{"id":"3513f03b-1228-448f-8f44-7f0eadaa26b9","arxiv_id":"2607.03172","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Partial amorphization of NiTi into crystal–amorphous superlattices introduces a continuous-then-first-order martensitic path, halves thermal hysteresis, and raises modulus and critical stress by 60–90%.","lead":"Molecular dynamics shows that partially amorphizing NiTi into crystal–amorphous superlattices cuts thermal hysteresis by ~60% and raises stiffness and transformation stress by 60–90%. The result offers a materials-design route for shape-memory alloys that avoids hunting for exotic coherent second phases.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"The continuous second-order-like pathway and quantitative hysteresis/modulus gains rest on unvalidated MEAM transferability to crystal–amorphous interfaces.","rationale":"The Reader correctly isolates the MEAM transferability issue as the weakest assumption. All reported Ms/Af values, the continuous pathway, retained-austenite snapshots, and the 60–90 % mechanical gains are direct numerical outputs of that single potential applied to crystal–amorphous interfaces never present in its training set. Relative trends inside the model are internally consistent and mechanistically plausible, so the comparative claims survive within the simulation framework; absolute design-rule status does not. No stronger internal inconsistency (e.g., thermodynamic violation or numerical artifact independent of the potential) is evident. Hence the verdict remains CONDITIONAL, with experimental validation or multi-potential confirmation required before the CAS design strategy can be treated as established.","tokens_in":11278,"tokens_out":568,"duration_ms":5894,"concrete_test":"Re-run the 53 % Xtal cooling–heating cycle (Fig. 2 protocol) with an independent NiTi potential that has been validated for amorphous or liquid NiTi (e.g., a recent machine-learning or EAM potential with published amorphous RDF/structure-factor agreement). If the continuous pre-Ms martensite rise disappears or the hysteresis reduction falls below ~40 %, the load-bearing interfacial mechanism is potential-specific and the design claim weakens.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The paper’s central claim—that partial amorphization introduces a continuous (second-order-like) martensite rise before the first-order jump, cuts thermal hysteresis from 275 K to 95–110 K, and raises modulus/critical stress 60–90 %—is generated entirely inside the 2NN MEAM potential of Ko et al. (Phys. Rev. B 92, 134107, 2015). That potential was fitted and validated for crystalline B2 \rightleftharpoons B19' energetics, lattice parameters, and transformation temperatures; it was never shown to reproduce amorphous NiTi structure factors, crystal–amorphous interfacial energies, or the local stress fields that the authors invoke as the origin of heterogeneous nucleation and retained austenite (see §2.1 construction protocol, PTM classification, and the continuous rise in Fig. 2 for 43 %/53 % Xtal). If the potential artificially softens or over-stabilizes the interface, both the two-stage pathway and the absolute magnitude of the property shifts become simulation artifacts rather than transferable design rules. The fully crystalline baseline already over-predicts experimental hysteresis (acknowledged in §3.1), so relative improvements inherit the same untested interfacial physics.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript uses large-scale MD simulations with the 2NN MEAM potential of Ko et al. to study temperature- and stress-induced martensitic transformations in crystal–amorphous superlattices (CAS) of NiTi with crystalline fractions of 100%, 66%, 53%, and 43%. Structures are generated by a selective melt–quench protocol that holds designated crystalline slabs at 500 K while amorphizing adjacent regions. Cooling–heating trajectories show that partial amorphization introduces an initial continuous (second-order-like) rise in martensite fraction before the conventional first-order jump, lowers Af while leaving Ms largely unchanged, and thereby reduces thermal hysteresis from 275 K (fully crystalline) to 95–110 K. Uniaxial loading–unloading at 500 K shows that elastic modulus and critical stress for stress-induced transformation rise by ~60–90% with increasing amorphous content, accompanied by larger mechanical hysteresis and remnant strain. The authors attribute these effects to heterogeneous nucleation at crystal–amorphous interfaces, retained austenite that seeds reverse transformation, and mechanical constraint by the amorphous phase, and propose CAS as a general microstructural design route for SMAs.","tokens_in":11582,"tokens_out":1439,"duration_ms":13433,"significance":"If the reported trends survive experimental scrutiny, the work offers a practical design strategy that bypasses the usual difficulty of finding thermodynamically compatible coherent second phases: simply amorphizing a fraction of the base alloy itself. The simultaneous reduction of thermal hysteresis and increase of modulus/critical stress would be attractive for applications that require both dimensional stability and high actuation stress. Within the MD literature the study is carefully executed—large cells (~1 M atoms), systematic variation of crystalline fraction, both thermal and mechanical pathways, and clear PTM-based visualization of nucleation and retained austenite. The relative improvements are internally consistent and free of circular fitting. The principal open question is transferability of the crystalline-fitted MEAM potential to crystal–amorphous interfaces, which limits the strength of the design claim until experimental or higher-fidelity validation is provided.","major_comments":[{"comment":"The central mechanistic claims (continuous second-order-like pathway, heterogeneous nucleation, retained austenite, and the quantitative 60–90% property shifts) rest entirely on the 2NN MEAM potential of Ko et al. (Phys. Rev. B 92, 134107, 2015). That potential was developed and validated for crystalline B2 ⇌ B19' energetics, lattice parameters, and transformation temperatures; the manuscript provides no evidence that it reproduces amorphous NiTi structure factors, crystal–amorphous interfacial energies, or the local stress fields invoked in §§3.1–3.2. Because the fully crystalline baseline already over-predicts experimental hysteresis (acknowledged in §3.1), the absolute magnitudes of the reported improvements inherit the same untested interfacial physics. At minimum the authors should (i) report radial distribution functions or structure factors of the amorphous regions against availab","section":null},{"comment":"Section 2.1 and SI Fig. 1: the selective melt–quench protocol (crystalline slabs held at 500 K while adjacent slabs are heated to 5000 K, annealed, and cooled) produces crystalline fractions that deviate substantially from the intended targets (43/53/66% instead of 25/50/75%). The authors correctly note interfacial reconstruction, yet they do not quantify interface width, residual crystallinity inside the “amorphous” slabs, or residual stress after NPT relaxation. Because the continuous martensite rise and the retained-austenite argument are attributed to these interfaces (Fig. 2, Fig. 3), a more rigorous structural characterization of the as-prepared interfaces is needed before the two-stage pathway can be confidently ascribed to crystal–amorphous boundaries rather than to residual crystalline nuclei or quench-induced defects.","section":null},{"comment":"Table 1 and §3.1: thermal hysteresis is defined solely as Af – Ms, with Ms taken as the onset of the first-order jump. For the 43% and 53% Xtal systems the continuous stage already produces a non-negligible martensite fraction before that jump. The paper should either (i) report an additional characteristic temperature for the continuous stage or (ii) demonstrate that the continuous stage is fully reversible and does not contribute to hysteresis. Without this clarification the claimed ~60% hysteresis reduction is only partially characterized.","section":null}],"minor_comments":[{"comment":"Figure 2 caption and Table 1: the caption refers to “Figure 1” for the temperature dependence; the correct reference is Figure 2. Also, the ±5 K uncertainties appear to be visual estimates; a brief statement of how they were obtained would help.","section":null},{"comment":"Section 2.3: the strain rate (5\times10^7 s^–1) and unloading times are stated, but no rate-sensitivity check is provided. A single additional run at a lower rate for one composition would strengthen the claim that the observed critical stresses are not rate artifacts.","section":null},{"comment":"SI Figure 4: the 0.2% offset construction is clear, yet the maximum pre-plateau stresses listed in Table 2 do not always coincide with the visual peaks in the main-text stress–strain curves of Figure 4; a short note reconciling the two would avoid confusion.","section":null},{"comment":"References: the experimental crystalline–amorphous NiTi nanocomposite of Hua et al. (Nature Nanotechnology 2021) is cited, but more recent experimental work on partially amorphized or nanocrystalline NiTi (e.g., shear-induced amorphization studies) could be mentioned to better situate the simulation results.","section":null},{"comment":"Throughout: “second-order-like” is used without a thermodynamic criterion (e.g., continuous free-energy derivative). A sentence clarifying that the term is phenomenological (continuous martensite-fraction rise) would prevent misinterpretation.","section":null}],"recommendation":"major_revision","confidential_remarks":"The work is a solid, well-executed MD study that fits the scope of a materials-science journal. The main risk is over-claiming transferability of a crystalline MEAM potential to crystal–amorphous interfaces; requiring the authors to add structural validation of the amorphous regions and a clearer discussion of limitations would bring the manuscript to an acceptable standard without demanding new experimental data. I see no citation or novelty issues."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a straightforward, well-run MD study that fills a real gap. Hua et al. already showed crystalline–amorphous NiTi can be strong and fatigue-resistant; these authors systematically vary crystalline fraction (43–100 %) and track both thermal and stress-induced martensite. The new pieces are the continuous-then-first-order thermal pathway in the partially amorphized cells, the drop in thermal hysteresis from 275 K to ~100 K, and the 60–90 % rises in modulus and critical stress. Those trends are internally consistent across the martensite-fraction curves, volume jumps, stress–strain loops, and PTM snapshots.\n\nThey do the mechanics cleanly: heterogeneous nucleation at the interfaces, retained austenite that seeds reverse transformation, and mechanical constraint from the amorphous layers. The fully crystalline baseline over-predicts experimental hysteresis (they say so), but the relative shifts with amorphous content are the useful result. No free parameters were tuned to hit a target hysteresis or modulus; the Ko MEAM potential is taken as-is.\n\nThe soft spot is exactly the one the stress-test flags: that potential was built for crystalline B2–B19' energetics. Transferability to amorphous structure, crystal–amorphous interfacial energy, and the local stress fields that supposedly drive the continuous stage is untested. If the interface is artificially soft or over-stabilized, both the two-stage pathway and the size of the property gains could be artifacts. High MD rates and defect-free single crystals also push absolute temperatures and stresses away from experiment. No code or trajectories are released, so independent checks are limited.\n\nStill, the comparative design rule is clear and the data support it inside the simulation framework. This is for people who work on SMA microstructural design or who run MD of martensite; they will get a usable set of trends and a concrete construction protocol. It deserves a serious referee, not a desk reject. I would engage with it, cite the relative trends with the usual MD caveats, and watch for experimental follow-up on the continuous stage.","headline":"Clean MD campaign that maps how partial amorphization changes NiTi martensite pathways and properties; the relative trends look solid inside the model, absolute numbers and interface physics less so.","tokens_in":12192,"tokens_out":525,"would_cite":true,"duration_ms":5882,"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":"Partial amorphization of NiTi creates crystal-amorphous superlattices that cut thermal hysteresis from 275 K to 95-110 K and raise elastic modulus and transformation stress by 60-90 percent.","keywords":["Shape Memory Alloys","Crystal-Amorphous Superlattices","Martensitic Transformation","Superelasticity","Molecular Dynamics Simulations","NiTi","Thermal Hysteresis"],"falsifier":"Prepare experimental NiTi specimens with comparable crystalline fractions and crystal-amorphous interfaces; if slow thermal cycling shows neither a continuous first stage nor a thermal-hysteresis drop to roughly 100 K, the predicted pathway and property gains are falsified.","tokens_in":12140,"feed_emoji":"⚙️","tokens_out":835,"duration_ms":19418,"temperature":0.7,"pith_summary":"This paper shows that turning part of a NiTi shape-memory alloy into an amorphous phase, forming crystal-amorphous superlattices, changes how the alloy transforms under temperature or stress. Molecular-dynamics simulations of structures with 43-66 percent crystalline fraction reveal an initial continuous, second-order-like martensitic stage at the interfaces before the usual abrupt first-order jump. The amorphous layers also leave some austenite untransformed, which seeds the reverse path and shrinks thermal hysteresis by roughly 60 percent, while mechanically constraining the crystals so that both stiffness and the critical stress for superelasticity rise 60-90 percent. A sympathetic reader cares because the second phase is made from the same material, removing the usual need to find a coherent foreign precipitate, and because the same microstructural lever can be used to tune reversibility, modulus and load capacity in other shape-memory systems.","feed_headline":"Amorphizing NiTi cuts thermal hysteresis nearly in half","feed_subtitle":"Crystal-amorphous layers also raise stiffness and transformation stress by 60-90 percent.","key_machinery":"Crystal-amorphous superlattices (CAS): layered NiTi structures whose crystal-amorphous interfaces supply heterogeneous nucleation sites, retain austenite nuclei, and impose mechanical constraint on the crystalline fraction.","core_discovery":"Partial amorphization of NiTi produces crystal-amorphous superlattices in which crystal-amorphous interfaces nucleate martensite continuously, retained austenite facilitates the reverse transformation, and the amorphous phase mechanically constrains the crystalline regions, thereby reducing thermal hysteresis from 275 K to 95-110 K and increasing both elastic modulus and critical stress for stress-induced transformation by approximately 60-90 percent relative to fully crystalline NiTi.","pith_inferences":["If experimental CAS-NiTi can be made by controlled amorphization routes (irradiation, severe deformation, or multilayer deposition) with similar interface density, the two-stage pathway and hysteresis reduction should appear under laboratory heating rates.","The continuous first stage may allow small-amplitude, low-hysteresis actuation useful for precision thermal sensors or micro-actuators.","Because the amorphous phase has the same composition as the crystal, CAS structures avoid the interfacial chemistry and coherency-strain problems that limit multiphase precipitate designs."],"forward_implications":["SMA designers can adjust hysteresis, modulus and critical stress by choosing crystalline fraction instead of alloying additions.","Lower thermal hysteresis reduces energy dissipation and improves cyclic fatigue life in actuators.","Higher transformation stress widens the load range of superelastic components.","The same partial-amorphization route can be applied to other shape-memory alloys that lack suitable coherent precipitates."],"fun_headline_variants":["Partial amorphization cuts NiTi hysteresis from 275 K to 95-110 K","Crystal-amorphous NiTi superlattices raise modulus and stress 60-90%","CAS-NiTi lowers thermal hysteresis via crystal-amorphous interfaces","Amorphizing NiTi layers enhance reversibility and boost stiffness","NiTi superlattices cut hysteresis while raising critical stress 60-90%"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The interatomic potential trained mainly on crystalline NiTi remains accurate for the stresses, nucleation, and continuous transformation pathway that appear at crystal-amorphous interfaces.","fun_headline_variants_meta":{"raw":{"variants":["Partial amorphization cuts NiTi hysteresis from 275 K to 95-110 K","Crystal-amorphous NiTi superlattices raise modulus and stress 60-90%","CAS-NiTi lowers thermal hysteresis via crystal-amorphous interfaces","Amorphizing NiTi layers enhance reversibility and boost stiffness","NiTi superlattices cut hysteresis while raising critical stress 60-90%"]},"model":"grok-4.5","effort":"low","cost_usd":0.009146,"raw_usage":{"total_tokens":2109,"prompt_tokens":829,"num_sources_used":0,"completion_tokens":86,"cost_in_usd_ticks":91460000,"prompt_tokens_details":{"text_tokens":829,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1194,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":829,"tokens_out":86,"duration_ms":9432,"temperature":1.0,"reasoning_tokens":1194,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-12T04:21:20.342575+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Prepare experimental NiTi specimens with comparable crystalline fractions and crystal-amorphous interfaces; if slow thermal cycling shows neither a continuous first stage nor a thermal-hysteresis drop to roughly 100 K, the predicted pathway and property gains are falsified.","supporting_citations":[],"review_version":1}