arXiv subjects

Raymundo Arroyave

Publications and source records attributed to Raymundo Arroyave.

At least 37 records · Page 2Linked to original sources

arXiv2024

Decoding Non-Linearity and Complexity: Deep Tabular Learning Approaches for Materials Science

Materials data, especially those related to high-temperature properties, pose significant challenges for machine learning models due to extreme skewness, wide feature ranges, modality, and complex relationships. While traditional models like tree-based ensembles (e.g., XGBoost, LightGBM) are commonly used for tabular data, they often struggle to fully capture the subtle interactions inherent in materials science data. In this study, we leverage deep learning techniques based on encoder-decoder architectures and attention-based models to handle these complexities. Our results demonstrate that XGBoost achieves the best loss value and the fastest trial duration, but deep encoder-decoder learning like Disjunctive Normal Form architecture (DNF-nets) offer competitive performance in capturing non-linear relationships, especially for highly skewed data distributions. However, convergence rates and trial durations for deep model such as CNN is slower, indicating areas for further optimization. The models introduced in this study offer robust and hybrid solutions for enhancing predictive accuracy in complex materials datasets.

cond-mat.mtrl-sci↗

arXiv2024

Hierarchical Gaussian Process-Based Bayesian Optimization for Materials Discovery in High Entropy Alloy Spaces

Bayesian optimization (BO) is a powerful and data-efficient method for iterative materials discovery and design, particularly valuable when prior knowledge is limited, underlying functional relationships are complex or unknown, and the cost of querying the materials space is significant. Traditional BO methodologies typically utilize conventional Gaussian Processes (cGPs) to model the relationships between material inputs and properties, as well as correlations within the input space. However, cGP-BO approaches often fall short in multi-objective optimization scenarios, where they are unable to fully exploit correlations between distinct material properties. Leveraging these correlations can significantly enhance the discovery process, as information about one property can inform and improve predictions about others. This study addresses this limitation by employing advanced kernel structures to capture and model multi-dimensional property correlations through multi-task (MTGPs) or deep Gaussian Processes (DGPs), thus accelerating the discovery process. We demonstrate the effectiveness of MTGP-BO and DGP-BO in rapidly and robustly solving complex materials design challenges that occur within the context of complex multi-objective optimization -- carried out by leveraging the pyiron workflow manager over FCC FeCrNiCoCu high entropy alloy (HEA) spaces, where traditional cGP-BO approaches fail. Furthermore, we highlight how the differential costs associated with querying various material properties can be strategically leveraged to make the materials discovery process more cost-efficient.

cond-mat.mtrl-sci↗

arXiv2024

Two-Shot Optimization of Compositionally Complex Refractory Alloys

In this paper, a synergistic computational/experimental approach is presented for the rapid discovery and characterization of novel alloys within the compositionally complex (i.e., "medium/high entropy") refractory alloy space of Ti-V-Nb-Mo-Hf-Ta-W. This was demonstrated via a material design cycle aimed at simultaneously maximizing the objective properties of high specific hardness (hardness normalized by density) and high specific elastic modulus (elastic modulus normalized by density). This framework utilizes high-throughput computational thermodynamics and intelligent filtering to first reduce the untenably large alloy space to a feasible size, followed by an iterative design cycle comprised of high-throughput synthesis, processing, and characterization in batch sizes of 24 alloys. After the first iteration, Bayesian optimization was utilized to inform selection of the next batch of 24 alloys. This paper demonstrates the benefit of using batch Bayesian optimization (BBO) in material design, as significant gains in the objective properties were observed after only two iterations or "shots" of the design cycle without using any prior knowledge or physical models of how the objective properties relate to the design inputs (i.e., composition). Specifically, the hypervolume of the Pareto front increased by 54% between the first and second iterations. Furthermore, 10 of the 24 alloys in the second iteration dominated all alloys from the first iteration.

cond-mat.mtrl-sci↗

arXiv2024

Data-driven study of composition-dependent phase compatibility in NiTi shape memory alloys

The martensitic transformation in NiTi-based Shape Memory Alloys (SMAs) provides a basis for shape memory effect and superelasticity, thereby enabling applications requiring solid-state actuation and large recoverable shape changes upon mechanical load cycling. In order to tailor the transformation to a particular application, the compositional dependence of properties in NiTi-based SMAs, such as martensitic transformation temperatures and hysteresis, has been exploited. However, the compositional design space is large and complex, and experimental studies are expensive. In this work, we develop an interpretable piecewise linear regression model that predicts the $λ_2$ parameter, a measure of compatibility between austenite and martensite phases, and an (indirect) factor that is well-correlated with martensitic transformation hysteresis, based on the chemical features derived from the alloy composition. The model is capable of predicting, for the first time, the type of martensitic transformation for a given alloy chemistry. The proposed model is validated by experimental data from the literature as well as in-house measurements. The results show that the model can effectively distinguish between $B19$ and $B19^{\prime}$ regions for any given composition in NiTi-based SMAs and accurately estimate the $λ_2$ parameter. Our analysis also reveals that the weighted average of the quotient of the first ionization energy and the Voronoi coordination number is a key compositional characteristic that correlates with the $λ_2$ parameter and thermodynamic responses, including the transformation hysteresis, martensite start temperature, and critical temperature. The work herein demonstrates the potential of data-driven methodologies for understanding and designing NiTi-based SMAs with desired transformation characteristics.

cond-mat.mtrl-sci↗

arXiv2023

Phase-Field Model of Silicon Carbide Growth During Isothermal Condition

Silicon carbide (SiC) emerges as a promising ceramic material for high-temperature structural applications, especially within the aerospace sector. The utilization of SiC-based ceramic matrix composites (CMCs) instead of superalloys in components like engine shrouds, combustors, and nozzles offers notable advantages, including a 25% improvement in fuel efficiency, over 10% enhanced thrust, and the capability to withstand up to 500$^{\circ}$C higher operating temperatures. Employing a CALPHAD-reinforced multi-phase-field model, our study delves into the evolution of the SiC layer under isothermal solidification conditions. By modeling the growth of SiC between liquid Si and solid C at 1450$^{\circ}$C, we compared results with experimental microstructures and quantitatively examined the evolution of SiC thickness over time. Efficient sampling across the entire model space mitigated uncertainty in high-temperature kinetic parameters, allowing us to predict a range of growth rates and morphologies for the SiC layer. The model accounts for parameter uncertainty stemming from limited experimental knowledge and successfully predicts relevant morphologies for the system. Experimental results validated the kinetic parameters of the simulations, offering valuable insights and potential constraints on the reaction kinetics. We further explored the significance of multi-phase-field model parameters on two key outputs, and found that the diffusion coefficient of liquid Si emerges as the most crucial parameter significantly impacting the SiC average layer thickness and grain count over time. This study provides valuable insights into the microstructure evolution of the Si-C binary system, offering pertinent information for the engineering of CMCs in industrial applications.

physics.app-ph↗

arXiv2023

A ductility metric for refractory-based multi-principal-element alloys

We propose a quantum-mechanical dimensionless metric, the local$-$lattice distortion (LLD), as a reliable predictor of ductility in refractory multi-principal-element alloys (RMPEAs). The LLD metric is based on electronegativity differences in localized chemical environments and combines atomic$-$scale displacements due to local lattice distortions with a weighted average of valence$-$electron count. To evaluate the effectiveness of this metric, we examined body$-$centered cubic (bcc) refractory alloys that exhibit ductile$-$to$-$brittle behavior. Our findings demonstrate that local$-$charge behavior can be tuned via composition to enhance ductility in RMPEAs. With finite$-$sized cell effects eliminated, the LLD metric accurately predicted the ductility of arbitrary alloys based on tensile$-$elongation experiments. To validate further, we qualitatively evaluated the ductility of two refractory RMPEAs, i.e., NbTaMoW and Mo$_{72}$W$_{13}Ta$_{10}Ti$_{2.5}Zr$_{2.5}, through the observation of crack formation under indentation, again showing excellent agreement with LLD predictions. A comparative study of three refractory alloys provides further insights into the electronic-structure origin of ductility in refractory RMPEAs. This proposed metric enables rapid and accurate assessment of ductility behavior in the vast RMPEA composition space.

cond-mat.mtrl-sci↗

arXiv2023

High-throughput Alloy and Process Design for Metal Additive Manufacturing

Designing alloys for additive manufacturing (AM) presents significant opportunities. Still, the chemical composition and processing conditions required for printability (ie., their suitability for fabrication via AM) are challenging to explore using solely experimental means. In this work, we develop a high-throughput (HTP) computational framework to guide the search for highly printable alloys and appropriate processing parameters. The framework uses material properties from state-of-the-art databases, processing parameters, and simulated melt pool profiles to predict process-induced defects, such as lack-of-fusion, keyholing, and balling. We accelerate the printability assessment using a deep learning surrogate for a thermal model, enabling a 1,000-fold acceleration in assessing the printability of a given alloy at no loss in accuracy when compared with conventional physics-based thermal models. We verify and validate the framework by constructing printability maps for the CoCrFeMnNi Cantor alloy system and comparing our predictions to an exhaustive 'in-house' database. The framework enables the systematic investigation of the printability of a wide range of alloys in the broader Co-Cr-Fe-Mn-Ni HEA system. We identified the most promising alloys that were suitable for high-temperature applications and had the narrowest solidification ranges, and that was the least susceptible to balling, hot-cracking, and the formation of macroscopic printing defects. A new metric for the global printability of an alloy is constructed and is further used for the ranking of candidate alloys. The proposed framework is expected to be integrated into ICME approaches to accelerate the discovery and optimization of novel high-performance, printable alloys.

cond-mat.mtrl-sci↗

arXiv2023

An Automated Fully-Computational Framework to Construct Printability Maps for Additively Manufactured Metal Alloys

In additive manufacturing, the optimal processing conditions need to be determined to fabricate porosity-free parts. For this purpose, the design space for an arbitrary alloy needs to be scoped and analyzed to identify the areas of defects for different laser power-scan speed combinations and can be visualized using a printability map. Constructing printability maps is typically a costly process due to the involvement of experiments, which restricts their application in high-throughput product design. To reduce the cost and effort of constructing printability maps, a fully computational framework is introduced in this work. The framework combines CALPHAD models and a reduced-order model to predict material properties. THen, an analytical thermal model, known as the Eagar-Tsai model, utilizes some of these materials' properties to calculate the melt pool geometry during the AM processes. In the end, printability maps are constructed using material properties, melt pool dimensions, and commonly used criteria for lack of fusion, balling, and keyholing defects. To validate the framework and its general application to laser powder-bed fusion alloys, five common additive manufacturing alloys are analyzed. Furthermore, NiTi-based alloys at three different compositions are evaluated to show the further extension of the framework to alloy systems at different compositions. The defect regions in these printability maps are validated with corresponding experimental observations to compare and benchmark the defect criteria and find the optimal criterion set with the maximum accuracy for each unique material composition. Furthermore, printability maps for NiTi that are obtained from our framework are used in conjunction with process maps resulting from a multi-model framework to guide the fabrication of defect-free additive manufactured parts with tailorable properties and performance.

cond-mat.mtrl-sci↗

arXiv2023

Efficient Propagation of Uncertainty via Reordering Monte Carlo Samples

Uncertainty analysis in the outcomes of model predictions is a key element in decision-based material design to establish confidence in the models and evaluate the fidelity of models. Uncertainty Propagation (UP) is a technique to determine model output uncertainties based on the uncertainty in its input variables. The most common and simplest approach to propagate the uncertainty from a model inputs to its outputs is by feeding a large number of samples to the model, known as Monte Carlo (MC) simulation which requires exhaustive sampling from the input variable distributions. However, MC simulations are impractical when models are computationally expensive. In this work, we investigate the hypothesis that while all samples are useful on average, some samples must be more useful than others. Thus, reordering MC samples and propagating more useful samples can lead to enhanced convergence in statistics of interest earlier and thus, reducing the computational burden of UP process. Here, we introduce a methodology to adaptively reorder MC samples and show how it results in reduction of computational expense of UP processes.

cs.LG↗

arXiv2023

An Interpretable Boosting-based Predictive Model for Transformation Temperatures of Shape Memory Alloys

In this study, we demonstrate how the incorporation of appropriate feature engineering together with the selection of a Machine Learning (ML) algorithm that best suits the available dataset, leads to the development of a predictive model for transformation temperatures that can be applied to a wide range of shape memory alloys. We develop a gradient boosting ML surrogate model capable of predicting Martensite Start, Martensite Finish, Austenite Start, and Austenite Finish transformation temperatures with an average accuracy of more than 95% by explicitly taking care of potential distribution changes when modeling different alloy systems. We included heat treatment, rolling, extrusion processing parameters, and alloy system categorical features in the model input features to achieve more accurate and realistic results. In addition, using Shapley values, which are calculated based on the average marginal contribution of features to all possible coalitions, this study was able to gain insights into the governing features and their effect on predicted transformation temperatures, providing a unique opportunity to examine the critical parameters and features in martensite transformation temperatures.

cond-mat.mtrl-sci↗

arXiv2022

Machine-learning enabled thermodynamic model for the design of new rare-earth compounds

We employ a descriptor based machine-learning approach to assess the effect of chemical alloying on formation-enthalpy of rare-earth intermetallics. Application of machine-learning approaches in rare-earth intermetallic design have been sparse due to limited availability of reliable datasets. In this work, we developed an `in-house' rare-earth database with more than 600$+$ compounds, each entry was populated with formation enthalpy and related atomic features using high-throughput density-functional theory (DFT). The SISSO (sure independence screening and sparsifying operator) based machine-learning method with meaningful atomic features was used for training and testing the formation enthalpies of rare earth compounds. The complex lattice function coupled with the machine-learning model was used to explore the effect of transition metal alloying on the energy stability of Ce based cubic Laves phases (MgCu$_{2}$ type). The SISSO predictions show good agreement with high-fidelity DFT calculations and X$-$ray powder diffraction measurements. Our study provides quantitative guidance for compositional considerations within a machine-learning model and discovering new metastable materials. The electronic-structure of Ce$-$Fe$-$Cu based compound was also analyzed in$-$depth to understand the electronic origin of phase stability. The interpretable analytical models in combination with density$-$functional theory and experiments provide a fast and reliable design guide for discovering technologically useful materials.

cond-mat.mtrl-sci↗

arXiv2021

Towards Stacking Fault Energy Engineering in FCC High Entropy Alloys

Stacking Fault Energy (SFE) is an intrinsic alloy property that governs much of the plastic deformation mechanisms observed in fcc alloys. While SFE has been recognized for many years as a key intrinsic mechanical property, its inference via experimental observations or prediction using, for example, computationally intensive first-principles methods is challenging. This difficulty precludes the explicit use of SFE as an alloy design parameter. In this work, we combine DFT calculations (with necessary configurational averaging), machine-learning (ML) and physics-based models to predict the SFE in the fcc CoCrFeMnNiV-Al high-entropy alloy space. The best-performing ML model is capable of accurately predicting the SFE of arbitrary compositions within this 7-element system. This efficient model along with a recently developed model to estimate intrinsic strength of fcc HEAs is used to explore the strength-SFE Pareto front, predicting new-candidate alloys with particularly interesting mechanical behavior.

cond-mat.mtrl-sci↗

arXiv2021

Pseudoelastic deformation in Mo-based refractory multi-principal element alloys

Phase diagrams supported by density functional theory methods can be crucial for designing high-entropy alloys that are subset of multi-principal$-$element alloys. We present phase and property analysis of quinary (MoW)$_{x}$Zr$_{y}$(TaTi)$_{1-x-y}$ refractory high-entropy alloys from combined Calculation of Phase Diagram (CALPHAD) and density-functional theory results, supplemented by molecular dynamics simulations. Both CALPHAD and density-functional theory analysis of phase stability indicates a Mo-W-rich region of this quinary has a stable single-phase body-centered-cubic structure. We report first quinary composition from Mo$-$W$-$Ta$-$Ti$-$Zr family of alloy with pseudo-elastic behavior, i.e., hysteresis in stress$-$strain. Our analysis shows that only Mo$-$W$-$rich compositions of Mo$-$W$-$Ta$-$Ti$-$Zr, i.e., Mo$+$W$\ge$ 85 at.%, show reproducible hysteresis in stress-strain responsible for pseudo-elastic behavior. The (MoW)$_{85}$Zr$_{7.5}$(TaTi)$_{7.5}$ was down-selected based on temperature-dependent phase diagram analysis and molecular dynamics simulations predicted elastic behavior that reveals twinning assisted pseudoelastic behavior. While mostly unexplored in body-centered-cubic crystals, twinning is a fundamental deformation mechanism that competes against dislocation slip in crystalline solids. This alloy shows identical cyclic deformation characteristics during uniaxial $\lt$100$\gt$ loading, i.e., the pseudoelasticity is isotropic in loading direction. Additionally, a temperature increase from 77 to 1500 K enhances the elastic strain recovery in load-unload cycles, offering possibly control to tune the pseudoelastic behavior.

cond-mat.mtrl-sci↗

arXiv2021

On the martensitic transformation in Fe$_{x}$Mn$_{80-x}$Co$_{10}$Cr$_{10}$ high-entropy alloy

High-entropy alloys (HEAs), and even medium-entropy alloys (MEAs), are an intriguing class of materials in that structure and property relations can be controlled via alloying and chemical disorder over wide ranges in the composition space. Employing density-functional theory combined with the coherent-potential approximation to average over all chemical configurations, we tune free energies between face-centered-cubic (fcc) and hexagonal-close-packed (hcp) phases in Fe$_{x}$Mn$_{80-x}$Co$_{10}$Cr$_{10}$ systems.~Within Fe-Mn-based alloys, we show that the martensitic transformation and chemical short-range order directly correlate with the fcc-hcp energy difference and stacking-fault energies, which are in quantitative agreement with recent experiments on a $x$=40~at.\% polycrystalline HEA/MEA. Our predictions are further confirmed by single-crystal measurements on a$x$=40at.\% using transmission-electron microscopy, selective-area diffraction, and electron-backscattered-diffraction mapping. The results herein offer an understanding of transformation-induced/twinning-induced plasticity (TRIP/TWIP) in this class of HEAs and a design guide for controlling the physics behind the TRIP effect at the electronic level.

cond-mat.mtrl-sci↗

arXiv2020

Deep Multimodal Transfer-Learned Regression in Data-Poor Domains

In many real-world applications of deep learning, estimation of a target may rely on various types of input data modes, such as audio-video, image-text, etc. This task can be further complicated by a lack of sufficient data. Here we propose a Deep Multimodal Transfer-Learned Regressor (DMTL-R) for multimodal learning of image and feature data in a deep regression architecture effective at predicting target parameters in data-poor domains. Our model is capable of fine-tuning a given set of pre-trained CNN weights on a small amount of training image data, while simultaneously conditioning on feature information from a complimentary data mode during network training, yielding more accurate single-target or multi-target regression than can be achieved using the images or the features alone. We present results using phase-field simulation microstructure images with an accompanying set of physical features, using pre-trained weights from various well-known CNN architectures, which demonstrate the efficacy of the proposed multimodal approach.

cs.CV↗

arXiv2020

Impact of Particle Arrays on Phase Separation Composition Patterns

We examine the symmetry-breaking effect of fixed constellations of particles on the surface-directed spinodal decomposition of binary blends in the presence of particles whose surfaces have a preferential affinity for one of the components. Our phase-field simulations indicate that the phase separation morphology in the presence of particle arrays can be tuned to have a continuous, droplet, lamellar, or hybrid morphology depending on the interparticle spacing, blend composition, and time. In particular, when the interparticle spacing is large compared to the spinodal wavelength, a transient target pattern composed of alternate rings of preferred and non-preferred phases emerge at early times, tending to adopt the symmetry of the particle configuration. We reveal that such target patterns stabilize for certain characteristic length, time, and composition scales characteristic of the pure phase separating mixture. To illustrate the general range of phenomena exhibited by mixture-particle systems, we simulate the effects of single-particle, multi-particle, and cluster-particle systems having multiple geometrical configurations of the particle characteristic of pattern substrates on phase separation. Our simulations show that tailoring the particle configuration, or substrate pattern configuration, a relative fluid-particle composition should allow the desirable control of the phase separation morphology as in block copolymer materials, but where the scales accessible to this approach of organizing phase-separated fluids usually are significantly larger. Limited experiments confirm the trends observed in our simulations, which should provide some guidance in engineering patterned blend and other mixtures of technological interest.

physics.chem-ph↗

arXiv2020

Finite Interface Dissipation Phase Field Modeling of Ni-Nb under Additive Manufacturing Conditions

During the laser powder bed fusion (L-PBF) process, the built part undergoes multiple rapid heating-cooling cycles, leading to complex microstructures with nonuniform properties. In the present work, a computational framework, which weakly couples a finite element thermal model to a non-equilibrium PF model was developed to investigate the rapid solidification microstructure of a Ni-Nb alloy during L-PBF. The framework is utilized to predict the spatial variation of the morphology and size of cellular segregation structure as well as the microsegregation in single-track melt pool microstructures obtained under different process conditions. A solidification map demonstrating the variation of microstructural features as a function of the temperature gradient and growth rate is presented. A planar to cellular transition is predicted in the majority of keyhole mode melt pools, while a planar interface is predominant in conduction mode melt pools. The predicted morphology and size of the cellular segregation structure agrees well with experimental measurements.

physics.app-ph↗

arXiv2019

The Effect of Chemical Disorder on Defect Formation and Migration in Disordered MAX Phases

MAX phases have attracted increased attention due to their unique combination of ceramic and metallic properties. Point-defects are known to play a vital role in the structural, electronic and transport properties of alloys in general and this system in particular. As some MAX phases have been shown to be stable in non-stoichiometric compositions, it is likely that such alloying effects will affect the behavior of lattice point defects. This problem, however, remains relatively unexplored. In this work, we investigate the alloying effect on the structural-stability, energy-stability, electronic-structure, and diffusion barrier of point defects in MAX phase alloys within a first-principles density functional theory framework. The vacancy (V$_{M}$, V$_{A}$, V$_{X}$) and antisite (M-A; M-X) defects are considered with M and A site disorder in (Zr-M)$_{2}$(AA${'}$)C, where M=Cr,Nb,Ti and AA${'}$=Al, Al-Sn, Pb-Bi. Our calculations suggest that the chemical disorder helps lower the V$_{A}$ formation energies compared to V$_{M}$ and V$_{X}$. The V$_{A}$ diffusion barrier is also significantly reduced for M-site disorder compared to their ordered counterpart. This is very important finding because reduced barrier height will ease the Al diffusion at high-operating temperatures, which will help the formation of passivating oxide layer (i.e., Al$_{2}$O$_{3}$ in aluminum-based MAX phases) and will slow down or stop the material degradation. We believe that our study will provide a fundamental understanding and an approach to tailor the key properties that can lead to the discovery of new MAX phases.

cond-mat.mtrl-sci↗