With AI, UMD Researcher Creates New Recipes for Biodegradable Plastics

Like a sprinkle of domestic magic, those pre-measured pods combining detergent, fabric softener and brightener made life a bit easier when they hit the market more than a decade ago.  But it’s unclear whether laundry pods, made of a plastic called polyvinyl alcohol (PVA), can fully biodegrade after standard wastewater treatment, potentially leading to less-charmed outcomes for ecosystems, wildlife and even human health.

That cleaning conundrum, along with impacts from thousands of other kinds of plastic in daily life–from discarded food wrap to empty bottles—prompted University of Maryland engineering researchers to scour the world of materials science for biodegradable alternatives: natural substances that mimic regular plastic without the potentially toxic leftovers.

“It’s a massive pollution problem with no single solution,” said chemical and biomolecular engineering Assistant Professor Po-Yen Chen. He is leading the project, which is supported by a $2 million grant from the National Science Foundation and stems from a UMD Grand Challenges Grant.

While the effects of plastic as pollutants aren’t well understood, studies suggest, for example, that exposure to plastic additives like BPA and phthalates can disrupt hormones and may trigger inflammation, oxidative stress and DNA damage, while some plastics may increase cancer risk. A new study even points to potential climate effects, finding the presence of these “microplastics” could be reducing oceans’ ability to absorb atmospheric carbon.

The United Nations Environment Programme estimates that worldwide, people produce more than 430 million metric tons of plastic each year, mostly for packaging; half is used just once and then thrown away (called single-use plastics), and about 11 tons of it ends up in our lakes, rivers and oceans. Some kinds persist in the environment for 400 years or more. And despite company promises and consumer efforts to recycle at the curb, less than 10% of plastic is salvaged for reuse.

Until recently, the biggest obstacle to Chen’s search for greener products has been time. To find the exact formulation that meets a product’s requirements for characteristics like strength, opacity and moisture management is a huge ask, and to test every combination and ratio of biodegradable substances is unfeasible by conventional trial-and-error research methods, he said: “Even with our current library of 23 natural components like cellulose, gelatin, carrageenan, gluten, and additives like sorbitol and glycerol, to test all possible combinations would take 1.8 million years.”

But here comes artificial intelligence (AI) to make it all possible. With AI, plus robotics and computer modeling, “we can do those same tests in just a few months,” Chen said. He likens the process to making soup, adding different amounts of different ingredients “until you get the very tastiest soup you can make”—but in this case robots are adjusting the recipe at lightning speed, and learning and applying what they learn about each ingredient as they go.

The feedback loops built into the process let the team’s computer models work two ways: They predict characteristics of the materials, like tensile strength and stickiness, based on their composition and also suggest suitable all-natural substitutes with features designated by the user. 

Chen’s team focuses on single-use plastics like food packaging, which dominate in both production and waste streams. With AI speedily mixing and matching to meet the product’s physical requirements, some clear successes have emerged: In a project with the FDA, the team developed a food film made with ingredients that include copper and chitosan (a type of carbohydrate derived from shellfish shells) that’s transparent, compostable, moisture-regulating and even antimicrobial. In a comparison study, the team’s alternative wrap significantly outperformed standard plastic wrap at preserving cucumbers, fending off mold and shriveling for 15 days, and gave avocados a longer shelf life, too.

Of course, such biodegradable alternatives have applications well beyond wrapping veggies. Chen’s lab aims to develop materials for all-natural office supplies like file folders and badge holders, as well as nonflammable battery packaging, UV-blocking packaging and even biodegradable medical delivery systems, to name a few. “We don’t expect to get rid of all the plastic already out there,” he said, “but we hope to vastly reduce future production, and pollution, by creating replacements we know to be safe.”

Scientists are just beginning to parse out the ramifications of exposure to plastics across a lifetime, and to plastic pollution in the environment. That includes physical items invading nature (think a plastic bag wrapped around a sea turtle’s flipper) down to the tiny bits of microplastics left after nature does her best to break items down. These have scientists especially concerned.

“What worries me most is how ubiquitous microplastics are and how little we know about what that means to ecosystems and to our health,” said Christine Knauss ’14, an assistant research scientist at the University of Maryland Center for Environmental Science who studies how oysters interact with and respond to microplastics in the Chesapeake Bay.

Like Chen, Knauss is using AI to accelerate her work—in this case speeding up the identification of microplastics in water and tissue samples. It’s helping paint a picture of where the microplastics are, how they move through the water system, how likely oysters are to be exposed to these pollutants, and whether microplastics affect their growth, development, and respiration.

While the oysters are proving quite hearty in the plastic-polluted bay, Knauss said it’s an open question how human bodies will respond to the material.

“These pollutants are everywhere, in our soil, water, food, and both indoor and outdoor air; scientists have even found them in our lung tissue,” she said. “So I’m excited to see researchers like Dr. Chen being innovative with new technologies to help reduce the scope of the problem.”

Well before more concerning correlations between plastic exposure and health come to light, “developing benign alternatives is one of the most important things we can do,” added Knauss. “There’s simply too much single-use plastic being produced and entering the environment.”