When cooled below the miscibility temperature,alloys decompose into precipitates rich in one of the two components.These precipitates are observed in TEM to have remarkable shapes and sizes ,and are considered important for material strength ( in metal alloys) and electron transport (in semiconductor alloys). Since a precipitate system contains N= 10**3 to 10**6 atoms,in the past,simple energy models( e.g,elastic springs plus an isotropic chemical energy) have been used to describe their evolution.. We are interested in applying LDA-quality energy functionals to this problem. .Since the system of precipitate +medium is structurally coherent,all that is needed is the knowledge of the energy of N atoms (A and B) as a function of their lattice configuration ,so that the energy of any shape and size of the precipitate can be calculated.However,one must know the energy corresponding to the (coherently) relaxed system.We show that this can be conveniently done via the "mixed Basis Cluster Expansion" containing pair as well as multi-body and strain terms deduced from LDA calculations of simple prototype A-B systems.This functional is then subjected to a Monte Carlo simulation of typically 10**5 atoms. I will show how this reproduces remarkably the shape VS size VS temperature behavior noted experimentally for Al-Zn alloys ,and how the physical accuracy of the underlying LDA hamiltonian permits a transparent analysis of the factors that control precipitate shapes.Thus,this method allows one to extend the atomistic LDA size scale to mesoscopic dimensions without compromizing the quality of the energy expression.
This work was done in collabiration with Stefan Muller,
Li.W.Wang and C.Wolverton and was supported by DOE-OS-DMS.
[Q14.002] An approach to linking lengthscales in dynamical magnetic simulations
V. V. Dobrovitski (Ames Laboratory, Iowa State University), M. I. Katsnelson (Institute of Metal Physics RAS), B. N. Harmon (Ames Laboratory, Iowa State University)
Multiscale phenomena which include several processes occuring simultaneously at different length scales and exchanging energy with each other, are widespread in magnetism (e.g., the nucleation of magnetization reversal at a defect). These processes, besides being of fundamenal significance, are important for applications, e.g., they govern the reversal of information bits in contemporary and prospective magnetic memory devices. For accurate computer simulations of the multiscale phenomena the relevant length scales should be coupled, i.e. modeled simultaneously and seamlessly, with the possibility of energy transfer between them.
In this work we present an approach to lengthscale coupling in the dynamic modeling of magnets. It allows a smooth transition from atomic to microscopic lengthscales. Tests results will be presented showing that the approach proposed can be applicable for realistic modeling of magnetic materials. Also, we will present the results of simulations performed on several model 1-D systems, where an atomic-scale variation of anisotropy and exchange is present in some regions, representing defects in real systems. Directions for future development (finite temperature and dissipation effects) will be discussed.
Research at the Ames Laboratory is supported by the Director
of the Office of Science, Basic Energy Sciences of the U.S.
Department of Energy.
[Q14.003] Towards a unified description of magnets on different length scales
Vladimir Antropov, Kirill Belashchenko (Ames Lab,Ames,IA), Mark Novotny (SCRI,FSU, Tellahassee,FL)
We discuss the problem of bridging different scales in
material science theory. Using the problem of domain wall
dynamics in a polytwinned magnet as an example we show how
beginning from first-principle calculations for the
microscopical structural elements with the characteristic
size of up to several nanometers one may proceed to the
description of properties determined by microstructural
features on the mesoscopical scale (typically of the order
of 100 nm). Using CoPt as a generic example we demonstrate
how the complicated hierarchy of interactions works in this
case and the way the corresponding parameters of such
interactions can be calculated. Uncertainties in definition
of such parameters can be a source of some concern in such
simulations spanning different length scales.
[Q14.004] Self Consistent Non-Linear I-V Characteristics of Au atomic Wires
H. Mehrez, Jeremy Taylor, Brian Larade, Hong Guo (Physics Department, McGill University, Montreal,QP H3A 2T8, Canada)
We carried out an ab initio self consistent
calculation for the current flowing through a small chain of
Au atoms sandwiched between two semi-infinite Au(100)
electrodes, as a function of bias voltage. Our theory is
based on density functional theory within LDA approximation
using fireball atomic basis set, with standard norm
conserving pseudopotential to define the atomic core, and
with nonequilibrium Green's function to construct charge
density under external bias for open device system.
This formalism is the core of our simulation package McDCAL
(McGill Device CALculator software). Our results are
compared with recent experimental data of I-V
characteristics of Au mechanical break junctions. Our
results suggest that due to a negligible charge transfer
from the electrodes to the atoms in the chain, the nonlinear
I-V curve is dominated by effects of the band structure of
the electrodes and the eigenstates of the isolated Au chain
(molecule).
[Q14.005] Multiscale Modeling of Plasticity in BCC Tantalum: Bridging Atomistic and Mesoscale Dislocation Simulations
Lin Yang, Meijie Tang, John Moriarty (Lawrence Livermore National Laboratory)
Plastic deformation in bcc metals at the low-temperature and high-strain-rate is controlled by the motion of a/2<111> screw dislocations. Understanding the fundamental process of screw dislocation motion in bcc metals has become important both to advance modern dislocation theory and to develop predictive multiscale simulations of crystal plasticity in these materials. The multiscale modeling approach presented here for bcc Ta is based on information passing, where results of simulations at the atomistic scale are used in simulations of plastic deformation at micron length scales via dislocation dynamics. The relevant single-dislocation processes for a/2<111> screw dislocations have been simulated by means of volume-dependent multi-ion interatomic potentials derived from model generalized pseudopotential theory (MGPT) and an accurate ab-initio data base, using a robust Green's Function technique that implements flexible boundary conditions.(L.H. Yang, P. Söderlind, and J.A. Moriarty, Philos. Mag. A (2001, in press).) The pressure- and stress-dependent results on the fundamental kink-pair energetics and orientation-dependent Peierls stress have then been used as input for meso-scale dislocation-dynamics simulations. The latter simulations are based on a screw-edge model in a discretized lattice, and screw dislocation mobilities represent critical input into the study of both non-Schmid effects (anisotropic plasticity) and high-pressure behavior.
[Q14.006] Real-space calculation of the effects of oxygen and boron impurities on cohesion in nickel aluminides
D. Djajaputra, B.R. Cooper (Dept. of Physics, West Virginia University, Morgantown, WV 26506-6315)
Understanding the physical factors which determine the
strength of intermetallic alloys is an important problem for
technology as well as for the science involved. In nickel
aluminides some impurity atoms, e.g. boron, act as a
cohesion enhancer which can improve the cohesion
substantially; while some other atoms, e.g. oxygen, can
destroy the cohesion, even when present in minute
concentration. It is thus important to study the interaction
between a single atomic impurity and its local intermetallic
environment. We have studied this problem by using a
combination of ab-initio and real-space (tight-binding)
methods. We use a full-potential linear muffin-tin orbital
(FP-LMTO) method to obtain an accurate set of tight-binding
parameters which are then used as input parameters for a
real-space computation using Haydock's local Green's
function (recursion) method (Solid State Physics 35,
216--294, 1980). This combined method allows us to escape
the limitation of the ab-initio supercell method and focus
on the interaction of the single impurity atom with its
nickel-aluminide environment. We will present the results of
our calculation using this method for oxygen and boron
impurities in NiAl and Ni_3Al.
[Q14.007] Ab-initio based modeling of thermal expansion of iron-aluminides.
Tatiana Seletskaia, Leonid Muratov, Bernard R. Cooper (West Virginia University)
The iron-aluminides Fe3Al and FeAl have been among the most
widely studied intermetallics because of their low cost, low
density, good corrosion resistance and exceptional strength
retention at high temperature. We present calculations of
thermal expansion for this alloys and effects of additives
such as molybdenum, chromium and niobium on thermal
expansion over the wide range of temperatures. For
temperature comparable with Debay temperature we used a
combination of quasi-harmonic approximation and
semi-empirical tight binding and/or ab-initio pseudo
potential calculations. While for higher temperature,
molecular dynamics calculations based on interatomic
potentials and semi-empirical tight binding Monte Carlo
simulations were used. Results of these different methods
were compared between themselves and to experimental values.
The parameters of the tight-binding Hamiltonian and
interatomic potentials were fitted to reproduce results
(total energy and band structures) of full-potential LMTO
calculations for several distorted lattices.
[Q14.008] Multiscale Modeling of Polymer-Clay Composites
Tibor F. Nagy, P. M. Duxbury (Michigan State University)
Nanosized clays, which consist of platelets with thickness 1 nm and width of 100-200 nm, can now be dispersed in certain polymers, for example nylon-6 or polypropylene. This type of composite material may exhibit significantly improved properties compared to that of the pure polymers. The mechanisms for these property improvements are not well understood and provide a challenge to the modeling community. We describe atomic and mesoscale lattice based methods designed to model these materials beginning at the atomic level and continuing to the macroscopic level. We use these methods to study diffusion, the elastic modulii and interfacial energies relevant to the processing and toughness of these materials.
[Q14.009] Multiscale Modeling of Plastic Deformation of Glassy Polymers
Rahmi Ozisik (Institute of Polymers, Swiss Federal Institute of Technology (ETH)), Kenneth J. Beers (Department of Chemical Engineering, Massachusetts Institute of Technology), Ulrich W. Suter (Institute of Polymers, Swiss Federal Institute of Technology (ETH))
The study of elastic and plastic deformations and the underlying changes in the atomistic scale associated with these deformations is very challenging because of the multitude of length and time scales involved. The research in multiscale modeling is focused upon the area of plastic deformation, a subject of great importance in the areas of impact resistance, dimensional and thermal stability, and physical aging. In the study of plastic deformation, molecular simulation by itself is insufficient because the region undergoing the deformation transforms into a different strain state than the surrounding material. The development of a method that couples the surrounding continuum with an atomistic inclusion has been the recent focus of our group. This study extends a previous method for static zero Kelvin simulations in which an inclusion is embedded in a finite element mesh such that the nodes on the surface of the inclusion move in concert with the deformation state of an atomistic cell employing periodic boundary conditions. This coupling method is now being prepared to study the effect of thermal processes in the plastic deformation of glassy polymers.