Accurate segmentation of microstructural images plays a critical role in quantitative materials analysis. However, it remains a challenging task due to the scarcity of annotated data and the inherent diversity of microstructures. While transfer learning is a promising solution to data scarcity, conventional approaches using models pre-trained on general-purpose datasets like ImageNet are often suboptimal due to their low similarity to microstructural images. To overcome this limitation, we propose a novel transfer learning approach utilizing synthetically generated Voronoi diagrams as a source dataset for equiaxed grain microstructure analysis. Unlike images produced by simulations or deep learning models, Voronoi diagrams can be created quickly with simple and intuitive parameters, while still resembling key structural features of equiaxed grain microstructures. This simplicity allows for efficient exploration of how generation conditions affect downstream model performance. Our results demonstrate that adding Gaussian noise to the Voronoi diagrams improves model performance, achieving a higher Intersection over Union (IoU) score than models pre-trained on ImageNet. Furthermore, the model learns microstructure-specific features such as triple junctions, enabling accurate grain boundary extraction even in low-contrast regions. This approach offers a simple and practical solution for pre-training task-specific models in materials science.

Au-Cu-Al ternary alloys are promising candidates for biomedical shape memory and superelastic applications; however, their composition-dependent behavior in bulk form has not been fully elucidated. In this study, bulk Au-Cu-Al alloys were fabricated over a wide compositional range within the single beta-phase region, and their martensitic transformation behavior, phase constitution, and mechanical properties were systematically investigated. Martensitic transformation was observed for all alloys. The martensitic transformation start temperature (Ms) decreased primarily with increasing Al content and valence electron concentration (e/a). Even at identical e/a values, both Ms and phase constitution exhibited a strong dependence on the Cu/Al ratio, with increasing Cu/ Al ratio leading to a systematic transition from the L21 parent phase to orthorhombic and subsequently monoclinic martensite. Mechanical properties were also strongly influenced by the Cu/Al ratio, and alloys in which monoclinic martensite was stabilized exhibited both higher strength and improved ductility. Furthermore, depending on composition, both superelasticity and twinning pseudoelasticity were observed; however, a clear trade-off between ductility and superelasticity at room-temperature was identified. These results demonstrate that the Cu/Al ratio is a key compositional parameter governing martensitic transformation behavior and mechanical properties, and provide fundamental guidelines for designing Au-based shape memory and superelastic alloys.

In this study, laser cavitation peening without coating (LCPwC) was employed to enhance the wear resistance of Ti-30Zr-5Mo alloy for implant applications by inducing a multiscale gradient surface structure through laser-water interaction and surface oxidation. LCPwC was performed at pulse densities of 10, 50, and 100 pulses/ mm2, producing an oxide layer comprising TiO2, Ti2O3, and ZrO2. With increasing pulse density, the TiO2 fraction decreased from 76.8% to 56.9%, whereas the Ti2O3 and ZrO2 fractions increased from 6.7% to 20.9% and from 5.1% to 13.8%, respectively. The oxide layer exhibited a nanograined structure, and the average grain size decreased from 12.1 to 9.1 nm before increasing slightly to 9.5 nm, suggesting initial grain refinement followed by limited grain growth at higher pulse densities. The surface roughness increased from 0.4 mu m for the untreated alloy to 1.5-4.0 mu m after treatment, while the water contact angle increased from 64.9 degrees to 121.2-130.1 degrees, indicating increased surface hydrophobicity. The surface microhardness increased from 290 HV for the untreated alloy to 360, 392, and 403 HV after treatment, and the residual compressive stress increased in magnitude from -74 to -179 MPa. Wear testing showed that the specimen treated at 10 pulses/mm2 exhibited improved tribological performance, with a 26.2% reduction in wear volume, a 50% reduction in wear mass, and a decrease in the coefficient of friction from 0.56 to 0.12. These improvements were mainly attributed to the formation of a protective oxide layer and the increase in surface hardness induced by LCPwC. Overall, LCPwC is an effective strategy for improving the tribological performance of Ti-30Zr-5Mo alloy and represents a promising approach for implant surface modification.

The convergence of pyroelectric (PE) and magnetoelectric (ME) technologies presents a transformative opportunity for advancing key areas of our increasingly digitized society, ranging from next-generation sensors and spintronic voltage control to highly efficient energy harvesting systems. In this study, we explore the synergistic effects of combining P(VDF-TrFE) piezopolymer with Ni–Mn–Ga magnetic shape memory alloy fillers. Our experimental results reveal a 2.5-fold increase in the PE coefficient due to the cubic-tetragonal phase transition of Ni–Mn–Ga particles, dramatically enhancing the intrinsic properties of the piezopolymer. Theoretical modeling confirms that this increase is driven by the mechanical stress imparted by the phase transition within the composite material. Moreover, the ME response under cyclic temperature variation (from 298 to 328 K) demonstrates an unprecedented leap, with the ME coefficient surging from 4.3 V cm‒1Oe‒1 to 20 V cm‒1Oe‒1 during narrow 2 K intervals, attributed to phase transitions in the Ni–Mn–Ga filler. Remarkably, even the lowest observed ME coefficient exceeds by two orders of magnitude the maximum reported for similar P(VDF-TrFE)-based composites, signaling a breakthrough in polymer-based ME materials. These findings open new horizons for future technological innovation, where the manipulation of structural phase transitions in composite materials can unlock extraordinary advancements in multifunctional devices. The significant leap in PE and ME performance underscores the potential for disruptive applications in energy, sensing, and spintronics.

Ni-Mn-Ga particle/silicone polymer composites, which exhibit a large magnetic field-induced strain (MFIS), along with excellent cyclic and long-term stability, hold great promise for use in applications in actuation and sensing applications. In this study, we introduce an innovative approach using soft magnetic, high-permeability, high-saturation magnetization FeNi particles to investigate their influence on the MFIS effect in Ni-Mn-Ga/silicone composites. We examined the magnetization and magnetostrain behaviors of the particle grids in various composites, which contained varying fillers and were fabricated through different processes, using a standard vibrating sample magnetometer and an X-ray micro-CT 3D imaging technique, respectively. Compared to doped-free composite, the FeNi-doped composite exhibited a much larger slope on the dependence of the “magnetostrain versus the magnetic field”, whereby showing a tendency to the enhanced value of magnetostrain at lower magnetic field, feature that is advantageous for applications.

In additive manufacturing (AM), inhomogeneous precipitates from variable thermal cycles degrade mechanical properties and fatigue life. We propose a novel large-small melt pool coupling strategy to fabricate NiTi shape memory alloy (SMA) with an architectured microstructure containing homogeneous coherent nanoprecipitates. This approach actively designs the spatiotemporal distribution of thermal cycles, where large melt pools ensure material forming while secondary low-energy lasers create smaller melt pools to enable precise in-situ heat treatment. Consequently, Ti4Ni2Ox nanoprecipitates transform from an initial intergranular network into a homogeneous distribution. The resulting NiTi SMA exhibits significantly enhanced low-cycle superelastic fatigue life, outperforming all reported AMed NiTi. This superior performance stems from synergistic effects of the architectured microstructure: during loading transfer between two characteristic zones governed by strain compatibility, the homogeneous nanoprecipitation strengthening in the in-situ heat-treated zones enhances superelastic recovery and delays microcracking, while cooperative deformation in the remelted zones minimizes damage accumulation. This methodology transforms the remelting process into a precise microstructural design tool, enabling the fabrication of high-performance complex engineering components.

Metastable beta-Ti shape memory alloys exhibit room-temperature superelasticity, yet recoverable strain is often limited by multi-variant martensitic transformation and plasticity. Here, the deformation behavior of a near < 001 >(beta) Ti-2.5Cr-8.5Sn (at%) single crystal under cyclic compression was clarified by in-situ optical microscopy. Direct correlation of microstructural evolution with the stress-strain response reveals two-stage yielding associated with stress-induced beta ->alpha '' transformation and subsequent dislocation slip in fully transformed martensite. Transformation is initially dominated by a preferred variant, whereas further loading promotes multi-variant interactions and geometric constraints, leading to an increased transformation plateau stress and residual strain accumulation. The microstructure progressively evolves into a single-variant martensitic state, followed by plastic yielding at similar to 897 MPa. Quantitative analysis identifies similar to 5.5% transformation strain, similar to 1.6% beta-phase elasticity, and similar to 2.4% martensite elasticity. Strain recovery upon unloading is primarily governed by reverse variant reorientation rather than reverse phase transformation.

This study investigates the fabrication of ultra-high Sn-containing metastable beta-type Ti-Cr-Sn alloys using Laser Powder Bed Fusion (L-PBF). By utilizing the rapid solidification inherent to L-PBF, retention of an ultra-high Sn metastable beta phase at room temperature was achieved in compositions that are difficult to obtain via conventional processing routes. To address a key challenge in L-PBF, the high material cost associated with pre-alloyed powders, a mixture of pure elemental powders was employed as a cost-effective alternative feedstock. The effects of laser power on the microstructure, phase constitution, crystallographic texture, and mechanical properties of the as-built specimens were systematically evaluated. The results demonstrate that L-PBF successfully produced a homogeneous ultra-high Sn beta solid solution while effectively suppressing the formation of omega phase and Ti3Sn compounds. Under the highest laser power condition of 360 W, complete melting and the formation of a single beta phase were achieved. The conventional unidirectional laser scanning strategy (X-scan), in which the scanning direction is kept constant between layers, resulted in <110>(beta) alignment along the build direction and <100>(beta) alignment along the scan direction. The specimen fabricated at 360 W exhibited the most homogeneous microstructure, together with appropriate strength, excellent ductility, and low Young's modulus. These findings demonstrate the feasibility of low-cost processing via L-PBF while achieving enhanced functional performance in high-Sn metastable beta-Ti alloys.

The phenomenological theory of martensitic crystallography (PTMC) successfully describes the crystallographic features of martensitic transformations in various metallic systems, including Ti-Ni and Au-Cd alloys. In this study, we investigate the stress-induced martensitic transformation (SIMT) in Ti-6Mo-10Al single crystals, with emphasis on the applicability of PTMC. Assuming conventional lattice correspondence and lattice parameters established for thermally induced martensite, we evaluate the PTMC using precise experiments. To ensure high experimental accuracy, we analyze stress-induced transformations in well-oriented single crystals using in-situ optical microscopy, SEM-ECC imaging, and SEM-EBSD under controlled compression conditions. Experimental analyses reveal habit planes and orientation relationships that deviate significantly from PTMC predictions, with discrepancies exceeding 10 degrees and 2 degrees, respectively. Moreover, no evidence of lattice-invariant deformation such as internal twinning is observed when using any of the aforementioned techniques. The transformation proceeds via the growth of a single martensitic variant along habit planes, without satisfying the invariant-plane conditions assumed by the PTMC. Eigenvalue analysis of the experimentally determined total shape deformation confirms the absence of invariant strain directions. These findings show that the PTMC, when parameterized by thermalmartensite inputs, fails to describe the SIMT behavior in Ti-6Mo-10Al alloy, despite its broad success in other systems. This study presents a rare and significant case of PTMC breakdown in an SIMT and suggests the necessity for new theoretical frameworks to explain such crystallographic anomalies.

Single-crystalline (SC) Ni-Mn-Ga ferromagnetic shape memory alloys (FSMAs) exhibit giant magnetic-field-induced strain (MFIS), making them attractive for solid-state actuators and energy conversion systems. However, their practical application is limited by brittleness and fabrication costs. Some composites have been investigated to solve these issues. In this study, a finite element method (FEM) model was developed to investigate the magnetomechanical behavior of sandwich-structured composites consisting of SC Ni-Mn-Ga particles embedded between silicone rubber layers and metallic plates. The influences of silicone rubber thickness, particle number, interparticle distance, and plate material on deformation behavior, stress-strain distribution, particle rotation, and magnetic response were systematically analyzed. The simulations reveal that increasing silicone rubber thickness suppresses the overall stroke due to enhanced elastic constraints, while particle interactions become increasingly significant in multiparticle configurations, leading to particle rotation and localized stress and strain concentrations. Furthermore, replacing diamagnetic Cu plates with ferromagnetic Fe plates generates repulsive magnetic forces that enhance composite deformation and improve magnetic actuation efficiency. The simulated deformation behavior shows good agreement with experimental observations, thereby validating the proposed model and providing valuable design guidelines for high-performance Ni-Mn-Ga composite actuators and solid-state energy conversion devices.

Precious-metal-based shape memory alloys (SMAs) such as Au-Cu-Al alloys exhibit unique functional properties, including Ni-free composition, high radiopacity, and tunable martensitic transformation temperatures, making them attractive for biomedical applications. However, their limited workability and high material cost pose significant challenges for conventional manufacturing and powder-based additive manufacturing routes. In this study, investment casting combined with LCD-based 3D-printed castable resin patterns is investigated as a near-net-shape fabrication route for Au-Cu-Al SMAs. Three alloy compositions, 50Au-25Cu-25Al, 50Au-28Cu-22Al, and 53Au-27Cu-18Al-2Fe, were fabricated and systematically evaluated in terms of dimensional accuracy, microstructure, chemical homogeneity, phase constitution, martensitic transformation behavior, and mechanical properties. The results demonstrate that thin-section cast components with complex geometries, including dog-bone tensile specimens and a stent prototype, can be successfully produced with dimensional accuracy on the order of ~300 µm and minimal compositional variation (<1 mol.%). X-ray diffraction and differential scanning calorimetry confirm composition-dependent phase constitutions and martensitic transformation temperatures near body temperature for selected alloys. Porosity and shrinkage defects are identified as the primary factors governing the reduced ductility of as-cast samples, while mold preheating effectively reduces defect density. Furthermore, Fe addition promotes a mixed ductile/intergranular fracture mode and significantly improves mechanical performance. These findings establish investment casting as a viable near-net-shape processing route for precious-metal-based functional materials and provide a practical pathway toward the manufacturing of Au-Cu-Al SMAs for biomedical applications.

Au-Cu-Al-based shape memory alloys (SMAs) are promising Ni-free biomaterial candidates because of their high radiopacity and tunable martensitic transformation temperatures. However, their limited workability, grain-boundary brittleness, and high material cost make the fabrication and mechanical evaluation of small biomedical-relevant components challenging. In this study, three Au-Cu-Al-based SMAs, 50Au-25Cu-25Al, 50Au-28Cu-22Al, and 53Au-27Cu-18Al-2Fe, were fabricated by investment casting using LCD-based 3D-printed castable resin patterns, and their microstructure, phase constitution, martensitic transformation behavior, casting defects, and tensile fracture behavior were systematically evaluated. Thin-section specimens with complex geometries, including dog-bone tensile specimens and a stent-like prototype, were produced with dimensional deviations on the order of approximately 300 µm and compositional variations below 1 mol.%. X-ray diffraction and differential scanning calorimetry revealed composition-dependent phase constitutions and martensitic transformation behavior, with selected alloys exhibiting transformation temperatures near body temperature. Tensile testing and fracture analysis showed that the reduced ductility of the as-cast alloys was mainly governed by porosity and shrinkage defects, which acted as preferential sites for damage initiation and premature fracture. Mold preheating reduced casting defect formation, whereas Fe addition changed the fracture behavior from predominantly intergranular fracture to a mixed ductile/intergranular fracture mode and improved tensile performance. These findings demonstrate that defect control is essential for improving the tensile reliability of investment-cast Au–Cu–Al-based SMAs and provide a processing and mechanical basis for future biomedical and device-level evaluations of Ni-free Au-based functional alloy components.

Achieving both deformability and high efficiency in thermoelectric materials (TEs) remains challenging, as most high-performance TEs are inherently brittle and rely on toxic tellurium. We demonstrate that off-stoichiometry in silver chalcogenides intrinsically tailors both transport and mechanical properties, enabling enhanced and bendable TEs without extrinsic doping. Adjusting the Ag-Se stoichiometry to yield Ag2Se1.04 (SeAg2Se) and Ag2.02Se (AgAg2Se) reveals a strong correlation among defect chemistry, structural stability, and mechanical adaptability. Notably, SeAg2Se sustains a compressive strain of up to 15% before fracture, underscoring its exceptional bendability. This ductile behavior originates from nanoscale Se inclusions that serve as internal stress relievers and phonon scatterers, leading to an ultralow lattice thermal conductivity of 0.3 W m-1 K-1 (480 K) and an enhanced zT of 0.7, indicative of a soft yet efficient thermoelectric material. The SeAg2Se single-leg outperforms its AgAg2Se counterpart, underscoring its potential as a tellurium-free thermoelectric generator (TEG). In addition, SeAg2Se demonstrates excellent cooling capability, achieving a maximum temperature difference of 49.1 K, comparable to that of commercial Bi2Te3. Intrinsic stoichiometric control provides a sustainable design strategy, where self-doping bridges mechanical toughness and thermoelectric efficiency, paving the way for durable, tellurium-free energy devices.

The effects of second phases on the microstructure, martensitic transformation, and mechanical properties of Au–Cu–Al ternary alloys were investigated across selected compositions. Alloys near the β-phase region formed second phases, identified as the α (fcc) phase or Au4Al, depending on composition. The introduction of these phases generally resulted in grain refinement, likely associated with a pinning effect of second-phase particles on grain boundary migration. The martensitic transformation temperature tended to increase in alloys containing second phases. This behavior is likely associated with compositional changes in the matrix caused by second-phase formation, particularly Cu depletion, with possible additional contributions from changes in Al content and microstructural constraints. Mechanical testing revealed that alloys containing the α phase exhibited improved ultimate tensile strength (UTS) and fracture strain compared with β single-phase alloys. In particular, the α phase formed along grain boundaries is suggested to contribute to the suppression of crack initiation and propagation, thereby enhancing ductility. In contrast, alloys containing Au4Al showed only limited improvement, possibly due to its low plastic deformability. These results indicate that the introduction of a ductile α phase effectively improves mechanical properties via grain refinement and grain boundary modification. However, compositions that promote α-phase formation tend to exhibit martensitic transformation temperatures above body temperature. Therefore, achieving a balance between enhanced ductility and transformation behavior near physiological temperature remains a key challenge in the Au–Cu–Al system.

Metastable beta-titanium alloys exhibit excellent mechanical properties and shape memory behavior, making them promising candidates for high-temperature shape memory alloys. Among the nanoscale precipitates that form during aging, the omega phase is well-known for inducing embrittlement. Recently, other nanoscale phases, such as the isothermal alpha '' phase (alpha iso ''), have been discovered. However, the formation mechanisms and mechanical impacts of these precipitates remain unclear. This study reviews the current understanding of nanoscale precipitates, including omega phase, O ' phase, and alpha iso '', with a focus on their formation mechanisms and microstructural evolution in beta-titanium alloys. Based on this review, we conducted experimental investigations on Ti-(3.5-5.0)Mo-(11,14)Al alloys. The alloys were solution-treated and aged at 573 K for 3.6 ks to study the formation behavior of alpha iso '' and its influence on mechanical properties. X-ray diffraction (XRD) analysis confirmed the formation of alpha iso '' in alloys with high Mo and Al concentrations, characterized by peak broadening. Scanning electron microscopy (SEM) revealed fine needle-like precipitates and surface undulations associated with alpha iso '' formation. Alloys without alpha iso '' exhibited no significant microstructural changes. Tensile testing revealed that the alloys containing alpha iso '' suffered significant ductility loss, confirming, for the first time, that alpha iso '' plays a critical role in embrittlement. This study provides further insights into the role of nanoscale precipitates in beta-titanium alloys, emphasizing the importance of controlling alpha iso '' formation. These findings highlight the need for composition optimization to suppress detrimental precipitates, ensuring the structural integrity of beta-titanium alloys for high-temperature applications.

Composite materials made of Ni–Mn–Ga particles and silicone polymer, which exhibit large magnetic field-induced strain (MFIS), good cyclic and long-term stabilities, are highly promising for use in actuation and sensing applications. Whereas MFIS of these composites and their individual particles has been thoroughly investigated in our previous research, MFIS of the ensemble of particles, which are crystallographically and spatially aligned in composite samples, known as the particle grid, has not yet been explored. Since MFIS of a particle grid determines the magnetic field-induced shape change of the entire composite, it was crucial to determine how the variation in the spatial dimensions of the grid affects the overall shape change of the composite. In this study, we examined the magnetization curves and magnetostrain behavior of particle grids in six composites, with different fillers such as Ni–Mn–Ga and Fe particles, produced through distinct manufacturing schemes. We used a standard vibrating sample magnetometer and an innovative method that utilized X-ray microCT 3D imaging. It was found that applying a magnetic field perpendicularly to the particle chains causes large strains along the main directions of the particle grid, which differ in magnitude from those observed for composites as a whole known from the literature. Both the effect of peculiar MFIS transfer from particles grid to the entire composite and the impact of Fe particles are examined in relation to the spatial distribution of the particles and their interaction with the polymer matrix.

Due to the global aging population, shape memory alloys (SMAs) have become an important area of focus in the biomedical and biomaterials fields, particularly for their promising applications in medical devices. While Ni-Ti SMAs have been widely used in biomedical applications due to their excellent shape deformation and recovery strains, concerns about their hypersensitivity remain. As a result, this study shifts its focus from Ni-Ti alloys to biocompatible n-Ti SMAs. Specifically, this work investigates Ti-based SMAs modified with neutral elements, examining 3 alloy systems: (i) Ti-Zr-Hf, (ii) Ti-Zr-Sn, and (iii) Ti-Zr-Hf-Sn. Among the alloys tested, the Ti-38Zr-10Hf alloy stood out as the most promising, exhibiting the lowest phase transformation temperature (Tp) of 802 K, making it an ideal candidate for further Sn addition. In terms of phase stabilization, the n-phase could not be stabilized at room temperature (RT) with just the addition of Zr and Hf (set (i) alloys). However, both the Ti-Zr-Sn (set (ii)) and Ti-Zr-Hf-Sn (set (iii)) alloys successfully stabilized the n-phase at RT. Notably, the addition of neutral elements did not significantly affect the lattice constant of the n-phase, suggesting that these alloys have high potential for lattice deformation strain. This indicates their suitability for SMA applications. Additionally, the micro Vickers hardness of the Sn-added alloys showed a clear minimum near the martensite transformation start temperature (Ms), which is a typical phenomenon in SMAs. The lowest hardness was observed when Sn was added in the range of 5-6 mol%. These findings suggest that n-Ti SMAs with neutral element additions may hold promise for future biomedical applications.

This study developed machine learning (ML) models to predict the mechanical properties of Ni-free β-type titanium shape memory alloys (SMAs). Using a dataset of 107 entries derived from both literature and laboratory experiments, we focused on predicting ultimate tensile strength (UTS) and elongation (EL). Key features, including Mo equivalent, bond order, and d-orbital energy level, were selected for the models through Pearson correlation maps and subset selection methods. Four ML algorithms—Linear Regression (LIN), Support Vector Regression (SVR), Random Forest Regression (RFR), and Gradient Boosting Regression (GBR)—were employed and evaluated using metrics like mean absolute error (MAE), mean squared error (MSE), and coefficient of determination (R2). The GBR model for EL showed the highest prediction accuracy (R2 = 0.998 for training and R2 = 0.817 for testing), whereas UTS predictions were less accurate (R2 < 0.6 for testing). Although the models were also adapted to predict yield stress (YS), their accuracy was reduced, with improvements seen when incorporating phase constitution information reflecting phase stability. The primary reasons for the discrepancy in this study include the small dataset size and the absence of microstructural features. This research demonstrates the potential of ML models in predicting the mechanical properties of β-type titanium SMAs, highlighting the importance of integrating domain-specific knowledge through feature engineering to overcome the challenge of small data sets, and to enhance accuracy and robustness.