a microstructural control variable and as a source of uncertainty in additive manufacturing models.","feed_headline":"Changing isotherm shape shifts melt-pool cell spacing ~15%","feed_subtitle":"3D phase-field runs show non-planar temperature fronts give finer cells, longer mushy zones, and less segregation.","key_machinery":"The load-bearing machinery is a quantitative phase-field model for dilute binary alloy solidification with a frozen temperature approximation: the temperature field is prescribed analytically and advected rigidly at constant velocity $V$ with constant gradient $G$, so the only thing that changes between runs is the shape of the isotherm surface. Five distorted isotherms are compared with the planar reference: Gaussian-noise-perturbed, sinusoidal, transverse (tilted by angle $\\varphi$), parabolically curved, and pulsed (planar isotherm during the laser-on period, uniform temperature during the laser-off period). The resulting microstructures are read out through solid-fraction profiles $f_s(z)$, the Euler characteristic $\\chi(z)$ of solid-liquid connectivity in transverse planes, mean cell spacing $\\lambda_c$, line concentration profiles, and the microsegregation ratio $k_v = c_s^*/c_{\\max}$; the zero crossing of $\\chi$ locates the bridging plane where the mushy zone transitions from liquid-like to solid-like.","core_discovery":"At fixed thermal gradient and growth velocity, the geometry of the temperature isotherm alone changes the cellular solidification morphology in three dimensions. Using phase-field simulations of a Ni-Nb alloy in the additive-manufacturing regime, the paper shows that planar, noise-perturbed, sinusoidal, transverse, curved, and pulsed isotherms produce measurably different outcomes: the steepness of the solid-fraction profile differs by up to ~60% between patterns, average cell spacing by ~15%, microsegregation (the ratio $k_v = c_s^*/c_{\\max}$) by ~14%, and the fraction of solute-rich droplets emitted from intercellular grooves by ~40%. Non-planar isotherms, especially sinusoidal and pulsed, produce finer cells and reduced microsegregation relative to the planar reference; the transverse isotherm produces the longest mushy zone and coarsest cells and tilts the cellular array. The paper concludes that isotherm shape, not only $G$ and $V$, belongs in the list of factors controlling melt-pool solidification microstructures.","pith_inferences":["Beyond the paper: because $G$ and $V$ were held fixed, the reported differences isolate isotherm geometry as an independent control; a natural test is to correlate measured melt-pool boundary curvature with local cell spacing in a single material across different beam shapes.","Beyond the paper: the Gaussian-noise case behaving like the clean planar case suggests random thermal fluctuations matter less than coherent isotherm distortion; thermal modeling may need to capture organized curvature rather than add stochastic noise.","Beyond the paper: the larger change in mushy-zone steepness (~60%) than in cell spacing (~15%) implies isotherm shape acts more on the deep grooved region than on tip selection; this could be tested by measuring intercellular groove depth and microporosity in samples built with different scan strategies.","Beyond the paper: coupling the frozen-temperature phase-field model to melt flow and latent heat would show whether the qualitative trends survive real isotherm deformation, a direction the paper itself identifies as future work."],"forward_implications":["Isotherm patterns should be included as a microstructural control variable in additive manufacturing models, alongside the nominal gradient and velocity.","Scan strategies that create sinusoidal or pulsed isotherms are expected to produce finer cells and less solute partitioning than planar-front approximations predict, consistent with experiments on sinusoidal hatching and pulsed beams.","Local melt-pool boundary curvature can tilt cellular growth away from the nominal gradient direction, explaining some experimentally observed misorientations.","Mushy-zone length and the position of the liquid-to-solid percolation transition shift with isotherm pattern, so defect-prone zones such as hot cracks and porosity may be manipulated by thermal pattern design.","Different droplet fractions emitted from cell grooves imply different amounts of secondary phase formation during terminal solidification, depending on isotherm pattern."],"supporting_citations":[{"why":"Supplies the quantitative phase-field model equations for phase and concentration evolution used in all simulations.","marker":"[34]"},{"why":"Provides the additive-manufacturing Ni-Nb phase-field modeling context and the regime used to set the thermal gradient and velocity.","marker":"[7]"},{"why":"Provides the earlier methodology for primary spacing and microsegregation in laser-deposited Ni-Nb that the simulations build on.","marker":"[24]"},{"why":"Supplies the power-spectrum measure of cell spacing used as the cross-check in Eq. (24).","marker":"[36]"},{"why":"Supplies the quasi-binary Inconel 718 alloy (Ni-Nb) phase diagram and material parameters used in the simulations.","marker":"[42]"},{"why":"Supplies the typical laser powder bed fusion values of thermal gradient and velocity used in the simulations.","marker":"[44]"},{"why":"Experiment with sinusoidal laser scanning in 316L stainless steel that observed refined cellular structures, used as qualitative comparison.","marker":"[22]"},{"why":"Pulsed versus continuous-wave laser experiments on Inconel 718 showing finer cells, reduced segregation, and tilted dendrites, used as the main experimental comparison.","marker":"[12]"},{"why":"Cluster-labeling algorithm used to compute the Euler characteristic and solid-liquid percolation statistics in the mushy zone.","marker":"[51]"}],"fun_headline_variants":["Isotherm shape alone shifts melt-pool cell spacing by 15%","Non-planar isotherms give finer cells and less segregation in melt pools","Isotherm geometry changes cellular morphology in 3D melt-pool simulations","Melt-pool isotherm pattern alters cell size and solute segregation","Small isotherm variations reshape melt-pool solidification cells"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a melt pool's temperature field can be represented as a rigidly translating isotherm of a fixed shape with constant gradient and velocity, with no feedback from latent heat, convection, or the solidifying interface; if real isotherms are coupled to melt flow and the moving laser, the quantitative differences reported here would not transfer to the real process.","fun_headline_variants_meta":{"raw":{"variants":["Isotherm shape alone shifts melt-pool cell spacing by 15%","Non-planar isotherms give finer cells and less segregation in melt pools","Isotherm geometry changes cellular morphology in 3D melt-pool simulations","Melt-pool isotherm pattern alters cell size and solute segregation","Small isotherm variations reshape melt-pool solidification cells"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001308,"raw_usage":{"total_tokens":5352,"prompt_tokens":984,"completion_tokens":4368,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":600,"completion_tokens_details":{"reasoning_tokens":4289}},"tokens_in":600,"tokens_out":4368,"duration_ms":39536,"temperature":1.0,"reasoning_tokens":4289,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T14:48:05.074587+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a directional-solidification experiment, or a coupled thermal-fluid solidification simulation, in which only the curvature of the temperature isotherm is changed while $G$ and $V$ stay fixed; if cell spacing, mushy-zone depth, and microsegregation do not shift by roughly the reported amounts (~15%, ~60%, and ~14%), the central claim fails. A practical version would compare two laser beam shapes or scan patterns that produce different melt-pool boundary curvatures but the same measured $G$ and $V$, and measure cell spacing and segregation in the solidified track.","supporting_citations":[{"cited_title":"Echebarria, R","cited_arxiv_id":null,"evidence_quote":"Supplies the quantitative phase-field model equations for phase and concentration evolution used in all simulations."},{"cited_title":"Karayagiz, L","cited_arxiv_id":null,"evidence_quote":"Provides the additive-manufacturing Ni-Nb phase-field modeling context and the regime used to set the thermal gradient and velocity."},{"cited_title":"Ghosh, L","cited_arxiv_id":null,"evidence_quote":"Provides the earlier methodology for primary spacing and microsegregation in laser-deposited Ni-Nb that the simulations build on."},{"cited_title":"Ghosh, N","cited_arxiv_id":null,"evidence_quote":"Supplies the power-spectrum measure of cell spacing used as the cross-check in Eq. (24)."},{"cited_title":"Knorovsky, M","cited_arxiv_id":null,"evidence_quote":"Supplies the quasi-binary Inconel 718 alloy (Ni-Nb) phase diagram and material parameters used in the simulations."},{"cited_title":"Ghosh, J","cited_arxiv_id":null,"evidence_quote":"Supplies the typical laser powder bed fusion values of thermal gradient and velocity used in the simulations."},{"cited_title":"Mussatto, R","cited_arxiv_id":null,"evidence_quote":"Experiment with sinusoidal laser scanning in 316L stainless steel that observed refined cellular structures, used as qualitative comparison."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Pulsed versus continuous-wave laser experiments on Inconel 718 showing finer cells, reduced segregation, and tilted dendrites, used as the main experimental comparison."},{"cited_title":"Hoshen, R","cited_arxiv_id":null,"evidence_quote":"Cluster-labeling algorithm used to compute the Euler characteristic and solid-liquid percolation statistics in the mushy zone."}],"review_version":1}