The goal of this program is to translate on-going research in the mechanical behavior of materials, at length scales ranging from atomic to continuum, into unique learning opportunities and situations for undergraduate and graduate engineering students. The program consists of three major parts: 1) the development of modules on specific topics that can be widely disseminated and used in a broad spectrum of undergraduate classes; 2) the development of two courses at the senior and graduate level on connecting atomic and continuum length scales through numerical simulations; and 3) the development of interactive simulation programs and scientific visualization through which students may gain experience and physical knowledge in the topic area.

This report describes the initial progress made toward these three goals in the first 9 months of the program. In summary, the progress to date includes: the development of much of the graphics associated with the proposed modules; some work on text/description associated with the modules; a course proposal and outline for the senior-level course to be offered in the Fall, 1998 semester; some work on developing convenient interfaces for interactive computing and visualization using existing computer codes. Progress is reported on each goal in separate sections below.

1) Module Development

We have proposed to develop eight educational modules on topics ranging from atomic structure of dislocations and cracks to continuum-level simulations of deformation. Each module is similar to a mini-course on the topic, and requires graphics and text to lead students through the topic in a coherent fashion. Progress on each of the eight modules is discussed below, and the concept for a ninth module stemming from our course development is presented. Overall, we will use an HTML format for ease of creation and uniformity among the PIs.

Module 1: Interatomic potentials, crystal structures, lattice stability

For this module we have compiled data on interatomic potentials for various metals and intermetallic systems. We have concentrated on embedded atom method interatomic potentials and have started a database of these potentials that will be part of the module. We have also compiled a database of various crystal structures that we will include in the modules in VRML format. This format allows the user to explore the structure in three dimensions using a web browser.

Figure 1. Effective pair potentials for various pure metals

Module 2: Relaxation around point defects

We have calculated vacancy relaxation behavior in NiAl and Ni3Al as examples of point defect relaxation. We plan to add a variety of other pure metals and include interstitials in this module. The development of interatomic forces for interstitials is under way.

Module 3: Structure of grain boundaries

We have started work on this module with a series of symmetrical tile boundaries that demonstrate the properties of coincident site lattice boundaries. These simulations will be an important part of this module. We have also started work on a graphical method to study and easily display the effects of rigid body translations of one grain with respect to another in the grain boundary structure and energy.

Module 4: Dislocation structures and motion

For this module, we have performed simulations of dislocations in edge, screw and mixed orientations which will also be incorporated into VRML format. We have also worked on a technique to color-map the strain field in the dislocation core region. We have also started to develop animations of dislocation motion resulting from increasingly larger applied stress. These will be incorporated to explain how dislocation motion in crystals results in plastic deformation.

Figure 2. [100] screw dislocation in NiAl with isostrain color contours around dislocation core

Module 5: Interaction of a point defect and a dislocation.

We performed a series of simulations of dislocation core structure with a vacancy in different positions within the core region. These simulations illustrate the energetics of the interaction between the dislocation and the vacancy. The cases for the interaction of a dislocation and an interstitial are under way.

Module 6: Fracture in pure metals

This module will contain simulations of fracture in a variety of materials. We have already completed cases showing ductile fracture with extensive emission of mobile dislocations from the crack tip and cases that are perfectly brittle with no plastic deformation.

Figure 3. Crack propagation Aluminum

Module 7: Adiabatic shear bands at high strain rates

An adiabatic shear band is a narrow region of intense plastic deformation that forms during high strain-rate deformations of most metals, many polymers and granular materials. Their study is important since they usually precede fractures in ductile materials. In this educational module, our goal is to introduce the student to the concept of adiabatic shear bands, show pictures of experimentally observed shear bands, depict numerical simulations, and when possible, enable the student to run a computer program interactively to delineate the effect of various material parameters on the initiation and development of adiabatic shear bands. We have collected all of the literature and the experimental results to be put into the module, and are in the process of getting the pictures scanned and the accompanying text polished to be put on the web in HTML format. Computer simulations of the initiation, development and growth of adiabatic shear bands in the penetration of depleted uranium and tungsten rods into steel targets are available at the http://www.sv.vt.edu/research/batra-stevens/pent.html. Fortran codes to analyze shear bands in one and two dimensional problems are operational. They need to be optimized so that students can use them interactively and view the results on the screen. We will have this module developed, tested and made available to our colleagues at other institutions by the end of summer 1998.

Module 8: Fracture of fiber-reinforced composites

Quick-time movies of the damage evolution and failure in fiber-reinforced composites exist, and cover a range of the relevant stochastic variable in the problem. These must be adapted for use in the module, and preliminary text and concept development are underway.

Module 9: Elastic moduli and deformation of polycrystals

We will develop a continuum-level finite-element model to investigate the connection between the stresses, strains, and elastic constants in an anisotropic single-crystal and its associated polycrystal. Effects of grain orientation and anisotropy on the local stress state will be highlighted, as well as the overall averaging of the deformation leading to isotropic effective elastic constants in random polycrystal aggregates. The onset of yielding in individual grains in the polycrystal due to critical resolved shear stresses along the slip planes will then be considered as a function of increasing applied stress (but with non-interacting grains) to demonstrate the onset of bulk plasticity in polycrystals.

2) Course Development

Our initial effort has focused on the development of the senior-level course, which will be offered Fall semester 1998. The Fall offering allows senior students to complete Senior Design Projects in the spring semester in areas of computer simulation of materials.

The senior-level course is organized conceptually as follows. Topics will be introduced conceptually at the atomic level; atom interactions and mechanical response will be studied. Then, the same problem at the continuum level will be introduced and the atomic-scale results and concepts will be recalled as motivation and/or constituent input to the large-scale computations. In this manner, the course material will alternate between atomic and micro/meso/macro scales within each topic and so continually reinforce the connections between length scales. An outline of the course follows.

Title: Computer Simulation of Mechanical Behavior of Materials

Description: This course is aimed at students in engineering and pure sciences who are interested in the simulation of material properties at many length scales. The connections between interatomic forces, atomic- scale deformations and defects, and macroscopic properties such as elastic constants, fracture, and yielding, will be presented. Contact with established concepts such as Voight and Reuss averages for elastic moduli of polycrystals and the Griffith concept of fracture will be made very naturally. The computational/simulation approaches to dealing with problems at each length scale will also be highlighted. Scientific visualization tools will be used throughout to demonstrate the physical phenomena being studied.

Outline:

I. Interatomic potentials, Newton's laws, simulations           10%

II. Elastic constants of single crystals                        20%
    a. Symmetry
    b. Local atomic deformations

III. Elastic behavior of polycrystals                           25%
    a. Continuum concepts
    b. Elasticity and finite-element implementation
    c. Local and average responses
    d. Analytic concepts at the macro scale

IV. Dislocations and yielding                                   20%
    a. Dislocation structure, Peierls stress
    b. Yielding in polycrystals

V. Cracks and fracture                                          25%
    a. Atomic-scale stress concentrations
    b. Crack-size dependence, crack-tip instabilities
    c. Potential energy release rate, surface energy,
         Griffith's concept at atomic scale
    d. Continuum models of cracks and crack growth
    e. Crack growth in polycrystals

3) Instructional Technology Module Delivery:

The third goal of this project is to implement the latest interactive computer simulation and visualization tools such that students can gain the most knowledge from the various modules listed above. Java web-based tools allow students to work in a more interactive environment and the recently completed CAVE virtual environment at Virginia Tech has already been used by students to gain a more fundamental understanding of structure-property relationships.

Visualization of simulation model results using CAVE Technology:

The NSF-ARI CAVE project, "Breaking Barriers in Research and Education Using CAVE Technology", was recently installed and became operational on December 17, 1997. In the summer of 1997 the PI of the NSF-ARI and NSF-CRCD projects spent the summer at NCSA (a NSF supercomputing center and CAVE partner with Virginia Tech) where the Atomview CAVE application was developed in partnership with NCSA. This CAVE partnership was formally extended as the NSF National Computational Science Alliance (NCSA) Partnership in Advanced Computational Infrastructure (PACI) where Virginia Tech is a member of the Enabling Technologies Team in Virtual Environments. Atomview has already been successfully used by NSF-CRCD team members in their research: see http://www.sv.vt.edu/future/vt-cave/VT/#proj. Results of this research have been demonstrated to students in the undergraduate curriculum at Virginia Tech using the Atomview CAVE application during spring semester 1998. From these intial efforts in the last two months, January - February 1998, we have learned how to create VRML files of complex structures that can either be distributed to our remote site collaborators through the web or loaded into the Virginia Tech CAVE for a more insightful immersive experience. We intend to further extend the use of the Atomview CAVE application for modules 1 through 6 where VRML is the common data format:It can either be download into a VRML player on a web browser which can then be used by the student interactively or it can be loaded into an immersive interactive environment such as an I-Desk or CAVE, both from the same web page. Here the emphasis is on organizing the content such that the students at remote sites who do not have CAVEs can still benefit from the use of interactive VRML web-based tools.

Interactive Web based Java Technology:

In collaboraton with SUN Microsystems Corporation and Visual Numerics, Inc. (VNI) we have developed two new Java web-based tools that will be incorporated into the NSF-CRCD modules. Visual Numerics Inc is an industrial partner on this NSF-CRCD project. These two Java web based tools are: 1) Network Programming Interface Builder (NPIB) and 2) Visualizer. NPIB allows for rapid development of interfaces that can be used by students to submit batch jobs to supercomputers and view the results as a VRML file. Visualizer allows for the rapid development of interactive interfaces where students can parametrically explore complex analytic or numerical models ( http://www.jwave.vt.edu). Both NPIB and Visualizer were proof of concept projects that were further developed into products by SUN and VNI as "Studio" and "J-Wave" respectively. Both SUN and VNI have extended funding ($18K) into the second year that will benefit this NSF-CRCD project. Our goal here is to focus more on the content and not further develop these tools. Future development of NSF-CRCD modules will now benefit from Sun's Studio and VNI's J-Wave Java technology.

Evaluation/Assessment:

Evaluation and assessment of modules will be implemented during the actual construction of the modules this summer. Greg Sherman from the Department of Teaching and Learning will substitute for John Burton, who is now the head of the Department of Teaching and Learning. John Burton was the original CoPI on the project.