Concurrent Atomistic-Continuum Studies of Interface/Dislocation Interactions in Nanolaminates and with He bubbles in Stainless Steels

(Georgia Institute of Technology, 2022-11-21) Selimov, Alex

The interactions between dislocations and interfaces between materials with different structures, chemistry, and/or orientation primarily mediate the mechanical properties of polycrystalline materials. Studying these interactions necessitates the consideration of non-ideal interface structures which can arise through interdiffusion of constituent species, the presence of defects such as voids or He bubbles, and the build-up of dislocation content from previous interactions. Analysis of these interface structures requires atomic resolution to capture due to the complex dislocation reactions that are not known a priori. Computational cost precludes application of atomistic methods for interfaces that have large characteristic lengths or for studying the accumulation of dislocation content from sequential interactions, both of which require large spatial domains. Instead, we desire a multi-scale modeling approach which renders the interface at full atomistic resolution while considering the bulk of the model with a lower cost description. The Concurrent Atomistic-Continuum (CAC) method is a coarse-graining atomistics approach that utilizes full atomistic resolution at locations with high degrees of atomic restructuring to preserve predictive accuracy, while using a coarse-grained description elsewhere to reduce degrees of freedom. CAC utilizes a unified model form for both coarse-grained and atomistic regions that depends only on the interatomic potential; an integral formulation finite element approach in CAC enables use of a discontinuous mesh which can accommodate dislocations and does not require bridging methods to join coarse-grained and atomistic regions. CAC will be applied to the study of the deformation of nanolaminate materials with semi-coherent interfaces to understand interface evolution as a result of the glide of threading dislocations. The relation between interface misfit dislocation density and resistance of the interface to dislocation glide will specifically be compared. CAC will also be used to study the sequential interactions of dislocations with He bubbles embedded within Σ3 grain boundaries in stainless steels. The He bubble/grain boundary dislocation reactions will provide insight into irradiation induced embrittlement. This dissertation will characterize mechanisms of dislocation/interface interactions and will highlight the effects of evolving interface structure on the interaction mechanisms while demonstrating the necessity of multi-scale modeling schemes to accurately estimate evolution of dislocations and interface structure and properties.
ATOMISTIC AND COARSE-GRAINED ATOMISTIC MODELING OF SOLUTE ORDERING EFFECTS ON DISLOCATION MIGRATION

(Georgia Institute of Technology, 2022-07-21) Chu, Kevin

Dislocation-mediated plasticity in alloy systems is strongly affected by the concentration and distribution of solute atoms. These effects manifest as changes in the strength or ductility of the alloy. Such microstructural variables are especially relevant in state-of-the-art alloy systems such as high entropy alloys or alloys produced by additive manufacturing (AM). The temperature and stress gradients induced during processing lead to solute configurations that can no longer be considered random. In 316L stainless steel, the AM process leads to the development of sub-grain solute segregation which in turn leads to dislocation cellular structures that simultaneously enhance strength and ductility. At certain target composition ranges in this system, heat treatment procedures lead to chemical short-range ordering (CSRO) and solute clustering which affect strength retention at high temperature. Targeted design of alloys for such improved mechanical properties requires a fundamental understanding of these dislocation-solute interactions at the nanoscale. Computational modeling is one such strategy that offers insight into the atomic scale mechanisms and can screen large parameter spaces more quickly than experiments. This work begins by building a bottom-up understanding of the solute composition and temperature dependent dislocation mobility in 316L from molecular dynamics (MD), equipping reduced-order modeling approaches with the ability to directly simulate dislocation cell structure formation. Then, a physics-based analytical model is developed and validated to predict the effects of varying CSRO on yield strength. While powerful as modeling approach, the computational cost of MD can become quite demanding when exploring multidimensional parameter spaces. An application of average-atom interatomic potentials to the Concurrent Atomistic-Continuum (CAC) coarse-grained atomistics method is presented that enables study of multicomponent alloy systems at reduced computational cost while preserving key nanoscale mechanisms. Finally, a coarse-grained CAC implementation of the nudged elastic band (NEB) method is validated as length and time scale bridging approach, again extending accessible parameter ranges through coarse graining while retaining a full atomistic description of the reaction pathway.
Probing Interfacial Dynamics in Solid-State Lithium Metal Batteries

(Georgia Institute of Technology, 2022-05-03) Lewis, John A.

Solid-state batteries (SSBs) are a promising technology to surpass the energy density and safety of conventional lithium-ion batteries. These devices replace the flammable liquid electrolyte with a more stable solid-state electrolyte (SSE) that can conduct lithium ions. The rigid mechanical properties of SSEs are also promising for enabling the energy dense lithium metal anode, which is plagued by dendrite formation and dead lithium in liquid electrolytes. Despite advances towards SSEs with high ionic conductivity, the understanding and control over solid electrode/SSE interfaces have emerged as major challenges in the development of SSBs. Chemo-mechanical degradation is expected to be more severe in SSBs compared to conventional liquid-electrolyte-batteries because the SSE cannot reconfigure like liquids. Understanding chemical transformations at interfaces, mechanical damage, and lithium filament growth is therefore critical for engineering SSBs. This dissertation investigated the underlying mechanisms of these interfacial phenomenon to better inform the design of SSBs. First, severe SSE decomposition caused by electrochemical side reactions with lithium metal was found when the reacted species exhibited mixed ionic-electronic conduction. Continuous decomposition ultimately resulted in fracture due to the build-up of internal stress, and this process was accelerated when operating at higher rates. Second, the dynamic evolution of Li/SSE interfaces was probed using operando X-ray tomography. 3D images of SSBs were obtained during operation, which were then processed using segmentation to quantify how phases change and link their behavior directly to the measured electrochemistry. Analysis revealed that the significant loss of interfacial contact was responsible for cell failure. Third, the relationships between unstable Li metal deposition and electrochemical parameters, such as current density and areal capacity, were investigated. A new metric called the threshold capacity was introduced and used to evaluate lithium deposition behavior in SSBs. Cycling of cells with areal capacity controlled to be well below the threshold capacity greatly improved cell lifetime, while approaching the threshold capacity resulted in rapid short circuiting. Fourth, the mechanisms of anode-free SSBs were investigated, in which lithium metal was deposited onto a bare copper current collector during the first charge. Stripping lithium from the copper current collector was found to cause significant degradation that severely limited cycling lifetime. Lastly, the energy densities of various battery chemistries were calculated at the cell-level. Alloy anodes in SSBs were shown to have competitive energy densities, and their mechanistic advantages over alloy anodes with liquid electrolytes and lithium metal SSBs are discussed.
Multiscale Modeling of Hydrogen Embrittlement

(Georgia Institute of Technology, 2022-04-04) Zirkle, Theodore

Hydrogen embrittlement is a long-standing issue in materials science and engineering with a multitude of competing hypotheses and theories. Despite advances in experimental and computational capabilities, common understanding of contributing phenomena has not yet been achieved. Hence, a more complete understanding of hydrogen embrittlement processes operating at multiple length and time scales is still an open challenge that justifies the current research. In this thesis, a unique approach is taken to incorporate a wide range of experimental, computational, and analytical approaches across multiple length scales to produce a mechanistically motivated hydrogen embrittlement model for fracture and fatigue. This research describes and simulates the complex interplay between hydrogen, hydrogen-related defects, dislocations, and dislocation substructures. The model is developed in a crystal plasticity context and implemented in a finite element framework to simulate the hydrogen embrittlement of austenitic stainless steels, structural materials important in energy applications. The proposed research extends current understanding through the development of: i. a physically-based crystal plasticity model developed to capture the evolution of dislocation substructure and material behavior during cyclic loading, ii. a hydrogen transport and trapping model that considers dislocation-mediated transport mechanisms and a more complete set of hydrogen traps, and iii. a fully coupled chemo-mechanical model to capture the effects of hydrogen in reducing crack tip ductility, leading to embrittlement effects.
HIGH THROUGHPUT MECHANICAL PROPERTY CHARACTERIZATION OF STRUCTURAL ALLOYS USING SPHERICAL MICROINDENTATION

(Georgia Institute of Technology, 2021-07-27) Bhat, Anirudh Srinivas

Spherical indentation has been shown to be a reliable high throughput alternative to capture the complete elastic-plastic response of polycrystalline metal alloys when using the Pathak-Kalidindi (P-K) protocol. However, yielding initiates subsurface and due to the hydrostatic pressure associated with the constraint of the surrounding material, this occurs at a higher stress than it would for the same material under uniaxial load. Hence, the indentation stress is higher than uniaxial stress and the two are related by a scaling factor, referred to as the constraint factor. In the current work, some of the open questions in the use of spherical indentation to extract uniaxial stress-strain curves are addressed. The dependence of the mechanical properties of the material and the indentation strain on the constraint factor is investigated using FEA and experiments. Based on the P-K indentation strain definition, revisions are proposed to the classical equations for: 1) the representative uniaxial strain, which relates the indentation strain to an equivalent uniaxial strain and 2) the non-dimensional strain, which relates the indentation strain to the constraint factor. From these revised definitions, an inverse method has been developed using FEA simulations to estimate the uniaxial stress-strain from spherical indentation stress-strain curves. The inverse method is verified with FEA generated indentation stress-strain curves and experiments conducted on Al7050 and Al6061 samples of varying strengths and hardening behaviors. Finally, spherical indentation of a material exhibiting anisotropic mechanical response is studied. The uniaxial and spherical indentation mechanical response of samples of Inconel 718 that were additively manufactured using electron beam melting are critically evaluated. The variation of the experimentally obtained elastic spherical indentation response as a function of the orientation of the material is verified using FEA simulations.
A Large Scale Computational Study of Fatigue Hot-Spots

(Georgia Institute of Technology, 2021-04-26) Muth, Adrienne

Formation of a fatigue crack at the subgrain scale is a statistically rare event, as plastic deformation at the microscale ranges from highly heterogeneous at low strain to homogeneous at high strain. Fatigue Indicator Parameters (FIPs) for Ti-6Al-4V are computed using crystal plasticity finite element modeling of uniaxial cyclic straining of ensembles of statistical volume elements for a range of distinct microstructures at several strain amplitudes and mean strain conditions. The selection of FIPs is informed by prior experimental studies. The sites of extreme value (EV) FIPs that are most likely to form and grow a fatigue crack are identified in these simulations, and 2-point spatial correlations are applied to investigate the higher dimensional influence of microstructure attributes in the neighborhood of these fatigue hot-spots. To reduce the high dimensionality of the associated 2-point correlations, principal component analysis is applied. A reduced-order model using an artificial neural network is used to classify EV FIP locations based on these neighborhood spatial correlations.
Uncertainty Informed Integrated Computational Materials Engineering for Design and Development of Fatigue Critical Alloys

(Georgia Institute of Technology, 2020-12-06) Whelan, Gary Francis

Uncertainty is intrinsically tied to decision-making in design. Process-Structure-Property (PSP) relations are central to development of new and improved materials. The multitude of PSP linkages for any performance objective can be explored using the top down, inductive design exploration method (IDEM). Each PS and SP linkage has associated uncertainty, arising both from the types of models or interpretation of experimental results used to form linkages, as well as model parameters. These uncertainties can propagate and significantly affect the decision-making process in design and development of materials for specific performance targets. Uncertainty quantification (UQ) can be a highly computationally expensive undertaking in materials design and development. In this research, computationally efficient protocols are developed to effectively incorporate UQ in the IDEM. The uncertainty associated with PS linkages is assigned based on existing literature results. Gaussian process (GP) surrogate models are developed for the various SP linkages of interest as lower order approximations of computational expensive computational materials science simulations (e.g., the crystal plasticity finite element method (CPFEM)). These GP models are used to propagate uncertainty in microstructure attributes to the quantities of interest associated with properties that are then optimized in design. These surrogate models are integrated into existing python IDEM (pyDEM) protocols in the form of mapping functions. In this work, novel protocols are developed and demonstrated for uncertainty-informed design and development of Ti-6Al-4V and Al7075-T6 microstructures for targeted performance requirements involving combinations of fatigue resistance, elastic stiffness, and yield strength.
Hydrogen effects on dislocation structures and interactions

(Georgia Institute of Technology, 2020-12-02) Costello, Luke L.

Hydrogen embrittlement (HE) is a complex process, in which the interactions of H atoms, vacancies, and dislocations lead to a macroscopic loss of ductility. Although this phenomenon is commonly observed, its microscopic origins remain unclear. In this thesis we study the process of HE, starting from the microscale, using atomistic simulations and modeling to connect with higher length scales. Passing information from physically realistic atomic scale simulations allows for improved understanding of the underlying mechanisms of H embrittlement and specifically how effects attributed to H contribute at the meso and macroscales. We study these contributions in three parts. First, a method for computing the distribution of H around an edge dislocation is presented and compared to an alternative approach. The presented method is then exercised in an example, studying the effect of H on the stacking fault width (SFW) of an extended edge dislocation. It is shown that H acts to decrease the SFW. Further, only the H very locally around and inside the dislocation cores and stacking fault contribute significantly to the observed decrease in the SFW. We then turn our focus to the role of H on the stabilization and clustering of vacancies. A new model is developed for the production of excess vacancies by plastic deformation. This model is integrated into an existing macroscale computational framework. Lastly, the role of H on the formation energy of vacancy clusters is studied using a hybrid molecular statics / Monte Carlo method. These calculations show that H tends to decrease the formation energy of vacancy clusters which leads to a decrease in the critical cluster size for void nucleation.
Hierarchical multiscale materials modeling: Calibration, uncertainty quantification, and decision support

(Georgia Institute of Technology, 2018-07-24) Tallman, Aaron E.

Computational material models help establish structure-property relationships by simulating properties, and are most effective when physically-based. The length and time scales of each simulation are constrained both by model type and computing power. Significant uncertainty can arise when models attempt to bridge across length and time scales, especially when using different model constructs. Hierarchical multiscale modeling (HMM) links models at different scales by informing parameters and form of higher scale models based on lower scale simulations, which can reduce uncertainty. The combination of diverse information sources in HMMs requires rigorous approaches to evaluate uncertainty propagation. In the pursuit of improved methods for empirical testing and development of model hierarchies, four approaches in which information is coordinated amongst multiple models are presented. (1) In a reconciled top-down and bottom-up approach, a likelihood-based model calibration method is proposed, and bcc Fe crystal plasticity (CP) is used to demonstrate the compatibility of information pathways. (2) A statistical volume element (SVE) ensemble-based homogenization scheme of two models of cartridge brass polycrystal plasticity is used to inform a Bammann-Chiesa-Johnson macroplasticity model with a local variation in parameters. The effects of SVE size and model form on the performance of the homogenization in bridging microstructure variability to macroscale uncertainty are explored. (3) A multiscale model development framework is outlined for the reduced order modeling of mesoscale variability in cartridge brass. The variability in SVE simulations is included with the results of a series of spherical microindentation experiments in a multiscale data collection. An initial study of the modeling involved in connecting the two length scales is performed. (4) In a CP-finite element method (FEM) based Materials Knowledge System model of -Ti, the influence of texture is considered. Texture is parameterized using generalized spherical harmonics. The CP-FEM model is used with polycrystalline SVE-ensembles to calibrate the MKS model across different textures, sampled according to an uncertainty reduction criterion. Results of the work suggest that data collection is an especially critical step in the formulation and deployment of hierarchical multiscale models. The use of bottom-up information in calibrating a multiscale model is shown to be susceptible to bias. A multiscale approach to coarse-grained simulations of polycrystals at the mesoscale is proposed. An approach to automating the data collection for a reduced-order model of microstructure sensitive response is shown to be competitive with manual data selection, prior to full optimization of the automated approach.
Effect of mesoscale inhomogeneities on planar shock response of materials

(Georgia Institute of Technology, 2017-04-28) Ferri, Brian Anthony

In all previous spall models, the source of spall failure in metals either comes from damage at the grain boundary or from void nucleation, growth, and coalescence. However, it has been observed in experiments that both phenomena occur in Aluminum 6061-T6, which is termed “combined failure” for the purposes of this thesis. Thus, the challenge undertaken in this thesis is to use a computational study to determine the role that each source of spall plays separately, and then in tandem to determine the traditional failure parameters for each source. The results of determining each failure model’s ideal parameters, which are representative of that source’s role in combined failure, is compared with data gathered from plate-flyer experiments to determine the accuracy of the model in both 1D and in 2D simulations. Sand is a heterogeneous granular material that has the capability of allowing a shock wave to propagate through it. The computational model and study presented in this thesis is phenomenologically similar, yet easier to conduct than a spall study on granular Aluminum. The study of sand using the same computational LS-DYNA method shows both an introduction to the process for completing the spall study on granular Aluminum, and it also yields interesting results in the wave phenomena as well as the effect of porosity on the average stress on the sand grains. With the conclusion of the sand study, the same process of creating the grain structure is applied to create the Aluminum grain structure for spall simulations, which are carried out in LS-DYNA using 2D cohesive elements. The results of the LS-DYNA Aluminum simulation are compared to both the 1D spall results as well as to the experimental data to determine model accuracy. The main findings from this thesis show that, first, a mutually exclusive combined failure linear relationship can be shown with the 1D simulation results, which gives insight into a method that could be used to choose a set of optimal failure parameters. Second, the 2D LS-DYNA homogeneous results had excellent agreement with the 1D homogeneous results, which gave confidence to the notion that the parametric studies in 1D simulations could be used to find parameter values that could be applied in the 2D models. Lastly, LS-DYNA was shown to be an effective way to simulate grain structure response to shock wave propagation and showed spall modeling was possible with 2D cohesive elements, which lays the groundwork for combined failure studies in 2D.