Vision-based quality assessment (QA) is essential in ensuring production quality in metal additive manufacturing (AM). Traditional machine learning-based QA requires dedicated models to be trained on specialized data, making the process expensive and time-consuming. In this work, we leverage the visual reasoning capabilities of cutting-edge vision-language models (VLMs) to perform QA without task-specific model training or fine-tuning. Application-specific knowledge is incorporated through in-context learning (ICL), which enables the model to adapt to tasks with only a small number of examples. Our results show that the ICL-assisted VLM can achieve QA accuracies close to those of specially trained ML models. Additionally, VLMs can provide human-interpretable justifications for their QA decisions, increasing human trust. To assess the faithfulness of VLM-generated linguistic justifications, we introduce two metrics, knowledge relevance and rationale validity, providing a principled framework for evaluating explanation quality.

Understanding the atomic structure of precipitate phases in shape memory alloys is critical to determining their structure-property relationships and developing high-performance shape memory alloys. However, experimental methods are limited in determining atomic configurations incases where the number of atoms per unit cell is very high, or the phase is small (few nms). While density functional theory (DFT) can aid in the accurate determination of a phase's crystallography, this is challenged by the number of candidate structures. Recently, a cubic phase was discovered during the heat treatment of a Hf-Ni-Ti alloy developed with improved tribological applications and rolling contact fatigue. We use DFT, machine learned interatomic potentials (MLIPs), and a genetic algorithm to identify likely configurations for the cubic phase. Likely candidate structures consistent with experimentally determined structural information were identified. Limitations of experimental microscopy methods, crystal simulation, and DFT-MLIP techniques are discussed.

Outstanding challenges must be addressed to mature additive manufacturing (AM) technologies for glass‐forming alloys (GFA) including increasing as‐built densities, limiting cracks and pores, controlling crystallinity within fusion and heat‐affected zones (HAZ), and increasing the size of process parameter windows that result in quality materials. The thermal cycles that are unique and specific to AM limit the applicability of research from casting literature under some circumstances. In this work, single‐track and multitrack laser glazing experiments performed on four unique suction cast glass‐forming compositions provide a means for screening the suitability and relative process window sizes for powder bed fusion–laser beam (PBF–LB) AM of four GFAs in the absence of available powders. The crystallinity of the fusion zone and the HAZ as well as the presence of cracks are observed as process parameters are varied for each alloy. Both conventionally good and conventionally bad glass formers show great potential for PBF–LB processability. Glass‐forming ability in and of itself is not a good predictor of PBF–LB printability.

Laser powder bed fusion (LPBF) additive manufacturing is widely used to fabricate geometrically complex metal components. In LPBF, a layer of metal powder is swept onto the build surface by a recoater blade and then melted with a laser; this process is repeated to build a part layer-by-layer. A common flaw in LPBF processes is overbuilding, where some region of a layer is built higher than the desired layer height. Collisions between the overbuild and the recoater blade on subsequent sweeps can lead to a persistent defect in the part or damage to the recoater blade. In this research, acoustic emission sensors mounted to the recoater blade subassembly of an EOS M290 machine are used to detect collisions with the recoater blade. This work presents the sensor configurations for and results of preliminary experiments conducted during powder sweeps only and during fabrication of small test coupons. [Work supported by NASA Space Technology Graduate Research Opportunities (Grant No. 80NSSC23K1213)]

A rate-dependent crystal-plasticity (CP) framework that captures the coupled phase transformation - plastic deformation behavior of shape memory alloys (SMAs) is presented. Here, different from previous models, the flow rule for martensitic phase transformation incorporates the entire deformation gradient for transformation, including the rotation. Predictions of transformation strain and variant selection of Nickel-Titanium (NiTi) using this model are directly compared with previous formulations that did not include the rotation. The results show that the rotation is essential to accurately calculate the single crystal and polycrystal micromechanics of variant selection and transformation strains of SMAs. The constitutive law formulation also includes current formulations for both slip and deformation twinning plasticity mechanisms, and the differences in transformation mechanisms are further shown to impact plasticity calculations through transformation-plasticity interactions. In addition to the advancement of the constitutive law, a computationally efficient implicit time integration scheme is given for numerical implementation and demonstrated using a user material subroutine (UMAT) in the commercial finite element code ABAQUS Standard. The proposed framework and the associated numerical protocols achieve stable solutions using strain increments on the order of 0.05 mm/mm in simulating inelastic deformations and strain increments 0.01 mm/mm in the elastic-inelastic transitions. Furthermore, the use of an analytic Jacobian results in stable convergence in fewer than 10 global Newton iterations while calculating solutions for elastic-inelastic transitions, making the computational benefits evident.

Cold spray additive manufacturing (CSAM) is an attractive solid-state bonding technique due to its rapid manufacturing rate and the ability to avoid deleterious effects found in solidification-based additive manufacturing. Unfortunately, CSAM of steel components has been difficult to date to the high strength of the steel particles which resists deformation and creates interparticle porosity. Herein, it is found adding softer Cu powder particles to steel (SS316) powder and utilizing a heat treatment can decrease the porosity of the as-sprayed structure while increasing the mechanical properties. The mixture results in an increased sprayability of the structure, as the Cu particles preferentially fill the pores, increasing the density. The microstructural evolution of the SS316 and Cu particles at the particle interfaces and interiors is investigated and reveals that the materials undergo a heterogeneous deformation route which facilitates the densification of the CSAM structure. Through annealing these components, the tensile strength increases and the density increases further. Both materials undergo microstructural recovery along with selected interdiffusion of elements which improves the metallurgical bonding. It is demonstrated that the heterogeneous deposition and microstructural evolution between the dissimilar materials can improve the overall component properties.

Timely part procurement is vital to the maintenance and performance of deployed military equipment. Yet, logistical hurdles can delay this process, which can compromise efficiency and mission success for the warfighter. Point-of-need part procurement through additive manufacturing (AM) is a means to circumvent these logistical challenges. An Integrated Computational Materials Engineering framework is presented as a means to validate and quantify the performance of AM replacement parts. Statistical modeling using a random forest network and finite element modeling were to inform the build design. Validation was performed by testing coupons extracted from each legacy replacement parts, as well as the new additively manufactured replacement parts through monotonic tensile and combined tension-torsion fatigue testing. Destructive full hinge assembly tests were also performed as part of the experimental characterization. Lastly, the collected experimental results were used to iterate on the part design to showcase the advantages AM offers for meeting new service requirements. Citation: T.G. Gallmeyer, J. Dahal, B.B. Kappes, A.P. Stebner, R.S. Thyagarajan, J.A. Miranda, A. Pilchak, J. Nuechterlein, “Systematic Development of Framework for Validation and Performance Quantification of Additively Manufactured (AM) Replacement Parts for Structural Steel Applications”, In Proceedings of the Ground Vehicle Systems Engineering and Technology Symposium (GVSETS), NDIA, Novi, MI, Aug. 13-15, 2019.

Additively manufactured lattice structures offer a high specific strength-to-density ratio in comparison to conventional, fully dense parts. The small feature size and intricate geometry of lattice structures make it challenging to directly characterize their mechanical properties. This study presents direct measurement of the elastic mechanical properties of an additively manufactured titanium alloy octet truss unit cell. Strut stress and strain were measured using high-energy X-ray diffraction during quasi-static compressive loading. The crystallographic stress-strain relationship was converted into the stress-strain relationship of the octet truss lattice unit cell using the known symmetry of the struts. The sample studied herein was best fit by an orthorhombic linear elastic stress-strain relationship. This technique shows promise for studying additively manufactured lattice structures at length scales not previously reported.

To ensure a strong adhesive bond, most standards and adhesive manufacturers specify a maximum adhesive gap of 1 mm when bonding fiber reinforced composite structures. In manufacturing large components, such as joining two halves of wind turbine blades, meeting this gap tolerance specification is impractical; gaps larger than 10 mm are common in large adhesively bonded composite structures using state-of-the-art manufacturing techniques. Currently, there is a lack of fundamental understanding of the failure mechanics of adhesive gaps larger than 3 mm. To create such understanding, glass fiber – acrylic thermoplastic composite panels bonded using different epoxy adhesives within single-lap joint samples with adhesive thicknesses of 0.1 mm, 0.3 mm, 1 mm, 3 mm, 5 mm, and 10 mm were sheared to failure. A transition from cohesive to adhesive failure was observed to occur about 1 mm to 3 mm joint thicknesses. Plotting the shear stress normalized by the ratio of the joint width to thickness as a function of the joint thickness normalized by the joint length is shown to result in the ability to fit simple empirically derived models of the cohesive-to-adhesive failure transition, regardless of the adhesive. Furthermore, using these normalized variables, all the observed cohesively failed specimens collapse to a single master curve, as do the adhesively failed specimens.

The necessity for recyclable materials in wind energy applications has fueled research in glass fiber reinforced thermoplastic composites to replace their thermoset counterparts. Toward demonstrating that infusible acrylic resins can replace epoxy based composite systems in wind blade manufacturing, comprehensive static test protocols were performed, and the resulting data are presented. Specifically, unidirectional and biaxial (±45) continuous E-glass reinforced thermoplastic and epoxy laminates were prepared in four and eight ply laminates. Physical properties were characterized including density, fiber volume fraction, and glass transition temperatures, together with mechanical properties for tensile, compression, and shear responses. Comprehensive evaluation of these data supports infusible acrylic thermoplastic resin systems as viable alternative prospects to replace epoxies in E-glass reinforced wind blades as verified by comparable results for both composite systems.

Machine learning is on a bit of a tear right now, with advances that are infiltrating nearly every aspect of our lives. In the domain of materials science, this wave seems to be growing into a tsunami. Yet, there are still real hurdles that we face to maximize its benefit. This Matter of Opinion, crafted as a result of a workshop hosted by researchers at Sandia National Laboratories and attended by a cadre of luminaries, briefly summarizes our perspective on these barriers. By recognizing these problems in a community forum, we can share the burden of their resolution together with a common purpose and coordinated effort.

Microstructures of aged nickel‑titanium‑hafnium shape memory alloys (SMAs) Ni50.3Ti41.2Hf8.5 and Ni50.3Ti43.7Hf6 (at.%) are examined using high-resolution transmission electron microscopy (HRTEM) techniques. Characterizations of the structures of austenite-martensite, martensite-martensite, and martensite-precipitate interfaces provide new insights into cubic B2 austenite–monoclinic B19′ martensite phase transformation and martensite reorientation mechanisms. Specifically, in the absence of external load, theoretically unfavorable (001) compound martensite twins nucleate at H-phase precipitate interfaces. At the interface of these twins and the austenite matrix, a zig-zag-rise structure follows an extension of superdislocation theory previously developed to explain analogous zig-zag structured interfaces in non-shape-memory martensitic alloys. Additionally, 〈011〉 type II martensite twins that are theoretically favorable to form were also observed within the austenite matrices. However, instead of forming defect-free habit planes accommodated by elastic strains as predicted by theory, misfit dislocations worked in cooperation with elastic strain fields to accommodate atomic misfits between the austenite and martensite phases, and the elastic strain fields were smaller at the interfaces. Finally, the atomic scale thermoelastic mechanism of <011> type II twins reorienting to form (001) compound twins within the martensite is proposed in a manner consistent with direct observations of such interfaces. In total, connections with previous work on other binary and ternary NiTi base alloys show that these new mechanistic understandings generalize to other B2 - B19′ SMAs that should not perform well according to the elasticity-based Phenomenological Theory of Martensite Crystallography, yet still exhibit robust cyclic thermomechanical performances due to transformation-plasticity interactions.

Nickel‐rich NiTiHf alloys that are heat treated to strengthen the microstructures with a dense distribution of Ni 4 Ti 3 nanoprecipitates exhibit very high strengths and good quasi‐static indentation resistance and rolling contact fatigue performances. To determine whether these properties are maintained at high rates of loading, in situ and recovery flyer plate impact shock experiments are performed on a Ni 54 Ti 45 Hf 1 alloy at impact velocities ranging from approximately 150 m s −1 (2.5 GPa) to 700 m s −1 (12.40 GPa). Analysis of shocked samples indicated less cracking is observed to emanate from spall failures resulting from impact velocities greater than 250 m s −1 (4.23 GPa), concurrent with observations of intragranular microbands within the microstructures. Analyses show clear evidence that, like responses to quasi‐static loading, martensitic phase transformation occurs upon shock compression in all cases. However, dissimilarly, for the higher impact velocities it reverses upon stress release, leaving behind microbands that show no evidence for retained martensite and within which the Ni 4 Ti 3 nanoprecipitates dissolved. These results indicate that strain‐rate dependence of these SMAs under shock loading is not only governed by the expected physics of rate‐dependence of the martensitic transformations themselves but may also be enhanced by inelastic deformation mechanisms that result in precipitate dissolution.

A solid state phase transformation of Ti-6Al-4V was studied using high speed in situ x-ray diffraction measurements made during rapid cooling of a cold metal transfer arc weld bead deposited onto a water cooled substrate. Analysis of body centered cubic (BCC) and hexagonal close packed (HCP) lattices revealed an abrupt, nonlinear shift in the lattice parameters of both phases just after the HCP phase had nucleated. Postmortem transmission electron microscopy confirmed that V diffusion was mostly suppressed during cooling. Together, these results indicate that at this cooling rate of approximately 104 K/s, which is representative of cooling rates of many additive manufacturing and welding processes, kinematic coherency of the BCC–HCP interfaces gives rise to the anomalous lattice expansion and contraction behaviors of both phases during the initial nucleation and growth stages of (mostly) martensitic transformation from BCC to HCP; the role of diffusion in such lattice anomalies is shown to be minimal.

There is a growing body of evidence that “in situ” modification of the microstructure of Nitinol due to pre-strain and mean strain strongly influence the fatigue properties of implanted medical devices and often, counterintuitively, for the better. The purpose of this review paper is to focus on the experimental evidence and to support these finding with micro-mechanical and metallurgical references. Pre-strain effects are well recognized in Nitinol thermal actuator fatigue behavior and applied in medical device applications but the effects of mean strain remain controversial. Detailed reviews of the pertinent literature indicate that pre-strains intensify texture, increase martensite volume fraction, and induce plasticity that may enhance or degrade fatigue performance. The majority of experimental studies demonstrate that fatigue cycling with mean strains within the two-phase region leads to longer lives compared to cycling in the linear elastic region or at mean strains greater than the end of the stress plateau. These effects are described in terms of martensite stabilization, cyclic phase change, and the decrease in composite modulus with increasing mean strain. These factors combine to result in alternating strain between approximately constant upper and lower plateau stresses to induce phase change.

Elastocaloric cooling, which exploits superelastic transitions of shape memory alloys to pump heat, has recently emerged as a frontrunner in alternative cooling technologies. Despite its intrinsic high efficiency, elastocaloric materials exhibit hysteresis associated with input work, a common attribute of caloric cooling materials. In this study, the authors created a Ni-Ti-based elastocaloric material by additive manufacturing nanocomposite materials using a laser directed-energy- deposition system. The material exhibited exceptional stability and unusual operational efficiency derived from the unique and intricate nanocomposite structures made by additive manufacturing. This demonstration shows the potential for using additive manufacturing to optimize caloric cooling by providing a highly desirable topology flexibility into materials components that serve as both refrigerants and heat exchangers.

This collection of case studies presents a brief introduction of fundamental concepts for four SMA-based applications in the aerospace, energy, and medical fields designed by students and facilitated by professionals. The Consortium for the Advancement of Shape Memory Alloy Research and Technology (CASMART) Student Design Challenge is used as an outreach strategy to promote the implementation of state-of-the-art designs with SMA technology and is meant to inspire the next generation of SMA research. Student design challenge teams address real-world problems facing the SMA community and receive guidance and feedback from CASMART members. Student teams’ hardware and materials deliverables had to meet basic function requirements specific to the application. Key results from seven teams (four hardware designs and three materials designs) highlight the design priorities, processes, and challenges raised during development. The hardware designs used NiTi wires shape set and implemented by the students into prototypes for deployment and reorientation mechanisms in small satellites, linear generators to save energy, and a self-apply tourniquet design. Materials development explored the processability and material properties of CuAl-based and NiTi-based alloys for passive actuators in a deployment and reorientation mechanism for a small satellite, energy recovery from waste heat, and a pseudoelastic spinal curvature correction device.

Plastic deformation of B2 Nickel Titanium is usually attributed to {110}< 001 > slip and {114}< 221 > deformation twinning. The most commonly observed hot-worked texture of these alloys, is defined by a {111}< vuw > gamma fiber and {hkl}< 110 > partial alpha fiber. A Visco-Plastic Self Consistent (VPSC) model was used to establish relationships between microscopic slip and twin activity with the observed macroscopic hot-rolling texture. This knowledge will better aid in modeling NiTi austenite plasticity. Since the primary slip modes in NiTi do not have five independent slip systems, and deformation that can be accommodated by twinning is limited, multiple deformation modes must contribute to NiTi ductility. It is shown that {110}< 001 > slip, {100}< 001 > slip and, {114}< 221 > twin deformation modes need to be active simultaneously to explain the observed textures. The relative CRSS ratios and hardening parameters were varied to study the effect of the deformation modes on the various texture components. Textures observed below 723 K and at less than 80% rolling reduction were simulated with deformation primarily accommodated on the {110}< 001 > slip mode and {114}< 221 > twinning mode. Textures observed at temperatures greater than 903 K and greater than 80% rolling reductions were captured in the simulations that included all three deformation modes. Activity on the {100}< 001 > slip mode strongly correlated with the {110}< 110 > texture component observed at high temperatures.

The deformation mechanisms that dictate the tribological performances of heat treated Ni55Ti45, Ni54Ti45Hf1 and Ni56Ti36Hf8 alloys are revealed through rolling contact fatigue (RCF) testing and transmission electron microscopy (TEM) analyses. Analysis of worn samples that passed a 5×108 cycle RCF runout condition shows that damage is primarily confined to deformation bands that propagate several hundred nanometers – to – several microns beneath the surface. These bands nucleate via localization of dislocation slip within the B2 austenite phase of the alloys. For the Ni55Ti45 and Ni54Ti45Hf1 samples, further damage and eventual spall failures occur by shearing and dissolution of the strengthening Ni4Ti3 nanoprecipitates within the deformation bands, followed by nanocrystallization that sometimes includes stress-induced nucleation of B19′ nanocrystals. Eventually, complete amorphization occurs prior to fracture. The relative 15 – 25% increase in RCF contact stress performance of the Ni56Ti36Hf8 alloy correlates with a more limited depth of damaged material beneath the wear surface; heavily damaged material beneath spall failure sites only extends 1.5 µm into the sample, compared to > 6 µm for Ni55Ti45 and Ni54Ti45Hf1 alloys. This superior RCF damage resistance of the Ni56Ti36Hf8 alloy results from a lower fraction of B2 matrix phase (≤ 13 %) that is highly confined by both cubic Ni-rich NiTiHf and H-phase nanoprecipitates. Specifically, this microstructure limits the widths of deformation bands within the B2 austenite phase to less than the size of the strengthening nanoprecipitates, which in turn inhibits precipitate shearing and dissolution processes that precede nanocrystallization and amorphization.