Challenges in applied materials engineering often originate from the variability of materials microstructure upon processing. Classic examples are seemingly small changes in rolling or sintering protocol that lead to dramatic changes in product quality. Inverse simulation of such processes can only be tackled once the forward problem can firmly be based on mechanistic understanding and experimental validation. While integral measurements without microscopic information are not particularly useful in this respect, microstructural analysis and micro-experiments, like nanoindentation or micro-compression testing, seem attractive. However, they are not without their own difficulties so that in-situ experiments or multiple experiments at various stages of a process are needed. Three examples will be introduced: (1) in-situ micro-compression testing of nanocrystalline nickel to determine the constitutive plastic deformation behavior, (2) rolling simulation of a dual phase steel to obtain texture evolution that then leads to mechanical anisotropy, (3) interrupted sintering experiments on strontium titanate including determination of the full three dimensional microstructure by diffraction contrast tomography.

The goal of this work is to numerically and experimentally establish the relationship between particle size and mechanical properties, independent of particle orientation or relative spacial relationships. Additionally, the work explores the effects of controlling the viscoelastic matrix material properties, which adds a substantial degree of flexibility to inverse materials computation. Matrix polymers can be readily modified by many means: varying molecular weight, choice and size of reinforcing particle, plasticizers, etc. A multi-material 3D printer is used to build the identical representative volume element (RVE) samples on length scales varying by three orders of magnitude. The 10% volume fraction RVE consists of glassy, thermoplastic, 1:5 aspect ratio ellipsoid particles in a clear, rubbery matrix. Utilizing the 3D printer, identical and reproducible specimens can be physically tested and compared to simulation. Viscoelastic mechanical properties of the bulk matrix and particle materials are obtained in a novel, self-consistent manner. Frequency domain tensile modulus and Poisson's ratio are measured from a tensile specimen, from which frequency domain Bulk and Shear modulus are inferred, then transformed into time domain Prony series. Large deformation ramp tests are also conducted to find parameters for 3 non-linear viscoelastic models: Elasto-plastic, Hyperelastic, and a distortion modified reduced time model. Each model is evaluated on the bulk material. Ramp and large deformation creep loading is simulated on the RVE's to establish the effects of different length scales on identical geometries. Surface strains from the computational results are compared to experiments from an under optical microscope test system utilizing digital image correlation to map the surface 2D strain field. The matrix and particle deformation can be experimentally resolved with up to 25 nm resolution of the displacement field. Additionally, inter-phase mechanical response is compared between simulation and experiment. Besides establishing effect of particle size independent of orientation, the effect of varying viscoelastic matrix properties will be simulated. Mechanical properties from other potential matrix materials will be applied to the model to explore their effect on composite strength and modulus. The authors would like to thank Prof. Dr. Z. Major of the Johannes Kepler University Linz, for his valuable input and for graciously providing the 3D specimens.

Inspired by recent experimental results suggesting that the heterogeneous distribution of the elastic modulus in bone tissue leads to increased toughness, we perform a computational investigation of a discrete particle system with stochastic elastic properties. We consider an elastic solid under plane strain conditions in uniaxial tension with a Young&’s modulus distribution modeled as a 2-d Gaussian process. Via a comparison to spectral methods with stochastic finite elements and Monte Carlo methods we validate the persistence of the Cauchy Born rule in a stochastic spring bead network of FCC-lattices. This validated stochastic molecular mechanics framework supports the inverse computation of local elastic properties, not accessible with continuum mechanics, to tailor global mechanical properties such as the fracture toughness. Specifically, Markov Chain Monte Carlo can be used to infer the elastic and geometric parameters. To this end, we analyze a flawed discrete particle system that models a biological tissue to assess the effect of heterogeneity on fracture properties. Our work sets the foundation for stochastic modeling in a micromechanical environment and unveils mechanisms by which mechanical behavior can be tailored due to increasingly heterogeneous mechanical properties. We gain insight into a universal natural design mechanism and pave the way for new advances in bio-inspired engineering.

In this presentation we consider the inverse problem of optimal reconstruction of state-dependent constitutive relations in complex multiphysics systems based on measurements. Assuming that the system is described by coupled nonlinear partial differential equations (PDEs), we will demonstrate that such problems can be formulated in terms of PDE-constrained optimization where solutions can be obtained using a suitable iterative gradient-based technique. A key element of this approach is the computation of the cost functional gradients which can be obtained from the solution of an adjoint system leading to a number of interesting questions at the level of numerical analysis and scientific computing. While the proposed approach is illustrated with a model problem from the area of fluid mechanics and heat transfer, it can be applied to diverse problems in solid mechanics and thermodynamics. We will also discuss a number of extensions, including the reconstruction of inertial manifolds in the phase-space formulation of dynamical systems. The presentation will contain elements of rigorous mathematical analysis alongside with results of large-scale numerical computations. [joint work with V. Bukshtynov] References: [1] V. Bukshtynov, O. Volkov, and B. Protas, "On Optimal Reconstruction of Constitutive Relations" Physica D 240, 1228-1244, 2011. [2] V. Bukshtynov and B. Protas, "Optimal Reconstruction of Material Properties in Complex Multiphysics Phenomena", (submitted), 2011.

Localization (opposite of homogenization) describes the spatial distribution of the response field of interest at the microscale for an imposed loading condition at the macroscale. A novel approach called Materials Knowledge Systems (MKS) has been formulated recently to build accurate, bi-directional, microstructure-property-processing linkages in hierarchical material systems to facilitate computationally efficient multi-scale modeling and simulation. This approach is built on the statistical continuum theories developed by Kroner that express the localization of the response field at the microscale using a series of highly complex convolution integrals, which have historically been evaluated analytically. The MKS approach dramatically improves the accuracy of these expressions by calibrating the convolution kernels in these expressions to results from previously validated physics-based models. In this paper, we demonstrate the power of this novel approach with selected case studies in multi-scale simulation of materials phenomenon.

Low-k dielectrics for Cu interconnects are used in microelectronics to reduce interconnect delay and decrease parasitic capacitance. Typically a hard TiN mask is used to pattern narrow dielectric lines. The large mismatch in elastic modulus (40:1 ratio) and large TiN mask stress (1 GPa ~ 5 GPa) can induce buckling during processing of thin (width < 50 nm), tall (height > 150 nm) lines. Buckling is determined by the mechanical properties of the low-k dielectric, the TiN stress, geometric details and random irregularities in geometry and properties. In this work, buckled dielectric lines were measured via atomic force microscopy (AFM) to determine their spatially varying deflection. Buckling was modeled as a weakly nonlinear stochastic phenomenon, so that the random deflections could be described via three statistical parameters: characteristic wavelength, mean-square amplitude and coherence length. Mechanical modeling was used to determine materials parameters from the statistical observation: low-k dielectric elastic modulus and TiN film stress. A simplified analytic model was used to determine trends and to demonstrate qualitative behavior and finally the finite element method was used for quantitative accuracy. Extracted parameters were compared with independently measured values using techniques such as contact resonance atomic force microscopy. Once confirmed, the stochastic mechanical model could be used to place design limits on line height and width to avoid dielectric line buckling.

In structural engineering, optimized performance requires the simultaneous combination of component design and the selection of a material with the best balance of properties. Many applications require strong, tough, lightweight materials, and polymers or polymer composites are ideal candidates. Unfortunately, polymers are complex highly nonlinear materials, where desired properties often follow inverse trends with respect to each other with changes in composition or morphology (Murphy&’s Law for polymer science). Component design or analysis is facilitated by simulation techniques such as dynamic finite element analysis, FEA, at the continuum scale, which generally use empirical datasets for material properties in constitutive equations to allow calculation of local stress and temperature fields under dynamic loading. At the opposite end of the multi-scale spectrum, atomistic simulations such as molecular dynamics, MD, are computationally demanding and provide little insight into how to make structural changes for engineering materials design at the molecular level. The objective here is to simplify the detailed molecular modeling process to allow in situ integration into dynamic FEA codes by deriving a self-consistent set of analytical equations for bulk nonlinear mechanical properties directly from ensemble average potential energy wells for intermolecular interactions. The potential energy well for energy as a function of molecular dimensions requires only a limited number of single valued parameters (such as molar mass and volume, cohesive energy, and degree of crystallinity), which can be calculated directly from the chemical and morphological structure of the material using molecular techniques such as MD or quantum methods or empirical tools such as connectivity indices. Thus, FEA simulation can be performed by direct input of a chemical structure, or the parameter sets can be optimized within the FEA simulation framework to best satisfy sets of application design criteria. For example, elastic moduli predictions as a function of temperature and rate allows direct and rapid simulation of nonlinear stress-strain relations over all temperature, rate, and strain space, even with physically unrealistic values accessed by FEA codes during convergence. Thus, molecular or nanoscale simulations are coupled directly into the continuum codes by replacing empirical constitutive relations. A specific example of prime interest would be optimization of biomechanical compatibility of polymers for orthopedic implants, where we are reducing the model for potentially very complex protein structures and compositions (such as derived from silk products) to a tractable set of parameters (order, disorder and hydrated fractions) and routines that can be optimized in an interactive biomechanics FEA code.

Statistical continuum mechanics strong contrast formulation was developed to evaluate the effective properties for a general multiscale nanocomposite comprising three different phases. The developed method can be applied to compute the effective thermal, electrical and permeability properties of a general anisotropic three phase nanocomposite by considering the individual properties for each phase together with their microstructural morphologies. In order to account for the microstructural morphology, statistical N-point correlation functions were utilized. As the microstructure can be fully reconstructed by N-point correlation functions, the proposed formulation is expandable up to N-point correlation functions. Practically, a third-order truncation of the expansion provides a sufficient approximation for the effective properties of the micro/nanostructures with high contrast between the properties of each phase. A third order truncation of the expansion exploits two-point and three-point microstructural correlation functions into the solution. Two and three phase nanocomposites were computationally reproduced by a stochastic Monte Carlo technique using random sequential addition method. Two-point microstructural correlation functions were evaluated numerically from the simulated representative volume element of the nanocomposite. As the evaluation of the three-point correlation functions directly from the simulated microstructure turned out to be computationally expensive; the three point correlation functions of the microstructure were further approximated by the two-point correlation functions. In order to assess the validity of the proposed method, the effective thermal conductivity of a three phase nanocomposite comprising a polymer matrix, carbon nanotubes (CNTs) and graphene nanoplatelets (GNPs) were treated for both isotropic and anisotropic formulations. The effective thermal properties of the same microstructures were also estimated by standard finite elements methods. Good agreement between the predicted properties via both constituted the validity of the proposed method. The developed method carries the appeals of the ready applicability to most general case, the consideration of the morphological effects on the effective properties, the strong contrast assumption for the properties, and the relatively lower computational expense.

Zr is an important metal for hydrogen storage and nuclear application, and the interaction between Zr and hydrogen (H) has raised great research interests among researchers. Regarding experimental and calculational studies of the well-known δ→ε transition, there are several controversies in the literature. In the present study, first-principles calculation reveals that the ZrHx phases (x=1, 1.25, 1.5, 1.75, and 2) with the cubic fluorite-type (fcc, δ phase) and face-centered-tetragonal (fct: ε phase, c/a<1; γ phase, c/a >1) structures are all energetically favorable with negative heats of formation of -30 to -56 kJ/(mol. H) and very small structural energy differences, while mechanical stability plays a more important role in determining the existence of various ZrHx phases. Calculation also shows that the intrinsic composition range of the δ→ε transition of ZrHx phases is xge;1.5, and that the fundamental mechanism of this transition is mechanical unstableness of the δ phase which will spontaneously transform into ε phase by means of the {110}<110> shear. Moreover, electronic structures show that the co-function of van Hove singularities and degenerate bands along several directions brings about the high level of density of states at or near the Fermi level and fundamentally induces the mechanical unstableness of the δ phase. The calculated results agree well with experimental observations and could clarify the controversies of the ZrHx phases in the literature. Ref.: F. Wang and H. R. Gong, International Journal of Hydrogen Energy 2012, 37: 9688-9695

For many years, computational techniques for the accurate treatment of contact problems in continuum mechanics have been developed and applied with great success to a large variety of problems. On the other hand, when approaching the nanoscale, a continuum viewpoint might be not justified anymore but atomistic simulations still allow for very detailed and accurate calculations. The relevant static and dynamic quantities that occur in a contact problem, such as normal forces, frictional forces and the area of interaction can be obtained from both types of simulation, although for some of these quantities a meaningful atomistic definition is not obvious and thus had to be developed to come close to the corresponding intuitive concepts. The goal of our study is to quantify the discrepancy between the continuum and the atomistic formulations on length scales in the nanometer regime. To this end, the two approaches are compared within a typical nanoindentation setup where a spherical indenter is pushed onto a polymeric self-assembled monolayer made of poly(ethylene oxide). In this vein, we try to extend existing endeavors that have been made to shed light on the limitations of continuum mechanical descriptions of contact problems when reaching the nanoscale. Here, our implementation of the continuum contact model is a finite element-scheme based on the penalty method, while atomistic computations are conducted by means of molecular dynamics. A challenging subproblem herein is to find a constitutive law for a layered material which exhibits strong anisotropy. Our approach relied on fitting common hyperelastic material models that are designed for anisotropic polymers and rubbers by employing a procedure of Cauchy-Born rule type.

Certain technological applications of graphene may be enabled by a sufficiently-detailed understanding of graphene&’s elastic and lattice-dynamical properties when subjected to in-plane strain of moderately large magnitude. In anisotropic crystals, the response to applied deformation depends on the orientation of the principal axes of strain with respect to directions of reference crystal symmetry, a consequence of Neumann&’s principle. Further, the strain, in general, has the additional effect of changing the symmetry of the deformed crystal. The representation of the crystal properties, therefore, requires due consideration of the lattice symmetry and its evolution with strain. We present symmetry-based representation schemes for constructing lattice response functions via invariant interpolation. The methods are demonstrated through two example applications. First, we employ the scheme to obtain the anisotropic strain energy density function of graphene corresponding to a subspace of the in-plane deformations: ε such that n. εn > 0 for all n isin; R2 (R2 is the Euclidean 2-space). Such a deformation subspace strictly precludes compression (and hence buckling), and only non-trivial cases of instability are investigated. We decompose the in-plane strain tensor into a dilatational part (isotropic areal stretching) and a deviatoric part. For pure areal dilatation, the lattice response is accurately captured by the Rydberg function expressed in the trace of log strain. For the deviatoric part, we fit ab-initio energies using polynomial functions of invariants of the log strain. From the response function thus constructed, we obtain a complete representation of stress-strain response that enables prediction of the deformation-dependence of long-wavelength elastic instabilities (loss of strong ellipticity). Extrapolating the response, we also find the locus in strain space corresponding to buckling instability i.e. n. σn le; 0 (σ is the Cauchy stress). In a second application, we use the invariant-based interpolation scheme to represent lattice vibrational spectra (i.e., all branches of phonon frequencies as functions of wave vector k in the entire Brillouin zone), as well as their evolution with strain ε. The representation of the evolution of vibrational spectra with strain is used to predict deformation-induced short-wavelength instabilities (soft-modes) in the crystal. By combining the results of the two example studies, we are able to quantify the deformation-dependence of all types of mechanical instability in the graphene lattice.

Collagen is the most abundant protein in the human body. Collagen exists intimately associated with the mineral hydroxyapatite in human bone. Multiscale modeling of mechanical behavior has extensive applications in mechanics of structures used in engineering due to extensive use of nanomaterials but biology exhibits an added complexity to the multiscale modeling efforts arising for the inherent hierarchy of structure that is a characteristic of biological systems. We present here a multiscale effort using ab initio methods, molecular dynamics and finite element to develop a multiscale mode of the collagen fibril in human bone. In this study we have linked molecular properties of bone with the mechanics of collagen fibril using a hierarchical multiscale scheme. Steered molecular dynamics (SMD) is used to investigate the nanomechanical property of the collagen molecule in the proximity of mineral and the collagen-mineral interface. These simulations show that the mechanical response of collagen is significantly altered by the hydroxyapatite. The finite element (FEM) model of a collagen fibril 50 nm diameter and a micron in length is built that mimics the staggered arrangement of collagen molecule resulting in the banded pattern along the of fibril. The mechanical properties of the collagen and the molecular interfaces used in FEM model is directly obtained from SMD. The finite element simulations are performed in the small displacement regime (<5% axial strain) and large displacement regime. Our simulations in the small-displacement regime show that the elastic modulus of collagen is significantly affected by the mineral proximity. We also present results of our simulation carried out in the large-displacement regime, as well as the nanomechanics of collagen fibril at various crosslink densities.

A bottom-up, multiscale modeling approach is presented to carry out high-fidelity virtual mechanical tests of composite materials and structures. The overall multiscale simulation scheme takes advantage of the fact that composite structures are made up of laminates which in turn are obtained by stacking individual plies with different fiber orientation. This leads to three different entities (ply, laminate and component) whose mechanical behavior is characterized by three different length scales, namely fiber diameter, ply and laminate thickness, respectively. Fiber diameters are of the order of 5-10 mu;m, while ply thicknesses are in the range 100-300 mu;m and standard laminates are several mm in thickness and above. This clear separation of length scales is very useful to carry out multiscale modeling by computing the properties of one entity (e.g. individual plies) at the relevant length scale, homogenizing the results into a constitutive model, and passing this information to the simulations at the next length scale to determine the mechanical behavior of the larger entity (e.g. laminate). This strategy is applied here to design composite laminates with optimum out-of-plane impact resistance and damage tolerance. Using optimisation tools based on evolutionary algorithms, non-conventional laminates (in which the plies are dispersed in different orientations through the 0-90° range instead of the classic orientations at 0°, 90°, and ±45°) were designed. They presented nearly the same in-plane and out-of-plane elastic properties as the conventional laminates from which they were derived but in these laminates with dispersed stacking sequences, the occurrence of matrix cracks in plies clustered at the same orientation is mitigated due to more efficient fiber bridging at neighboring plies, leading to an alleviation of stresses in the transverse direction and a load transfer to the stronger longitudinal direction. In addition, it was possible to optimize the location and propagation of delaminations in order to maximize the Compression-After-Impact resistance of these non-conventional laminates. The predictions of the optimization process by means of multiscale modeling were confirmed by the experimental results.

Random fiber networks are structural elements in many biological and man-made materials. The prediction of mechanical properties of such structures is desirable in many applications. In this work we use simulations to investigate the mechanical behavior of fiber networks, to develop constitutive representations for these structures and to identify the microstructural parameters that play a role in the global scale mechanics. The conditions under which the system of fibers can be treated as a continuum and those in which its discrete nature is important will be discussed. For some types of structures, significant size effects are observed and quantified. Their influence on the constitutive behavior on the network scale will be reviewed. Embedding the network in an elastic matrix modifies significantly the network response and this is important in most biological applications. This effect will be reviewed and discussed.

Considerable effort has been made to manufacture carbon nanotube (CNT) assemblies to realize remarkable nano-scale properties of CNTs in micro/macro-scales. Among these, CNT yarns are recognized as most beneficial structure because of various and significant potentials to superelastic muscles, flexible touch screens, transparent electrodes, supercapacitors, energy harvesters, etc. To design or investigate those futuristic devices using CNT yarns, it is essential to understand the internal structure of CNT yarns and their influences over the electrical and mechanical behavior of CNT yarns. In this study, the mechanical behavior of CNT yarns, which were produced by twisting CNT sheets drawn from CNT forest, has been modeled under continuum mechanic framework. Considering the manufacturing processes, a CNT yarn was assumed as a wrapping assembly of CNT sheets. Furthermore, CNT yarns showed the nonlinear mechanical behavior due to the nonlinear mechanical behavior of individual CNT and the helical structure of the yarn. Given that pores were formed inside the CNT yarn during the yarn manufacturing process, CNT yarns were modeled using a neo-Hookean composite model with aligned continuous cylindrical pores in the finite elasticity regime. The material parameters necessary for the material law were obtained by using CNT arrangement within the yarn, the classical staple yarn theory, and molecular dynamics. To validate the current modeling approach, the tensile and hysteretic behavior of CNT yarns were calculated and compared with experiments.

The present study investigates the nonlinear elastic properties of a single-layer molybdenum disulfide crystal through experiment, finite element modeling, and density functional theory in a multiscale material model validated via atomic force microscopy nanoindentation. The model provides a platform to validate first-principles derivation of continuum fifth-order elastic constants for in-plane stiffness and intrinsic strength usingbased upon density functional theory. The derived higher order elastic constants are used in a finite element model to predict the breaking strength of two-dimensional molybdenum disulfide. To validate the multiscale model, single-layer molybdenum disulfide crystals are suspended over circular holes that were etched on a silicon oxide surface. The free-standing circular monatomically-thin crystalline layers of molybdenum disulfide are loaded at the center with an atomic force microscope until fracture occurs. The load-displacement curve is used to determine the pretension and linear-elastic response of the crystal. The force at which fracture occurs gives insight into the intrinsic strength and higher order elastic constants of the crystal. The multiscale finite element model accurately predicts both the elastic properties and the intrinsic breaking strength of monatomically-thin molybdenum disulfide. The study bridges the gap between density functional theory and finite element analysis with experimental evidence.

The spall fracture of materials resulting from the interaction of two rarefaction waves reflected from the free surfaces is an important phenomenon measured in plate impact shock propagation experiments. The spall strength calculated from the pull-back velocity is treated as materials characteristics at high strain rate loading and simulated using phenomenological damage models based on stress (energy) release or void nucleation, growth, and coalescence (VNGC). The existing models predict pull-back velocity implying homogeneous spall plane and are often found to show mesh size dependence. On the contrary, the spall planes in recovered samples are found to be heterogeneous and the pull-back velocity in high resolution measurements reveal length scale effects not predicted by existing models. The present work focuses on simulating spall fracture using the stress (energy) release, VNGC, and intra-grain fracture approaches for one dimensional (1D) uni-axial strain simulations of the spall event. Furthermore, the VNGC approach is coupled with the intra-grain fracture for two dimensional (2D) mesoscale simulations to gain insight into the length scale signatures recorded in the measurements. The results will be presented to show the effects of model parameters on simulated pull-back velocity and the maximum tensile stress at spall plane compared to the measured data and spall strength. It is concluded that more advanced experimental techniques are required to resolve different failure phenomena and arrive at unique set of model parameters for forward simulation of the spall fracture in materials.

Graphene nanostructures have recently generated heightened interest primarily due to its high strength-to-weight ratio, unique thermal and physio-chemical properties. Improvements in the thermal conductivity of large-scale composite structures were observed with the addition of graphene nanostructures, even at low volume fraction. Therefore, to understand and built efficient composite systems with quality thermal characteristics; it is necessary to develop theoretical formulations and associated computational procedures that take into account the physical processes that occur at different spatial and temporal scales. In this study, a multiscale mathematical framework and simulation strategy was used to estimate the thermo-mechanical characteristics of advanced -composite structures having embedded nano-graphene systems. In most of the previous theoretical works a perfect bonding between the nanographene and the surrounding materials was assumed; however, there exists an interphase layer that contributes significantly to the interfacial thermal resistance. This interface resistance in nano-systems was one of the most important factors that lead to the large variation in thermal conductivities reported in literature. In this work, the thermal properties of multilayered nanographene structures were estimated using molecular dynamics (MD) simulations. The MD simulations were carried out at various temperatures by rescaling the velocities of the carbon atoms. The properties obtained from MD were scaled to the macro-scale using continuum thermo-mechanical formulations. The results show a significant effect of the interfacial thermal resistance on the effective thermal conductivity of the overall composite system. The developed multiscale model is also used to study the thermal behavior of the structure with varying graphene concentrations. Acknowledgement: This work was supported by the generous faculty start-up funds from The University of Alabama.

Recrystallization of the grain structure of metals into nano-sized grains by using mechanical means, has received wide attention in the last two decades. It is well known that materials with a fine-grain crystal structure have favorable properties compared to the same materials with coarse-grained crystal structure. Surface Mechanical Attrition Treatment (SMAT), a technique developed in the early part of this decade, has been successfully used to recrystallize the surface grains of metals into nanocrystals of the order of 10 to 100 nanometers from their original grain sizes of 10 to 30 microns. Resulting enhancement in surface properties has quite interesting applications, varying from materials with improved fatigue resistance to medical devices. However, most of these materials are highly unstable, especially in the high temperature applications. In this study, the deformation and stability of nanostructured metals are investigated using Molecular Dynamics (MD) simulations. Several strategies to introduce dislocations into the model are considered and its influence on the material behavior is investigated. Single crystal FCC metals are considered in the study. A criterion to evaluate how well these strategies represent the realistic nanostructure of the material is derived. Using a representative nanostructure, the deformation of these materials under a localized loading, such as nanoindentation, is simulated. The simulated result is compared with experimental data for verification of the model. Effect of grain size on the deformation behavior is studied by considering different initial nanostructure. By studying the variation in the deformation behavior for material under several temperatures, the stability of the nanostructured materials is also investigated as a function of temperature.

Previously, colloidal structures inside solutions were mostly considered as rigid bodies or linear springs. However, recent experimental results show a strongly nonlinear mechanical response of the large clusters. In this contribution, the nonlinear elastic behavior of the colloidal structures inside polymeric solutions is investigated. We present a model for deformation induced evolution of the aggregate structure with respect to inter-particle interactions. To this end, the inter-particle forces were categorized into central and lateral (non-central) forces, where the lateral forces were considered responsible for rotation of the particles around each other. The directional stiffness of a colloidal structure is described by the concept of the backbone chain, which is a unique path between two ends of the colloidal structure that carries the main portion of load applied to the structure. The mechanical behavior of the backbone chain is formulated so that it sensitively depends on the aggregate geometry, deformation history and moreover, on the nature and the strength of the inter-particle interactions. The model can further be generalized to other types of colloidal structures with central and lateral inter-particle forces, e.g. blood clots.

First-principles predictions of molecular and thus material properties require a quantum mechanical description of the elementary particles of chemistry, i.e., of electrons and atomic nuclei. Such a theoretical framework can be the basis for a multi-scale modeling approach in materials science. Unfortunately, almost all present-day attempts to quantitative property predictions have to apply a brute-force forward approach that requires structural input (the nuclear configuration of the structural model) to which then the property is assigned in an electronic structure calculation with some quantum chemical method chosen. As a result, computational screening for structures with desired properties becomes possible, but is cumbersome. It would, however, be desirable to streamline these calculations by devising methods that produce (structural) information for pre-defined properties [1]. Along these lines, we have studied various approaches that may come in handy for achieving this goal. These approaches range from studying quantum-mechanical embedding theories [2] to inverse methods for the calculation of molecular properties. As an example for the latter, we have developed quantum chemical methods for the inverse solution of the vibrational problem within the harmonic approximation [3]. Considering the fact that one would calculate in the standard (forward) approach all spectral features of which many are often not needed for answering a specific scientific question (e.g., for assigning a molecular structure based on a diagnostic vibrational band), we developed a subspace iteration protocol called Mode-Tracking [4] that allows us to selectively converge pre-defined molecular vibrations to normal modes. This is convenient for diagnostic vibrations clearly visible (often isolated) in the spectrum. However, since it yields only a very limited part of the vibrational spectrum, a fingerprint identification is not possible. To overcome this problem, we devised the Intensity-Tracking algorithm [5] that distributes spectral intensity of any type of vibrational spectroscopy to only those normal modes that pick up intensity. In order to accomplish this, it was necessary to rigorously define Intensity Carrying Modes [5]. References: [1] M. Reiher, Chimia 63 (2009) 140. [2] K. Kiewisch, G. Eickerling, M. Reiher, J. Neugebauer, J. Chem. Phys. 128 (2008) 044114; S. Fux, K. Kiewisch, C.R. Jacob, J. Neugebauer, M. Reiher, Chem. Phys. Lett. 461 (2008) 353; S. Fux, C. R. Jacob, J. Neugebauer, L. Visscher, M. Reiher, J. Chem. Phys. 132 (2010) 164101. [3] C. Herrmann, J. Neugebauer, M. Reiher, New J. Chem. 31 (2007) 818; K. Kiewisch, S. Luber, J. Neugebauer, M. Reiher, Chimia 63 (2009) 270. [4] M. Reiher, J. Neugebauer, J. Chem. Phys. 118 (2003) 1634. [5] K. Kiewisch, J. Neugebauer, M. Reiher, Chem. Phys. 129 (2008) 204103; S. Luber, J. Neugebauer, M. Reiher, J. Chem. Phys. 130 (2009) 064105; S. Luber, M. Reiher, ChemPhysChem 10 (2009) 2049.

The development of accurate and transferable models to describe multicomponent systems is attracting considerable interest in materials modelling. Tight-binding models derived from the density functional theory potentially provide an accurate and systematic approach to this problem. The present contribution outlines our attempts to develop such models for transition-metal-based complex alloys, and efforts to extend their applicability towards larger-scale molecular dynamics simulations within a linear-scaling framework.

We present a scale-bridging approach for modelling the integral elastic response of polycrystalline composites that is based on a combination of (i) first-principles calculations of thermodynamic phase stability and single-crystal elastic constants, and (ii) multi-scale homogenization schemes developed for composite aggregates. The modeling is used as a forward theory-guided materials design strategy to identify alloys with application-dictated elastic properties. As an example, Ti alloys that aim at elastically matching human bones in biomedical applications are discussed. The theoretical results reveal decisive influence of the multi-phase character of Ti-Nb composites on their properties. Quantum-mechanically determined compositional trends clearly demonstrate how the thermodynamic stability of phases is connected with their elastic properties imposing strict limits on macroscopic properties that may be realistically achieved. Our study concludes with a sensitivity analysis probing relation between the output (homogenized elastic moduli) and the input (ab initio calculated single-crystalline elastic constants).

An important prerequisite for developing inverse tools in materials science is a highly accurate description between the specific atomic structure such as chemical composition and coordination of the atoms and their properties on the macroscopic scale. Only if the underlying relations between atomistic and macroscopic scale are known inversion schemes that use as input macroscopic materials properties and provide the corresponding atomistic structure become feasible. While originally such approaches have been mainly restricted to T=0 K calculations on the atomistic level due to the high computational demand of the ab initio calculations recent developments in the combination of accurate first principles calculations with thermodynamic or kinetic concepts opens the door to tackle even advanced engineering problems. Key to these studies is the highly accurate determination of thermodynamic quantities at finite temperatures. In the first part of the talk it will be shown how novel sampling strategies allow to obtain an amazing coarse graining in configuration space resulting in a reduction from 107 to a few hundred configurations. This enormous reduction permits to employ highly converged density-functional theory calculations thus providing the basis for accurately determining all relevant temperature dependent free energy contributions such as harmonic and anharmonic vibrations, magnetic excitations or defect creation. The flexibility and the predictive power of this approach will be discussed in the second part of the talk for a few examples relevant to the design and understanding of metallic alloys with superior mechanical properties: Determination of mechanisms that provide an adaptive and self-organized response of the materials microstructure on external forces and unraveling hidden relations between competing materials properties.

Effective masses of carriers play an important role in many applications of semiconductors and insulators, e.g., (i) low effective mass is needed for high speed integrated circuits, (ii) in field-effect transistors and transparent conducting oxides, low effective mass is crucial to gain high mobility, and (iii) for thermoelectrics, both thermopower and conductivity are closely related to effective mass. Surprisingly, there are no compelling design principles for deliberately modifying materials to get target effective mass other than the rule from the k-dot-p theory that the electron effective mass is proportional to the band gap. It does not tell us how the effective masses of holes are, and also the simple k-dot-p theory is expected to break down when numerous bands interacting. Standard DFT theory is very poor in predicting effective masses due to its serious underestimation of band gap and effective mass. Here we propose to use the atomistic screened-pseudopotential calculations, which correct many "DFT errors", along with genetic algorithms, to explore the underlying rules governing effective mass and inversely design structures with target effective masses. We include the effects of strain, disordered alloys, atomic ordering into superstructures, and quantum confinement by nanostructuring, and apply these to group IV and III-V semiconductors. An intriguing behavior is found for hole effective masses: the effective masses of heavy and light holes show a discontinuity under uniaxial and biaxial strains. We further find that the effective mass of heavy hole is strongly dependent on the atomic disordering/ordering, whereas the effective masses of electron and light hole are not. We probed the possibility of inversely designing effective masses in the disordered alloy and ordered superlattice structures. The present work aims to discover structures and materials with target effective masses (e.g., minimal, maximal, or given anisotropy), as well as design principles to guide effective mass engineering. This work is funded by DOE through Energy Frontier Research Center, Center for Inverse Design.

Nano-particles of metal clusters or metal oxides are of great importance in science and engineering, as a bridge between bulk and atomistic structures. Also they are of interest in applications such as catalysis, energy storage and relate fields. The environmental stability and chemical reactions that these particles undergo is therefore of importance. Using reactive molecular dynamics (ReaxFF), computational modeling and simulations of nano-particle oxidation and oxide breakdown have been conducted, in order to characterize the evolution of structure and chemical compositions of metal nano-particles, such as copper and iron. Reactive simulations allow charge transfer between metal cations and chloride or oxygen anions, incorporating molecular association or dissociation. Due to lower coordination, nano-particles interact with nearby chemical elements rapidly compared to bulk counterparts providing efficient performance over molecular dynamics approach. We have also studied their surface interaction in chloride containing environments to simulate corrosion phenomena and will be further discussed in the presentation. The role of point defects in influencing the oxide breakdown will be highlighted.

Discovery of new materials and processing methods relies on multi-scale models parameterized by accurate thermodynamic potentials, kinetic parameters, and constitutive relations. First principles prediction of materials properties has become an important strategy to parameterize these models for a range of technologically important materials. This has been enabled by mathematical frameworks, such as the cluster expansion formalism or harmonic lattice dynamics, that form a link between ab initio calculations and finite-temperature macroscopic variables via the construction of effective crystal Hamiltonians, which can be studied with techniques from statistical physics. Unfortunately, major challenges persist in predicting thermodynamic and kinetic properties of high-temperature materials, especially those that are predicted to be mechanically unstable at zero Kelvin. In this talk, we present a generalized, automated method to construct strongly anharmonic effective Hamiltonians by exploiting symmetry of the ideal crystal. Use of these effective Hamiltonians within Monte Carlo simulations enables prediction of macroscopic material properties at high temperature and under a range of stress/strain conditions. We demonstrate the usefulness of such methods to explore structural (martensitic) phase transformations and nonlinear elastic properties at finite temperature.

Recently, lithium ion batteries have been subject to intense scientific study due to growing demand arising from their utilization in portable electronics, electric vehicles, etc. Most cathode materials in lithium ion batteries involve a two-phase process during charging and discharging, and the rate of these processes is typically limited by the slow interface mobility. We have undertaken to modeling how lithium diffusion in the interface region affects the motion of the phase boundary. We have developed a multi-physics computational method suitable for predicting time evolution of the driven interface. In this method, we calculate formation energies and migration energy barriers by ab initio methods, which are then approximated by cluster expansions. Monte Carlo calculation is further employed to obtain thermodynamics and kinetic information, e.g., anisotropic interfacial energy, and mobility, which are used to parameterize continuum modeling of the charging and discharging processes. We test this methodology on spinel Li1+xTi2O4. Elastic effects are incorporated into the calculations to determine the effect of variations in modulus and strain on stress concentrations and failure modes within the material.

The volume optimization was performed followed by the calculation of electronic structure and magnetic properties on Co2MSi. The structural properties were calculated based on the Murnaghan&’s equation of state. The calculation of electronic structure was based on full potential linear augmented plane wave (FP-LAPW) method and exchange correlation. Results of density of states (DOS) and band structures shows the half-metallicity of Co2YSi and Co2ZrSi and with integer value of magnetic moments 1.003 µB not; and 1.996 µB respectively. The total magnetic moment of Co2NbSi was found to be 1.9114 µB small deviation from the integer value, the Fermi energy is not located exactly at the middle of the gap in spin down channel, even though there exist a gap. GGA fails to give the half-metallicity in Co2NbSi, so it has been treated within LSDA+U. For computation a computer code Wien2k has been employed. Key words: GGA, LSDA+U, half-metallicity, DOS and band structure. PACS numbers: --71.15.Mb, 71.15.m, 71.20.-b, 71.15.Ap. References 1. Murnaghan F D (1944) Proc. Natl. Acad. Sci. USA, 30 244 2. P. Blaha, K. Schwarz, G. K. H. Madsen, D. Kvasnicka, J. Luitz, K. Schwarz (2008). An Augmented Plane Wave Plus Local Orbitals Program for Calculating Crystal Properties: Wien2K User&’s Guide, Techn. Universitat Wien, Wien., 1-108. 3. H. C. Kandpal, G. H. Fecher and C. Felser, Calculated electronic and magnetic properties of the half-metallic, transition metal based Heusler compounds, (2006) J. Phys. D: Appl. Phys. 40, 1507-1523. 4. J. Y. Jiu and J. I. Lee, (2007) J. Kor. Phys Soc. 51 155-158.

We investigated the spin-transfer torque effects in a full-Heusler Co2FeAl0.5Si0.5 (CFAS) alloy and a half-metallic Heusler Co2MnSi (CMS) alloy spin-valve nanopillars through micromagnetic simulation. It is shown that different magnitudes of current densities and directions of external magnetic fields give rise to a shift of resistance hysteretic loop and a variable range of switching. A two-step switching hysteresis loop due to the fourfold in-plane magnetocrystalline anisotropy of CFAS layers was obtained. The -90° state and 90° state of two-step switching which could not be detected due to the same resistance in the experiment. Thus we firstly reported the spin transfer multi-step magnetization switching through changing the magnetic anisotropy in a full-Heusler CFAS alloy nanopillar. Furthermore, we also investigated the spin-torque oscillator in a half-metallic Heusler CMS alloy spin-valve nanopillar using micromagnetic simulations. Although it is known that the out-of-plane precession (OPP) usually has larger power output than in-plane precession (IPP), only IPP mode was observed in experiments of CMS material. Our simulations predicted that the OPP mode can be obtained under the condition of initial antiparallel state, a small external magnetic field, and large current density.

Cellulose is the most abundant hydrocarbon in the world. The chemical derivatives of cellulose have much attention recently from the point of view of its renewable and biodegradable properties. Cellulose triacetate (CTA), which is obtained from cellulose by acetylation, is one of the most common cellulose derivatives. CTA has been used as a variety of commercial products such as films and fibers for many years. Recently, there has been widespread use of CTA in new high-technology application fields. Particularly, optical film with highly controlled optical properties is the most important product in application field of CTA. However, the relation between molecular structure of CTA in the film and the optical properties has never been investigated yet. In this work, we investigate the relation of molecular structure of CTA to the optical properties of CTA film using computational chemistry method. First, some models of CTA film with amorphous structure which have different molecular orientation and molecular conformation were built up by Molecular Dynamics (MD) method. Then whole refractive index of CTA film model was estimated as a sum of refractive indices of CTA monomer which is calculated by molecular orbital method. Some refractive indices of CTA films which have different molecular orientation and molecular conformation were discussed.

Polycrystalline silicon carbide (SiC) has tremendous potential as a lightweight structural material provided its fracture toughness and tensile strength could be significantly improved. Field assisted sintering techniques (FAST) allow for the rapid consolidation of nanocrystalline ceramic materials at lower temperatures, thus minimizing grain growth and allowing for the incorporation of organic reinforcements in ceramic composites. Both a nanocrystalline microstructure and the incorporation of tensile members have been shown to improve the fracture toughness of SiC materials. Key to the development of improved SiC materials and composites is optimization of sintering based fabrication methods. We are performing multi-scale computational simulations of early- and intermediate-stage FAST sintering of nanocrystalline SiC to better understand, and then engineer, the sintering process. We have developed continuum models to predict the thermal, electric, and displacement fields inside the sintering chamber. These provide boundary and initial conditions for large scale molecular dynamics simulations, which reveal the mechanisms that contribute to densification. Several consolidation mechanisms were observed during each stage of sintering, with the rate limiting mechanism dependent upon temperature, pressure and grain size. This research helps lay the technical foundation for development of a lightweight structural ceramic matrix composite.

We have studied the structural properties of the clean (001) surface of Cu3N in the anti-ReO3 structure (space group Pm3m) using Density Functional Theory (DFT). A small relaxation was present: the first interlayer distance, is contracted by ~ 1.4%, while the second one is expanded by 0.55%. We have also studied the adsorption and diffusion of a Pd ad-atom. On the surface, the most stable configuration corresponds to the H4 site, with the Pd atom bonded to four Cu atoms of the first layer. Adsorption on top of a N atom is metastable with a slightly higher energy of 0.27 eV. The saddle point corresponds to the bridge site, with an energy barrier of 0.55eV. Occupation of the vacant site at the center of the cube results in more favorable energies. A configuration in which the Pd atom occupies the C2 site, and the Cu atom (initially at C2 as in the bulk) is displaced to the center of the unit cell and it is 1,13 eV more stable than the H4 site, while if the Pd atoms occupies the center site it is in the ground state with an energy 2,11 eV more stable than it is when in the H4 site.

Polymer self-consistent field theory (SCFT) simulation has been a useful tool to study the phase behaviors of many polymer/copolymer systems. This versatile method not only helps us to verify the different phases observed from experiments, but also helps us to explore the broad unknown parameter spaces that have not been fully resolved. The key to the success of the SCFT simulations largely relies on a robust, accurate, and efficient partial differential equation (PDE) solver, since the essence of the SCFT simulations is solving the Fokker-Planck (Diffusion) equations for the partition function propagators first described by Edwards. Pseudo Spectrum (PS) methods have been a standard PDE solver for SCFT simulations due to its accuracy and efficiency; however, since the PS method applies the Fourier operator in the reciprocal space, this method is only semi-local, preventing a local refinement algorithm. In contrast, the Lattice Boltzmann (LB) method, which was originally developed for solving hydrodynamic equations, is a purely local PDE solver. As a result, local refinement within the LB algorithm is feasible and could find many important applications in SCFT theory where interfaces are to be described accurately saving computational time and solving the equations with a high resolution only where needed. Such a scheme would be beneficial, for example, in large-scale simulations where only a small part of the system consists of modulated phases, but the rest is homogeneous, or when we are dealing with interactions between nanoscopic objects and polymers. Furthermore, the LB scheme is robust, easily parallelizable, and stable. We have implemented such a scheme for diblock copolymers and directly compared our LB PDE solver with the PS method. Our findings show that the LB method is extremely accurate we also demonstrate local refinement using the LB method for SCFT polymer simulations reproducing exactly previous findings for the self-assembly of block copolymers within commensurate templates. Lastly, the benchmarks for the LB SCFT simulations running on graphic processing units (GPUs) are provided. We find that our LB algorithm has up to 60x speed up comparing to the PS algorithm running on the contemporary CPUs, and up to 10x speed up comparing to the PS algorithm running on GPUs.

Phase change materials, which play an important role in the development of rewritable media and nonvolatile computer memory, work on rapid and reversible change between crystalline and amorphous states. In the past, phase change materials based on the Ge-Sb-Te (GST) family have been received considerable attention for memory device applications. In the GST family, Ge2Sb2Te5 has superior performance in terms of speed and stability. GST has two crystalline phases: a stable hexagonal high temperature phase and a metastable cubic phase. While the structures are now relatively well understood, the dynamics of phase transitions is largely unknown. Here we employ ab initio molecular dynamics simulations to study the density and momentum fluctuations during phase transformation in Ge2Sb2Te5. We start from the hexagonal structure with the stacking sequence Te-Ge-Te-Sb-Te-Te-Sb-Te-Ge that corresponds to the lowest total energy. The amorphous structure is then obtained by melting and quenching the crystalline structure. The dynamical fluctuations of the phase change transitions are then computed by evaluating the density correlator (intermediate scattering function), current correlations and the dynamic structure function.

Recently GaN-based nitrides and alloys have attracted attentions in light-emitting diodes (LEDs) and laser diodes (LDs). In particular, the InGaN/GaN multi-quantum wells structures are researched for device applications such as high-power blue (5mW) and green (3mW) LEDs [1, 2]. For optimization of emission efficiency in such devices, a detailed understanding of local strain conditions and material composition within the InGaN/GaN MQWs are required. Thus, it is important to obtain the accurate strain of materials on an atomic scale for the desired device applications. Local strains in nanometer-scale size devices were measured using convergent beam electron diffraction (CBED) method due to its excellent spatial resolution. Here, we introduce an improved method for measuring lattice distortion in InGaN/GaN MQWs using CBED. Experimental high order Laue zone (HOLZ) lines in the central CBED disks are compared with simulated lines because the position of HOLZ lines are very sensitive to lattice parameters. We used kinematic theory to simulate HOLZ lines. Experimental CBED were carried out with off-zone axis which tilted from low zone axis to reduce dynamic effects. In order to improve reliability of the results, we evaluated the sensitivity of each HOLZ line to each simulated lattice parameter. In addition, we considered the possibility of dynamic effects in experimental data compared with dynamic simulation. Based on these results, we decided the weights of each HOLZ line. The method how to apply the weights to CBED pattern matching and the effects of applying the weights on results of CBED pattern matching will be discussed. Through this method, a new automatically CBED pattern matching will provide the lattice strain of InGaN/GaN MQWs related to electronic properties. References: [1] S. Nakamura et al, Jpn. J. Appl. Phys. 35, pp. 74-76 (1996) [2] S. Nakamura et al, Jpn. J. Appl. Phys. 34, pp. 797-799 (1995)

Molecular dynamics has enabled a mechanism-focused viewpoint that has brought forth many explanations for mechanical properties bulk phenomena. It is massively parallelizable which enables the study of systems are larger in space. Upscaling in time is inherently less straightforward as regardless of the parallelization spatially, the requirement of a small timestep ~1fs has limited MD from the study of longer duration phenomena such as statically mechanical failure, materials aging, and creep. Recent advances have been made with the introduction of metadynamics in which a hybrid approach of energy minimization and dynamic simulation is used to overcome the time limitation while maintain physicality in the system. This study is concerned with the application of this method to predict time scales and morphology of statistical failure of graphene and 2-dimensional Silicon. Additionally, this work addresses the concerns of unrealistic strain rates that many hold for MD simulations when they are compared with experimental results. The potential for graphene, in electronics, heat transfer interfaces, and structural materials, more With great application driven interest for these materials, this work provides guidance in terms of durability and performance potential in realistic settings. The traditional Autonomous Basin Climbing (ABC) method has the problem of not taking into the account the degenerate states. In this work we represent a new algorithm in order to overcome such shortcomings and explore the energy landscape in an unbiased nature. In this method the penalty functions (frequencies) are distributed randomly over the degrees of freedom. Each time a new state is discovered the system is reset to the initial state and new random penalty frequencies are reassigned. The technique presented here has proven robust and capable of addressing previous concerns of accuracy and extendibility.

A genetic algorithm (GA) is coupled with a crystal plasticity based finite element (CPFE) in order to optimize the microstructure of a wrought magnesium alloy for maximum strength and ductility. The variables used in the simulations are the grain size in the sub-micron range, the crystallographic texture, and the grain misorientation distribution. The grains are assumed to contain an initial population of deformation twins with a spacing, which is considered another microstructural variable. It is assumed that plastic deformation occurs mainly by dislocation slip where the hardness of a slip system is scaled by the length-scale of the twin. In addition to the conventional slip systems, dislocation glide along a compression twin plane is also assumed to exist. The crystal plasticity code calculates the stress-strain response of the microstructure as a function of the microstructural parameters and the length-scale of the features. A failure criterion based on a critical strain and a critical hydrostatic stress is used to define failure. In the genetic code each "chromosome" is defined in terms of 3*N genes where N is the number of grains in the system and each gene carries the three Euler angles sequentially for each grain. The fitness function is defined in terms of the strength and the ductility. The initial population contains a random assignment of grain orientations and twin spacings. The CPFE code and the GA are coupled in parallel so that new generations are created and analyzed dynamically. The genetic operations of mutation and crossover are performed and the fitness of a new generation in terms of strength and ductility are computed using crystal plasticity. By using a Pareto-frontal approach, the microstructural features that provide the maximum strength and ductility combinations are obtained. The implications of the computations for the design of a wrought magnesium alloy are discussed.

The current standard material for high energy density capacitor dielectrics is biaxially oriented polypropylene (BOPP), which has a remarkably high electrical breakdown strength and large band gap, but low dielectric constant. The envisaged next generation capacitor dielectric with even higher electrostatic energy storage capacity should provide high dielectric constant, while still preserving the insulating characteristics of BOPP (such as large band gap and high electrical breakdown strength). To meet these growing needs, we use high throughput DFT calculations in combination with cluster expansion to solve the “inverse problem”, namely, identification of classes of polymers with large dielectric constant and band gaps. In our approach, we consider various possible local chemical modifications to polyethylene (PE), which would result in much larger dielectric constant than PE, while still preserve its insulating properties. PE is chemically similar to BOPP but with a much simpler physical structure. To be specific, we allow the -CH2- unit in the PE backbone segment to be replaced by -NH-, -C(=O)-, -C6H4- (benzene), -C4H2S- (thiophene), -C(=S)-, or -O- units in a systematic, progressive, and exhaustive manner. All of these units are commonly seen in polymer backbones. Different combinations of these units will allow us to explore various polymer systems, such as polyesters, polyamides, polyethers, polyureas, etc. In the present study, a newly developed high throughput method is used to accurately estimate the dielectric constant of the chemically modified PE chains for a set of limited compositions and configurations through carrying out density functional perturbation theory (DFPT) computations in combination with effective medium theory. Furthermore, by successfully parameterizing a cluster expansion from the DFT computed data set we are able to predict the dielectric constant of systems spanning a much larger part of the configurational and compositional space. Based on this screening strategy, we have identified several polymer sub-classes which offer a better tradeoff between the dielectric constant and the band gap.

We applied Inverse Design approach to predict and realize previously undocumented stable V-IX-IV semiconductors with V= V, Nb, Ta, IX=Co, Rh, Ir, and IV= C, Si, Ge, Sn, Pb. This V-IX-IV group contains 45 possible compounds, out of which 15 are documented and 30 are missing from citable sources. We are particularly interested in these V-IX-IV compounds, because based on the crystal field theory, it is expected to include new semiconductors, despite the constituent elements being largely metallic. We theoretically determined which of the 30 missing materials are stable and which ones are unstable. Then we experimentally synthesized and characterized TaCoSn, which is one of the newly predicted stable V-IX-IV materials. For each of these V-IX-IV materials we theoretically calculated (i) the lowest energy crystal structure, and (ii) the thermodynamic stability region of each compound with respect to decomposition. For these calculations we used the first principles thermodynamics approach [Advanced Functional Materials 22, 1425, 2012]. Out of 30 missing V-IX-IV materials, we predict that 9 are stable, 20 are unstable, and 1 is too close to call. Among the 9 predicted new thermodynamically stable V-IX-IV compounds, there are 2 silicides, 3 germanides, and 4 stannides. All 18 stable silicides and germanides (except TaIrGe) assume orthorhombic MgSrSi-type structure, and all 6 stable V-IX-IV stannides (and TaIrGe) appear in the cubic AgMgAs-type crystal structure. Sn-based V-IX-IV materials made of all metallic elements in the AgMgAs-type structure are semiconductors, while the Si- and Ge-based V-IX-IV materials in MgSrSi-type structure are metals. We have then experimentally synthesized and characterized one of the predicted stable V-IX-IV materials, TaCoSn. Thin films were deposited using combinatorial radio-frequency co-sputtering from Ta, Co and Sn targets on fused silica substrates heated to 600C in Ar atmosphere. Positions and intensities of the XRD peaks of TaCoSn thin films are consistent with the theoretically predicted AgMgAs-type structure and Oh-coordinated Co atoms. Small amount of metallic impurity (possibly TaCo2Sn) obscured experimental measurement of absorption spectrum of the thin films. TaCoSn bulk powders were synthesized from Ta and Co nanopowders in evacuated fused quartz ampoules at 600C for 10 days. The products of the reaction included TaCoSn as the major phase, but with Co2Ta, CoSn, and CoSn2 impurities. Despite the presence of impurities, it was possible to do the XRD refinement and determine the experimental lattice parameter of TaCoSn of 0.594 nm, as compared to the 0.597 nm from theoretical prediction. Overall, we believe that this computationally driven Inverse Design approach to new materials is better suited to the modern world with the ever-decreasing price and abundance of computational resources. This work is supported by the U.S. DOE, as a part of the “Center for Inverse Design” EFRC

In Inverse Design of materials we generally first the functionality or property required, then we search for the material or configuration that has this functionality. Different circumstances require different approaches to this Inverse Design problem, all requiring a theoretical/computational step upfront. Modality 1 applies to cases where a single material system is considered, such as Si/Ge of GaAs/AlAs , but despite the chemical simplicity , the system can have a very large ( in fact, astronomic) number of spatial configurations that can in principle be realized . The strategy applied to modality 1 problems involves (a) calculating the final property “on the fly” for various assumed ( equilibrium or non-equilibrium) configurations [or via a surrogate model such as cluster expansion ] ,guided by (b) Genetic Algorithm or simulated annealing to find the configuration with the target property ,followed by (c) laboratory realization of the “best of class” case. Modality 2 applies to cases where numerous material systems need to be considered (such as different ternary semiconductors documented in ICSD ), and the materials are known, and so the crystal structure are also generally documented and need not be predicted. This modality then deals with materials where the structure and composition are known from previous studies, but the properties are unknown. Physical properties or functionalities sought can include, photovoltaic absorption , transparent -conductivity , photo electrochemical water-splitting ability , or thermoelectricity. The strategy applied to Modality 2 involves (a) identification of a calculable metric that is simpler to compute than the final property sought, but can act as a marker/descriptor for it .Then,(b) the marker is calculated, high throughput style ,for all the chemical compounds in the group at the knows crystal structure ,the result are sorted, and “best of class” identified and (c) laboratory realization of best of class attempted. Modality 3 applies to cases where numerous material systems need to be considered, but they were not previously known. Examples include compounds that can be hypothesized ,but are undocumented in ICSD . This modality then deals with materials where the structure and composition are unknown and so do the properties . The strategy applied to Modality 3 involves (a) predict the stable crystal structure using either a fixed-list of candidate structure-types or the “Global Space Group Optimization” which can start from zero knowledge on the Lattice-vectors and weikpf positions, (b) examine the stability of the predicted stable crystal structure with respect to decomposition to any possible sub-system .Once a stable new material has been identified it can be subjected to the strategy of modality 1 or 2 . This work is done in collaboration with all experimental and theoretical scientists at the Energy Frontier Research Center on Inverse Design.

It is a timely goal in the biological and materials sciences to computationally design novel compounds that exhibit specific chemical properties and are straightforward to synthesize. Some of the most relevant and promising materials properties depend explicitly on atomistic details, rendering an atomistic resolution of any employed simulation model mandatory. Alas, even when using high-performance computing, brute force high-throughput screening of all the possible compounds is beyond any capacity for all but the simplest systems and properties due to the combinatorial nature of chemical compound space (compositional, constitutional, and conformational isomers). Consequently, when it comes to properties or systems that require first principles calculations, a successful optimization algorithm must not only make a trade-off between sufficient accuracy of applied models and computational speed, but must also aim for rapid convergence in terms of number of compounds "visited". I will present recent contributions related to the latter aspect. More specifically, I will explain the notion of fractional nuclear charges and their use for calculating ``alchemical'' property gradients in chemical compound space [1]. Examples for molecular interaction energies and frontier orbital eigenvalues will be given [2]. Finally, I will discuss promising preliminary applications to the design of catalysts and heat transfer fluids. [1] OAvL, R. Lins, U. Rothlisberger, Phys Rev Lett (2005); OAvL, M. Tuckerman, J Chem Phys (2006); OAvL, J Chem Phys (2009); [2] OAvL, M. Tuckerman, J Chem Theory Comput (2007); V. Marcon, OAvL, D. Andrienko, J Chem Phys (2007); [3] D. Sheppard, G. Henkelman, OAvL, J Chem Phys (2010); S. Jayaraman, A. Thompson, OAvL Phys Rev E (2011).

Advanced materials with new functionalities often require a combination of several materials properties which are difficult to achieve simultaneously. In this situation, the Inverse Design of Materials (for a given desired property, find the material that has it) can greatly accelerate the materials discovery and development. Transparent conducting oxides (TCO) are an example where the desired functionality requires counter-indicated materials properties, i.e., transparency and conductivity. While good n-type TCOs, such as In2O3:Sn (ITO) and SnO2:F (FTO), are known and in commercial use, achieving p-type conductivity in transparent wide-gap oxides is much more difficult. We address this problem in the Inverse Design approach, which tightly couples theory and experiment, and integrates the three steps of (1) formulating design principles, (2) screening of a wider range of materials for multi-target properties, and (3) the final realization and validation of new materials by design. As a design principle, we proposed d5 oxides, which show a strong p-d coupling not unlike the prototypical Cu containing p-type TCO like, e.g., CuAlO2. Starting from a larger number of ternary Mn(II)-oxides (d5 high-spin) as candidate materials, including A2MnO4 (A=Ga, In, Cr) in the spinel structure, Mn2BO4 (B=Si, Ge, Sn) in both olivine and inverse spinel structures, and MnSnO3 in the ilmenite structure, We use ab-initio theory methods to down-select the candidate materials. The target properties include (i) thermodynamic stability under accessible growth conditions (ii) a sufficiently large band gap and absorption threshold, (iii) a sufficiently light effective hole mass (iv) the absence of hole-killer defects, (v) the absence of hole self-trapping (i.e., band- rather than small-polaron conductivity), and (vi) p-type doping by suitable impurities. Experimental synthesis and characterization of optical and electrical properties is performed for Cr2MnO4 and Mn2SnO4 in order to validate and realize the predicted materials. We identify Cr2MnO4 as a thermodynamically stable, wide-gap, and p-type dopable oxide with the characteristics of a band-conductor for holes (no carrier self-trapping). Since, however, Cr2MnO4 lacks a native intrinsic hole-producer (acceptor-type defect), we screen for suitable dopant impurities that would have sufficiently high solubility and shallow acceptor ionization energies, finding by theory that Li should be the best dopant. Bulk electrical measurements confirmed the expected insulating behavior of pure Cr2MnO4 and Mn2SnO4, and demonstrated a dramatic increase in the conductivity of Cr2MnO4 after Li-doping. This presentation highlights the individual steps in the Inverse Design approach that lead from a desired functionality to the realization of a best-of-class material. Supported by the US Department of Energy, Office of Basic Energy Sciences as part of an Energy Frontier Research Center.

The development of new energy technologies is an urgent endeavor that advanced modeling can help to accelerate. But as capabilities in theory and modeling continue to progress, it becomes ever more difficult to integrate the most sophisticated theoretical knowledge into experimentally verified, quantitative, commercially relevant models. Among scientists working in physical chemistry, there is also a frequent disconnect between experimentalists and theorists, based on a mutual distrust of the others' methods. But there may be peace to be had through statistics. From a purely scientific perspective, the use of ab initio methods to define probability distributions for physical parameters, in conjunction with scientifically accurate models applied to experimental observations, combines the insight of first-principles methods with the grounded reality of experiment. This leads to the best available estimate of physical quantities, while honestly accounting for the uncertainty associated with both theoretical and experimental approaches. From an engineering standpoint, a rigorous apportionment of error between parameters and model forms enables the quantitative assessment of uncertainty due to upscaling and model extrapolation, thereby potentially enabling and justifying the use of the most sophisticated physical understandings. Bayesian methodologies for model calibration and uncertainty propagation developed in the context of the Department of Energy's Carbon Capture Simulation Initiative will be presented, and their roles in the bridging of scales between quantum, chemical kinetic and process/device scales demonstrated.

The III-V system of semiconducting alloys (Al, Ga, In cations and N, P, As anions) is investigated through the use of a generalized cluster expansion. This method provides a general expression relating the state of atomic order of an alloy to various tensor-valued properties, such as static strain, elastic constants, carrier masses and dielectric properties. The unknown coefficients in this general expansion are determined via first-principles calculations. The resulting structure-property relationships can provide helpful guidance in the tailored design of optoelectronic devices.

The electronic structure of binary and ternary AlxBiyNz nanoclusters where x+y+z le; 5 is studied using the density functional theory (DFT), with the pure functional M06L of Truhlar and Zhao, with atomic basis 6-311+G(3df) for Al and N (this include 3 sets of d diffuse functions), and the effective core potential (ECP) LANL2DZdp (a double zeta basis with polarization and diffuse functions) for Bi. The initial geometries are obtained by Montecarlo technique to find global minimum using simple Hartree Fock method with the ECP LANL2DZ over all the atoms in the clusters. These geometries are refined with the mentioned DFT method; the resultant clusters are re-optimized to find the lowest energy geometries. Our results are compared with those obtained experimentally by other groups for some binary AlxBiy neutral and anionic clusters, and characterized by photoelectron spectroscopy and time of flight mass spectroscopy. We also compare with other authors results using DFT and other perturbative methods. Our results show good behavior and are consistent with previous works. The analysis is done in terms of orbitals, charge distribution, vertical detachment energy, ionization potential and electron affinity. As shown experimentally N5 clusters break into two clusters N2 and N3. This research has received funding from the European Community Seven Framework Programme (FP7-NMP-2010-EU-MEXICO) and CONACYT under grant agreements 263878 and 125141, respectively.

Force-biased Monte Carlo is a useful addition to a Molecular Dynamics code, since the force algorithm is already available so that a few extra lines are sufficient to give access to an entirely different and complementary simulation technique. MC and MD can take turns in an alternating sequence, where MD takes care of the violent changes and MC handles the equilibration processes. In this presentation we show how the particular uniform-acceptance MC version that we use - i.e., all stochastic moves are accepted - outperforms MD in diffusion, phase transformation (Si), and nanotube growth (C). It also outperforms Metropolis MC. The method appears to work always, does not depend on event detection, and is not sensitive to, e.g., very low activation barriers. New in this work are the proof of a rigid statistical mechanics underpinning of the method and the introduction of statistical time, allowing this MC technique to provide unbiased estimates of activation kinetics. We will also discuss recent methodic refinements and indications that the method may in fact not “always” work.

Van der Waals (vdW) dispersion interactions are essential for determining the structure, stability, and function for a wide variety of molecules and materials. However, the non-local vdW interactions are missing in the widely employed density functional theory calculations, that are typically used as the "base" in multiscale modeling approaches. In condensed matter, cohesion is thought to arise mainly from covalent bonding and/or electrostatic interactions, hence it is typically assumed that vdW forces play a minor role. We have recently developed a set of efficient methods for an accurate description of (screened) vdW interactions in molecules, solids, and interfaces [2]. We discuss the challenges for an accurate theoretical modeling of vdW interactions in condensed matter systems and demonstrate that vdW interactions play a significant role for their cohesive properties. Applications will be presented for a few fundamental systems: phase diagram of ice, cohesion in ionic and semiconducting solids, and stability of hybrid inorganic/organic interfaces. For all these cases it is found that vdW interactions play a noticeable if not crucial role, not just for quantitative energies but also for the qualitative behavior. Ongoing work and the remaining challenges in the modeling of vdW interactions will be discussed. [1] A. Tkatchenko et al., MRS Bulletin 35, 435 (2010). [2] Phys. Rev. Lett. 102, 073005 (2009); Phys. Rev. Lett. 107, 185701 (2011); Phys. Rev. Lett. 107, 245501 (2011); Phys. Rev. Lett. 108, 146103 (2012).

An algorithm based on kinetic Monte Carlo (KMC) for spinodal decomposition of a simple binary mixture was presented. For the algorithm, we considered free energy barrier calculated from the phase field model. We also considered that diffusive jump of particles among discrete compartments is constrained by the free energy barrier. These considerations were combined to develop an efficient KMC algorithm namely free energy-limited next reaction method (FENRM). By constructing the diffusive master equation for the discretized system, we also demonstrated that the biased diffusive jump governed by Boltzmann distribution is mathematically in accordance with the governing spatio-temporal differential equation known as the Cahn-Hilliard equation of for the spinodal decomposition of a binary mixture. Computer simulations with different initial conditions based on the FENRM exhibited typical temporal evolution of microstructures in the course of the phase separation governed by spinodal decomposition. The physical validity of the proposed algorithm is also examined by comparing with different numerical calculations. The presented KMC algorithm demonstrated its ability to describe dynamics of unstable system by capturing most of the critical characters of the microstructure evolution and its dynamic properties. We expect that the presented stochastic algorithm can be extended and applied to simulate more complicated dynamic systems involving unstable multi-phase multi-component mixture and reaction-diffusion system with phase separation.

It is about 25 years since density-functional theory has become applicable with sufficient accuracy to real materials. This quarter of a century of expertise has enabled the DFT community to gain a good insight in the strength and weaknesses of DFT predictions. The community is now clearly evolving from explanatory to predictive studies, with “computational materials design” as the often-promised ultimate goal. In this contribution, we will discuss how to quantify the expected accuracy of DFT-predictions. Only when knowing the “error bar” on DFT predictions, computational materials design efforts make sense [1]. Not all materials design problems are equally suited for a computational (DFT) approach. We will spend some thoughts on the question “to which kind of design questions DFT can make useful contributions?”. We will identify some of these questions in the context of materials development for nuclear fusion applications (post-ITER), and we will report on a case study in which a computational strategy is used to narrow the search space for candidate materials for the first wall of fusion reactors. [1] K. Lejaeghere et al., "Error bars for solid-state density-functional theory predictions: an overview by means of the ground-state elemental crystals", http://arxiv.org/abs/1204.2733v2

Complex phenomena arise at solid-liquid interfaces, leading to surface induced changes that are not only important for the solid but also for the liquid. In that respect, water plays an important role in a number of interfacial phenomena encountered in biological, chemical and physical processes. Water properties at the interface can be quite different with respect to bulk properties and have been subject of recent research efforts. Here we present some simulation results on solid/water, namely quartz/water and alumina/water and water/vapor interfaces based on Density Functional Theory. We aim to address the molecular details of the solvation structure and to include the electronic polarization effects. We calculate the acidity of oxide surfaces and we discuss the water structure at the interface to interpret recent experimental results from surface sensitive Sum Frequency Generation.

Pyrochlore oxides exhibit many interesting physical phenomena, for example, with a 3d transition element at B-site and a rare earth at A-site, magnetic behavior from paramagnetism to ferro- or antiferro-magnetism may be observed, the electrical nature varies from insulator to conductor, the structure of pyrochlore can be transformed to anion-deficient fluorite structure by disordering of similar A-site and B-site atoms. Due to the difficulty and high-cost in micro-scale experiments, molecular dynamics simulations of multiphysics attract much attention on the study of material properties and behaviors. An atomistic model, capable of simulating thermomechanical-electromagnetic coupling, is briefly introduced in this presentation. It is based on Coulomb-Buckingham potential and upgraded Nose-Hoover thermostat. It has been verified that, without any external field, total energy (kinetic plus potential), linear momentum, and angular momentum are conserved. For a class of materials, A2B2O7 pyrochlore oxides, we have a transferable database. One may design a pyrochlore oxide x [A2B3O7] + (x-1) [C2D2O7], where is the design variable, such that it acquires certain specified physical properties. In other words, this model serves the purpose of material design. This model also enables the study of the propagation of wave with both acoustic and optical modes, mechanically induced electric field and temperature, and effects of electromagnetic wave. Numerical results of a few sample problems and their physical meanings will be discussed.

A new model, based on statistical thermodynamics and kinetics theories, is introduced to simulate the physical properties of different metastable states in liquid. In a statistical thermodynamic view, a physical state is fixed in equilibrium as long as the macroscopic variables of the system are determined, such as temperature (T), pressure (P) and particle number (N) [1][2]. However, according to the kinetic route of development, while this physical state illustrates the most probable and stable microstates distribution, there are many more metastable states that the system could fill in on the way of approaching its equilibrium. Since the state-of-the-art theory cannot fully explain or predict the final physical state, we thus introduce, in this work, a new "molecule cluster" model to simulate the situation aforementioned. A liquid system, as it were, is composed of various clusters, in which a specific number of molecules are bounded by weak interactions within the clusters. In different physical states, however, the distribution of these molecule clusters varies, in terms of the number, the type and the proportion of molecules [3]. We point out that, to determine the physical state of liquid in equilibrium, macroscopic physical properties, including volume [4], dielectric constant, refractive index and optical rotation, are required. In our experiments, different kinetic routes (e.g., temperature, mixing sequence, etc.) are designed for the same system under final specific conditions. We suggest that the macroscopic properties are tested to demonstrate the many metastable states of the system. Our results verified the accuracy of this "molecule cluster" model. [1] L. D. Landau, E. M. Lifshitz: "Statistical Physics", Oxford, Pergamon Press, 1958. [2] T. B. Schroslash;der, N. P. Bailey, U. R. Pedersen, N. Gnan, J. C. Dyre: "Pressure-energy correlations in liquids. III. Statistical mechanics and thermodynamics of liquid with hidden scale invariance", Chem. Phys. 131, 234503 (2009). [3] P. J. Flory, J. Am: "Statistical Thermodynamics of Liquid Mixtures", Chem. Soc. 87, 9 (1965). [4] R. K. Shukla, A. K. Shukla, R. D. Rai, J. Pandey: "Statistical Thermodynamics and Excess Volume of Quaternary Liquid Mixtures", Chem. Phys. 93, 11 (1989).

Interfacial stiffnesses and energies are computed by capillary wave analysis of simulated inhomogeneous solid-fluid systems at coexistence. Results for Ni (fcc) and Ti (bcc) (modeled with EAM potentials) are compared to the ones obtained for Lennard-Jones (LJ), hard-sphere (HS) and Yukawa systems. The latter system is particularly interesting because it exhibits both fcc and bcc crystalline phases. In agreement with the findings of Laird [J. Chem. Phys. 115, 2887 (2001)], we observe that in reduced units the orientationally averaged interfacial free energy γ₀ for different systems shows similar values of γ₀~0.6k_BT_m/σ² for the fcc-fluid interface and γ₀~0.3k_BT_m/σ² for the bcc-fluid interface where T_m is the coexistence temperature and σ the particle diameter. However, in the Yukawa system the reduced interfacial free energies are much lower. We relate these results to the density difference between the coexisting phases and correlate this density difference with the entropic contribution to the interfacial free energy.

Modeling the mesoscale (~100 nm) morphology of polymeric semiconductors is challenging. On the one hand, substantial coarse-graining is required for addressing efficiently the large time and length scales involved. On the other hand, the mesoscale structuring of these materials is significantly affected by the local details of molecular structure (e.g. side chains) and interactions (e.g. pi-pi stacking). Thus albeit coarse-grained the models must account for these features, contained in the finer non-resolved scales. Towards meeting this challenge, we present here a strategy for developing such models through a combination of systematic coarse-graining [1] with more phenomenological descriptions. As a case study we consider P3HT systems. The polymer architecture is represented through a continuum analog of the rotational isomeric state model where each coarse-grained monomer represents a single hexathiophene unit. The bonded interactions (i.e. torsional and bending potentials) are obtained from underlying atomistic configurations through the efficient VOTCA package [2]. The non-bonded interactions of the coarse-grained hexathiophenes are captured by combining soft DPD-like isotropic and orientation-dependent potentials [3]. The former control the compressibility, while the latter lead to liquid-crystalline mesophases with a stack-like structuring of the material. With the above model, Monte Carlo simulations of systems with device-relevant dimensions (~ 50 nm) were performed. We will discuss the effect of the molecular weight on the morphology of the material. In the stack-like structures we focus on the conformations of the polymer chains, due to their importance for charge transport properties. The global molecular shape in the stacks will be quantified from the analysis of the molecular gyration tenzor. [1] C. Peter and K. Kremer Soft Matter 2009, 5, 4357. [2] Ruehle et al J. Chem. Theory Comput. 2011, 7, 3335. [3] K. Daoulas, V. Ruehle, K. Kremer J. Phys.: Condens. Matter 2012 in press

Hydrogen has detrimental influence on the stability of many metals and steels, as it leads to embrittlement and therefore to fracture. As a prototype of such a material we investigate Nickel, where a hydride can form, depending on the hydrogen concentration, temperature and local stress state. This effect is amplified by an effective attractive hydrogen-hydrogen interaction. We present a model which allows to bridge the atomistic level of Monte Carlo simulations of the Ni-H two-phase system to the continuum level. Taking into account configurational entropy, an attractive hydrogen-hydrogen interaction, mechanical deformations and interfacial effects, we obtain a fully quantitative agreement with a precision of the order of 20 meV per atom in the chemical potential, without the need for any additional adjustable parameter. We find that nonlinear elastic effects are crucial for a complete understanding of the phenomena, including the prediction of the phase diagram with and without elastic effects. The results of this analysis is then used in mesoscale simulations for the formation of hydride zones around crack tips. To this end we simulate the diffusive flux of hydrogen towards the tensile regions of the crack, and investigate the growth of the hydride zone and its steady state size depending on the stress intensity factor and the hydrogen supply. The results are in agreement with analytical scaling relations.

The coarsening kinetics of Cu-rich precipitates in an Fe-rich matrix for thermally aged Fe-Cu-Ni-Mn alloys at temperatures above 700°C is studied using a kinetic Monte Carlo (KMC) and a phase-field method (PFM), connecting them via appropriate sequential parameter transfer. In a three dimensional domain, we first use KMC simulations to study the early stage of the system evolution which involves nucleation, growth and coarsening and then we employ PFM to study the late stage coarsening behaviour. We use a quantitative PFM based on the grand chemical potential formulation, which uses the final microstructure from KMC as initial configuration and CALPHAD data for the thermodynamic description of the system. We show that our phase-field model can be validated quantitatively for the Gibbs-Thomson effect and it also predicts the coarsening kinetics correctly. It is found that the kinetics closely follow the LSW (Lifshitz, Slyozov, Wagner) temporal power law whereas the coarsening rate constant increases with an increase in volume fraction of precipitates.

Liquid Metal Embrittlement (LME) has a destabilizing effect on many metallic components which undergo the process of hot dip galvanizing, where samples are subjected to thin coating films. Specifically, grain boundaries are prone to the penetration by thin liquid films at the surface, most prominent example probably the embrittlement of Aluminium samples subjected to thin Gallium films. In LME experiments, different stages of the embrittlement process have been observed. In the present study we focus on the initial stage where elastic contributions are not yet present. Our model describes how wetting affects the grain boundary melting process at this stage. The combination of analytic and numeric methods allows us to predict the penetration velocity into the grain boundary and the thickening of the penetrating film. These predictions cover fundamental aspects of the incubation stage and provide a firm basis to describe also the dominant factors for the onset of the subsequent phases of the embrittlement process.

Composite materials, composed of rod-like particles in a polymeric melt are of great relevance in materials science. In the resulting product, the rods increase the stiffness and stability of the material. If they are aligned, they provide an anisotropic elastic modulus, allowing to develop tailored materials for specific applications. One typical example are fibers in polymer melts used in molding processes. In theories of dissolved rods, like the Folgar-Tucker theory or the Hess-Doi theory, a spatially homogenous distribution of rods is assumed. We combine the Hess-Doi model with the phase field theory in order to study systems with inhomogenous rod concentrations. With this method, molecular aspects are considered in simulations on long time and length scales study. The method provides a relation between microscopic material properties and the spatial and orientational rod distribution. We study the development of coexisting regions of high and low rod concentration and the corresponding orientational order. The combination of rod concentration and alignment leads to a complex phase behavior. If the surface energy of the domains depends on the orientation of the rods, anisotropic nucleation is found (as known from tactoids in lyotropic liquid crystals) and spinodal decomposition changes qualitatively. In various examples, we demonstrate the versatile properties of the system.

A correct description of hydrogen diffusion in metals is a prerequisite for understanding the phenomenon of hydrogen embrittlement. The H diffusion in bulk materials has been studied extensively in the past both experimentally and theoretically. Nevertheless, the knowledge of diffusion processes in distorted environments, e.g. in the vicinity of extended crystal defects such as dislocations, grain boundaries or crack tips, is still limited. In this multiscale study, we investigate the H diffusion in uniformly strained lattices of α-Fe. First, classical diffusion barriers are obtained using accurate first-principles calculations, based on the density functional theory and a semi-empirical tight binding approach. Since the H diffusion at low and ambient temperatures is affected by quantum effects, the classical barriers are corrected using a path-integral approach as well as more approximate theories. Finally, the quantum-corrected diffusion barriers are used as input in kinetic Monte Carlo simulations to obtain diffusion coefficients under various loading conditions. Our results show that the H diffusivity is indeed strongly influenced by the lattice distortions and depends sensitively on the type of deformation. This research is funded by the European Commission through Contract No. NMP.2010.2.5-1.263335 (MultiHy).

Pure Ti has a body-centered-cubic (bcc) structure, the β phase, at high temperatures and a hexagonal-close-packed (hcp) structure, α phase, at low temperatures. Ti alloys for structural applications, however, usually contain α+β two-phase microstructures. Different thermo-mechanical processing routes produce a wide spectrum of complex (α+β) two-phase microstructures from fully lamellar structure (or basket-weave and Widmanstätten structures) to bimodal (duplex) structure containing lamellae with primary α phases displaying globular morphology. This presentation will discuss the understanding and prediction of phase transformation kinetic pathways leading to different microstructures in binary or ternary Ti alloys using both phase-field simulations and thermodynamic stability analysis. The strain energy contribution is taken into account by assuming the Burgers orientation relation between β and α phases and Khachaturyan&’s microelasticity theory. The mechanisms of grain boundary α plate formation and α variant selection behavior in polycrystalline β grain structures will be discussed.

There has been recent interest in constructing realistic polycrystalline microstructures based on a given grain size distribution. Such configurations may be analyzed using various simulation techniques to develop structure-property relations, which have implications for property optimization with respect to microstructure. Methods based on the reverse Monte Carlo algorithm have been shown to be successful in reproducing size distributions; however, such methods have not yet been related to physical growth processes. In the present study, a reverse Monte Carlo algorithm is developed that is consistent with the nucleation and growth mechanism. Applied to a specific material system with known nucleation and growth rates vs temperature, our method may determine a thermal history that would produce a given grain size distribution.

We use data over an extended Q range from 0.01 to 100/A from the recently commissioned NIMROD instrument at the ISIS pulsed neutron source to drive a multiscale inverse modelling procedure to gain insight in to the phase transformations of polymer systems. The first level of our procedure is atomistic and we use internal coordinates (Bond length, bond angles and torsion angles) to define the polymer chain in full atomistic detail. Values were assigned to each internal coordinate within the chain using a stochastic Monte Carlo method in which the probabilities were drawn from distributions representing the possible range of values. For example bond lengths were represented by a Gaussian Distribution defined by 2 values, namely a mean value and a width. Using this approach random chain configurations could be rapidly built and the intrachain structure factor calculated utilising a small set of parameters. We evaluated each structure factor using the chi squared test. Parameters representing the probability distribution functions were systematically varies using a grid search to find the values which gave the lowest chi squared test .use the structure factor for Q > 3/A to determine the details of the chain conformation in the molten phase. This process was repeated for data over the same extended Q range obtained at lower temperatures where the polymer was expected to crystallise. Polymer crystallise via chain-folded thin lamellae crystals. Such crystals give rise to an intense peak at Q ~ 0.03/A. This scattering was calculated using a lamellar stack model, coarse grained from the single chain structure. The crystal peaks at Q ~ 1.5/A were calculated from the crystal structure scaled by the volume of crystallisable material obtained by coarse graining the single chain conformation. We report on the effectiveness of this approach using data obtained on the crystallisation from the melt phase of perdeuterated polyethyleneoxide. The objective here is follow the three key length scales; the chain folded lamellar thickness of ~ 10nm), the crystal unit cell ~ 1nm and the detail of the chain conformation is ~ 0.1nm. The availability of data over the extended Q range available on NIMROD which relates to these three length scales transforms our ability to make progress with this challenge problem.

Deposition of semiconductor films is a key process for production of thin-film solar cells, such as CdTe or CIGS cells. In order to optimize photovoltaic properties of the film a comprehensive model of the deposition process should be build, which can relate deposition conditions and film properties. We have developed a multiscale model of deposition of CdTe film in close space sublimation (CSS) process. The model is based on kinetic Monte Carlo method on the rigid lattice, in which each site can be occupied by either Cd or Te atom. The model tabulates the energy of the site as a function of its local environment. These energies were obtained from first-principles calculates and then approximated with analytical formulas. Based on determined energies of each site we performed exchange (diffusion) processes using Metropolis algorithm. In addition the model included adsorption and desorption processes of Cd and Te2 species. The results of the model show that a steady-state structure of the surface layer is formed during film growth. The model can reproduce transition from film deposition to film etching depending on external conditions. Moreover, the model can predict deposition rates for non-stoichiometric gas compositions.

A recently introduced hybrid Potts-Phase Field method has demonstrated the ability to evolve microstructural networks in conjunction with compositional fields tied to specific phases. In this method, Monte Carlo Potts methods are used to evolve the microstructure while phase field methods are used to evolve the composition, and the two fields are coupled through free energy functionals. By handling the microstructure with Potts and the composition with phase field, the solutions balance efficiency with accuracy. More recent developments of the model allow different alloy systems to be simulated by using thermodynamic databases and kinetic quantities to dictate the behavior. The framework is designed around energy functionals and as such, can be applied to systems with arbitrary numbers of phases. Future extensions include the handling of thermal diffusion and extensions to multi-component systems. While relatively new, this framework represents an important advance in inverse-materials solution, because it is computationally efficient, and systems can be solved quickly in lower dimensions before fully simulating conditions in higher dimensions.