Symposium Organizers
Mark Asta University of California
Alex Umantsev Fayetteville State University
Joerg Neugebauer Max-Planck-Institut fuer Eisenforschung
LL1: Mechanical Properties
Session Chairs
Monday PM, November 30, 2009
Room 313 (Hynes)
9:30 AM - **LL1.1
Making and Breaking of Chemical Bonds under Mechanical Load.
Peter Gumbsch 1 2 , Michael Moseler 1 2 , Lars Pastewka 1 2
1 , Fraunhofer-Institut fuer Werkstoffmechanik IWM, Freiburg Germany, 2 IZBS, University of Karlsruhe, Karlsruhe Germany
Show AbstractBrittle fracture as well as adhesion, or friction and wear are prominent examples for mechanical problems with clearly observable macroscopic consequences that are directly related to the processes of the formation and destruction of chemical bonds. Modelling such processes requires to propagate the atomistic information through the scales to obtain macroscopic information. If chemical specificity and chemical accuracy are required, only very few approaches are available.I will quickly review latest achievments in concurrent coupling techniques, in particular the "learn on the fly" (LOTF) technique with applications to brittle fracture of diamond and silicon. The main focus of my talk will then be the simulation of friction and wear processes between amorphous diamond-like carbon DLC films. Sequential techniques must be applied there to first obtain reasonable starting configurations for the atomic structure and topology of the films. Then different levels of approximations are required to assess the evolution of the friction contacts. Considerable attention must be paid there to extracting relevant information from large scale atomistic simulations, which in turn first requires an atomistic model for the hydrocarbons that can describe well the making and breaking of the atomic bonds. I will introduce such a new potential, report about comparison to large-scale tight binding simulations and present results for the evolution of an atomistically determined friction coefficient during running-in of such a contact.
10:00 AM - LL1.2
The Heterogeneous Multiscale Method for Dynamics of Solids with Applications to Brittle Cracks.
Jerry Yang 1 , Xiantao Li 2
1 , Rochester Institute of Technology, Rochester, New York, United States, 2 , Penn State University, University Park, Pennsylvania, United States
Show AbstractWe present a multiscale method for the modeling of dynamics of crystalline solids. The method employs the continuum elastodynamics model to introduce loading conditions and capture elastic waves, and near isolated defects, molecular dynamics (MD) model is used to resolve the local structure at the atomic scale. The coupling of the two models is achieved based on the framework of the heterogeneous multiscale method (HMM) and a consistent coupling condition with special treatment of the MD boundary condition. Application to the dynamics of a brittle crack under various loading conditions is presented. Elastic waves are observed to pass through the interface from atomistic region to the continuum region and reversely. Thresholds of strength and duration of shock waves to launch the crack opening are quantitatively studied and related to the inertia effect of crack tips.
10:15 AM - LL1.3
Multiscale Simulations of Low Speed Fracture Instabilities in Brittle Materials.
Noam Bernstein 1 , James Kermode 2 , Gabor Csanyi 3
1 Center for Computational Materials Science, Naval Research Lab, Washington, District of Columbia, United States, 2 Department of Physics, King's College London, London United Kingdom, 3 Engineering Laboratory, University of Cambridge, Cambridge United Kingdom
Show AbstractWhile it is well known that cracks in brittle materials are unstableat high speeds, we have recently shown that covalently bonded,nominally brittle materials such as silicon show instabilities at lowspeeds as well [1]. These instabilities arise from changes in theatomic structure of the crack tip that lead to macroscopic featureson the crack surface. We study this process at the atomic scale usingstate-of-the-art computer simulations dynamically coupling afirst-principles quantum-mechanical (QM) description of bonding at thecrack tip to a much larger system described with an interatomicpotential (IP). We present results on the structure and energetics ofcrack-tip reconstructions and instabilities in a range of materials,including silicon carbide, diamond, graphene, and silica. The last ofthese is particularly challenging, because of the partially ionicnature of bonding and resulting electrostatic coupling between the QMand IP region. We conclude that even very brittle single-crystalmaterials can have a complex crack tip atomic structure, and thatatomic scale rearrangements can lead to macroscopic changes in crackmorphology.[1] J. R. Kermode et al., Nature 455, 1224 (2008).
10:30 AM - LL1.4
Understanding Embrittlement in Metals: A Multiscale Study of the Hydrogen-enhanced Local Plasticity (HELP) Mechanism.
Johann von Pezold 1 , Liverios Lymperakis 1 , Joerg Neugebauer 1
1 Computational Materials Design, Max-Planck-Institut fuer Eisenforschung GmbH, Duesseldorf Germany
Show AbstractThe embrittlement of metals by H is a long-standing problem, whose underlying mechanisms are still largely unclear. In this study we consider the atomistic basis of the HELP mechanism. According to this mechanism interstitial H shields dislocation-dislocation interactions, resulting in increased dislocation densities and eventually the nucleation of cracks in regions of high H concentrations.Using a combination of density-functional theory calculations, semi-empirical EAM potentials and an effective lattice gas Hamiltonian we determine the effect of H on the stress field around edge dislocations in fcc metals. In particular, the effect of H-H interactions on the H distribution around edge dislocations in fcc metals was determined using Monte-Carlo simulations in the grand canonical ensemble. Depending on the strength of the H-H interactions, already at rather modest bulk H concentrations a hydride phase is formed in the vicinity of the dislocation core. Our results further show that the formation of the new phase induces a highly anisotropic stress response: A significant reduction in the shear stress along the glide plane of the dislocation is observed, while normal to the glide plane the shear stress is predominantly increased. An important consequence of the new phase is a weakened stress field along the glide plane that induces slip planarity and reduced dislocation-dislocation separations in dislocation pile-ups, giving rise to a substantial stress accumulation and finally the onset of localised plastic fracture.
10:45 AM - LL1.5
Crack Growth by Surface Diffusion in Viscoelastic Media.
Robert Spatschek 1 , Efim Brener 2 , Denis Pilipenko 2
1 ICAMS, Ruhr University Bochum, Bochum Germany, 2 IFF, Research Center Juelich, Juelich Germany
Show AbstractDissipation plays a central role in fracture, since typically only a small fraction on the elastic energy is used to create the surfaces of the advancing crack. Whereas in brittle materials dissipation takes place mainly close to the crack surfaces, in materials with a more viscous behavior an extended zone of bulk dissipation can form around the crack.We discuss steady state crack growth in the spirit of a free boundary problem, where growth of the crack is modeled as a surface diffusion process. It turns out that mode I and mode III situations are very different from each other: In particular, mode III exhibits a pronounced transition towards unstable crack growth at higher driving forces, and the behavior close to the Griffith point is determined entirely through crack surface dissipation, whereas in mode I the fracture energy is renormalized due to a remaining finite viscous dissipation. Intermediate mixed-mode scenarios allow steady state crack growth with higher velocities than for pure mode I.
11:30 AM - LL1.6
Molecular Dynamics Simulation Study of Shock-induced Twinning in fcc Bicrystals.
Shijin Zhao 1 2
1 Institute for Materials Science, Shanghai University, Shanghai China, 2 Theoretical Division, Los Alamos National Laboratory, Los Alamos, New Mexico, United States
Show AbstractWe study shock wave propagation in fcc bicrystals of aluminum by means of a recently introduced molecular dynamics technique, which captures the initial shock transit as well as the subsequent long time scale relaxation process. Both elastic and plastic shock fronts can be clearly identified in the initial shock transit, with an underdriven plastic wave lagging behind the elastic front. Large shear stresses generated behind the elastic shock front are greatly relieved by the partial twinning, fcc-hcp structural transition and crystal rotation behind the plastic shock front. We observe in the subsequent NVE simulation a partial-to-perfect twinning transition in the bicrystals, which results in a sudden drop in the overall pressure and a steep increase in the overall temperature.
11:45 AM - LL1.7
Magnesium < a > and < c+a > Dislocation Cores: Comparison of First-Principles and Embedded-Atom-Potential Methods Predictions.
Thomas Nogaret 1 , Joseph Yasi 2 , Louis Hector 3 , Dallas Trinkle 4 , William Curtin 1
1 Division of Engineering, Brown University, Providence, Rhode Island, United States, 2 Physics, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States, 3 General Motors R&D Center, General Motors , Warren, Michigan, United States, 4 Materials Science and Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States
Show AbstractThe use of Magnesium alloys increases due to their light weight. However their formability is poor due to their HCP structure: they deform easily along the < a > axis via < a > dislocations gliding in basal planes, but the deformation along the < c > axis is difficult due to the high stresses required for the motion of < c+a > dislocations and twinning dislocations in pyramidal planes. The pyramidal deformation modes are not well understood and constitute a great challenge for material scientists.We performed first principles and EAM potential calculations of gamma surfaces and < a > dislocation core properties in basal and prism planes, and the results were compared. One of the tested EAM potentials was found in good agreement with ab-initio calculations and used to study < c+a > dislocations in pyramidal planes. The (1-101) and (11-22) EAM potential gamma surfaces were compared to ab-initio calculations. New low energy dislocation core structures were observed and the effects of non-glide stresses on Peierls stresses were studied.
12:00 PM - LL1.8
Suzuki Effect and Stacking Fault Energies for Cu Based Binary Alloys Using First-Principles Results of Segregation Energy.
Tokuteru Uesugi 1 , Kenji Higashi 1
1 Department of Materials Science, Graduate School of Engineering, Osaka Prefecture University, Sakai Japan
Show AbstractThe stacking fault energy in fcc alloys is one of the most important factors in determining mechanical properties. Lowering the stacking fault energy increases the width of the extended dislocation and decreases the mobility of the extended dislocations. The mobility of the extended dislocations, in turn, relate to macroscopic phenomena such as the creep resistance and work hardening. Thus, it is an important goal of materials science to determine the value of the stacking fault energy for various alloys and to relate it to the mechanical properties in the macroscopic scale. In this work, the stacking fault energy for pure Cu and the segregation energy of solute atoms such as Al and Zn in the staking fault were calculated from the first-principles calculations. We presented numerical results for the stacking fault energy for the Cu based binary alloys using the results of the first-principles calculations as input parameters to an expression in the equilibrium state at 623 K. The expression for the stacking fault energy as a function of concentration for fcc binary alloys was based on three approximations and one condition: two atomic layers as the equivalent segregation sites, nointeracting solutes, a temperature-independent segregation energy, and an equilibrium state following Suzuki’s work. These numerical results are in good agreement with the experimental values at low concentration. The discrepancy between the numerical and the experimental results at high concentration most likely arises from the approximation regarding nointeracting solutes.
12:15 PM - LL1.9
First-principles Study of Solute Strengthening in Aluminum Alloys.
Gerard Paul Leyson 1 , Louis Hector 2 , William Curtin 1
1 , Brown University, Providence, Rhode Island, United States, 2 , General Motors, Warren, Michigan, United States
Show AbstractThe strengthening of Aluminum by substitutional solute atoms (Li, Mg, Si, Cu, Ge and Cr) is predicted using first principles calculations and analytic theory. Solute energies in and around an edge dislocation core are first calculated using density functional theory and a flexible boundary condition method [1]. These solute energies are then used within an analytic model that derives from concepts first presented by Labusch [2] to predict the pinning of a dislocation within a random field of solutes. Finally, the critical shear stress to overcome the pinning forces of the solutes is computed. The analysis demonstrates the role of the core solutes relative to the “far-field” solutes in determining the strengthening. Quantitative comparisons with experiments are made for several cases. [1] Woodward, C., Trinkle, D.R., Hector, L.G., Olmsted, D.L., 2008. Phys. Rev. Lett. 100, 045507 [2] Labush, R., 1970. Phys. Status Solidi 41, 659
12:30 PM - LL1.10
Stress Effects on Grain Boundary Wetting Angles.
Nan Wang 1 , Alain Karma 1 , Robert Spatschek 2
1 Phsysics Department and Center for Interdisciplinary Research on Complex System, Northeastern University, Boston, Massachusetts, United States, 2 Interdisciplinary Centre for Advanced Materials Simulation (ICAMS), Ruhr-Universitat, Bochum Germany
Show AbstractGrain boundary wetting plays an important role in a wide range of materials science problems. Equilibrium dihedral angles at solid/solid/liquid triple lines are traditionally predicted using a Young's condition together with values for the solid-liquid interfacial energy and grain boundary energy. Based on a scaling analysis of interfacial and elastic energy near triple lines, it has been previously argued that stress should not affect dihedral angles. Yet, paradoxically, a finite amount of stress can cause an apparent breakdown of equilibrium at triple lines and drive the penetration of the liquid along grain boundaries, as manifest in problems ranging from liquation cracking to liquid metal embrittlement. We present the results of a phase-field study of stress effects on grain boundary wetting that sheds light on this conundrum, with the main conclusion that dihedral angles are affected by stress for grain sizes of practical relevance.
12:45 PM - LL1.11
Identification of Descriptors for Oxygen Reduction Reaction on Solid Oxide Fuel Cell Cathodes.
Dane Morgan 2 1 , Yueh-Lin Lee 1
2 Department of Materials Science and Engineering, University of Wisconsin-Madison, Madison, Wisconsin, United States, 1 Materials Science Program, University of Wisconsin-Madison, Madison, Wisconsin, United States
Show AbstractPerovskites are the major class of materials used for modern solid oxide fuel cell (SOFC) cathodes and have the ability to catalyze the oxygen reduction reaction (ORR) on their surfaces. However, difficulties in performing in-situ characterization of well-controlled samples means that the rate limiting steps and structure-property relationships underlying ORR on these materials are not understood. In particular, to date it has not been possible to find a simple set of descriptors that can be correlated to the ORR activity. A descriptor based approach has been very valuable in understanding many reactions, including the ORR [1], on metal catalysts (e.g. d-band center descriptor). In this talk we use an ab initio based approach to identify a descriptor for the ORR in perovskite SOFC cathodes. Energetics of key steps in the SOFC ORR are calculated for LaBO3 (B= Mn, Fe, Co, and Ni) systems and correlated with oxygen surface binding, oxygen surface vacancy formation, and oxygen band center. Reasonably good linear relationships suggest that these quantities could be effective descriptors for the ORR on SOFC perovskite cathodes.[1] J. K. Norskov, et al., Origin of the overpotential for oxygen reduction at a fuel-cell cathode, Journal of Physical Chemistry B 108, 17886 (2004).
LL2: Microstructure Formation and Evolution
Session Chairs
Long-Qing Chen
Alain Karma
Monday PM, November 30, 2009
Room 313 (Hynes)
2:30 PM - **LL2.1
Multiscale Modeling and Simulation of Solidification with Crystal Defects.
Alain Karma 1 , Robert Spatschek 2
1 Department of Physics, Northeastern University, Boston, Massachusetts, United States, 2 Interdisciplinary Center for Advanced Materials Simulation, Ruhr University, Bochum Germany
Show AbstractSolidification microstructures are generally polycrystalline and accompanied by defects in the form of tangles of dislocations, vacancies, and grain boundaries.Twin dendrite growth, polygonization, grain refinement, and hot cracking are just a few examples were crystal defects can strongly influence the microstructure by altering the dynamics or coalescence of solid-liquid interfaces at different stages of solidification, and in the presence of internal or external stresses. This talk will describe a new Ginzburg-Landau model of polycrystalline solidification formulated in terms of complex amplitudes of crystal density waves. This model is rooted in models of pattern formation and can also be derived from classical density functional theory by a multiple scale analysis. This approach has the atomic scale resolution of molecular dynamics and the phase-field-crystal method, thereby making it possible to describe generally the interaction of solid-liquid interfaces with a variety of crystal defects. At the same time, it retains much of the flexibility of the conventional phase-field method for constructing free-energy functionals for different alloy systems as well as for carrying out quantitative and efficient simulations. While formulated for solidification, this model also offers interesting prospects to model microstructural evolution at high homologous temperature with elastic interactions.
3:00 PM - LL2.2
A Phase-field Study of Ternary Multi-phase Microstructures.
Daniel Cogswell 1 , W. Craig Carter 1
1 Materials Science and Engineering, MIT, Cambridge, Massachusetts, United States
Show AbstractA new multi-phase, multi-component, computational microstructure model was developed. The model addresses physical and computational deficiencies with previous multi-phase models and was used to simulate several phenomena for the first time with phase-field techniques. Composition and phase gradients were included in the free energy functional. Chemical potential gradients calculated from the free energy functional served as the basis for the derivation of biharmonic nonlinear diffusion equations. A linear interpolation function was used to handle phase fractions and avoid problematic pair-wise treatment of phase boundaries. A cutoff barrier function was introduced as a less restrictive generalization of the pair-wise double well function. Parameters in the model were related to physical quantities such as free energy, surface energies between phases, and diffusivity. Four and five phase ternary free energy landscapes were used to simulate eutectic reactions, nucleation and growth, premelting, intergranular films, the appearance of transient and reactive phases, transient liquid bonding, and kaleidoscopic microstructure growth. Slow diffusion in the solid phases coupled with fast diffusion in the liquid phase was found to have a dramatic effect on microstructure evolution. Several numerical challenges were overcome and an adaptive implicit-explicit numerical scheme was developed for simulating coarsening at long times.
3:15 PM - LL2.3
Atomistic Study of Phase Transitions in Single and Nanocrstalline Fe-Cu: Structure, Shape and Plasticity.
Paul Erhart 1 , Babak Sadigh 1 , Alexander Stukowski 2 , Jaime Marian 1 , Alfredo Caro 1
1 Chemistry, Materials, Earth and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, California, United States, 2 Institute for Materials Science, Technische Universitaet Darmstadt, Darmstadt Germany
Show AbstractWe have recently developed a hybrid Monte-Carlo/molecular dynamics (MD/MC) algorithm which is based on the so-called variance-constrained semigrandcanonical ensemble. For the first time, this tool enables us to study the phase transitions of Cu precipitates in ferritic Fe-Cu alloys without constraints on structure, shape or chemical composition. In single crystalline Fe-Cu at 700 K Cu precipitates containing less than ~18,000 atoms are found to adopt the bcc crystal structure and to form spherical clusters. For precipitates with more than ~28,000 atoms a multiply-twinned 9R structure is observed and the clusters assume an elongated "cigar-like" shape with a size-dependent aspect ratio between 1.7 and 2.3. The transition is associated with a hysteresis which extends approximately from 18,000 to 28,000 atoms. The critical size at which the structural transition is observed to be strongly temperature dependent with larger bcc clusters observed at higher temperatures. We show this behavior to be caused by a dynamical stabilization of bcc-Cu. Using the equilibrium structures obtained from our hybrid MD/MC algorithm, we have simulated the interaction of screw dislocations with Cu precipitates as a function of size and structure. Finally, we have also applied our methodology to study the precipitation in nano-crystalline Fe-Cu alloys. Copper shows a strong tendency to seggregate to grain boundaries and grain boundary junctions. The resulting precipitates are observed to undergo a size-dependent phase transition which --in stark contrast to bulk Fe-Cu -- does not proceed via a multiply-twinned 9R-structure but leads directly to fcc-Cu precipitates.
3:30 PM - LL2.4
An Atomic Scale Investigation of the Heterogeneous Nucleation of Solid Phase Aluminum from its Molten State.
Junsheng Wang 1 , Andrew Horsfield 1 , Stefano Angioletti-Uberti 1 , Peter Lee 1
1 Department of Materials, Imperial College London, London, London, United Kingdom
Show AbstractTo improve aluminum alloy properties we need to understand the mechanisms by which liquid Al solidifies. A key step is the heterogeneous nucleation of the solid at the surface of small particles. TiB2 has been found experimentally to be very effective as a heterogeneous nucleus for solid Al. However, there is still controversy about the precise phase evolution sequence from TiB2 to solid Al at an atomistic rearrangement scale. Experimentally, many people have found that TiB2 can become an efficient nucleus only when a small amount of Ti is added. This has been explained by the formation of an intermediate Al3Ti at the surface of TiB2, followed by the subsequent production of the primary Al phase. In this study, Molecular Dynamics simulations are used to investigate the nucleation of solid Al at the surface of Al3Ti. We will discuss the complexities of constructing a robust potential on the basis of results from Density Functional Theory, and will show that the predicted sequence of events is consistent with the known phase diagram for Ti and Al.
4:15 PM - **LL2.5
Phase-field Approach to Integrated Phase and Grain Microstructure Evolution.
Saswata Bhattacharya 1 , Taewook Heo 1 , Long-Qing Chen 1
1 , Penn State University, University Park, Pennsylvania, United States
Show AbstractMost materials in engineering applications are polycrystalline, containing grains of different crystallographic orientations separated by grain boundaries. To predict the kinetics of phase transformations such as precipitation reactions and the accompanying microstructure evolution in the presence of grain boundaries is significantly more challenging than those in a uniform single crystal. Precipitation reactions in a polycrystalline material involve the complicated coupling among a number of different processes: solute segregation or depletion near grain boundaries, grain boundary migration, precipitate nucleation, growth and coarsening. Furthermore, the elastic moduli for a polycrystal are always spatially inhomogeneous: each grain has different elastic modulus and the elastic constants in the grain boundary regions are generally different from those inside the grains. In this presentation, a phase-field model will be presented for modeling solute segregation and precipitation of second-phase particles in a polycrystal in the presence of the elastic strain with inhomogeneous modulus. Examples to be discussed include segregation behavior at grain boundaries in the presence of coherent precipitates inside grains, the morphological evolution during isostructural phase separation in the presence of grain boundaries, and precipitation of tetragonal particles in polycrystalline cubic materials.
4:45 PM - LL2.6
Premelting of Body-Centered-Cubic Bicrystals.
Ari Adland 1 2 , Alain Karma 1 2 , Mark Asta 3 , David Olmstead 1 2 , Dorel Buta 3
1 Physics, Northeastern University, Boston, Massachusetts, United States, 2 Center for Inter-Disciplinary Research on Complex Systems, Northeastern University, Boston, Massachusetts, United States, 3 Department of Chemical Engineering and Materials Science, University of California, Davis, Davis, California, United States
Show AbstractThe presence of liquid films at grain boundaries below the bulk melting point can alter macroscopic properties of polycrystalline solids and dramatically reduce their resistance to shear stresses. The phase-field crystal (PFC) model is used to investigate the premelting behavior ofsymmetric tilt boundaries in body-centered-cubic (bcc) bicrystals as a function of misorientation and liquid correlation length that determines the solid-liquid interface width.A continuous premelting transition characterized by a diverging liquid film thickness at the bulk melting point is only found for interface widths larger than some threshold. Above this threshold, the range of misorientation for which this continuous transition occurs increases with interface width. The results are compared to molecular dynamics (MD) simulations for parameters of Fe where both PFC and MD simulations predict continuous premelting transitions over finite ranges of misorientation.The comparison sheds light on the role of capillary fluctuations in the determination of short-range-forces between crystal-melt interfaces from different grains.
5:00 PM - LL2.7
Modeling of Magnetic Thin Film with Misfit Dislocations.
Nirand Pisutha-Arnond 1 , Bo Yang 2 , Mark Asta 2 , Katsuyo Thornton 1
1 Department of Materials Science and Engineering, University Of Michigan, Ann Arbor, Michigan, United States, 2 Department of Chemical Engineering and Materials Science, University of California, Davis, Davis, California, United States
Show AbstractWe present a misfit dislocation model based on the Peierls-Nabarro formulation to study dislocation structures within heteroepitaxial Fe films grown on Mo(110) and W(110) substrates. The continuum model calculates the elastic field originating from the misfit dislocation array within a film of finite thickness. We use the stacking fault energy of the Fe/Mo and Fe/W systems from ab initio calculations as an input to the model. The interfacial energy stemming from the misfit dislocation is calculated and compared with the experimental estimation. By allowing the dislocation spacing to vary and by including the effect of homogeneous strain, the equilibrium dislocation spacing as a function of film thickness is obtained. We relate these results to the surface instability mechanism as well as the occurrence of the metastable height observed in Fe/Mo and Fe/W systems.
5:15 PM - LL2.8
Magnetic Grain Boundaries in Nickel.
Jan Kuriplach 1 , Oksana Melikhova 1 , Mojmir Sob 2 3
1 Low Temperature Physics, Charles University, Prague Czechia, 2 Department of Chemistry, Masaryk University, Brno Czechia, 3 , Institute of Physics of Materials, Brno Czechia
Show AbstractExploring connections between macroscopic characteristics of materials and their microscopic structure in atomistic dimensions is certainly an important topic in contemporary solid-state physics and materials science. A better understanding of the relations between macroscopic properties of solids and their structure provides new knowledge needed for development of materials with better technological properties. Internal interfaces, such as grain boundaries (GBs), are important elements of microstructure in polycrystalline solids, which have been widely used as engineering materials. Till now, however, fundamental interactions that determine the structure, stability and other important properties of GBs have not been fully understood, especially in magnetic materials. Theoretical calculations are, in this respect, very helpful as they can provide microscopic information that is often hardly accessible experimentally. In the present work, we employ an ab initio pseudopotential technique and concentrate on selected coherent GBs, such as Σ=5 (210) and Σ=19 (331), in nickel. We investigate first their stability and magnetism of atoms in the vicinity of GBs. The GB stability appears not to be significantly influenced by magnetism. On the other hand, we can observe an enhancement of atomic magnetic moments in the vicinity of grain boundaries. Furthermore, we examine an interaction of studied GBs with point defects. Namely, these are vacancies and sulphur and antimony impurities which tend to segregate at GBs in Ni. Both the grain boundary energy and segregation enthalpy are affected by the presence of a vacancy. In some cases, the so called vacancy delocalization at GBs, which is demonstrated as a substantial reduction of the vacancy free volume, is observed. Finally, the magnetic properties of GBs are also influenced by the interaction with point defects. These findings indicate that GB imperfections need to be seriously taken into account when calculating GB properties.
5:30 PM - LL2.9
Direct Approach to Atomistic ab initio Studies of Precipitate Growth in Alloys.
Flemming Ehlers 1 , Randi Holmestad 1 , Sigmund Andersen 2 , Calin Marioara 2
1 Department of Physics, Norwegian University of Science and Technology, NTNU, Trondheim Norway, 2 , SINTEF Materials and Chemistry, Trondheim Norway
Show AbstractA dramatic gain in the knowledge of precipitate formation, composition, and evolution in alloys has been achieved in the recent years with improvement of transmission electron microscopy techniques for direct structural imaging [1]. A detailed understanding of the microstructure is often essential for control and manipulation of materials properties: an important example for metals is the significant hardening of Al alloys by particular precipitates from a sequence strongly dependent on alloying element concentration and the treatment of the material [2].The wealth of experimental information provides a playground for theory in the context of elucidating precipitate growth mechanisms and influence on the host material. A head-on approach to atomistic modelling of these phenomena using an ab initio based scheme is conventionally deemed highly desired but impractical. The basic argument is that the system of any reasonably sized (i.e. realistic) and well isolated microstructure will simply contain too many atoms.We will challenge this conventional view: it is argued that most of the atoms of the above mentioned system do not play an active role in the growth discussion, hence need not be included in the modelling. Subsequently, a model system is presented which offers a highly accurate description of the interface between the host lattice and a microstructure of an arbitrary size. When used in conjunction with other approaches already available, this model system offers a direct approach to atomistic ab initio studies of microstructure growth.A general introduction to the modelling scheme will be presented, with the particular application being the main hardening precipitate beta'' in the Al-Mg-Si alloy.[1] K. W. Urban, Nature Mater. 8, 260 (2009).[2] C. D. Marioara, S. J. Andersen, H. W. Zandbergen, and R. Holmestad, Metal. Mater. Trans. A 36A, 691 (2005).
5:45 PM - LL2.10
Modeling Plastic Interactions in HCP Crystals.
Alejandro Diaz Ortiz 1 , Ruslan Kurta 1 , Volodymyr Bugaev 1 , Helmut Dosch 2
1 , Max Planck Institute for Metals Research, Stuttgart Germany, 2 , DESY, Hamburg Germany
Show AbstractA promising avenue for the intelligent design of materials involves the construction of maps relating structural information with physical properties. The mapping of short ranged interactions have been already accomplished by a variety of schemes, but the long-range interactions arising from the atomic-size mismatch have resisted a proper description in complex systems. This is an unfortunate gap since the description elastic interactions are the sine qua non to understand important materials processes, such as hardening by precipitate formation. Here we introduce a method to map long-range strain-induced interactions that enables large-scale simulations in materials with complex crystalline structures. We demonstrate our approach by calculating the long-wave strain energy of hcp-based Ti alloys for a variety of impurities.
Symposium Organizers
Mark Asta University of California
Alex Umantsev Fayetteville State University
Joerg Neugebauer Max-Planck-Institut fuer Eisenforschung
LL3: Defects and Radiation Damage in Steels and Nuclear Materials
Session Chairs
Joerg Neugebauer
Francois Willaim
Tuesday AM, December 01, 2009
Room 313 (Hynes)
9:30 AM - **LL3.1
Straight and Kinked Dislocations in Fe from First Principles.
Lisa Ventelon 1 , Emmanuel Clouet 1 , Francois Willaime 1
1 SRMP, CEA, Gif-sur-Yvette France
Show AbstractThe long range elastic interactions around dislocations make their investigation from first principles rather challenging. Focussing on screw dislocations in bcc iron, we have revisited and compared the two types of cell geometries used for such simulations: the periodic or dipole approach, and the cluster approach including with flexible boundary conditions. The non-degenerate core structure obtained in ab initio calculations has been analyzed in details, revealing a dilatation effect. Taking it into account in an anisotropic elasticity model, allows explaining the cell-size dependence of the energetics obtained within the dipole approach [1]. The Peierls potential obtained in ab initio suggests that the metastable core configuration at halfway position in the Peierls barrier, predicted by empirical potential, does not exist. We show how to construct tri-periodic cells optimized to study kinked dislocations [2]. 1. E. Clouet, L. Ventelon and F. Willaime, Phys. Rev. Lett. 102(2009) 0555022. L. Ventelon, F. Willaime and P. Leyronnas, J. Nucl. Mat. 386-388 (2009) 26.
10:00 AM - LL3.2
Solution, Mobility and Clustering with Vacancies of Al in Fe.
Hakim Amara 1 , Chu Chun Fu 2 , Frederic Soisson 2 , Philippe Maugis 3 4
1 , ONERA-CNRS, Chatillon France, 2 , CEA, Saclay France, 3 , CIRIMAT, Toulouse France, 4 , Arcelor Research, Metz France
Show AbstractThe development of low density steel is a very promising alternative in order to meet the industrial demand for high-performance material. In particular, Fe-Al based alloys show interesting properties for a new class of high-Al steels owing to high-temperatures corrosion resistance, mechanical strength, and relative low density. However, their mechanical properties are related to point defects and their concentration. It is well known that upon rapid quenching from elevated temperatures, iron aluminides retain a high concentration of thermal vacancies (V), which frozen, increase their yield strength and hardness at room temperatureThe aim of the present work is to study the properties of vacancies and defects in a FeAl system by mainly focusing on very dilute system containing fewAl atoms. Up to now, there are few studies related to the role of structural defects on the electronic structures, magnetic properties and atomic bonding in very dilute system, which should also help to give a better understanding of their behavior and mechanical properties. In particular, as these materials are sensitive to quench-in vacancies, the combination of small radius Al atoms with vacancies has not been fully addressed and is of a particular interest. Thus, it is expected that the presence of AlV clusters can be formed and then affect the properties of the material. The attempt of the present work is to investigate the stability and mobility of AlV clusters, at atomic scale, by combining ab initio calculations and Monte Carlo simulations.
10:15 AM - LL3.3
An Atomistic Study of Martensitic Phases in Dilute Fe-based Solid Solutions.
Alexander Udyansky 1 , Johann von Pezold 1 , Vladimir Bugaev 2 , Martin Friak 1 , Joerg Neugebauer 1
1 Computational Materials Design, Max-Planck-Institut für Eisenforschung GmbH, Düsseldorf Germany, 2 Low-dimensional and Metastable Materials, Max-Planck-Institut für Metallforschung, Stuttgart Germany
Show AbstractThe tetragonal states of interstitial Fe-based solid solutions are known as martensite. Martensitic transformations are not only underlying steel strengthening but are also the origin of unusual mechanical properties such as reversible strain, superelasticity and superplasticity as observed in a range of technologically important materials, including shape memory alloys and transformation-toughened ceramics. Martensitic transformations can be triggered/controlled e.g. by adding low concentrations of interstitial impurities such as carbon, nitrogen or oxygen. The stability of such low concentration tetragonal phases depends on both short-range chemical and long-range elastic interactions. The long-range limit is difficult to capture by standard atomistic approaches. We therefore employ the reciprocal space Krivoglaz-Kanzaki [1] force concept to calculate all relevant elastic interactions between interstitials in dilute Fe-based solid solutions. The short-range chemical interactions, as well as the parameters entering the analytical approach for the description of the elastic interaction are obtained atomistically using density functional theory (DFT) in the generalized gradient approximation (GGA). Applying this approach to technologically important Fe-based solid solutions allowed us to construct the temperature/interstitial concentration phase diagrams. An analysis of these diagrams showed that both long-range elastic and short-range chemical interactions need to be taken into account. Based on the computed phase diagrams we get a direct insight into the stability and formation of martensite: specifically tetragonal states are predicted to be preferred also at low C concentrations due to a thermodynamically driven [2] orientational ordering of carbon interstitials [3]. Further, the critical concentration for the cubic-tetragonal transition at room temperature is found in excellent agreement with recent experimental data. The developed multi-scale methodology allows to study long-range elastic defect-defect interactions even with rather modest supercell sizes making it an ideal tool in combination with modern DFT approaches.[1] M. A. Krivoglaz, X-Ray and Neutron Diffraction in Nonideal Crystals (Berlin: Springer, 1996) H. Kanzaki, J. Phys. Chem. Solids 2 24 (1957)[2] G. V. Kurdjumov and A. G. Khachaturyan, Metall. Trans. 3, 1069 1972[3] A. Udyansky, J. von Pezold, V. N. Bugaev, M. Friák and J. Neugebauer PRB 79, 224112 (2009)
10:30 AM - LL3.4
Multiscale Modeling of the Resistivity Recovery Experiments in Alpha-Iron using Event-based Kinetic Monte-Carlo.
Thomas Luypaert 1 , Donev Aleksandar 2 , Vasily Bulatov 2 , Mihai-Cosmin Marinica 3 , Athenes Manuel 3
1 Mechanical Engineering, Massachussets Institute of Technology, Cambridge, Massachusetts, United States, 2 , Lawrence Livermore National Laboratory, Livermore, California, United States, 3 DEN/DMN Service de Recherches de Métallurgie Physique, CEA Saclay, Gif s/ Yvette France
Show AbstractPredicting the microstructural evolution of the radiation damage in materials requires handling the physics of infrequent-event systems where several time-scales are involved. The event-based Kinetic Monte-Carlo (KMC) method is essential for simulating the kinetics of materials under irradiation since it allows carrying out simulations over large time scales and high irradiation doses. Recently, a new event Monte-Carlo algorithm was proposed [1] which goes beyond the binary collision approximation used in the past (e.g. JERK method [2,3]). Using the theory of first-passage processes and time dependent Green's functions, the diffusion N-body problem is split into independent single- and two-body propagations circumventing numerous diffusion hops used in standard Monte Carlo simulations. This new method, namely, the First Passage KMC (FPKMC) algorithm, solves exactly the diffusion problems and is extremely efficient (order N). These two approaches were applied in the case of the isochronal resistivity recovery experiment in alpha-iron irradiated by electrons. Some aspects of this experiment were successfully reproduced by a multiscale modelling approach based on the image at 0 K of the energy landscape of the system [2]. The stability and mobility of small self-defect clusters determined by ab initio methods were the input data for an event based Kinetic Monte Carlo model used to explore the defect population evolution during the annealing and to extract the resistivity recovery peaks. However, some high doses and/or high temperature peaks are not very well reproduced. Using FPKMC we go beyond this vision including the finite temperature effects. To explore the energetic landscape we use an eigenvector following method for systematic search of saddle points and transition pathways on a given potential energy surface (recently improved version [4] of activation relaxation technique nouveau [5]). The effect of finite temperature is taken into account using the lattice dynamics [6].1.T. Oppelstrup et al. Phys. Rev. Lett., 97, 230602, (2006); T. Oppelstrup et al, arXiv:0905.3575 (2009) A. Donev et al, arXiv:0905.3576, (2009)2.A. Barbu et al, Phil. Mag. 85, 541 (2005)3.C.C. Fu et al. Nature Mat. 4, 68, (2005)4.E. Cances et al, J. Chem. Phys., 130, 114711 (2009)5.G.T. Bakerma et al, Phys. Rev. Lett., 77, 4358, (1996).6.Terentyev D.A. et al, 100, 145503 (2009)
11:15 AM - LL3.5
Quantum and Thermal Effects in Hydrogen Diffusion in BCC Iron: A Path-Integral Molecular Dynamics Study.
Hajime Kimizuka 1 , Hideki Mori 1 , Hiroki Ushida 1 , Shigenobu Ogata 1
1 Department of Mechanical Science and Bioengineering, Graduate School of Engineering Science, Osaka University, Toyonaka, Osaka, Japan
Show AbstractWe analyze the diffusion behavior of interstitial hydrogen in bcc iron theoretically using path-integral centroid molecular dynamics (CMD) method, which can describe the real-time evolution of particles based on quantum statistical mechanics. In the present approach, the embedded-atom-method (EAM) potential model for the iron-hydrogen interaction is newly developed to reproduce the ab initio minimum energy path of hydrogen migration based on the density-functional-theory (DFT) data in the literature. This potential model allows us to describe the accurate “bare” potential surface, and the effective centroid potential surface can be obtained in quantum regime by incorporating the path-integral average.Time evolutions of mean-square displacements of hydrogen atoms in the bulk iron are calculated at temperatures of 100-1000 K, and then diffusion coefficients and activation energies of hydrogen migration are evaluated. The obtained results are in excellent agreement with experimental measurements over a wide temperature range. It clearly indicates that our approach is valid and effective for describing the nonlinear temperature dependence of the hydrogen diffusion behavior in the metal. In order to characterize the quantum effects on the hydrogen diffusion process, the CMD results are compared with those obtained from classical molecular dynamics (MD) method. By taking into account of quantum effects, the activation energy is significantly reduced and diffusion process is accelerated even at ambient temperatures. At low temperatures (below 500 K), quantum effects are dramatically enhanced as a temperature decreases, and thus the CMD values of activation energies become quite lower than the classical MD values. This leads to much higher hydrogen diffusivity in the quantum system than the classical system. On the other hand, we find that the quantum effects can be almost ignored at high temperatures over 500 K in the present case. This result suggests that hydrogen diffusion is approximately classical in this temperature region.These facts indicate that the quantum effects can play a significant role in hydrogen diffusivity over a wide temperature range in bcc iron. In this study, the hydrogen motions in the vicinity of a point defect and a screw-dislocation core are also investigated to evaluate the hydrogen-trapping effects by using our approach. It is noteworthy that no clear anisotropy of hydrogen diffusion is observed along the dislocation lines in bcc iron.
11:30 AM - LL3.6
Aiding the Design of Radiation Resistant Materials with Multiphysics Simulations of Damage Processes.
C. Race 1 , D. Mason 1 , J. le Page 1 , M. Finnis 1 2 , W. Foulkes 1 , A. Sutton 1
1 Department of Physics, Imperial College London, London United Kingdom, 2 Department of Materials, Imperial College London, London United Kingdom
Show AbstractThe design of metals and alloys resistant to radiation damage involves the physics of electronic excitations and the creation of defects and microstructure. During irradiation damage of metals by high energy particles, energy is exchanged between ions and electrons. Such "non-adiabatic" processes violate the Born-Oppenheimer approximation, on which all classical interatomic potentials rest. By treating the electrons of a metal explicitly and quantum mechanically we are able to explore the influence of electronic excitations on the ionic motion during irradiation damage. Simple theories suggest that moving ions should feel a damping force proportional to their velocity and directly opposed to it. In contrast, our simulations of a forced oscillating ion have revealed the full complexity of this force: in reality it is anisotropic and dependent on the ion velocity and local atomic environment. A large set of collision cascade simulations has allowed us to explore the form of the damping force further. We have a means of testing various schemes in the literature for incorporating such a force within molecular dynamics (MD) against our semi-classical evolution with explicitly modelled electrons. We find that a model in which the damping force is dependent upon the local electron density is superior to a simple fixed damping model. We also find that applying a lower kinetic energy cut-off for the damping force results in a worse model. Such cut-offs are frequently applied, but have only a poor physical justification. A detailed examination of the nature of the forces reveals that there is much scope for further improving the electronic force models within MD.
11:45 AM - LL3.7
The Relevance of fcc/bcc Interface Structure to Interface Properties: Investigation from Atomistic Modeling.
Xiang-Yang Liu 1 , R. Hoagland 1 , J. Wang 1 , M. Demkowicz 2 , B. Uberuaga 1 , A. Voter 1 , T. Germann 1 , M. Nastasi 1 , A. Misra 1
1 , Los Alamos National Laboratory, Los Alamos, New Mexico, United States, 2 Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
Show AbstractNanolayered Cu-Nb composites exhibit high strength and enhanced radiation damage tolerance. To understand the relevance of interface structure to interface properties in fcc-bcc systems, tunable potentials offer a fairly simple way to selectively vary parameters independently. In this work, the parameterization of the EAM interatomic potentials in fcc/bcc systems is modified to understand the interface properties. We first change the dilute heat of mixing between Cu and Nb and investigate the effects on interface structure, defect formation energies and shear resistance. To understand the interface behavior in different lattice geometries, the relative lattice constants between Cu and the bcc crystal phase were varied. The point defect energetic at these interfaces in the Kurdjumov-Sachs orientation relation is studied. Simulations of collision cascade at these interfaces using molecular dynamics (MD) and accelerated MD methods are performed to predict the radiation damage tolerance.
12:00 PM - LL3.8
Simulating Xe Redistribution in UO2±x with Heterogeneous Grain Boundary Micro-structures.
David Andersson 1 , Pankaj Nerikar 1 , Neil Carlson 1 , Blas Uberuaga 1 , Christopher Stanek 1
1 , Los Alamos National Laboratory, Los Alamos, New Mexico, United States
Show AbstractFrom an engineering perspective, the formation and redistribution of fission gases are critical determinants of nuclear fuel performance and in particular limit the extent of burnup. Most fission gases have low solubility in the UO2±x fuel matrix and as a result there is a significant driving force for segregation of gas atoms to heterogeneities such as grain boundaries and subsequently for nucleation of gas bubbles. These effects are most pronounced for large fission gas atoms, which specifically include Xe, which is the focus of this study. Segregation to grain boundaries is often assumed to be followed by more rapid release to the fuel plenum, either via fast diffusion of individual gas atoms or via transport of nucleated gas bubbles. This implies that the first controlling step for fission gas release is diffusion of individual gas atoms through the fuel matrix to existing bubbles or grain boundaries (sinks), a process that is governed by the activation energy for bulk diffusion of gas atoms, the driving force for segregation to existing sinks (bubbles or grain boundaries) and their saturation limit. In separate studies we used atomistic simulations, based on both empirical potentials and density functional theory, to determine Xe bulk diffusion as function of UO2±x stoichiometry, the sink strengths for segregation of Xe to different types of grain boundaries as well as the sink strength variation as function of the Xe loading. Additionally, these studies have established the spatial range of the Xe interaction field with grain boundaries. Altogether these findings suggest that in order to properly model Xe redistribution we need to account for, not only the position of grain boundaries, but also the distribution of various types of grain boundaries. In this study we have developed a thermodynamic description of Xe in micro-structurally heterogeneous UO2±x fuels, i.e. a model that accounts for the existence of various types of grain boundaries, each exhibiting unique properties, as well as the local concentration of Xe atoms. We demonstrate how such a model can be formulated and parameterized using results from the atomistic simulations described above. The thermodynamic model is then applied in conjunction with calculated Xe mobilities to derive a transport model that explicitly accounts for the Xe interaction field with grain boundaries. This model is then solved for a number of grain boundary distributions having different character and grain sizes by using a sharp interface model. We also present results for case studies that assume different release mechanisms and rates from the grain boundaries as well simplistic models for bubble nucleation. Finally we discuss generalizations of the sharp interface model to a phase field model capable of describing a wider range of phenomena.
12:15 PM - LL3.9
Computer Modeling of the Role of Symmetric and Asymmetric Tilt Grain Boundaries in Improving Radiation Tolerance.
Xian-Ming Bai 1 , Richard Hoagland 1 , Michael Nastasi 1 , Arthur Voter 1 , Blas Uberuaga 1
1 , Los Alamos National Laboratory, Los Alamos, New Mexico, United States
Show AbstractSymmetric and Asymmetric sigma 11 tilt grain boundaries (GBs) in copper are used as model systems to study the effects of detailed GB structure on radiation tolerance. Molecular dynamics is used to simulate the cascade-induced defect production phase of the radiation damage event. We have examined the damage produced as a function of both the original distance between the primary knock-on atom (PKA) and the GB as well as the energy of the PKA. Multiple simulations were performed to obtain good statistics of the number of defects. We have found that, compared to a single crystal, both GBs show the potential of enhancing radiation tolerance. The radiation resistance is sensitive to the initial PKA distance from a GB and there is an optimal PKA distance for defect absorption, implying that the optimal grain size depends on the PKA energy and thus the irradiation spectrum. Molecular statics was used to calculate the defect formation energies at the GBs relative to the bulk crystal. Finally, accelerated molecular dynamics was used to investigate the long-time annealing of defects produced near the GBs.
12:30 PM - LL3.10
Grain Boundary Structures and Influence on Segregation Properties in Uranium Dioxide.
Pankaj Nerikar 1 , Blas Uberuaga 1 , Chris Stanek 1 , David Andersson 1 , Susan Sinnott 2 , Simon Phillpot 2
1 MST-8 Structure/Property Relations, Los Alamos National Lab, Los Alamos, New Mexico, United States, 2 Materials Science and Engineering Department, University of Florida, Gainesville, Florida, United States
Show AbstractUranium dioxide (UO2) is the standard nuclear fuel in pressurized water reactors. Fission gases such as xenon (Xe) migrate to grain boundaries and cause swelling of the fuel. The structure of grain boundaries in UO2 and the propensity of Xe to segregate to boundaries of different structure is explored in this work using empirical potentials and density functional theory. The specific boundaries studied were symmetric Σ 5 tilt, Σ 5 twist, and an amorphous boundary. Surprisingly, we found the energy of segregation to be very sensitive to the local atomic environment of the solute atom in the host and that there is a substantial difference in the overall segregation propensity to the three boundaries selected. Possible implications of this study on Xe diffusion and in predicting macroscopic fuel behavior are discussed. This work was supported in part by the DOE-BES Computational Materials Science Network.
12:45 PM - LL3.11
Predicted Energies and Structures Associated With the Mixed Calcium Strontium Fluorapatites.
Robin Grimes 1 , Emily Michie 1 , Elly Jay 1
1 , Imperial College London, London United Kingdom
Show AbstractAtomic scale local density functional simulations and configurational averaging are used to predict the energies and lattice parameters associated with mixed calcium/strontium fluorapatites; CaxSr10-x(PO4)6F2. In particular, the partition of Sr2+ and Ca2+ ions between the 6h and 4f cation sites is established across the entire compositional range; 0 ≤ x ≤ 10 in steps of 1. Particularly around the mid-composition, large numbers of distinct configurations must be simulated. The resulting data is used to generate lattice parameters and lattice volumes, which are analyzed as a function of Ca2+ to Sr2+ concentration and particular cation site distributions. The predicted internal energy of mixing between the end members is used to discuss the available experimental data. The ab initio results are then compared with equivalent simulations carried out using interatomic potential parameters.At low Sr2+ ion concentrations there is only a slight energetic preference predicted for a Sr2+ ion to occupy a 6h site rather than a 4f site. Consequently the distribution of Sr2+ ions over the 6h and 4f sites approaches a random distribution. Since there are more 6h sites than 4f, this means that the majority of Sr2+ ions occupy 6h sites. As the Sr2+ ion concentration increases, there is a greater energetic preference for an individual Sr2+ ion to occupy a 6h rather than a 4f site and this translates into a greater overall preference for Sr2+ ions to be observed at 6h sites. The internal energies for solution calculated using the ab initio approach predict a strongly asymmetric curve across the compositional spectrum, which can be used to interpret experimental observations. The classical model is less successful; although it does reproduce the basic shape the detail is less satisfactory. Reasons for this will be discussed.We predict that a strong preference for Sr2+ ions to occupy 6h sites, will result in a nonlinear increase in the “a” lattice parameter but an opposite nonlinear increase in the “c” lattice parameter. Consequently there is, remarkably, an overall linear increase in volume upon Sr2+ substitution irrespective of the Sr2+ ion distribution. However, the predicted configurational average occupation values leads to a distribution of Sr2+ ions over 6h and 4f sites which is sufficiently close to random that an essentially linear change in lattice parameters is expected.
LL4: Recent Methodological Developments
Session Chairs
Tuesday PM, December 01, 2009
Room 313 (Hynes)
2:30 PM - **LL4.1
Bond-order Potentials for Bridging the Electronic to Atomistic Modelling Hierarchies in Materials Science.
Ralf Drautz 1
1 ICAMS, Ruhr-Universität Bochum, Bochum Germany
Show AbstractThe derivation of robust interatomic potentials is a key step for bridging from the electronic to the atomistic modelling hierarchies. In this talk I will present an analytic interatomic bond-order potential (BOP) that depends explicitly on the valence of the transition metal element [1]. This analytic potential predicts the structural trend from hcp to bcc to hcp to fcc that is observed across the non-magnetic 4d and 5d transition metal series. The potential also describes the different ferromagnetic moments of the alpha (bcc), gamma (fcc) and epsilon (hcp) phase of the 3d transition metal iron, the difference between the ferromagnetic and anti-ferromagnetic states as well as non-collinear spin-configurations. In addition, the potential includes a correct description of alloy bonding within its remit. I will show how the potential is derived from the tight-binding electronic structure as a systematic extension of the second-moment Finnis-Sinclair potential to include higher moments and will dicuss the application of the potential to modelling topologically close-packed phases in Ni-based superalloys. [1] R. Drautz and D.G. Pettifor, Phys. Rev. B 74, 174117 (2006).
3:00 PM - LL4.2
The Angular-Dependent Embedded Atom Method Potential for Atomistic Modeling of Metal-Covalent Systems.
Avinash Dongare 1 2 , Douglas Irving 1 , Leonid Zhigilei 3 , Arunachalam Rajendran 4 , Bruce LaMattina 5 , Mohammed Zikry 2 , Donald Brenner 1
1 Materials Science and Engineering, North Carolina State University, Raleigh, North Carolina, United States, 2 Department of Mechanical Engineering, North Carolina State University, Raleigh, North Carolina, United States, 3 Materials Science and Engineering, University of Virginia, Charlottesville, Virginia, United States, 4 Department of Mechanical Engineering, University of Mississippi, University, Mississippi, United States, 5 , U. S. Army Research Office, Research Triangle Park, North Carolina, United States
Show AbstractAtomic-scale modeling of many practically important multi-component systems requires interatomic potentials capable of providing an adequate description of interactions with mixed types of atomic bonding. We present a new computationally efficient Angular-dependent EAM (A-EAM) interatomic potential developed by combining the Embedded Atom Method (EAM) potential for metals with the Stillinger – Weber (SW) potential for Si/Ge in a compatible functional form. The cross metal-covalent interactions are fitted to reproduce the energies and structural characteristics of several representative bulk structures and small clusters as obtained from Density Functional Theory (DFT) calculations. The first applications of the A-EAM potential to investigate effects of intermixing and segregation in the Au-Si-Ge ternary system as well as the mechanical properties of metal-covalent (Al/Si) interfaces at the atomic scale using molecular dynamics and Monte Carlo simulation techniques will be presented. The combined potential proves to be computationally efficient and suitable for large-scale MD simulations of metal–Si/Ge systems, while retaining the properties of the pure components as predicted by the original SW and EAM potentials. The framework of the A-EAM potential also allows for an extension to the Tersoff potential opening possibilities for modeling of metal-carbon and metal-SiC systems.
3:15 PM - LL4.3
The Development of a Magnetic Potential for BCC Fe.
Samuele Chiesa 1 , Peter Derlet 2 , Sergei Dudarev 3 , Mark Gilbert 3 , Helena Van Swygenhoven 1
1 NUM/ASQ, PSI, Villigen Switzerland, 2 NUM/CMT, PSI, Villigen Switzerland, 3 EURATOM/UKAEA Fusion Association, Culham Science Centre, Abingdon, Oxfordshire, United Kingdom
Show AbstractThere has been considerable activity within the radiation damage community to develop reliable and transferrable empirical potentials for pure alpha-Fe for use in the atomistic study of defect creation, evolution and mobility. Such an effort has been driven by the availability of spin-resolved DFT calculations that expand the ab-initio materials database of pure alpha-Fe to include energy data of defects such as interstitial and vacancy clusters and dislocation structures. The fundamental role that ferromagnetism plays in alpha-Fe has been further confirmed by these calculations where, unique to all other BCC transition metals, the <110> dumbbell single interstitial defect has a significantly lower formation energy than the <111> dumbbell or a crowdion. The Magnetic Potential (MP) is the first attempt to explicitly include ferromagnetism within an EAM formalism [J. Phys.: Cond. Matt. 17, 7097 (2005)]. We applied a trial and error approach to optimize the MP within the available range of ab-initio data on defect and magnetic properties. To control anharmonicity, experimental third order elastic constants have been included in the fitting algorithm, and results from the recently developed Frenkel-Kontorova multi-string model were applied successfully to control the core structure of the <111> screw dislocation. Limitations of the MP formalism to extrapolate the ab initio database can be easily investigated within our method: the focus is on the mobility of the non-degenerate screw dislocation, self-interstitial properties, as well as equilibrium properties the FCC and BCC phases. An optimized short range version of the MP has been selected. When considering a multiscale model, a natural question arises: how sensitive are thermodynamical quantities to the static properties they where fitted to? We try to answer this question by comparing full dynamical, quasi-static and static calculations of the vibrational free energy of the <110> self-interstitial dumbbell defect in BCC Fe for a range of modern empirical potentials including the MP and the optimized version. It is found that, depending on the empirical potential, the harmonic approximation for the vibrational free energy is justified especially for empirical potentials that have been fitted to third order elastic constants. The unique applicability range of such calculations for BCC Fe is also discussed given that with rising temperature spin fluctuations become increasingly important. The work was partly funded by the UK EPSRC and EURATOM.
4:00 PM - LL4.4
Enhancing Molecular Dynamics to Capture Electronic Effects.
N. Modine 1 , R. Jones 2 , D. Olmsted 1 , J. Templeton 2 , G. Wagner 2 , R. Hatcher 3 , M. Beck 4
1 , Sandia National Laboratories, Albuquerque, New Mexico, United States, 2 , Sandia National Laboratories, Livermore, California, United States, 3 , Lockheed Martin Advanced Technology Laboratories, Cherry Hill, New Jersey, United States, 4 , Vanderbilt University, Nashville, Tennessee, United States
Show AbstractIn modeling non-equilibrium thermal transport in solids, classical molecular dynamics (MD) has the primary strength of explicitly representing phonon modes and the defects that scatter phonons. On the other hand, electrons and their role in energy transport are missing. In nanoscale and nanostructured systems, the behavior of the system is complicated further by phonon-confinement, ballistic transport, and discrete defect scattering effects. These effects are absent in phenomenological models of heat transport, but naturally captured by MD. Our goal is to couple a MD treatment of the ionic subsystem with a partial differential equation (PDE)-based model of the electronic subsystem in order to accurately capture the aggregate behavior of nanoscale systems. Along these lines, we have enhanced the LAMMPS MD package by coupling the ionic motions to a finite element (FE) based representation of electronic charge and heat transport. The coupling between the subsystems occurs via a local version of the two-temperature model that allows the ionic and electronic subsystems to exchange energy and eventually come into equilibrium. Key parameters describing the coupling between the electronic and ionic subsystems are calculated using Time Dependent Density Functional Theory (TDDFT). These TDDFT calculations can be either explicit (i.e., energy is actually transferred between the electrons and ions during the simulation) or implicit (i.e., the rate of energy transfer is determined via the fluctuation dissipation theorem). Initial demonstrations of our approach and capabilities have focused on heat transport in nanowires and carbon nanotubes, and these results will be discussed.Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under contract DE- AC04-94AL85000.
4:15 PM - LL4.5
Force Matching Via Local Geometry Mapping.
Ljubomir Miljacic 1 , Donald Ellis 2 , Axel van de Walle 1
1 Material Science, California Institute of Technology, Pasadena, California, United States, 2 Physics and Astronomy, Northwestern University, Evanston, Illinois, United States
Show AbstractImproving a force-field quality by matching the forces to a more accurate theoretical model on a finite set of system configurations is a long standing problem in the physics of materials. We address this problem via mapping a complex many-body situation to a much reduced description of important local geometries. Assuming that these can be described in a structured way, mainly related to how electrons are expected to respond to their continuously changing many-body environment, the matching of the forces on the involved atoms is expressed in terms of a small number of additional, problem specific degrees of freedom, the “supercoordinates” of the force field parameters. When the less accurate model is a classical Molecular Dynamics (MD) model, the supercoordinates can be based only on local atomic positions. When it is an ab-initio one capable of producing relevant MD trajectories, the supercoordinates can be also based on local electronic structure, as provided by the running quantum engine; here, the force matching introduces a corrective classical force-field UC, added to maximally reproduce a superior quality ab-initio forces. We applied this strategy in calculating Equation of State of Ta and Fe, needed to understand and model the high-energy-density dynamic response in metals, as it arises in hypervelocity impact experiments. The significant presence of electron correlation in these transition metals raises a problem of accuracy of the underlying ab-initio MD model, for which we used Density Functional Theory (DFT) with corrective pair-wise UC terms. Combining DFT using the B3LYP hybrid functional with QMC calculations provided the higher accuracy model. To reduce the ab-initio demand in the wide P-T ranges, we fit available gas-liquid data to the Peng-Robinson model [1]. We also tested this strategy on a system of a water molecule broadly interacting with hematite surface and a 66% reduction in the force mismatch, between a simple atomistic AMBER force field and a DFT-based model, was achieved [2]. [1] Ind. Eng. Chem., Fundam 15, 59 (1976)[2] L. Miljacic, D.E. Ellis, “Force matching via local geometry mapping method”, in preparation.
4:30 PM - LL4.6
Self-learning Synchronous Parallel Kinetic Monte Carlo for Discrete Systems.
Enrique Martinez 1 , Jaime Marian 2 , Paul Monasterio 3 , Malvin Kalos 2
1 , IMDEA Materiales, Madrid Spain, 2 , Lawrence Livermore National Laboratory, Livermore, California, United States, 3 , Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
Show AbstractA new kinetic Monte Carlo algorithm for discrete systems is presented. Its development is based on the continuum diffusion-reaction algorithm presented in ref. [1], where it was demonstrated that perfect synchronicity is achievable by introducing a set of null events that simplify the implementation considerably. The phase transition kinetics of the Ising model have been studied and the critical exponents calculated for a system up to 1e9 spins. We have avoided boundary conflicts by using a sublattice separation method such that events take place at the same time in not-neighboring subdomains. We analyze the bias as a function of the different parameters that control it and show that it can be kept basically negligible. Concerning scalability measures, the introduction of the null events makes possible the implementation of a self-learning algorithm that adjust the set of null events in time avoiding global communications and increasing the efficiency of the algorithm.[1] J. Comp. Phys. 227 (2008) 3804-3823
4:45 PM - LL4.7
A Concurrent Multiscale Method for Coupling Atomistic and Continuum Models at Finite Temperatures.
Catalin Picu 1 2 , Nithin Mathew 1 2 , Max Bloomfield 2 , Mark Shephard 2
1 , Rensselaer Polytechnic Institute, Troy, New York, United States, 2 Scientific Computation Research Center, Rensselaer Polytechnic Institute, Troy, New York, United States
Show AbstractA concurrent multi-scale modeling method for finite temperature simulation of solids is introduced. The objective is to represent far from equilibrium phenomena using an atomistic model and near equilibrium phenomena using a continuum model, the domain being partitioned in discrete and continuum regions, respectively. An overlay sub-domain is defined between the two regions. The overlay’s role is to provide proper coupling of mechanical phenomena between the discrete and continuum descriptions and to partition the phonon spectrum, therefore performing the mapping into the continuum notions of thermal and mechanical energy. This is achieved by using the generalized Langevin equation of motion in the interface, which functions as a low pass filter. The problem of spurious reflections of high frequency components from the interface of the two models is avoided by proper decoupling of the modes and ensuring the continuity of displacements in the interface region. The underlying thermal problem is coupled by imposing a flux input into atomistic and using a temperature boundary condition for continuum. Proper flow of energy between the continuum and atomistic models is ensured by this bi-directional coupling.
5:00 PM - LL4.8
Identification of Parameters for Continuum Modeling of Phase Transformations in Real Materials.
Alex Umantsev 1
1 Natural Sciences, Fayetteville State University, Fayetteville, North Carolina, United States
Show AbstractAs the materials science modeling community is moving into a new era of quantitative modeling and design of real materials (e.g. multicomponent alloys), it is important to assess the challenges that we will be facing. One of those is obtaining reliable material parameters for the model. The problem is that now we are dealing with parameters which cannot be easily identified in experiments and obtained through direct measurements. For example, continuum modeling methodology uses the free energy of the system expanded in powers of the order parameter and its gradients. But the coefficients of expansion are not measurable parameters. The first attempts to find these coefficients consisted in guessing their (T, P)-dependence and comparing the theoretical values of the specific heat and compressibility with the experimental values. This strategy did not work always. Then researchers tried another strategy: extracting the bulk free-energy and surface quantities (e.g. tension) from the model and comparing them with the experimental or computable counterparts. This strategy meets certain challenges which, together with other approaches, will be discussed in the presentation.
5:15 PM - LL4.9
A Linear Scaling Self-consistent Charge Transfer Tight-binding Code for Molecular Dynamics Simulations of Organic Molecular Materials.
Marc Cawkwell 1 , Edward Sanville 1 , Nicolas Bock 1 , Matt Challacombe 1 , Anders Niklasson 1 , Thomas Sewell 2 , Dana Dattelbaum 3 , Stephen Sheffield 3
1 Theoretical Division, Los Alamos National Laboratory, Los Alamos, New Mexico, United States, 2 Department of Chemistry, University of Missouri-Columbia, Columbia, Missouri, United States, 3 Dynamic and Energetic Materials Division, Los Alamos National Laboratory, Los Alamos, New Mexico, United States
Show AbstractCohesion in organic molecular materials is complex since it involves covalent intramolecular bonding, ionic bonding between atomic partial charges, and van der Waals interactions. We have developed a computational method and supporting code that is based on the self-consistent charge transfer tight-binding approximation and which enables all of these interactions to be modeled simultaneously and rapidly with high accuracy and transferability. We employ either purification algorithms or the recursive expansion of the finite temperature Fermi operator to calculate the bond-order. Through utilizing the sparsity of the bond-order matrices typical of molecular systems, these methods enable us to construct a linear scaling scheme for the computation of the bond order. Furthermore, both methods are easily parallelized and we will demonstrate that they can be adapted in a straightforward manner to hybrid computational architectures including LANL’s Roadrunner supercomputer. Coulombic interactions between atomic partial charges are computed in parallel using Ewald or particle-mesh methods and van der Waals bonding is described empirically by the use of long range Buckingham potentials. Finally, by propagating quantities from one time step to the next in a time reversible way through the extended Born-Oppenheimer molecular dynamics formalism we ensure rigorous energy conservation as well as a decrease in computation time by over two orders of magnitude. In combination, all of these methods enable a rapid computation of forces that lends the code toward large-scale, long duration molecular dynamics simulations. We will demonstrate these capabilities by detailing an experimental and theoretical study of shock-induced chemistry in simple liquid hydrocarbons. In particular, we will show that our potential for hydrocarbons yields an equation of state that is in excellent agreement with available experimental data and that molecular dynamics simulations identify mechanisms through with organic molecular materials polymerize and/or decompose under shock compression.
5:30 PM - **LL4.10
First-principles Approaches to Materials for Alternative Energies.
Christopher Wolverton 1
1 Materials Science and Engineering, Northwestern University, Evanston, Illinois, United States
Show AbstractMany of the key technological problems associated with alternative energies may be traced back to the lack of suitable materials. The materials discovery process may be greatly aided by the use of computational methods, particular those atomistic methods based on density functional theory. In this talk, we present an overview of recent work on energy-related materials from a density-functional based approach. We briefly highlight applications in the areas of energy storage, thermoelectric materials, materials for solar thermochemical water splitting, and nuclear materials.
LL5: Poster Session: Multiphysics Modeling in Materials Design
Session Chairs
Wednesday AM, December 02, 2009
Exhibit Hall D (Hynes)
9:00 PM - LL5.1
Modelling the Nucleation of Crystals at Surfaces.
Richard Sear 1 , Amanda Page 1 , Piyapong Asanithi 1 , Alan Dalton 1
1 Department of Physics, University of Surrey, Guildford, Surrey, United Kingdom
Show AbstractCrystallisation is of great importance to a wide range of industries, from pharmaceuticals to bulk chemicals. It starts with nucleation, the formation and growth of a microscopic nucleus of the new crystalline phase. This is an activated process, it can take minutes or hours or even days, timescales far in excess of the ns timescales accessible with direct computer simulation. Here we will use the recently developed Forward Flux Sampling algorithm of Allen and coworkers to calculate nucleation rates of a simple model of methane at inert surfaces such as graphite. In experiment, nucleation is almost always heterogeneous, i.e., it occurs in contact with a surface, often that of an impurity. Our simulation results are consistent with this. We find that nucleation can occur at surfaces at supersaturations less than 10% of those at which homogeneous nucleation occurs. I will also present results showing how engineering the surface geometry can control where nucleation occurs, the defects that form in the growing crystal, the orientation of the crystal that forms, and potentially the polymorph that forms.
9:00 PM - LL5.10
General Photon Counting Model for Beam Splitters and Optoelectronic Devices.
Teppo Haeyrynen 1 , Jani Oksanen 1 , Jukka Tulkki 1
1 Department of Biomedical Engineering and Computational Science , Helsinki University of Technology, Espoo Finland
Show AbstractMeasurement of a photon from a quantum light field may significantly alter the photon statistics of the field. Depending on the initial field type the detection process may produce a field which has a greater, an equal or a smaller expectation value of the number of photons than the initial field. For example the number of photons in a thermal light field doubles during a single photon subtraction process. This counter intuitive but theoretically recognized behavior due to the measurement back action has only recently been verified experimentally by Parigi et al. [Science 317, 1890 (2007)]. In the quantum trajectory approach the detection process is modeled using two quantum operators: the one-count operator describes the photon absorption and the no-count operator describes the evolution of the field when no photons are absorbed. It is assumed that during a differential time interval [t, t+dt] only these two trajectories are possible. In the standard model [J. Mod. Optic. 28, 981, 1981] (SD model) the one-count operator is given by the bosonic annihilation and creation operators so that the detection rate is directly proportional to the number of photons. Starting from these two operators we derive new quantum operators [J. Phys. B 42, 145506, 2009] that describe the detection of exactly one and at least one photon during a non-differential time interval [t, t+T]. The former corresponds to measurement using a resolving detector (RD) (i.e. the detector detects each of the photons) and the latter corresponds to measurement using a non-resolving detector (NRD) (i.e. the detector cannot distinguish if it has absorbed one or more photons). The difference between our model and the SD model is that our count operators take into account also the trajectories that correspond to absorbing two or more photons. Therefore, our model is not limited to low intensity fields or weak field-detector coupling. We show that our model reproduces the experimental results of Parigi et al. and besides of being exact for the weak quantum fields our model is also exact at the classical limit of strong fields. Furthermore, our model reflects the results of the beam splitter setup and the damping at cavity fields. In addition of applying our model to cavity photon counting and beam splitters we also investigate the energy transfer between two semiconductor devices by using the quantum jump operators that correspond to emission, absorption and mirror losses, and transfer from one cavity to the other. The results are applied to model the heat transfer in a new class of heat pumps/coolers that are based on electroluminescent cooling of semiconductor and use photons as their working fluid.
9:00 PM - LL5.11
First-principles Approach to the Antiferrodistortive Phase Transition in SrTiO3.
Yi Wang 1 , ShunLi Shang 1 , Long-Qing Chen 1 , Zi-Kui Liu 1
1 Materials Science & Engineering, Penn State, University Park, Pennsylvania, United States
Show Abstract Using a system of 1.4142a by 1.4142a by 8a 80-atom supercell, the antiferrodistortive (AFD) phase transition in SrTiO3 are addressed with a first-principles formulation of the Helmholtz energy based on the partition function. With increasing temperature, we see a specific heat peak at a transition temperature ~100 K, due to the thermodynamic fluctuations among many crystal configurations resulted from the staggered oxygen cage rotations at finite temperature. In fact, the staggered oxygen cage rotations play a role to overcome the domain wall energy in the 001 direction. We therefore attribute the specific heat peak as a signal of the AFD transition in the concept of order-disorder transition.
9:00 PM - LL5.12
Spin Exchange Coupling and Nonconventional Superconductivity in BaFe2As2.
Yi Wang 1 , Shunli Shang 1 , Long-Qing Chen 1 , Zi-Kui Liu 1
1 Materials Science & Engineering, Penn State, University Park, Pennsylvania, United States
Show AbstractThe surprising discovery of the relatively high-temperature superconductivity in iron pnictides in 2008 has it more enigmatic of the origin of high-temperature superconductivity. There is no widely accepted theory on the subject after over 100,000 published papers since the finding of a lanthanum-barium-copper oxide ceramic in 1986. Other examples include heavy Fermion systems and ferromagetic superconductors whose superconductivity cannot be understood in the existing framework of BCS theory. Hereby we developed a first-principles approach to spin fluctuations and apply it to BaFe2As2. In contrast to existing theories, we find that it is the spin exchange coupling in the inter-plane c direction that dictates the spin density wave ordering. We quantitatively predicted the pressure dependence of the spin density wave ordering, the Schottky anomaly, and the temperature dependence of thermal populations of spin structures, all in agreement with available experimental data. The proposed approach is a step towards understanding non-conventional superconductivity.
9:00 PM - LL5.13
First-principles Modeling of Hydrogen Spillover on a Carbon-supported Palladium Cluster.
Samir Mushrif 1 , Alejandro Rey 1 , Gilles Peslherbe 2
1 Chemical Engineering, McGill University, Montreal, Quebec, Canada, 2 Chemistry and Biochemistry, Centre for Research in Molecular Modeling (CERMM) and Concordia University, Montreal, Quebec, Canada
Show AbstractHydrogen spillover is suggested to occur in hydrogen catalyzed reactions on carbon supported transition metal catalysts and is also believed to be a promising technique to increase the hydrogen storage capacity of lightweight, metal-doped, porous carbon materials. Recent experimental investigations have observed the spillover phenomenon directly for the first time and they have also demonstrated the presence of spillover hydrogen on the carbon support at room temperature. The spillover process consists of dissociative chemisorption of H2 on a transition metal, migration of atomic hydrogen on the metal cluster and then to the support and subsequently its diffusion from the surface to the bulk. Theoretical investigations of the spillover phenomenon in the past have studied these different steps individually and it is believed that each of these steps is associated with an activation energy barrier. In the present work, we perform first-principles calculations of the dynamics of the entire spillover process of hydrogen, at room temperature, on a tetrahedral palladium cluster supported on coronene, a model polycyclic aromatic hydrocarbon. First-principles calculations are performed using the planewave-pseudopotential implementation of the density functional theory and the Car-Parrinello scheme is used for molecular dynamics. The free energy surface, as a function of the reaction coordinates of the process, is also reconstructed using the metadynamics technique.
9:00 PM - LL5.14
New Insights into the Ignition Dynamics of Ni-Al Nanocomposites from Multi-Scale Modeling.
Christopher O'Brien 1 , Xiaoyin Ji 1 , Douglas Irving 1 , Donald Brenner 1
1 Materials Science and Engineering, North Carolina State University, Raleigh, North Carolina, United States
Show AbstractWe have studied the ignition of Ni-Al nanocomposites induced by Joule heating using a novel modeling method that couples numerical solutions to continuum current and heat flow equations to a molecular dynamics simulation. A variety of initial structures and applied voltages were used to explore the effect of geometry and heating rate on ignition mechanisms. Our simulations show a multi-step process that depends strongly on the Joule heating via the applied voltage, and that is largely independent of geometry. In the initial step the Al constituent is Joule melted, which is followed by a rapid intermixing at the Ni-Al interface that is accompanied by chemical energy release. Subsequent Ni-Al inter-mixing (and associated energy release) then becomes limited by the diffusion of Ni into the melt. Melting of the Ni constituent during this process either by Joule heating or by energy release from the inter-mixing (or a combination of both) is found to significantly increase the intermixing rate and hence enhance power output. Reactive nanocomposites such as the Ni-Al system explored here offer an opportunity for energy storage within a structural material, and in principle control over power output. Our simulations demonstrate that the latter can be achieved through the applied voltage, which can be used to control the temperature and hence the melting of the Ni and Al nano-constituents, and that differences in our geometries at the nanoscale play a relatively lesser role in controlling power output at the time and length scales of our simulation. This work has been supported by ARO grant W911NF-08-1-0423. Prof. C. Padgett is thanked for helpful discussions.
9:00 PM - LL5.15
First Principles Based Simulations of Composition Fluctuation and Residual Stress Distribution in InGaN/GaN MQW.
Kiichiro Mukose 1 , Daisuke Mori 1 , Masatoshi Sano 1
1 Faculty of Engineering, Tokyo University of Science, Tokyo Japan
Show AbstractIn the ternary alloy InGaN, the indium composition has been known to show spatial inhomogeneity in various growth conditions. And this composition fluctuation has been considered to be able to form the quantum dots structures in InGaN quantum wells those localize excitons and influence the spontaneous emission rate. In this study, by a new simulation method based on first principles, quantitative relations between the spatial variations of the indium composition and those of the residual stress in InGaN/GaN multiple quantum wells (MQW) were investigated theoretically. The simulation models of InGaN/GaN MQW contain triangular pillar-shaped cells, where the composition ratio, the strain and the stress in the each cell follow an equation of state which has been determined by ab-initio electronic structure calculations. The volume and shape of the each cell were optimized to minimize the free energy of the system, and the residual stress distribution was simulated. By the results of our simulations, quantitative relations between the composition fluctuation and the dislocation density in InGaN/GaN MQW has been studied. It is considered that quantum dots up to about 10 nm in thickness and 70 nm in diameter can be free from dislocation defects. It is also found that the difference of the type of the short-range chemical ordering of the indium and gallium atoms, which must depend on growth conditions, influence the residual stress distribution and the dislocation density considerably.
9:00 PM - LL5.16
Phase Field Modeling of Diffusion and Morphology Instabilities in Thermal Barrier Coating Systems.
Jie Deng 1 , Karim Ahmed 1 , Anter El-Azab 1
1 Scientific Computing, Florida State University, Tallahassee, Florida, United States
Show AbstractThe development of thermal barrier coating (TBC) systems with well controlled and predictable performance is crucial for making breakthroughs in many high-temperature application areas. Related research and development efforts, however, have been highly experimentally oriented and little modeling effort is being conducted in parallel to understand the complex nature of bonding and failure mechanisms in TBC systems. At the root of these failure mechanisms is the interlayer and surface diffusion driven by high temperature, stress gradients and chemical composition differences between different components of the TBC systems. In this presentation we will discuss the critical aspects of connection between diffusional processes, microstructural and morphological changes and failure initiation in TBC systems, and present phase field modeling results for the densification and porosity changes and inter-layer diffusion in typical TBC systems.
9:00 PM - LL5.17
Absorption of CO2 on Solid Materials, A Simulation Model.
Francisco Jaramillo 1 , Hector Dominguez 1
1 Rheology, Instituto de Investigaciones en Materiales, Mexico, Distrito Federal, Mexico
Show AbstractPower generation industries produce carbon dioxide (CO2) at hight temperature. The CO2 presence in the earth atmosphere has increased its heat trapping capability called the greenhouse effect. In order to have a cleaner environment the CO2 emissions must be reduced.Using the reactive canonical Monte Carlo (RCMC) simulation technique the behavior of a solid material under an atmosphere of CO2 was simulated and studied. Our model consists of a pore formed by two parallel fcc crystal plates separated by a distance h in z direction and between them a CO2 fluid. All the interactions between molecules were taken as Lennard – Jones 12-6 potential. Periodic boundary conditions in x and y directions were considered and the Lorentz – Berthelot mixing rules were used to obtain the crossed chemical species interaction parameters.To compare and validate the model we used ceramic material Lithium oxide (Li2O) parameters as a CO2 solid retainer. In this case it exists a chemical reaction between Li2O and CO2 to form Lithium carbonate (Li2CO3). Results show that by this chemical mechanism Li2O can be used for CO2 retention at hight temperatures (up to 1600 °C). The results tendencies reveal that CO2 absorption on Li2O increases as the temperature rises. When the CO2 – Li2O energy parameter is increased by a factor of 3 the absorption reduces in 13 % and when the energy parameter is reduced by a factor of 0.55 the absorption increases in 8 %. Local density profiles show that for certain temperatures there are physical association between the fluid and the solid surface.
9:00 PM - LL5.18
Flow Control by Smart Nanofluidic Channels: Continuum Modeling and its Comparison with MD/DPD Simulations.
Zhengmin Li 1
1 , N/A, Westborough, Massachusetts, United States
Show AbstractContinuum modeling combining the mean-field analytic theory and Brinkman equation is used to study the flow through cylinder and slit nano-channel grafted with polymer chains. The relations between brush layer thickness and permeability obtained via continuum modeling are compared with those via molecular dynamics and dissipative particle dynamics simulation respectively and some quantitative agreements are observed in addition to the qualitative agreements. The advantages and disadvantages of each method in the study on the flow through polymer-grafted nano-channel are briefly discussed. A possible strategy to combine MD/DPD simulation and continuum modeling for multi-scale modeling on flow through polymer-grafted nano-channel is proposed.
9:00 PM - LL5.19
O-vacancy and Proton Migration in the Barium Zirconate Doped by Yttrium.
Dae-Hee Kim 1 , Yong-Chan Jeong 1 , Dae-Hyun Kim 1 , Hwa-Il Seo 2 , Jong-Sung Park 3 , Byung-Kook Kim 3 , Yeong-Cheol Kim 1
1 Department of Materials Engineering, Korea University of Technology & Education, Chonan Korea (the Republic of), 2 School of Information Technology, Korea University of Technology & Education, Chonan Korea (the Republic of), 3 Center for Energy Materials Research, Korea Institute of Science and Technology, Seoul Korea (the Republic of)
Show AbstractAmong perovskite oxide materials, rare earth doped barium zirconate (RE-BZO, RE-BaZrO3) has been researched as a proton conducting material for application in a variety of electrochemical devices, including solid oxide fuel cells (SOFCs), sensors, electrolysis cells, and hydrogen pumps. We studied the proton migration path in BaZr1-xYxO3-0.5x (BZYO) using density functional theory (DFT). Y atoms preferred to be located far away from each other when two Y atoms were substituted for two Zr atoms to produce an O-vacancy in the 2×2×2 BZO structure. The migration energy barrier of the O-vacancy from the first nearest site to the second nearest site was 0.5 eV, while those from the first to the first, and the second to the second nearest site were 1 eV. Two protons, one located near the O atom of the Zr-O-Zr bond and the other located near the O atom of the Zr-O-Y bond, were energetically the most favorable when a water molecule was incorporated into the BZYO structure. The migration energy of the proton was 0.59 eV when the proton moved from the O atom of the Y-O-Zr bond to the O atom of a neighboring Y-O-Zr bond around the Y atom. This result was well matched to the experimental data.
9:00 PM - LL5.2
A Molecular Dynamics Study of the Growth and Melting of Metal Nanoparticles in Melt.
Yusuke Watanabe 1 , Yasushi Shibuta 1 , Toshio Suzuki 1
1 , The University of Tokyo, Tokyo Japan
Show AbstractThe growth and melting of metal nanoparticles in melt have been investigated as functions of particle size and temperature by a classical molecular dynamics simulation. There is a critical radius that demarcates the growth or melting of a nanoparticle in the liquid [1], which is in agreement with the classical theory of nucleation. The critical radius became smaller with decreasing temperature. Also, for a given nanoparticle radius, there is a critical temperature that demarcates the growth or melting of nanoparticle in the liquid. This critical temperature corresponds to the melting point including the curvature effect of the solid–liquid interface. The depression of the melting point due to the curvature effect is proportional to the inverse of the particle radius in melt. By comparison with the depression of the melting point of a freestanding nanoparticle [2], it was found that the Gibbs–Thomson coefficient should be estimated from the proportional constant of the depression of the melting point of the nanoparticle in melt.[1] Y. Shibuta, Y. Watanabe, T. Suzuki, Chem. Phys. Lett., 475 (2009) 264.[2] Y. Shibuta, T. Suzuki, J. Chem. Phys., 129 (2008) 144102.
9:00 PM - LL5.20
Effect of ZnO Addition on the Proton Conduction of the Barium Zirconate.
Yong-Chan Jeong 1 , Dae-Hee Kim 1 , Dae-Hyun Kim 1 , Hwa-Il Seo 1 , Jong-Sung Park 2 , Byung-Kook Kim 3 , Yeong-Cheol Kim 3
1 Department of Materials Engineering, Korea University of Technology & Education, Chonan Korea (the Republic of), 2 School of Information Technology, Korea University of Technology & Education, Chonan Korea (the Republic of), 3 Center for Energy Materials Research, Korea Institute of Science and Technology, Seoul Korea (the Republic of)
Show AbstractAmong perovskite oxide materials, rare earth doped barium zirconate (RE-BZO, RE-BaZrO3) has been researched as a proton conducting material for application in a variety of electrochemical devices, including solid oxide fuel cells (SOFCs), sensors, electrolysis cells, and hydrogen pumps. Zinc oxide (ZnO) has been added to improve the sinterability of BaZrO3. However, some adverse effects of the ZnO addition on proton conductivity have been reported in the literature. We therefore studied the effects of ZnO on the migration of the proton in BaZr1-xZnxH2O3-x using density functional theory (DFT). An oxygen vacancy can be produced between Zr and Zr or between Zr and Zn when a Zn atom is substituted for a Zr atom in the BaZrO3. The oxygen vacancy located between Zr and Zn was energetically more favorable than that between Zr and Zr. When an H2O molecule was incorporated into BaZr1-xZnxO3-x, two protons were attached to the O atoms in the structure. The two protons preferred to be located near the Zn atom energetically, indicating that ZnO, the sintering aid, attracted and trapped protons. This result explains the adverse effect of ZnO on proton conductivity.
9:00 PM - LL5.21
Study of Dielectric Constants of Ultrafine Silver Nanoparticles from First Principles.
Yi He 1 , Taofang Zeng 1
1 Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
Show AbstractDielectric constants and optical absorptivity of eight ultrafine silver nanoparticles are simulated using the density functional theory. The intraband and interband transitions of each silver nanoparticle are analyzed by a 2-dimensional plot method. The imaginary part of dielectric constants arising from intraband transitions are fitted to the Lorentz model by a simple statistical approach. A semi-empirical analytical equation is established to calculate dielectric constants of particles ranging from nanometers to macroscopic size. The size effect on the dielectric constants arising from the interband transition is also investigated, and a faster convergence to the bulk is observed.
9:00 PM - LL5.22
Electrical Conduction and Heat Transfer Simulations of Nanocrystalline Silicon Microwires Including the Thermoelectric Thomson Effect.
Gokhan Bakan 1 , Helena Silva 1 , Ali Gokirmak 1
1 Electrical and Computer Engineering, University of Connecticut, Storrs, Connecticut, United States
Show AbstractSeebeck (S) and Peltier (Π) effects are two commonly known thermoelectric phenomena, which are linked by Kelvin relations, established by William Thomson in 1851. Same relations also introduce another thermoelectric phenomenon, namely the Thomson effect. The Thomson effect is responsible for reversible heating/cooling of a uniform material depending on the material’s Thomson coefficient (β = T*dS/dT, T: absolute temperature), electric current, and temperature gradient. We present mathematical modeling of nanocrystalline silicon (nc-Si) microwires [1-4] using COMSOL Multiphysics, accounting for the Thomson effect. The simulations are motivated by SEM images of wires showing asymmetric self-heating when large electric current densities (J > 10 MA/cm2) are forced through the wires.Our model uses heat transfer and electrical conduction modules of the simulation tool. These two modules can handle Poisson’s equation for electrical potential and Fourier’s heat transfer equation; however, they do not include the thermoelectric effects. Thus, the Seebeck electric field (S*grad(T)) and Thomson heat (β*J●grad(T)) are introduced in the electrical conduction and heat transfer modules, respectively. Here, J is redefined in terms of both applied and Seebeck electric fields. Electrical, thermal, and thermoelectric parameters are defined on a large range of temperature (250 – 1690 K). Simulations results including thermoelectric effects verify that a considerable asymmetry in temperature profile is present due to the Thomson effect. Peak temperature on n-type wires shifts towards the end of the wire with lower electrical potential, where electrons enter, consistent with the distribution of light intensity along the wires in the experiments. Shift in the peak temperature location is greater for higher temperatures and current densities. Transient analysis of the wires suggests that steady state is achieved in 10 ns when the applied current does not melt the wire. Melting temperature of silicon (1690 K) is reached if the current is further increased. Simulation results on p-type wires show higher temperature on the side where the holes enter the wire. [1] G. Bakan, A. Cywar, H. Silva and A. Gokirmak, "Melting and crystallization of nanocrystalline silicon microwires through rapid self-heating," Appl. Phys. Lett., 2009. [2] G. Bakan, A. Cywar, C. Boztug, M. Akbulut, H. Silva and A. Gokirmak, "Annealing of nanocrystalline silicon micro-bridges with electrical stress," in Mater. Res. Soc. Symp. Proc. Fall 2008, 2009, pp. LL03-25. [3] A. Cywar, G. Bakan, C. Boztug, H. Silva and A. Gokirmak, "Phase-change oscillations in silicon microwires," Appl. Phys. Lett., vol. 94, pp. 072111, 2009. [4] C. Boztug, G. Bakan, M. Akbulut, N. Henry, A. Gokirmak and H. Silva, "Numerical modeling of electrothermal effects in silicon nanowires," in Mater. Res. Soc. Symp. Proc. Spring 2008, 2008, pp. R04-11.
9:00 PM - LL5.24
Thermo-mechanical Properties of Polyurea and Polyurethane.
Jean Njoroge 1 , Tahir Cagin 1
1 , Texas A&M university , College Station , Texas, United States
Show AbstractMolecular dynamics (MD) simulations were used to predict the properties of amorphous polyurethane (PUR) and polyurea (PU). These polymers are good candidates as matrix materials for developing nanocomposites for blast and shock protection in both civilian and defense applications. Determination of ultimate thermo-mechanical properties of matrix material is critical for development of better polymer nanocomposites for these applications. We have calculated thermodynamic, structural and mechanical properties of these polymers including bulk, shear and Young’s modulus (at various pressures and elevated temperature) and the glass transition temperature (at near-ambient pressure). The model systems were obtained via thermal and pressure annealing of the densely packed amorphous PUR and PU. We employed widely used Dreiding force fields, with the exponential six forms of van der Waals interactions. The electrostatic interactions were evaluated using Ewald Summation Method where the atomic charges are charges determined from charge equilibration method. Keywords: Molecular dynamics; polyurea; polyurethane
9:00 PM - LL5.25
Hydrogen Diffusion Behavior around Lattice Defects in BCC Iron Based on Phase-Field Modeling.
Hideki Mori 1 , Hajime Kimizuka 1 , Shigenobu Ogata 1
1 Mechanical Science and Bioengineering, Osaka University, Osaka Japan
Show AbstractAtomic simulation shows that the hydrogen atoms attenuate the chemical-bonding state of metals and are strongly trapped at lattice defects in metals. On the other hand, the elastic analysis clarifies that the state of defects and the hydrogen-concentration distribution affect each other through the stress fields. For understanding the hydrogen-embrittlement phenomena in steel, it is important to evaluate the cooperative interactions among combinations of hydrogen atoms, hydrogen-trapping defects (dislocation, grain boundary, surface etc.) and stresses during the deformation process. We evaluate the distribution behavior of hydrogen concentration around dislocations and grain boundaries based on a phase-field (PF) microelasticity theory. By using the PF theory, we can take into account both the long-range elastic interactions and short-range chemical interactions between the defects and hydrogen concentration. To obtain the physical parameters included in the PF free-energy functional, the interaction energies between a hydrogen atom and dislocations and grain boundaries are quantitatively determined using an embedded-atom-method (EAM) and density-functional-theory (DFT) calculations. Based on these data, we investigate an evolution of the hydrogen diffusion and concentration around dislocations and grain boundaries. It is clearly observed that the hydrogen is significantly localized and concentrated around dislocation cores and grain boundaries, so that the remarkable difference exists in hydrogen concentration between in the bulk region and in the vicinity of defects, ranging from several weight ppm to several thousands weight ppm. Also, the spatial profile of trapped hydrogen around dislocations strongly depends on the stress field produced by dislocations.
9:00 PM - LL5.26
Multi-scale Studies of Wax-Cellular Solid Composites for Controllable Stiffness Robotic Elements.
Nadia Cheng 1 , Arvind Gopinath 1
1 , Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
Show AbstractWe are investigating the use of a new class of materials for realizing soft robots. Specifically, meso-scale composites—composed of cellular solids impregnated with active fluids—were designed to have controllable stiffness to take the form of a continuous body of a soft robot that embodies rigid, load-bearing components that can transform into morphable, compliant ones. Research has been focused on developing a specific composite: an open-cell flexible foam impregnated with wax. The unique concept of using wax as a structural element is that it can be thermally-activated to undergo drastic changes in viscosity and stiffness. Therefore, this novel composite can exhibit a combination of solid and fluid-like behavior. Additionally, “shape memory” and wickability characteristics enable the wax-foam composite to be self-restoring after being deformed.To understand their performance capabilities as load-bearing structures, the composites have been characterized in their solid states both analytically—using quantitative mesoscale modeling—and experimentally. The effective elastic modulus of the composites was varied using two methods: 1) by varying the relative volume of wax contained in the foam scaffold, and 2) by designing various cellular geometries of the foam component of the composite. The latter was achieved by 3D printing soft cellular materials and computationally optimizing their designs to achieve desired performance specifications. For both methods, extensions of classical composite theory yield scalings for the effective compression modulus. These scaling predictions were validated by experimental results from compression tests performed on the composites. Finite element analysis at the macro and meso scales delivers a more detailed characterization of composition-structure relationships and mesoscopic material response.
9:00 PM - LL5.3
First Principles Calculations of Nitrogen Atomic Position Effects on Elastic Properties of Aluminum Oxynitride (AlON) Spinel.
Iskander Batyrev 1 , Betsy Rice 1 , James McCauley 1
1 , US Army Research Laboratory, Aberdeen Proving Ground, Maryland, United States
Show AbstractThe elastic properties of aluminum oxynitride spinel (AlON) have been calculated from first principles. We have assumed an “ideal” stoichiometry of cubic AlON with 35.7 mole % AlN using the constant anion model, and assuming an Al vacancy on the octahedral site: Al8IVAl15VI ■VIO27 N5. The local density approximation (LDA) and the general gradient approximation (GGA) projected augmented waves (PAWs) were used for the simulations. Independent strains were applied to a unit cell, parameterizing the total energy as a function of the strain and calculated elastic constants. The purpose of the calculations is to determine if the location and/or segregation of N atoms in the unit cell affects the elastic properties of AlON. The calculations have been carried out for two general structural models of nitrogen atomic arrangements in the unit cell: 1) random to clustered and 2) adjacent to and not adjacent to the Al octahedral vacancy. It was found that using LDA or CGA in PAWs had slightly different results, however: LDA results in ~ 10% higher equilibrium volume and bulk modulus of AlON compared with the GGA. A clustered distribution of N atoms away from the Al vacancy has ~ 1 eV per 55 atoms higher total energy than for a random distribution and about 5% higher elastic constants. A configuration with randomly distributed N atoms next to the Al vacancy has almost equal total energy of the system, compared to when N atoms are randomly distributed not next to the vacancy. A clustered distribution of N atoms adjacent to the Al vacancy results in a slight gain of energy (~20 meV) compared with a random distribution of N atoms and a significant gain of energy compared with a cluster distribution of N atom not adjacent to the Al vacancy (1.11 eV). The cluster distribution results in higher bulk modulus compared with a random distribution of N atoms. The calculated elastic constants are in overall agreement with experimental measurements. The variation of C11, C12 and C44 for random and cluster distributions of N atoms is in the range of distribution of experimental data.
9:00 PM - LL5.4
First-principles Prediction of Unknown Metastable Crystalline Phase of Gallium.
Maurice de Koning 1 , Alex Antonelli 1 , Diego Alejandro Carvajal Jara 1
1 Instituto de Física "Gleb Wataghin", Universidade Estadual de Campinas, Campinas Brazil
Show AbstractWe report evidence for the existence of an unknown metastable crystalline phase of Gallium (Ga) by the combination of classical molecular dynamics (MD) simulations and density-functional theory (DFT) calculations. The MD simulations, based on a modified embedded-atom (MEAM) potential, reveal the unknown crystalline form through a first-order phase transition originating from the Cmca symmetric α-Ga phase under hydrostatic tension. Subsequently, the DFT calculations using two different GGA functionals are employed to verify its stability and determine its electronic structure. The structure of the orthorhombic phase is described by symmetry group Cmcm and shows a dimer arrangement resemblant of the α-Ga phase. A first-order phase transition from α-Ga to the unknown phase is estimated to occur at -1.3 GPa.
9:00 PM - LL5.5
Ab Initio Calculations of Vacancy Rearrangement and Substitution of Copper by Palladium in Copper Nitride.
Maria Moreno-Armenta 1 , Gerardo Soto 1 , Noboru Takeuchi 1
1 Nanoestructuras, universidad Nacional Autonoma de Mexico, Ensenada, Baja California, Mexico
Show AbstractThe study of copper nitride is important because it could be used as a host to create intercalated alloys, and as a material for optoelectronic recording. Its crystalline structure is cubic anti-ReO3, which leaves vacancies in the cube-center sites. It has been proposed that these vacancies can be filled out by foreign atoms [1-3], and even Cu [4], to form intercalated alloys. In this work we present first principles calculations of vacancy reorganization and the substitution of copper by palladium atoms in copper nitride. The calculations were performed using the DFT plane-wave pseudopotential approach within the generalized gradient approximation as implemented in the Quantum Espresso package [5]. To calculate the effects of low amounts of foreigner atoms, we have used a 2x2x2 Cu3N-like cubic supercell with an equivalent reduced symmetry. Small Pd concentrations and vacancy rearrangements in the copper sublattice were created in these low-symmetry cells. We considered some possible situations: (1) the occupation of the already present vacancies by copper and (2) by palladium; (3) the relocation of vacancies in the copper sublattice; and (4) the substitution of copper by small quantities of palladium in the copper sublattice. The equilibrium volumes and energies after relaxation of the resulting structures are compared to those of intrinsic copper nitride. The lattice constants decrease if the vacancies are interchanged with copper sites and there are new combinations for the vacancies positions. This switching of positions makes no difference if the extra atoms are palladiums or copper. The lattice parameter remains almost the same as in a 2x2x2 Cu3N-like supercell, even though it has 34 atoms instead of 32. We observed that changes to the ideal Cu3N stoichiometry result in a shift from a semiconductor to a metallic or semimetallic state. Keywords: Transition metal nitrides; copper nitride; ab initio; interstitial alloys; DFT.AcknowledgementsThis work was supported by the Supercomputer Center DGSCA-UNAM, DGAPA and CONACyT Grants IN108908 and 89768 respectively.References[4] Moreno-Armenta, Lopez, Takeuchi, Solid State Science, 9 (2007) 166.[5] Pierson and Horwat, Scripta Materialita, 568 (2008).[6] A. Ji, C. Li, Z. Cao, Appl. Phys. Lett. 89, 252120 (2006).[7] Moreno-Armenta, Martinez, Takeuchi Solid State Sciences 6 (2003) 9.[11] Baroni, Corso, Girocoli, Giannozzi, Cavazzoni, Ballabio, Scandolo, Chiarotti, Focher, Pasquarello, et al (http//www.democritos.it)
9:00 PM - LL5.6
Multiphysics Modeling of Grain Boundaries in Multicrystalline Silicon.
Hiroshi Mizuseki 1 , Ambigapathy Suvitha 1 , Ryoji Sahara 1 , Yoshiyuki Kawazoe 1
1 , Institute for Materials Research, Tohoku Univ., Sendai, Miyagi, Japan
Show AbstractMulticrystalline silicon (mc-Si) is widely used as a solar cell material because of its low production cost, even though the energy conversion efficiency of mc-Si solar cells is lower than that of single-crystalline Si solar cells due to the random orientations of the crystal grains in the former. Optimization of the grain-boundary structures of mc-Si is a key issue to achieving high efficiency, because these regions act as recombination centers for carriers in solar cell materials. Multicrystalline Si with artificially-controlled grain orientations has been proposed as a means of reducing the number of electrically-active grain boundaries that lead to undesirable carrier recombination.[1] In the present study, we evaluate the ‘grain-boundary-energies’ at the grain boundaries between <100>, <110>, <111> and <112>-oriented grains by using the Tersoff potential.[2] First, we prepare two silicon crystals which have the same crystal orientations. Then, we independently rotate each crystal about the same direction, such as the <110>-oriented direction. Finally, we combine both crystals to form a grain boundary in a spherical sample, in which one hemisphere of the sphere model corresponds to the first crystal and the other hemisphere corresponds to the second crystal. Moreover, the effects of impurity (Cr, Fe, Cu & Ni) at the grain boundary plane as point defects on the electronic structures of Coincidence Site Lattice grain boundaries such as Σ 3, Σ 5, and Σ 9 with 96, 80, and 288 atoms model have been examined using the density-functional theory with the plane-wave pseudopotential method. This work was partially supported by New Energy and Industrial Technology Development Organization (NEDO) of Japan.[1] N. Usami et al., Jpn. J. Appl. Phys. 45 (2006) 1734. [2] J. Tersoff, Phys. Rev. B39 (1989) 5566.
9:00 PM - LL5.7
Anti-ferroelectric PbZrO3 (001) Surfaces: An ab initio Study.
Ghanshyam Pilania 1 , Rampi Ramprasad 1
1 CMBE, University of Connecticut, Storrs, Connecticut, United States
Show AbstractPerovskite oxides of ABO3 type constitute an important class of materials which display a wide variety of electronic properties and complex structural instabilities such as ferroelectric, anti-ferroelectric, ferrodistortive, etc. These properties result due to a delicate balance between short range repulsive forces and long range coulombic forces and get modified significantly close to a surface. Although detailed experimental and theoretical studies of the surfaces of various ferroelectric perovskites such as BaTiO3, PbTiO3, SrTiO3 etc., have been performed, such studies of an anti-ferroelectric system such as PbZrO3 has not been reported to date.We have carried out first-principles total energy calculations of bulk and (001) surfaces of PbZrO3. The ground state for bulk PbZrO3 is determined to be the anti-ferroelectric orthorhombic phase, with the ferroelectric rhombohedral and paraelectric cubic phases being 0.14 and 0.39 eV per formula unit higher in energy, respectively. PbO- and ZrO2-terminated (001) surfaces were investigated; these surfaces were either clean or decorated with adsorbed hydroxyl species. The hydroxyl species were intended to model PbZrO3 nanoparticle surfaces that were highly oxidized. Surface relaxations, in-plane anti-ferroelectric distortions and modifications to the electronic structure due to the surfaces and the hydroxyl adsorbates on the surfaces were investigated. We found that in the case of the PbO-terminated clean (001) surface, the nature of the anti-ferroelectric distortions in the surface regions is almost identical to that in bulk PbZrO3. In contrast, the ZrO2-terminated (001) surface favors anti-ferroelectric distortions significantly different from those found in the bulk PbZrO3. However, in both of these cases no surface states were found in the electronic structure of the surface region. Hydroxyl-covered surfaces, intended to model excessively oxidized (001) surfaces, display markedly different behavior with significant changes to the surface geometry as well as electronic structure. Adsorbate-induced surface states in the band gap are indicative of the possible metallic nature of these surfaces.
Symposium Organizers
Mark Asta University of California
Alex Umantsev Fayetteville State University
Joerg Neugebauer Max-Planck-Institut fuer Eisenforschung
LL6: Meso- and Macro-Scale Simulations
Session Chairs
R. Spatschek
Alex Umantsev
Wednesday AM, December 02, 2009
Room 313 (Hynes)
9:30 AM - **LL6.1
Pattern Formation During Diffusion Limited Transformations in Solids.
M. Fleck 1 , C. Hueter 1 , D. Pilipenko 1 , R. Spatschek 2 , E. Brener 1
1 Institut fur Festkorperforschung, Forschungzentrum Juelich, Juelich Germany, 2 Interdisciplinary Centre for Advanced Materials Simulation (ICAMS), Ruhr-Universitaet Bochum, Bochum Germany
Show AbstractKey feature of many metallurgical procedures to improve materials properties is the formation of very complex microstructures due to solid-solid phase transformation processes. We develop a description of diffusion limited growth in solid-solid transformations, which are strongly influenced by elastic effects. Density differences and structural transformations provoke stresses at interfaces, which affect the phase equilibrium conditions. We formulate equations for the interface kinetics similar to dendritic growth and study the growth of a stable phase from a metastable solid in both a channel geometry and in free space. We perform sharp interface calculations based on Green’s function methods and phase field simulations, supplemented by analytical investigations. For pure dilatational transformations we find a single growing finger with symmetry breaking at higher driving forces, whereas for shear transformations the emergence of twin structures can be favorable. We predict the steady state shapes and propagation velocities, which can be higher than in conventional dendritic growth.
10:00 AM - LL6.2
Nucleation and Successive Microstructure Evolution via Phasefield and Phasefield Crystal Method.
Heike Emmerich 1
1 Computational Materials Engineering, RWTH Aachen University, Aachen Germany
Show AbstractIt is well known, that the mechanical material properties of a material sample after solidification are strongly tied to its microstructure structure. Nevertheless, the precise laws governing the initial stage of this structuring process, i.e. nucleation and the successive transiental microstructure evolution scenario's, are still far from being fully understood.Here we show, that the phase field method [1], which originally established itself to tackle the free boundary problem given by microstructure evolution in a multiphysics environment, can also be employed to investigate the energetics of heterogenous nucleation in a solidifying sample [2]. Moreover it is demonstrated, how the phasefield crystal method can shade more light in open questions regarding a quantitative formulation of nucleation statistics to thereby simulate the phase transition phenomena in solidification from nucleation to crystallization in larger domains thoroughly [3,4].In the talk special emphasis is put on pointing out for what kind of input the phase-field respectively the phase-field crystal approach depend on ab-initio input to achieve a multi-scale simulation approach.[1] see e.g. H. Emmerich, Advances in Physics 57, 1 (2008)and references therein[2] H. Emmerich, R. Siquieri, J. Phys.: Condens. Matter 18, 11121 (2006)[3] E. Doernberg, R. Siquieri, R. Schmid-Fetzer, H. Emmerich accepted for publication in JPCM[4] R. Prieler, B. Verleye, R. Haberkern, D. Li, H. Emmerich, submitted to JPCM
10:15 AM - LL6.3
A Numerical Method for the Phase Field Crystal Model.
Victor Chan 1 , Nirand Pisutha-Arnond 1 , Mrinal Iyer 2 , Vikram Gavini 2 , Katsuyo Thornton 1
1 Material Science and Engineering, University of Michigan, Ann Arbor, Ann Arbor, Michigan, United States, 2 Mechanical Engineering, University of Michigan, Ann Arbor, Ann Arbor, Michigan, United States
Show AbstractThe phase field crystal (PFC) model is a technique for simulating crystal growth at atomic spatial scales on diffusive time scales and is a promising method for studying nano/microstructural evolution. We present an alternative numerical method that can be used to evolve the PFC equation. This method includes a fit of the atomic correlation function that captures its detail beyond the first peak. The model is evaluated against the existing PFC/classical density functional theory models, as well as experimental data. The model is used to simulate liquid-solid and solid systems and the liquid-solid transition of copper and iron. Properties such as bulk moduli and interfacial energies are calculated and compared with other model predictions and experimental results.
10:30 AM - LL6.4
Diffuse Interface Model and Simulations of Kirkendall-Effect-Induced Deformation.
Hui-Chia Yu 1 , Katsuyo Thornton 1
1 Materials Science and Engineering, University of Michigan, Ann Arbor, Michigan, United States
Show AbstractThe Kirkendall effect stems from the difference between the exchange rates of the atomic species and vacancies in a substitutional alloy. Vacancies that mediate diffusion are generated and eliminated at their sources and sinks, resulting in lattice shift and deformation. In the conventional treatment, these vacancy sources and sinks are assumed to be distributed everywhere in a solid and maintain vacancy concentration at a constant, uniform equilibrium value throughout the solid. We employ a smooth boundary method to impose a no-flux boundary condition at the solid surfaces to simulate interdiffusion and Kirkendall-effect-induced deformation. This method was first applied to the traditional interdiffusion model coupled with a linear visco-plastic deformation model. Expansion, contraction, bending, as well as a complex combination of these types of deformations were studied. The new method was also applied to a system in which vacancy sources and sinks are placed only at the surface to simulate interdiffusion and resultant shape changes of a single crystalline solid without bulk vacancy sources.
11:30 AM - **LL6.6
Predicting Phase Transition Behavior of Nanoscale Olivine Cathodes by Diffuse-Interface Modeling.
W. Craig Carter 1 , Ming Tang 2 , Yu-Hua Kao 1 , Nonglak Meethong 3 , Yet-Ming Chiang 1
1 Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States, 2 , Lawrence Livermore National Laboratory, Livermore, California, United States, 3 , Khon Kaen University, Khon Kaen Thailand
Show AbstractOlivine type lithium metal phostphates have emerged as important cathode materials for high-power lithium ion rechargeable batteries. Bulk olivine materials display a phase transition between a lithiated and delithiated phase during electrochemical cycling at room temperature. However, recent experimental observations indicate that the phase stability is significantly altered in nanoscale particle size regime, as exemplified by the formation of amorphous phase in nano-particles upon charge/discharge. A diffuse-interface modeling approach was employed to explain and predict such size-dependent phase transition behavior. Our model considers various contributions towards a cathode particle's free energy including volume free energy, surface energy and elastic strain energy. Nucleation energy barrier calculations and kinetic simulations of the delithiation process were carried out with the model. It is predicted that below a critical particle size, there exists an intermediate overpotential window within which amorphization is kinetically favorable, while the conventional crystalline phase transition remains the dominant transition pathway outside this window overpotentials. Such predictions have been confirmed by recent experiments.
12:00 PM - LL6.7
Stress-induced Morphological Instabilities at the Nanoscale Examined Using the Phase Field Crystal Approach.
Kuo-An Wu 1 , Peter Voorhees 1
1 Materials Science & Engineering, Northwestern University, Evanston, Illinois, United States
Show AbstractThe stress-induced morphological instability of an interface is examined with the recently developed phase field crystal (PFC) approach. Our results show that the interface width, generally assumed to be zero thickness in classical theory, is a crucial length scale for phenomena at the nanoscale. We find that the critical wavenumber of the instability deviates from the prediction of the continuum theory when the length scale of the instability is on the order of the interface width. In addition, we find that large stresses induce nonlinear elastic effects that alter both the wavelength of the instability and the interfacial morphology. This nonlinear elastic effect is generic and thus will be observable during heteroepitaxy. Finally, we show that the relaxational dynamics employed in the PFC model can capture quantitatively the stress field that accompanies the diffusional evolution of interfaces.
12:15 PM - LL6.8
Self-Consistent Field Theory Study of Block Copolymer / Nanoparticle Composites.
Kahyun Hur 1 , Richard Hennig 1 , Fernando Escobedo 2 , Ulrich Wiesner 1
1 Materials Science and Engineering, Cornell University, Ithaca, New York, United States, 2 Chemical and Biomolecular Engineering, Cornell University, Ithaca, New York, United States
Show AbstractFunctionalized mesoporus materials are attracting much attention for their potential applications in electronics, optics, and energy. One feasible route for their synthesis is block copolymer (BCP)/nanoparticle (NP) self-assembly. BCP self-assembly enables synthesizing novel materials with ordered nanostructures of NPs. However, a full understanding of the underlying physics that governs the morphologies of BCP/NP composites is still a challenge. Reliable prediction of structure and quantitative analysis of properties by theoretical and simulation approaches may provide helpful insights. In this work, we study BCP/ NP self-assembly using a combination of self-consistent field theory (SCFT) and density functional theory (DFT). Long range Coulomb and hard sphere interactions between NPs embedded in BCPs are implemented exploiting the SCFT/DFT approach. We find that exothermic mixing of BCPs and NPs triggers a transition from lamellar to inverse hexagonal morphology consistent with experiments in the dense particle regime. Coulomb interactions between NPs significantly affect BCP/NP composite structure, and induce a surprising NP lattice within the phase separated domains.
12:30 PM - LL6.9
Mesoscale Simulations of Self-Assembly of Arbitrarily-Shaped Particles in Bulk and at Fluid-Fluid Interfaces.
Paul Millett 1 , Yu Wang 2
1 , Idaho National Laboratory, Idaho Falls, ID 83415, Idaho, United States, 2 Materials Science and Engineering, Michigan Tech, Houghton , Michigan, United States
Show AbstractRecent advances in the ability to experimentally control the size, shape, and composition of nanoparticles have significantly broadened the possibilities to create novel mesoscale structures as a result of their “bottom-up” assembly. A particularly efficient approach to facilitate various assembly dynamics in colloidal systems is to control the collective electrostatic interactions by tuning the particle charge density and/or dipole moment as well as the application of an external electric field. Here, we present a novel mesoscale simulation approach that utilizes diffuse interface fields to capture and investigate the dynamic assembly processes for arbitrarily shaped particles with arbitrary charge density and/or dipole moment. Illustrative results demonstrating the method’s ability to predict a wide variety of colloidal crystal structures, with a particular focus on binary lattices consisting of positively- and negatively-charged particles, will be presented. Furthermore, this mesoscale approach has also been extended to include the capillary forces experienced by particles segregated at fluid-fluid interfaces. We will present simulations of the complex evolutions of multi-phase fluid mixtures in which particle absorption to the interfaces significantly alters the dynamics of system.
LL7: Thermodynamics and Nucleation; Point Defects
Session Chairs
Wednesday PM, December 02, 2009
Room 313 (Hynes)
2:30 PM - **LL7.1
Computational Phase Studies: Deriving Free Energies and Phase Transitions from First Principles.
Tilmann Hickel 1
1 Computational Materials Design, MPI für Eisenforschung GmbH, Düsseldorf Germany
Show AbstractDue to their direct relation to many technologically relevant properties and processes in metals, the accurate prediction of phase diagrams is highly significant in computational materials science. Well established simulation tools (such as, e.g., CALPHAD) have been very successful in predicting complex phase diagrams, if sufficient experimental input data are available. Density functional theory in combination with mesoscopic/macroscopic thermodynamic concepts, on the other hand, provides a promising approach to determine phase transition temperatures for well defined material systems from first principles only. However, the key for a reliable prediction of phase diagrams from ab initio or for providing supplementary information to CALPHAD is the availability of efficient and highly accurate theoretical tools to determine all contributions to the free energy including harmonic/anharmonic vibrations, electronic/magnetic excitations and their combinations. We have therefore performed an extensive and systematic study of the capabilities of present day implementations (xc-functionals) of densitiy functional theory in determining ab initio free energies for metals. Lattice vibrations, which yield the dominant contribution to the free energy of elementary, non-magnetic materials, can be determined within the quasiharmonic approximation. We show for a large set of metals that the thus derived thermodynamic properties are in excellent agreement with available experimental data [1]. For magnetic materials such as iron we have developed a proper quantum-mechanical treatment of magnetic excitations, since classical simulations showed significant shortcoming in determining magnetic free energies [2]. An integrated approach, combining electronic, vibrational, and magnetic effects, leads us to an extremely high accuracy of only a few meV for the free energy of the considered metals. Using the thus determined free energies, we successfully predict martensitic phase transition temperatures in complex materials such as steels and shape memory alloys [3]. References: [1] B. Grabowski, T. Hickel and J. Neugebauer, Phys. Rev. B 76, 024309 (2007).[2] F. Körmann, A. Dick, B. Grabowski, B. Hallstedt, T. Hickel, J. Neugebauer, Phys. Rev. B 78, 235302 (2008). [3] M.A. Uijttewaal, T. Hickel, J. Neugebauer, M.E. Gruner, P. Entel, Phys. Rev. Lett. 102, 035702 (2009).
3:00 PM - LL7.2
Anharmonic Effects on Pressure-Temperature Phase Diagram of ZnO — A First-Principles Molecular Dynamics Study.
Yuji Takahashi 1 , Atsuto Seko 2 , Isao Tanaka 1 3
1 Department of Materials Science and Engineering, Kyoto University, Kyoto Japan, 2 Pioneering Research Unit for Next Generation, Kyoto University, Kyoto Japan, 3 Nanostructures Research Laboratory, Japan Fine Ceramics Center, Nagoya Japan
Show AbstractZnO crystallizes in wurtzite structure at ambient conditions. It undergoes a phase transition to rocksalt structure at 6 - 8 GPa. According to experimental works, the transition pressure decreases with increasing temperature. A first-principles lattice-dynamics calculation based on the quasi-harmonic approximation (QHA) in the present authors’ group[1] has reproduced the temperature dependence of the transition pressure only in the low temperature region below approximately 700 K. As the temperature increases, the computed transition pressure significantly deviated from the experimental values. In order to elucidate the mechanism of the deviation, we examine the effect of anhamonic lattice vibration on the theoretical transition pressure in the present study. In QHA, the anharmonicity is restricted to thermal expansion. In other words, only volume-dependent quadratic terms in the potential energy are considered. In our previous study[1], vibrational free energy at the given volume was computed using phonon density of states within QHA. The other part of the anharmonicity due to cubic and higher-order terms were neglected. In the present study, we perform a set of first principles molecular dynamics (FPMD) calculations in order to evaluate the magnitude of the anharmonicity. Calculations are made with a plane-wave basis PAW method as implemented in VASP code. Harmonic Phonon spectra are computed with a frozen phonon method using fropho code. FPMD calculations are made with the NEV ensemble. The vibrational free energy is evaluated from the internal energy by MD using thermodynamic integration scheme. The anharmonic part is obtained by subtracting the classical harmonic vibrational free energy term from the vibrational free energy by MD. The vibrational free energy is computed in wide ranges of volume and temperature for two phases of ZnO. Then the transition pressure is determined as a function of temperature. The computed transition pressure shows better agreement with experimental results at high temperatures when the anharmonic effects are included with the procedure as described above. The anharmonic effects decrease the free energy of the wurtzite phase and increase that of the rocksalt phase when compared at the same temperature. It therefore stabilizes the wurtzite phase and increases the transition pressure. The transition pressure increases by 0.3 GPa at 500 K, 1 GPa at 1000 K and 3 GPa at 1500 K. Recalling the experimental transition pressure in the range of 6 − 8 GPa, the present result implies that the inclusion of the anharmonic effects beyond QHA is essential for quantitative prediction of the transition pressure or pressure-temperature phase diagram in the high temperature region. [1] A. Seko, F. Oba, A. Kuwabara, and I. Tanaka, Phys. Rev. B 72, (2005).
3:15 PM - LL7.3
Ab initio Up To The Melting Point: Efficient Sampling Strategies of Anharmonic Free Energy Contributions in Bulk and for Vacancies.
Blazej Grabowski 1 , Lars Ismer 1 , Tilmann Hickel 1 , Joerg Neugebauer 1
1 , Max-Planck-Institut, Düsseldorf Germany
Show AbstractTemperature dependent thermodynamic phase diagrams of bulk crystallographic structures are important tools in materials design. Traditionally, these diagrams are constructed based on experimentally determined free energies for elementary and a few binary and ternary phases together with suitable interpolation schemes such as the CALPHAD approach. In the past years, missing experimental data, as e.g. for metastable or even unstable phases, has been computed based on T = 0 K ab initio (density functional theory; DFT) total energies [1]. A study of whether the presently available DFT approaches are sufficiently accurate to allow predicting the temperature dependence was largely hampered by the high accuracy needed (better 1 meV/atom). A particular challenge is the determination of the anharmonic contribution: for a statistically converged result more than 106 configurations are needed.
We have therefore developed methods which allow an efficient determination of all relevant free energy contributions for elementary non-magnetic metals [2,3]. Using fully DFT based coarse graining techniques in configuration space we succeeded in reducing the number of configurations from 106 to a few hundreds at the highest and thus computationally most expensive level. Using this approach we were able to guarantee a numerical accuracy of better 1 meV/atom - remaining errors can thus be exclusively related to the xc-functional. These highly accurate results, including all relevant excitation mechanisms up to the melting point, allowed e.g. to tackle a long standing debate about the dominating physical mechanisms determining the isobaric heat capacity of aluminum close to the melting point [3].
[1] H.L. Skriver, Phys. Rev. B 31, 1909 (1985).[2] B. Grabowski, T. Hickel, and J. Neugebauer, Phys. Rev. B 76, 024309 (2007).[3] B. Grabowski, L. Ismer, T. Hickel, and J. Neugebauer, Phys. Rev. B 79, 134106 (2009); see also the Viewpoint in Physics: G. Grimvall, Physics 2, 28 (2009).
4:00 PM - LL7.4
Evolution of the Size Distribution and Shape of Nanoparticles and Nanorods: Experiments, Modeling and Simulation.
Rajdip Bandyopadhyaya 1 , Mani Ethayaraja 1
1 Chemical Engg. Dept., Indian Institute of Technology Bombay, Mumbai India
Show AbstractSpherical, solid nanoparticles and cylindrical nanorods are the basic building blocks of different complex nanostructures of interest to the material science community. They are currently actively investigated for various interesting applications - starting from bulk additives and catalysts on the one hand to sophisticated sensors and optoelectronic devices, on the other end. To optimize the performance, one needs to a-priori predict the size, shape and composition of these nanostructures, as the relevant physical and chemical properties are directly determined by the quantum size effect and the specific surface area of the nano-sized objects. To this end, we have modelled a series of increasingly complex nanostructures - starting from single component, spherical nanoparticles, to core-shell, two component spherical nanoparticles, leading finally to cylindrical nanorods. Our models are based on time scale analysis of the various elementary steps namely, material transport, solubilization, reaction, nucleation, growth, coagulation etc. Time scales analysis has identified the relevant rate limiting processes, based on which we have proposed mechanisms of formation and growth of any given nanostructure.With this mechanism in place, we have developed both a deterministic population balance equation (PBE) based modelling framework and a stochastic kinetic Monte Carlo (MC) simulation algorithm. These two complimentary techniques have been applied to predict size distribution of core nanoparticles (like calcium carbonate, calcium hydroxide, cadmium sulphide etc.), shell thickness of core-shell particles (like cadmium sulphide-zinc sulphide) and aspect ratio (length to diameter) distribution of nanorods (like zinc oxide, zinc sulphide etc.). The predictions have been compared in each case with experimental data of our group, or from literature, thus validating the underlying mechanism of nanoparticle or nanorod formation. This has thrown light on how the individual steps can be tailored in different experimental scenario to achieve nanostructures with a desirable size or shape. The PBE models and MC simulations are fairly general; we have applied these to understand the dynamics of nanostructure formation in both bulk liquid solutions or in self-assembled templates, like water-in-oil microemulsions. Our framework can be suitably extended to predict nanobelts, ribbons, helices and tripod or tetrapod shapes, thus completing the gamut of experimentally obtained different exotic nanostructures.Thus given any new experimental system, our modeling and simulation framework enables one to select an appropriate mechanism (based on a simple time-scale analysis of the above processes); and then from it obtain a-priori prediction of the size and shape evolution of the expected nanostructure.
4:15 PM - LL7.5
Size-Dependent Nucleation Kinetics at Non-planar Nanowire Growth Interfaces.
Tomorr Haxhimali 1 2 , Mark Asta 2 , Dorel Buta 2 , Jeffrey Hoyt 3 , Peter Voorhees 1
1 Material Science & Engineering, Northwestern University , Evanston, Illinois, United States, 2 Chemical Engineering & Material Science , University of California, Davis, California, United States, 3 Materials Science and Engineering, McMaster University, Hamilton, Ontario, Canada
Show AbstractIn catalyst-mediated nanowire growth, kinetic processes at the nanowire-catalyst interface can play an important role in governing wire compositions, morphologies and growth rates.In this talk we present a theoretical and computational study on the growth mechanism of solid wires and nanowires at the solid-liquid interface by the Vapour-Liquid-Solid (VLS) ans Vapour-Solid-Solid (VSS) method. The theoretical model was developed based on insights derived from atomistic simulations employing a model potential for elemental Si, with the driving force for growth applied by undercooling. Molecular-dynamics simulation established that in equilibrium the wire-catalyst interface is non-planar with a (111) faceted orientation bounded by curved (rough) orientations near the solid-liquid-vacuum contact line. The curved portions of the solid-liquid interface lead to a large capillaryundercooling which increases in magnitude with decreasing nanowirediameter. In non-equilibrium growth simulations, the interface shape ispreserved. Simulated growth rates display a strong dependence on nanowire diameter, consistent with a size-dependent barrier for facet nucleation. Growth is observed to proceed in a layer-by-layer mode with arate limited by the nucleation of new (111) terraces. For a given drivingforce, measured as the undercooling below the capillary-correctedcoexistence temperature, the growth rates are observed to increase withdecreasing nanowire diameter. These results are interpreted to reflect asize dependence of the barrier for terrace nucleation. The origin of thiseffect and its consequences for the synthesis of nanowires from liquidalloy catalysts in the vapour-liquid-solid growth method will be discussed.The kinetic theory that we present for nanowire growth is applicable to both pure systems(as in the MD simulations) and alloys (as in experiment) with non-planar interfaces.For pure system the above theoretical model predicts exactly the same form for the dependence ofgrowth velocity on driving force as in MD simulations. In the limit of small driving force, as in experiment, this theory predictsthat linear mobility decays exponentially while increasing the size.In a general context, this study establishes that system size-dependent growth rates are an intrinsic feature of systems with non-planar interfaces.
4:30 PM - LL7.6
Kinetic Path of Precipitation in Si-Ge-P Alloy Studied by Kinetic Monte Carlo Simulations.
Celine Hin 1 , Mildred Dresselhaus 2 , Zhifeng Ren 3 , Gang Chen 1
1 Mechanical Engineering, MIT, Cambridge, Massachusetts, United States, 2 Physics, MIT, Cambridge, Massachusetts, United States, 3 Physics, Boston College, Cambridge, Massachusetts, United States
Show AbstractRecently, nanostructured bulk materials like n-type SiGe with 30 nm grain size exhibit a figure of merit of 1.3, which is 40% higher than that of sample with a 1 µm grain size.1,2,3 The ZT enhancement results from a significant reduction of the thermal conductivity in the nanostructured samples, mainly due to the increased phonon scattering at numerous interfaces. Besides, in a wide range of temperatures between 600 ° and 1000 °C, the ZT value is maintained above 1.0. This makes the Si-Ge-P material much more useful as high-performance thermoelectric material for power generation with a large temperature difference generated by such as solar radiation, radioisotope and waste heat. In this presentation, Kinetic Monte Carlo simulations, based on parameters obtained with density functional theory in the gradient generalized approximation and experimental data, is used to understsand the Si-Ge-P precipitation kinetic path. The simulation involves realistic diffusion mechanisms by vacancy jumps, and a point defect source which drives the vacancy concentration towards its equilibrium value, during isothermal and anisothermal heat treatments.4 Depending on alloy and thermal history conditions, Monte Carlo simulations predict different kinetic behavior, including segregation of phosphorus at grain boundary as well as competition between homogeneous and heterogeneous precipitation of Si-Ge-P precipitates.
4:45 PM - LL7.7
Molecular Dynamics Simulation of Nucleation Process.
Ramanarayan Hariharaputran 1 , Pavlo Rutkevych 1 , David Wu 1
1 Large Scale Complex Systems, Institute of High Performance Computing, Singapore Singapore
Show AbstractThe process of nucleation plays a significant role in many phase transformations; hence understanding the physics of nucleation is important for designing and processing advanced materials. In the study of nucleation, insights from experimental results are limited due to finite resolution of the observation tools, while theories such as classical nucleation theory (CNT) are handicapped by its assumptions.In this talk, we present our studies of nucleation using molecular dynamics (MD) simulation.Traditionally, equilibrium MD is used to simulate nucleation at very high driving forces due to limitation in computational power. We propose non-equilibrium modifications of MD methodology to address artifacts arising from high driving forces. We use this tool to study nucleation in unary systems and compare the results to the predictions of CNT.
5:00 PM - LL7.8
Stability and Work Function of TiCxN1-x Alloy Surfaces.
Hong Zhu 1 , Rampi Ramprasad 1
1 Department of Chemical, Materials and Biomolecular Engineering, University of Connecticut, Storrs, Connecticut, United States
Show AbstractTransition metal nitrides and carbides are being considered and optimized as replacements for the poly-Si gate electrodes in next-generation transistors composed of high dielectric constant, or “high-K”, dielectrics. It is expected that the currently widely used Si substrate/SiO2/poly-Si gate stack in most transistors will gradually be replaced by the Si substrate/high-K dielectric/metal stack. The choice of the metal electrode is guided by its effective work function when in contact with the dielectric, which is determined by the true (or vacuum) work function of the metal itself and the dipole moment at the metal-dielectric interface. The specific effective work functions for the metal gate to be useful in p-type and n-type transistors, respectively, are 4.1 eV and 5.2 eV, such that the metal Fermi level is aligned with the valence and conduction band edges, respectively, of the underlying Si substrate. In the past, such a “tunability” has been achieved by p-doping or n-doping the poly-Si electrode. Migration to a suitable metal electrode that can display both values of the work function thus requires a control over the vacuum work function (e.g., through composition modulation) as well as a control over the metal-dielectric interface chemistry.In this work, we have attempted to arrive at a comprehensive understanding of the vacuum work function of TiCxN1-x alloys through ab initio density functional theory (DFT) simulations combined with statistical thermodynamics (the latter introduced to determine the most probable surface terminations as a function of alloy composition). We have modeled the TiCxN1-x alloy as a uniform solid solution between TiC and TiN. The stability of the solid solution has been assessed, the most stable surface for each composition has been identified, and finally, the work function values for all surfaces considered as a function of composition have been determined. Trends in the work function variation with surface type are explained in terms of surface dipole moments. One of the interesting findings of this work is that the vacuum work function value for the most stable surfaces varies from 3.2 eV for pure TiN to 4.6 eV for pure TiC, accompanied by a smooth variation between these end-point values for intermediate alloy compositions. In summary, with the combination of the surface stability and work function results, the most probable work function for each stable alloy composition has been determined.
5:15 PM - LL7.9
Theory of Defect Distribution at Semiconductor Interfaces Based on Ab-initio Thermodynamics.
Christoph Freysoldt 1 , Joerg Neugebauer 1
1 Computational Materials Design, MPI für Eisenforschung, Düsseldorf Germany
Show AbstractPoint defects such as vacancies, interstitials, or impurities critically influence the electrical properties of semiconductor materials. In this context, ab initio atomistic calculations are increasingly used to complement experiment, providing e.g. microscopic formation energies and charge transition levels of point defects and their complexes. In order to go beyond the atomistic length scale, the ab initio data is used in conjunction with thermodynamical concepts to predict defect concentrations in continuum models, which can then be applied to realistic device structures.A problem usually treated with continuum models is band bending: At interfaces, space charge zones develop that align the Fermi levels of the two adjacent materials. If simplifying approximations fail, the band bending problem must be solved numerically. The presence of mobile defects as well as defect reactions influence the screening of internal and external potentials. The relevant length scales (between a few nm and a few 10 μm) and concentrations may span many orders of magnitude. Kinetic simulations of defect diffusion and reactions are able to cope with this problem in principle, but the steady state is often reached only after long simulation times.We have therefore developed a new approach that allows to determine the thermodynamical equilibrium directly. For this, we vary the local chemical potentials of immobile elements with a gradient-based optimization scheme until the boundary conditions (total concentration of elements) are met. The formalism allows to treat inert and reactive defects as well as mobile and immobile species simultaneously. By coupling the equilibrium solver to a one-dimensional Poisson integration scheme with adaptive length step, the band bending problem can be treated very efficiently. We also discuss the possibility to employ the equilibrium concept for treating fast processes in kinetic simulations. We present results for the H distribution at GaN p-n-junctions as well as hydrogen accumulation at Mg3N2-based inversion domain boundaries.
5:30 PM - LL7.10
Predictive Point Defect Chemistry and Conductivity in Rutile TiO_2.
Xin Li 1 , Mike Finnis 3 , Simon Phillpot 2 , Susan Sinnott 2 , Beth Dickey 1
1 Materials Science and Engineering, Pennsylvania State University, University Park, Pennsylvania, United States, 3 Department of Materials, Imperial College London, London United Kingdom, 2 Materials Science and Engineering, University of Florida, Gainesville, Florida, United States
Show Abstract In ionic materials, the concentrations of charged lattice point defects and their concomitant compensation by electronic defects often govern the materials’ electronic conductivity. Since the point defect equilibria are functions of both temperature and oxygen activity in oxides, the charge carrier concentrations and conductivity are also strong functions of these external variables. While point defect equilibria in ionic materials are generally extracted via application of the mass-action law of thermodynamics to experimental conductivity data, there is a need to develop computational approaches to predict the defect equilibria a priori. In this study, we calculate point defect energetics in TiO_2 by the supercell method within the density functional theory (DFT) framework using the generalized gradient approximation (GGA) functional. Supercell convergence is carefully examined up to a supercell size of 3x3x5 TiO_2 unit cells, and the defect energies are corrected for the artificial electrostatic interactions by the electrostatic potential alignment method. Charge localization around the defects is quantitatively analyzed using Bader analysis. The calculated formation energy of a Frenkel pair is about 1 eV lower than the experimental value, which we conclude from our DFT-based phonon calculations is due to the temperature-dependant phonon energies, which are non-trivial at high temperatures. The DFT-calculated defect energies and temperature-dependant phonon energies are then combined with thermodynamic data to extrapolate the defect free energies as a function of oxygen partial pressure (PO_2), temperature (T) and Fermi level. With the application of electroneutrality we then calculate Brower diagrams for TiO_2, which give the concentrations of all possible defects and the Fermi Level as a function of temperature and PO_2. The results predict the experimentally observed n-type behavior of TiO_2 at high T and low PO_2, and p-type behavior at low T and high PO_2. In addition, the results predict the instability of the TiO_2 lattice at very high temperatures and low PO2, where Magneli phases are experimentally observed to form. Using experimental values for the electronic mobilities, the electronic conductivity is finally calculated and found to be in good agreement with experimental data for both pure and doped rutile TiO_2, predicting changes in the dominant type of point defects with T and PO_2.
5:45 PM - LL7.11
Kinetics of O Defect Diffusion in Amorphous HfO2 from First Principles.
Chunguang Tang 1 , Rampi Ramprasad 1
1 , University of Connecticut, Storrs, Connecticut, United States
Show AbstractDriven by a need for device miniaturization in the microelectronic industry, Hf-based high-permittivity materials, such as HfO2, have attracted attention due to their potential application as gate dielectrics. However, diffusion of point defects in HfO2 and their accumulation at interfaces is found to result in the creation of undesirable interfacial phases such as Hf silicides, silicates and silica between HfO2 and the Si substrate. Moreover, accumulation of charged point defects at the electrode-HfO2 interfaces causes device performance issues, such as threshold voltage shifts. Although experimentally deposited HfO2 films are amorphous, computational efforts that address these aspects mainly focus on monoclinic HfO2 due to the simplicity afforded by crystalline materials. Atomic-level diffusion studies of point defects in monoclinic HfO2 indicate that the O interstitials are far more mobile than O vacancies, while experiments indicate that O vacancies are more prevalent and dangerous.In an attempt to bridge this apparent gap between computations and experiments, we have undertaken migration barrier studies of neutral and charged O vacancies and interstitials in amorphous HfO2. All calculations were performed using first principles density functional theory, with the amorphous HfO2 model generated using the melt-and-quench technique [D. Ceresoli and D. Vanderbilt, Phys. Rev. B 74, 125108]. Long-range diffusion pathways composed of elementary site-to-site hops, spanning the entire amorphous supercell were defined, and activation barriers for a neutral or charged point defect traversing the entire supercell was calculated. Consistent with prior work, we found that the preferred charge states of O vacancies and interstitials are positive and negative, respectively, for a wide range of electronic chemical potentials. For O vacancies, we found that the 2+ charged state has more than a 1 eV lower migration barrier than the corresponding neutral state for each of the elementary steps considered. The local chemistry and charge distribution around the vacancy sites are found to be responsible for the large barrier drop. In the case of O interstitials, barrier reduction due to negative charging of the interstitial defect was found to be small, with the net effect being that O vacancies in the 2+ charge state are the most mobile species. Consideration of amorphous environments and charging of defects leads us to the conclusion that O vacancies are in fact the more dominant of the migrating species. This work thus underlines the importance of local chemistry in treatments of defect dynamics.
Symposium Organizers
Mark Asta University of California
Alex Umantsev Fayetteville State University
Joerg Neugebauer Max-Planck-Institut fuer Eisenforschung
LL8: Soft Matter; Alloy and Glass Design
Session Chairs
Thursday AM, December 03, 2009
Room 313 (Hynes)
9:30 AM - **LL8.1
Adaptive Resolution Simulations: Towards Open Systems Molecular Dynamics Simulations.
Kurt Kremer 1
1 , MPI for Polymer Research, Mainz Germany
Show AbstractThe relation between atomistic structure, architecture, molecular weight and material properties is a basic concern of modern soft matter science. This longstanding aim by now goes far beyond standard properties of bulk materials. A typical additional focus is on surface interface aspects or the relation between structure and function in nanoscopic molecular assemblies. This all implies a thorough understanding on many length and correspondingly time scales ranging from (sub)-atomic to macroscopic. Traditionally computer simulations have been separated in two main groups, namely simplified models to deal with generic or universal aspects, i.e. critical exponents, of polymers and those employing classical force field simulations with (almost) all atomistic detail, i.e. for the diffusion of small additives in small “sample”. To progress further adaptive schemes have to be developed, which allow for a free exchange of particles (atoms, molecules) between the different levels of resolution. First attempts towards this direction will be presented in this lecture. We study model systems, which display a spatially variable resolution with a free exchange of particles between the different regimes, ranging from atomistic resolution to coarse grained and continuum.
10:00 AM - LL8.2
Binding of Peptides (CR3-1, S2) to Platelets and their Intercalation in a Nano-clay Layer by a Coarse-grained Monte Carlo Simulation.
Ras Pandey 1 , Barry Farmer 2
1 Physics and Astronomy, University of Southern Mississippi, Hattiesburg, Mississippi, United States, 2 Materials and Manufacturing Directorate, Air Force Research Laboratory, Wright Patterson Air Force Base, Dayton, Ohio, United States
Show AbstractMonte Carlo simulations are performed to study binding of peptides (CR3-1: Trp-Pro-Ser-Ser-Tyr-Leu-Ser-Pro-Lle-Pro-Tyr-Ser and S2: His-Gly-Lle-Asn-Thr-Thr-Lys-Pro-Phe-Lys-ser-Val) to clay platelets on a cubic lattice. The Platelet executes its stochastic motion in a solvent matrix of mobile peptide chains. Bond-fluctuation description is used to model both clay platelet and the peptide chains. A platelet is described by a set of nodes tethered together in a flexible sheet. The peptides are represented by chains of tethered residue nodes, each with its own interaction characteristics in a specific sequence each appropriate for CR3-1 and S2. Although the atomistic details within the amino acid groups of peptides are ignored, their specificity is incorporated via an interaction matrix guided by atomistic simulations, X-ray crystallographic data, and hydrophobicity of amino acids. Residue–residue and residue-clay interactions are considered via interaction between each residue and the platelet nodes. Binding probability of CR3-1 and S2 to the clay platelet is examined by analyzing the mobility of each residue, their energy, and correlation profiling the proximity to the platelet. The plate is then replaced by a stack of sheets, a model for the layer of clay platelets. The intercalation of peptide chains in the clay galleries are under-investigation along with the analysis of exfoliation and dispersion of the platelets as a function of peptide concentration. Some of the results of our on-going computer simulations will be presented as data become available.
10:15 AM - LL8.3
A Dissipative Particle Dynamics Model of Biofilm Growth.
Zhijie Xu 1 , Meakin Paul 2
1 Energy Resource Recovery & Management, Idaho National Lab, Idaho Falls, Idaho, United States, 2 Center for Advanced Modeling and Simulation, Idaho National Laboratory , Idaho Falls, Idaho, United States
Show AbstractA dissipative particle dynamics model (DPD) that has been developed to quantitatively simulate the structural development of biofilm controlled by advective and diffusive substrate (nutrient) transport, growth/decay kinetics and hydrodynamic interactions with flowing liquid will be described. In the DPD model, three different types of particle were used to represent the biofilm, the liquid and the substrate, and the walls on which the biofilm grows were represented by a fourth type of particle. An important advantage of particle models, including DPD, is that topological changes such as biofilm fragmentation and reattachment of biofilm fragments can be reliable simulated without complex algorithms or additional computational overhead. Results obtained from simulations that were performed to investigate the effects of biofilm rheomechanical properties and flow conditions on biofilm morphology and growth will be presented.
11:30 AM - LL8.5
Prediction of the Stability of the MAX Phases from First Principles.
Vicki Keast 1 , Stuart Harris 1 , Daniel Smith 1
1 School of Mathematical and Physical Sciences, The University of Newcastle, Callaghan, New South Wales, Australia
Show AbstractThe Mn+1AXn phases (where M is a transition metal, A is a group A element, X is carbon or nitrogen and n = 1,2,3…) have seen a recent increase in attention due to the development of new methods for phase-pure fabrication. Their unique combination of metallic and ceramic properties makes them of interest for a variety of high-temperature applications or in other extreme environments. One of the unusual features of this set of alloys is that only certain values of n have been reported and there does not appear to be any systematic behaviour. For example, for the Ti-Si-C system we have n = 2, 3, for Ti-Al-N we have n = 1, 3 and for Ti-Al-C we have n = 1, 2. The reason why the insertion or removal of a transition metal carbide or nitride layer should or should not be preferred is not immediately apparent.Density functional theory (DFT) has been used extensively to study the Mn+1AXn phases, but the question of which of the possible phases can be expected to occur has been largely unexplored. The formation energy (the total energy minus the energy of the constituent elements) is usually found to be negative, favouring phase formation. However, this information alone is insufficient to predict the actual occurrence of the compounds. In order to answer this question, the calculated total energy must be compared to that of the competing phases, as determined from the ternary phase diagrams. In this work, DFT was used to predict the stability of the different phases by comparing their total energy to that the appropriate competing phases. Five systems (Ti-Al-C, Ti-Si-C, Ti-Al-N, Ti-Si-N, Cr-Al-C) have been studied for n=1-4. Both the α and β structures were considered. Structural optimisations were performed for all structures, including lattice parameters and internal coordinates (where applicable). In all cases the structural parameters are in agreement with experimental data or previous calculations, within the expected limits of accuracy of DFT. The calculations were very successful in predicting the observed occurrences of the Mn+1AXn phases for the 20 phases studied. The energy of the Mn+1AXn phases was found to be lower than the competing phases when the occurrence of that phase has been reported and higher when it has not. The results suggest that the dominant contribution to the phase stability is the total electronic energy. The single exception to the success was the prediction that Ti4SiC3 would not occur (albeit with a very small energy difference). However, as this alloy has only been observed by thin film fabrication, it is likely it is a metastable phase. On the basis of our calculations, the phases that also have very small energy differences and may warrant further efforts at thin-film fabrication are Ti2SiC, Ti2SiN and Ti3AlN2. None of the M5AX4 phases were predicted to occur and in all cases the α-phases were found to be more energetically favourable than the β-phases.
11:45 AM - LL8.6
Theory-guided Design of Bone-matched Ti-based Multi-phase Biomaterials.
Martin Friak 1 , William Counts 1 , Duancheng Ma 1 , Benedikt Sander 1 , Dierk Raabe 1 , Joerg Neugebauer 1
1 Computational Materials Design, Max Planck Institute for Iron Research, Duesseldorf Germany
Show AbstractWe present a new multi-disciplinary strategy for the theory-guided multi-scale design of Ti-based two-phase materials with bio-application-tailored properties. The approach is based on the combination of (i) ab-initio calculations of thermodynamical and elastic properties as a theoretical guidance in selecting the optimum chemical composition and (ii) experimental verification. Our study has been focused on Ti-biomaterials as the improvement of hip transplants is severely hampered by the lack of suitable materials which are biocompatible in terms of non-toxicity and mechanical properties matched to the bone. The aim of our research has been therefore to identify metallurgical trends for non-poisonous Ti-based alloys employing quantum-mechanical calculations. Specifically, density functional theory (DFT), a plane wave basis set and PAW pseudopotentials have been used. As a first step the thermodynamic stability of selected Ti-binaries has been determined employing special quasirandom structures. Second, the Young modulus of the thermodynamically stable alloys has been calculated and an alloy composition that maximally matches human bone has been selected. Guided by the theoretical calculations of phase stability and elastic properties, selected binaries were actually melted, cast, and heat treated to a homogeneous state. The samples have been experimentally characterized by X-ray methods, electron microscopy including crystallographic (EBSD) and chemical (EDX) analysis, and mechanically tested using ultrasound measurements. The experimental data obtained in these experiments are in excellent agreement with the theoretical predictions.
12:00 PM - LL8.7
A Combined Ab Initio/CALPHAD Modeling of Phase Diagrams in Systems with Complex Phases.
Mojmir Sob 1 2 , Jana Pavlu 1 2 , Jan Vrestal 1 , Ales Kroupa 2
1 Dept. of Chemistry, Masaryk University, Brno Czechia, 2 , Institute of Physics of Materials, Academy of Sciences of the Czech Republic, Brno Czechia
Show AbstractWe address the application of ab initio electronic structure calculations to modelling of phase equilibria and to construction of phase diagrams in binary and ternary iron systems containing complex intermetallic phases (Laves phases, sigma phase) in combination with the phenomenological CALPHAD (CALculations of PHAse Diagrams) method. This approach is briefly outlined and utilization of ab initio total energy differences (lattice stabilities) for sigma phase is shown in Fe-Cr, Fe-Cr-Ni and Fe-Cr-Mo systems. Our description of sigma phase is included into a thermodynamic database for calculation of phase equilibria in steels, developed earlier by some of the authors. Theoretical results obtained by new and older models are compared with new experimental data for superaustenitic (high Ni and Cr) steels. In conclusion, some problems and challenges in application of ab initio results in the CALPHAD approach are outlined.
12:15 PM - LL8.8
Designing Glasses with Tunable Structure and Properties by Computer Simulation.
Liping Huang 1 , Fenglin Yuan 1 , Qing Zhao 1
1 Materials Science and Engineering, Rensselaer Polytechnic Institute, Troy, New York, United States
Show AbstractA normal solid becomes stiffer when squeezed and softer when heated. In contrast, silica glass behaves the opposite way: its elastic moduli decrease upon compression and increase upon heating. Silica glass is also known to densify under compression and radiations. These have been long-standing mysteries in materials science. Using molecular dynamics simulation, we uncovered the structural origins of the anomalous thermo-mechanical behaviors and mechanisms of permanent densification in silica glass. Accordingly, these anomalies can be attributed to localized structural transitions, analogous to those that occur in the crystalline counterparts. The irreversible densification in silica glass is achieved through structural transition involving bond breaking and re-formation under a combination of high pressure and temperature. We further revealed that the anomalous thermo-mechanical behaviors are inherently connected to the ability of the glass to undergo permanent densification. Our computer simulations demonstrate that by processing in ways that gradually eliminates anomalous thermo-mechanical behaviors, degree of the glass to undergo densification can be eventually eradicated. This provides the conceptual foundation for the bottom-up design of new glasses with tunable structure and properties.
12:30 PM - LL8.9
First-principles Computational Thermodynamics of Light-weight Additions to Ti.
Giancarlo Trimarchi 1 , Donald Shih 3 , Dongwon Shin 2 , Chirs Wolverton 2 , Arthur Freeman 1 2
1 Dept. of Physics, Northwestern University, Evanston, Illinois, United States, 3 Boeing Research and Technology, The Boeing Company, St. Louis, Missouri, United States, 2 Dept. of Materials Science and Engineering, Northwestern University, Evanston, Illinois, United States
Show AbstractTitanium alloys play an important role in aerospace structures, mainly due to their high specific strength, specific modulus, and excellent resistance to corrosion and oxidation. These benefits can be further enhanced by designing and developing much lower density Ti alloys, while maintaining acceptable mechanical performance. As a key part of multiscale modeling and simulation in this lighter Ti alloy design effort, density functional theory based first principles calculations and phase diagram calculations have been used. Among the elements to consider in the design of new light Ti alloys, Mg represents an appealing choice being isostructural to the hexagonal close-packed α-Ti phase. However, Mg has a limited solubility in α-Ti, e.g., ~1.5 at % at 600-700 °C, and nearly zero at room temperature. A detailed ab-initio study combined with thermodynamical modeling schemes has been conducted to characterize the solubility of Mg in α-Ti. In addition, as a way to further increase the solubility of Mg in α-Ti, and to also achieve a larger weight reduction, first principles calculations on selected ternary Ti-Mg-X systems, where X=Al and Si, were performed. The microstructure of Ti alloys used in aerospace usually consists of α-Ti and body-centered cubic β-Ti phases in various morphologies and volume fractions, thus making weight reduction in both such phases necessary. Therefore, in order to devise comprehensive design rules for light-weight Ti alloys, ab-initio modeling of the solubility of light elements in β-Ti phase and phase diagram calculations will also carried out.Supported by The Boeing Company
12:45 PM - LL8.10
Icosahedral Self-aggregation and Slow Dynamics: A Mechanism for Metallic Glass Formation.
Shaogang Hao 1 , Cai-Zhuang Wang 1 , Maozhi Li 1 , Kai-Ming Ho 1
1 , Iowa State University, Ames, Iowa, United States
Show AbstractUnderstanding of the slow dynamics and glass transition in metallic alloy systems has been an important and longstanding challenge in material science. From ab initio molecular dynamics simulations, we found that icosahedral short-range order (ISRO) have significantly strong tendency to be self-aggregated in a number of different high temperature metallic liquids of good glass-forming alloys. As ISRO can be linked to slow dynamics in Zr1-xCux system, we found that the slowing down effect is much more severe in aggregated icosahedral clusters. Coupled with the tendency of icosahedral clusters to aggregate into string-like networks, we developed a lattice model to study the dynamics and motif of the system on a larger scale. We observed non-exponential dynamics behavior. Our work sheds light on the microscopic nature of dynamical heterogeneity and glass transition in metallic alloys, and provides a fundamental for the facilitated lattice model.
LL9: Electrochemistry and Semiconductors
Session Chairs
Thursday PM, December 03, 2009
Room 313 (Hynes)
2:30 PM - **LL9.1
High-Throughput Computing for Constrained Property Optimization.
Gerbrand Ceder 1
1 Materials Science and Engineering, MIT, Cambridge, Massachusetts, United States
Show AbstractDesigning materials that can be used in practical technology often requires the simultaneous optimization of multiple properties under the constraint of others. Such a constrained optimization problem can only be executed computationally if ab initio methods are developed for most of the critical properties. I will show that for many materials applications in energy, this is the case. By implementing ab initio methods in a high-throughput computing environment and coupling it to databases of all existing materials, we have been able to rapidly search for new materials for several technologies, including lithium batteries, mercury absorption materials, radiation detectors and alkaline batteries. I will show that in several of these domains it is possible to search rather exhaustively across the domain of inorganic materials. Data mining and statistical knowledge methods to extra simple heuristic models have played a key role in these design problems to come up with better materials. We find that in practice, the constraints one is designing under (stability, cost, lack of reactivity with environment, etc.) are more limiting than the optimization of the key engineering property. One of the key problems, in large-scale computational materials design, is the prediction of the structure of novel compounds. I will show how that can be solved with simple statistical knowledge methods which learn from the large amount of data already available.
3:00 PM - LL9.2
Computing Electrochemical Impedance of Solid Electrolyte from Fluctuations.
Eunseok Lee 1 , Fritz Prinz 1 , Wei Cai 1
1 Mechanical Engineering, Stanford University, Stanford, California, United States
Show AbstractWe introduce a new method to compute the electrochemical impedance of solid electrolyte from kinetic Monte Carlo simulation using the fluctuation-dissipation theorem. This method is much more efficient than the conventional computational method, because the impedance at all frequencies can be calculated from only one simulation. Using this efficient method, we could systemically study the effect of doping concentration and cation dopant distribution to the ionic conductivity of yttria-stabilized zirconia. Cluster expansion method (CEM) is used to improve the accuracy of the energy barrier model for vacancy diffusion. The ground state energy is computed for several ionic configurations by density functional theory, and the energy is fitted by many-body ionic clusters. Distinguished from previous attempts of CEM, we focus on the vacancy-oriented ionic clusters to emphasize the effect of vacancy and use larger supercells to reduce the interference with images of ions. The change of impedance is discussed when the effective cluster interactions are added to the energy barrier model of vacancy diffusion.
3:15 PM - LL9.3
Simulating Diamond Cubic Anisotropy Using Extended Cahn-Hilliard Model.
Solmaz Torabi 1 , Zhengzheng Hu 2 , Steven Wise 3 , John Lowengrub 2
1 Material Science and Eng., UCI, Irvine, California, United States, 2 Mathematics, UCI, Irvine, California, United States, 3 mathematics, University of Tennessee, knoxville, Tennessee, United States
Show AbstractSelf-assembly semiconductor nanostructures such as quantum- dots are a promising inexpensive and effective approach to manufacture novel nanoscale electronic devices. Producing such quantum-dot-based devices, however, is still challenging. We study the influence of surface, strain energies on heteroepitaxial thin film growth.We are studying an alternate way of simulating anisotropy for the surface energy in the crystallographic systems. In most of the literature, surface energy anisotropy is modeled by some function of the normal angle. Even though, this can produce some desired results but has some limitations. For example, for different crystals, one needs to guess the correct function that produces the right surface energy and consequently correct facets at the equilibrium. This can be a difficult task. Due to this kind of limitations, most of the researchers have only studied the simple cubic crystals in their models and have not advanced to other crystallographic and more complex systems. One natural way of presenting anisotropy in the interfacial energy is by using the higher order terms in the Taylor expansion of the free energy. It has been shown (Cahn & Hilliard 1958) that to the second order, the system will only have isotropic properties. We claim that we can produce all the crystallographic systems by adding higher order terms to the energy. For instance, only adding the 4th order terms can produce all the cubic systems. This extra order will result to three distinguished 4th rank terms with 4th rank tensor factors. Considering the cubic symmetry, these tensors each has three independent elements. Some terms will be canceled, and at the end we get 5 different parameters that each is affecting the anisotropy. By choosing the right parameters we can model the preferred interfacial energy for our desired system, which in our case is the SiGe/Si thin film, with diamond cubic crystallography. This type of model has previously studied (Abinandanan & Heider 2001) but to our knowledge no one has ever studied this method with including all the terms in the systems, both linear and nonlinear. For hexagonal system, in addition, one needs to add the 6th order terms from the expansion.In our diffuse interface model, we use Cahn-Hilliard type equation for the growth and coarsening of the system. One advantage of this new anisotropic energy is that it has the intrinsic regularization. In other models, strong anisotropy will cause sharp corners to form and the equations become ill-posed. In our model, even the strong anisotropic Cahn-Hilliard equations are always well-posed. We present numerical results using an adaptive, nonlinear multigrid finite-difference method. Finally, in order to model the misfit and displacement strains, we add the elastic energy to our diffuse model. We can predict different Qdot shapes, such as pyramids and domes, based on the strength of the elastic interactions.
3:30 PM - LL9.4
Multiscale Properties of Galvanostatic Electrochemical Deposition.
Matteo Nicoli 1 , Mathis Plapp 2 , Mario Castro 3 , Rodolfo Cuerno 1
1 Matemáticas, Universidad Carlos III de Madrid, Leganés, Madrid, Spain, 2 Laboratoire de Physique de la Matière Condensée , Ecole Polytechnique, Palaiseau, Paris, France, 3 Escuela Técnica Superior de Ingeniería, Universidad Pontificia Comillas, Madrid, Madrid, Spain
Show AbstractSurface roughening is a problem attracting a huge interest from different fields of science. From the viewpoint of materials science, great efforts have been devoted to characterize and control the growth of surfaces, motivated by the many technological applications. On the fundamental side, many works coming from the statistical mechanics community have addressed the existence of universality classes for surface growth and their implications for the interpretation of experimental findings [1]. The apparent divorce between both approaches may be due to the fact that universality classes are expected to apply only asymptotically (long times and large scales) which are often difficult to attain experimentally. The richer variety of morphologies and transient behaviors observed experimentally are difficult to address in this theoretical framework, thus requiring a numerical approach to the problem. The so-called phase-field models are an ideal framework to analyze the problem in depth.In this work, we propose a mixed strategy to address the morphological and scaling properties of aggregates produced by electrochemical deposition (ECD). We analyze in depth the linear stage of surface growth in terms of physically motivated parameters and study, with the help of a phase-field formulation of the problem, the role of instabilities and fluctuations in the pattern-formation process. Recently, we have introduced a phase-field model with fluctuations in order to integrate numerically a moving boundary model including all the most relevant effects that occur in the bulk and at the surface of a galvanostatic ECD cell [2-3]. In order to achieve an accurate quantitative comparison between the model and the ECD experiments we have used a multi-grid algorithm for the resolution of the different length scales present in the system. This algorithm allow us to reach a system size for which we can observe many growing structures at the surface, giving us the possibility to measure the roughness properties of the effective interface created from the interaction of these structures. Through this methodology we can infer the scaling and the universal properties of several ECD experiments that can be found in the literature.[1] R. Cuerno and L. Vázquez, in Advances in Condensed Matter and Statistical Physics edited by E. Korutcheva and R. Cuerno (Nova Science, New York, 2004). [2] M. Nicoli, M. Castro, and R. Cuerno, Phys. Rev. E 78, 21601 (2008).[3] M. Nicoli, M. Castro, and R. Cuerno, J. Stat. Mech. P02036 (2009).
4:30 PM - LL9.6
Numerical Optimization of Light Emitting Diodes for High Efficiency Operation.
Oskari Heikkila 1 , Jukka Tulkki 1
1 Biomedical Engineering and Computational Science, Helsinki University of Technology, Espoo Finland
Show AbstractRecent development has drawn interest to light emitting diodes (LEDs) as compact and efficient solid state light sources for applications ranging from general illumination to telecommunications. The current research and development of LEDs mainly focuses on increasing optical power obtained per unit area and unit cost. However, more interest should be paid on the ultimate limit of LED efficiency, which in terms of wall plug efficiency may in principle exceed unity. We investigate how efficient the LEDs can ultimately be, how the efficiency of LEDs is determined and how much of the efficiency is sacrificed when high optical intensities are sought for. These limits are studied by using extensive numerical simulations.The operation of LEDs is simulated using a multiphysics model that accounts for the macroscopic carrier transport, photon extraction and heat exchange. The junction temperature is solved self consistently along with the carrier transport equations to include the effect of temperature on various material parameters. All relevant radiative and nonradiative recombination mechanisms in III-V semiconductors are taken into account in the calculations. The model provides a framework which can be used to simulate general semiconductor structures with material properties obtained equally well from theoretical calculations or from measurements.The model is currently employed to simulate simple bulk AlGaAs-GaAs double heterostructure LEDs which are expected to be currently capable of the most efficient operation. The model is used to calculate e.g. the external quantum efficiency (EQE) and the wall plug efficiency to find the optimal operating range. Also comparative calculations are made using simplified analytical models. It will be shown that the optimal operating point is not necessarily found at the maximum of EQE because of the contribution of lattice heat to the photon energy. Also the plausibility of electroluminescent cooling in realistic LED structures is discussed. The efficiency droop limiting the high power operation is analyzed to find the underlying mechanisms in double heterojunction structures. The model is also used to predict a limiting current at which an insufficiently cooled LED becomes thermally unstable. Simple measures to increase the efficiency, reduce the power consumption and lower the junction temperature of LEDs are introduced based on the results, and experimental setups for measuring various material and device parameters are proposed.
4:45 PM - LL9.7
Where are Nature's Missing Ternary Oxides? Combining Machine Learning with ab initio Computations to Accelerate New Compounds Discovery.
Geoffroy Hautier 1 , Christopher Fischer 1 , Anubhav Jain 1 , Gerbrand Ceder 1
1 Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
Show AbstractThe discovery of new compounds is an essential step in achieving breakthroughs in technologically important materials science fields (e.g. energy storage, catalysis, or radiation detection). This discovery process can be however very slow, relying on a combination of chemical intuition and serendipity. In this talk, we will present how a crystal structure prediction technique based on a combination of machine learning and ab initio computations can dramatically accelerate the search for new materials, predicting hundreds of new compounds with a relatively limited computational budget.While computational crystal structure prediction has made dramatic progress recently, most of the techniques available are not very well suited to a broad large-scale search for new compounds. These methods indeed assume a previous knowledge of the compositions of interest and require significant computational resources. We will show in this talk how an approach proposed by Fischer et al. can be used to find new compounds in thousands of systems in a matter of a few months. This crystal structure prediction technique is based on the correlations present in nature between crystal structures at different compositions (for instance the crystal structures present in a binary and a ternary composition). Using an experimental crystal structure database, a probabilistic model can be built taking mathematically into account all of these correlations. This model can then identify the likely crystal structures for a given composition, as well the compositions for which new crystal structures are likely to be found. Finally, these suggested structures can be computed via ab initio methods to test if they are indeed thermodynamically stable with respect to known compounds.As a test case of our approach, we will present results on ternary oxides. After building a probabilistic model using the Inorganic Crystal Structure Database (ICSD), we identified and tested with ab initio methods the most likely “unknown” ternary oxides against an extensive database of thousands of ab initio calculations on known structures in the ICSD. This large scale search identified around 300 new compounds using only a limited computational budget. We will present some of these new compounds, compare to experimental data, and analyze the chemical spaces in which most new compounds tend to be found.
5:00 PM - LL9.8
Novel Mechanism of Reversible Photoinduced Magnetism in Prussian Blue Analogs.
Mukul Kabir 1 , Krystyn J. Van Vliet 1
1 Department of Materials Science and Technology, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States
Show AbstractPrussian blue (PB) crystals are metallorganic materials that exhibit several modes of inducible changes in oxidation state, and thus changes in physical and functional properties. Reversible switching of magnetic state by external stimuli has been demonstrated experimentally for PB and PB analogs, with many potential applications including information storage and processing. For example, at low temperature, illumination of visible light (500-700 nm) induces bulk magnetization in KCo[Fe(CN)6]:5H2O, which can be eliminated by near-IR light (~ 1319 nm).[1] However, the molecular-scale mechanisms by which such magnetic state switching occurs have not yet been properly understood, and are required to facilitate design of this and other such magnetic switch-inducible materials. Here we report our simulations and proposed switching mechanism for this PB analog crystal. We apply density functional theory, corrected for on-site Coulomb interaction (DFT+U), to describe the structure and energetic ground state of anhydrate KCo[Fe(CN)6]. This ground state has large octahedral ligand-field splitting, and consequently exhibits a low-spin structure. We find that the low spin to high-spin transition is a two step process involving charge transfer between the metal atoms via the carbon-nitrogen ligand, followed by electron transfer within the cobalt atom. This spin crossover results in an approximately 10% increase in carbon-nitrogen bond length, and proceeds via a strong Jahn-Teller active metastable intermediate-spin state. We find that the magnetization and demagnetization requires input energy of 1.15 and 0.9 eV, respectively, which is in good agreement with experimental observations. We also demonstrate that this magnetization can be tuned and even increased by expansion of the Fe-C octahedra in this PB analog crystal, through means such as applied hydrostatic strain. References:[1] Sato, T. Iyoda, A. Fujishima, and K. Hashimoto, Science 272, 704 (1996).
5:15 PM - LL9.9
Statistical Phonon Transport Model of Thermal Transport in Silicon.
Thomas Brown 1 , Edward Hensel 2
1 Microsystems Engineering, Rochester Institute of Technology, Rochester, New York, United States, 2 Mechanical Engineering, Rochester Institute of Technology, Rochester, New York, United States
Show AbstractThermal transport in crystalline materials at various length scales can be modeled by the Boltzmann transport equation (BTE). However, difficulties associated with solving the BTE for non-simplified cases have limited its use. A statistical phonon transport model is presented that solves the BTE in a statistical framework that incorporates a unique state-based phonon transport methodology. Anisotropy of the first Brillouin zone is captured by utilizing directionally-dependent dispersion curves obtained from lattice dynamics calculations. Elastic and inelastic scattering processes are accounted for with a rigorous implementation of phonon energy and pseudo-momentum conservation. Thermal transport results of silicon from the diffuse to ballistic regimes are compared with experimental data from literature.