Re-Envisioning Materials Science Education Through Atomic-Level Computational Modeling

Jacob Kelter; Jonathan Daniel Emery

Download Paper | Permalink

Conference: 2024 ASEE Annual Conference & Exposition
Location: Portland, Oregon
Publication Date: June 23, 2024
Start Date: June 23, 2024
End Date: July 12, 2024
Conference Session: Materials Division (MATS) Technical Session 1
Tagged Division: Materials Division (MATS)
Permanent URL: https://peer.asee.org/47920

Request a correction

Paper Authors

biography

Jacob Kelter Northwestern University

visit author page

Jacob Kelter is a PhD candidate at Northwestern University in the joint program between computer science and learning sciences. His research focuses on using agent-based modeling for science education and computational social science research, both related

visit author page

biography

Jonathan Daniel Emery Northwestern University

visit author page

Jonathan Emery is an Associate Professor of Instruction in Materials Science and Engineering at Northwestern University.

visit author page

Download Paper | Permalink

Abstract

Computational tools are transforming materials science and engineering (MSE). There is a consensus in the field that both undergraduate and graduate MSE education need to adapt to this transformation. Despite the consensus, introductory courses still largely lack computational modeling, and upper-level courses typically only use it for problem solving, not as a core representation of phenomena.

As computation has revolutionized the practice of MSE, a substantial body of educational research has demonstrated the power of computational tools in science education, particularly agent/atomic-based modeling (ABM). One of ABM’s powerful learning affordances is the perspective of emergence it foregrounds. Traditionally, differential equations, which do not explicitly represent atomic-scale mechanisms, are used to model the dynamics of macro-level variables. In contrast, ABMs model discrete atoms and their micro-level interactions from which macro-level properties emerge. This shift in perspective, paired with interactive and visual simulations, is useful not only for doing science but also in supporting students to develop deep conceptual understanding. Atomistic computational techniques should therefore not be reserved for advanced classes. Rather, they should be used from the beginning to teach and learn the fundamentals.

To this end, we re-designed the introductory MSE course to make use of ABM for nearly every topic. Many phenomena in MSE can be modeled and understood using a small number of atomistic computational techniques, allowing most of the topics covered in an introductory course to be taught using computational representations without overloading the curriculum. The conceptual structure of these techniques is often simple. With the right computational tools, the computational models can be made visual and interactive to enable active student exploration and model-based inquiry learning.

In this paper, we start with background on computation in MSE, ABM in education, and learning theories that underly our design, including constructionism and the learning cycle. We then outline the structure of our redesigned course, including all the topics covered and which computational techniques we use to model them. Next, we describe the general structure of individual modules within the curriculum and the iterative design process we have engaged to create them. Each module is built around one or more interactive ABMs of the target phenomenon (implemented with the widely used NetLogo modeling platform) and includes learning activities for students and explanatory text and/or video. Two modules, on atomic bonding and diffusion, will be described as examples. A brief analysis of student outcomes for the diffusion unit will illustrate how learning can be assessed to include understanding of emergence from atomic mechanisms. Students were given a free response exam question that involved predicting the evolution of a concentration profile (macro-level), sketching atomic level motion (micro-level), and explaining how the concentration profile behavior emerges from that motion. Student answers were coded for correctness and misconceptions. While most students showed strong understanding on all three parts, some misconceptions and hybrid-conceptions remained.

Citation
Format

Kelter, J., & Emery, J. D. (2024, June), Re-Envisioning Materials Science Education Through Atomic-Level Computational Modeling Paper presented at 2024 ASEE Annual Conference & Exposition, Portland, Oregon. https://peer.asee.org/47920

TY - CPAPER
AB - Computational tools are transforming materials science and engineering (MSE). There is a consensus in the field that both undergraduate and graduate MSE education need to adapt to this transformation. Despite the consensus, introductory courses still largely lack computational modeling, and upper-level courses typically only use it for problem solving, not as a core representation of phenomena.

As computation has revolutionized the practice of MSE, a substantial body of educational research has demonstrated the power of computational tools in science education, particularly agent/atomic-based modeling (ABM). One of ABM’s powerful learning affordances is the perspective of emergence it foregrounds. Traditionally, differential equations, which do not explicitly represent atomic-scale mechanisms, are used to model the dynamics of macro-level variables. In contrast, ABMs model discrete atoms and their micro-level interactions from which macro-level properties emerge. This shift in perspective, paired with interactive and visual simulations, is useful not only for doing science but also in supporting students to develop deep conceptual understanding. Atomistic computational techniques should therefore not be reserved for advanced classes. Rather, they should be used from the beginning to teach and learn the fundamentals.

To this end, we re-designed the introductory MSE course to make use of ABM for nearly every topic. Many phenomena in MSE can be modeled and understood using a small number of atomistic computational techniques, allowing most of the topics covered in an introductory course to be taught using computational representations without overloading the curriculum. The conceptual structure of these techniques is often simple. With the right computational tools, the computational models can be made visual and interactive to enable active student exploration and model-based inquiry learning.

In this paper, we start with background on computation in MSE, ABM in education, and learning theories that underly our design, including constructionism and the learning cycle. We then outline the structure of our redesigned course, including all the topics covered and which computational techniques we use to model them. Next, we describe the general structure of individual modules within the curriculum and the iterative design process we have engaged to create them. Each module is built around one or more interactive ABMs of the target phenomenon (implemented with the widely used NetLogo modeling platform) and includes learning activities for students and explanatory text and/or video. Two modules, on atomic bonding and diffusion, will be described as examples. A brief analysis of student outcomes for the diffusion unit will illustrate how learning can be assessed to include understanding of emergence from atomic mechanisms. Students were given a free response exam question that involved predicting the evolution of a concentration profile (macro-level), sketching atomic level motion (micro-level), and explaining how the concentration profile behavior emerges from that motion. Student answers were coded for correctness and misconceptions. While most students showed strong understanding on all three parts, some misconceptions and hybrid-conceptions remained.
AU - Jacob Kelter
AU - Jonathan Daniel Emery
CY - Portland, Oregon
DA - 2024/06/23
PB - ASEE Conferences
TI - Re-Envisioning Materials Science Education Through Atomic-Level Computational Modeling
UR - https://peer.asee.org/47920
ER -