{"id":"81241835-ed83-4dcf-98df-3be4b0a44171","arxiv_id":"2411.18642","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"In ductile functionally graded metals, the angle and width of the gradient determine whether a crack deflects, continues, or stops, with plasticity changing behavior compared with brittle predictions.","lead":"The paper simulates cracks growing through metal parts with a graded composition, using a model that includes plastic flow. It maps how the angle and width of the graded zone determine whether a crack turns, passes through, or stops.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Three-regime map rests on only 12 simulations with no convergence study, no outcome table, and no objective classification; the geometry-only claim is not yet established.","rationale":"I read the central claim as the three-regime map plus the plastic-strain prediction. The phase-field/J2 formulation is standard, and the qualitative comparison between brittle and ductile response is a useful contribution, so I would not reject the paper. But the load-bearing part of the claim is the regime diagram, and its empirical basis within the paper is thin. The reader’s chosen weakest assumption, linear interpolation of properties, is real and acknowledged in Section 3, but even granting that assumption the 12-case design is insufficient to define regime boundaries. Sparse sampling, no table, no error bars, and no convergence or resolution study mean the “regimes” may be artifacts of the discretization or of subjective reading of plastic-strain plots. The claimed predictive use of plastic strain is also post hoc: Section 4.2 describes signatures after the crack behavior is known, rather than a criterion that could anticipate it. I therefore recommend keeping the CONDITIONAL verdict: the paper needs a full outcome table, a convergence check, and an objective classification before the central claim can be accepted. This is a partial agreement with the reader, who correctly identified the property-profile idealization but did not foreground the numerical-support problem.","tokens_in":9664,"tokens_out":5358,"duration_ms":57362,"concrete_test":"For all 12 cases, publish the (θ, W/L) outcome table with an automated classification metric, such as final crack deflection angle and normalized crack extension. Then rerun the three representative cases in Figure 8 at one additional BSAMR refinement level and with ξ halved; if any case changes category, or if the Figure 7 boundaries shift by more than one parameter cell, the regime map is not converged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing concern is not the material-property interpolation (which the authors explicitly flag in Section 3) but the evidentiary status of the central regime map. Section 4.3 claims “three separated regimes for crack behavior… defined and separated only by material geometry” on the basis of 12 simulations (3 angles × 4 widths) yielding 2 deflections, 8 propagations, and 2 arrests. No table reports which (θ, W) pair produced which outcome, so Figure 7’s boundaries cannot be checked. No convergence study is presented: the phase-field length scale ξ = 1.0×10^-5 is never varied, and with a 64×64 base mesh the resolution of the crack scale is not demonstrated. Crack behavior is classified qualitatively from plastic-strain plots, with no quantitative criterion separating deflection from propagation from arrest. Because the regime boundaries are inferred from sparse, unreplicated, subjectively classified points, the claim that regimes are separated “only by material geometry” is currently a plausible hypothesis rather than a supported result; the conclusion even concedes that the behavior depends on material choice, which undercuts the geometry-only formulation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a phase-field ductile fracture model with J2 plasticity implemented in the Alamo finite-difference solver and applies it to mode-I crack propagation through a functionally graded interface between two aluminum alloys (Al 3003 and Al 7075). The parametric study varies the interface angle θ (18°, 45°, 72°) and gradient width W/L (1/6, 1/3, 1, 2), and reports three crack behaviors: deflection, continued propagation, and arrest. The authors claim these regimes are separated only by material geometry, and that the variance of accumulated plastic strain can be used to predict the crack path. The paper also demonstrates one qualitative difference between brittle and ductile predictions for a single geometry.","tokens_in":9849,"tokens_out":4122,"duration_ms":38665,"significance":"If the three-regime map is established, it would provide practically useful guidance for designing ductile FGM interfaces that deflect or arrest cracks. The work addresses an understudied problem, uses material constants from published alloy data rather than fitting parameters to match crack paths, and provides a qualitative brittle-versus-ductile comparison. However, the evidence presented is currently insufficient to support the geometry-only claim: the study is purely numerical, has no experimental validation, no mesh-convergence study, and no objective classification criterion for the reported regimes.","major_comments":[{"comment":"The central claim of three geometry-defined regimes is based on only 12 simulations, yet the paper does not report which (θ, W/L) pair produced which outcome. The text says there were two deflections, eight propagations, and two arrests, but without a table or labeled scatter plot the boundaries drawn in Figure 7 cannot be checked. Please include a table listing all 12 cases with their observed outcome.","section":"§4.3, Fig. 7"},{"comment":"No mesh-convergence or phase-field length-scale convergence study is presented. The regularization length ξ = 1.0×10^-5 is fixed and the base mesh is 64×64; with six AMR levels the effective resolution near the crack is not quantified. Since the classification into deflection, propagation, and arrest depends on whether the crack tip advances or turns, the outcomes could change with discretization or with ξ. A convergence study (varying ξ and mesh refinement) is required to support the regime map.","section":"§3, §2.2.1"},{"comment":"The classification of crack behavior into deflection, continued propagation, and arrest is made qualitatively from plots of plastic strain. No quantitative criterion is defined, so different readers could classify borderline cases differently. Please provide an objective rule (for example, based on final crack-tip position, crack deflection angle, or a threshold in crack advance) and apply it consistently to all 12 simulations.","section":"§4.2"},{"comment":"The conclusion states that the observed behavior 'also depends on the choice of materials', whereas Section 4.3 claims the regimes are 'defined and separated only by material geometry'. These statements are in tension. If material choice is an additional governing parameter, the regime map should be presented as valid for the specific Al-3003/Al-7075 pair studied here, and a broader parameter study is needed before claiming geometry-only separation.","section":"§5 vs. §4.3"},{"comment":"The linear interpolation of all mechanical properties across the gradient is acknowledged as a simplification, and the paper correctly cites experimental studies showing nonlinear property variation. However, because the entire regime map is derived from this property profile, the sensitivity of the deflection/propagation/arrest boundaries to the interpolation choice should be tested (for example, by repeating a subset of simulations with a nonlinear profile). Without such a test, the geometry-only claim is contingent on a modeling assumption that the authors themselves flag as uncertain.","section":"§3"}],"minor_comments":[{"comment":"The sentence 'Phase field (PF) modeling has been shown to robust at predicting crack growth' contains a grammatical error; 'to robust' should be 'to be robust'.","section":"§1"},{"comment":"The angle θ is described as the 'angle of interface' in Section 3 but as the 'angle with which the crack approaches the gradient' in Section 4.3. Please define the angle consistently in the setup and label it clearly in Figure 1.","section":"§3, Fig. 1"},{"comment":"The caption of Figure 2 says 'Variance in phase field crack growth' but the figure compares a brittle simulation and a ductile simulation; the caption should state that comparison explicitly.","section":"§4.1, Fig. 2"},{"comment":"The text refers to 'the more brittle of materials' without identifying which material that is. Since Material 1 has the lower fracture energy Gc, it should be named as the brittle/less-tough material and Material 2 as the tougher material.","section":"§4.2"},{"comment":"The hardening modulus is introduced as H̄ in Eq. (6) but appears as H in the plastic-energy term of Eq. (7); please unify the notation.","section":"§2.1, Eq. (6)-(7)"},{"comment":"Figure 7 is a schematic regime diagram. Please add axis labels and, if possible, plot the 12 data points on it so that the reader can see which cases support the inferred boundaries.","section":"§4.3, Fig. 7"},{"comment":"The manuscript would benefit from a careful proofreading pass for typographical and grammatical issues (for example, 'arresting' versus 'arrest', and inconsistent use of 'gradient' and 'interface').","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely and practical problem, and the modeling framework is a reasonable starting point. The main obstacle is evidentiary: the central three-regime claim is supported by a small number of un-tabulated, qualitatively classified simulations without convergence checks. These issues are fixable within the scope of the manuscript, so I do not recommend rejection, but the paper needs substantial strengthening before it can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper is a clean, well-scoped computational study of crack propagation inside a ductile FGM gradient, and the three-regime picture is worth attention. The evidence behind the picture is thinner than the claims, so treat the regime map as a hypothesis, not a result.\n\nThe phase-field ductile equations are standard, and the authors don't pretend otherwise. What is new is the systematic parameter sweep—three crack incidence angles and four gradient widths, all for a crack propagating inside the gradient region of a ductile FGM. That is a manufacturing-relevant configuration that brittle-FGM and ductile-phase-field papers have not covered. The material constants come from published alloy data (Al 3003 and 7075), no parameters are tuned to reproduce crack paths, and the brittle-versus-ductile comparison in Figure 2 is a genuine model prediction. The authors also deserve credit for explicitly flagging the linear-interpolation idealization in Section 3; it is a simplification, not a hidden one.\n\nThe soft spots are real, though. The central claim—three regimes 'defined and separated only by material geometry'—rests on 12 simulations: 2 deflections, 8 propagations, 2 arrests. There is no table telling the reader which (θ, W) pair gave which outcome, so Figure 7's boundaries cannot be checked. There is no mesh-convergence study; the phase-field length scale is never varied. And the classification of deflection versus propagation versus arrest is qualitative, read off plastic-strain plots. The paper's own conclusion concedes the behavior depends on material choice, which undercuts the 'only by material geometry' phrasing. The accumulated-plastic-strain correlation is asserted more than quantified. These are addressable issues rather than fatal ones, but they make the regime map a well-motivated hypothesis rather than a demonstrated result.\n\nWho gets value: computational fracture folks and additive-manufacturing designers thinking about crack-arresting gradients. It is a useful paper to have in the literature, but it needs a revised version with an outcome table, a convergence study, and an objective classification criterion before I would fully trust the map. I would send it to peer review—it deserves a serious referee—and the referee should push on exactly those points. If the authors can supply the missing support, the three-regime picture would be a solid engineering contribution.","headline":"Clean computational study with a plausible three-regime map that currently rests on twelve qualitative simulations; deserves review but needs more evidence.","tokens_in":2,"tokens_out":3065,"would_cite":false,"duration_ms":54441,"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":"Crack behavior in ductile functionally graded materials falls into three regimes—deflection, continued propagation, and arrest—determined only by gradient geometry, with accumulated plastic strain predicting the path.","keywords":["phase field fracture","ductile fracture","functionally graded materials","J2 plasticity","crack deflection","crack arrest","directed energy deposition","accumulated plastic strain"],"falsifier":"Measure the actual property profile across a directed-energy-deposition aluminum gradient and rerun the same phase-field simulations at $\\theta = 45^\\circ$ and $W/L = 1/3$; if the crack no longer continues through the gradient, or the plastic-strain signature changes, the geometry-only regime classification is refuted.","tokens_in":9446,"feed_emoji":"⚙️","tokens_out":9922,"duration_ms":85958,"temperature":0.7,"pith_summary":"This paper tries to establish that cracks running through the graded interface of a ductile functionally graded metal behave differently from cracks in brittle graded materials, and that the difference is large enough that plasticity cannot be left out of fracture predictions. Using phase-field simulations of a Mode-I crack crossing a gradient between two aluminum alloys, the authors find exactly three behaviors—deflection, continued propagation, and arrest—and argue that these regimes are set only by two geometric parameters: the crack's angle of incidence to the gradient and the gradient's width. They further claim that the time history of accumulated plastic strain carries a recognizable signature for each regime, so plastic strain can be used to predict the crack path. If this is right, narrow gradients with steep property changes can be engineered to deflect or arrest cracks in additively manufactured metal components, while wide gradients are unlikely to deflect cracks at all.","feed_headline":"Three geometry rules decide how cracks move through graded metals","feed_subtitle":"A crack either bounces off, cuts through, or stops, depending on gradient width and angle.","key_machinery":"The load-bearing machinery is a phase-field ductile fracture model coupled to J2 plasticity with linear hardening. A regularized field $c(x)$ smooths the sharp crack over a length scale $\\xi$; the energy functional splits strain energy into tensile and compressive parts, subtracts plastic strain from elastic strain, and adds a plastic-dissipation term $(\\sigma_Y + \\frac{1}{2}H\\alpha)\\alpha$ to the fracture energy. Plastic strain $\\alpha$ evolves through a radial-return algorithm. The crack field is driven by a relaxational evolution equation whose driving force is tensile strain energy plus plastic dissipation minus fracture resistance, with irreversibility enforced by a history field. A near-singular finite difference solver with block-structured adaptive mesh refinement resolves the crack tip and the gradient interface at the required scale. The output that carries the argument is the accumulated plastic strain field $\\alpha$: its time history is what separates deflection, propagation, and arrest.","core_discovery":"The central claim is a geometry-based classification of crack behavior inside ductile functionally graded materials. In a Mode-I phase-field fracture model with J2 plasticity and linear hardening, a crack entering a gradient from the more brittle side toward the tougher side either turns back before crossing, keeps growing straight through the gradient, or stops entirely. The same geometry and loading without plasticity can produce a different outcome: in one illustrative case a brittle crack deflects while the ductile crack continues through, so plastic dissipation changes the crack driving force. Across twelve parameter combinations (three incidence angles and four gradient widths), the three regimes occupy separated regions of the angle–width plane, and each has a distinct accumulated-plastic-strain signature: a rise then fall for deflection, a mild inflection then sustained rise for propagation, and a sharp rise then drop for arrest. The authors conclude that resistance to plastic strain accumulation is a defining driving factor in crack behavior and that variance in plastic strain can be used to predict crack path.","pith_inferences":["If the regime map is as geometry-dominated as claimed, it implies a design chart for additively manufactured part certification: for a fixed alloy pair, angle and width alone select the expected failure mode, which could be checked with instrumented fracture tests.","The linear property interpolation is the assumption most likely to break the map; real directed-energy-deposition gradients can have sigmoidal or locally sharp property profiles, and those profiles could shift or blur the regime boundaries, especially for narrow widths.","The plastic-strain signature suggests an experimental bridge: full-field strain measurements could classify crack outcome by matching the rise-fall or rise-drop pattern before visible deflection or arrest occurs.","The regime separation rests on only twelve parameter combinations, so a denser sweep or an analytical energy-balance criterion could test whether the boundaries are as sharp as claimed."],"forward_implications":["Plasticity must be included in fracture predictions for ductile FGMs: a brittle-only model can predict crack deflection where the ductile simulation shows continued propagation through the gradient.","Narrow gradients with a high rate of change in material properties are the route to engineered crack-deflection or crack-arrest behavior in additively manufactured ductile FGMs.","Wide gradients reduce crack deflection, so laminate-like narrow gradient structures hold more promise than wide graded regions for crack-resistant design.","Accumulated plastic strain can be used as a predictive diagnostic: the qualitative shape of its time history tells which regime a crack is entering before the path is fully decided.","For common nonstructural FGM uses such as coefficient-of-thermal-expansion matching or magnetic tailoring, the bulk fracture behavior is not substantially affected by the gradient when the gradient is wide."],"supporting_citations":[{"why":"Supplies the energy-balance fracture criterion on which the phase-field energy functional is built.","marker":"[23]"},{"why":"Supplies the variational energy-minimization formulation of brittle fracture that the ductile model extends.","marker":"[24]"},{"why":"Supplies a phase-field formulation for ductile fracture with plastic degradation that this model adapts.","marker":"[16]"},{"why":"Provides the brittle-FGM phase-field study whose crack behavior the ductile results are contrasted against.","marker":"[26]"},{"why":"Supplies the massively parallel block-structured adaptive mesh refinement solver infrastructure used for the simulations.","marker":"[31]"},{"why":"Provides the earlier ductile phase-field treatment of multi-materials and FGMs that this work extends to cracks within the graded region.","marker":"[37]"},{"why":"Supplies the quartic degradation and geometric functions used in the crack field evolution.","marker":"[41]"},{"why":"Supplies the radial-return algorithm used for the J2 plasticity updates.","marker":"[42]"},{"why":"Supplies the aluminum alloy material property data used to set the two material phases.","marker":"[43]"},{"why":"Documents nonlinear property variation in gradient interfaces, the behavior the paper's linear interpolation assumption departs from.","marker":"[44]"}],"fun_headline_variants":["Three geometry rules govern crack fate in graded metals","Cracks in graded metals: three outcomes from angle and width","Plastic strain maps which way cracks go in ductile FGMs","Angle and width predict three crack behaviors in graded metals","Crack deflection, propagation, or arrest: set by gradient geometry"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire regime map assumes that all mechanical properties vary linearly across the gradient interface, while real manufactured gradient interfaces are known to vary nonlinearly.","fun_headline_variants_meta":{"raw":{"variants":["Three geometry rules govern crack fate in graded metals","Cracks in graded metals: three outcomes from angle and width","Plastic strain maps which way cracks go in ductile FGMs","Angle and width predict three crack behaviors in graded metals","Crack deflection, propagation, or arrest: set by gradient geometry"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000406,"raw_usage":{"total_tokens":2067,"prompt_tokens":857,"completion_tokens":1210,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":473,"completion_tokens_details":{"reasoning_tokens":1127}},"tokens_in":473,"tokens_out":1210,"duration_ms":10880,"temperature":1.0,"reasoning_tokens":1127,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T12:40:50.447068+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the actual property profile across a directed-energy-deposition aluminum gradient and rerun the same phase-field simulations at $\\theta = 45^\\circ$ and $W/L = 1/3$; if the crack no longer continues through the gradient, or the plastic-strain signature changes, the geometry-only regime classification is refuted.","supporting_citations":[{"cited_title":"Vi. the phenomena of rupture and flow in solids,","cited_arxiv_id":null,"evidence_quote":"Supplies the energy-balance fracture criterion on which the phase-field energy functional is built."},{"cited_title":"Revisiting brittle fracture as an energy minimization problem,","cited_arxiv_id":null,"evidence_quote":"Supplies the variational energy-minimization formulation of brittle fracture that the ductile model extends."},{"cited_title":"A phase-field formulation for fracture in ductile materials: Finite deformation balance law derivation, plastic degradation, and stress triaxiality effects,","cited_arxiv_id":null,"evidence_quote":"Supplies a phase-field formulation for ductile fracture with plastic degradation that this model adapts."},{"cited_title":"Phase field modelling of crack propagation in functionally graded materials,","cited_arxiv_id":null,"evidence_quote":"Provides the brittle-FGM phase-field study whose crack behavior the ductile results are contrasted against."},{"cited_title":"Massively parallel finite difference elasticity using block-structured adaptive mesh refinement with a geometric multigrid solver,","cited_arxiv_id":null,"evidence_quote":"Supplies the massively parallel block-structured adaptive mesh refinement solver infrastructure used for the simulations."},{"cited_title":"Phase-field ductile fracture analysis of multi-materials and functionally graded composites through numerical and experimental methods,","cited_arxiv_id":null,"evidence_quote":"Provides the earlier ductile phase-field treatment of multi-materials and FGMs that this work extends to cracks within the graded region."},{"cited_title":"On degradation functions in phase field fracture models,","cited_arxiv_id":null,"evidence_quote":"Supplies the quartic degradation and geometric functions used in the crack field evolution."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the radial-return algorithm used for the J2 plasticity updates."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the aluminum alloy material property data used to set the two material phases."},{"cited_title":"Gradient material mechanics: perspectives and prospects,","cited_arxiv_id":null,"evidence_quote":"Documents nonlinear property variation in gradient interfaces, the behavior the paper's linear interpolation assumption departs from."}],"review_version":1}