Martensitic Transformation in Crystal-Amorphous Superlattices of NiTi Shape Memory Alloy

Bhavna Singh; Shivam Tripathi

REVIEW 3 major objections 5 minor 9 references

Martensitic Transformation in Crystal-Amorphous Superlattices of NiTi Shape Memory Alloy

T0 review · 3 major / 5 minor · reviewed 2026-07-12 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read 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.

desk verdict 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. read the letter →

arxiv 2607.03172 v1 pith:ATFEJNTK submitted 2026-07-03 cond-mat.mtrl-sci

Bhavna Singh , Shivam Tripathi This is my paper · ORCID

classification cond-mat.mtrl-sci

keywords ShapeMemoryAlloysCrystal-AmorphousSuperlatticesMartensiticTransformationSuperelasticityMolecularDynamicsSimulationsNiTiThermalHysteresis

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

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.

What carries the argument

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.

What would settle it

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.

Watch

Extended reading notes

Core claim

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.

Load-bearing premise

The interatomic potential trained mainly on crystalline NiTi remains accurate for the stresses, nucleation, and continuous transformation pathway that appear at crystal-amorphous interfaces.

Editorial extensions

If this is right

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.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

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.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the 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.

They 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.

The 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.

Still, 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.

Referee Report

3 major / 5 minor

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.

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 (3)

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 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.
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.

minor comments (5)

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 2.3: the strain rate (5 imes10^7 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.
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.
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.
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.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: MD outputs (Ms/Af, moduli, critical stresses) are measured from trajectories under a fixed external potential, not fitted or definitionally forced.

full rationale

The paper is a forward molecular-dynamics campaign. CAS structures are built by a prescribed thermal protocol (Section 2.1), then cooled/heated or loaded/unloaded; martensite fractions, cell volumes, stress–strain curves, elastic moduli, and critical stresses are extracted as observables (PTM, 0.2 % offset, linear fit 0–0.03 strain). The 2NN MEAM potential is taken unchanged from Ko et al. (Phys. Rev. B 2015) and is not re-parameterized to any target hysteresis or modulus reported here. No equation equates a claimed prediction to a quantity defined by the same fit; no uniqueness theorem or ansatz is imported from the authors’ prior work to force the two-stage pathway or the 60–90 % property shifts. Self-citations (e.g., Tripathi et al. on Mg-Sc or precipitates) appear only as background motivation and are not load-bearing for the numerical results. The acknowledged over-prediction of crystalline hysteresis relative to experiment is an external-validation issue, not circularity. Consequently the derivation chain is self-contained against its own inputs and scores 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claims rest on standard MD machinery plus one domain-critical potential and several simulation-protocol choices. No new physical entities are postulated. Free parameters are the usual MD rates and the post-reconstruction crystalline fractions; they affect absolute numbers but the paper’s comparative claims are made across a fixed protocol.

free parameters (4)

cooling/heating rate
Fixed at 5 K/ps for all thermal cycles; orders of magnitude above experiment and known to shift Ms/Af. Comparative trends assume the shift is similar across crystalline fractions.
strain rate
Fixed at 5×10^7 s^−1 for uniaxial loading; affects critical stress and hysteresis magnitude.
amorphization anneal temperature and duration
5000 K for 1000 ps chosen by hand to melt designated regions; controls interface roughness and final crystalline fraction after reconstruction.
actual crystalline fractions after reconstruction
Intended 25/50/75% became ~43/53/66% via diffusion at interfaces; these realized fractions are the independent variable for all reported trends.

assumptions (4)

domain assumption The 2NN MEAM Ni–Ti potential of Ko, Grabowski & Neugebauer (2015) correctly describes both bulk martensitic transformation and crystal–amorphous interface energetics/stresses.
Invoked for every trajectory (§2); potential was fitted and validated mainly on crystalline phases and bulk MT temperatures.
domain assumption Polyhedral Template Matching (PTM) with the chosen RMSD cutoff reliably partitions atoms into austenite (BCC), martensite (HCP/FCC), and amorphous (unidentified).
Used to define martensite fraction, Ms, Af, and all snapshots (Figs. 1–3, 6); cutoff not reported.
domain assumption Periodic boundary conditions and a single-crystal, defect-free starting cell are representative enough for relative trends in hysteresis and modulus versus amorphous fraction.
Stated in §2; authors note that absolute hysteresis exceeds polycrystal experiments for this reason.
ad hoc to paper The melt–quench protocol that holds crystalline slabs at 500 K while melting adjacent slabs produces interfaces that are physically analogous to experimental crystalline–amorphous nanocomposites.
Construction method in §2.1; no experimental interface characterization is compared.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Martensitic Transformation in Crystal-Amorphous Superlattices of NiTi Shape Memory Alloy." pith.science (2026). https://pith.science/paper/ATFEJNTK

@misc{pith2026260703172,
  author       = {Pith},
  title        = {Pith review of: Martensitic Transformation in Crystal-Amorphous Superlattices of NiTi Shape Memory Alloy},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ATFEJNTK}},
  note         = {Machine review of arXiv:2607.03172}
}

read the original abstract

Shape memory alloys (SMAs) exhibit unique thermo-mechanical properties arising from reversible martensitic transformation. Tailoring these properties through coherent second phases is effective but is often limited by the difficulty of identifying thermodynamically compatible phases. Crystal-amorphous superlattices (CAS), in which the second phase is derived from the base material itself, provide an attractive alternative. However, the influence of partial amorphization on martensitic transformation and thermo-mechanical behavior remains largely unexplored. Here, large-scale molecular dynamics simulations are used to investigate temperature- and stress-induced martensitic transformations in CAS-NiTi with different crystalline phase fractions. Partial amorphization fundamentally modifies the transformation pathway, introducing an initial continuous (second-order-like) transformation before the conventional first-order martensitic transformation. The amorphous phase also enhances transformation reversibility, reducing the thermal hysteresis from 275 K in fully crystalline NiTi to 95-110 K in CAS-NiTi. Simultaneously, the elastic modulus and the critical stress for stress-induced martensitic transformation increase by approximately 60-90 percent, depending on the crystalline fraction. These improvements originate from heterogeneous nucleation at crystal-amorphous interfaces, retained austenite that promotes the reverse transformation, and mechanical constraint imposed by the amorphous phase. These findings establish crystal-amorphous superlattices as a promising microstructural design strategy for tailoring the thermo-mechanical performance of shape memory alloys.

Figures

Figures reproduced from arXiv: 2607.03172 by the authors.

**Figure 1.** Crystal–amorphous superlattice (CAS-NiTi) structures after amorphization and subsequent NPT relaxation at 500 K for (a) 66% Xtal, (b) 53% Xtal, and (c) 43% Xtal systems. Atoms are color-coded according to the Polyhedral Template Matching (PTM) analysis: austenite (blue), amorphous (grey), and martensite (red). 2.2. Temperature-induced martensitic transformation To investigate the temperature-induced martensitic tran… view at source ↗

**Figure 2.** Martensite fraction as a function of temperature during heating and cooling simulations for various CAS-NiTi. Green, blue, red, and black colors are used for 43%, 53%, 66%, and 100% Xtal CAS-NiTi. The introduction of an amorphous phase has little influence on Ms but significantly lowers Af, indicating enhanced stabilization of the austenite phase. As a result, the thermal hysteresis (Af - Ms) is substantially reduce… view at source ↗

**Figure 3.** ), In contrast, the CAS-NiTi systems develop multiple martensite variants (see [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Stress-strain response during stress-induced martensitic transformation in 100% Xtal, 66% Xtal, 53% Xtal, and 43% Xtal CAS-NiTi systems at 500 K. Uniaxial loading was applied along (a) the [001] direction, perpendicular to the crystal-amorphous interface, and (b) the […

**Figure 5.** Figure 5: Comparison of (a) mechanical hysteresis (area under stress-strain curve) and (b) remnant strain for the investigated CAS-NiTi systems following unloading from maximum strains of 12% and 8% under uniaxial tensile loading along the [001] direction at 500 K The incorporat…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

9 extracted references

[1]

This transformation underpins several interesting thermo-mechanical characteristics including shape memory behavior, superelasticity, and pseudo-plasticity

Introduction SMAs belong to a special class of materials that undergo solid-solid reversible martensitic transformation between a high-temperature, high-symmetry austenite phase and a low-temperature, low-symmetry martensite phase. This transformation underpins several interesting thermo-mechanical characteristics including shape memory behavior, superela...
[2]

Atomic structures were visualized using the Open Visualization Tool (OVITO) [31]

Simulation Details MD simulations were performed using the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) [30]. Atomic structures were visualized using the Open Visualization Tool (OVITO) [31]. We used the Parrinello-Rahman barostat [32] and the Nose-Hoover thermostat [33-34], with a time step of 1 fs for all of our simulations. Period...
[3]

and [010] directions are crystallographically equivalent and both are parallel to the interface, only the [100] direction was considered. Starting from an NPT-relaxed CAS-NiTi structures, we performed loading simulations by elongating the simulation cell along the tensile axis [100] and [001] up to 12% strain, using a strain rate of 5x107 s-1. The lateral...
[4]

Temperature-induced martensitic transformation We performed heating and cooling simulations to study thermally induced martensitic transformation

Results and Discussion 3.1. Temperature-induced martensitic transformation We performed heating and cooling simulations to study thermally induced martensitic transformation. Figure 2 shows the evolution of the martensitic phase fraction as a function of temperature during these simulations. The NPT-relaxed structures at 500 K are predominantly in the aus...
[5]

Self-accommodation in martensite

Conclusions In this study, we investigated the temperature and stress-induced martensitic transformation in various crystal-amorphous superlattices. Our simulations reveal that the presence of amorphous phase fundamentally modifies the transformation pathway by introducing a two-stage transformation consisting of an initial continuous (second-order-like) ...

2024
[6]

A review of shape memory alloys in MEMS devices and biomedical applications

Chaudhary, Kedarnath, Vikrant K. Haribhakta, and Pradeep V . Jadhav. "A review of shape memory alloys in MEMS devices and biomedical applications." Materials Today: Proceedings (2024). 14. Lendlein, Andreas, and Robert Langer. "Biodegradable, elastic shape-memory polymers for potential biomedical applications." Science 296.5573 (2002): 1673-1676. 15. El F...

2024
[7]

Uncovering the role of nanoscale precipitates on martensitic transformation and superelasticity

Tripathi, Shivam, et al. "Uncovering the role of nanoscale precipitates on martensitic transformation and superelasticity." Acta Materialia 229 (2022): 117790. 27. Chen, Haiyang, et al. "Unprecedented non-hysteretic superelasticity of [001]-oriented NiCoFeGa single crystals." Nature materials 19.7 (2020): 712-7 28. Hua, Peng, et al. "Nanocomposite NiTi sh...

2022
[8]

Study on grain size dependence of shape memory effect in nanocrystalline NiTi shape memory alloys with grain size below 20 nm based on molecular dynamics simulation

Zhang, Yanqiu, et al. "Study on grain size dependence of shape memory effect in nanocrystalline NiTi shape memory alloys with grain size below 20 nm based on molecular dynamics simulation." Materials Today Nano 30 (2025): 100610. 40. Li, Zhu, et al. "Atomic scale modeling of the coherent strain field surrounding Ni4Ti3 precipitate and its effects on therm...

2025

Show all 9 references

[9]

Stress-induced martensitic transformation SI Figure 4. Representative loading stress–strain curves of the investigated CAS-NiTi systems showing the procedure used to determine the elastic modulus, the 0.2% offset critical stress for stress-induced martensitic transformation, a...

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.

[1] [1]

This transformation underpins several interesting thermo-mechanical characteristics including shape memory behavior, superelasticity, and pseudo-plasticity

Introduction SMAs belong to a special class of materials that undergo solid-solid reversible martensitic transformation between a high-temperature, high-symmetry austenite phase and a low-temperature, low-symmetry martensite phase. This transformation underpins several interesting thermo-mechanical characteristics including shape memory behavior, superela...

[2] [2]

Atomic structures were visualized using the Open Visualization Tool (OVITO) [31]

Simulation Details MD simulations were performed using the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) [30]. Atomic structures were visualized using the Open Visualization Tool (OVITO) [31]. We used the Parrinello-Rahman barostat [32] and the Nose-Hoover thermostat [33-34], with a time step of 1 fs for all of our simulations. Period...

[3] [3]

and [010] directions are crystallographically equivalent and both are parallel to the interface, only the [100] direction was considered. Starting from an NPT-relaxed CAS-NiTi structures, we performed loading simulations by elongating the simulation cell along the tensile axis [100] and [001] up to 12% strain, using a strain rate of 5x107 s-1. The lateral...

[4] [4]

Temperature-induced martensitic transformation We performed heating and cooling simulations to study thermally induced martensitic transformation

Results and Discussion 3.1. Temperature-induced martensitic transformation We performed heating and cooling simulations to study thermally induced martensitic transformation. Figure 2 shows the evolution of the martensitic phase fraction as a function of temperature during these simulations. The NPT-relaxed structures at 500 K are predominantly in the aus...

[5] [5]

Self-accommodation in martensite

Conclusions In this study, we investigated the temperature and stress-induced martensitic transformation in various crystal-amorphous superlattices. Our simulations reveal that the presence of amorphous phase fundamentally modifies the transformation pathway by introducing a two-stage transformation consisting of an initial continuous (second-order-like) ...

2024

[6] [6]

A review of shape memory alloys in MEMS devices and biomedical applications

Chaudhary, Kedarnath, Vikrant K. Haribhakta, and Pradeep V . Jadhav. "A review of shape memory alloys in MEMS devices and biomedical applications." Materials Today: Proceedings (2024). 14. Lendlein, Andreas, and Robert Langer. "Biodegradable, elastic shape-memory polymers for potential biomedical applications." Science 296.5573 (2002): 1673-1676. 15. El F...

2024

[7] [7]

Uncovering the role of nanoscale precipitates on martensitic transformation and superelasticity

Tripathi, Shivam, et al. "Uncovering the role of nanoscale precipitates on martensitic transformation and superelasticity." Acta Materialia 229 (2022): 117790. 27. Chen, Haiyang, et al. "Unprecedented non-hysteretic superelasticity of [001]-oriented NiCoFeGa single crystals." Nature materials 19.7 (2020): 712-7 28. Hua, Peng, et al. "Nanocomposite NiTi sh...

2022

[8] [8]

Study on grain size dependence of shape memory effect in nanocrystalline NiTi shape memory alloys with grain size below 20 nm based on molecular dynamics simulation

Zhang, Yanqiu, et al. "Study on grain size dependence of shape memory effect in nanocrystalline NiTi shape memory alloys with grain size below 20 nm based on molecular dynamics simulation." Materials Today Nano 30 (2025): 100610. 40. Li, Zhu, et al. "Atomic scale modeling of the coherent strain field surrounding Ni4Ti3 precipitate and its effects on therm...

2025