Tag Archives: molecular-workbench

More things, more people, more easily

“The most exciting thing about the Next-Generation Molecular Workbench is that it lets us to do more things for more people, more easily.” -Chad Dorsey, CEO & President of the Concord Consortium

Have you seen the latest Next-Generation MW interactives? We’ve taken the physics-based interactive simulations and made them better. The Next-Generation Molecular Workbench is accessible to anyone with a browser and works on multiple devices (desktops, laptops, tablets and phones).

With our easy-to-share interactives (just click the “Share” link in the top right corner of any interactive), you can develop your own activities on your blog or teacher page. CK-12 and edX already have!

And best of all, the code is free and open source. Want to change the model? You can get access to the code and tweak it.

Learn more in our latest video from the Molecular Workbench team!

Share and embed—easily!

One of the key features of our Next-Generation Molecular Workbench is the ability to easily share and embed interactives in blog posts, learning management systems, emails and more—wherever you can paste a weblink or HTML code. Just two simple steps will have you sharing your favorite interactives with all your friends and colleagues in no time flat!

  1. Click the Share link at the top of an interactive.
  2. Copy and paste the link into Facebook, Google+, Twitter, Pinterest or wherever you want to share the interactive.

Want to embed the interactive in your own blog or web page instead?

  1. Click the Share link at the top of an interactive.
  2. Copy the HTML and paste the iframe code where you want the interactive to appear.

Sharing and embedding Next-Generation Molecular Workbench interactives

Learn more about how easy it is to share interactives.

We want to make it easy for you to learn and teach with accurate scientific models.  We’ve gotten it down to two steps. Now it’s up to you to share your favorite interactives far and wide. 🙂

Explore currently available interactives.

Share with us: which are your favorite interactives and why? What interactives do you want to see?


Modeling Physical Behavior with an Atomic Engine

Our Next-Generation Molecular Workbench (MW) software usually models molecular dynamics—from states of matter and phase changes to diffusion and gas laws. Recently, we adapted the Molecular Dynamics 2D engine to model macroscale physics mechanics as well, including pendulums and springs.

In order to scale up the models from microscopic to macroscopic, we employ specific unit-scaling conventions. The Next-Generation Molecular Workbench (MW) engine simulates molecular behavior by treating atoms as particles that obey Newton’s laws. For example, the bond between two atoms is treated as a spring that obeys Hooke’s law, and electrostatic interactions between charged ions follow Coulomb’s Law.

Dipole-dipole interactions simulated using Coulomb’s Law.

At the microscale, the Next-Generation MW engine calculates the forces between molecules or atoms using atomic mass units (amu), nanometers (10−9 meters) and femtoseconds (10-15 seconds), and depicts their motion. To simulate macroscopic particles that follow the same laws, we can imagine them as microscopic particles with masses in amu, distance in nanometers, and timescales measured in femtoseconds. Once the Next-Generation MW engine calculates the movement of these atomic-scale particles, we simply multiply the length, mass and time units by the correct scaling factors. This motion satisfies the same physical laws as the atomic motion but is now measured in meters, kilograms and seconds.

In the pendulum simulation below, the Next-Generation MW engine models the behavior of a pendulum by treating it as two atoms connected by a very stiff bond with a very long equilibrium length. The topmost atom is restrained to become a “pivot” while the bottom atom “swings” because of the stiff bond. Once the engine has calculated the force using the atomic-scale units, it converts the mass, velocity and acceleration to the appropriate units for large, physical objects like the pendulum.

Large-scale physical behavior simulated with a molecular dynamics engine.

In order to appropriately model the physical behavior of a pendulum or a spring, we use specific scaling constants. Independent scaling constants for mass, distance and time enable us to convert nanometers to meters, atomic mass units to kilograms and femtoseconds to model seconds. Using the same scaling constants, we can derive other physical conversions, such as elementary charge unit to Coulomb. In order to make one model second pass for every real second, we adjusted the amount of model time between each page refresh. We also chose to simulate a gravitation field—a feature usually absent in molecular dynamics simulators—because it is relevant to macroscopic phenomena.

From microscale to macroscale, the Next-Generation Molecular Workbench engine is a powerful modeling tool that we can use to simulate a wide variety of biological, chemical, and physical phenomena.  Find more simulations at mw.concord.org/nextgen/interactives.

9 Highlights of 2012

It was a great year for the Concord Consortium!

  1. We won a Smaller Business Association of New England (SBANE) Innovation Award!
  2. Next-Generation Molecular Workbench interactives starred in the MIT MOOC (Massive Open Online Course) “Introduction to Solid State Chemistry” through a new collaboration with edX.
  3. Chad Dorsey described our vision of deeply digital education at the national Cyberlearning Research Summit.
  4. Six new projects were funded by the National Science Foundation: InquirySpace, Understanding Sub-Microscopic Interactions, High-Adventure Science: Earth’s Systems and Sustainability, GeniVille, Graph Literacy, and Sensing Science.
  5. The What Works Clearinghouse (WWC), a federally funded organization that scans educational research for high-quality studies, recognized our Technology Enhanced Elementary and Middle School Science (TEEMSS) software and materials.
  6. The Concord Consortium Collection was accessioned into the National Science Digital Library (NSDL).
  7. Our debut webcast featured Chad Dorsey, speaking about the scientific and engineering practices of the Next Generation Science Standards and our free, technology-based activities.
  8. We had two fabulous Google Summer of Code students.
  9. Our staff population increased by 10%, thanks to our new Software Portfolio and Project Manager Jen Goree, Web Developer Parker Morse, and Software Developer Tom Dyer, who just started (technically in 2013, but we’re so excited, we’ve included him on this 2012 list)!

2013 promises to be another great year! Follow us on Facebook, Twitter, Google+, and subscribe to our mailing list to receive print or email news updates.

Hitting the Wall

Gas laws are generally taught in high school chemistry. Students learn that Boyle’s law, for instance, can be expressed as P1V1=P2V2, where P is pressure and V is volume.

From the equation, it’s clear that there is an inverse relationship between the gas pressure and volume, but do students understand the molecular mechanism behind this relationship?

Since students are programmed to plug and chug, if you give them, say, P1, V1, and P2, they can find the numeric value of V2. Although students can get the correct answer, teachers have told us that their students don’t really understand the gas laws because they don’t have a mental model of what’s happening. Gases are, after all, invisible! Nor can students see volume or pressure.

Molecular Workbench makes the gases, volume, and pressure visible. With a new set of Next-Generation Molecular Workbench interactives, students can experiment with increasing the pressure on a gas to see why the gas volume decreases.

The “What is Pressure?” interactive (above) shows the inside (yellow atoms) and outside (pink atoms) of a balloon. (Even the velocities of the individual atoms are visible with vectors!) The green barrier represents the wall of the balloon.

Students learn that pressure is nothing more than molecular collisions with a barrier. In the beginning, atoms hitting the balloon wall on either side move it just a tiny bit—transferring some of their kinetic energy to the barrier. At equilibrium, the balloon wall remains (relatively) stationary. (Go ahead and run it to see!)

But if you add atoms to the balloon, the balloon wall moves out; more atoms means that there is increased pressure pushing outwards on the barrier. Since the number of atoms on the outside of the balloon hasn’t changed, the pressure pushing inwards is the same as it was before. With unbalanced forces, you get net movement.

With barriers, we can also measure the pressure caused by those molecular collisions.

In the “Volume-Pressure Relationship” interactive (above), students see a visual representation of Boyle’s law.

Other models allow students to investigate all the relationships of Charles’s law (V1T2=V2T1), Gay-Lussac’s law (P1/T1=P2/T2), and Avogadro’s law (V1/n1=V2/n2).

And, of course, all of these relationships together make up the Ideal Gas Law (PV=nRT). Explore gas laws today with some HTML5 molecular models!

Optimizing short-range and long-range atomic interactions

[Editor’s note: Piotr Janik (janikpiotrek@gmail.com) was a Google Summer of Code 2012 student at the Concord Consortium and is now a consultant working on our Next-Generation Molecular Workbench.]

Some time ago we described the core engine used in Molecular Workbench and our attempts to speed it up. At that time we focused mainly on the low-level optimization connected with reducing the number of necessary multiplications. This promising early work encouraged us to think even more about performance.

We next reviewed existing algorithms in the core of the molecular dynamics engine. To make a long story short, atoms interact with each other using two kinds of forces:

  • Lennard-Jones forces (repulsion and short-range attraction)
  • Coulomb forces (electrostatic and long-range attraction)

Atomic interactions are pairwise, meaning that we have to calculate forces between each pair of atoms while using the basic, naive algorithm. Having n atoms, we must perform about n^2/2 calculations. “The Big O” notation can be used and the computational complexity can be described as O(N^2), which means that the execution time of calculations grows very fast as the number of atoms used in the simulation increases. This is definitely an unwanted effect, but fortunately there are ways to reduce the complexity.

Solutions are different for short-range and long-range forces, so let’s start with short-range. “Short-range” means that atoms interact only while they are quite close to each other. Let’s use rCut as a symbol for the interaction maximum distance. So, one obvious optimization would be to limit calculations to pairs of atoms that are closer to each other than rCut. How? There are two popular approaches—cell lists and Verlet (neighbor) list algorithms.

The cell lists algorithm is based on the concept that we can divide the simulation area into smaller boxes or cells. Each cell dimension is equal to the maximum range of interaction between atoms—rCut. So, while calculating interactions for a given atom, it’s enough to take into account only atoms from the same box and its closest neighbors. Atoms in other boxes are too far to interact with this atom. This is both simple and effective, reducing computational complexity to O(N)! Note that it’s C * O(N) with a pretty significant C, unfortunately.

However, while calculating interactions between atoms in neighboring cells, still only 16% of atoms that we take into account are interacting! This is a waste of resources and where we find room for further optimizations. So, what about creating a list for each atom, which contains only atoms actually interacting with it? This Verlet or neighbor list algorithm as it’s called works well. The only problem is that we have to be smart about updating these lists, as atoms constantly change their position and, thus, their “neighborhood.” We can slightly extend these lists to also include some atoms outside the area of interaction. So each list should include atoms closer than rCut + d from the given atom, where d defines a buffer area size. Because of that, lists need to be updated only when the maximum displacement of some atom, measured since the moment of the previous lists update, is bigger than d. If it’s smaller, neighbor lists are still valid. Lists can be updated using the normal, naive algorithm (which still leaves the complexity O(N^2)), or even better, using the cell lists algorithm presented above! This ensures complexity O(N) and greatly reduces inefficiencies of the cell lists approach.

We’re also working on long-range forces optimization. Since we can no longer use the assumption that atoms interact only when they are close to each other, we can’t rely on the optimization strategies above. The algorithms are now more complicated. The problem of the electrostatic interaction is akin to a problem of gravitational interactions (called N-body problem), popular in astrophysics. One of the most common algorithms for speed-up of such calculations is the Barnes-Hut algorithm. We tried to implement it, but the overhead connected with creating additional data structures was bigger than potential performance gains. The reason is that the number of charged atoms we use in our models is too small to see the advantage of such an approach. As a result, we left our naive algorithms for long-range interactions, which perform better due to their simplicity.

However, we successfully implemented both short-range optimizations in Next-Generation Molecular Workbench and the results are spectacular. The speed-up varies from 20% for really small models (where the number of atoms is less than 50) to 700% for bigger ones (where the number of atoms is about 250). This is the really significant improvement and made complex models usable. As you can see, conceptual, algorithmic optimizations really matter!

We’re still thinking about further optimizations, both low level and algorithmic. Stay tuned as the Next-Generation MW is getting more and more computational power!

Improving Speed at the Heart of Molecular Workbench

At the heart of Molecular Workbench’s modeling of atomic interactions is a profoundly important but fundamentally simple concept:

At close distances, atoms attract each other until they get so close that they repel.

Here’s a demo of that concept: two atoms interacting. Drag the green atom to various locations near and far from the purple atom and watch what happens as the two atoms approach each other and move apart. (If you’re wondering why the atom slows down and stops, the answer is that we apply an artificial damping force to the green atom in order to make it easier for you to “grab” it and play with it.)

This concept is called intermolecular attraction. Molecular Workbench (MW) uses an approximate formula for calculating the intermolecular potential that was originally proposed by John Lennard-Jones (in 1924!) and is now called the “Lennard-Jones potential” or L-J for short.

Lennard-Jones potential

Here you see the L-J potential as a graph. The horizontal axis shows distance between two atoms, the vertical axis is the net intermolecular energy, with regions of negative slope indicating that the resulting force is repulsive, and regions of positive slope indicating attraction. This graph shows that these atoms will attract to each other if they are more than 2.3 radii apart, but begin to repel sharply at distances less than that value as shown by the steep rise in the curve.

This interaction is not just a fundamental concept in physics and chemistry. It also is quite central to the MW simulation engine. We typically do this calculation tens or hundreds of thousands times per second, especially when we have many atoms interacting, such as here:

If you study our code, going deeper and deeper, you’ll peel away layers of the coding onion, until you get to the very center of this model, which calculates the L-J force between just one pair of atoms. (Did we mention that the reason the L-J approximation is used is that it’s considered relatively fast to calculate?) It does this for each of the many pairs of interacting atoms, repeating over and over, with each time-tick of our simulation.

As it turns out, the L-J formula is still computationally demanding, requiring calculating 6th and 12th powers. (That’s the theory. In practice, the form that is most convenient for our code happens to use the 8th and 14th powers. That also speeds things up–but we’ll save that for another blog post.) So when we’re looking for ways to make our code run faster and our model run better, the L-J calculation is a prime place to look!

In our first pass at improving speed (in MW Classic), we converted the 6th and 12th power calculations to simpler repetitive multiplications:

X2 = X * X

X3 = X2 * X

X6 = X3 * X3

X12 = X6 * X6

That gave us just 4 multiplication tasks, instead of 16 (X6 has 5 multiplications, X12 has 11).  This reduced the calculation demand to 25%. The model ran faster and smoother.

Recently, we tried another method to improve speed: using a look-up table, with pre-calculated values. We computed the L-J values for each element type for dozens of typical interatomic distances and put them into a table. We thought this would save computational time because the software could simply look up the values in the table without any multiplication.

We tested the two methods (multiplication vs. look-up table) over many iterations and found that the look-up table was slower! As we investigated further, we saw that accessing the look-up table had its own overhead in terms of use of cache memory and data transfer. This was evidence of a trend in improving computer speeds over the past few years: computation speeds improve at a faster rate than memory access speeds. They both are faster, but computation (i.e., all that multiplying we have to do) has gotten the improvement edge.

So, after this testing, we went back to the efficient multiplication approach. This illustrates our basic approach: creative thinking validated by empirical data.

We will continue to develop and test creative ways to speed the software and improve the user experience, especially as we move to support a greater variety of learning activities and computer platforms.

Exhibit Booth at BCCE Conference and free MW buttons

Just got back from the Biennial Conference on Chemical Education (BCCE 2012), where I participated in a symposium titled “Web-Based Resources for Chemical Education.” About 60 people attended to learn about Molecular Workbench and other online tools and resources. One of the audience questions was about future availability of Molecular Workbench on the iPad and other tablets. Our latest work on the HTML5/JavaScript next-generation MW project, generously funded by Google.org, will address exactly this. We’ll be bringing much of the Java-based Classic MW to the browser, so that any device running a modern Web browser will be able to run our newest interactives and activities.

I didn’t get to attend many of the other sessions at BCCE because much of my time was spent staffing Concord Consortium’s exhibit booth to disseminate our free software. Jeanne Hurtz and I spoke with hundreds of people who stopped by our booth to hear about the current MW capabilities and see a next-generation MW model running on a tablet. We gave away about 350 MW buttons, but have a few left. If you’d like one of your own, please stop by our office at 25 Love Lane in Concord, MA, to pick one up!

It was great to share the excitement of MW’s potential and versatility with so many new people. We heard from many (surprised) guests at our booth: “This is free?”  Yes! And so is the button.

Flexible textbooks

We’re in the midst of a remarkable transition in education – a change that will give teachers more flexibility in the resources they use in their classroom.

The growing role of digital textbooks is gaining momentum. Major publishers are not just converting their textbooks to digital format, they’re also reconceptualizing them, adding a more diverse array of embedded interactives and providing states and districts with the option to pick and choose sections to meet local educational goals.

Think about this for a moment.

We are used to the monolithic textbook package – a basal textbook, lab manuals, CDs and other ancillaries. Each major publisher offers its package. States and districts decide which publisher’s package to purchase. End of story.

But that world is changing. A district might choose several chapters from one publisher and other chapters from a second publisher. From a third publisher, they might select a lab manual that is especially engaging for their students. And they might select multiple online resources to extend student learning.

From the teacher’s perspective, this is potentially liberating. Instead of working through the standard textbook and its aligned support materials, teachers have a richer set of options. They can select resources based on personal expertise, knowledge of their students, teaching style and familiarity with the growing array of digital interactives.

How does Molecular Workbench fit in? MW helps students understand fundamental principles of physics, chemistry and biology, yet it hasn’t always been clear how to fit this into the classroom, as it might seem a diversion from the flow of the textbook.

With a more flexible approach to teaching and learning, science teachers will be able to easily integrate the power of atomic and molecular simulations into their classrooms. This will not be an aberration, but the new norm.

This change will take a few years to fully play out, but it is a welcome transition away from the dominance of the standard, one-size-fits-all textbook and towards freedom to use a robust set of resources – including Molecular Workbench.