ILP Institute InsiderFebruary 6, 2013
The Chemical Industry Remix
Taking chemical reactions digital to bring innovation to an industry looking for its next revolution.
With the increasing pace of technological innovation in the digital age, it’s hard to think of an industry that hasn’t been transformed over the past few decades, but MIT Professor of Chemical Engineering Bill Green has one in his sights.
“Most of the commercial technologies in the chemical industry were developed back in the 50’s and 60’s,” he says. “They had slide rules. They had no computers, they had kind of primitive chromatography, and they were smart guys, they figured this out, but there’s no reason to think that any of those systems are actually optimal.”
“Back in the 50’s and 60’s energy was very cheap, so they never used that as a constraint,” Green explains, “and there were no environmental laws really, so they never used that as a constraint either. Now there’s a chance to go back and say, well, if you want to have something that’s environmentally responsive and economical in a time when energy is expensive then maybe you want to change the whole process.”
But using the methods that developed the old processes, actually cooking every possible combination of chemicals to see what happens, it would take as much time to reinvent the chemical industry as it took to invent it in the first place. In fact, in Professor Green’s view, it’s precisely these slow, old-fashioned practices that are inhibiting innovation in energy and many other chemical-related fields.
“You might have to put a lot of experimental work in and find out that a system’s not going to work,” Green says. “The research managers don’t want to do that. That’s one of the reasons why the chemical industry and the energy industry have much lower innovation rates than for example the computer science industry.”
That’s why Green and his lab, the Green Group, are taking chemical reactions digital.
“Since about the late 90’s, it’s been possible to actually compute chemical reaction rates on a computer by solving the Schrodinger equation,” he says. “By doing that, then we can predict how things will react without having to do the experiments, and so we can pre-screen a lot of possibilities and focus our attentions on the systems that have a good chance to succeed.”
Knowing which systems have a good chance to succeed could give research managers a running start in the laboratory and kick start the pace of innovation in an industry looking for its next revolution.
The Potential in Prediction
All the Green Group’s research projects focus around energy, but within energy, they cast a wide net, working on everything from biofuels and hydrogen to gasification and new engine design. Using their predictive models, the Green Group has compared isomers of Butanol, a potential supplement or replacement for ethanol as the main biofuel in the U.S.; they’ve collaborated with the automotive industry to test fuels for a new engine type that doesn’t use a flame; and they’ve joined forces with numerous energy industry partners to examine more efficient uses of and treatments for fossil fuels and their waste products.
But even with all the potential in each of their individual projects to impact the energy industry, Professor Green sees the greatest potential in developing the capacity for predicting chemical reactions itself. “The capability to be able to predict reactions,” he says, “it opens up so many doors.”
Because it turns out, there’s a lot we still don’t know about everyday chemical phenomena. “People know if you cook something in a certain way, exactly right, you can make a cake,” Professor Green says, “but no one can tell you how to make a cake on a computer. It’s not possible yet because we don’t understand all the chemical reactions that happen in baking a cake.”
According to Green, the same is true of most commercial chemical processes. “Sometimes people find a recipe that works, but you never really know if it’s the best recipe, and you don’t know what all the different alternatives are, and so you can’t intelligently design the system to be as efficient as possible.”
With a truly predictive capability for modeling chemical reactions, all that could change. It would become possible for a computer program to redesign old chemistry and even discover new reactions. “The computer could exhaustively find out all the different possibilities,” Green says. “It’s theoretically possible to do it, and so I’m very excited to pursue that and try to figure out how to automate it.”
Professor Green envisions a day when a chemical engineer will be able to propose a product, feed it into their computer model, and find out the range of what’s chemically possible without having to set foot in a lab. “It’ll tell what product you’ll get and how much pollution you’re going to make and you can just go, oh, that’s not good, let’s change the design and try it again,” Green says. “I think that’s really where the future is.”
MIT: A Home for Industry
The Green Group’s numerous industrial collaborations provide the lab with an endless array of problems on which to test and refine their computer models as they work towards perfecting their predictive capabilities. Through the ILP and the MIT Energy Initiative, Professor Green has forged partnerships on projects large and small, involving multidisciplinary teams from across the institute or a dedicated group of students exploring a singular research need.
“The engineering students, they want to do something that’s useful and practical and real, not just theoretical,” Green says. “They feel like they’re really doing something that matters, someone really cares, this is not just for publishing a journal paper, but there’s a chance that what they invent might actually become something soon in the real world.”
In Green’s view, opportunities for that kind of real-world collaboration don’t happen just anywhere. “I’m fortunate to be at MIT,” Green says. “MIT has a long history of working really well with industry. The industry that comes here to work with us, they’re serious, they want to work with us, and our students want to work with them, it’s really a marriage made in heaven.”
More ILP News
- Better Machine Learning January 19, 2015
- Watching How the Money Flows January 12, 2015
- Making More than Pretty Pictures December 30, 2014
- MISTI: The Global Internship Program December 16, 2014
- Evolutionary Approaches to Big Data Problems December 8, 2014
- Commercializing A New Generation of Polymer Coatings December 1, 2014