Advanced Computational Analysis of Cobalt-Based Superalloys through Crystal Plasticity.

This study introduces an advanced computational method aimed at accelerating continuum-scale processes using crystal plasticity approaches to predict mechanical responses in cobalt-based superalloys. The framework integrates two levels, namely, sub-grain and homogenized, at the meso-scale through crystal plasticity finite element (CPFE) platforms. The model is applicable across a temperature range from room temperature up to 900 °C, accommodating various dislocation mechanisms in the microstructure. The sub-grain level explicitly incorporates precipitates and employs a dislocation density-based constitutive model that is size-dependent. In contrast, the homogenized level utilizes an activation energy-based constitutive model, implicitly representing the γ' phase for efficiency in computations. This level considers the effects of composition and morphology on mechanical properties, demonstrating the potential for cobalt-based superalloys to rival nickel-based superalloys. The study aims to investigate the impacts of elements including tungsten, tantalum, titanium, and chromium through the homogenized constitutive model. The model accounts for the locking mechanism to address the cross-slip of screw dislocations at lower temperatures as well as the glide and climb mechanism to simulate diffusions at higher temperatures. The model's validity is established across diverse compositions and morphologies, as well as various temperatures, through comparison with experimental data. This advanced computational framework not only enables accurate predictions of mechanical responses in cobalt-based superalloys across a wide temperature range, but also provides valuable insights into the design and optimization of these materials for high-temperature applications.

MPpredictor: An Artificial Intelligence-Driven Web Tool for Composition-Based Material Property Prediction.

The applications of artificial intelligence, machine learning, and deep learning techniques in the field of materials science are becoming increasingly common due to their promising abilities to extract and utilize data-driven information from available data and accelerate materials discovery and design for future applications. In an attempt to assist with this process, we deploy predictive models for multiple material properties, given the composition of the material. The deep learning models described here are built using a cross-property deep transfer learning technique, which leverages source models trained on large data sets to build target models on small data sets with different properties. We deploy these models in an online software tool that takes a number of material compositions as input, performs preprocessing to generate composition-based attributes for each material, and feeds them into the predictive models to obtain up to 41 different material property values. The material property predictor is available online at http://ai.eecs.northwestern.edu/MPpredictor.

Multi-Scale Crystal Plasticity Model of Creep Responses in Nickel-Based Superalloys.

The current study focuses on the modeling of two-phase γ-γ' nickel-based superalloys, utilizing multi-scale approaches to simulate and predict the creep behaviors through crystal plasticity finite element (CPFE) platforms. The multi-scale framework links two distinct levels of the spatial spectrum, namely, sub-grain and homogenized scales, capturing the complexity of the system responses as a function of a tractable set of geometric and physical parameters. The model considers two dominant features of γ' morphology and composition. The γ' morphology is simulated using three parameters describing the average size, volume fraction, and shape. The sub-grain level is expressed by a size-dependent, dislocation density-based constitutive model in the CPFE framework with the explicit depiction of γ-γ' morphology as the building block of the homogenized scale. The homogenized scale is developed as an activation energy-based crystal plasticity model reflecting intrinsic composition and morphology effects. The model incorporates the functional configuration of the constitutive parameters characterized over the sub-grain γ-γ' microstructural morphology. The developed homogenized model significantly expedites the computational processes due to the nature of the parameterized representation of the dominant factors while retains reliable accuracy. Anti-Phase Boundary (APB) shearing and, glide-climb dislocation mechanisms are incorporated in the constitutive model which will become active based on the energies associated with the dislocations. The homogenized constitutive model addresses the thermo-mechanical behavior of nickel-based superalloys for an extensive temperature domain and encompasses orientation dependence as well as the loading condition of tension-compression asymmetry aspects. The model is validated for diverse compositions, temperatures, and orientations based on previously reported data of single crystalline nickel-based superalloy.

Cross-property deep transfer learning framework for enhanced predictive analytics on small materials data.

Artificial intelligence (AI) and machine learning (ML) have been increasingly used in materials science to build predictive models and accelerate discovery. For selected properties, availability of large databases has also facilitated application of deep learning (DL) and transfer learning (TL). However, unavailability of large datasets for a majority of properties prohibits widespread application of DL/TL. We present a cross-property deep-transfer-learning framework that leverages models trained on large datasets to build models on small datasets of different properties. We test the proposed framework on 39 computational and two experimental datasets and find that the TL models with only elemental fractions as input outperform ML/DL models trained from scratch even when they are allowed to use physical attributes as input, for 27/39 (≈ 69%) computational and both the experimental datasets. We believe that the proposed framework can be widely useful to tackle the small data challenge in applying AI/ML in materials science.

Design Strategy for Additive Manufacturing Ti-Al-Fe alloys with Calculation of Phase Diagram Method.

Calculation of Phase Diagram (CALPHAD) method was used to develop an additive manufacturing (AM) titanium alloy design strategy for new alloys in the Ti-Al-Fe system. This design strategy includes five key criteria addressing AM-processing suitability, mechanical properties and heat treatment capability. The methods for evaluating the design criteria were determined considering both physical implications and computational efficiency. Available experimental results agree well with CALPHAD predictions in the design map. CALPHAD method in this work can be applied for high-throughput early-stage assessments in large composition space for new AM titanium alloy development.

Author Correction: Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Learning to Predict Crystal Plasticity at the Nanoscale: Deep Residual Networks and Size Effects in Uniaxial Compression Discrete Dislocation Simulations.

The density and configurational changes of crystal dislocations during plastic deformation influence the mechanical properties of materials. These influences have become clearest in nanoscale experiments, in terms of strength, hardness and work hardening size effects in small volumes. The mechanical characterization of a model crystal may be cast as an inverse problem of deducing the defect population characteristics (density, correlations) in small volumes from the mechanical behavior. In this work, we demonstrate how a deep residual network can be used to deduce the dislocation characteristics of a sample of interest using only its surface strain profiles at small deformations, and then statistically predict the mechanical response of size-affected samples at larger deformations. As a testbed of our approach, we utilize high-throughput discrete dislocation simulations for systems of widths that range from nano- to micro- meters. We show that the proposed deep learning model significantly outperforms a traditional machine learning model, as well as accurately produces statistical predictions of the size effects in samples of various widths. By visualizing the filters in convolutional layers and saliency maps, we find that the proposed model is able to learn the significant features of sample strain profiles.

Nitrogen Effects in Additively Manufactured Martensitic Stainless Steels: Conventional Thermal Processing and Comparison with Wrought.

The microstructures of additively manufactured (AM) precipitation-hardenable stainless steels 17-4 and 15-5 were investigated and compared to those of conventionally produced materials. The residual N found in N2-atomized 17-4 powder feedstock is inherited by the additively produced material, and has dramatic effects on phase stability, microstructure, and microstructural evolution. Nitrogen is a known austenite stabilizing element, and the as-built microstructure of AM 17-4 can contain up to 90 pct or more retained austenite, compared to the nearly 100 pct martensite structure of wrought 17-4. Even after homogenization and solutionization heat treatments, AM 17-4 contains 5 to 20 pct retained austenite. In contrast, AM 15-5 and Ar-atomized AM 17-4 contain<5 pct retained austenite in the as-built condition, and this level is further decreased following post-build thermal processing. Computational thermodynamics-based calculations qualitatively describe the observed depression in the martensite start temperature and martensite stability as a function of N-content, but require further refinements to become quantitative. A significant increase in the volume fraction of fine-scale carbide precipitates attributed to the high N-content of AM 17-4 is also hypothesized to give rise to additional activation barriers for the dislocation motion required for martensite nucleation and subsequent growth. An increase in the volume fraction of carbide/nitride precipitates is observed in AM 15-5, although they do not inhibit martensite formation to the extent observed in AM 17-4.

Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning.

The current predictive modeling techniques applied to Density Functional Theory (DFT) computations have helped accelerate the process of materials discovery by providing significantly faster methods to scan materials candidates, thereby reducing the search space for future DFT computations and experiments. However, in addition to prediction error against DFT-computed properties, such predictive models also inherit the DFT-computation discrepancies against experimentally measured properties. To address this challenge, we demonstrate that using deep transfer learning, existing large DFT-computational data sets (such as the Open Quantum Materials Database (OQMD)) can be leveraged together with other smaller DFT-computed data sets as well as available experimental observations to build robust prediction models. We build a highly accurate model for predicting formation energy of materials from their compositions; using an experimental data set of [Formula: see text] observations, the proposed approach yields a mean absolute error (MAE) of [Formula: see text] eV/atom, which is significantly better than existing machine learning (ML) prediction modeling based on DFT computations and is comparable to the MAE of DFT-computation itself.

Effect of heat treatment on the microstructural evolution of a nickel-based superalloy additive-manufactured by laser powder bed fusion.

Elemental segregation is a ubiquitous phenomenon in additive-manufactured (AM) parts due to solute rejection and redistribution during the solidification process. Using electron microscopy, in situ synchrotron X-ray scattering and diffraction, and thermodynamic modeling, we reveal that in an AM nickel-based superalloy, Inconel 625, stress-relief heat treatment leads to the growth of unwanted δ-phase precipitates on a time scale much faster than that in wrought alloys (minutes versus tens to hundreds of hours). The root cause for this behavior is the elemental segregation that results in local compositions of AM alloys outside the bounds of the allowable range set for wrought alloys. In situ small angle scattering experiments reveal that platelet-shaped δ phase precipitates grow continuously and preferentially along their lateral dimensions during stress-relief heat treatment, while the thickness dimension reaches a plateau very quickly. In situ XRD experiments reveal that nucleation and growth of δ-phase precipitates occur within 5 min during stress-relief heat treatment, indicating a low nucleation barrier and a short incubation time. An activation energy for the growth of δ phase was found to be (131.04 ± 0.69) kJ mol-1. We further demonstrate that a subsequent homogenization heat treatment can effectively homogenize the AM alloy and remove the deleterious δ phase. The combined experimental and modeling methodology in this work can be extended to elucidate the phase evolution during heat treatments in a broad range of AM materials.

Generating Domain Terminologies using Root- and Rule-Based Terms.

Motivated by the need for flexible, intuitive, reusable, and normalized terminology for guiding search and building ontologies, we present a general approach for generating sets of such terminologies from natural language documents. The terms that this approach generates are root- and rule-based terms, generated by a series of rules designed to be flexible, to evolve, and, perhaps most important, to protect against ambiguity and standardize semantically similar but syntactically distinct phrases to a normal form. This approach combines several linguistic and computational methods that can be automated with the help of training sets to quickly and consistently extract normalized terms. We discuss how this can be extended as natural language technologies improve and how the strategy applies to common use-cases such as search, document entry and archiving, and identifying, tracking, and predicting scientific and technological trends.

The Effect of Oxygen on Phase Equilibria in the Ti-V System: Impacts on the AM Processing of Ti Alloys.

Oxygen is always a constituent in "real" titanium alloys including titanium alloy powders used for powder-based additive manufacturing (AM). In addition, oxygen uptake during powder handling and printing is hard to control and, hence, it is important to understand and predict how oxygen is affecting the microstructure. Therefore, oxygen is included in the evaluation of the thermodynamic properties of the titanium-vanadium system employing the CALculation of PHAse Diagrams method and a complete model of the O-Ti-V system is presented. The ß-transus temperature is calculated to increase with increasing oxygen content whereas the extension of the α-Ti phase field into the binary is calculated to decrease, which explains the low vanadium solubilities measured in some experimental works. In addition, the critical temperature of the metastable miscibility gap of the ß-phase is calculated to increase to above room temperature when oxygen is added. The effects of oxygen additions on phase fractions, martensite and ω formation temperatures are discussed, along with the impacts these changes may have on AM of titanium alloys.

Systems Design Approach to Low-Cost Coinage Materials.

A systems approach within an Integrated Computational Materials Engineering framework was used to design three new low-cost seamless replacement coinage alloys to reduce the raw material of the current US coinage alloys. Maintaining compatibility with current coinage materials required matching the currently used alloy properties of yield strength, work-hardening behavior, electrical conductivity, color, corrosion resistance and wear resistance. In addition, the design alloys were required to use current production processes. CALPHAD-based models for electrical conductivity and color were developed to integrate into the system design. Three prototype alloys were designed, produced and characterized. The design process highlighted the trade-off between minimizing the raw material costs and achieving the desired color properties. Characterization of the three prototype alloys showed good agreement with the design goals.

Homogenization Kinetics of a Nickel-based Superalloy Produced by Powder Bed Fusion Laser Sintering.

Additively manufactured (AM) metal components often exhibit fine dendritic microstructures and elemental segregation due to the initial rapid solidification and subsequent melting and cooling during the build process, which without homogenization would adversely affect materials performance. In this letter, we report in situ observation of the homogenization kinetics of an AM nickel-based superalloy using synchrotron small angle X-ray scattering. The identified kinetic time scale is in good agreement with thermodynamic diffusion simulation predictions using microstructural dimensions acquired by ex situ scanning electron microscopy. These findings could serve as a recipe for predicting, observing, and validating homogenization treatments in AM materials.

A Computational Framework for Material Design.

A computational framework is proposed that enables the integration of experimental and computational data, a variety of user-selected models, and a computer algorithm to direct a design optimization. To demonstrate this framework, a sample design of a ternary Ni-Al-Cr alloy with a high work-to-necking ratio is presented. This design example illustrates how CALPHAD phase-based, composition and temperature-dependent phase equilibria calculations and precipitation models are coupled with models for elastic and plastic deformation to calculate the stress-strain curves. A genetic algorithm then directs the search within a specific set of composition and processing constraints for the ideal composition and processing profile to optimize the mechanical properties. The initial demonstration of the framework provides a potential solution to initiate the material design process in a large space of composition and processing conditions. This framework can also be used in similar material systems or adapted for other material classes.

In Situ Structural Characterization of Ageing Kinetics in Aluminum Alloy 2024 across Angstrom-to-Micrometer Length Scales.

The precipitate structure and precipitation kinetics in an Al-Cu-Mg alloy (AA2024) aged at 190 °C, 208 °C, and 226 °C have been studied using ex situ Transmission Electron Microscopy (TEM) and in situ synchrotron-based, combined ultra-small angle X-ray scattering, small angle X-ray scattering (SAXS), and wide angle X-ray scattering (WAXS) across a length scale from sub-Angstrom to several micrometers. TEM brings information concerning the nature, morphology, and size of the precipitates while SAXS and WAXS provide qualitative and quantitative information concerning the time-dependent size and volume fraction evolution of the precipitates at different stages of the precipitation sequence. Within the experimental time resolution, precipitation at these ageing temperatures involves dissolution of nanometer-sized small clusters and formation of the planar S phase precipitates. Using a three-parameter scattering model constructed on the basis of TEM results, we established the temperature-dependent kinetics for the cluster-dissolution and S-phase formation processes simultaneously. These two processes are shown to have different kinetic rates, with the cluster-dissolution rate approximately double the S-phase formation rate. We identified a dissolution activation energy at (149.5 ± 14.6) kJ mol-1, which translates to (1.55 ± 0.15) eV/atom, as well as an activation energy for the formation of S precipitates at (129.2 ± 5.4) kJ mol-1, i.e. (1.33 ± 0.06) eV/atom. Importantly, the SAXS/WAXS results show the absence of an intermediate Guinier-Preston Bagaryatsky 2 (GPB2)/S″ phase in the samples under the experimental ageing conditions. These results are further validated by precipitation simulations that are based on Langer-Schwartz theory and a Kampmann-Wagner numerical method.