Table of contents

Molten salts have been proposed as heat carrier media in the nuclear and concentrating solar power plants. Due to their high melting temperature, solidification of the salts is expected to occur during routine and accidental scenarios. Furthermore, passive safety systems based on the solidification of these salts are being studied. The following article presents new developments in the modeling of eutectic molten salts by means of a multiphase, multicomponent, phase-field model. Besides, an application of this methodology for the eutectic solidification process of the ternary system LiF–KF–NaF is presented. The model predictions are compared with a newly developed semi-analytical solution for directional eutectic solidification at stable growth rate. A good qualitative agreement is obtained between the two approaches. The results obtained with the phase-field model are then used for calculating the homogenized properties of the solid phase distribution. These properties can then be included in a mixture macroscale model, more suitable for industrial applications.

Structure maps predict the crystal structure of a compound from the knowledge of constituent elements and chemical composition. We recently developed a highly predictive, three-dimensional structure map for stoichiometric binary sp-d-valent compounds. Here we show that the descriptors of this structure map are transferable to off-stoichiometric compounds with similar predictive power. We furthermore demonstrate that the descriptors are suitable for ternary prototypes. In particular, we construct a three-dimensional structure map for 129 prototypical crystal structures for ternary compounds. The crystal structure is predicted correctly with a probability of 78%. With a confidence of 95% the correct crystal structure is among the three most likely crystal structures predicted by the structure map.

In this paper we propose a novel approach to accelerate the multi-scale simulation of metal plasticity. In macroscopic zones of nearly homogeneous strain responses, the evolution of plastic anisotropy at each finite element integration point can be approximated from the properties at a representative point within a zone. We show how these zones can be identified by a clustering algorithm and can be utilised to reduce the computational cost of the simulation. We present and analyse the results obtained for two test cases.

In the present study, dynamic recrystallization (DRX) of IN718 superalloy was modeled based on the experimental data as well as modified cellular automaton (CA) methods. To verify the ability of the presented models for predicting the microstructural features, the electron backscattered diffraction technique was used. Results showed that both models are capable of estimating the mean grain size and volume fraction of DRX grains. Furthermore, the macroscopic flow behavior of the material, and texture evolution of the structure were simulated using the proposed CA model. In this model, the lattice rotation of grains was determined assuming that each grain obeys the rules of single crystal deformation, and the morphological texture was obtained using a topological module in the CA approach.

Magnesium-aluminum (Mg-Al) intermetallic compounds that form as precipitates can significantly influence the mechanical properties of Mg-Al alloys. A computational evaluation of known and unknown Mg-Al intermetallic compounds could help design new Mg-Al alloy microstructures with optimal properties. Here, we employ the cluster-expansion method with energies efficiently calculated with orbital-free density functional theory (OFDFT) and predict a new, metastable intermetallic compound Mg₃Al with a D0₁₉ hexagonal structure that is slightly more stable than an alternative L1₂ cubic structure. We apply Kohn-Sham DFT (KSDFT) to accurately evaluate various metastability criteria for D0₁₉ and L1₂ Mg₃Al, including Born's criterion and phonon dispersion. We show that both Mg₃Al crystalline phases satisfy the metastability criteria and hence should be at least metastable. We further compare ductility metrics for D0₁₉ and L1₂ Mg₃Al to that of hexagonal-close-packed Mg by computing Pugh's ratio and generalized stacking fault energies. The ductility is predicted to follow the order: D0₁₉ Mg₃Al > L1₂ Mg₃Al > Mg, based on the highest Pugh's ratio and the lowest unstable stacking and twinning fault energies of D0₁₉ Mg₃Al compared to that of Mg. We also predict a very low antiphase boundary energy for Mg₃Al and therefore expect D0₁₉ Mg₃Al to be beneficial for improving the ductility of Mg-rich Mg-Al alloys. A computational design of Mg-Al alloy microstructures may become possible by combining the strengths of both OFDFT and KSDFT, i.e., the efficiency of the former and the accuracy of the latter, as demonstrated here.

We propose a new scheme based on machine learning for the efficient screening in grain-boundary (GB) engineering. A set of results obtained from first-principles calculations based on density functional theory (DFT) for a small number of GB systems is used as a training data set. In our scheme, by partitioning the total energy into atomic energies using a local-energy analysis scheme, we can increase the training data set significantly. We use atomic radial distribution functions and additional structural features as atom descriptors to predict atomic energies and GB energies simultaneously using the least absolute shrinkage and selection operator, which is a recent standard regression technique in statistical machine learning. In the test study with fcc-Al [110] symmetric tilt GBs, we could achieve enough predictive accuracy to understand energy changes at and near GBs at a glance, even if we collected training data from only 10 GB systems. The present scheme can emulate time-consuming DFT calculations for large GB systems with negligible computational costs, and thus enable the fast screening of possible alternative GB systems.

Residual stresses develop in thin film interconnects mainly as a result of deposition conditions and multiple thermal loading cycles during the manufacturing flow. Understanding the relation between the distribution of residual stress and the interconnect microstructure is of key importance to manage the nucleation and growth of defects that can lead to failure under reliability testing and use conditions. Dislocation dynamics simulations are performed in nanocrystalline copper subjected to cyclic loading to quantify the distribution of residual stresses as a function of grain misorientation and grain size distribution. The outcomes of this work help to evaluate the effect of microstructure in thin films failure by identifying potential voiding sites. Furthermore, the simulations show how dislocation structures are influenced by texture and grain size distribution that affect the residual stress. For example, when dislocation loops reach the opposite grain boundary during loading, these dislocations remain locked during unloading.

The thermal stability of nanocrystalline Ni due to small additions of Mo or W (up to 1 at%) was investigated in computer simulations by means of a combined Monte Carlo (MC)/molecular dynamics (MD) two-steps approach. In the first step, energy-biased on-lattice MC revealed segregation of the alloying elements to grain boundaries. However, the condition for the thermodynamic stability of these nanocrystalline Ni alloys (zero grain boundary energy) was not fulfilled. Subsequently, MD simulations were carried out for up to 0.5 μs at 1000 K. At this temperature, grain growth was hindered for minimum global concentrations of 0.5 at% W and 0.7 at% Mo, thus preserving most of the nanocrystalline structure. This is in clear contrast to a pure Ni model system, for which the transformation into a monocrystal was observed in MD simulations within 0.2 μs at the same temperature. These results suggest that grain boundary segregation of low-soluble alloying elements in low-alloyed systems can produce high-temperature metastable nanocrystalline materials. MD simulations carried out at 1200 K for 1 at% Mo/W showed significant grain boundary migration accompanied by some degree of solute diffusion, thus providing additional evidence that solute drag mostly contributed to the nanostructure stability observed at lower temperature.

The first order phase transitions in binary alloys were simulated basing on the Cahn–Hilliard equation for metastable states with mobility depending on the local composition. The simulation was carried out utilizing the semi-implicit Fourier spectral method for 3D fragment of a solid solution satisfying the regular solution approximation. We defined kinetics of the main characteristics of phase distribution: nucleation rate, average size, concentration of precipitates and autocorrelation function etc. Peculiarities of different stages of binary alloy decomposition (nucleation, diffusion growth and coarsening) were analyzed both for constant and variable mobility.

Ti-6Al-4V is an alloy of titanium that dominates titanium usage in applications ranging from mass-produced consumer goods to high-end aerospace parts. The material's structure on a microscale is known to affect its mechanical properties but these effects are not fully understood. Specifically, this work will address the effects of low volume fraction intergranular β phase on Ti-6Al-4V's mechanical response during the transition from elastic to plastic deformation. A crystal plasticity-based finite element model is used to fully resolve the deformation of the β phase for the first time. This high fidelity model captures mechanisms difficult to access via experiments or lower fidelity models. The results are used to assess lower fidelity modeling assumptions and identify phenomena that have ramifications for failure of the material.

The behavior of primary static recrystallization (SRX) in a NiTiFe shape memory alloy (SMA) subjected to cold canning compression was investigated using the coupling crystal plasticity finite element method (CPFEM) with the cellular automaton (CA) method, where the distribution of the dislocation density and the deformed grain topology quantified by CPFEM were used as the input for the subsequent SRX simulation performed using the CA method. The simulation results were confirmed by the experimental ones in terms of microstructures, average grain size and recrystallization fraction, which indicates that the proposed coupling method is well able to describe the SRX behavior of the NiTiFe SMA. The results show that the dislocation density exhibits an inhomogeneous distribution in the deformed sample and the recrystallization nuclei mainly concentrate on zones where the dislocation density is relatively higher. An increase in the compressive deformation degree leads to an increase in nucleation rate and a decrease in grain boundary spaces in the compression direction, which reduces the growth spaces for the SRX nuclei and impedes their further growth. In addition, both the mechanisms of local grain refinement in the incomplete SRX and the influence of compressive deformation degree on the grain size of SRX were vividly illustrated by the corresponding physical models.

Reliable and predictive relationships between fundamental microstructural material properties and observable macroscopic mechanical behaviour are needed for the successful design of new materials. In this study we establish a link between physical properties that are defined on the atomic level and the deformation mechanisms of slip planes and interfaces that govern the mechanical behaviour of a metallic material. To accomplish this, the shear instability energy Γ is introduced, which can be determined via quantum mechanical ab initio calculations or other atomistic methods. The concept is based on a multilayer generalised stacking fault energy calculation and can be applied to distinguish the different shear deformation mechanisms occurring at TiAl interfaces during finite-temperature molecular dynamics simulations. We use the new parameter Γ to construct a deformation mechanism map for different interfaces occurring in this intermetallic. Furthermore, Γ can be used to convert the results of ab initio density functional theory calculations into those obtained with an embedded atom method type potential for TiAl. We propose to include this new physical parameter into material databases to apply it for the design of materials and microstructures, which so far mainly relies on single-crystal values for the unstable and stable stacking fault energy.

We demonstrate how spatial cross-correlation image analysis can be used to characterize the strain rate dependence of simulated plastic damage for three systems, two copper bi-crystals containing complementary tilt grain boundaries, and a copper crystal without a grain boundary strained along the [011] direction. Distributions of cross-correlation coefficients (CCs) generated within the same system and strain rate are used to characterize the range of different types of damage observed, while CC distributions generated between the same system but at different strain rates indicate the degree of similarity of the damage generated between strain rates. For both bi-crystals, the CC distributions indicate a broader range of damage configurations as strain rate decreased. For a 15° tilt angle CC distributions generated between the damage configurations at the highest and the lowest strain rates indicate a common set of damage configurations, while for the 45° tilt angle the same analysis suggested comparatively fewer similar damage configurations. In contrast, lower strain rates for the system without initial grain boundaries resulted in far fewer distinct damage configurations and a high degree of matching between the similar configurations. In this case the damage is composed of ordered stacking faults along the (111) planes.

In the current study, the viscoelastic properties of the free-standing DPPC lipid bilayer are investigated using coarse-grained molecular dynamics (CG-MD) and inverse finite element (FE) methods. As the first step, the CG-MD method is employed to simulate the loading/relaxation of a free-standing DPPC lipid bilayer in an indentation experiment. Then the experiment is simulated using the FE method, in which viscoelastic properties of the bilayer are chosen by a genetic algorithm. At each optimization step, the force–time curve is extracted and evaluated with respect to the curve obtained from the CG-MD simulation. The optimization process is continued until a sufficiently good accordance is acquired between the force–time curves obtained from the FE and CG-MD simulations. The material's behavior in the FE simulation is represented by a two-term Prony model which comprises three unknown constants; the instantaneous Young's modulus, the steady-state Young's modulus and the relaxation time constant, which are obtained through optimization. The effects of various simulation parameters, such as indentation speed, the shape of the indenter, the size of the bilayer and temperature, on the viscoelastic properties of the bilayer are also studied and discussed.

Properties and distribution of the products formed during the hydration of cementitious composite at the microlevel are investigated using a nanoindentation technique. First, numerical nanoindentation using nonlinear contact mechanics is carried out on three different phase compositions of cement paste, viz. mono-phase Tri-calcium Silicate (C₃S), Di-calcium Silicate (C₂S) and Calcium-Silicate-Hydrate (CSH) individually), bi-phase (C₃S-CSH, C₂S-CSH) and multi-phase (more than 10 individual phases including water pores). To reflect the multi-phase characteristics of hydrating cement composite, a discretized multi-phase microstructural model of cement composite during the progression of hydration is developed. Further, a grid indentation technique for simulated nanoindentation is established, and employed to evaluate the mechanical characteristics of the hydrated multi-phase cement paste. The properties obtained from the numerical studies are compared with those obtained from experimental grid nanoindentation. The influence of composition and distribution of individual phase properties on the properties obtained from indentation are closely investigated. The study paves the way to establishing the procedure for simulated grid nanoindentation to evaluate the mechanical properties of heterogeneous composites, and facilitates the design of experimental nanoindentation.

Atomistic simulations of bicrystal samples containing a grain boundary are used to examine the effect of hydrogen atoms on the nucleation of intergranular cracks in Ni. Specifically, the theoretical strength is obtained by rigid separation of the two crystals above and below the GB and the yield strength (point of dislocation emission) is obtained by standard tension testing normal to the GB. These strengths are computed in pure Ni and Ni with H segregated to the grain boundaries under conditions typical of H embrittlement in Ni, and in artificially highly-H-saturated states. In all GBs studied here, the theoretical strength $\hat{\sigma }$ is not significantly reduced by the presence of the hydrogen atoms. Similarly, with the exception of the Ni ${\rm{\Sigma }}27(115)\langle 110\rangle$ boundary, the yield strength ${\sigma }_{{\rm{y}}}$ is not significantly altered by the presence of segregated H atoms. In all cases, the theoretical strengths are ∼25 GPa and the yield strengths are ∼10 GPa, so that (i) the theoretical strength is always well above the yield strength, with or without H, and (ii) both strengths are far above the bulk plastic flow stress, ${\sigma }_{{\rm{y}}}^{{\rm{B}}}$ of Ni and Ni alloys. Significant reductions in fracture energy (25%–45%) are only achieved for some of the artificially high-H-segregation cases and then only when all the H around the GB is allow to diffuse locally to the fracture surface, which corresponds to unlikely out-of-equilibrium segregation plus local kinetics. Complementing recent work showing that H does not change the ability of GB cracks to emit dislocations and blunt, the present work indicates that equilibrium segregation of hydrogen atoms to GBs has little effect on lowering the GB strength and energy, and so does not significantly facilitate nucleation of intergranular cracks.