Molecular Workbench and the Rise of the MOOC


MOOCs are all the rage. However, don’t worry if you don’t recognize the acronym. It popped up only last fall, when Stanford offered—free of charge—a graduate-level course in artificial intelligence. Over 160,000 students from 190 countries signed up, defining what is quickly becoming a new genre in online education: the Massive Open Online Course or MOOC.

This genre expanded further last spring as MIT attracted over 150,000 students to its “Circuits and Electronics” MOOC. The graduating group made the course’s broad appeal clear—7,157 strong and ranging in age from 14 to 74. The promise of such courses led MIT, Harvard and the University of California, Berkeley to form the nonprofit partnership edX. Now, a collaboration between edX and the Concord Consortium is forging new ground for how MOOCs can be used to teach science.

On October 15, MIT opened a MOOC through edX entitled “Introduction to Solid State Chemistry.” This MOOC, in effect a chemistry approach to materials science, is an online version of one of MIT’s longest-running and highest-enrollment traditional courses. It covers the relationship between such microscopic features as electronic structure, chemical bonding and atomic order and the resulting atomic arrangement and macroscopic properties of various crystalline and amorphous solids, including metals, ceramics, semiconductors, polymers and even proteins.

This course’s strong focus on intermolecular interactions and their importance in giving rise to material properties made integration with the Molecular Workbench a natural choice. Molecular Workbench’s computational simulation capabilities turn the invisible world of atoms and molecules into an accurate environment for student exploration and experimentation.

Molecular Workbench was one of the first software engines to use research-grade algorithms to simulate atomic motions, forces and interactions for generating real-time visualizations of atomic-level behavior. A grant from Google.org is enabling us to move this molecular dynamics engine to the Web, an advance that permits integrations such as this collaboration with edX.

Exploring intermolecular forces

The core of this MIT chemistry MOOC is the study of why solids, as opposed to liquids and gases, hang together, a phenomenon that derives directly from the fundamental force between charged particles. Memorizing simple rules such as “like charges repel and unlike charges attract” doesn’t begin to capture the complexity of intermolecular forces in the real world. For the MIT MOOC we used Molecular Workbench to create four interactives that help students experience and understand these nuances:

Intermolecular Attractions

Intermolecular Attractions. Oppositely charged atoms (blue and red) attract to each other while neutral atoms (gray) attract to everything. Dashed lines show the attractions.

In this introductory activity, students explore patterns of attractive forces between molecules by dragging polar and non-polar molecules around (Figure 1). Dashed lines appear where there are attractions between molecules. Through direct manipulation they explore the surprising notion that neutral atoms attract to everything.

Differences in Attractive Force

Differences in Attractive Force. Students pull on an imaginary handle (green star attached with springs) to test how strongly pairs of molecules attract to each other.

In this activity, students choose between different pairs of molecules—two non-polar molecules, two polar molecules (Figure 2) and a non-polar and polar molecule—and then pull on one of them. Through these experiments, students determine the relative strengths of attraction among combinations of polar and nonpolar molecules.

Phase Change

Phase Change. Non-polar (gray) and polar (blue and red) substances can be heated and cooled at the same time to compare melting and boiling points.

Here, students make the connection between the unseen molecular forces they have just explored and the familiar macroscopic phenomenon of phase change. Using a temperature control, students “boil” two different solids, one composed of non-polar molecules, the other of polar ones (Figure 3).

Solubility

Solubility. Polar (blue and red) and non-polar (gray) substances separate when mixed together, but are still attracted to each other.

This activity illustrates the point that atomic-level models can connect properties that seem to be entirely unrelated by underscoring the little-understood connection between boiling point and solubility. Students know inherently that oil and water don’t mix. However, their investigations have shown that at the atomic level everything attracts to everything else. Students explore this seeming puzzle by mixing a non-polar substance (like oil) with a polar substance (like water). While the two do attract, the stronger attractions between polar water molecules “squeeze” the oil molecules out from between them, causing the two to separate (Figure 4).

Bringing labs into MOOCs

These interactives serve as initial examples that offer one view into a much larger question: how can MOOCs be used best to teach science content? Most MOOCs today are a fascinating hybrid of the new and the familiar. At their core they are simply a series of videotaped lectures given by the professor. Students “attend” these lectures whenever they want, at their own pace, starting, stopping, repeating or skipping over sections at their discretion. They are given weekly assignments to complete and a discussion forum in which to ask and answer questions. Over time this forum becomes a specialized Wikipedia devoted to the material of the course.

Since they are by definition completely online courses, MOOCs naturally lack in-person laboratory components that students would receive on campus. Many of the first widely popular MOOCs in this quickly rising field were offerings in computer science from groups such as Stanford, Coursera or Udacity. These had the advantage that student programming can act as the “lab,” and highly structured computer languages can be adapted to exercises that can be easily auto-scored. Recent MOOCs offered in other domains simply do not attempt to face this challenge, hewing instead to typical modes of lecture and occasional written homework responses.

However, with the capabilities of models and simulations such as those from the Molecular Workbench, we can provide “lab” experiences in areas beyond what is even possible in physical laboratories. In this case, the exploration occurs at the molecular level, but simulations can permit students to explore multi-scale phenomena such as genetics, long-ranging phenomena such as evolution or large-scale phenomena such as climate change just as easily. MIT has also been a leader in this realm, having created a simulated electronics environment specifically for its initial MOOC offering.

New frontiers

Using models or simulations for such applications opens the door to much more interesting possibilities such as a reliable means for evaluating student understanding. Researchers at the Concord Consortium are working on logging student interactions with the interactives and providing feedback based on their actions. In the long run, we expect to be able to use students’ manipulations of the interactives as a reliable marker for their understanding of the underlying scientific principles and their facility with the practices of science and engineering. By taking advantage of possibilities such as this, MOOCs may indeed become the platform of the future for education, and bring sophisticated learning about science, math and engineering to millions around the world.

Paul Horwitz (phorwitz@concord.org) is a senior scientist.
Dan Damelin (ddamelin@concord.org) is a technology and curriculum developer.
Chad Dorsey (cdorsey@concord.org) is President of the Concord Consortium.

This material is based upon work supported by Google.org.