Ferroelectric ceramics are broadly used in applications including actuators, sensors and information storage. An understanding of the microstructual evolution and domain dynamics is vital for predicting the performance and reliability of such devices. The underlying mechanism responsible for ferroelectric constitutive response is ferroelectric domain wall motion, domain switching and the interactions of domain walls with other material defects. In this work, a combined theoretical and numerical modeling framework is developed to investigate the nucleation and growth of domains in a single crystal of ferroelectric material. The phase-field approach, applying the material electrical polarization as the order parameter, is used as the theoretical modeling framework to allow for a detailed accounting of the electromechanical processes. The finite element method is used for the numerical solution technique. In order to obtain a better understanding of the energetics of fracture within the phase-field setting, the J-integral is modified to include the energies associated with the order parameter. Also, the J- integral is applied to determine the crack-tip energy release rate for common sets of electromechanical crack-face boundary conditions. The calculations confirm that only true equilibrium states exhibit path-independence of J, and that domain structures near crack tips may be responsible for allowing positive energy release rate during purely electrical loading. The small deformation assumption is prevalent in the phase-field modeling approach, and is used in the previously described calculations. The analysis of large deformations will introduce the concept of Maxwell stresses, which are assumed to be higher order effects that can be neglected in the small deformation theory. However, in order to investigate the material response of soft dielectric elastomers undergoing large mechanical deformation and electric field, which are employed in electrically driven actuator devices, manipulators and energy harvesters, a finite deformation theory is incorporated in the phase-field model. To describe the material free energy, compressible Neo-Hookean and Gent models are used. The Jaumann rate of the polarization is used as the objective polarization rate to make the description of the dissipation frame indifferent. To illustrate the theory, electromechanical instabilities in composite materials with different inclusions will be studied using the finite element methods. / text

Indiana University-Purdue University Indianapolis (IUPUI) / Sintering is a primary particulate manufacturing technology to provide densification and strength for ceramics and many metals. A persistent problem in this manufacturing technology has been to maintain the quality of the manufactured parts. This can be attributed to the various sources of uncertainty present during the manufacturing process. In this work, a two-particle phase-field model has been analyzed which simulates microstructure evolution during the solid-state sintering process. The sources of uncertainty have been considered as the two input parameters surface diffusivity and inter-particle distance. The response quantity of interest (QOI) has been selected as the size of the neck region that develops between the two particles. Two different cases with equal and unequal sized particles were studied. It was observed that the neck size increased with increasing surface diffusivity and decreased with increasing inter-particle distance irrespective of particle size. Sensitivity analysis found that the inter-particle distance has more influence on variation in neck size than that of surface diffusivity. The machine-learning algorithm Gaussian Process Regression was used to create the surrogate model of the QOI. Bayesian Optimization method was used to find optimal values of the input parameters. For equal-sized particles, optimization using Probability of Improvement provided optimal values of surface diffusivity and inter-particle distance as 23.8268 and 40.0001, respectively. The Expected Improvement as an acquisition function gave optimal values 23.9874 and 40.7428, respectively. For unequal sized particles, optimal design values from Probability of Improvement were 23.9700 and 33.3005 for surface diffusivity and inter-particle distance, respectively, while those from Expected Improvement were 23.9893 and 33.9627. The optimization results from the two different acquisition functions seemed to be in good agreement with each other. The results also validated the fact that surface diffusivity should be higher and inter-particle distance should be lower for achieving larger neck size and better mechanical properties of the material.

Zirconia based ceramics are strong, hard, inert, and smooth, with low thermal conductivity and good biocompatibility. Such properties made zirconia ceramics an ideal material for different applications form thermal barrier coatings (TBCs) to biomedicine applications like femoral implants and dental bridges. However, this unusual versatility of excellent properties would be mediated by the metastable tetragonal (or cubic) transformation to the stable monoclinic phase after a certain exposure at service temperatures. This transformation from tetragonal to monoclinic, known as LTD (low temperature degradation) in biomedical application, proceeds by propagation of martensite, which corresponds to transformation twinning. As such, tetragonal to monoclinic transformation is highly sensitive to mechanical and chemomechanical stresses. It is known in fact that this transformation is the source of the fracture toughening in stabilized zirconia as it occurs at the stress concentration regions ahead of the crack tip. This dissertation is an attempt to provide a kinetic-based model for tetragonal to monoclinic transformation in zirconia. We used the phase field technique to capture the temporal and spatial evolution of monoclinic phase. In addition to morphological patterns, we were able to calculate the developed internal stresses during tetragonal to monoclinic transformation. The model was started form the two dimensional single crystal then was expanded to the two dimensional polycrystalline and finally to the three dimensional single crystal. The model is able to predict the most physical properties associated with tetragonal to monoclinic transformation in zirconia including: morphological patterns, transformation toughening, shape memory effect, pseudoelasticity, surface uplift, and variants impingement. The model was benched marked with several experimental works. The good agreements between simulation results and experimental data, make the model a reliable tool for predicting tetragonal to monoclinic transformation in the cases we lack experimental observations.

<p> This thesis details work done in the development of the phase field model which allows simulation of elasticity with diffuse interfaces and the extension of a thin interface analysis developed by previous authors to study non-dilute ideal alloys. These models are coupled with a new finite difference adaptive mesh algorithm to efficiently simulate a variety of physical systems. The finite difference adaptive mesh algorithm is shown to be at worse 4-5 times faster than an equivalent finite element method on a per node basis. In addition to this increase in speed for explicit solvers in the code, an iterative solver used to compute elastic fields is found to converge in O(N) time for a dynamically growing precipitate, where N is the number of nodes on the adaptive mesh. A previous phase field formulation is extended such as to make possible the study of non-ideal binary alloys with complex phase diagrams. A phase field model is also derived for a free energy that incorporates an elastic free energy and is used to investigate the competitive development of solid state structures in which the kinetic transfer rate of atoms from the parent phase to the precipitate phase is large. This results in the growth of solid state dendrites. The morphological effects of competing surface anisotropy and anisotropy in the elastic modulus tensor is analyzed. It is shown that the transition from surfaceenergy driven dendrites to elastically driven dendrites depends on the magnitudes of the surface energy anisotropy coefficient (E4 ) and the anisotropy of the elastic tensor (β) as well as on the super saturation of the particle and therefore to a specific Mullins-Sekerka onset radius. The transition point of this competitive process is predicted from these three controlling parameters. </p> / Thesis / Doctor of Philosophy (PhD)

The mechanical performance evaluation of asphalt binder has always been a challenging issue for pavement engineers. Recently, the Phase Field Method (PFM) has emerged as a powerful computational tool to simulate the microstructure evolution of asphalt binder. PFM analyzes the structure from the free energy aspect and can provide a view of the whole microstructure evolution process. In this dissertation, asphalt binder performance is analyzed by PFM in three aspects: first, the relationship between asphalt chemistry and performance is investigated. The components of asphalt are simplified to three: asphaltene, resin and oil. Simulation results show that phase separation will occur under certain thermal conditions and result in an uneven distribution of residual thermal stress. Second, asphalt cracking is analyzed by PFM. The traditional approach to analyze crack propagation is Classic Fracture Mechanics first proposed by Griffith, which needs to clearly depict the crack front conditions and may cause complex cracking topologies. PFM describes the microstructure using a phase-field variable which assumes positive one in the intact solid and negative one in the crack void. The fracture toughness is modeled as the surface energy stored in the diffuse interface between the intact solid and crack void. To account for the growth of cracks, a non-conserved Allen-Cahn equation is adopted to evolve the phase-field variable. The energy based formulation of the phase-field method handles the competition between the growth of surface energy and release of elastic energy in a natural way: the crack propagation is a result of the energy minimization in the direction of the steepest descent. Both the linear elasticity and phase-field equation are solved in a unified finite element frame work, which is implemented in the commercial software COMSOL. Different crack mode simulations are performed for validation. It was discovered that the onset of crack propagation agrees very well with the Griffith criterion and experimental results. Third, asphalt self-healing phenomenon is studied based on the Atomic Force Microscopy (AFM) technology. The self-healing mechanism is simulated in two ways: thermodynamic approach and mechanical approach. Cahn-Hilliard dynamics and Allen-Cahn dynamics are adopted, respectively. / Ph. D.

The size and morphology of the dendrite growth patterns are simulated for laser powder deposition of Ti-Nb alloys under steady-state and transient growth conditions. A phase field model using an adaptive grid technique was employed to simulate the steady-state growth of dendrites on rather small domains, in which fixed local solidification conditions are present. For simulation of dendrite growth patterns at transient conditions, a cellular automaton model was used along with a virtual front tracking technique on larger domains, containing various initial orientations of the solid-liquid (SL) interface. To obtain the required input thermal data, i.e., the temporal distribution of temperature, a finite element analysis was performed along with a novel numerical approach for the real-time addition of new deposition material in each time step, thus building the deposition geometry momentarily. Using the output of the thermal model, the motion and morphology of the SL interface was determined through tracking the isotherm of the solidification temperature. First, in this study, the appropriate set of processing parameters was found through an optimization process using a new concept, laser supplied energy Es, which combines the effects of the energy and powder density in the process. With the developed analytical/experimental procedure, crack and pore-free coatings of Ti-Nb with continuous beads were produced by examining the effects of a few sets of processing parameters, including laser power, laser scan velocity, laser beam diameter and powder feed rate. The results of the thermal model for the optimized set of parameters matched with the thermocouple temperature measurements with only ~5% deviation. The thermal model was able to predict realistic profiles for the temporal development of deposition geometry, thus predicting meaningful morphologies of the SL interface. The model output was easily treated for extraction of local processing parameters, such as the temperature gradient and solidification velocity. These data are very useful when simulating the dendrite growth patterns at steady-state conditions in directional solidification of selected regions in the microstructure. In order to define transient growth conditions, the simulated distribution of temperature can be also directly fed into the microstructure model at each solution time step. Phase field simulations of steady-state growth of dendrites during directional solidification showed a remarkable agreement with the experimental observations for the local dendrite arm spacing across the microstructure. Also qualitatively agreeing with the experiment, the simulated dendrite spacing exhibited a minimum around the mid-height region of the microstructure, which is explained by the counter effect of the temperature gradient and solidification velocity along the height of the sample. On a large domain containing different initial orientations of the SL interface, cellular automaton simulations for transient growth patterns of dendrites could reproduce most qualitative features observed in the microstructure. The dendrite arm spacing gradually decreased from the top of the microstructure. The competition was won by the dendrites growing in areas with higher cooling rates, i.e., in the regions closer to the top of the microstructure. The secondary arms of the primary dendrites, which are initially inclined on the vertical axis, grew extensively only along the overall growth direction and eventually became primary arms in some cases.

Homogenization of alloying elements is desired during sintering of powder metallurgy components. The redistribution processes such as penetration of liquid phase into the interparticle/grain boundaries of solid particles and subsequent solid-state  diffusion of alloy element(s) in the base powder, are important for the effective homogenization of alloy element(s) during liquid phase sintering of the mixed powders. The aim of this study is to increase the understanding of alloy element redistribution processes and their effect on the dimensional properties of the compact by means of numerical and experimental techniques. The phase field model coupled with Navier-Stokes equations is used for the simulations of dynamic wetting of millimeter- and micrometer-sized metal drops and liquid phase penetration into interparticle boundaries. The simulations of solid particle rearrangement under the action of capillary forces exerted by the liquid phase are carried out by using the equilibrium equation for a linear elastic material. Thermodynamic and kinetic calculations are performed to predict the phase diagram and the diffusion distances respectively. The test materials used for the experimental studies are three different powder mixes; Fe-2%Cu, Fe-2%Cu-0.5%C, and Fe-2%(Cu-2%Ni-1.5%Si)-0.5%C. Light optical microscopy, energy dispersive X-ray spectroscopy and dilatometry are used to study the microstructure, kinetics of the liquid phase penetration, solid-state diffusion of the Cu, and the dimensional changes during sintering. The wetting simulations are verified by matching the spreading experiments of millimeter-sized metal drops and it is observed that wetting kinetics is much faster for a micrometer-sized drop compared to the millimeter-sized drop. The simulations predicted the liquid phase penetration kinetics and the motion of solid particles during the primary rearrangement stage of liquid phase sintering in agreement with the analytical model. Microscopy revealed that the C addition delayed the penetration of the Cu rich liquid phase into interparticle/grain boundaries of Fe particles, especially into the grain boundaries of large Fe particles, and consequently the Cu diffusion in Fe is also delayed. We propose that the relatively lower magnitude of the sudden volumetric expansion in the master alloy system could be due to the continuous melting of liquid forming master alloy particles. / <p>QC 20140515</p>

The size and morphology of the dendrite growth patterns are simulated for laser powder deposition of Ti-Nb alloys under steady-state and transient growth conditions. A phase field model using an adaptive grid technique was employed to simulate the steady-state growth of dendrites on rather small domains, in which fixed local solidification conditions are present. For simulation of dendrite growth patterns at transient conditions, a cellular automaton model was used along with a virtual front tracking technique on larger domains, containing various initial orientations of the solid-liquid (SL) interface. To obtain the required input thermal data, i.e., the temporal distribution of temperature, a finite element analysis was performed along with a novel numerical approach for the real-time addition of new deposition material in each time step, thus building the deposition geometry momentarily. Using the output of the thermal model, the motion and morphology of the SL interface was determined through tracking the isotherm of the solidification temperature. First, in this study, the appropriate set of processing parameters was found through an optimization process using a new concept, laser supplied energy Es, which combines the effects of the energy and powder density in the process. With the developed analytical/experimental procedure, crack and pore-free coatings of Ti-Nb with continuous beads were produced by examining the effects of a few sets of processing parameters, including laser power, laser scan velocity, laser beam diameter and powder feed rate. The results of the thermal model for the optimized set of parameters matched with the thermocouple temperature measurements with only ~5% deviation. The thermal model was able to predict realistic profiles for the temporal development of deposition geometry, thus predicting meaningful morphologies of the SL interface. The model output was easily treated for extraction of local processing parameters, such as the temperature gradient and solidification velocity. These data are very useful when simulating the dendrite growth patterns at steady-state conditions in directional solidification of selected regions in the microstructure. In order to define transient growth conditions, the simulated distribution of temperature can be also directly fed into the microstructure model at each solution time step. Phase field simulations of steady-state growth of dendrites during directional solidification showed a remarkable agreement with the experimental observations for the local dendrite arm spacing across the microstructure. Also qualitatively agreeing with the experiment, the simulated dendrite spacing exhibited a minimum around the mid-height region of the microstructure, which is explained by the counter effect of the temperature gradient and solidification velocity along the height of the sample. On a large domain containing different initial orientations of the SL interface, cellular automaton simulations for transient growth patterns of dendrites could reproduce most qualitative features observed in the microstructure. The dendrite arm spacing gradually decreased from the top of the microstructure. The competition was won by the dendrites growing in areas with higher cooling rates, i.e., in the regions closer to the top of the microstructure. The secondary arms of the primary dendrites, which are initially inclined on the vertical axis, grew extensively only along the overall growth direction and eventually became primary arms in some cases.

Sintering is a primary particulate manufacturing technology to provide densification and strength for ceramics and many metals. A persistent problem in this manufacturing technology has been to maintain the quality of the manufactured parts. This can be attributed to the various sources of uncertainty present during the manufacturing process. In this work, a two-particle phase-field model has been analyzed which simulates microstructure evolution during the solid-state sintering process. The sources of uncertainty have been considered as the two input parameters surface diffusivity and inter-particle distance. The response quantity of interest (QOI) has been selected as the size of the neck region that develops between the two particles. Two different cases with equal and unequal sized particles were studied. It was observed that the neck size increased with increasing surface diffusivity and decreased with increasing inter-particle distance irrespective of particle size. Sensitivity analysis found that the inter-particle distance has more influence on variation in neck size than that of surface diffusivity. The machine-learning algorithm Gaussian Process Regression was used to create the surrogate model of the QOI. Bayesian Optimization method was used to find optimal values of the input parameters. For equal-sized particles, optimization using Probability of Improvement provided optimal values of surface diffusivity and inter-particle distance as 23.8268 and 40.0001, respectively. The Expected Improvement as an acquisition function gave optimal values 23.9874 and 40.7428, respectively. For unequal sized particles, optimal design values from Probability of Improvement were 23.9700 and 33.3005 for surface diffusivity and inter-particle distance, respectively, while those from Expected Improvement were 23.9893 and 33.9627. The optimization results from the two different acquisition functions seemed to be in good agreement with each other. The results also validated the fact that surface diffusivity should be higher and inter-particle distance should be lower for achieving larger neck size and better mechanical properties of the material.

This thesis develops a new quantitative multi-phase field model for polycrystalline solidification of binary alloys. We extend the thin interface formalism of Karma and co-workers to multiple order parameters. This makes it possible to model segregation and interface kinetics during equiaxed dendritic growth quantitatively, a feature presently lacking from polycrystalline or multi-phase solidification models. We study dendrite tip speed convergence as a function of interface width during free dendritic growth. We then analyze the steady state and grain coalescence properties of the model. It is shown that the model captures the correct physics of back diffusion and repulsive grain boundary coalescence as outlined by Rappaz and co-workers. Finally, the model is applied to simulate solidification and coarsening in delta-ferrite solidification. / Thesis / Master of Applied Science (MASc)